diff options
156 files changed, 12168 insertions, 351 deletions
diff --git a/Documentation/ia64/xen.txt b/Documentation/ia64/xen.txt new file mode 100644 index 0000000..c61a99f --- /dev/null +++ b/Documentation/ia64/xen.txt @@ -0,0 +1,183 @@ + Recipe for getting/building/running Xen/ia64 with pv_ops + -------------------------------------------------------- + +This recipe describes how to get xen-ia64 source and build it, +and run domU with pv_ops. + +============ +Requirements +============ + + - python + - mercurial + it (aka "hg") is an open-source source code + management software. See the below. + http://www.selenic.com/mercurial/wiki/ + - git + - bridge-utils + +================================= +Getting and Building Xen and Dom0 +================================= + + My environment is; + Machine : Tiger4 + Domain0 OS : RHEL5 + DomainU OS : RHEL5 + + 1. Download source + # hg clone http://xenbits.xensource.com/ext/ia64/xen-unstable.hg + # cd xen-unstable.hg + # hg clone http://xenbits.xensource.com/ext/ia64/linux-2.6.18-xen.hg + + 2. # make world + + 3. # make install-tools + + 4. copy kernels and xen + # cp xen/xen.gz /boot/efi/efi/redhat/ + # cp build-linux-2.6.18-xen_ia64/vmlinux.gz \ + /boot/efi/efi/redhat/vmlinuz-2.6.18.8-xen + + 5. make initrd for Dom0/DomU + # make -C linux-2.6.18-xen.hg ARCH=ia64 modules_install \ + O=$(/bin/pwd)/build-linux-2.6.18-xen_ia64 + # mkinitrd -f /boot/efi/efi/redhat/initrd-2.6.18.8-xen.img \ + 2.6.18.8-xen --builtin mptspi --builtin mptbase \ + --builtin mptscsih --builtin uhci-hcd --builtin ohci-hcd \ + --builtin ehci-hcd + +================================ +Making a disk image for guest OS +================================ + + 1. make file + # dd if=/dev/zero of=/root/rhel5.img bs=1M seek=4096 count=0 + # mke2fs -F -j /root/rhel5.img + # mount -o loop /root/rhel5.img /mnt + # cp -ax /{dev,var,etc,usr,bin,sbin,lib} /mnt + # mkdir /mnt/{root,proc,sys,home,tmp} + + Note: You may miss some device files. If so, please create them + with mknod. Or you can use tar instead of cp. + + 2. modify DomU's fstab + # vi /mnt/etc/fstab + /dev/xvda1 / ext3 defaults 1 1 + none /dev/pts devpts gid=5,mode=620 0 0 + none /dev/shm tmpfs defaults 0 0 + none /proc proc defaults 0 0 + none /sys sysfs defaults 0 0 + + 3. modify inittab + set runlevel to 3 to avoid X trying to start + # vi /mnt/etc/inittab + id:3:initdefault: + Start a getty on the hvc0 console + X0:2345:respawn:/sbin/mingetty hvc0 + tty1-6 mingetty can be commented out + + 4. add hvc0 into /etc/securetty + # vi /mnt/etc/securetty (add hvc0) + + 5. umount + # umount /mnt + +FYI, virt-manager can also make a disk image for guest OS. +It's GUI tools and easy to make it. + +================== +Boot Xen & Domain0 +================== + + 1. replace elilo + elilo of RHEL5 can boot Xen and Dom0. + If you use old elilo (e.g RHEL4), please download from the below + http://elilo.sourceforge.net/cgi-bin/blosxom + and copy into /boot/efi/efi/redhat/ + # cp elilo-3.6-ia64.efi /boot/efi/efi/redhat/elilo.efi + + 2. modify elilo.conf (like the below) + # vi /boot/efi/efi/redhat/elilo.conf + prompt + timeout=20 + default=xen + relocatable + + image=vmlinuz-2.6.18.8-xen + label=xen + vmm=xen.gz + initrd=initrd-2.6.18.8-xen.img + read-only + append=" -- rhgb root=/dev/sda2" + +The append options before "--" are for xen hypervisor, +the options after "--" are for dom0. + +FYI, your machine may need console options like +"com1=19200,8n1 console=vga,com1". For example, +append="com1=19200,8n1 console=vga,com1 -- rhgb console=tty0 \ +console=ttyS0 root=/dev/sda2" + +===================================== +Getting and Building domU with pv_ops +===================================== + + 1. get pv_ops tree + # git clone http://people.valinux.co.jp/~yamahata/xen-ia64/linux-2.6-xen-ia64.git/ + + 2. git branch (if necessary) + # cd linux-2.6-xen-ia64/ + # git checkout -b your_branch origin/xen-ia64-domu-minimal-2008may19 + (Note: The current branch is xen-ia64-domu-minimal-2008may19. + But you would find the new branch. You can see with + "git branch -r" to get the branch lists. + http://people.valinux.co.jp/~yamahata/xen-ia64/for_eagl/linux-2.6-ia64-pv-ops.git/ + is also available. The tree is based on + git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6 test) + + + 3. copy .config for pv_ops of domU + # cp arch/ia64/configs/xen_domu_wip_defconfig .config + + 4. make kernel with pv_ops + # make oldconfig + # make + + 5. install the kernel and initrd + # cp vmlinux.gz /boot/efi/efi/redhat/vmlinuz-2.6-pv_ops-xenU + # make modules_install + # mkinitrd -f /boot/efi/efi/redhat/initrd-2.6-pv_ops-xenU.img \ + 2.6.26-rc3xen-ia64-08941-g1b12161 --builtin mptspi \ + --builtin mptbase --builtin mptscsih --builtin uhci-hcd \ + --builtin ohci-hcd --builtin ehci-hcd + +======================== +Boot DomainU with pv_ops +======================== + + 1. make config of DomU + # vi /etc/xen/rhel5 + kernel = "/boot/efi/efi/redhat/vmlinuz-2.6-pv_ops-xenU" + ramdisk = "/boot/efi/efi/redhat/initrd-2.6-pv_ops-xenU.img" + vcpus = 1 + memory = 512 + name = "rhel5" + disk = [ 'file:/root/rhel5.img,xvda1,w' ] + root = "/dev/xvda1 ro" + extra= "rhgb console=hvc0" + + 2. After boot xen and dom0, start xend + # /etc/init.d/xend start + ( In the debugging case, # XEND_DEBUG=1 xend trace_start ) + + 3. start domU + # xm create -c rhel5 + +========= +Reference +========= +- Wiki of Xen/IA64 upstream merge + http://wiki.xensource.com/xenwiki/XenIA64/UpstreamMerge + +Written by Akio Takebe <takebe_akio@jp.fujitsu.com> on 28 May 2008 diff --git a/MAINTAINERS b/MAINTAINERS index 5c3f79c..1d11b56 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2928,9 +2928,9 @@ S: Maintained NETEFFECT IWARP RNIC DRIVER (IW_NES) P: Faisal Latif -M: flatif@neteffect.com +M: faisal.latif@intel.com P: Chien Tung -M: ctung@neteffect.com +M: chien.tin.tung@intel.com L: general@lists.openfabrics.org W: http://www.neteffect.com S: Supported diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index df39d20..bc0fcde 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -356,7 +356,7 @@ config ARCH_IXP4XX select GENERIC_GPIO select GENERIC_TIME select GENERIC_CLOCKEVENTS - select ZONE_DMA if PCI + select DMABOUNCE if PCI help Support for Intel's IXP4XX (XScale) family of processors. diff --git a/arch/arm/common/Kconfig b/arch/arm/common/Kconfig index 2e32acc..86b5e69 100644 --- a/arch/arm/common/Kconfig +++ b/arch/arm/common/Kconfig @@ -13,10 +13,10 @@ config ICST307 config SA1111 bool select DMABOUNCE if !ARCH_PXA - select ZONE_DMA if !ARCH_PXA config DMABOUNCE bool + select ZONE_DMA config TIMER_ACORN bool diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c index fb86f24..47ccec9 100644 --- a/arch/arm/common/sa1111.c +++ b/arch/arm/common/sa1111.c @@ -581,6 +581,7 @@ sa1111_init_one_child(struct sa1111 *sachip, struct resource *parent, goto out; } +#ifdef CONFIG_DMABOUNCE /* * If the parent device has a DMA mask associated with it, * propagate it down to the children. @@ -598,6 +599,7 @@ sa1111_init_one_child(struct sa1111 *sachip, struct resource *parent, } } } +#endif out: return ret; @@ -937,7 +939,7 @@ static int sa1111_resume(struct platform_device *dev) #define sa1111_resume NULL #endif -static int sa1111_probe(struct platform_device *pdev) +static int __devinit sa1111_probe(struct platform_device *pdev) { struct resource *mem; int irq; diff --git a/arch/arm/configs/trizeps4_defconfig b/arch/arm/configs/trizeps4_defconfig index 8b7a431..9033d14 100644 --- a/arch/arm/configs/trizeps4_defconfig +++ b/arch/arm/configs/trizeps4_defconfig @@ -147,6 +147,7 @@ CONFIG_ARCH_PXA=y # CONFIG_MACH_MAINSTONE is not set # CONFIG_ARCH_PXA_IDP is not set # CONFIG_PXA_SHARPSL is not set +CONFIG_TRIZEPS_PXA=y CONFIG_MACH_TRIZEPS4=y CONFIG_MACH_TRIZEPS4_CONXS=y # CONFIG_MACH_TRIZEPS4_ANY is not set diff --git a/arch/arm/mach-clps711x/include/mach/memory.h b/arch/arm/mach-clps711x/include/mach/memory.h index 71c2fa7..98ec30c 100644 --- a/arch/arm/mach-clps711x/include/mach/memory.h +++ b/arch/arm/mach-clps711x/include/mach/memory.h @@ -89,6 +89,8 @@ * node 3: 0xd8000000 - 0xdfffffff */ #define NODE_MEM_SIZE_BITS 24 +#define SECTION_SIZE_BITS 24 +#define MAX_PHYSMEM_BITS 32 #endif diff --git a/arch/arm/mach-ixp4xx/Kconfig b/arch/arm/mach-ixp4xx/Kconfig index db8b5fe..2c5a02b 100644 --- a/arch/arm/mach-ixp4xx/Kconfig +++ b/arch/arm/mach-ixp4xx/Kconfig @@ -167,11 +167,6 @@ config MACH_GTWX5715 comment "IXP4xx Options" -config DMABOUNCE - bool - default y - depends on PCI - config IXP4XX_INDIRECT_PCI bool "Use indirect PCI memory access" depends on PCI diff --git a/arch/arm/mach-kirkwood/common.c b/arch/arm/mach-kirkwood/common.c index 85cad05..0bb1fbd 100644 --- a/arch/arm/mach-kirkwood/common.c +++ b/arch/arm/mach-kirkwood/common.c @@ -16,6 +16,7 @@ #include <linux/mv643xx_eth.h> #include <linux/ata_platform.h> #include <linux/spi/orion_spi.h> +#include <net/dsa.h> #include <asm/page.h> #include <asm/timex.h> #include <asm/mach/map.h> @@ -152,6 +153,40 @@ void __init kirkwood_ge00_init(struct mv643xx_eth_platform_data *eth_data) /***************************************************************************** + * Ethernet switch + ****************************************************************************/ +static struct resource kirkwood_switch_resources[] = { + { + .start = 0, + .end = 0, + .flags = IORESOURCE_IRQ, + }, +}; + +static struct platform_device kirkwood_switch_device = { + .name = "dsa", + .id = 0, + .num_resources = 0, + .resource = kirkwood_switch_resources, +}; + +void __init kirkwood_ge00_switch_init(struct dsa_platform_data *d, int irq) +{ + if (irq != NO_IRQ) { + kirkwood_switch_resources[0].start = irq; + kirkwood_switch_resources[0].end = irq; + kirkwood_switch_device.num_resources = 1; + } + + d->mii_bus = &kirkwood_ge00_shared.dev; + d->netdev = &kirkwood_ge00.dev; + kirkwood_switch_device.dev.platform_data = d; + + platform_device_register(&kirkwood_switch_device); +} + + +/***************************************************************************** * SoC RTC ****************************************************************************/ static struct resource kirkwood_rtc_resource = { diff --git a/arch/arm/mach-kirkwood/common.h b/arch/arm/mach-kirkwood/common.h index 8fa0f6a..5774632 100644 --- a/arch/arm/mach-kirkwood/common.h +++ b/arch/arm/mach-kirkwood/common.h @@ -11,6 +11,7 @@ #ifndef __ARCH_KIRKWOOD_COMMON_H #define __ARCH_KIRKWOOD_COMMON_H +struct dsa_platform_data; struct mv643xx_eth_platform_data; struct mv_sata_platform_data; @@ -29,6 +30,7 @@ void kirkwood_pcie_id(u32 *dev, u32 *rev); void kirkwood_ehci_init(void); void kirkwood_ge00_init(struct mv643xx_eth_platform_data *eth_data); +void kirkwood_ge00_switch_init(struct dsa_platform_data *d, int irq); void kirkwood_pcie_init(void); void kirkwood_rtc_init(void); void kirkwood_sata_init(struct mv_sata_platform_data *sata_data); diff --git a/arch/arm/mach-kirkwood/rd88f6281-setup.c b/arch/arm/mach-kirkwood/rd88f6281-setup.c index f785093..175054a 100644 --- a/arch/arm/mach-kirkwood/rd88f6281-setup.c +++ b/arch/arm/mach-kirkwood/rd88f6281-setup.c @@ -19,6 +19,7 @@ #include <linux/ata_platform.h> #include <linux/mv643xx_eth.h> #include <linux/ethtool.h> +#include <net/dsa.h> #include <asm/mach-types.h> #include <asm/mach/arch.h> #include <asm/mach/pci.h> @@ -74,6 +75,15 @@ static struct mv643xx_eth_platform_data rd88f6281_ge00_data = { .duplex = DUPLEX_FULL, }; +static struct dsa_platform_data rd88f6281_switch_data = { + .port_names[0] = "lan1", + .port_names[1] = "lan2", + .port_names[2] = "lan3", + .port_names[3] = "lan4", + .port_names[4] = "wan", + .port_names[5] = "cpu", +}; + static struct mv_sata_platform_data rd88f6281_sata_data = { .n_ports = 2, }; @@ -87,6 +97,7 @@ static void __init rd88f6281_init(void) kirkwood_ehci_init(); kirkwood_ge00_init(&rd88f6281_ge00_data); + kirkwood_ge00_switch_init(&rd88f6281_switch_data, NO_IRQ); kirkwood_rtc_init(); kirkwood_sata_init(&rd88f6281_sata_data); kirkwood_uart0_init(); diff --git a/arch/arm/mach-mv78xx0/db78x00-bp-setup.c b/arch/arm/mach-mv78xx0/db78x00-bp-setup.c index 49f434c..2e285bb 100644 --- a/arch/arm/mach-mv78xx0/db78x00-bp-setup.c +++ b/arch/arm/mach-mv78xx0/db78x00-bp-setup.c @@ -13,6 +13,7 @@ #include <linux/platform_device.h> #include <linux/ata_platform.h> #include <linux/mv643xx_eth.h> +#include <linux/ethtool.h> #include <mach/mv78xx0.h> #include <asm/mach-types.h> #include <asm/mach/arch.h> @@ -28,10 +29,14 @@ static struct mv643xx_eth_platform_data db78x00_ge01_data = { static struct mv643xx_eth_platform_data db78x00_ge10_data = { .phy_addr = MV643XX_ETH_PHY_NONE, + .speed = SPEED_1000, + .duplex = DUPLEX_FULL, }; static struct mv643xx_eth_platform_data db78x00_ge11_data = { .phy_addr = MV643XX_ETH_PHY_NONE, + .speed = SPEED_1000, + .duplex = DUPLEX_FULL, }; static struct mv_sata_platform_data db78x00_sata_data = { diff --git a/arch/arm/mach-orion5x/common.c b/arch/arm/mach-orion5x/common.c index 9625ef5..437065c 100644 --- a/arch/arm/mach-orion5x/common.c +++ b/arch/arm/mach-orion5x/common.c @@ -19,6 +19,7 @@ #include <linux/mv643xx_i2c.h> #include <linux/ata_platform.h> #include <linux/spi/orion_spi.h> +#include <net/dsa.h> #include <asm/page.h> #include <asm/setup.h> #include <asm/timex.h> @@ -198,6 +199,40 @@ void __init orion5x_eth_init(struct mv643xx_eth_platform_data *eth_data) /***************************************************************************** + * Ethernet switch + ****************************************************************************/ +static struct resource orion5x_switch_resources[] = { + { + .start = 0, + .end = 0, + .flags = IORESOURCE_IRQ, + }, +}; + +static struct platform_device orion5x_switch_device = { + .name = "dsa", + .id = 0, + .num_resources = 0, + .resource = orion5x_switch_resources, +}; + +void __init orion5x_eth_switch_init(struct dsa_platform_data *d, int irq) +{ + if (irq != NO_IRQ) { + orion5x_switch_resources[0].start = irq; + orion5x_switch_resources[0].end = irq; + orion5x_switch_device.num_resources = 1; + } + + d->mii_bus = &orion5x_eth_shared.dev; + d->netdev = &orion5x_eth.dev; + orion5x_switch_device.dev.platform_data = d; + + platform_device_register(&orion5x_switch_device); +} + + +/***************************************************************************** * I2C ****************************************************************************/ static struct mv64xxx_i2c_pdata orion5x_i2c_pdata = { @@ -275,7 +310,8 @@ void __init orion5x_sata_init(struct mv_sata_platform_data *sata_data) * SPI ****************************************************************************/ static struct orion_spi_info orion5x_spi_plat_data = { - .tclk = 0, + .tclk = 0, + .enable_clock_fix = 1, }; static struct resource orion5x_spi_resources[] = { diff --git a/arch/arm/mach-orion5x/common.h b/arch/arm/mach-orion5x/common.h index 1f8b2da..a000c7c 100644 --- a/arch/arm/mach-orion5x/common.h +++ b/arch/arm/mach-orion5x/common.h @@ -1,6 +1,7 @@ #ifndef __ARCH_ORION5X_COMMON_H #define __ARCH_ORION5X_COMMON_H +struct dsa_platform_data; struct mv643xx_eth_platform_data; struct mv_sata_platform_data; @@ -29,6 +30,7 @@ void orion5x_setup_pcie_wa_win(u32 base, u32 size); void orion5x_ehci0_init(void); void orion5x_ehci1_init(void); void orion5x_eth_init(struct mv643xx_eth_platform_data *eth_data); +void orion5x_eth_switch_init(struct dsa_platform_data *d, int irq); void orion5x_i2c_init(void); void orion5x_sata_init(struct mv_sata_platform_data *sata_data); void orion5x_spi_init(void); diff --git a/arch/arm/mach-orion5x/rd88f5181l-fxo-setup.c b/arch/arm/mach-orion5x/rd88f5181l-fxo-setup.c index 500cdada..15f5323 100644 --- a/arch/arm/mach-orion5x/rd88f5181l-fxo-setup.c +++ b/arch/arm/mach-orion5x/rd88f5181l-fxo-setup.c @@ -16,6 +16,7 @@ #include <linux/mtd/physmap.h> #include <linux/mv643xx_eth.h> #include <linux/ethtool.h> +#include <net/dsa.h> #include <asm/mach-types.h> #include <asm/gpio.h> #include <asm/leds.h> @@ -93,6 +94,15 @@ static struct mv643xx_eth_platform_data rd88f5181l_fxo_eth_data = { .duplex = DUPLEX_FULL, }; +static struct dsa_platform_data rd88f5181l_fxo_switch_data = { + .port_names[0] = "lan2", + .port_names[1] = "lan1", + .port_names[2] = "wan", + .port_names[3] = "cpu", + .port_names[5] = "lan4", + .port_names[7] = "lan3", +}; + static void __init rd88f5181l_fxo_init(void) { /* @@ -107,6 +117,7 @@ static void __init rd88f5181l_fxo_init(void) */ orion5x_ehci0_init(); orion5x_eth_init(&rd88f5181l_fxo_eth_data); + orion5x_eth_switch_init(&rd88f5181l_fxo_switch_data, NO_IRQ); orion5x_uart0_init(); orion5x_setup_dev_boot_win(RD88F5181L_FXO_NOR_BOOT_BASE, diff --git a/arch/arm/mach-orion5x/rd88f5181l-ge-setup.c b/arch/arm/mach-orion5x/rd88f5181l-ge-setup.c index ebde814..8ad3934 100644 --- a/arch/arm/mach-orion5x/rd88f5181l-ge-setup.c +++ b/arch/arm/mach-orion5x/rd88f5181l-ge-setup.c @@ -17,6 +17,7 @@ #include <linux/mv643xx_eth.h> #include <linux/ethtool.h> #include <linux/i2c.h> +#include <net/dsa.h> #include <asm/mach-types.h> #include <asm/gpio.h> #include <asm/leds.h> @@ -94,6 +95,15 @@ static struct mv643xx_eth_platform_data rd88f5181l_ge_eth_data = { .duplex = DUPLEX_FULL, }; +static struct dsa_platform_data rd88f5181l_ge_switch_data = { + .port_names[0] = "lan2", + .port_names[1] = "lan1", + .port_names[2] = "wan", + .port_names[3] = "cpu", + .port_names[5] = "lan4", + .port_names[7] = "lan3", +}; + static struct i2c_board_info __initdata rd88f5181l_ge_i2c_rtc = { I2C_BOARD_INFO("ds1338", 0x68), }; @@ -112,6 +122,7 @@ static void __init rd88f5181l_ge_init(void) */ orion5x_ehci0_init(); orion5x_eth_init(&rd88f5181l_ge_eth_data); + orion5x_eth_switch_init(&rd88f5181l_ge_switch_data, gpio_to_irq(8)); orion5x_i2c_init(); orion5x_uart0_init(); diff --git a/arch/arm/mach-orion5x/rd88f6183ap-ge-setup.c b/arch/arm/mach-orion5x/rd88f6183ap-ge-setup.c index 40e0495..262e25e 100644 --- a/arch/arm/mach-orion5x/rd88f6183ap-ge-setup.c +++ b/arch/arm/mach-orion5x/rd88f6183ap-ge-setup.c @@ -19,6 +19,7 @@ #include <linux/spi/orion_spi.h> #include <linux/spi/flash.h> #include <linux/ethtool.h> +#include <net/dsa.h> #include <asm/mach-types.h> #include <asm/gpio.h> #include <asm/leds.h> @@ -34,6 +35,15 @@ static struct mv643xx_eth_platform_data rd88f6183ap_ge_eth_data = { .duplex = DUPLEX_FULL, }; +static struct dsa_platform_data rd88f6183ap_ge_switch_data = { + .port_names[0] = "lan1", + .port_names[1] = "lan2", + .port_names[2] = "lan3", + .port_names[3] = "lan4", + .port_names[4] = "wan", + .port_names[5] = "cpu", +}; + static struct mtd_partition rd88f6183ap_ge_partitions[] = { { .name = "kernel", @@ -79,6 +89,7 @@ static void __init rd88f6183ap_ge_init(void) */ orion5x_ehci0_init(); orion5x_eth_init(&rd88f6183ap_ge_eth_data); + orion5x_eth_switch_init(&rd88f6183ap_ge_switch_data, gpio_to_irq(3)); spi_register_board_info(rd88f6183ap_ge_spi_slave_info, ARRAY_SIZE(rd88f6183ap_ge_spi_slave_info)); orion5x_spi_init(); diff --git a/arch/arm/mach-orion5x/wrt350n-v2-setup.c b/arch/arm/mach-orion5x/wrt350n-v2-setup.c index 9a4fd52..cc8f892 100644 --- a/arch/arm/mach-orion5x/wrt350n-v2-setup.c +++ b/arch/arm/mach-orion5x/wrt350n-v2-setup.c @@ -15,6 +15,7 @@ #include <linux/mtd/physmap.h> #include <linux/mv643xx_eth.h> #include <linux/ethtool.h> +#include <net/dsa.h> #include <asm/mach-types.h> #include <asm/gpio.h> #include <asm/mach/arch.h> @@ -105,6 +106,15 @@ static struct mv643xx_eth_platform_data wrt350n_v2_eth_data = { .duplex = DUPLEX_FULL, }; +static struct dsa_platform_data wrt350n_v2_switch_data = { + .port_names[0] = "lan2", + .port_names[1] = "lan1", + .port_names[2] = "wan", + .port_names[3] = "cpu", + .port_names[5] = "lan3", + .port_names[7] = "lan4", +}; + static void __init wrt350n_v2_init(void) { /* @@ -119,6 +129,7 @@ static void __init wrt350n_v2_init(void) */ orion5x_ehci0_init(); orion5x_eth_init(&wrt350n_v2_eth_data); + orion5x_eth_switch_init(&wrt350n_v2_switch_data, NO_IRQ); orion5x_uart0_init(); orion5x_setup_dev_boot_win(WRT350N_V2_NOR_BOOT_BASE, diff --git a/arch/arm/mach-pxa/Kconfig b/arch/arm/mach-pxa/Kconfig index f27f6b3..f781873 100644 --- a/arch/arm/mach-pxa/Kconfig +++ b/arch/arm/mach-pxa/Kconfig @@ -257,7 +257,6 @@ config MACH_ARMCORE bool "CompuLab CM-X255/CM-X270 modules" select PXA27x select IWMMXT - select ZONE_DMA if PCI select PXA25x select PXA_SSP diff --git a/arch/arm/mach-pxa/include/mach/irqs.h b/arch/arm/mach-pxa/include/mach/irqs.h index 9c163e1..32bb4a2 100644 --- a/arch/arm/mach-pxa/include/mach/irqs.h +++ b/arch/arm/mach-pxa/include/mach/irqs.h @@ -9,7 +9,8 @@ * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. */ - +#ifndef __ASM_MACH_IRQS_H +#define __ASM_MACH_IRQS_H #ifdef CONFIG_PXA_HAVE_ISA_IRQS #define PXA_ISA_IRQ(x) (x) @@ -264,3 +265,5 @@ #endif #endif /* CONFIG_PCI_HOST_ITE8152 */ + +#endif /* __ASM_MACH_IRQS_H */ diff --git a/arch/arm/mach-pxa/include/mach/spitz.h b/arch/arm/mach-pxa/include/mach/spitz.h index 31ac26b..e8488df 100644 --- a/arch/arm/mach-pxa/include/mach/spitz.h +++ b/arch/arm/mach-pxa/include/mach/spitz.h @@ -142,7 +142,7 @@ #define SPITZ_SCP2_GPIO_BASE (NR_BUILTIN_GPIO + 12) #define SPITZ_GPIO_IR_ON (SPITZ_SCP2_GPIO_BASE + 0) -#define SPITZ_GPIO_AKIN_PULLUP (SPITZ_SCP2_GPIO_BASE + 1 +#define SPITZ_GPIO_AKIN_PULLUP (SPITZ_SCP2_GPIO_BASE + 1) #define SPITZ_GPIO_RESERVED_1 (SPITZ_SCP2_GPIO_BASE + 2) #define SPITZ_GPIO_RESERVED_2 (SPITZ_SCP2_GPIO_BASE + 3) #define SPITZ_GPIO_RESERVED_3 (SPITZ_SCP2_GPIO_BASE + 4) diff --git a/arch/arm/mach-pxa/pwm.c b/arch/arm/mach-pxa/pwm.c index 316cd98..74e2ead 100644 --- a/arch/arm/mach-pxa/pwm.c +++ b/arch/arm/mach-pxa/pwm.c @@ -60,7 +60,7 @@ int pwm_config(struct pwm_device *pwm, int duty_ns, int period_ns) do_div(c, 1000000000); period_cycles = c; - if (period_cycles < 0) + if (period_cycles < 1) period_cycles = 1; prescale = (period_cycles - 1) / 1024; pv = period_cycles / (prescale + 1) - 1; diff --git a/arch/arm/mach-pxa/trizeps4.c b/arch/arm/mach-pxa/trizeps4.c index a13dbf3..a72e3ad 100644 --- a/arch/arm/mach-pxa/trizeps4.c +++ b/arch/arm/mach-pxa/trizeps4.c @@ -399,7 +399,7 @@ static void trizeps4_irda_transceiver_mode(struct device *dev, int mode) /* Switch mode */ if (mode & IR_SIRMODE) trizeps_conxs_ircr &= ~ConXS_IRCR_MODE; /* Slow mode */ - else if (mode & IR_FIRMODE) { + else if (mode & IR_FIRMODE) trizeps_conxs_ircr |= ConXS_IRCR_MODE; /* Fast mode */ /* Switch power */ diff --git a/arch/arm/mach-s3c2443/clock.c b/arch/arm/mach-s3c2443/clock.c index 2f60bf6..f854e73 100644 --- a/arch/arm/mach-s3c2443/clock.c +++ b/arch/arm/mach-s3c2443/clock.c @@ -1033,8 +1033,7 @@ void __init s3c2443_init_clocks(int xtal) fclk = pll / s3c2443_fclk_div(clkdiv0); hclk = s3c2443_prediv_getrate(&clk_prediv); - hclk = hclk / s3c2443_get_hdiv(clkdiv0); - hclk = hclk / ((clkdiv0 & S3C2443_CLKDIV0_HALF_HCLK) ? 2 : 1); + hclk /= s3c2443_get_hdiv(clkdiv0); pclk = hclk / ((clkdiv0 & S3C2443_CLKDIV0_HALF_PCLK) ? 2 : 1); s3c24xx_setup_clocks(xtal, fclk, hclk, pclk); diff --git a/arch/arm/mm/cache-v4.S b/arch/arm/mm/cache-v4.S index 33926c9..5786adf1 100644 --- a/arch/arm/mm/cache-v4.S +++ b/arch/arm/mm/cache-v4.S @@ -29,7 +29,7 @@ ENTRY(v4_flush_user_cache_all) * Clean and invalidate the entire cache. */ ENTRY(v4_flush_kern_cache_all) -#ifdef CPU_CP15 +#ifdef CONFIG_CPU_CP15 mov r0, #0 mcr p15, 0, r0, c7, c7, 0 @ flush ID cache mov pc, lr @@ -48,7 +48,7 @@ ENTRY(v4_flush_kern_cache_all) * - flags - vma_area_struct flags describing address space */ ENTRY(v4_flush_user_cache_range) -#ifdef CPU_CP15 +#ifdef CONFIG_CPU_CP15 mov ip, #0 mcreq p15, 0, ip, c7, c7, 0 @ flush ID cache mov pc, lr @@ -116,7 +116,7 @@ ENTRY(v4_dma_inv_range) * - end - virtual end address */ ENTRY(v4_dma_flush_range) -#ifdef CPU_CP15 +#ifdef CONFIG_CPU_CP15 mov r0, #0 mcr p15, 0, r0, c7, c7, 0 @ flush ID cache #endif diff --git a/arch/arm/plat-s3c24xx/pwm-clock.c b/arch/arm/plat-s3c24xx/pwm-clock.c index b8e854f..3fad68a 100644 --- a/arch/arm/plat-s3c24xx/pwm-clock.c +++ b/arch/arm/plat-s3c24xx/pwm-clock.c @@ -315,7 +315,7 @@ static int clk_pwm_tin_set_parent(struct clk *clk, struct clk *parent) if (parent == s3c24xx_pwmclk_tclk(id)) bits = S3C2410_TCFG1_MUX_TCLK << shift; else if (parent == s3c24xx_pwmclk_tdiv(id)) - bits = clk_pwm_tdiv_bits(to_tdiv(clk)) << shift; + bits = clk_pwm_tdiv_bits(to_tdiv(parent)) << shift; else return -EINVAL; diff --git a/arch/arm/plat-s3c24xx/pwm.c b/arch/arm/plat-s3c24xx/pwm.c index feb770f..ec56b88 100644 --- a/arch/arm/plat-s3c24xx/pwm.c +++ b/arch/arm/plat-s3c24xx/pwm.c @@ -56,7 +56,7 @@ static struct clk *clk_scaler[2]; } \ } -#define DEFINE_TIMER(_tmr_no, _irq) \ +#define DEFINE_S3C_TIMER(_tmr_no, _irq) \ .name = "s3c24xx-pwm", \ .id = _tmr_no, \ .num_resources = TIMER_RESOURCE_SIZE, \ @@ -67,11 +67,11 @@ static struct clk *clk_scaler[2]; */ struct platform_device s3c_device_timer[] = { - [0] = { DEFINE_TIMER(0, IRQ_TIMER0) }, - [1] = { DEFINE_TIMER(1, IRQ_TIMER1) }, - [2] = { DEFINE_TIMER(2, IRQ_TIMER2) }, - [3] = { DEFINE_TIMER(3, IRQ_TIMER3) }, - [4] = { DEFINE_TIMER(4, IRQ_TIMER4) }, + [0] = { DEFINE_S3C_TIMER(0, IRQ_TIMER0) }, + [1] = { DEFINE_S3C_TIMER(1, IRQ_TIMER1) }, + [2] = { DEFINE_S3C_TIMER(2, IRQ_TIMER2) }, + [3] = { DEFINE_S3C_TIMER(3, IRQ_TIMER3) }, + [4] = { DEFINE_S3C_TIMER(4, IRQ_TIMER4) }, }; static inline int pwm_is_tdiv(struct pwm_device *pwm) diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 912c57d..27eec714 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -23,6 +23,7 @@ config IA64 select HAVE_KRETPROBES select HAVE_DMA_ATTRS select HAVE_KVM + select HAVE_ARCH_TRACEHOOK default y help The Itanium Processor Family is Intel's 64-bit successor to @@ -110,6 +111,33 @@ config AUDIT_ARCH bool default y +menuconfig PARAVIRT_GUEST + bool "Paravirtualized guest support" + help + Say Y here to get to see options related to running Linux under + various hypervisors. This option alone does not add any kernel code. + + If you say N, all options in this submenu will be skipped and disabled. + +if PARAVIRT_GUEST + +config PARAVIRT + bool "Enable paravirtualization code" + depends on PARAVIRT_GUEST + default y + bool + default y + help + This changes the kernel so it can modify itself when it is run + under a hypervisor, potentially improving performance significantly + over full virtualization. However, when run without a hypervisor + the kernel is theoretically slower and slightly larger. + + +source "arch/ia64/xen/Kconfig" + +endif + choice prompt "System type" default IA64_GENERIC @@ -119,6 +147,7 @@ config IA64_GENERIC select NUMA select ACPI_NUMA select SWIOTLB + select PCI_MSI help This selects the system type of your hardware. A "generic" kernel will run on any supported IA-64 system. However, if you configure @@ -126,11 +155,13 @@ config IA64_GENERIC generic For any supported IA-64 system DIG-compliant For DIG ("Developer's Interface Guide") compliant systems + DIG+Intel+IOMMU For DIG systems with Intel IOMMU HP-zx1/sx1000 For HP systems HP-zx1/sx1000+swiotlb For HP systems with (broken) DMA-constrained devices. SGI-SN2 For SGI Altix systems SGI-UV For SGI UV systems Ski-simulator For the HP simulator <http://www.hpl.hp.com/research/linux/ski/> + Xen-domU For xen domU system If you don't know what to do, choose "generic". @@ -138,6 +169,11 @@ config IA64_DIG bool "DIG-compliant" select SWIOTLB +config IA64_DIG_VTD + bool "DIG+Intel+IOMMU" + select DMAR + select PCI_MSI + config IA64_HP_ZX1 bool "HP-zx1/sx1000" help @@ -181,6 +217,10 @@ config IA64_HP_SIM bool "Ski-simulator" select SWIOTLB +config IA64_XEN_GUEST + bool "Xen guest" + depends on XEN + endchoice choice @@ -583,6 +623,16 @@ source "drivers/pci/hotplug/Kconfig" source "drivers/pcmcia/Kconfig" +config DMAR + bool "Support for DMA Remapping Devices (EXPERIMENTAL)" + depends on IA64_GENERIC && ACPI && EXPERIMENTAL + help + DMA remapping (DMAR) devices support enables independent address + translations for Direct Memory Access (DMA) from devices. + These DMA remapping devices are reported via ACPI tables + and include PCI device scope covered by these DMA + remapping devices. + endmenu endif diff --git a/arch/ia64/Makefile b/arch/ia64/Makefile index 905d25b..58a7e46a 100644 --- a/arch/ia64/Makefile +++ b/arch/ia64/Makefile @@ -53,12 +53,15 @@ libs-y += arch/ia64/lib/ core-y += arch/ia64/kernel/ arch/ia64/mm/ core-$(CONFIG_IA32_SUPPORT) += arch/ia64/ia32/ core-$(CONFIG_IA64_DIG) += arch/ia64/dig/ +core-$(CONFIG_IA64_DIG_VTD) += arch/ia64/dig/ core-$(CONFIG_IA64_GENERIC) += arch/ia64/dig/ core-$(CONFIG_IA64_HP_ZX1) += arch/ia64/dig/ core-$(CONFIG_IA64_HP_ZX1_SWIOTLB) += arch/ia64/dig/ +core-$(CONFIG_IA64_XEN_GUEST) += arch/ia64/dig/ core-$(CONFIG_IA64_SGI_SN2) += arch/ia64/sn/ core-$(CONFIG_IA64_SGI_UV) += arch/ia64/uv/ core-$(CONFIG_KVM) += arch/ia64/kvm/ +core-$(CONFIG_XEN) += arch/ia64/xen/ drivers-$(CONFIG_PCI) += arch/ia64/pci/ drivers-$(CONFIG_IA64_HP_SIM) += arch/ia64/hp/sim/ diff --git a/arch/ia64/configs/generic_defconfig b/arch/ia64/configs/generic_defconfig index 9f48397..e05f9e1 100644 --- a/arch/ia64/configs/generic_defconfig +++ b/arch/ia64/configs/generic_defconfig @@ -233,6 +233,8 @@ CONFIG_DMIID=y CONFIG_BINFMT_ELF=y CONFIG_BINFMT_MISC=m +# CONFIG_DMAR is not set + # # Power management and ACPI # diff --git a/arch/ia64/configs/tiger_defconfig b/arch/ia64/configs/tiger_defconfig index 797acf9..c522edf 100644 --- a/arch/ia64/configs/tiger_defconfig +++ b/arch/ia64/configs/tiger_defconfig @@ -172,6 +172,8 @@ CONFIG_DMIID=y CONFIG_BINFMT_ELF=y CONFIG_BINFMT_MISC=m +# CONFIG_DMAR is not set + # # Power management and ACPI # diff --git a/arch/ia64/dig/Makefile b/arch/ia64/dig/Makefile index 971cd78..5c02838 100644 --- a/arch/ia64/dig/Makefile +++ b/arch/ia64/dig/Makefile @@ -6,4 +6,9 @@ # obj-y := setup.o +ifeq ($(CONFIG_DMAR), y) +obj-$(CONFIG_IA64_GENERIC) += machvec.o machvec_vtd.o dig_vtd_iommu.o +else obj-$(CONFIG_IA64_GENERIC) += machvec.o +endif +obj-$(CONFIG_IA64_DIG_VTD) += dig_vtd_iommu.o diff --git a/arch/ia64/dig/dig_vtd_iommu.c b/arch/ia64/dig/dig_vtd_iommu.c new file mode 100644 index 0000000..1c8a079 --- /dev/null +++ b/arch/ia64/dig/dig_vtd_iommu.c @@ -0,0 +1,59 @@ +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/intel-iommu.h> + +void * +vtd_alloc_coherent(struct device *dev, size_t size, dma_addr_t *dma_handle, + gfp_t flags) +{ + return intel_alloc_coherent(dev, size, dma_handle, flags); +} +EXPORT_SYMBOL_GPL(vtd_alloc_coherent); + +void +vtd_free_coherent(struct device *dev, size_t size, void *vaddr, + dma_addr_t dma_handle) +{ + intel_free_coherent(dev, size, vaddr, dma_handle); +} +EXPORT_SYMBOL_GPL(vtd_free_coherent); + +dma_addr_t +vtd_map_single_attrs(struct device *dev, void *addr, size_t size, + int dir, struct dma_attrs *attrs) +{ + return intel_map_single(dev, (phys_addr_t)addr, size, dir); +} +EXPORT_SYMBOL_GPL(vtd_map_single_attrs); + +void +vtd_unmap_single_attrs(struct device *dev, dma_addr_t iova, size_t size, + int dir, struct dma_attrs *attrs) +{ + intel_unmap_single(dev, iova, size, dir); +} +EXPORT_SYMBOL_GPL(vtd_unmap_single_attrs); + +int +vtd_map_sg_attrs(struct device *dev, struct scatterlist *sglist, int nents, + int dir, struct dma_attrs *attrs) +{ + return intel_map_sg(dev, sglist, nents, dir); +} +EXPORT_SYMBOL_GPL(vtd_map_sg_attrs); + +void +vtd_unmap_sg_attrs(struct device *dev, struct scatterlist *sglist, + int nents, int dir, struct dma_attrs *attrs) +{ + intel_unmap_sg(dev, sglist, nents, dir); +} +EXPORT_SYMBOL_GPL(vtd_unmap_sg_attrs); + +int +vtd_dma_mapping_error(struct device *dev, dma_addr_t dma_addr) +{ + return 0; +} +EXPORT_SYMBOL_GPL(vtd_dma_mapping_error); diff --git a/arch/ia64/dig/machvec_vtd.c b/arch/ia64/dig/machvec_vtd.c new file mode 100644 index 0000000..7cd3eb4 --- /dev/null +++ b/arch/ia64/dig/machvec_vtd.c @@ -0,0 +1,3 @@ +#define MACHVEC_PLATFORM_NAME dig_vtd +#define MACHVEC_PLATFORM_HEADER <asm/machvec_dig_vtd.h> +#include <asm/machvec_init.h> diff --git a/arch/ia64/ia32/ia32_entry.S b/arch/ia64/ia32/ia32_entry.S index 53505bb..a8cf199 100644 --- a/arch/ia64/ia32/ia32_entry.S +++ b/arch/ia64/ia32/ia32_entry.S @@ -108,6 +108,11 @@ GLOBAL_ENTRY(ia32_trace_syscall) ;; st8 [r2]=r3 // initialize return code to -ENOSYS br.call.sptk.few rp=syscall_trace_enter // give parent a chance to catch syscall args + cmp.lt p6,p0=r8,r0 // check tracehook + adds r2=IA64_PT_REGS_R8_OFFSET+16,sp // r2 = &pt_regs.r8 + ;; +(p6) st8.spill [r2]=r8 // store return value in slot for r8 +(p6) br.spnt.few .ret4 .ret2: // Need to reload arguments (they may be changed by the tracing process) adds r2=IA64_PT_REGS_R1_OFFSET+16,sp // r2 = &pt_regs.r1 adds r3=IA64_PT_REGS_R13_OFFSET+16,sp // r3 = &pt_regs.r13 @@ -199,10 +204,10 @@ ia32_syscall_table: data8 sys_setuid /* 16-bit version */ data8 sys_getuid /* 16-bit version */ data8 compat_sys_stime /* 25 */ - data8 sys32_ptrace + data8 compat_sys_ptrace data8 sys32_alarm data8 sys_ni_syscall - data8 sys32_pause + data8 sys_pause data8 compat_sys_utime /* 30 */ data8 sys_ni_syscall /* old stty syscall holder */ data8 sys_ni_syscall /* old gtty syscall holder */ @@ -215,7 +220,7 @@ ia32_syscall_table: data8 sys_mkdir data8 sys_rmdir /* 40 */ data8 sys_dup - data8 sys32_pipe + data8 sys_pipe data8 compat_sys_times data8 sys_ni_syscall /* old prof syscall holder */ data8 sys32_brk /* 45 */ diff --git a/arch/ia64/ia32/sys_ia32.c b/arch/ia64/ia32/sys_ia32.c index f4430bb..5e92ae0 100644 --- a/arch/ia64/ia32/sys_ia32.c +++ b/arch/ia64/ia32/sys_ia32.c @@ -1098,21 +1098,6 @@ sys32_mremap (unsigned int addr, unsigned int old_len, unsigned int new_len, return ret; } -asmlinkage long -sys32_pipe (int __user *fd) -{ - int retval; - int fds[2]; - - retval = do_pipe_flags(fds, 0); - if (retval) - goto out; - if (copy_to_user(fd, fds, sizeof(fds))) - retval = -EFAULT; - out: - return retval; -} - asmlinkage unsigned long sys32_alarm (unsigned int seconds) { @@ -1209,25 +1194,6 @@ sys32_waitpid (int pid, unsigned int *stat_addr, int options) return compat_sys_wait4(pid, stat_addr, options, NULL); } -static unsigned int -ia32_peek (struct task_struct *child, unsigned long addr, unsigned int *val) -{ - size_t copied; - unsigned int ret; - - copied = access_process_vm(child, addr, val, sizeof(*val), 0); - return (copied != sizeof(ret)) ? -EIO : 0; -} - -static unsigned int -ia32_poke (struct task_struct *child, unsigned long addr, unsigned int val) -{ - - if (access_process_vm(child, addr, &val, sizeof(val), 1) != sizeof(val)) - return -EIO; - return 0; -} - /* * The order in which registers are stored in the ptrace regs structure */ @@ -1525,49 +1491,15 @@ restore_ia32_fpxstate (struct task_struct *tsk, struct ia32_user_fxsr_struct __u return 0; } -asmlinkage long -sys32_ptrace (int request, pid_t pid, unsigned int addr, unsigned int data) +long compat_arch_ptrace(struct task_struct *child, compat_long_t request, + compat_ulong_t caddr, compat_ulong_t cdata) { - struct task_struct *child; - unsigned int value, tmp; + unsigned long addr = caddr; + unsigned long data = cdata; + unsigned int tmp; long i, ret; - lock_kernel(); - if (request == PTRACE_TRACEME) { - ret = ptrace_traceme(); - goto out; - } - - child = ptrace_get_task_struct(pid); - if (IS_ERR(child)) { - ret = PTR_ERR(child); - goto out; - } - - if (request == PTRACE_ATTACH) { - ret = sys_ptrace(request, pid, addr, data); - goto out_tsk; - } - - ret = ptrace_check_attach(child, request == PTRACE_KILL); - if (ret < 0) - goto out_tsk; - switch (request) { - case PTRACE_PEEKTEXT: - case PTRACE_PEEKDATA: /* read word at location addr */ - ret = ia32_peek(child, addr, &value); - if (ret == 0) - ret = put_user(value, (unsigned int __user *) compat_ptr(data)); - else - ret = -EIO; - goto out_tsk; - - case PTRACE_POKETEXT: - case PTRACE_POKEDATA: /* write the word at location addr */ - ret = ia32_poke(child, addr, data); - goto out_tsk; - case PTRACE_PEEKUSR: /* read word at addr in USER area */ ret = -EIO; if ((addr & 3) || addr > 17*sizeof(int)) @@ -1632,27 +1564,9 @@ sys32_ptrace (int request, pid_t pid, unsigned int addr, unsigned int data) compat_ptr(data)); break; - case PTRACE_GETEVENTMSG: - ret = put_user(child->ptrace_message, (unsigned int __user *) compat_ptr(data)); - break; - - case PTRACE_SYSCALL: /* continue, stop after next syscall */ - case PTRACE_CONT: /* restart after signal. */ - case PTRACE_KILL: - case PTRACE_SINGLESTEP: /* execute chile for one instruction */ - case PTRACE_DETACH: /* detach a process */ - ret = sys_ptrace(request, pid, addr, data); - break; - default: - ret = ptrace_request(child, request, addr, data); - break; - + return compat_ptrace_request(child, request, caddr, cdata); } - out_tsk: - put_task_struct(child); - out: - unlock_kernel(); return ret; } @@ -1704,14 +1618,6 @@ out: } asmlinkage int -sys32_pause (void) -{ - current->state = TASK_INTERRUPTIBLE; - schedule(); - return -ERESTARTNOHAND; -} - -asmlinkage int sys32_msync (unsigned int start, unsigned int len, int flags) { unsigned int addr; diff --git a/arch/ia64/include/asm/break.h b/arch/ia64/include/asm/break.h index f034020..e90c40e 100644 --- a/arch/ia64/include/asm/break.h +++ b/arch/ia64/include/asm/break.h @@ -20,4 +20,13 @@ */ #define __IA64_BREAK_SYSCALL 0x100000 +/* + * Xen specific break numbers: + */ +#define __IA64_XEN_HYPERCALL 0x1000 +/* [__IA64_XEN_HYPERPRIVOP_START, __IA64_XEN_HYPERPRIVOP_MAX] is used + for xen hyperprivops */ +#define __IA64_XEN_HYPERPRIVOP_START 0x1 +#define __IA64_XEN_HYPERPRIVOP_MAX 0x1a + #endif /* _ASM_IA64_BREAK_H */ diff --git a/arch/ia64/include/asm/cacheflush.h b/arch/ia64/include/asm/cacheflush.h index afcfbda..c8ce271 100644 --- a/arch/ia64/include/asm/cacheflush.h +++ b/arch/ia64/include/asm/cacheflush.h @@ -34,6 +34,8 @@ do { \ #define flush_dcache_mmap_unlock(mapping) do { } while (0) extern void flush_icache_range (unsigned long start, unsigned long end); +extern void clflush_cache_range(void *addr, int size); + #define flush_icache_user_range(vma, page, user_addr, len) \ do { \ diff --git a/arch/ia64/include/asm/device.h b/arch/ia64/include/asm/device.h index 3db6daf..41ab85d 100644 --- a/arch/ia64/include/asm/device.h +++ b/arch/ia64/include/asm/device.h @@ -10,6 +10,9 @@ struct dev_archdata { #ifdef CONFIG_ACPI void *acpi_handle; #endif +#ifdef CONFIG_DMAR + void *iommu; /* hook for IOMMU specific extension */ +#endif }; #endif /* _ASM_IA64_DEVICE_H */ diff --git a/arch/ia64/include/asm/dma-mapping.h b/arch/ia64/include/asm/dma-mapping.h index 06ff1ba..bbab7e2 100644 --- a/arch/ia64/include/asm/dma-mapping.h +++ b/arch/ia64/include/asm/dma-mapping.h @@ -7,6 +7,49 @@ */ #include <asm/machvec.h> #include <linux/scatterlist.h> +#include <asm/swiotlb.h> + +struct dma_mapping_ops { + int (*mapping_error)(struct device *dev, + dma_addr_t dma_addr); + void* (*alloc_coherent)(struct device *dev, size_t size, + dma_addr_t *dma_handle, gfp_t gfp); + void (*free_coherent)(struct device *dev, size_t size, + void *vaddr, dma_addr_t dma_handle); + dma_addr_t (*map_single)(struct device *hwdev, unsigned long ptr, + size_t size, int direction); + void (*unmap_single)(struct device *dev, dma_addr_t addr, + size_t size, int direction); + void (*sync_single_for_cpu)(struct device *hwdev, + dma_addr_t dma_handle, size_t size, + int direction); + void (*sync_single_for_device)(struct device *hwdev, + dma_addr_t dma_handle, size_t size, + int direction); + void (*sync_single_range_for_cpu)(struct device *hwdev, + dma_addr_t dma_handle, unsigned long offset, + size_t size, int direction); + void (*sync_single_range_for_device)(struct device *hwdev, + dma_addr_t dma_handle, unsigned long offset, + size_t size, int direction); + void (*sync_sg_for_cpu)(struct device *hwdev, + struct scatterlist *sg, int nelems, + int direction); + void (*sync_sg_for_device)(struct device *hwdev, + struct scatterlist *sg, int nelems, + int direction); + int (*map_sg)(struct device *hwdev, struct scatterlist *sg, + int nents, int direction); + void (*unmap_sg)(struct device *hwdev, + struct scatterlist *sg, int nents, + int direction); + int (*dma_supported_op)(struct device *hwdev, u64 mask); + int is_phys; +}; + +extern struct dma_mapping_ops *dma_ops; +extern struct ia64_machine_vector ia64_mv; +extern void set_iommu_machvec(void); #define dma_alloc_coherent(dev, size, handle, gfp) \ platform_dma_alloc_coherent(dev, size, handle, (gfp) | GFP_DMA) @@ -96,4 +139,11 @@ dma_cache_sync (struct device *dev, void *vaddr, size_t size, #define dma_is_consistent(d, h) (1) /* all we do is coherent memory... */ +static inline struct dma_mapping_ops *get_dma_ops(struct device *dev) +{ + return dma_ops; +} + + + #endif /* _ASM_IA64_DMA_MAPPING_H */ diff --git a/arch/ia64/include/asm/iommu.h b/arch/ia64/include/asm/iommu.h new file mode 100644 index 0000000..5fb2bb9 --- /dev/null +++ b/arch/ia64/include/asm/iommu.h @@ -0,0 +1,16 @@ +#ifndef _ASM_IA64_IOMMU_H +#define _ASM_IA64_IOMMU_H 1 + +#define cpu_has_x2apic 0 +/* 10 seconds */ +#define DMAR_OPERATION_TIMEOUT (((cycles_t) local_cpu_data->itc_freq)*10) + +extern void pci_iommu_shutdown(void); +extern void no_iommu_init(void); +extern int force_iommu, no_iommu; +extern int iommu_detected; +extern void iommu_dma_init(void); +extern void machvec_init(const char *name); +extern int forbid_dac; + +#endif diff --git a/arch/ia64/include/asm/kregs.h b/arch/ia64/include/asm/kregs.h index aefcdfe..39e65f6 100644 --- a/arch/ia64/include/asm/kregs.h +++ b/arch/ia64/include/asm/kregs.h @@ -32,7 +32,7 @@ #define IA64_TR_CURRENT_STACK 1 /* dtr1: maps kernel's memory- & register-stacks */ #define IA64_TR_ALLOC_BASE 2 /* itr&dtr: Base of dynamic TR resource*/ -#define IA64_TR_ALLOC_MAX 32 /* Max number for dynamic use*/ +#define IA64_TR_ALLOC_MAX 64 /* Max number for dynamic use*/ /* Processor status register bits: */ #define IA64_PSR_BE_BIT 1 diff --git a/arch/ia64/include/asm/machvec.h b/arch/ia64/include/asm/machvec.h index 2b850cc..1ea28bc 100644 --- a/arch/ia64/include/asm/machvec.h +++ b/arch/ia64/include/asm/machvec.h @@ -120,6 +120,8 @@ extern void machvec_tlb_migrate_finish (struct mm_struct *); # include <asm/machvec_hpsim.h> # elif defined (CONFIG_IA64_DIG) # include <asm/machvec_dig.h> +# elif defined(CONFIG_IA64_DIG_VTD) +# include <asm/machvec_dig_vtd.h> # elif defined (CONFIG_IA64_HP_ZX1) # include <asm/machvec_hpzx1.h> # elif defined (CONFIG_IA64_HP_ZX1_SWIOTLB) @@ -128,6 +130,8 @@ extern void machvec_tlb_migrate_finish (struct mm_struct *); # include <asm/machvec_sn2.h> # elif defined (CONFIG_IA64_SGI_UV) # include <asm/machvec_uv.h> +# elif defined (CONFIG_IA64_XEN_GUEST) +# include <asm/machvec_xen.h> # elif defined (CONFIG_IA64_GENERIC) # ifdef MACHVEC_PLATFORM_HEADER diff --git a/arch/ia64/include/asm/machvec_dig_vtd.h b/arch/ia64/include/asm/machvec_dig_vtd.h new file mode 100644 index 0000000..3400b56 --- /dev/null +++ b/arch/ia64/include/asm/machvec_dig_vtd.h @@ -0,0 +1,38 @@ +#ifndef _ASM_IA64_MACHVEC_DIG_VTD_h +#define _ASM_IA64_MACHVEC_DIG_VTD_h + +extern ia64_mv_setup_t dig_setup; +extern ia64_mv_dma_alloc_coherent vtd_alloc_coherent; +extern ia64_mv_dma_free_coherent vtd_free_coherent; +extern ia64_mv_dma_map_single_attrs vtd_map_single_attrs; +extern ia64_mv_dma_unmap_single_attrs vtd_unmap_single_attrs; +extern ia64_mv_dma_map_sg_attrs vtd_map_sg_attrs; +extern ia64_mv_dma_unmap_sg_attrs vtd_unmap_sg_attrs; +extern ia64_mv_dma_supported iommu_dma_supported; +extern ia64_mv_dma_mapping_error vtd_dma_mapping_error; +extern ia64_mv_dma_init pci_iommu_alloc; + +/* + * This stuff has dual use! + * + * For a generic kernel, the macros are used to initialize the + * platform's machvec structure. When compiling a non-generic kernel, + * the macros are used directly. + */ +#define platform_name "dig_vtd" +#define platform_setup dig_setup +#define platform_dma_init pci_iommu_alloc +#define platform_dma_alloc_coherent vtd_alloc_coherent +#define platform_dma_free_coherent vtd_free_coherent +#define platform_dma_map_single_attrs vtd_map_single_attrs +#define platform_dma_unmap_single_attrs vtd_unmap_single_attrs +#define platform_dma_map_sg_attrs vtd_map_sg_attrs +#define platform_dma_unmap_sg_attrs vtd_unmap_sg_attrs +#define platform_dma_sync_single_for_cpu machvec_dma_sync_single +#define platform_dma_sync_sg_for_cpu machvec_dma_sync_sg +#define platform_dma_sync_single_for_device machvec_dma_sync_single +#define platform_dma_sync_sg_for_device machvec_dma_sync_sg +#define platform_dma_supported iommu_dma_supported +#define platform_dma_mapping_error vtd_dma_mapping_error + +#endif /* _ASM_IA64_MACHVEC_DIG_VTD_h */ diff --git a/arch/ia64/include/asm/machvec_init.h b/arch/ia64/include/asm/machvec_init.h index 7f21249..ef964b2 100644 --- a/arch/ia64/include/asm/machvec_init.h +++ b/arch/ia64/include/asm/machvec_init.h @@ -1,3 +1,4 @@ +#include <asm/iommu.h> #include <asm/machvec.h> extern ia64_mv_send_ipi_t ia64_send_ipi; diff --git a/arch/ia64/include/asm/machvec_xen.h b/arch/ia64/include/asm/machvec_xen.h new file mode 100644 index 0000000..55f9228 --- /dev/null +++ b/arch/ia64/include/asm/machvec_xen.h @@ -0,0 +1,22 @@ +#ifndef _ASM_IA64_MACHVEC_XEN_h +#define _ASM_IA64_MACHVEC_XEN_h + +extern ia64_mv_setup_t dig_setup; +extern ia64_mv_cpu_init_t xen_cpu_init; +extern ia64_mv_irq_init_t xen_irq_init; +extern ia64_mv_send_ipi_t xen_platform_send_ipi; + +/* + * This stuff has dual use! + * + * For a generic kernel, the macros are used to initialize the + * platform's machvec structure. When compiling a non-generic kernel, + * the macros are used directly. + */ +#define platform_name "xen" +#define platform_setup dig_setup +#define platform_cpu_init xen_cpu_init +#define platform_irq_init xen_irq_init +#define platform_send_ipi xen_platform_send_ipi + +#endif /* _ASM_IA64_MACHVEC_XEN_h */ diff --git a/arch/ia64/include/asm/meminit.h b/arch/ia64/include/asm/meminit.h index 7245a57..6bc96ee 100644 --- a/arch/ia64/include/asm/meminit.h +++ b/arch/ia64/include/asm/meminit.h @@ -18,10 +18,11 @@ * - crash dumping code reserved region * - Kernel memory map built from EFI memory map * - ELF core header + * - xen start info if CONFIG_XEN * * More could be added if necessary */ -#define IA64_MAX_RSVD_REGIONS 8 +#define IA64_MAX_RSVD_REGIONS 9 struct rsvd_region { unsigned long start; /* virtual address of beginning of element */ diff --git a/arch/ia64/include/asm/native/inst.h b/arch/ia64/include/asm/native/inst.h index c8efbf7..0a1026c 100644 --- a/arch/ia64/include/asm/native/inst.h +++ b/arch/ia64/include/asm/native/inst.h @@ -36,8 +36,13 @@ ;; \ movl clob = PARAVIRT_POISON; \ ;; +# define CLOBBER_PRED(pred_clob) \ + ;; \ + cmp.eq pred_clob, p0 = r0, r0 \ + ;; #else -# define CLOBBER(clob) /* nothing */ +# define CLOBBER(clob) /* nothing */ +# define CLOBBER_PRED(pred_clob) /* nothing */ #endif #define MOV_FROM_IFA(reg) \ @@ -136,7 +141,8 @@ #define SSM_PSR_I(pred, pred_clob, clob) \ (pred) ssm psr.i \ - CLOBBER(clob) + CLOBBER(clob) \ + CLOBBER_PRED(pred_clob) #define RSM_PSR_I(pred, clob0, clob1) \ (pred) rsm psr.i \ diff --git a/arch/ia64/include/asm/native/pvchk_inst.h b/arch/ia64/include/asm/native/pvchk_inst.h new file mode 100644 index 0000000..b8e6eb1 --- /dev/null +++ b/arch/ia64/include/asm/native/pvchk_inst.h @@ -0,0 +1,263 @@ +#ifndef _ASM_NATIVE_PVCHK_INST_H +#define _ASM_NATIVE_PVCHK_INST_H + +/****************************************************************************** + * arch/ia64/include/asm/native/pvchk_inst.h + * Checker for paravirtualizations of privileged operations. + * + * Copyright (C) 2005 Hewlett-Packard Co + * Dan Magenheimer <dan.magenheimer@hp.com> + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +/********************************************** + * Instructions paravirtualized for correctness + **********************************************/ + +/* "fc" and "thash" are privilege-sensitive instructions, meaning they + * may have different semantics depending on whether they are executed + * at PL0 vs PL!=0. When paravirtualized, these instructions mustn't + * be allowed to execute directly, lest incorrect semantics result. + */ + +#define fc .error "fc should not be used directly." +#define thash .error "thash should not be used directly." + +/* Note that "ttag" and "cover" are also privilege-sensitive; "ttag" + * is not currently used (though it may be in a long-format VHPT system!) + * and the semantics of cover only change if psr.ic is off which is very + * rare (and currently non-existent outside of assembly code + */ +#define ttag .error "ttag should not be used directly." +#define cover .error "cover should not be used directly." + +/* There are also privilege-sensitive registers. These registers are + * readable at any privilege level but only writable at PL0. + */ +#define cpuid .error "cpuid should not be used directly." +#define pmd .error "pmd should not be used directly." + +/* + * mov ar.eflag = + * mov = ar.eflag + */ + +/********************************************** + * Instructions paravirtualized for performance + **********************************************/ +/* + * Those instructions include '.' which can't be handled by cpp. + * or can't be handled by cpp easily. + * They are handled by sed instead of cpp. + */ + +/* for .S + * itc.i + * itc.d + * + * bsw.0 + * bsw.1 + * + * ssm psr.ic | PSR_DEFAULT_BITS + * ssm psr.ic + * rsm psr.ic + * ssm psr.i + * rsm psr.i + * rsm psr.i | psr.ic + * rsm psr.dt + * ssm psr.dt + * + * mov = cr.ifa + * mov = cr.itir + * mov = cr.isr + * mov = cr.iha + * mov = cr.ipsr + * mov = cr.iim + * mov = cr.iip + * mov = cr.ivr + * mov = psr + * + * mov cr.ifa = + * mov cr.itir = + * mov cr.iha = + * mov cr.ipsr = + * mov cr.ifs = + * mov cr.iip = + * mov cr.kr = + */ + +/* for intrinsics + * ssm psr.i + * rsm psr.i + * mov = psr + * mov = ivr + * mov = tpr + * mov cr.itm = + * mov eoi = + * mov rr[] = + * mov = rr[] + * mov = kr + * mov kr = + * ptc.ga + */ + +/************************************************************* + * define paravirtualized instrcution macros as nop to ingore. + * and check whether arguments are appropriate. + *************************************************************/ + +/* check whether reg is a regular register */ +.macro is_rreg_in reg + .ifc "\reg", "r0" + nop 0 + .exitm + .endif + ;; + mov \reg = r0 + ;; +.endm +#define IS_RREG_IN(reg) is_rreg_in reg ; + +#define IS_RREG_OUT(reg) \ + ;; \ + mov reg = r0 \ + ;; + +#define IS_RREG_CLOB(reg) IS_RREG_OUT(reg) + +/* check whether pred is a predicate register */ +#define IS_PRED_IN(pred) \ + ;; \ + (pred) nop 0 \ + ;; + +#define IS_PRED_OUT(pred) \ + ;; \ + cmp.eq pred, p0 = r0, r0 \ + ;; + +#define IS_PRED_CLOB(pred) IS_PRED_OUT(pred) + + +#define DO_SAVE_MIN(__COVER, SAVE_IFS, EXTRA, WORKAROUND) \ + nop 0 +#define MOV_FROM_IFA(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_ITIR(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_ISR(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_IHA(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_IPSR(pred, reg) \ + IS_PRED_IN(pred) \ + IS_RREG_OUT(reg) +#define MOV_FROM_IIM(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_IIP(reg) \ + IS_RREG_OUT(reg) +#define MOV_FROM_IVR(reg, clob) \ + IS_RREG_OUT(reg) \ + IS_RREG_CLOB(clob) +#define MOV_FROM_PSR(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_OUT(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_IFA(reg, clob) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_ITIR(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_IHA(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_IPSR(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_IFS(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_IIP(reg, clob) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define MOV_TO_KR(kr, reg, clob0, clob1) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) +#define ITC_I(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define ITC_D(pred, reg, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define ITC_I_AND_D(pred_i, pred_d, reg, clob) \ + IS_PRED_IN(pred_i) \ + IS_PRED_IN(pred_d) \ + IS_RREG_IN(reg) \ + IS_RREG_CLOB(clob) +#define THASH(pred, reg0, reg1, clob) \ + IS_PRED_IN(pred) \ + IS_RREG_OUT(reg0) \ + IS_RREG_IN(reg1) \ + IS_RREG_CLOB(clob) +#define SSM_PSR_IC_AND_DEFAULT_BITS_AND_SRLZ_I(clob0, clob1) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) +#define SSM_PSR_IC_AND_SRLZ_D(clob0, clob1) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) +#define RSM_PSR_IC(clob) \ + IS_RREG_CLOB(clob) +#define SSM_PSR_I(pred, pred_clob, clob) \ + IS_PRED_IN(pred) \ + IS_PRED_CLOB(pred_clob) \ + IS_RREG_CLOB(clob) +#define RSM_PSR_I(pred, clob0, clob1) \ + IS_PRED_IN(pred) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) +#define RSM_PSR_I_IC(clob0, clob1, clob2) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) \ + IS_RREG_CLOB(clob2) +#define RSM_PSR_DT \ + nop 0 +#define SSM_PSR_DT_AND_SRLZ_I \ + nop 0 +#define BSW_0(clob0, clob1, clob2) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) \ + IS_RREG_CLOB(clob2) +#define BSW_1(clob0, clob1) \ + IS_RREG_CLOB(clob0) \ + IS_RREG_CLOB(clob1) +#define COVER \ + nop 0 +#define RFI \ + br.ret.sptk.many rp /* defining nop causes dependency error */ + +#endif /* _ASM_NATIVE_PVCHK_INST_H */ diff --git a/arch/ia64/include/asm/paravirt.h b/arch/ia64/include/asm/paravirt.h index 660cab0..2bf3636 100644 --- a/arch/ia64/include/asm/paravirt.h +++ b/arch/ia64/include/asm/paravirt.h @@ -117,7 +117,7 @@ static inline void paravirt_post_smp_prepare_boot_cpu(void) struct pv_iosapic_ops { void (*pcat_compat_init)(void); - struct irq_chip *(*get_irq_chip)(unsigned long trigger); + struct irq_chip *(*__get_irq_chip)(unsigned long trigger); unsigned int (*__read)(char __iomem *iosapic, unsigned int reg); void (*__write)(char __iomem *iosapic, unsigned int reg, u32 val); @@ -135,7 +135,7 @@ iosapic_pcat_compat_init(void) static inline struct irq_chip* iosapic_get_irq_chip(unsigned long trigger) { - return pv_iosapic_ops.get_irq_chip(trigger); + return pv_iosapic_ops.__get_irq_chip(trigger); } static inline unsigned int diff --git a/arch/ia64/include/asm/pci.h b/arch/ia64/include/asm/pci.h index ce342fb..1d660d8 100644 --- a/arch/ia64/include/asm/pci.h +++ b/arch/ia64/include/asm/pci.h @@ -156,4 +156,7 @@ static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel) return channel ? isa_irq_to_vector(15) : isa_irq_to_vector(14); } +#ifdef CONFIG_DMAR +extern void pci_iommu_alloc(void); +#endif #endif /* _ASM_IA64_PCI_H */ diff --git a/arch/ia64/include/asm/ptrace.h b/arch/ia64/include/asm/ptrace.h index 15f8dcf..6417c1e 100644 --- a/arch/ia64/include/asm/ptrace.h +++ b/arch/ia64/include/asm/ptrace.h @@ -240,6 +240,12 @@ struct switch_stack { */ # define instruction_pointer(regs) ((regs)->cr_iip + ia64_psr(regs)->ri) +static inline unsigned long user_stack_pointer(struct pt_regs *regs) +{ + /* FIXME: should this be bspstore + nr_dirty regs? */ + return regs->ar_bspstore; +} + #define regs_return_value(regs) ((regs)->r8) /* Conserve space in histogram by encoding slot bits in address @@ -319,6 +325,8 @@ struct switch_stack { #define arch_has_block_step() (1) extern void user_enable_block_step(struct task_struct *); +#define __ARCH_WANT_COMPAT_SYS_PTRACE + #endif /* !__KERNEL__ */ /* pt_all_user_regs is used for PTRACE_GETREGS PTRACE_SETREGS */ diff --git a/arch/ia64/include/asm/pvclock-abi.h b/arch/ia64/include/asm/pvclock-abi.h new file mode 100644 index 0000000..44ef9ef --- /dev/null +++ b/arch/ia64/include/asm/pvclock-abi.h @@ -0,0 +1,48 @@ +/* + * same structure to x86's + * Hopefully asm-x86/pvclock-abi.h would be moved to somewhere more generic. + * For now, define same duplicated definitions. + */ + +#ifndef _ASM_IA64__PVCLOCK_ABI_H +#define _ASM_IA64__PVCLOCK_ABI_H +#ifndef __ASSEMBLY__ + +/* + * These structs MUST NOT be changed. + * They are the ABI between hypervisor and guest OS. + * Both Xen and KVM are using this. + * + * pvclock_vcpu_time_info holds the system time and the tsc timestamp + * of the last update. So the guest can use the tsc delta to get a + * more precise system time. There is one per virtual cpu. + * + * pvclock_wall_clock references the point in time when the system + * time was zero (usually boot time), thus the guest calculates the + * current wall clock by adding the system time. + * + * Protocol for the "version" fields is: hypervisor raises it (making + * it uneven) before it starts updating the fields and raises it again + * (making it even) when it is done. Thus the guest can make sure the + * time values it got are consistent by checking the version before + * and after reading them. + */ + +struct pvclock_vcpu_time_info { + u32 version; + u32 pad0; + u64 tsc_timestamp; + u64 system_time; + u32 tsc_to_system_mul; + s8 tsc_shift; + u8 pad[3]; +} __attribute__((__packed__)); /* 32 bytes */ + +struct pvclock_wall_clock { + u32 version; + u32 sec; + u32 nsec; +} __attribute__((__packed__)); + +#endif /* __ASSEMBLY__ */ +#endif /* _ASM_IA64__PVCLOCK_ABI_H */ diff --git a/arch/ia64/include/asm/swiotlb.h b/arch/ia64/include/asm/swiotlb.h new file mode 100644 index 0000000..fb79423 --- /dev/null +++ b/arch/ia64/include/asm/swiotlb.h @@ -0,0 +1,56 @@ +#ifndef ASM_IA64__SWIOTLB_H +#define ASM_IA64__SWIOTLB_H + +#include <linux/dma-mapping.h> + +/* SWIOTLB interface */ + +extern dma_addr_t swiotlb_map_single(struct device *hwdev, void *ptr, + size_t size, int dir); +extern void *swiotlb_alloc_coherent(struct device *hwdev, size_t size, + dma_addr_t *dma_handle, gfp_t flags); +extern void swiotlb_unmap_single(struct device *hwdev, dma_addr_t dev_addr, + size_t size, int dir); +extern void swiotlb_sync_single_for_cpu(struct device *hwdev, + dma_addr_t dev_addr, + size_t size, int dir); +extern void swiotlb_sync_single_for_device(struct device *hwdev, + dma_addr_t dev_addr, + size_t size, int dir); +extern void swiotlb_sync_single_range_for_cpu(struct device *hwdev, + dma_addr_t dev_addr, + unsigned long offset, + size_t size, int dir); +extern void swiotlb_sync_single_range_for_device(struct device *hwdev, + dma_addr_t dev_addr, + unsigned long offset, + size_t size, int dir); +extern void swiotlb_sync_sg_for_cpu(struct device *hwdev, + struct scatterlist *sg, int nelems, + int dir); +extern void swiotlb_sync_sg_for_device(struct device *hwdev, + struct scatterlist *sg, int nelems, + int dir); +extern int swiotlb_map_sg(struct device *hwdev, struct scatterlist *sg, + int nents, int direction); +extern void swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sg, + int nents, int direction); +extern int swiotlb_dma_mapping_error(struct device *hwdev, dma_addr_t dma_addr); +extern void swiotlb_free_coherent(struct device *hwdev, size_t size, + void *vaddr, dma_addr_t dma_handle); +extern int swiotlb_dma_supported(struct device *hwdev, u64 mask); +extern void swiotlb_init(void); + +extern int swiotlb_force; + +#ifdef CONFIG_SWIOTLB +extern int swiotlb; +extern void pci_swiotlb_init(void); +#else +#define swiotlb 0 +static inline void pci_swiotlb_init(void) +{ +} +#endif + +#endif /* ASM_IA64__SWIOTLB_H */ diff --git a/arch/ia64/include/asm/sync_bitops.h b/arch/ia64/include/asm/sync_bitops.h new file mode 100644 index 0000000..593c12e --- /dev/null +++ b/arch/ia64/include/asm/sync_bitops.h @@ -0,0 +1,51 @@ +#ifndef _ASM_IA64_SYNC_BITOPS_H +#define _ASM_IA64_SYNC_BITOPS_H + +/* + * Copyright (C) 2008 Isaku Yamahata <yamahata at valinux co jp> + * + * Based on synch_bitops.h which Dan Magenhaimer wrote. + * + * bit operations which provide guaranteed strong synchronisation + * when communicating with Xen or other guest OSes running on other CPUs. + */ + +static inline void sync_set_bit(int nr, volatile void *addr) +{ + set_bit(nr, addr); +} + +static inline void sync_clear_bit(int nr, volatile void *addr) +{ + clear_bit(nr, addr); +} + +static inline void sync_change_bit(int nr, volatile void *addr) +{ + change_bit(nr, addr); +} + +static inline int sync_test_and_set_bit(int nr, volatile void *addr) +{ + return test_and_set_bit(nr, addr); +} + +static inline int sync_test_and_clear_bit(int nr, volatile void *addr) +{ + return test_and_clear_bit(nr, addr); +} + +static inline int sync_test_and_change_bit(int nr, volatile void *addr) +{ + return test_and_change_bit(nr, addr); +} + +static inline int sync_test_bit(int nr, const volatile void *addr) +{ + return test_bit(nr, addr); +} + +#define sync_cmpxchg(ptr, old, new) \ + ((__typeof__(*(ptr)))cmpxchg_acq((ptr), (old), (new))) + +#endif /* _ASM_IA64_SYNC_BITOPS_H */ diff --git a/arch/ia64/include/asm/syscall.h b/arch/ia64/include/asm/syscall.h new file mode 100644 index 0000000..2f758a4 --- /dev/null +++ b/arch/ia64/include/asm/syscall.h @@ -0,0 +1,163 @@ +/* + * Access to user system call parameters and results + * + * Copyright (C) 2008 Intel Corp. Shaohua Li <shaohua.li@intel.com> + * + * This copyrighted material is made available to anyone wishing to use, + * modify, copy, or redistribute it subject to the terms and conditions + * of the GNU General Public License v.2. + * + * See asm-generic/syscall.h for descriptions of what we must do here. + */ + +#ifndef _ASM_SYSCALL_H +#define _ASM_SYSCALL_H 1 + +#include <linux/sched.h> +#include <linux/err.h> + +static inline long syscall_get_nr(struct task_struct *task, + struct pt_regs *regs) +{ + if ((long)regs->cr_ifs < 0) /* Not a syscall */ + return -1; + +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) + return regs->r1; +#endif + + return regs->r15; +} + +static inline void syscall_rollback(struct task_struct *task, + struct pt_regs *regs) +{ +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) + regs->r8 = regs->r1; +#endif + + /* do nothing */ +} + +static inline long syscall_get_error(struct task_struct *task, + struct pt_regs *regs) +{ +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) + return regs->r8; +#endif + + return regs->r10 == -1 ? regs->r8:0; +} + +static inline long syscall_get_return_value(struct task_struct *task, + struct pt_regs *regs) +{ + return regs->r8; +} + +static inline void syscall_set_return_value(struct task_struct *task, + struct pt_regs *regs, + int error, long val) +{ +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) { + regs->r8 = (long) error ? error : val; + return; + } +#endif + + if (error) { + /* error < 0, but ia64 uses > 0 return value */ + regs->r8 = -error; + regs->r10 = -1; + } else { + regs->r8 = val; + regs->r10 = 0; + } +} + +extern void ia64_syscall_get_set_arguments(struct task_struct *task, + struct pt_regs *regs, unsigned int i, unsigned int n, + unsigned long *args, int rw); +static inline void syscall_get_arguments(struct task_struct *task, + struct pt_regs *regs, + unsigned int i, unsigned int n, + unsigned long *args) +{ + BUG_ON(i + n > 6); + +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) { + switch (i + n) { + case 6: + if (!n--) break; + *args++ = regs->r13; + case 5: + if (!n--) break; + *args++ = regs->r15; + case 4: + if (!n--) break; + *args++ = regs->r14; + case 3: + if (!n--) break; + *args++ = regs->r10; + case 2: + if (!n--) break; + *args++ = regs->r9; + case 1: + if (!n--) break; + *args++ = regs->r11; + case 0: + if (!n--) break; + default: + BUG(); + break; + } + + return; + } +#endif + ia64_syscall_get_set_arguments(task, regs, i, n, args, 0); +} + +static inline void syscall_set_arguments(struct task_struct *task, + struct pt_regs *regs, + unsigned int i, unsigned int n, + unsigned long *args) +{ + BUG_ON(i + n > 6); + +#ifdef CONFIG_IA32_SUPPORT + if (IS_IA32_PROCESS(regs)) { + switch (i + n) { + case 6: + if (!n--) break; + regs->r13 = *args++; + case 5: + if (!n--) break; + regs->r15 = *args++; + case 4: + if (!n--) break; + regs->r14 = *args++; + case 3: + if (!n--) break; + regs->r10 = *args++; + case 2: + if (!n--) break; + regs->r9 = *args++; + case 1: + if (!n--) break; + regs->r11 = *args++; + case 0: + if (!n--) break; + } + + return; + } +#endif + ia64_syscall_get_set_arguments(task, regs, i, n, args, 1); +} +#endif /* _ASM_SYSCALL_H */ diff --git a/arch/ia64/include/asm/thread_info.h b/arch/ia64/include/asm/thread_info.h index 7c60fcd..ae69226 100644 --- a/arch/ia64/include/asm/thread_info.h +++ b/arch/ia64/include/asm/thread_info.h @@ -87,9 +87,6 @@ struct thread_info { #define alloc_task_struct() ((struct task_struct *)__get_free_pages(GFP_KERNEL | __GFP_COMP, KERNEL_STACK_SIZE_ORDER)) #define free_task_struct(tsk) free_pages((unsigned long) (tsk), KERNEL_STACK_SIZE_ORDER) -#define tsk_set_notify_resume(tsk) \ - set_ti_thread_flag(task_thread_info(tsk), TIF_NOTIFY_RESUME) -extern void tsk_clear_notify_resume(struct task_struct *tsk); #endif /* !__ASSEMBLY */ /* diff --git a/arch/ia64/include/asm/timex.h b/arch/ia64/include/asm/timex.h index 05a6baf..4e03cfe 100644 --- a/arch/ia64/include/asm/timex.h +++ b/arch/ia64/include/asm/timex.h @@ -39,4 +39,6 @@ get_cycles (void) return ret; } +extern void ia64_cpu_local_tick (void); + #endif /* _ASM_IA64_TIMEX_H */ diff --git a/arch/ia64/include/asm/unistd.h b/arch/ia64/include/asm/unistd.h index d535833..f791576 100644 --- a/arch/ia64/include/asm/unistd.h +++ b/arch/ia64/include/asm/unistd.h @@ -337,6 +337,7 @@ # define __ARCH_WANT_SYS_NICE # define __ARCH_WANT_SYS_OLD_GETRLIMIT # define __ARCH_WANT_SYS_OLDUMOUNT +# define __ARCH_WANT_SYS_PAUSE # define __ARCH_WANT_SYS_SIGPENDING # define __ARCH_WANT_SYS_SIGPROCMASK # define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND diff --git a/arch/ia64/include/asm/xen/events.h b/arch/ia64/include/asm/xen/events.h new file mode 100644 index 0000000..7324878 --- /dev/null +++ b/arch/ia64/include/asm/xen/events.h @@ -0,0 +1,50 @@ +/****************************************************************************** + * arch/ia64/include/asm/xen/events.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ +#ifndef _ASM_IA64_XEN_EVENTS_H +#define _ASM_IA64_XEN_EVENTS_H + +enum ipi_vector { + XEN_RESCHEDULE_VECTOR, + XEN_IPI_VECTOR, + XEN_CMCP_VECTOR, + XEN_CPEP_VECTOR, + + XEN_NR_IPIS, +}; + +static inline int xen_irqs_disabled(struct pt_regs *regs) +{ + return !(ia64_psr(regs)->i); +} + +static inline void xen_do_IRQ(int irq, struct pt_regs *regs) +{ + struct pt_regs *old_regs; + old_regs = set_irq_regs(regs); + irq_enter(); + __do_IRQ(irq); + irq_exit(); + set_irq_regs(old_regs); +} +#define irq_ctx_init(cpu) do { } while (0) + +#endif /* _ASM_IA64_XEN_EVENTS_H */ diff --git a/arch/ia64/include/asm/xen/grant_table.h b/arch/ia64/include/asm/xen/grant_table.h new file mode 100644 index 0000000..2b1fae0 --- /dev/null +++ b/arch/ia64/include/asm/xen/grant_table.h @@ -0,0 +1,29 @@ +/****************************************************************************** + * arch/ia64/include/asm/xen/grant_table.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#ifndef _ASM_IA64_XEN_GRANT_TABLE_H +#define _ASM_IA64_XEN_GRANT_TABLE_H + +struct vm_struct *xen_alloc_vm_area(unsigned long size); +void xen_free_vm_area(struct vm_struct *area); + +#endif /* _ASM_IA64_XEN_GRANT_TABLE_H */ diff --git a/arch/ia64/include/asm/xen/hypercall.h b/arch/ia64/include/asm/xen/hypercall.h new file mode 100644 index 0000000..96fc623 --- /dev/null +++ b/arch/ia64/include/asm/xen/hypercall.h @@ -0,0 +1,265 @@ +/****************************************************************************** + * hypercall.h + * + * Linux-specific hypervisor handling. + * + * Copyright (c) 2002-2004, K A Fraser + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License version 2 + * as published by the Free Software Foundation; or, when distributed + * separately from the Linux kernel or incorporated into other + * software packages, subject to the following license: + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this source file (the "Software"), to deal in the Software without + * restriction, including without limitation the rights to use, copy, modify, + * merge, publish, distribute, sublicense, and/or sell copies of the Software, + * and to permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ + +#ifndef _ASM_IA64_XEN_HYPERCALL_H +#define _ASM_IA64_XEN_HYPERCALL_H + +#include <xen/interface/xen.h> +#include <xen/interface/physdev.h> +#include <xen/interface/sched.h> +#include <asm/xen/xcom_hcall.h> +struct xencomm_handle; +extern unsigned long __hypercall(unsigned long a1, unsigned long a2, + unsigned long a3, unsigned long a4, + unsigned long a5, unsigned long cmd); + +/* + * Assembler stubs for hyper-calls. + */ + +#define _hypercall0(type, name) \ +({ \ + long __res; \ + __res = __hypercall(0, 0, 0, 0, 0, __HYPERVISOR_##name);\ + (type)__res; \ +}) + +#define _hypercall1(type, name, a1) \ +({ \ + long __res; \ + __res = __hypercall((unsigned long)a1, \ + 0, 0, 0, 0, __HYPERVISOR_##name); \ + (type)__res; \ +}) + +#define _hypercall2(type, name, a1, a2) \ +({ \ + long __res; \ + __res = __hypercall((unsigned long)a1, \ + (unsigned long)a2, \ + 0, 0, 0, __HYPERVISOR_##name); \ + (type)__res; \ +}) + +#define _hypercall3(type, name, a1, a2, a3) \ +({ \ + long __res; \ + __res = __hypercall((unsigned long)a1, \ + (unsigned long)a2, \ + (unsigned long)a3, \ + 0, 0, __HYPERVISOR_##name); \ + (type)__res; \ +}) + +#define _hypercall4(type, name, a1, a2, a3, a4) \ +({ \ + long __res; \ + __res = __hypercall((unsigned long)a1, \ + (unsigned long)a2, \ + (unsigned long)a3, \ + (unsigned long)a4, \ + 0, __HYPERVISOR_##name); \ + (type)__res; \ +}) + +#define _hypercall5(type, name, a1, a2, a3, a4, a5) \ +({ \ + long __res; \ + __res = __hypercall((unsigned long)a1, \ + (unsigned long)a2, \ + (unsigned long)a3, \ + (unsigned long)a4, \ + (unsigned long)a5, \ + __HYPERVISOR_##name); \ + (type)__res; \ +}) + + +static inline int +xencomm_arch_hypercall_sched_op(int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, sched_op_new, cmd, arg); +} + +static inline long +HYPERVISOR_set_timer_op(u64 timeout) +{ + unsigned long timeout_hi = (unsigned long)(timeout >> 32); + unsigned long timeout_lo = (unsigned long)timeout; + return _hypercall2(long, set_timer_op, timeout_lo, timeout_hi); +} + +static inline int +xencomm_arch_hypercall_multicall(struct xencomm_handle *call_list, + int nr_calls) +{ + return _hypercall2(int, multicall, call_list, nr_calls); +} + +static inline int +xencomm_arch_hypercall_memory_op(unsigned int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, memory_op, cmd, arg); +} + +static inline int +xencomm_arch_hypercall_event_channel_op(int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, event_channel_op, cmd, arg); +} + +static inline int +xencomm_arch_hypercall_xen_version(int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, xen_version, cmd, arg); +} + +static inline int +xencomm_arch_hypercall_console_io(int cmd, int count, + struct xencomm_handle *str) +{ + return _hypercall3(int, console_io, cmd, count, str); +} + +static inline int +xencomm_arch_hypercall_physdev_op(int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, physdev_op, cmd, arg); +} + +static inline int +xencomm_arch_hypercall_grant_table_op(unsigned int cmd, + struct xencomm_handle *uop, + unsigned int count) +{ + return _hypercall3(int, grant_table_op, cmd, uop, count); +} + +int HYPERVISOR_grant_table_op(unsigned int cmd, void *uop, unsigned int count); + +extern int xencomm_arch_hypercall_suspend(struct xencomm_handle *arg); + +static inline int +xencomm_arch_hypercall_callback_op(int cmd, struct xencomm_handle *arg) +{ + return _hypercall2(int, callback_op, cmd, arg); +} + +static inline long +xencomm_arch_hypercall_vcpu_op(int cmd, int cpu, void *arg) +{ + return _hypercall3(long, vcpu_op, cmd, cpu, arg); +} + +static inline int +HYPERVISOR_physdev_op(int cmd, void *arg) +{ + switch (cmd) { + case PHYSDEVOP_eoi: + return _hypercall1(int, ia64_fast_eoi, + ((struct physdev_eoi *)arg)->irq); + default: + return xencomm_hypercall_physdev_op(cmd, arg); + } +} + +static inline long +xencomm_arch_hypercall_opt_feature(struct xencomm_handle *arg) +{ + return _hypercall1(long, opt_feature, arg); +} + +/* for balloon driver */ +#define HYPERVISOR_update_va_mapping(va, new_val, flags) (0) + +/* Use xencomm to do hypercalls. */ +#define HYPERVISOR_sched_op xencomm_hypercall_sched_op +#define HYPERVISOR_event_channel_op xencomm_hypercall_event_channel_op +#define HYPERVISOR_callback_op xencomm_hypercall_callback_op +#define HYPERVISOR_multicall xencomm_hypercall_multicall +#define HYPERVISOR_xen_version xencomm_hypercall_xen_version +#define HYPERVISOR_console_io xencomm_hypercall_console_io +#define HYPERVISOR_memory_op xencomm_hypercall_memory_op +#define HYPERVISOR_suspend xencomm_hypercall_suspend +#define HYPERVISOR_vcpu_op xencomm_hypercall_vcpu_op +#define HYPERVISOR_opt_feature xencomm_hypercall_opt_feature + +/* to compile gnttab_copy_grant_page() in drivers/xen/core/gnttab.c */ +#define HYPERVISOR_mmu_update(req, count, success_count, domid) ({ BUG(); 0; }) + +static inline int +HYPERVISOR_shutdown( + unsigned int reason) +{ + struct sched_shutdown sched_shutdown = { + .reason = reason + }; + + int rc = HYPERVISOR_sched_op(SCHEDOP_shutdown, &sched_shutdown); + + return rc; +} + +/* for netfront.c, netback.c */ +#define MULTI_UVMFLAGS_INDEX 0 /* XXX any value */ + +static inline void +MULTI_update_va_mapping( + struct multicall_entry *mcl, unsigned long va, + pte_t new_val, unsigned long flags) +{ + mcl->op = __HYPERVISOR_update_va_mapping; + mcl->result = 0; +} + +static inline void +MULTI_grant_table_op(struct multicall_entry *mcl, unsigned int cmd, + void *uop, unsigned int count) +{ + mcl->op = __HYPERVISOR_grant_table_op; + mcl->args[0] = cmd; + mcl->args[1] = (unsigned long)uop; + mcl->args[2] = count; +} + +static inline void +MULTI_mmu_update(struct multicall_entry *mcl, struct mmu_update *req, + int count, int *success_count, domid_t domid) +{ + mcl->op = __HYPERVISOR_mmu_update; + mcl->args[0] = (unsigned long)req; + mcl->args[1] = count; + mcl->args[2] = (unsigned long)success_count; + mcl->args[3] = domid; +} + +#endif /* _ASM_IA64_XEN_HYPERCALL_H */ diff --git a/arch/ia64/include/asm/xen/hypervisor.h b/arch/ia64/include/asm/xen/hypervisor.h new file mode 100644 index 0000000..7a804e8 --- /dev/null +++ b/arch/ia64/include/asm/xen/hypervisor.h @@ -0,0 +1,89 @@ +/****************************************************************************** + * hypervisor.h + * + * Linux-specific hypervisor handling. + * + * Copyright (c) 2002-2004, K A Fraser + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License version 2 + * as published by the Free Software Foundation; or, when distributed + * separately from the Linux kernel or incorporated into other + * software packages, subject to the following license: + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this source file (the "Software"), to deal in the Software without + * restriction, including without limitation the rights to use, copy, modify, + * merge, publish, distribute, sublicense, and/or sell copies of the Software, + * and to permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ + +#ifndef _ASM_IA64_XEN_HYPERVISOR_H +#define _ASM_IA64_XEN_HYPERVISOR_H + +#ifdef CONFIG_XEN + +#include <linux/init.h> +#include <xen/interface/xen.h> +#include <xen/interface/version.h> /* to compile feature.c */ +#include <xen/features.h> /* to comiple xen-netfront.c */ +#include <asm/xen/hypercall.h> + +/* xen_domain_type is set before executing any C code by early_xen_setup */ +enum xen_domain_type { + XEN_NATIVE, + XEN_PV_DOMAIN, + XEN_HVM_DOMAIN, +}; + +extern enum xen_domain_type xen_domain_type; + +#define xen_domain() (xen_domain_type != XEN_NATIVE) +#define xen_pv_domain() (xen_domain_type == XEN_PV_DOMAIN) +#define xen_initial_domain() (xen_pv_domain() && \ + (xen_start_info->flags & SIF_INITDOMAIN)) +#define xen_hvm_domain() (xen_domain_type == XEN_HVM_DOMAIN) + +/* deprecated. remove this */ +#define is_running_on_xen() (xen_domain_type == XEN_PV_DOMAIN) + +extern struct shared_info *HYPERVISOR_shared_info; +extern struct start_info *xen_start_info; + +void __init xen_setup_vcpu_info_placement(void); +void force_evtchn_callback(void); + +/* for drivers/xen/balloon/balloon.c */ +#ifdef CONFIG_XEN_SCRUB_PAGES +#define scrub_pages(_p, _n) memset((void *)(_p), 0, (_n) << PAGE_SHIFT) +#else +#define scrub_pages(_p, _n) ((void)0) +#endif + +/* For setup_arch() in arch/ia64/kernel/setup.c */ +void xen_ia64_enable_opt_feature(void); + +#else /* CONFIG_XEN */ + +#define xen_domain() (0) +#define xen_pv_domain() (0) +#define xen_initial_domain() (0) +#define xen_hvm_domain() (0) +#define is_running_on_xen() (0) /* deprecated. remove this */ +#endif + +#define is_initial_xendomain() (0) /* deprecated. remove this */ + +#endif /* _ASM_IA64_XEN_HYPERVISOR_H */ diff --git a/arch/ia64/include/asm/xen/inst.h b/arch/ia64/include/asm/xen/inst.h new file mode 100644 index 0000000..19c2ae1 --- /dev/null +++ b/arch/ia64/include/asm/xen/inst.h @@ -0,0 +1,458 @@ +/****************************************************************************** + * arch/ia64/include/asm/xen/inst.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <asm/xen/privop.h> + +#define ia64_ivt xen_ivt +#define DO_SAVE_MIN XEN_DO_SAVE_MIN + +#define __paravirt_switch_to xen_switch_to +#define __paravirt_leave_syscall xen_leave_syscall +#define __paravirt_work_processed_syscall xen_work_processed_syscall +#define __paravirt_leave_kernel xen_leave_kernel +#define __paravirt_pending_syscall_end xen_work_pending_syscall_end +#define __paravirt_work_processed_syscall_target \ + xen_work_processed_syscall + +#define MOV_FROM_IFA(reg) \ + movl reg = XSI_IFA; \ + ;; \ + ld8 reg = [reg] + +#define MOV_FROM_ITIR(reg) \ + movl reg = XSI_ITIR; \ + ;; \ + ld8 reg = [reg] + +#define MOV_FROM_ISR(reg) \ + movl reg = XSI_ISR; \ + ;; \ + ld8 reg = [reg] + +#define MOV_FROM_IHA(reg) \ + movl reg = XSI_IHA; \ + ;; \ + ld8 reg = [reg] + +#define MOV_FROM_IPSR(pred, reg) \ +(pred) movl reg = XSI_IPSR; \ + ;; \ +(pred) ld8 reg = [reg] + +#define MOV_FROM_IIM(reg) \ + movl reg = XSI_IIM; \ + ;; \ + ld8 reg = [reg] + +#define MOV_FROM_IIP(reg) \ + movl reg = XSI_IIP; \ + ;; \ + ld8 reg = [reg] + +.macro __MOV_FROM_IVR reg, clob + .ifc "\reg", "r8" + XEN_HYPER_GET_IVR + .exitm + .endif + .ifc "\clob", "r8" + XEN_HYPER_GET_IVR + ;; + mov \reg = r8 + .exitm + .endif + + mov \clob = r8 + ;; + XEN_HYPER_GET_IVR + ;; + mov \reg = r8 + ;; + mov r8 = \clob +.endm +#define MOV_FROM_IVR(reg, clob) __MOV_FROM_IVR reg, clob + +.macro __MOV_FROM_PSR pred, reg, clob + .ifc "\reg", "r8" + (\pred) XEN_HYPER_GET_PSR; + .exitm + .endif + .ifc "\clob", "r8" + (\pred) XEN_HYPER_GET_PSR + ;; + (\pred) mov \reg = r8 + .exitm + .endif + + (\pred) mov \clob = r8 + (\pred) XEN_HYPER_GET_PSR + ;; + (\pred) mov \reg = r8 + (\pred) mov r8 = \clob +.endm +#define MOV_FROM_PSR(pred, reg, clob) __MOV_FROM_PSR pred, reg, clob + + +#define MOV_TO_IFA(reg, clob) \ + movl clob = XSI_IFA; \ + ;; \ + st8 [clob] = reg \ + +#define MOV_TO_ITIR(pred, reg, clob) \ +(pred) movl clob = XSI_ITIR; \ + ;; \ +(pred) st8 [clob] = reg + +#define MOV_TO_IHA(pred, reg, clob) \ +(pred) movl clob = XSI_IHA; \ + ;; \ +(pred) st8 [clob] = reg + +#define MOV_TO_IPSR(pred, reg, clob) \ +(pred) movl clob = XSI_IPSR; \ + ;; \ +(pred) st8 [clob] = reg; \ + ;; + +#define MOV_TO_IFS(pred, reg, clob) \ +(pred) movl clob = XSI_IFS; \ + ;; \ +(pred) st8 [clob] = reg; \ + ;; + +#define MOV_TO_IIP(reg, clob) \ + movl clob = XSI_IIP; \ + ;; \ + st8 [clob] = reg + +.macro ____MOV_TO_KR kr, reg, clob0, clob1 + .ifc "\clob0", "r9" + .error "clob0 \clob0 must not be r9" + .endif + .ifc "\clob1", "r8" + .error "clob1 \clob1 must not be r8" + .endif + + .ifnc "\reg", "r9" + .ifnc "\clob1", "r9" + mov \clob1 = r9 + .endif + mov r9 = \reg + .endif + .ifnc "\clob0", "r8" + mov \clob0 = r8 + .endif + mov r8 = \kr + ;; + XEN_HYPER_SET_KR + + .ifnc "\reg", "r9" + .ifnc "\clob1", "r9" + mov r9 = \clob1 + .endif + .endif + .ifnc "\clob0", "r8" + mov r8 = \clob0 + .endif +.endm + +.macro __MOV_TO_KR kr, reg, clob0, clob1 + .ifc "\clob0", "r9" + ____MOV_TO_KR \kr, \reg, \clob1, \clob0 + .exitm + .endif + .ifc "\clob1", "r8" + ____MOV_TO_KR \kr, \reg, \clob1, \clob0 + .exitm + .endif + + ____MOV_TO_KR \kr, \reg, \clob0, \clob1 +.endm + +#define MOV_TO_KR(kr, reg, clob0, clob1) \ + __MOV_TO_KR IA64_KR_ ## kr, reg, clob0, clob1 + + +.macro __ITC_I pred, reg, clob + .ifc "\reg", "r8" + (\pred) XEN_HYPER_ITC_I + .exitm + .endif + .ifc "\clob", "r8" + (\pred) mov r8 = \reg + ;; + (\pred) XEN_HYPER_ITC_I + .exitm + .endif + + (\pred) mov \clob = r8 + (\pred) mov r8 = \reg + ;; + (\pred) XEN_HYPER_ITC_I + ;; + (\pred) mov r8 = \clob + ;; +.endm +#define ITC_I(pred, reg, clob) __ITC_I pred, reg, clob + +.macro __ITC_D pred, reg, clob + .ifc "\reg", "r8" + (\pred) XEN_HYPER_ITC_D + ;; + .exitm + .endif + .ifc "\clob", "r8" + (\pred) mov r8 = \reg + ;; + (\pred) XEN_HYPER_ITC_D + ;; + .exitm + .endif + + (\pred) mov \clob = r8 + (\pred) mov r8 = \reg + ;; + (\pred) XEN_HYPER_ITC_D + ;; + (\pred) mov r8 = \clob + ;; +.endm +#define ITC_D(pred, reg, clob) __ITC_D pred, reg, clob + +.macro __ITC_I_AND_D pred_i, pred_d, reg, clob + .ifc "\reg", "r8" + (\pred_i)XEN_HYPER_ITC_I + ;; + (\pred_d)XEN_HYPER_ITC_D + ;; + .exitm + .endif + .ifc "\clob", "r8" + mov r8 = \reg + ;; + (\pred_i)XEN_HYPER_ITC_I + ;; + (\pred_d)XEN_HYPER_ITC_D + ;; + .exitm + .endif + + mov \clob = r8 + mov r8 = \reg + ;; + (\pred_i)XEN_HYPER_ITC_I + ;; + (\pred_d)XEN_HYPER_ITC_D + ;; + mov r8 = \clob + ;; +.endm +#define ITC_I_AND_D(pred_i, pred_d, reg, clob) \ + __ITC_I_AND_D pred_i, pred_d, reg, clob + +.macro __THASH pred, reg0, reg1, clob + .ifc "\reg0", "r8" + (\pred) mov r8 = \reg1 + (\pred) XEN_HYPER_THASH + .exitm + .endc + .ifc "\reg1", "r8" + (\pred) XEN_HYPER_THASH + ;; + (\pred) mov \reg0 = r8 + ;; + .exitm + .endif + .ifc "\clob", "r8" + (\pred) mov r8 = \reg1 + (\pred) XEN_HYPER_THASH + ;; + (\pred) mov \reg0 = r8 + ;; + .exitm + .endif + + (\pred) mov \clob = r8 + (\pred) mov r8 = \reg1 + (\pred) XEN_HYPER_THASH + ;; + (\pred) mov \reg0 = r8 + (\pred) mov r8 = \clob + ;; +.endm +#define THASH(pred, reg0, reg1, clob) __THASH pred, reg0, reg1, clob + +#define SSM_PSR_IC_AND_DEFAULT_BITS_AND_SRLZ_I(clob0, clob1) \ + mov clob0 = 1; \ + movl clob1 = XSI_PSR_IC; \ + ;; \ + st4 [clob1] = clob0 \ + ;; + +#define SSM_PSR_IC_AND_SRLZ_D(clob0, clob1) \ + ;; \ + srlz.d; \ + mov clob1 = 1; \ + movl clob0 = XSI_PSR_IC; \ + ;; \ + st4 [clob0] = clob1 + +#define RSM_PSR_IC(clob) \ + movl clob = XSI_PSR_IC; \ + ;; \ + st4 [clob] = r0; \ + ;; + +/* pred will be clobbered */ +#define MASK_TO_PEND_OFS (-1) +#define SSM_PSR_I(pred, pred_clob, clob) \ +(pred) movl clob = XSI_PSR_I_ADDR \ + ;; \ +(pred) ld8 clob = [clob] \ + ;; \ + /* if (pred) vpsr.i = 1 */ \ + /* if (pred) (vcpu->vcpu_info->evtchn_upcall_mask)=0 */ \ +(pred) st1 [clob] = r0, MASK_TO_PEND_OFS \ + ;; \ + /* if (vcpu->vcpu_info->evtchn_upcall_pending) */ \ +(pred) ld1 clob = [clob] \ + ;; \ +(pred) cmp.ne.unc pred_clob, p0 = clob, r0 \ + ;; \ +(pred_clob)XEN_HYPER_SSM_I /* do areal ssm psr.i */ + +#define RSM_PSR_I(pred, clob0, clob1) \ + movl clob0 = XSI_PSR_I_ADDR; \ + mov clob1 = 1; \ + ;; \ + ld8 clob0 = [clob0]; \ + ;; \ +(pred) st1 [clob0] = clob1 + +#define RSM_PSR_I_IC(clob0, clob1, clob2) \ + movl clob0 = XSI_PSR_I_ADDR; \ + movl clob1 = XSI_PSR_IC; \ + ;; \ + ld8 clob0 = [clob0]; \ + mov clob2 = 1; \ + ;; \ + /* note: clears both vpsr.i and vpsr.ic! */ \ + st1 [clob0] = clob2; \ + st4 [clob1] = r0; \ + ;; + +#define RSM_PSR_DT \ + XEN_HYPER_RSM_PSR_DT + +#define SSM_PSR_DT_AND_SRLZ_I \ + XEN_HYPER_SSM_PSR_DT + +#define BSW_0(clob0, clob1, clob2) \ + ;; \ + /* r16-r31 all now hold bank1 values */ \ + mov clob2 = ar.unat; \ + movl clob0 = XSI_BANK1_R16; \ + movl clob1 = XSI_BANK1_R16 + 8; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r16, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r17, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r18, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r19, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r20, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r21, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r22, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r23, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r24, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r25, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r26, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r27, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r28, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r29, 16; \ + ;; \ +.mem.offset 0, 0; st8.spill [clob0] = r30, 16; \ +.mem.offset 8, 0; st8.spill [clob1] = r31, 16; \ + ;; \ + mov clob1 = ar.unat; \ + movl clob0 = XSI_B1NAT; \ + ;; \ + st8 [clob0] = clob1; \ + mov ar.unat = clob2; \ + movl clob0 = XSI_BANKNUM; \ + ;; \ + st4 [clob0] = r0 + + + /* FIXME: THIS CODE IS NOT NaT SAFE! */ +#define XEN_BSW_1(clob) \ + mov clob = ar.unat; \ + movl r30 = XSI_B1NAT; \ + ;; \ + ld8 r30 = [r30]; \ + mov r31 = 1; \ + ;; \ + mov ar.unat = r30; \ + movl r30 = XSI_BANKNUM; \ + ;; \ + st4 [r30] = r31; \ + movl r30 = XSI_BANK1_R16; \ + movl r31 = XSI_BANK1_R16+8; \ + ;; \ + ld8.fill r16 = [r30], 16; \ + ld8.fill r17 = [r31], 16; \ + ;; \ + ld8.fill r18 = [r30], 16; \ + ld8.fill r19 = [r31], 16; \ + ;; \ + ld8.fill r20 = [r30], 16; \ + ld8.fill r21 = [r31], 16; \ + ;; \ + ld8.fill r22 = [r30], 16; \ + ld8.fill r23 = [r31], 16; \ + ;; \ + ld8.fill r24 = [r30], 16; \ + ld8.fill r25 = [r31], 16; \ + ;; \ + ld8.fill r26 = [r30], 16; \ + ld8.fill r27 = [r31], 16; \ + ;; \ + ld8.fill r28 = [r30], 16; \ + ld8.fill r29 = [r31], 16; \ + ;; \ + ld8.fill r30 = [r30]; \ + ld8.fill r31 = [r31]; \ + ;; \ + mov ar.unat = clob + +#define BSW_1(clob0, clob1) XEN_BSW_1(clob1) + + +#define COVER \ + XEN_HYPER_COVER + +#define RFI \ + XEN_HYPER_RFI; \ + dv_serialize_data diff --git a/arch/ia64/include/asm/xen/interface.h b/arch/ia64/include/asm/xen/interface.h new file mode 100644 index 0000000..f00fab40 --- /dev/null +++ b/arch/ia64/include/asm/xen/interface.h @@ -0,0 +1,346 @@ +/****************************************************************************** + * arch-ia64/hypervisor-if.h + * + * Guest OS interface to IA64 Xen. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Copyright by those who contributed. (in alphabetical order) + * + * Anthony Xu <anthony.xu@intel.com> + * Eddie Dong <eddie.dong@intel.com> + * Fred Yang <fred.yang@intel.com> + * Kevin Tian <kevin.tian@intel.com> + * Alex Williamson <alex.williamson@hp.com> + * Chris Wright <chrisw@sous-sol.org> + * Christian Limpach <Christian.Limpach@cl.cam.ac.uk> + * Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com> + * Hollis Blanchard <hollisb@us.ibm.com> + * Isaku Yamahata <yamahata@valinux.co.jp> + * Jan Beulich <jbeulich@novell.com> + * John Levon <john.levon@sun.com> + * Kazuhiro Suzuki <kaz@jp.fujitsu.com> + * Keir Fraser <keir.fraser@citrix.com> + * Kouya Shimura <kouya@jp.fujitsu.com> + * Masaki Kanno <kanno.masaki@jp.fujitsu.com> + * Matt Chapman <matthewc@hp.com> + * Matthew Chapman <matthewc@hp.com> + * Samuel Thibault <samuel.thibault@eu.citrix.com> + * Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com> + * Tristan Gingold <tgingold@free.fr> + * Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com> + * Yutaka Ezaki <yutaka.ezaki@jp.fujitsu.com> + * Zhang Xin <xing.z.zhang@intel.com> + * Zhang xiantao <xiantao.zhang@intel.com> + * dan.magenheimer@hp.com + * ian.pratt@cl.cam.ac.uk + * michael.fetterman@cl.cam.ac.uk + */ + +#ifndef _ASM_IA64_XEN_INTERFACE_H +#define _ASM_IA64_XEN_INTERFACE_H + +#define __DEFINE_GUEST_HANDLE(name, type) \ + typedef struct { type *p; } __guest_handle_ ## name + +#define DEFINE_GUEST_HANDLE_STRUCT(name) \ + __DEFINE_GUEST_HANDLE(name, struct name) +#define DEFINE_GUEST_HANDLE(name) __DEFINE_GUEST_HANDLE(name, name) +#define GUEST_HANDLE(name) __guest_handle_ ## name +#define GUEST_HANDLE_64(name) GUEST_HANDLE(name) +#define set_xen_guest_handle(hnd, val) do { (hnd).p = val; } while (0) + +#ifndef __ASSEMBLY__ +/* Guest handles for primitive C types. */ +__DEFINE_GUEST_HANDLE(uchar, unsigned char); +__DEFINE_GUEST_HANDLE(uint, unsigned int); +__DEFINE_GUEST_HANDLE(ulong, unsigned long); +__DEFINE_GUEST_HANDLE(u64, unsigned long); +DEFINE_GUEST_HANDLE(char); +DEFINE_GUEST_HANDLE(int); +DEFINE_GUEST_HANDLE(long); +DEFINE_GUEST_HANDLE(void); + +typedef unsigned long xen_pfn_t; +DEFINE_GUEST_HANDLE(xen_pfn_t); +#define PRI_xen_pfn "lx" +#endif + +/* Arch specific VIRQs definition */ +#define VIRQ_ITC VIRQ_ARCH_0 /* V. Virtual itc timer */ +#define VIRQ_MCA_CMC VIRQ_ARCH_1 /* MCA cmc interrupt */ +#define VIRQ_MCA_CPE VIRQ_ARCH_2 /* MCA cpe interrupt */ + +/* Maximum number of virtual CPUs in multi-processor guests. */ +/* keep sizeof(struct shared_page) <= PAGE_SIZE. + * this is checked in arch/ia64/xen/hypervisor.c. */ +#define MAX_VIRT_CPUS 64 + +#ifndef __ASSEMBLY__ + +#define INVALID_MFN (~0UL) + +union vac { + unsigned long value; + struct { + int a_int:1; + int a_from_int_cr:1; + int a_to_int_cr:1; + int a_from_psr:1; + int a_from_cpuid:1; + int a_cover:1; + int a_bsw:1; + long reserved:57; + }; +}; + +union vdc { + unsigned long value; + struct { + int d_vmsw:1; + int d_extint:1; + int d_ibr_dbr:1; + int d_pmc:1; + int d_to_pmd:1; + int d_itm:1; + long reserved:58; + }; +}; + +struct mapped_regs { + union vac vac; + union vdc vdc; + unsigned long virt_env_vaddr; + unsigned long reserved1[29]; + unsigned long vhpi; + unsigned long reserved2[95]; + union { + unsigned long vgr[16]; + unsigned long bank1_regs[16]; /* bank1 regs (r16-r31) + when bank0 active */ + }; + union { + unsigned long vbgr[16]; + unsigned long bank0_regs[16]; /* bank0 regs (r16-r31) + when bank1 active */ + }; + unsigned long vnat; + unsigned long vbnat; + unsigned long vcpuid[5]; + unsigned long reserved3[11]; + unsigned long vpsr; + unsigned long vpr; + unsigned long reserved4[76]; + union { + unsigned long vcr[128]; + struct { + unsigned long dcr; /* CR0 */ + unsigned long itm; + unsigned long iva; + unsigned long rsv1[5]; + unsigned long pta; /* CR8 */ + unsigned long rsv2[7]; + unsigned long ipsr; /* CR16 */ + unsigned long isr; + unsigned long rsv3; + unsigned long iip; + unsigned long ifa; + unsigned long itir; + unsigned long iipa; + unsigned long ifs; + unsigned long iim; /* CR24 */ + unsigned long iha; + unsigned long rsv4[38]; + unsigned long lid; /* CR64 */ + unsigned long ivr; + unsigned long tpr; + unsigned long eoi; + unsigned long irr[4]; + unsigned long itv; /* CR72 */ + unsigned long pmv; + unsigned long cmcv; + unsigned long rsv5[5]; + unsigned long lrr0; /* CR80 */ + unsigned long lrr1; + unsigned long rsv6[46]; + }; + }; + union { + unsigned long reserved5[128]; + struct { + unsigned long precover_ifs; + unsigned long unat; /* not sure if this is needed + until NaT arch is done */ + int interrupt_collection_enabled; /* virtual psr.ic */ + + /* virtual interrupt deliverable flag is + * evtchn_upcall_mask in shared info area now. + * interrupt_mask_addr is the address + * of evtchn_upcall_mask for current vcpu + */ + unsigned char *interrupt_mask_addr; + int pending_interruption; + unsigned char vpsr_pp; + unsigned char vpsr_dfh; + unsigned char hpsr_dfh; + unsigned char hpsr_mfh; + unsigned long reserved5_1[4]; + int metaphysical_mode; /* 1 = use metaphys mapping + 0 = use virtual */ + int banknum; /* 0 or 1, which virtual + register bank is active */ + unsigned long rrs[8]; /* region registers */ + unsigned long krs[8]; /* kernel registers */ + unsigned long tmp[16]; /* temp registers + (e.g. for hyperprivops) */ + }; + }; +}; + +struct arch_vcpu_info { + /* nothing */ +}; + +/* + * This structure is used for magic page in domain pseudo physical address + * space and the result of XENMEM_machine_memory_map. + * As the XENMEM_machine_memory_map result, + * xen_memory_map::nr_entries indicates the size in bytes + * including struct xen_ia64_memmap_info. Not the number of entries. + */ +struct xen_ia64_memmap_info { + uint64_t efi_memmap_size; /* size of EFI memory map */ + uint64_t efi_memdesc_size; /* size of an EFI memory map + * descriptor */ + uint32_t efi_memdesc_version; /* memory descriptor version */ + void *memdesc[0]; /* array of efi_memory_desc_t */ +}; + +struct arch_shared_info { + /* PFN of the start_info page. */ + unsigned long start_info_pfn; + + /* Interrupt vector for event channel. */ + int evtchn_vector; + + /* PFN of memmap_info page */ + unsigned int memmap_info_num_pages; /* currently only = 1 case is + supported. */ + unsigned long memmap_info_pfn; + + uint64_t pad[31]; +}; + +struct xen_callback { + unsigned long ip; +}; +typedef struct xen_callback xen_callback_t; + +#endif /* !__ASSEMBLY__ */ + +/* Size of the shared_info area (this is not related to page size). */ +#define XSI_SHIFT 14 +#define XSI_SIZE (1 << XSI_SHIFT) +/* Log size of mapped_regs area (64 KB - only 4KB is used). */ +#define XMAPPEDREGS_SHIFT 12 +#define XMAPPEDREGS_SIZE (1 << XMAPPEDREGS_SHIFT) +/* Offset of XASI (Xen arch shared info) wrt XSI_BASE. */ +#define XMAPPEDREGS_OFS XSI_SIZE + +/* Hyperprivops. */ +#define HYPERPRIVOP_START 0x1 +#define HYPERPRIVOP_RFI (HYPERPRIVOP_START + 0x0) +#define HYPERPRIVOP_RSM_DT (HYPERPRIVOP_START + 0x1) +#define HYPERPRIVOP_SSM_DT (HYPERPRIVOP_START + 0x2) +#define HYPERPRIVOP_COVER (HYPERPRIVOP_START + 0x3) +#define HYPERPRIVOP_ITC_D (HYPERPRIVOP_START + 0x4) +#define HYPERPRIVOP_ITC_I (HYPERPRIVOP_START + 0x5) +#define HYPERPRIVOP_SSM_I (HYPERPRIVOP_START + 0x6) +#define HYPERPRIVOP_GET_IVR (HYPERPRIVOP_START + 0x7) +#define HYPERPRIVOP_GET_TPR (HYPERPRIVOP_START + 0x8) +#define HYPERPRIVOP_SET_TPR (HYPERPRIVOP_START + 0x9) +#define HYPERPRIVOP_EOI (HYPERPRIVOP_START + 0xa) +#define HYPERPRIVOP_SET_ITM (HYPERPRIVOP_START + 0xb) +#define HYPERPRIVOP_THASH (HYPERPRIVOP_START + 0xc) +#define HYPERPRIVOP_PTC_GA (HYPERPRIVOP_START + 0xd) +#define HYPERPRIVOP_ITR_D (HYPERPRIVOP_START + 0xe) +#define HYPERPRIVOP_GET_RR (HYPERPRIVOP_START + 0xf) +#define HYPERPRIVOP_SET_RR (HYPERPRIVOP_START + 0x10) +#define HYPERPRIVOP_SET_KR (HYPERPRIVOP_START + 0x11) +#define HYPERPRIVOP_FC (HYPERPRIVOP_START + 0x12) +#define HYPERPRIVOP_GET_CPUID (HYPERPRIVOP_START + 0x13) +#define HYPERPRIVOP_GET_PMD (HYPERPRIVOP_START + 0x14) +#define HYPERPRIVOP_GET_EFLAG (HYPERPRIVOP_START + 0x15) +#define HYPERPRIVOP_SET_EFLAG (HYPERPRIVOP_START + 0x16) +#define HYPERPRIVOP_RSM_BE (HYPERPRIVOP_START + 0x17) +#define HYPERPRIVOP_GET_PSR (HYPERPRIVOP_START + 0x18) +#define HYPERPRIVOP_SET_RR0_TO_RR4 (HYPERPRIVOP_START + 0x19) +#define HYPERPRIVOP_MAX (0x1a) + +/* Fast and light hypercalls. */ +#define __HYPERVISOR_ia64_fast_eoi __HYPERVISOR_arch_1 + +/* Xencomm macros. */ +#define XENCOMM_INLINE_MASK 0xf800000000000000UL +#define XENCOMM_INLINE_FLAG 0x8000000000000000UL + +#ifndef __ASSEMBLY__ + +/* + * Optimization features. + * The hypervisor may do some special optimizations for guests. This hypercall + * can be used to switch on/of these special optimizations. + */ +#define __HYPERVISOR_opt_feature 0x700UL + +#define XEN_IA64_OPTF_OFF 0x0 +#define XEN_IA64_OPTF_ON 0x1 + +/* + * If this feature is switched on, the hypervisor inserts the + * tlb entries without calling the guests traphandler. + * This is useful in guests using region 7 for identity mapping + * like the linux kernel does. + */ +#define XEN_IA64_OPTF_IDENT_MAP_REG7 1 + +/* Identity mapping of region 4 addresses in HVM. */ +#define XEN_IA64_OPTF_IDENT_MAP_REG4 2 + +/* Identity mapping of region 5 addresses in HVM. */ +#define XEN_IA64_OPTF_IDENT_MAP_REG5 3 + +#define XEN_IA64_OPTF_IDENT_MAP_NOT_SET (0) + +struct xen_ia64_opt_feature { + unsigned long cmd; /* Which feature */ + unsigned char on; /* Switch feature on/off */ + union { + struct { + /* The page protection bit mask of the pte. + * This will be or'ed with the pte. */ + unsigned long pgprot; + unsigned long key; /* A protection key for itir.*/ + }; + }; +}; + +#endif /* __ASSEMBLY__ */ + +#endif /* _ASM_IA64_XEN_INTERFACE_H */ diff --git a/arch/ia64/include/asm/xen/irq.h b/arch/ia64/include/asm/xen/irq.h new file mode 100644 index 0000000..a904509 --- /dev/null +++ b/arch/ia64/include/asm/xen/irq.h @@ -0,0 +1,44 @@ +/****************************************************************************** + * arch/ia64/include/asm/xen/irq.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#ifndef _ASM_IA64_XEN_IRQ_H +#define _ASM_IA64_XEN_IRQ_H + +/* + * The flat IRQ space is divided into two regions: + * 1. A one-to-one mapping of real physical IRQs. This space is only used + * if we have physical device-access privilege. This region is at the + * start of the IRQ space so that existing device drivers do not need + * to be modified to translate physical IRQ numbers into our IRQ space. + * 3. A dynamic mapping of inter-domain and Xen-sourced virtual IRQs. These + * are bound using the provided bind/unbind functions. + */ + +#define XEN_PIRQ_BASE 0 +#define XEN_NR_PIRQS 256 + +#define XEN_DYNIRQ_BASE (XEN_PIRQ_BASE + XEN_NR_PIRQS) +#define XEN_NR_DYNIRQS (NR_CPUS * 8) + +#define XEN_NR_IRQS (XEN_NR_PIRQS + XEN_NR_DYNIRQS) + +#endif /* _ASM_IA64_XEN_IRQ_H */ diff --git a/arch/ia64/include/asm/xen/minstate.h b/arch/ia64/include/asm/xen/minstate.h new file mode 100644 index 0000000..4d92d9b --- /dev/null +++ b/arch/ia64/include/asm/xen/minstate.h @@ -0,0 +1,134 @@ +/* + * DO_SAVE_MIN switches to the kernel stacks (if necessary) and saves + * the minimum state necessary that allows us to turn psr.ic back + * on. + * + * Assumed state upon entry: + * psr.ic: off + * r31: contains saved predicates (pr) + * + * Upon exit, the state is as follows: + * psr.ic: off + * r2 = points to &pt_regs.r16 + * r8 = contents of ar.ccv + * r9 = contents of ar.csd + * r10 = contents of ar.ssd + * r11 = FPSR_DEFAULT + * r12 = kernel sp (kernel virtual address) + * r13 = points to current task_struct (kernel virtual address) + * p15 = TRUE if psr.i is set in cr.ipsr + * predicate registers (other than p2, p3, and p15), b6, r3, r14, r15: + * preserved + * CONFIG_XEN note: p6/p7 are not preserved + * + * Note that psr.ic is NOT turned on by this macro. This is so that + * we can pass interruption state as arguments to a handler. + */ +#define XEN_DO_SAVE_MIN(__COVER,SAVE_IFS,EXTRA,WORKAROUND) \ + mov r16=IA64_KR(CURRENT); /* M */ \ + mov r27=ar.rsc; /* M */ \ + mov r20=r1; /* A */ \ + mov r25=ar.unat; /* M */ \ + MOV_FROM_IPSR(p0,r29); /* M */ \ + MOV_FROM_IIP(r28); /* M */ \ + mov r21=ar.fpsr; /* M */ \ + mov r26=ar.pfs; /* I */ \ + __COVER; /* B;; (or nothing) */ \ + adds r16=IA64_TASK_THREAD_ON_USTACK_OFFSET,r16; \ + ;; \ + ld1 r17=[r16]; /* load current->thread.on_ustack flag */ \ + st1 [r16]=r0; /* clear current->thread.on_ustack flag */ \ + adds r1=-IA64_TASK_THREAD_ON_USTACK_OFFSET,r16 \ + /* switch from user to kernel RBS: */ \ + ;; \ + invala; /* M */ \ + /* SAVE_IFS;*/ /* see xen special handling below */ \ + cmp.eq pKStk,pUStk=r0,r17; /* are we in kernel mode already? */ \ + ;; \ +(pUStk) mov ar.rsc=0; /* set enforced lazy mode, pl 0, little-endian, loadrs=0 */ \ + ;; \ +(pUStk) mov.m r24=ar.rnat; \ +(pUStk) addl r22=IA64_RBS_OFFSET,r1; /* compute base of RBS */ \ +(pKStk) mov r1=sp; /* get sp */ \ + ;; \ +(pUStk) lfetch.fault.excl.nt1 [r22]; \ +(pUStk) addl r1=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r1; /* compute base of memory stack */ \ +(pUStk) mov r23=ar.bspstore; /* save ar.bspstore */ \ + ;; \ +(pUStk) mov ar.bspstore=r22; /* switch to kernel RBS */ \ +(pKStk) addl r1=-IA64_PT_REGS_SIZE,r1; /* if in kernel mode, use sp (r12) */ \ + ;; \ +(pUStk) mov r18=ar.bsp; \ +(pUStk) mov ar.rsc=0x3; /* set eager mode, pl 0, little-endian, loadrs=0 */ \ + adds r17=2*L1_CACHE_BYTES,r1; /* really: biggest cache-line size */ \ + adds r16=PT(CR_IPSR),r1; \ + ;; \ + lfetch.fault.excl.nt1 [r17],L1_CACHE_BYTES; \ + st8 [r16]=r29; /* save cr.ipsr */ \ + ;; \ + lfetch.fault.excl.nt1 [r17]; \ + tbit.nz p15,p0=r29,IA64_PSR_I_BIT; \ + mov r29=b0 \ + ;; \ + WORKAROUND; \ + adds r16=PT(R8),r1; /* initialize first base pointer */ \ + adds r17=PT(R9),r1; /* initialize second base pointer */ \ +(pKStk) mov r18=r0; /* make sure r18 isn't NaT */ \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r8,16; \ +.mem.offset 8,0; st8.spill [r17]=r9,16; \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r10,24; \ + movl r8=XSI_PRECOVER_IFS; \ +.mem.offset 8,0; st8.spill [r17]=r11,24; \ + ;; \ + /* xen special handling for possibly lazy cover */ \ + /* SAVE_MIN case in dispatch_ia32_handler: mov r30=r0 */ \ + ld8 r30=[r8]; \ +(pUStk) sub r18=r18,r22; /* r18=RSE.ndirty*8 */ \ + st8 [r16]=r28,16; /* save cr.iip */ \ + ;; \ + st8 [r17]=r30,16; /* save cr.ifs */ \ + mov r8=ar.ccv; \ + mov r9=ar.csd; \ + mov r10=ar.ssd; \ + movl r11=FPSR_DEFAULT; /* L-unit */ \ + ;; \ + st8 [r16]=r25,16; /* save ar.unat */ \ + st8 [r17]=r26,16; /* save ar.pfs */ \ + shl r18=r18,16; /* compute ar.rsc to be used for "loadrs" */ \ + ;; \ + st8 [r16]=r27,16; /* save ar.rsc */ \ +(pUStk) st8 [r17]=r24,16; /* save ar.rnat */ \ +(pKStk) adds r17=16,r17; /* skip over ar_rnat field */ \ + ;; /* avoid RAW on r16 & r17 */ \ +(pUStk) st8 [r16]=r23,16; /* save ar.bspstore */ \ + st8 [r17]=r31,16; /* save predicates */ \ +(pKStk) adds r16=16,r16; /* skip over ar_bspstore field */ \ + ;; \ + st8 [r16]=r29,16; /* save b0 */ \ + st8 [r17]=r18,16; /* save ar.rsc value for "loadrs" */ \ + cmp.eq pNonSys,pSys=r0,r0 /* initialize pSys=0, pNonSys=1 */ \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r20,16; /* save original r1 */ \ +.mem.offset 8,0; st8.spill [r17]=r12,16; \ + adds r12=-16,r1; /* switch to kernel memory stack (with 16 bytes of scratch) */ \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r13,16; \ +.mem.offset 8,0; st8.spill [r17]=r21,16; /* save ar.fpsr */ \ + mov r13=IA64_KR(CURRENT); /* establish `current' */ \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r15,16; \ +.mem.offset 8,0; st8.spill [r17]=r14,16; \ + ;; \ +.mem.offset 0,0; st8.spill [r16]=r2,16; \ +.mem.offset 8,0; st8.spill [r17]=r3,16; \ + ACCOUNT_GET_STAMP \ + adds r2=IA64_PT_REGS_R16_OFFSET,r1; \ + ;; \ + EXTRA; \ + movl r1=__gp; /* establish kernel global pointer */ \ + ;; \ + ACCOUNT_SYS_ENTER \ + BSW_1(r3,r14); /* switch back to bank 1 (must be last in insn group) */ \ + ;; diff --git a/arch/ia64/include/asm/xen/page.h b/arch/ia64/include/asm/xen/page.h new file mode 100644 index 0000000..03441a78 --- /dev/null +++ b/arch/ia64/include/asm/xen/page.h @@ -0,0 +1,65 @@ +/****************************************************************************** + * arch/ia64/include/asm/xen/page.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#ifndef _ASM_IA64_XEN_PAGE_H +#define _ASM_IA64_XEN_PAGE_H + +#define INVALID_P2M_ENTRY (~0UL) + +static inline unsigned long mfn_to_pfn(unsigned long mfn) +{ + return mfn; +} + +static inline unsigned long pfn_to_mfn(unsigned long pfn) +{ + return pfn; +} + +#define phys_to_machine_mapping_valid(_x) (1) + +static inline void *mfn_to_virt(unsigned long mfn) +{ + return __va(mfn << PAGE_SHIFT); +} + +static inline unsigned long virt_to_mfn(void *virt) +{ + return __pa(virt) >> PAGE_SHIFT; +} + +/* for tpmfront.c */ +static inline unsigned long virt_to_machine(void *virt) +{ + return __pa(virt); +} + +static inline void set_phys_to_machine(unsigned long pfn, unsigned long mfn) +{ + /* nothing */ +} + +#define pte_mfn(_x) pte_pfn(_x) +#define mfn_pte(_x, _y) __pte_ma(0) /* unmodified use */ +#define __pte_ma(_x) ((pte_t) {(_x)}) /* unmodified use */ + +#endif /* _ASM_IA64_XEN_PAGE_H */ diff --git a/arch/ia64/include/asm/xen/privop.h b/arch/ia64/include/asm/xen/privop.h new file mode 100644 index 0000000..71ec754 --- /dev/null +++ b/arch/ia64/include/asm/xen/privop.h @@ -0,0 +1,129 @@ +#ifndef _ASM_IA64_XEN_PRIVOP_H +#define _ASM_IA64_XEN_PRIVOP_H + +/* + * Copyright (C) 2005 Hewlett-Packard Co + * Dan Magenheimer <dan.magenheimer@hp.com> + * + * Paravirtualizations of privileged operations for Xen/ia64 + * + * + * inline privop and paravirt_alt support + * Copyright (c) 2007 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + */ + +#ifndef __ASSEMBLY__ +#include <linux/types.h> /* arch-ia64.h requires uint64_t */ +#endif +#include <asm/xen/interface.h> + +/* At 1 MB, before per-cpu space but still addressable using addl instead + of movl. */ +#define XSI_BASE 0xfffffffffff00000 + +/* Address of mapped regs. */ +#define XMAPPEDREGS_BASE (XSI_BASE + XSI_SIZE) + +#ifdef __ASSEMBLY__ +#define XEN_HYPER_RFI break HYPERPRIVOP_RFI +#define XEN_HYPER_RSM_PSR_DT break HYPERPRIVOP_RSM_DT +#define XEN_HYPER_SSM_PSR_DT break HYPERPRIVOP_SSM_DT +#define XEN_HYPER_COVER break HYPERPRIVOP_COVER +#define XEN_HYPER_ITC_D break HYPERPRIVOP_ITC_D +#define XEN_HYPER_ITC_I break HYPERPRIVOP_ITC_I +#define XEN_HYPER_SSM_I break HYPERPRIVOP_SSM_I +#define XEN_HYPER_GET_IVR break HYPERPRIVOP_GET_IVR +#define XEN_HYPER_THASH break HYPERPRIVOP_THASH +#define XEN_HYPER_ITR_D break HYPERPRIVOP_ITR_D +#define XEN_HYPER_SET_KR break HYPERPRIVOP_SET_KR +#define XEN_HYPER_GET_PSR break HYPERPRIVOP_GET_PSR +#define XEN_HYPER_SET_RR0_TO_RR4 break HYPERPRIVOP_SET_RR0_TO_RR4 + +#define XSI_IFS (XSI_BASE + XSI_IFS_OFS) +#define XSI_PRECOVER_IFS (XSI_BASE + XSI_PRECOVER_IFS_OFS) +#define XSI_IFA (XSI_BASE + XSI_IFA_OFS) +#define XSI_ISR (XSI_BASE + XSI_ISR_OFS) +#define XSI_IIM (XSI_BASE + XSI_IIM_OFS) +#define XSI_ITIR (XSI_BASE + XSI_ITIR_OFS) +#define XSI_PSR_I_ADDR (XSI_BASE + XSI_PSR_I_ADDR_OFS) +#define XSI_PSR_IC (XSI_BASE + XSI_PSR_IC_OFS) +#define XSI_IPSR (XSI_BASE + XSI_IPSR_OFS) +#define XSI_IIP (XSI_BASE + XSI_IIP_OFS) +#define XSI_B1NAT (XSI_BASE + XSI_B1NATS_OFS) +#define XSI_BANK1_R16 (XSI_BASE + XSI_BANK1_R16_OFS) +#define XSI_BANKNUM (XSI_BASE + XSI_BANKNUM_OFS) +#define XSI_IHA (XSI_BASE + XSI_IHA_OFS) +#endif + +#ifndef __ASSEMBLY__ + +/************************************************/ +/* Instructions paravirtualized for correctness */ +/************************************************/ + +/* "fc" and "thash" are privilege-sensitive instructions, meaning they + * may have different semantics depending on whether they are executed + * at PL0 vs PL!=0. When paravirtualized, these instructions mustn't + * be allowed to execute directly, lest incorrect semantics result. */ +extern void xen_fc(unsigned long addr); +extern unsigned long xen_thash(unsigned long addr); + +/* Note that "ttag" and "cover" are also privilege-sensitive; "ttag" + * is not currently used (though it may be in a long-format VHPT system!) + * and the semantics of cover only change if psr.ic is off which is very + * rare (and currently non-existent outside of assembly code */ + +/* There are also privilege-sensitive registers. These registers are + * readable at any privilege level but only writable at PL0. */ +extern unsigned long xen_get_cpuid(int index); +extern unsigned long xen_get_pmd(int index); + +extern unsigned long xen_get_eflag(void); /* see xen_ia64_getreg */ +extern void xen_set_eflag(unsigned long); /* see xen_ia64_setreg */ + +/************************************************/ +/* Instructions paravirtualized for performance */ +/************************************************/ + +/* Xen uses memory-mapped virtual privileged registers for access to many + * performance-sensitive privileged registers. Some, like the processor + * status register (psr), are broken up into multiple memory locations. + * Others, like "pend", are abstractions based on privileged registers. + * "Pend" is guaranteed to be set if reading cr.ivr would return a + * (non-spurious) interrupt. */ +#define XEN_MAPPEDREGS ((struct mapped_regs *)XMAPPEDREGS_BASE) + +#define XSI_PSR_I \ + (*XEN_MAPPEDREGS->interrupt_mask_addr) +#define xen_get_virtual_psr_i() \ + (!XSI_PSR_I) +#define xen_set_virtual_psr_i(_val) \ + ({ XSI_PSR_I = (uint8_t)(_val) ? 0 : 1; }) +#define xen_set_virtual_psr_ic(_val) \ + ({ XEN_MAPPEDREGS->interrupt_collection_enabled = _val ? 1 : 0; }) +#define xen_get_virtual_pend() \ + (*(((uint8_t *)XEN_MAPPEDREGS->interrupt_mask_addr) - 1)) + +/* Although all privileged operations can be left to trap and will + * be properly handled by Xen, some are frequent enough that we use + * hyperprivops for performance. */ +extern unsigned long xen_get_psr(void); +extern unsigned long xen_get_ivr(void); +extern unsigned long xen_get_tpr(void); +extern void xen_hyper_ssm_i(void); +extern void xen_set_itm(unsigned long); +extern void xen_set_tpr(unsigned long); +extern void xen_eoi(unsigned long); +extern unsigned long xen_get_rr(unsigned long index); +extern void xen_set_rr(unsigned long index, unsigned long val); +extern void xen_set_rr0_to_rr4(unsigned long val0, unsigned long val1, + unsigned long val2, unsigned long val3, + unsigned long val4); +extern void xen_set_kr(unsigned long index, unsigned long val); +extern void xen_ptcga(unsigned long addr, unsigned long size); + +#endif /* !__ASSEMBLY__ */ + +#endif /* _ASM_IA64_XEN_PRIVOP_H */ diff --git a/arch/ia64/include/asm/xen/xcom_hcall.h b/arch/ia64/include/asm/xen/xcom_hcall.h new file mode 100644 index 0000000..20b2950 --- /dev/null +++ b/arch/ia64/include/asm/xen/xcom_hcall.h @@ -0,0 +1,51 @@ +/* + * Copyright (C) 2006 Tristan Gingold <tristan.gingold@bull.net>, Bull SAS + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#ifndef _ASM_IA64_XEN_XCOM_HCALL_H +#define _ASM_IA64_XEN_XCOM_HCALL_H + +/* These function creates inline or mini descriptor for the parameters and + calls the corresponding xencomm_arch_hypercall_X. + Architectures should defines HYPERVISOR_xxx as xencomm_hypercall_xxx unless + they want to use their own wrapper. */ +extern int xencomm_hypercall_console_io(int cmd, int count, char *str); + +extern int xencomm_hypercall_event_channel_op(int cmd, void *op); + +extern int xencomm_hypercall_xen_version(int cmd, void *arg); + +extern int xencomm_hypercall_physdev_op(int cmd, void *op); + +extern int xencomm_hypercall_grant_table_op(unsigned int cmd, void *op, + unsigned int count); + +extern int xencomm_hypercall_sched_op(int cmd, void *arg); + +extern int xencomm_hypercall_multicall(void *call_list, int nr_calls); + +extern int xencomm_hypercall_callback_op(int cmd, void *arg); + +extern int xencomm_hypercall_memory_op(unsigned int cmd, void *arg); + +extern int xencomm_hypercall_suspend(unsigned long srec); + +extern long xencomm_hypercall_vcpu_op(int cmd, int cpu, void *arg); + +extern long xencomm_hypercall_opt_feature(void *arg); + +#endif /* _ASM_IA64_XEN_XCOM_HCALL_H */ diff --git a/arch/ia64/include/asm/xen/xencomm.h b/arch/ia64/include/asm/xen/xencomm.h new file mode 100644 index 0000000..cded677 --- /dev/null +++ b/arch/ia64/include/asm/xen/xencomm.h @@ -0,0 +1,42 @@ +/* + * Copyright (C) 2006 Hollis Blanchard <hollisb@us.ibm.com>, IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#ifndef _ASM_IA64_XEN_XENCOMM_H +#define _ASM_IA64_XEN_XENCOMM_H + +#include <xen/xencomm.h> +#include <asm/pgtable.h> + +/* Must be called before any hypercall. */ +extern void xencomm_initialize(void); +extern int xencomm_is_initialized(void); + +/* Check if virtual contiguity means physical contiguity + * where the passed address is a pointer value in virtual address. + * On ia64, identity mapping area in region 7 or the piece of region 5 + * that is mapped by itr[IA64_TR_KERNEL]/dtr[IA64_TR_KERNEL] + */ +static inline int xencomm_is_phys_contiguous(unsigned long addr) +{ + return (PAGE_OFFSET <= addr && + addr < (PAGE_OFFSET + (1UL << IA64_MAX_PHYS_BITS))) || + (KERNEL_START <= addr && + addr < KERNEL_START + KERNEL_TR_PAGE_SIZE); +} + +#endif /* _ASM_IA64_XEN_XENCOMM_H */ diff --git a/arch/ia64/kernel/Makefile b/arch/ia64/kernel/Makefile index 87fea11..c381ea9 100644 --- a/arch/ia64/kernel/Makefile +++ b/arch/ia64/kernel/Makefile @@ -42,6 +42,10 @@ obj-$(CONFIG_IA64_ESI) += esi.o ifneq ($(CONFIG_IA64_ESI),) obj-y += esi_stub.o # must be in kernel proper endif +obj-$(CONFIG_DMAR) += pci-dma.o +ifeq ($(CONFIG_DMAR), y) +obj-$(CONFIG_SWIOTLB) += pci-swiotlb.o +endif # The gate DSO image is built using a special linker script. targets += gate.so gate-syms.o @@ -112,5 +116,23 @@ clean-files += $(objtree)/include/asm-ia64/nr-irqs.h ASM_PARAVIRT_OBJS = ivt.o entry.o define paravirtualized_native AFLAGS_$(1) += -D__IA64_ASM_PARAVIRTUALIZED_NATIVE +AFLAGS_pvchk-sed-$(1) += -D__IA64_ASM_PARAVIRTUALIZED_PVCHECK +extra-y += pvchk-$(1) endef $(foreach obj,$(ASM_PARAVIRT_OBJS),$(eval $(call paravirtualized_native,$(obj)))) + +# +# Checker for paravirtualizations of privileged operations. +# +quiet_cmd_pv_check_sed = PVCHK $@ +define cmd_pv_check_sed + sed -f $(srctree)/arch/$(SRCARCH)/scripts/pvcheck.sed $< > $@ +endef + +$(obj)/pvchk-sed-%.s: $(src)/%.S $(srctree)/arch/$(SRCARCH)/scripts/pvcheck.sed FORCE + $(call if_changed_dep,as_s_S) +$(obj)/pvchk-%.s: $(obj)/pvchk-sed-%.s FORCE + $(call if_changed,pv_check_sed) +$(obj)/pvchk-%.o: $(obj)/pvchk-%.s FORCE + $(call if_changed,as_o_S) +.PRECIOUS: $(obj)/pvchk-sed-%.s $(obj)/pvchk-%.s $(obj)/pvchk-%.o diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c index 5d1eb7e..0635015 100644 --- a/arch/ia64/kernel/acpi.c +++ b/arch/ia64/kernel/acpi.c @@ -52,6 +52,7 @@ #include <asm/numa.h> #include <asm/sal.h> #include <asm/cyclone.h> +#include <asm/xen/hypervisor.h> #define BAD_MADT_ENTRY(entry, end) ( \ (!entry) || (unsigned long)entry + sizeof(*entry) > end || \ @@ -91,6 +92,9 @@ acpi_get_sysname(void) struct acpi_table_rsdp *rsdp; struct acpi_table_xsdt *xsdt; struct acpi_table_header *hdr; +#ifdef CONFIG_DMAR + u64 i, nentries; +#endif rsdp_phys = acpi_find_rsdp(); if (!rsdp_phys) { @@ -121,7 +125,21 @@ acpi_get_sysname(void) return "uv"; else return "sn2"; + } else if (xen_pv_domain() && !strcmp(hdr->oem_id, "XEN")) { + return "xen"; + } + +#ifdef CONFIG_DMAR + /* Look for Intel IOMMU */ + nentries = (hdr->length - sizeof(*hdr)) / + sizeof(xsdt->table_offset_entry[0]); + for (i = 0; i < nentries; i++) { + hdr = __va(xsdt->table_offset_entry[i]); + if (strncmp(hdr->signature, ACPI_SIG_DMAR, + sizeof(ACPI_SIG_DMAR) - 1) == 0) + return "dig_vtd"; } +#endif return "dig"; #else @@ -137,6 +155,10 @@ acpi_get_sysname(void) return "uv"; # elif defined (CONFIG_IA64_DIG) return "dig"; +# elif defined (CONFIG_IA64_XEN_GUEST) + return "xen"; +# elif defined(CONFIG_IA64_DIG_VTD) + return "dig_vtd"; # else # error Unknown platform. Fix acpi.c. # endif diff --git a/arch/ia64/kernel/asm-offsets.c b/arch/ia64/kernel/asm-offsets.c index 94c44b1..742dbb1 100644 --- a/arch/ia64/kernel/asm-offsets.c +++ b/arch/ia64/kernel/asm-offsets.c @@ -16,6 +16,9 @@ #include <asm/sigcontext.h> #include <asm/mca.h> +#include <asm/xen/interface.h> +#include <asm/xen/hypervisor.h> + #include "../kernel/sigframe.h" #include "../kernel/fsyscall_gtod_data.h" @@ -286,4 +289,32 @@ void foo(void) offsetof (struct itc_jitter_data_t, itc_jitter)); DEFINE(IA64_ITC_LASTCYCLE_OFFSET, offsetof (struct itc_jitter_data_t, itc_lastcycle)); + +#ifdef CONFIG_XEN + BLANK(); + + DEFINE(XEN_NATIVE_ASM, XEN_NATIVE); + DEFINE(XEN_PV_DOMAIN_ASM, XEN_PV_DOMAIN); + +#define DEFINE_MAPPED_REG_OFS(sym, field) \ + DEFINE(sym, (XMAPPEDREGS_OFS + offsetof(struct mapped_regs, field))) + + DEFINE_MAPPED_REG_OFS(XSI_PSR_I_ADDR_OFS, interrupt_mask_addr); + DEFINE_MAPPED_REG_OFS(XSI_IPSR_OFS, ipsr); + DEFINE_MAPPED_REG_OFS(XSI_IIP_OFS, iip); + DEFINE_MAPPED_REG_OFS(XSI_IFS_OFS, ifs); + DEFINE_MAPPED_REG_OFS(XSI_PRECOVER_IFS_OFS, precover_ifs); + DEFINE_MAPPED_REG_OFS(XSI_ISR_OFS, isr); + DEFINE_MAPPED_REG_OFS(XSI_IFA_OFS, ifa); + DEFINE_MAPPED_REG_OFS(XSI_IIPA_OFS, iipa); + DEFINE_MAPPED_REG_OFS(XSI_IIM_OFS, iim); + DEFINE_MAPPED_REG_OFS(XSI_IHA_OFS, iha); + DEFINE_MAPPED_REG_OFS(XSI_ITIR_OFS, itir); + DEFINE_MAPPED_REG_OFS(XSI_PSR_IC_OFS, interrupt_collection_enabled); + DEFINE_MAPPED_REG_OFS(XSI_BANKNUM_OFS, banknum); + DEFINE_MAPPED_REG_OFS(XSI_BANK0_R16_OFS, bank0_regs[0]); + DEFINE_MAPPED_REG_OFS(XSI_BANK1_R16_OFS, bank1_regs[0]); + DEFINE_MAPPED_REG_OFS(XSI_B0NATS_OFS, vbnat); + DEFINE_MAPPED_REG_OFS(XSI_B1NATS_OFS, vnat); +#endif /* CONFIG_XEN */ } diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index 0dd6c14..7ef0c59 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -534,6 +534,11 @@ GLOBAL_ENTRY(ia64_trace_syscall) stf.spill [r16]=f10 stf.spill [r17]=f11 br.call.sptk.many rp=syscall_trace_enter // give parent a chance to catch syscall args + cmp.lt p6,p0=r8,r0 // check tracehook + adds r2=PT(R8)+16,sp // r2 = &pt_regs.r8 + adds r3=PT(R10)+16,sp // r3 = &pt_regs.r10 + mov r10=0 +(p6) br.cond.sptk strace_error // syscall failed -> adds r16=PT(F6)+16,sp adds r17=PT(F7)+16,sp ;; diff --git a/arch/ia64/kernel/ivt.S b/arch/ia64/kernel/ivt.S index 416a952..f675d8e3 100644 --- a/arch/ia64/kernel/ivt.S +++ b/arch/ia64/kernel/ivt.S @@ -580,7 +580,7 @@ ENTRY(dirty_bit) mov b0=r29 // restore b0 ;; st8 [r17]=r18 // store back updated PTE - itc.d r18 // install updated PTE + ITC_D(p0, r18, r16) // install updated PTE #endif mov pr=r31,-1 // restore pr RFI @@ -646,7 +646,7 @@ ENTRY(iaccess_bit) mov b0=r29 // restore b0 ;; st8 [r17]=r18 // store back updated PTE - itc.i r18 // install updated PTE + ITC_I(p0, r18, r16) // install updated PTE #endif /* !CONFIG_SMP */ mov pr=r31,-1 RFI @@ -698,7 +698,7 @@ ENTRY(daccess_bit) or r18=_PAGE_A,r18 // set the accessed bit ;; st8 [r17]=r18 // store back updated PTE - itc.d r18 // install updated PTE + ITC_D(p0, r18, r16) // install updated PTE #endif mov b0=r29 // restore b0 mov pr=r31,-1 diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c index 60c6ef6..702a09c 100644 --- a/arch/ia64/kernel/msi_ia64.c +++ b/arch/ia64/kernel/msi_ia64.c @@ -5,6 +5,7 @@ #include <linux/pci.h> #include <linux/irq.h> #include <linux/msi.h> +#include <linux/dmar.h> #include <asm/smp.h> /* @@ -162,3 +163,82 @@ void arch_teardown_msi_irq(unsigned int irq) return ia64_teardown_msi_irq(irq); } + +#ifdef CONFIG_DMAR +#ifdef CONFIG_SMP +static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) +{ + struct irq_cfg *cfg = irq_cfg + irq; + struct msi_msg msg; + int cpu = first_cpu(mask); + + + if (!cpu_online(cpu)) + return; + + if (irq_prepare_move(irq, cpu)) + return; + + dmar_msi_read(irq, &msg); + + msg.data &= ~MSI_DATA_VECTOR_MASK; + msg.data |= MSI_DATA_VECTOR(cfg->vector); + msg.address_lo &= ~MSI_ADDR_DESTID_MASK; + msg.address_lo |= MSI_ADDR_DESTID_CPU(cpu_physical_id(cpu)); + + dmar_msi_write(irq, &msg); + irq_desc[irq].affinity = mask; +} +#endif /* CONFIG_SMP */ + +struct irq_chip dmar_msi_type = { + .name = "DMAR_MSI", + .unmask = dmar_msi_unmask, + .mask = dmar_msi_mask, + .ack = ia64_ack_msi_irq, +#ifdef CONFIG_SMP + .set_affinity = dmar_msi_set_affinity, +#endif + .retrigger = ia64_msi_retrigger_irq, +}; + +static int +msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_msg *msg) +{ + struct irq_cfg *cfg = irq_cfg + irq; + unsigned dest; + cpumask_t mask; + + cpus_and(mask, irq_to_domain(irq), cpu_online_map); + dest = cpu_physical_id(first_cpu(mask)); + + msg->address_hi = 0; + msg->address_lo = + MSI_ADDR_HEADER | + MSI_ADDR_DESTMODE_PHYS | + MSI_ADDR_REDIRECTION_CPU | + MSI_ADDR_DESTID_CPU(dest); + + msg->data = + MSI_DATA_TRIGGER_EDGE | + MSI_DATA_LEVEL_ASSERT | + MSI_DATA_DELIVERY_FIXED | + MSI_DATA_VECTOR(cfg->vector); + return 0; +} + +int arch_setup_dmar_msi(unsigned int irq) +{ + int ret; + struct msi_msg msg; + + ret = msi_compose_msg(NULL, irq, &msg); + if (ret < 0) + return ret; + dmar_msi_write(irq, &msg); + set_irq_chip_and_handler_name(irq, &dmar_msi_type, handle_edge_irq, + "edge"); + return 0; +} +#endif /* CONFIG_DMAR */ + diff --git a/arch/ia64/kernel/nr-irqs.c b/arch/ia64/kernel/nr-irqs.c index 8273afc..ee56457 100644 --- a/arch/ia64/kernel/nr-irqs.c +++ b/arch/ia64/kernel/nr-irqs.c @@ -10,6 +10,7 @@ #include <linux/kbuild.h> #include <linux/threads.h> #include <asm/native/irq.h> +#include <asm/xen/irq.h> void foo(void) { diff --git a/arch/ia64/kernel/paravirt.c b/arch/ia64/kernel/paravirt.c index afaf5b9..de35d8e 100644 --- a/arch/ia64/kernel/paravirt.c +++ b/arch/ia64/kernel/paravirt.c @@ -332,7 +332,7 @@ ia64_native_iosapic_write(char __iomem *iosapic, unsigned int reg, u32 val) struct pv_iosapic_ops pv_iosapic_ops = { .pcat_compat_init = ia64_native_iosapic_pcat_compat_init, - .get_irq_chip = ia64_native_iosapic_get_irq_chip, + .__get_irq_chip = ia64_native_iosapic_get_irq_chip, .__read = ia64_native_iosapic_read, .__write = ia64_native_iosapic_write, diff --git a/arch/ia64/kernel/paravirt_inst.h b/arch/ia64/kernel/paravirt_inst.h index 5cad6fb..64d6d81 100644 --- a/arch/ia64/kernel/paravirt_inst.h +++ b/arch/ia64/kernel/paravirt_inst.h @@ -20,7 +20,9 @@ * */ -#ifdef __IA64_ASM_PARAVIRTUALIZED_XEN +#ifdef __IA64_ASM_PARAVIRTUALIZED_PVCHECK +#include <asm/native/pvchk_inst.h> +#elif defined(__IA64_ASM_PARAVIRTUALIZED_XEN) #include <asm/xen/inst.h> #include <asm/xen/minstate.h> #else diff --git a/arch/ia64/kernel/pci-dma.c b/arch/ia64/kernel/pci-dma.c new file mode 100644 index 0000000..10a75b5 --- /dev/null +++ b/arch/ia64/kernel/pci-dma.c @@ -0,0 +1,129 @@ +/* + * Dynamic DMA mapping support. + */ + +#include <linux/types.h> +#include <linux/mm.h> +#include <linux/string.h> +#include <linux/pci.h> +#include <linux/module.h> +#include <linux/dmar.h> +#include <asm/iommu.h> +#include <asm/machvec.h> +#include <linux/dma-mapping.h> + +#include <asm/machvec.h> +#include <asm/system.h> + +#ifdef CONFIG_DMAR + +#include <linux/kernel.h> +#include <linux/string.h> + +#include <asm/page.h> +#include <asm/iommu.h> + +dma_addr_t bad_dma_address __read_mostly; +EXPORT_SYMBOL(bad_dma_address); + +static int iommu_sac_force __read_mostly; + +int no_iommu __read_mostly; +#ifdef CONFIG_IOMMU_DEBUG +int force_iommu __read_mostly = 1; +#else +int force_iommu __read_mostly; +#endif + +/* Set this to 1 if there is a HW IOMMU in the system */ +int iommu_detected __read_mostly; + +/* Dummy device used for NULL arguments (normally ISA). Better would + be probably a smaller DMA mask, but this is bug-to-bug compatible + to i386. */ +struct device fallback_dev = { + .bus_id = "fallback device", + .coherent_dma_mask = DMA_32BIT_MASK, + .dma_mask = &fallback_dev.coherent_dma_mask, +}; + +void __init pci_iommu_alloc(void) +{ + /* + * The order of these functions is important for + * fall-back/fail-over reasons + */ + detect_intel_iommu(); + +#ifdef CONFIG_SWIOTLB + pci_swiotlb_init(); +#endif +} + +static int __init pci_iommu_init(void) +{ + if (iommu_detected) + intel_iommu_init(); + + return 0; +} + +/* Must execute after PCI subsystem */ +fs_initcall(pci_iommu_init); + +void pci_iommu_shutdown(void) +{ + return; +} + +void __init +iommu_dma_init(void) +{ + return; +} + +struct dma_mapping_ops *dma_ops; +EXPORT_SYMBOL(dma_ops); + +int iommu_dma_supported(struct device *dev, u64 mask) +{ + struct dma_mapping_ops *ops = get_dma_ops(dev); + +#ifdef CONFIG_PCI + if (mask > 0xffffffff && forbid_dac > 0) { + dev_info(dev, "Disallowing DAC for device\n"); + return 0; + } +#endif + + if (ops->dma_supported_op) + return ops->dma_supported_op(dev, mask); + + /* Copied from i386. Doesn't make much sense, because it will + only work for pci_alloc_coherent. + The caller just has to use GFP_DMA in this case. */ + if (mask < DMA_24BIT_MASK) + return 0; + + /* Tell the device to use SAC when IOMMU force is on. This + allows the driver to use cheaper accesses in some cases. + + Problem with this is that if we overflow the IOMMU area and + return DAC as fallback address the device may not handle it + correctly. + + As a special case some controllers have a 39bit address + mode that is as efficient as 32bit (aic79xx). Don't force + SAC for these. Assume all masks <= 40 bits are of this + type. Normally this doesn't make any difference, but gives + more gentle handling of IOMMU overflow. */ + if (iommu_sac_force && (mask >= DMA_40BIT_MASK)) { + dev_info(dev, "Force SAC with mask %lx\n", mask); + return 0; + } + + return 1; +} +EXPORT_SYMBOL(iommu_dma_supported); + +#endif diff --git a/arch/ia64/kernel/pci-swiotlb.c b/arch/ia64/kernel/pci-swiotlb.c new file mode 100644 index 0000000..16c5051 --- /dev/null +++ b/arch/ia64/kernel/pci-swiotlb.c @@ -0,0 +1,46 @@ +/* Glue code to lib/swiotlb.c */ + +#include <linux/pci.h> +#include <linux/cache.h> +#include <linux/module.h> +#include <linux/dma-mapping.h> + +#include <asm/swiotlb.h> +#include <asm/dma.h> +#include <asm/iommu.h> +#include <asm/machvec.h> + +int swiotlb __read_mostly; +EXPORT_SYMBOL(swiotlb); + +struct dma_mapping_ops swiotlb_dma_ops = { + .mapping_error = swiotlb_dma_mapping_error, + .alloc_coherent = swiotlb_alloc_coherent, + .free_coherent = swiotlb_free_coherent, + .map_single = swiotlb_map_single, + .unmap_single = swiotlb_unmap_single, + .sync_single_for_cpu = swiotlb_sync_single_for_cpu, + .sync_single_for_device = swiotlb_sync_single_for_device, + .sync_single_range_for_cpu = swiotlb_sync_single_range_for_cpu, + .sync_single_range_for_device = swiotlb_sync_single_range_for_device, + .sync_sg_for_cpu = swiotlb_sync_sg_for_cpu, + .sync_sg_for_device = swiotlb_sync_sg_for_device, + .map_sg = swiotlb_map_sg, + .unmap_sg = swiotlb_unmap_sg, + .dma_supported_op = swiotlb_dma_supported, +}; + +void __init pci_swiotlb_init(void) +{ + if (!iommu_detected) { +#ifdef CONFIG_IA64_GENERIC + swiotlb = 1; + printk(KERN_INFO "PCI-DMA: Re-initialize machine vector.\n"); + machvec_init("dig"); + swiotlb_init(); + dma_ops = &swiotlb_dma_ops; +#else + panic("Unable to find Intel IOMMU"); +#endif + } +} diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c index fc8f350..ada4605 100644 --- a/arch/ia64/kernel/perfmon.c +++ b/arch/ia64/kernel/perfmon.c @@ -40,6 +40,7 @@ #include <linux/capability.h> #include <linux/rcupdate.h> #include <linux/completion.h> +#include <linux/tracehook.h> #include <asm/errno.h> #include <asm/intrinsics.h> @@ -3684,7 +3685,7 @@ pfm_restart(pfm_context_t *ctx, void *arg, int count, struct pt_regs *regs) PFM_SET_WORK_PENDING(task, 1); - tsk_set_notify_resume(task); + set_notify_resume(task); /* * XXX: send reschedule if task runs on another CPU @@ -5044,8 +5045,6 @@ pfm_handle_work(void) PFM_SET_WORK_PENDING(current, 0); - tsk_clear_notify_resume(current); - regs = task_pt_regs(current); /* @@ -5414,7 +5413,7 @@ pfm_overflow_handler(struct task_struct *task, pfm_context_t *ctx, u64 pmc0, str * when coming from ctxsw, current still points to the * previous task, therefore we must work with task and not current. */ - tsk_set_notify_resume(task); + set_notify_resume(task); } /* * defer until state is changed (shorten spin window). the context is locked diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c index 3ab8373..c571627 100644 --- a/arch/ia64/kernel/process.c +++ b/arch/ia64/kernel/process.c @@ -28,6 +28,7 @@ #include <linux/delay.h> #include <linux/kdebug.h> #include <linux/utsname.h> +#include <linux/tracehook.h> #include <asm/cpu.h> #include <asm/delay.h> @@ -160,21 +161,6 @@ show_regs (struct pt_regs *regs) show_stack(NULL, NULL); } -void tsk_clear_notify_resume(struct task_struct *tsk) -{ -#ifdef CONFIG_PERFMON - if (tsk->thread.pfm_needs_checking) - return; -#endif - if (test_ti_thread_flag(task_thread_info(tsk), TIF_RESTORE_RSE)) - return; - clear_ti_thread_flag(task_thread_info(tsk), TIF_NOTIFY_RESUME); -} - -/* - * do_notify_resume_user(): - * Called from notify_resume_user at entry.S, with interrupts disabled. - */ void do_notify_resume_user(sigset_t *unused, struct sigscratch *scr, long in_syscall) { @@ -203,6 +189,11 @@ do_notify_resume_user(sigset_t *unused, struct sigscratch *scr, long in_syscall) ia64_do_signal(scr, in_syscall); } + if (test_thread_flag(TIF_NOTIFY_RESUME)) { + clear_thread_flag(TIF_NOTIFY_RESUME); + tracehook_notify_resume(&scr->pt); + } + /* copy user rbs to kernel rbs */ if (unlikely(test_thread_flag(TIF_RESTORE_RSE))) { local_irq_enable(); /* force interrupt enable */ @@ -251,7 +242,6 @@ default_idle (void) /* We don't actually take CPU down, just spin without interrupts. */ static inline void play_dead(void) { - extern void ia64_cpu_local_tick (void); unsigned int this_cpu = smp_processor_id(); /* Ack it */ diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c index 2a9943b..92c9689 100644 --- a/arch/ia64/kernel/ptrace.c +++ b/arch/ia64/kernel/ptrace.c @@ -22,6 +22,7 @@ #include <linux/signal.h> #include <linux/regset.h> #include <linux/elf.h> +#include <linux/tracehook.h> #include <asm/pgtable.h> #include <asm/processor.h> @@ -603,7 +604,7 @@ void ia64_ptrace_stop(void) { if (test_and_set_tsk_thread_flag(current, TIF_RESTORE_RSE)) return; - tsk_set_notify_resume(current); + set_notify_resume(current); unw_init_running(do_sync_rbs, ia64_sync_user_rbs); } @@ -613,7 +614,6 @@ void ia64_ptrace_stop(void) void ia64_sync_krbs(void) { clear_tsk_thread_flag(current, TIF_RESTORE_RSE); - tsk_clear_notify_resume(current); unw_init_running(do_sync_rbs, ia64_sync_kernel_rbs); } @@ -644,7 +644,7 @@ ptrace_attach_sync_user_rbs (struct task_struct *child) spin_lock_irq(&child->sighand->siglock); if (child->state == TASK_STOPPED && !test_and_set_tsk_thread_flag(child, TIF_RESTORE_RSE)) { - tsk_set_notify_resume(child); + set_notify_resume(child); child->state = TASK_TRACED; stopped = 1; @@ -1232,37 +1232,16 @@ arch_ptrace (struct task_struct *child, long request, long addr, long data) } -static void -syscall_trace (void) -{ - /* - * The 0x80 provides a way for the tracing parent to - * distinguish between a syscall stop and SIGTRAP delivery. - */ - ptrace_notify(SIGTRAP - | ((current->ptrace & PT_TRACESYSGOOD) ? 0x80 : 0)); - - /* - * This isn't the same as continuing with a signal, but it - * will do for normal use. strace only continues with a - * signal if the stopping signal is not SIGTRAP. -brl - */ - if (current->exit_code) { - send_sig(current->exit_code, current, 1); - current->exit_code = 0; - } -} - /* "asmlinkage" so the input arguments are preserved... */ -asmlinkage void +asmlinkage long syscall_trace_enter (long arg0, long arg1, long arg2, long arg3, long arg4, long arg5, long arg6, long arg7, struct pt_regs regs) { - if (test_thread_flag(TIF_SYSCALL_TRACE) - && (current->ptrace & PT_PTRACED)) - syscall_trace(); + if (test_thread_flag(TIF_SYSCALL_TRACE)) + if (tracehook_report_syscall_entry(®s)) + return -ENOSYS; /* copy user rbs to kernel rbs */ if (test_thread_flag(TIF_RESTORE_RSE)) @@ -1283,6 +1262,7 @@ syscall_trace_enter (long arg0, long arg1, long arg2, long arg3, audit_syscall_entry(arch, syscall, arg0, arg1, arg2, arg3); } + return 0; } /* "asmlinkage" so the input arguments are preserved... */ @@ -1292,6 +1272,8 @@ syscall_trace_leave (long arg0, long arg1, long arg2, long arg3, long arg4, long arg5, long arg6, long arg7, struct pt_regs regs) { + int step; + if (unlikely(current->audit_context)) { int success = AUDITSC_RESULT(regs.r10); long result = regs.r8; @@ -1301,10 +1283,9 @@ syscall_trace_leave (long arg0, long arg1, long arg2, long arg3, audit_syscall_exit(success, result); } - if ((test_thread_flag(TIF_SYSCALL_TRACE) - || test_thread_flag(TIF_SINGLESTEP)) - && (current->ptrace & PT_PTRACED)) - syscall_trace(); + step = test_thread_flag(TIF_SINGLESTEP); + if (step || test_thread_flag(TIF_SYSCALL_TRACE)) + tracehook_report_syscall_exit(®s, step); /* copy user rbs to kernel rbs */ if (test_thread_flag(TIF_RESTORE_RSE)) @@ -1940,7 +1921,7 @@ gpregs_writeback(struct task_struct *target, { if (test_and_set_tsk_thread_flag(target, TIF_RESTORE_RSE)) return 0; - tsk_set_notify_resume(target); + set_notify_resume(target); return do_regset_call(do_gpregs_writeback, target, regset, 0, 0, NULL, NULL); } @@ -2199,3 +2180,68 @@ const struct user_regset_view *task_user_regset_view(struct task_struct *tsk) #endif return &user_ia64_view; } + +struct syscall_get_set_args { + unsigned int i; + unsigned int n; + unsigned long *args; + struct pt_regs *regs; + int rw; +}; + +static void syscall_get_set_args_cb(struct unw_frame_info *info, void *data) +{ + struct syscall_get_set_args *args = data; + struct pt_regs *pt = args->regs; + unsigned long *krbs, cfm, ndirty; + int i, count; + + if (unw_unwind_to_user(info) < 0) + return; + + cfm = pt->cr_ifs; + krbs = (unsigned long *)info->task + IA64_RBS_OFFSET/8; + ndirty = ia64_rse_num_regs(krbs, krbs + (pt->loadrs >> 19)); + + count = 0; + if (in_syscall(pt)) + count = min_t(int, args->n, cfm & 0x7f); + + for (i = 0; i < count; i++) { + if (args->rw) + *ia64_rse_skip_regs(krbs, ndirty + i + args->i) = + args->args[i]; + else + args->args[i] = *ia64_rse_skip_regs(krbs, + ndirty + i + args->i); + } + + if (!args->rw) { + while (i < args->n) { + args->args[i] = 0; + i++; + } + } +} + +void ia64_syscall_get_set_arguments(struct task_struct *task, + struct pt_regs *regs, unsigned int i, unsigned int n, + unsigned long *args, int rw) +{ + struct syscall_get_set_args data = { + .i = i, + .n = n, + .args = args, + .regs = regs, + .rw = rw, + }; + + if (task == current) + unw_init_running(syscall_get_set_args_cb, &data); + else { + struct unw_frame_info ufi; + memset(&ufi, 0, sizeof(ufi)); + unw_init_from_blocked_task(&ufi, task); + syscall_get_set_args_cb(&ufi, &data); + } +} diff --git a/arch/ia64/kernel/setup.c b/arch/ia64/kernel/setup.c index 916ba89..ae79117 100644 --- a/arch/ia64/kernel/setup.c +++ b/arch/ia64/kernel/setup.c @@ -116,6 +116,13 @@ unsigned int num_io_spaces; */ #define I_CACHE_STRIDE_SHIFT 5 /* Safest way to go: 32 bytes by 32 bytes */ unsigned long ia64_i_cache_stride_shift = ~0; +/* + * "clflush_cache_range()" needs to know what processor dependent stride size to + * use when it flushes cache lines including both d-cache and i-cache. + */ +/* Safest way to go: 32 bytes by 32 bytes */ +#define CACHE_STRIDE_SHIFT 5 +unsigned long ia64_cache_stride_shift = ~0; /* * The merge_mask variable needs to be set to (max(iommu_page_size(iommu)) - 1). This @@ -852,13 +859,14 @@ setup_per_cpu_areas (void) } /* - * Calculate the max. cache line size. + * Do the following calculations: * - * In addition, the minimum of the i-cache stride sizes is calculated for - * "flush_icache_range()". + * 1. the max. cache line size. + * 2. the minimum of the i-cache stride sizes for "flush_icache_range()". + * 3. the minimum of the cache stride sizes for "clflush_cache_range()". */ static void __cpuinit -get_max_cacheline_size (void) +get_cache_info(void) { unsigned long line_size, max = 1; u64 l, levels, unique_caches; @@ -872,12 +880,14 @@ get_max_cacheline_size (void) max = SMP_CACHE_BYTES; /* Safest setup for "flush_icache_range()" */ ia64_i_cache_stride_shift = I_CACHE_STRIDE_SHIFT; + /* Safest setup for "clflush_cache_range()" */ + ia64_cache_stride_shift = CACHE_STRIDE_SHIFT; goto out; } for (l = 0; l < levels; ++l) { - status = ia64_pal_cache_config_info(l, /* cache_type (data_or_unified)= */ 2, - &cci); + /* cache_type (data_or_unified)=2 */ + status = ia64_pal_cache_config_info(l, 2, &cci); if (status != 0) { printk(KERN_ERR "%s: ia64_pal_cache_config_info(l=%lu, 2) failed (status=%ld)\n", @@ -885,15 +895,21 @@ get_max_cacheline_size (void) max = SMP_CACHE_BYTES; /* The safest setup for "flush_icache_range()" */ cci.pcci_stride = I_CACHE_STRIDE_SHIFT; + /* The safest setup for "clflush_cache_range()" */ + ia64_cache_stride_shift = CACHE_STRIDE_SHIFT; cci.pcci_unified = 1; + } else { + if (cci.pcci_stride < ia64_cache_stride_shift) + ia64_cache_stride_shift = cci.pcci_stride; + + line_size = 1 << cci.pcci_line_size; + if (line_size > max) + max = line_size; } - line_size = 1 << cci.pcci_line_size; - if (line_size > max) - max = line_size; + if (!cci.pcci_unified) { - status = ia64_pal_cache_config_info(l, - /* cache_type (instruction)= */ 1, - &cci); + /* cache_type (instruction)=1*/ + status = ia64_pal_cache_config_info(l, 1, &cci); if (status != 0) { printk(KERN_ERR "%s: ia64_pal_cache_config_info(l=%lu, 1) failed (status=%ld)\n", @@ -947,7 +963,7 @@ cpu_init (void) } #endif - get_max_cacheline_size(); + get_cache_info(); /* * We can't pass "local_cpu_data" to identify_cpu() because we haven't called diff --git a/arch/ia64/kernel/signal.c b/arch/ia64/kernel/signal.c index 19c5a78..e12500a 100644 --- a/arch/ia64/kernel/signal.c +++ b/arch/ia64/kernel/signal.c @@ -11,6 +11,7 @@ #include <linux/kernel.h> #include <linux/mm.h> #include <linux/ptrace.h> +#include <linux/tracehook.h> #include <linux/sched.h> #include <linux/signal.h> #include <linux/smp.h> @@ -439,6 +440,13 @@ handle_signal (unsigned long sig, struct k_sigaction *ka, siginfo_t *info, sigse sigaddset(¤t->blocked, sig); recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); + + /* + * Let tracing know that we've done the handler setup. + */ + tracehook_signal_handler(sig, info, ka, &scr->pt, + test_thread_flag(TIF_SINGLESTEP)); + return 1; } diff --git a/arch/ia64/lib/flush.S b/arch/ia64/lib/flush.S index 2a0d27f..1d8c888 100644 --- a/arch/ia64/lib/flush.S +++ b/arch/ia64/lib/flush.S @@ -60,3 +60,58 @@ GLOBAL_ENTRY(flush_icache_range) mov ar.lc=r3 // restore ar.lc br.ret.sptk.many rp END(flush_icache_range) + + /* + * clflush_cache_range(start,size) + * + * Flush cache lines from start to start+size-1. + * + * Must deal with range from start to start+size-1 but nothing else + * (need to be careful not to touch addresses that may be + * unmapped). + * + * Note: "in0" and "in1" are preserved for debugging purposes. + */ + .section .kprobes.text,"ax" +GLOBAL_ENTRY(clflush_cache_range) + + .prologue + alloc r2=ar.pfs,2,0,0,0 + movl r3=ia64_cache_stride_shift + mov r21=1 + add r22=in1,in0 + ;; + ld8 r20=[r3] // r20: stride shift + sub r22=r22,r0,1 // last byte address + ;; + shr.u r23=in0,r20 // start / (stride size) + shr.u r22=r22,r20 // (last byte address) / (stride size) + shl r21=r21,r20 // r21: stride size of the i-cache(s) + ;; + sub r8=r22,r23 // number of strides - 1 + shl r24=r23,r20 // r24: addresses for "fc" = + // "start" rounded down to stride + // boundary + .save ar.lc,r3 + mov r3=ar.lc // save ar.lc + ;; + + .body + mov ar.lc=r8 + ;; + /* + * 32 byte aligned loop, even number of (actually 2) bundles + */ +.Loop_fc: + fc r24 // issuable on M0 only + add r24=r21,r24 // we flush "stride size" bytes per iteration + nop.i 0 + br.cloop.sptk.few .Loop_fc + ;; + sync.i + ;; + srlz.i + ;; + mov ar.lc=r3 // restore ar.lc + br.ret.sptk.many rp +END(clflush_cache_range) diff --git a/arch/ia64/mm/tlb.c b/arch/ia64/mm/tlb.c index 8caf424..bd9818a 100644 --- a/arch/ia64/mm/tlb.c +++ b/arch/ia64/mm/tlb.c @@ -362,9 +362,13 @@ ia64_tlb_init (void) per_cpu(ia64_tr_num, cpu) = vm_info_1.pal_vm_info_1_s.max_dtr_entry+1; if (per_cpu(ia64_tr_num, cpu) > IA64_TR_ALLOC_MAX) { + static int justonce = 1; per_cpu(ia64_tr_num, cpu) = IA64_TR_ALLOC_MAX; - printk(KERN_DEBUG "TR register number exceeds IA64_TR_ALLOC_MAX!" - "IA64_TR_ALLOC_MAX should be extended\n"); + if (justonce) { + justonce = 0; + printk(KERN_DEBUG "TR register number exceeds " + "IA64_TR_ALLOC_MAX!\n"); + } } } diff --git a/arch/ia64/scripts/pvcheck.sed b/arch/ia64/scripts/pvcheck.sed new file mode 100644 index 0000000..ba66ac2 --- /dev/null +++ b/arch/ia64/scripts/pvcheck.sed @@ -0,0 +1,32 @@ +# +# Checker for paravirtualizations of privileged operations. +# +s/ssm.*psr\.ic.*/.warning \"ssm psr.ic should not be used directly\"/g +s/rsm.*psr\.ic.*/.warning \"rsm psr.ic should not be used directly\"/g +s/ssm.*psr\.i.*/.warning \"ssm psr.i should not be used directly\"/g +s/rsm.*psr\.i.*/.warning \"rsm psr.i should not be used directly\"/g +s/ssm.*psr\.dt.*/.warning \"ssm psr.dt should not be used directly\"/g +s/rsm.*psr\.dt.*/.warning \"rsm psr.dt should not be used directly\"/g +s/mov.*=.*cr\.ifa/.warning \"cr.ifa should not used directly\"/g +s/mov.*=.*cr\.itir/.warning \"cr.itir should not used directly\"/g +s/mov.*=.*cr\.isr/.warning \"cr.isr should not used directly\"/g +s/mov.*=.*cr\.iha/.warning \"cr.iha should not used directly\"/g +s/mov.*=.*cr\.ipsr/.warning \"cr.ipsr should not used directly\"/g +s/mov.*=.*cr\.iim/.warning \"cr.iim should not used directly\"/g +s/mov.*=.*cr\.iip/.warning \"cr.iip should not used directly\"/g +s/mov.*=.*cr\.ivr/.warning \"cr.ivr should not used directly\"/g +s/mov.*=[^\.]*psr/.warning \"psr should not used directly\"/g # avoid ar.fpsr +s/mov.*=.*ar\.eflags/.warning \"ar.eflags should not used directly\"/g +s/mov.*cr\.ifa.*=.*/.warning \"cr.ifa should not used directly\"/g +s/mov.*cr\.itir.*=.*/.warning \"cr.itir should not used directly\"/g +s/mov.*cr\.iha.*=.*/.warning \"cr.iha should not used directly\"/g +s/mov.*cr\.ipsr.*=.*/.warning \"cr.ipsr should not used directly\"/g +s/mov.*cr\.ifs.*=.*/.warning \"cr.ifs should not used directly\"/g +s/mov.*cr\.iip.*=.*/.warning \"cr.iip should not used directly\"/g +s/mov.*cr\.kr.*=.*/.warning \"cr.kr should not used directly\"/g +s/mov.*ar\.eflags.*=.*/.warning \"ar.eflags should not used directly\"/g +s/itc\.i.*/.warning \"itc.i should not be used directly.\"/g +s/itc\.d.*/.warning \"itc.d should not be used directly.\"/g +s/bsw\.0/.warning \"bsw.0 should not be used directly.\"/g +s/bsw\.1/.warning \"bsw.1 should not be used directly.\"/g +s/ptc\.ga.*/.warning \"ptc.ga should not be used directly.\"/g diff --git a/arch/ia64/xen/Kconfig b/arch/ia64/xen/Kconfig new file mode 100644 index 0000000..f1683a2 --- /dev/null +++ b/arch/ia64/xen/Kconfig @@ -0,0 +1,26 @@ +# +# This Kconfig describes xen/ia64 options +# + +config XEN + bool "Xen hypervisor support" + default y + depends on PARAVIRT && MCKINLEY && IA64_PAGE_SIZE_16KB && EXPERIMENTAL + select XEN_XENCOMM + select NO_IDLE_HZ + + # those are required to save/restore. + select ARCH_SUSPEND_POSSIBLE + select SUSPEND + select PM_SLEEP + help + Enable Xen hypervisor support. Resulting kernel runs + both as a guest OS on Xen and natively on hardware. + +config XEN_XENCOMM + depends on XEN + bool + +config NO_IDLE_HZ + depends on XEN + bool diff --git a/arch/ia64/xen/Makefile b/arch/ia64/xen/Makefile new file mode 100644 index 0000000..0ad0224 --- /dev/null +++ b/arch/ia64/xen/Makefile @@ -0,0 +1,22 @@ +# +# Makefile for Xen components +# + +obj-y := hypercall.o xenivt.o xensetup.o xen_pv_ops.o irq_xen.o \ + hypervisor.o xencomm.o xcom_hcall.o grant-table.o time.o suspend.o + +obj-$(CONFIG_IA64_GENERIC) += machvec.o + +AFLAGS_xenivt.o += -D__IA64_ASM_PARAVIRTUALIZED_XEN + +# xen multi compile +ASM_PARAVIRT_MULTI_COMPILE_SRCS = ivt.S entry.S +ASM_PARAVIRT_OBJS = $(addprefix xen-,$(ASM_PARAVIRT_MULTI_COMPILE_SRCS:.S=.o)) +obj-y += $(ASM_PARAVIRT_OBJS) +define paravirtualized_xen +AFLAGS_$(1) += -D__IA64_ASM_PARAVIRTUALIZED_XEN +endef +$(foreach o,$(ASM_PARAVIRT_OBJS),$(eval $(call paravirtualized_xen,$(o)))) + +$(obj)/xen-%.o: $(src)/../kernel/%.S FORCE + $(call if_changed_dep,as_o_S) diff --git a/arch/ia64/xen/grant-table.c b/arch/ia64/xen/grant-table.c new file mode 100644 index 0000000..777dd9a --- /dev/null +++ b/arch/ia64/xen/grant-table.c @@ -0,0 +1,155 @@ +/****************************************************************************** + * arch/ia64/xen/grant-table.c + * + * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <linux/module.h> +#include <linux/vmalloc.h> +#include <linux/mm.h> + +#include <xen/interface/xen.h> +#include <xen/interface/memory.h> +#include <xen/grant_table.h> + +#include <asm/xen/hypervisor.h> + +struct vm_struct *xen_alloc_vm_area(unsigned long size) +{ + int order; + unsigned long virt; + unsigned long nr_pages; + struct vm_struct *area; + + order = get_order(size); + virt = __get_free_pages(GFP_KERNEL, order); + if (virt == 0) + goto err0; + nr_pages = 1 << order; + scrub_pages(virt, nr_pages); + + area = kmalloc(sizeof(*area), GFP_KERNEL); + if (area == NULL) + goto err1; + + area->flags = VM_IOREMAP; + area->addr = (void *)virt; + area->size = size; + area->pages = NULL; + area->nr_pages = nr_pages; + area->phys_addr = 0; /* xenbus_map_ring_valloc uses this field! */ + + return area; + +err1: + free_pages(virt, order); +err0: + return NULL; +} +EXPORT_SYMBOL_GPL(xen_alloc_vm_area); + +void xen_free_vm_area(struct vm_struct *area) +{ + unsigned int order = get_order(area->size); + unsigned long i; + unsigned long phys_addr = __pa(area->addr); + + /* This area is used for foreign page mappping. + * So underlying machine page may not be assigned. */ + for (i = 0; i < (1 << order); i++) { + unsigned long ret; + unsigned long gpfn = (phys_addr >> PAGE_SHIFT) + i; + struct xen_memory_reservation reservation = { + .nr_extents = 1, + .address_bits = 0, + .extent_order = 0, + .domid = DOMID_SELF + }; + set_xen_guest_handle(reservation.extent_start, &gpfn); + ret = HYPERVISOR_memory_op(XENMEM_populate_physmap, + &reservation); + BUG_ON(ret != 1); + } + free_pages((unsigned long)area->addr, order); + kfree(area); +} +EXPORT_SYMBOL_GPL(xen_free_vm_area); + + +/**************************************************************************** + * grant table hack + * cmd: GNTTABOP_xxx + */ + +int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes, + unsigned long max_nr_gframes, + struct grant_entry **__shared) +{ + *__shared = __va(frames[0] << PAGE_SHIFT); + return 0; +} + +void arch_gnttab_unmap_shared(struct grant_entry *shared, + unsigned long nr_gframes) +{ + /* nothing */ +} + +static void +gnttab_map_grant_ref_pre(struct gnttab_map_grant_ref *uop) +{ + uint32_t flags; + + flags = uop->flags; + + if (flags & GNTMAP_host_map) { + if (flags & GNTMAP_application_map) { + printk(KERN_DEBUG + "GNTMAP_application_map is not supported yet: " + "flags 0x%x\n", flags); + BUG(); + } + if (flags & GNTMAP_contains_pte) { + printk(KERN_DEBUG + "GNTMAP_contains_pte is not supported yet: " + "flags 0x%x\n", flags); + BUG(); + } + } else if (flags & GNTMAP_device_map) { + printk("GNTMAP_device_map is not supported yet 0x%x\n", flags); + BUG(); /* not yet. actually this flag is not used. */ + } else { + BUG(); + } +} + +int +HYPERVISOR_grant_table_op(unsigned int cmd, void *uop, unsigned int count) +{ + if (cmd == GNTTABOP_map_grant_ref) { + unsigned int i; + for (i = 0; i < count; i++) { + gnttab_map_grant_ref_pre( + (struct gnttab_map_grant_ref *)uop + i); + } + } + return xencomm_hypercall_grant_table_op(cmd, uop, count); +} + +EXPORT_SYMBOL(HYPERVISOR_grant_table_op); diff --git a/arch/ia64/xen/hypercall.S b/arch/ia64/xen/hypercall.S new file mode 100644 index 0000000..d4ff0b9 --- /dev/null +++ b/arch/ia64/xen/hypercall.S @@ -0,0 +1,91 @@ +/* + * Support routines for Xen hypercalls + * + * Copyright (C) 2005 Dan Magenheimer <dan.magenheimer@hp.com> + * Copyright (C) 2008 Yaozu (Eddie) Dong <eddie.dong@intel.com> + */ + +#include <asm/asmmacro.h> +#include <asm/intrinsics.h> +#include <asm/xen/privop.h> + +/* + * Hypercalls without parameter. + */ +#define __HCALL0(name,hcall) \ + GLOBAL_ENTRY(name); \ + break hcall; \ + br.ret.sptk.many rp; \ + END(name) + +/* + * Hypercalls with 1 parameter. + */ +#define __HCALL1(name,hcall) \ + GLOBAL_ENTRY(name); \ + mov r8=r32; \ + break hcall; \ + br.ret.sptk.many rp; \ + END(name) + +/* + * Hypercalls with 2 parameters. + */ +#define __HCALL2(name,hcall) \ + GLOBAL_ENTRY(name); \ + mov r8=r32; \ + mov r9=r33; \ + break hcall; \ + br.ret.sptk.many rp; \ + END(name) + +__HCALL0(xen_get_psr, HYPERPRIVOP_GET_PSR) +__HCALL0(xen_get_ivr, HYPERPRIVOP_GET_IVR) +__HCALL0(xen_get_tpr, HYPERPRIVOP_GET_TPR) +__HCALL0(xen_hyper_ssm_i, HYPERPRIVOP_SSM_I) + +__HCALL1(xen_set_tpr, HYPERPRIVOP_SET_TPR) +__HCALL1(xen_eoi, HYPERPRIVOP_EOI) +__HCALL1(xen_thash, HYPERPRIVOP_THASH) +__HCALL1(xen_set_itm, HYPERPRIVOP_SET_ITM) +__HCALL1(xen_get_rr, HYPERPRIVOP_GET_RR) +__HCALL1(xen_fc, HYPERPRIVOP_FC) +__HCALL1(xen_get_cpuid, HYPERPRIVOP_GET_CPUID) +__HCALL1(xen_get_pmd, HYPERPRIVOP_GET_PMD) + +__HCALL2(xen_ptcga, HYPERPRIVOP_PTC_GA) +__HCALL2(xen_set_rr, HYPERPRIVOP_SET_RR) +__HCALL2(xen_set_kr, HYPERPRIVOP_SET_KR) + +#ifdef CONFIG_IA32_SUPPORT +__HCALL1(xen_get_eflag, HYPERPRIVOP_GET_EFLAG) +__HCALL1(xen_set_eflag, HYPERPRIVOP_SET_EFLAG) // refer SDM vol1 3.1.8 +#endif /* CONFIG_IA32_SUPPORT */ + +GLOBAL_ENTRY(xen_set_rr0_to_rr4) + mov r8=r32 + mov r9=r33 + mov r10=r34 + mov r11=r35 + mov r14=r36 + XEN_HYPER_SET_RR0_TO_RR4 + br.ret.sptk.many rp + ;; +END(xen_set_rr0_to_rr4) + +GLOBAL_ENTRY(xen_send_ipi) + mov r14=r32 + mov r15=r33 + mov r2=0x400 + break 0x1000 + ;; + br.ret.sptk.many rp + ;; +END(xen_send_ipi) + +GLOBAL_ENTRY(__hypercall) + mov r2=r37 + break 0x1000 + br.ret.sptk.many b0 + ;; +END(__hypercall) diff --git a/arch/ia64/xen/hypervisor.c b/arch/ia64/xen/hypervisor.c new file mode 100644 index 0000000..cac4d97 --- /dev/null +++ b/arch/ia64/xen/hypervisor.c @@ -0,0 +1,96 @@ +/****************************************************************************** + * arch/ia64/xen/hypervisor.c + * + * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <linux/efi.h> +#include <asm/xen/hypervisor.h> +#include <asm/xen/privop.h> + +#include "irq_xen.h" + +struct shared_info *HYPERVISOR_shared_info __read_mostly = + (struct shared_info *)XSI_BASE; +EXPORT_SYMBOL(HYPERVISOR_shared_info); + +DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu); + +struct start_info *xen_start_info; +EXPORT_SYMBOL(xen_start_info); + +EXPORT_SYMBOL(xen_domain_type); + +EXPORT_SYMBOL(__hypercall); + +/* Stolen from arch/x86/xen/enlighten.c */ +/* + * Flag to determine whether vcpu info placement is available on all + * VCPUs. We assume it is to start with, and then set it to zero on + * the first failure. This is because it can succeed on some VCPUs + * and not others, since it can involve hypervisor memory allocation, + * or because the guest failed to guarantee all the appropriate + * constraints on all VCPUs (ie buffer can't cross a page boundary). + * + * Note that any particular CPU may be using a placed vcpu structure, + * but we can only optimise if the all are. + * + * 0: not available, 1: available + */ + +static void __init xen_vcpu_setup(int cpu) +{ + /* + * WARNING: + * before changing MAX_VIRT_CPUS, + * check that shared_info fits on a page + */ + BUILD_BUG_ON(sizeof(struct shared_info) > PAGE_SIZE); + per_cpu(xen_vcpu, cpu) = &HYPERVISOR_shared_info->vcpu_info[cpu]; +} + +void __init xen_setup_vcpu_info_placement(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + xen_vcpu_setup(cpu); +} + +void __cpuinit +xen_cpu_init(void) +{ + xen_smp_intr_init(); +} + +/************************************************************************** + * opt feature + */ +void +xen_ia64_enable_opt_feature(void) +{ + /* Enable region 7 identity map optimizations in Xen */ + struct xen_ia64_opt_feature optf; + + optf.cmd = XEN_IA64_OPTF_IDENT_MAP_REG7; + optf.on = XEN_IA64_OPTF_ON; + optf.pgprot = pgprot_val(PAGE_KERNEL); + optf.key = 0; /* No key on linux. */ + HYPERVISOR_opt_feature(&optf); +} diff --git a/arch/ia64/xen/irq_xen.c b/arch/ia64/xen/irq_xen.c new file mode 100644 index 0000000..af93aad --- /dev/null +++ b/arch/ia64/xen/irq_xen.c @@ -0,0 +1,435 @@ +/****************************************************************************** + * arch/ia64/xen/irq_xen.c + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <linux/cpu.h> + +#include <xen/interface/xen.h> +#include <xen/interface/callback.h> +#include <xen/events.h> + +#include <asm/xen/privop.h> + +#include "irq_xen.h" + +/*************************************************************************** + * pv_irq_ops + * irq operations + */ + +static int +xen_assign_irq_vector(int irq) +{ + struct physdev_irq irq_op; + + irq_op.irq = irq; + if (HYPERVISOR_physdev_op(PHYSDEVOP_alloc_irq_vector, &irq_op)) + return -ENOSPC; + + return irq_op.vector; +} + +static void +xen_free_irq_vector(int vector) +{ + struct physdev_irq irq_op; + + if (vector < IA64_FIRST_DEVICE_VECTOR || + vector > IA64_LAST_DEVICE_VECTOR) + return; + + irq_op.vector = vector; + if (HYPERVISOR_physdev_op(PHYSDEVOP_free_irq_vector, &irq_op)) + printk(KERN_WARNING "%s: xen_free_irq_vecotr fail vector=%d\n", + __func__, vector); +} + + +static DEFINE_PER_CPU(int, timer_irq) = -1; +static DEFINE_PER_CPU(int, ipi_irq) = -1; +static DEFINE_PER_CPU(int, resched_irq) = -1; +static DEFINE_PER_CPU(int, cmc_irq) = -1; +static DEFINE_PER_CPU(int, cmcp_irq) = -1; +static DEFINE_PER_CPU(int, cpep_irq) = -1; +#define NAME_SIZE 15 +static DEFINE_PER_CPU(char[NAME_SIZE], timer_name); +static DEFINE_PER_CPU(char[NAME_SIZE], ipi_name); +static DEFINE_PER_CPU(char[NAME_SIZE], resched_name); +static DEFINE_PER_CPU(char[NAME_SIZE], cmc_name); +static DEFINE_PER_CPU(char[NAME_SIZE], cmcp_name); +static DEFINE_PER_CPU(char[NAME_SIZE], cpep_name); +#undef NAME_SIZE + +struct saved_irq { + unsigned int irq; + struct irqaction *action; +}; +/* 16 should be far optimistic value, since only several percpu irqs + * are registered early. + */ +#define MAX_LATE_IRQ 16 +static struct saved_irq saved_percpu_irqs[MAX_LATE_IRQ]; +static unsigned short late_irq_cnt; +static unsigned short saved_irq_cnt; +static int xen_slab_ready; + +#ifdef CONFIG_SMP +/* Dummy stub. Though we may check XEN_RESCHEDULE_VECTOR before __do_IRQ, + * it ends up to issue several memory accesses upon percpu data and + * thus adds unnecessary traffic to other paths. + */ +static irqreturn_t +xen_dummy_handler(int irq, void *dev_id) +{ + + return IRQ_HANDLED; +} + +static struct irqaction xen_ipi_irqaction = { + .handler = handle_IPI, + .flags = IRQF_DISABLED, + .name = "IPI" +}; + +static struct irqaction xen_resched_irqaction = { + .handler = xen_dummy_handler, + .flags = IRQF_DISABLED, + .name = "resched" +}; + +static struct irqaction xen_tlb_irqaction = { + .handler = xen_dummy_handler, + .flags = IRQF_DISABLED, + .name = "tlb_flush" +}; +#endif + +/* + * This is xen version percpu irq registration, which needs bind + * to xen specific evtchn sub-system. One trick here is that xen + * evtchn binding interface depends on kmalloc because related + * port needs to be freed at device/cpu down. So we cache the + * registration on BSP before slab is ready and then deal them + * at later point. For rest instances happening after slab ready, + * we hook them to xen evtchn immediately. + * + * FIXME: MCA is not supported by far, and thus "nomca" boot param is + * required. + */ +static void +__xen_register_percpu_irq(unsigned int cpu, unsigned int vec, + struct irqaction *action, int save) +{ + irq_desc_t *desc; + int irq = 0; + + if (xen_slab_ready) { + switch (vec) { + case IA64_TIMER_VECTOR: + snprintf(per_cpu(timer_name, cpu), + sizeof(per_cpu(timer_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_virq_to_irqhandler(VIRQ_ITC, cpu, + action->handler, action->flags, + per_cpu(timer_name, cpu), action->dev_id); + per_cpu(timer_irq, cpu) = irq; + break; + case IA64_IPI_RESCHEDULE: + snprintf(per_cpu(resched_name, cpu), + sizeof(per_cpu(resched_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_ipi_to_irqhandler(XEN_RESCHEDULE_VECTOR, cpu, + action->handler, action->flags, + per_cpu(resched_name, cpu), action->dev_id); + per_cpu(resched_irq, cpu) = irq; + break; + case IA64_IPI_VECTOR: + snprintf(per_cpu(ipi_name, cpu), + sizeof(per_cpu(ipi_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_ipi_to_irqhandler(XEN_IPI_VECTOR, cpu, + action->handler, action->flags, + per_cpu(ipi_name, cpu), action->dev_id); + per_cpu(ipi_irq, cpu) = irq; + break; + case IA64_CMC_VECTOR: + snprintf(per_cpu(cmc_name, cpu), + sizeof(per_cpu(cmc_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_virq_to_irqhandler(VIRQ_MCA_CMC, cpu, + action->handler, + action->flags, + per_cpu(cmc_name, cpu), + action->dev_id); + per_cpu(cmc_irq, cpu) = irq; + break; + case IA64_CMCP_VECTOR: + snprintf(per_cpu(cmcp_name, cpu), + sizeof(per_cpu(cmcp_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_ipi_to_irqhandler(XEN_CMCP_VECTOR, cpu, + action->handler, + action->flags, + per_cpu(cmcp_name, cpu), + action->dev_id); + per_cpu(cmcp_irq, cpu) = irq; + break; + case IA64_CPEP_VECTOR: + snprintf(per_cpu(cpep_name, cpu), + sizeof(per_cpu(cpep_name, cpu)), + "%s%d", action->name, cpu); + irq = bind_ipi_to_irqhandler(XEN_CPEP_VECTOR, cpu, + action->handler, + action->flags, + per_cpu(cpep_name, cpu), + action->dev_id); + per_cpu(cpep_irq, cpu) = irq; + break; + case IA64_CPE_VECTOR: + case IA64_MCA_RENDEZ_VECTOR: + case IA64_PERFMON_VECTOR: + case IA64_MCA_WAKEUP_VECTOR: + case IA64_SPURIOUS_INT_VECTOR: + /* No need to complain, these aren't supported. */ + break; + default: + printk(KERN_WARNING "Percpu irq %d is unsupported " + "by xen!\n", vec); + break; + } + BUG_ON(irq < 0); + + if (irq > 0) { + /* + * Mark percpu. Without this, migrate_irqs() will + * mark the interrupt for migrations and trigger it + * on cpu hotplug. + */ + desc = irq_desc + irq; + desc->status |= IRQ_PER_CPU; + } + } + + /* For BSP, we cache registered percpu irqs, and then re-walk + * them when initializing APs + */ + if (!cpu && save) { + BUG_ON(saved_irq_cnt == MAX_LATE_IRQ); + saved_percpu_irqs[saved_irq_cnt].irq = vec; + saved_percpu_irqs[saved_irq_cnt].action = action; + saved_irq_cnt++; + if (!xen_slab_ready) + late_irq_cnt++; + } +} + +static void +xen_register_percpu_irq(ia64_vector vec, struct irqaction *action) +{ + __xen_register_percpu_irq(smp_processor_id(), vec, action, 1); +} + +static void +xen_bind_early_percpu_irq(void) +{ + int i; + + xen_slab_ready = 1; + /* There's no race when accessing this cached array, since only + * BSP will face with such step shortly + */ + for (i = 0; i < late_irq_cnt; i++) + __xen_register_percpu_irq(smp_processor_id(), + saved_percpu_irqs[i].irq, + saved_percpu_irqs[i].action, 0); +} + +/* FIXME: There's no obvious point to check whether slab is ready. So + * a hack is used here by utilizing a late time hook. + */ + +#ifdef CONFIG_HOTPLUG_CPU +static int __devinit +unbind_evtchn_callback(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + unsigned int cpu = (unsigned long)hcpu; + + if (action == CPU_DEAD) { + /* Unregister evtchn. */ + if (per_cpu(cpep_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(cpep_irq, cpu), NULL); + per_cpu(cpep_irq, cpu) = -1; + } + if (per_cpu(cmcp_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(cmcp_irq, cpu), NULL); + per_cpu(cmcp_irq, cpu) = -1; + } + if (per_cpu(cmc_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(cmc_irq, cpu), NULL); + per_cpu(cmc_irq, cpu) = -1; + } + if (per_cpu(ipi_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(ipi_irq, cpu), NULL); + per_cpu(ipi_irq, cpu) = -1; + } + if (per_cpu(resched_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(resched_irq, cpu), + NULL); + per_cpu(resched_irq, cpu) = -1; + } + if (per_cpu(timer_irq, cpu) >= 0) { + unbind_from_irqhandler(per_cpu(timer_irq, cpu), NULL); + per_cpu(timer_irq, cpu) = -1; + } + } + return NOTIFY_OK; +} + +static struct notifier_block unbind_evtchn_notifier = { + .notifier_call = unbind_evtchn_callback, + .priority = 0 +}; +#endif + +void xen_smp_intr_init_early(unsigned int cpu) +{ +#ifdef CONFIG_SMP + unsigned int i; + + for (i = 0; i < saved_irq_cnt; i++) + __xen_register_percpu_irq(cpu, saved_percpu_irqs[i].irq, + saved_percpu_irqs[i].action, 0); +#endif +} + +void xen_smp_intr_init(void) +{ +#ifdef CONFIG_SMP + unsigned int cpu = smp_processor_id(); + struct callback_register event = { + .type = CALLBACKTYPE_event, + .address = { .ip = (unsigned long)&xen_event_callback }, + }; + + if (cpu == 0) { + /* Initialization was already done for boot cpu. */ +#ifdef CONFIG_HOTPLUG_CPU + /* Register the notifier only once. */ + register_cpu_notifier(&unbind_evtchn_notifier); +#endif + return; + } + + /* This should be piggyback when setup vcpu guest context */ + BUG_ON(HYPERVISOR_callback_op(CALLBACKOP_register, &event)); +#endif /* CONFIG_SMP */ +} + +void __init +xen_irq_init(void) +{ + struct callback_register event = { + .type = CALLBACKTYPE_event, + .address = { .ip = (unsigned long)&xen_event_callback }, + }; + + xen_init_IRQ(); + BUG_ON(HYPERVISOR_callback_op(CALLBACKOP_register, &event)); + late_time_init = xen_bind_early_percpu_irq; +} + +void +xen_platform_send_ipi(int cpu, int vector, int delivery_mode, int redirect) +{ +#ifdef CONFIG_SMP + /* TODO: we need to call vcpu_up here */ + if (unlikely(vector == ap_wakeup_vector)) { + /* XXX + * This should be in __cpu_up(cpu) in ia64 smpboot.c + * like x86. But don't want to modify it, + * keep it untouched. + */ + xen_smp_intr_init_early(cpu); + + xen_send_ipi(cpu, vector); + /* vcpu_prepare_and_up(cpu); */ + return; + } +#endif + + switch (vector) { + case IA64_IPI_VECTOR: + xen_send_IPI_one(cpu, XEN_IPI_VECTOR); + break; + case IA64_IPI_RESCHEDULE: + xen_send_IPI_one(cpu, XEN_RESCHEDULE_VECTOR); + break; + case IA64_CMCP_VECTOR: + xen_send_IPI_one(cpu, XEN_CMCP_VECTOR); + break; + case IA64_CPEP_VECTOR: + xen_send_IPI_one(cpu, XEN_CPEP_VECTOR); + break; + case IA64_TIMER_VECTOR: { + /* this is used only once by check_sal_cache_flush() + at boot time */ + static int used = 0; + if (!used) { + xen_send_ipi(cpu, IA64_TIMER_VECTOR); + used = 1; + break; + } + /* fallthrough */ + } + default: + printk(KERN_WARNING "Unsupported IPI type 0x%x\n", + vector); + notify_remote_via_irq(0); /* defaults to 0 irq */ + break; + } +} + +static void __init +xen_register_ipi(void) +{ +#ifdef CONFIG_SMP + register_percpu_irq(IA64_IPI_VECTOR, &xen_ipi_irqaction); + register_percpu_irq(IA64_IPI_RESCHEDULE, &xen_resched_irqaction); + register_percpu_irq(IA64_IPI_LOCAL_TLB_FLUSH, &xen_tlb_irqaction); +#endif +} + +static void +xen_resend_irq(unsigned int vector) +{ + (void)resend_irq_on_evtchn(vector); +} + +const struct pv_irq_ops xen_irq_ops __initdata = { + .register_ipi = xen_register_ipi, + + .assign_irq_vector = xen_assign_irq_vector, + .free_irq_vector = xen_free_irq_vector, + .register_percpu_irq = xen_register_percpu_irq, + + .resend_irq = xen_resend_irq, +}; diff --git a/arch/ia64/xen/irq_xen.h b/arch/ia64/xen/irq_xen.h new file mode 100644 index 0000000..26110f3 --- /dev/null +++ b/arch/ia64/xen/irq_xen.h @@ -0,0 +1,34 @@ +/****************************************************************************** + * arch/ia64/xen/irq_xen.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#ifndef IRQ_XEN_H +#define IRQ_XEN_H + +extern void (*late_time_init)(void); +extern char xen_event_callback; +void __init xen_init_IRQ(void); + +extern const struct pv_irq_ops xen_irq_ops __initdata; +extern void xen_smp_intr_init(void); +extern void xen_send_ipi(int cpu, int vec); + +#endif /* IRQ_XEN_H */ diff --git a/arch/ia64/xen/machvec.c b/arch/ia64/xen/machvec.c new file mode 100644 index 0000000..4ad588a --- /dev/null +++ b/arch/ia64/xen/machvec.c @@ -0,0 +1,4 @@ +#define MACHVEC_PLATFORM_NAME xen +#define MACHVEC_PLATFORM_HEADER <asm/machvec_xen.h> +#include <asm/machvec_init.h> + diff --git a/arch/ia64/xen/suspend.c b/arch/ia64/xen/suspend.c new file mode 100644 index 0000000..fd66b04 --- /dev/null +++ b/arch/ia64/xen/suspend.c @@ -0,0 +1,64 @@ +/****************************************************************************** + * arch/ia64/xen/suspend.c + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + * suspend/resume + */ + +#include <xen/xen-ops.h> +#include <asm/xen/hypervisor.h> +#include "time.h" + +void +xen_mm_pin_all(void) +{ + /* nothing */ +} + +void +xen_mm_unpin_all(void) +{ + /* nothing */ +} + +void xen_pre_device_suspend(void) +{ + /* nothing */ +} + +void +xen_pre_suspend() +{ + /* nothing */ +} + +void +xen_post_suspend(int suspend_cancelled) +{ + if (suspend_cancelled) + return; + + xen_ia64_enable_opt_feature(); + /* add more if necessary */ +} + +void xen_arch_resume(void) +{ + xen_timer_resume_on_aps(); +} diff --git a/arch/ia64/xen/time.c b/arch/ia64/xen/time.c new file mode 100644 index 0000000..d15a94c --- /dev/null +++ b/arch/ia64/xen/time.c @@ -0,0 +1,213 @@ +/****************************************************************************** + * arch/ia64/xen/time.c + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <linux/delay.h> +#include <linux/kernel_stat.h> +#include <linux/posix-timers.h> +#include <linux/irq.h> +#include <linux/clocksource.h> + +#include <asm/timex.h> + +#include <asm/xen/hypervisor.h> + +#include <xen/interface/vcpu.h> + +#include "../kernel/fsyscall_gtod_data.h" + +DEFINE_PER_CPU(struct vcpu_runstate_info, runstate); +DEFINE_PER_CPU(unsigned long, processed_stolen_time); +DEFINE_PER_CPU(unsigned long, processed_blocked_time); + +/* taken from i386/kernel/time-xen.c */ +static void xen_init_missing_ticks_accounting(int cpu) +{ + struct vcpu_register_runstate_memory_area area; + struct vcpu_runstate_info *runstate = &per_cpu(runstate, cpu); + int rc; + + memset(runstate, 0, sizeof(*runstate)); + + area.addr.v = runstate; + rc = HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area, cpu, + &area); + WARN_ON(rc && rc != -ENOSYS); + + per_cpu(processed_blocked_time, cpu) = runstate->time[RUNSTATE_blocked]; + per_cpu(processed_stolen_time, cpu) = runstate->time[RUNSTATE_runnable] + + runstate->time[RUNSTATE_offline]; +} + +/* + * Runstate accounting + */ +/* stolen from arch/x86/xen/time.c */ +static void get_runstate_snapshot(struct vcpu_runstate_info *res) +{ + u64 state_time; + struct vcpu_runstate_info *state; + + BUG_ON(preemptible()); + + state = &__get_cpu_var(runstate); + + /* + * The runstate info is always updated by the hypervisor on + * the current CPU, so there's no need to use anything + * stronger than a compiler barrier when fetching it. + */ + do { + state_time = state->state_entry_time; + rmb(); + *res = *state; + rmb(); + } while (state->state_entry_time != state_time); +} + +#define NS_PER_TICK (1000000000LL/HZ) + +static unsigned long +consider_steal_time(unsigned long new_itm) +{ + unsigned long stolen, blocked; + unsigned long delta_itm = 0, stolentick = 0; + int cpu = smp_processor_id(); + struct vcpu_runstate_info runstate; + struct task_struct *p = current; + + get_runstate_snapshot(&runstate); + + /* + * Check for vcpu migration effect + * In this case, itc value is reversed. + * This causes huge stolen value. + * This function just checks and reject this effect. + */ + if (!time_after_eq(runstate.time[RUNSTATE_blocked], + per_cpu(processed_blocked_time, cpu))) + blocked = 0; + + if (!time_after_eq(runstate.time[RUNSTATE_runnable] + + runstate.time[RUNSTATE_offline], + per_cpu(processed_stolen_time, cpu))) + stolen = 0; + + if (!time_after(delta_itm + new_itm, ia64_get_itc())) + stolentick = ia64_get_itc() - new_itm; + + do_div(stolentick, NS_PER_TICK); + stolentick++; + + do_div(stolen, NS_PER_TICK); + + if (stolen > stolentick) + stolen = stolentick; + + stolentick -= stolen; + do_div(blocked, NS_PER_TICK); + + if (blocked > stolentick) + blocked = stolentick; + + if (stolen > 0 || blocked > 0) { + account_steal_time(NULL, jiffies_to_cputime(stolen)); + account_steal_time(idle_task(cpu), jiffies_to_cputime(blocked)); + run_local_timers(); + + if (rcu_pending(cpu)) + rcu_check_callbacks(cpu, user_mode(get_irq_regs())); + + scheduler_tick(); + run_posix_cpu_timers(p); + delta_itm += local_cpu_data->itm_delta * (stolen + blocked); + + if (cpu == time_keeper_id) { + write_seqlock(&xtime_lock); + do_timer(stolen + blocked); + local_cpu_data->itm_next = delta_itm + new_itm; + write_sequnlock(&xtime_lock); + } else { + local_cpu_data->itm_next = delta_itm + new_itm; + } + per_cpu(processed_stolen_time, cpu) += NS_PER_TICK * stolen; + per_cpu(processed_blocked_time, cpu) += NS_PER_TICK * blocked; + } + return delta_itm; +} + +static int xen_do_steal_accounting(unsigned long *new_itm) +{ + unsigned long delta_itm; + delta_itm = consider_steal_time(*new_itm); + *new_itm += delta_itm; + if (time_after(*new_itm, ia64_get_itc()) && delta_itm) + return 1; + + return 0; +} + +static void xen_itc_jitter_data_reset(void) +{ + u64 lcycle, ret; + + do { + lcycle = itc_jitter_data.itc_lastcycle; + ret = cmpxchg(&itc_jitter_data.itc_lastcycle, lcycle, 0); + } while (unlikely(ret != lcycle)); +} + +struct pv_time_ops xen_time_ops __initdata = { + .init_missing_ticks_accounting = xen_init_missing_ticks_accounting, + .do_steal_accounting = xen_do_steal_accounting, + .clocksource_resume = xen_itc_jitter_data_reset, +}; + +/* Called after suspend, to resume time. */ +static void xen_local_tick_resume(void) +{ + /* Just trigger a tick. */ + ia64_cpu_local_tick(); + touch_softlockup_watchdog(); +} + +void +xen_timer_resume(void) +{ + unsigned int cpu; + + xen_local_tick_resume(); + + for_each_online_cpu(cpu) + xen_init_missing_ticks_accounting(cpu); +} + +static void ia64_cpu_local_tick_fn(void *unused) +{ + xen_local_tick_resume(); + xen_init_missing_ticks_accounting(smp_processor_id()); +} + +void +xen_timer_resume_on_aps(void) +{ + smp_call_function(&ia64_cpu_local_tick_fn, NULL, 1); +} diff --git a/arch/ia64/xen/time.h b/arch/ia64/xen/time.h new file mode 100644 index 0000000..f98d7e1 --- /dev/null +++ b/arch/ia64/xen/time.h @@ -0,0 +1,24 @@ +/****************************************************************************** + * arch/ia64/xen/time.h + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +extern struct pv_time_ops xen_time_ops __initdata; +void xen_timer_resume_on_aps(void); diff --git a/arch/ia64/xen/xcom_hcall.c b/arch/ia64/xen/xcom_hcall.c new file mode 100644 index 0000000..ccaf743 --- /dev/null +++ b/arch/ia64/xen/xcom_hcall.c @@ -0,0 +1,441 @@ +/* + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * + * Tristan Gingold <tristan.gingold@bull.net> + * + * Copyright (c) 2007 + * Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * consolidate mini and inline version. + */ + +#include <linux/module.h> +#include <xen/interface/xen.h> +#include <xen/interface/memory.h> +#include <xen/interface/grant_table.h> +#include <xen/interface/callback.h> +#include <xen/interface/vcpu.h> +#include <asm/xen/hypervisor.h> +#include <asm/xen/xencomm.h> + +/* Xencomm notes: + * This file defines hypercalls to be used by xencomm. The hypercalls simply + * create inlines or mini descriptors for pointers and then call the raw arch + * hypercall xencomm_arch_hypercall_XXX + * + * If the arch wants to directly use these hypercalls, simply define macros + * in asm/xen/hypercall.h, eg: + * #define HYPERVISOR_sched_op xencomm_hypercall_sched_op + * + * The arch may also define HYPERVISOR_xxx as a function and do more operations + * before/after doing the hypercall. + * + * Note: because only inline or mini descriptors are created these functions + * must only be called with in kernel memory parameters. + */ + +int +xencomm_hypercall_console_io(int cmd, int count, char *str) +{ + /* xen early printk uses console io hypercall before + * xencomm initialization. In that case, we just ignore it. + */ + if (!xencomm_is_initialized()) + return 0; + + return xencomm_arch_hypercall_console_io + (cmd, count, xencomm_map_no_alloc(str, count)); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_console_io); + +int +xencomm_hypercall_event_channel_op(int cmd, void *op) +{ + struct xencomm_handle *desc; + desc = xencomm_map_no_alloc(op, sizeof(struct evtchn_op)); + if (desc == NULL) + return -EINVAL; + + return xencomm_arch_hypercall_event_channel_op(cmd, desc); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_event_channel_op); + +int +xencomm_hypercall_xen_version(int cmd, void *arg) +{ + struct xencomm_handle *desc; + unsigned int argsize; + + switch (cmd) { + case XENVER_version: + /* do not actually pass an argument */ + return xencomm_arch_hypercall_xen_version(cmd, 0); + case XENVER_extraversion: + argsize = sizeof(struct xen_extraversion); + break; + case XENVER_compile_info: + argsize = sizeof(struct xen_compile_info); + break; + case XENVER_capabilities: + argsize = sizeof(struct xen_capabilities_info); + break; + case XENVER_changeset: + argsize = sizeof(struct xen_changeset_info); + break; + case XENVER_platform_parameters: + argsize = sizeof(struct xen_platform_parameters); + break; + case XENVER_get_features: + argsize = (arg == NULL) ? 0 : sizeof(struct xen_feature_info); + break; + + default: + printk(KERN_DEBUG + "%s: unknown version op %d\n", __func__, cmd); + return -ENOSYS; + } + + desc = xencomm_map_no_alloc(arg, argsize); + if (desc == NULL) + return -EINVAL; + + return xencomm_arch_hypercall_xen_version(cmd, desc); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_xen_version); + +int +xencomm_hypercall_physdev_op(int cmd, void *op) +{ + unsigned int argsize; + + switch (cmd) { + case PHYSDEVOP_apic_read: + case PHYSDEVOP_apic_write: + argsize = sizeof(struct physdev_apic); + break; + case PHYSDEVOP_alloc_irq_vector: + case PHYSDEVOP_free_irq_vector: + argsize = sizeof(struct physdev_irq); + break; + case PHYSDEVOP_irq_status_query: + argsize = sizeof(struct physdev_irq_status_query); + break; + + default: + printk(KERN_DEBUG + "%s: unknown physdev op %d\n", __func__, cmd); + return -ENOSYS; + } + + return xencomm_arch_hypercall_physdev_op + (cmd, xencomm_map_no_alloc(op, argsize)); +} + +static int +xencommize_grant_table_op(struct xencomm_mini **xc_area, + unsigned int cmd, void *op, unsigned int count, + struct xencomm_handle **desc) +{ + struct xencomm_handle *desc1; + unsigned int argsize; + + switch (cmd) { + case GNTTABOP_map_grant_ref: + argsize = sizeof(struct gnttab_map_grant_ref); + break; + case GNTTABOP_unmap_grant_ref: + argsize = sizeof(struct gnttab_unmap_grant_ref); + break; + case GNTTABOP_setup_table: + { + struct gnttab_setup_table *setup = op; + + argsize = sizeof(*setup); + + if (count != 1) + return -EINVAL; + desc1 = __xencomm_map_no_alloc + (xen_guest_handle(setup->frame_list), + setup->nr_frames * + sizeof(*xen_guest_handle(setup->frame_list)), + *xc_area); + if (desc1 == NULL) + return -EINVAL; + (*xc_area)++; + set_xen_guest_handle(setup->frame_list, (void *)desc1); + break; + } + case GNTTABOP_dump_table: + argsize = sizeof(struct gnttab_dump_table); + break; + case GNTTABOP_transfer: + argsize = sizeof(struct gnttab_transfer); + break; + case GNTTABOP_copy: + argsize = sizeof(struct gnttab_copy); + break; + case GNTTABOP_query_size: + argsize = sizeof(struct gnttab_query_size); + break; + default: + printk(KERN_DEBUG "%s: unknown hypercall grant table op %d\n", + __func__, cmd); + BUG(); + } + + *desc = __xencomm_map_no_alloc(op, count * argsize, *xc_area); + if (*desc == NULL) + return -EINVAL; + (*xc_area)++; + + return 0; +} + +int +xencomm_hypercall_grant_table_op(unsigned int cmd, void *op, + unsigned int count) +{ + int rc; + struct xencomm_handle *desc; + XENCOMM_MINI_ALIGNED(xc_area, 2); + + rc = xencommize_grant_table_op(&xc_area, cmd, op, count, &desc); + if (rc) + return rc; + + return xencomm_arch_hypercall_grant_table_op(cmd, desc, count); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_grant_table_op); + +int +xencomm_hypercall_sched_op(int cmd, void *arg) +{ + struct xencomm_handle *desc; + unsigned int argsize; + + switch (cmd) { + case SCHEDOP_yield: + case SCHEDOP_block: + argsize = 0; + break; + case SCHEDOP_shutdown: + argsize = sizeof(struct sched_shutdown); + break; + case SCHEDOP_poll: + { + struct sched_poll *poll = arg; + struct xencomm_handle *ports; + + argsize = sizeof(struct sched_poll); + ports = xencomm_map_no_alloc(xen_guest_handle(poll->ports), + sizeof(*xen_guest_handle(poll->ports))); + + set_xen_guest_handle(poll->ports, (void *)ports); + break; + } + default: + printk(KERN_DEBUG "%s: unknown sched op %d\n", __func__, cmd); + return -ENOSYS; + } + + desc = xencomm_map_no_alloc(arg, argsize); + if (desc == NULL) + return -EINVAL; + + return xencomm_arch_hypercall_sched_op(cmd, desc); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_sched_op); + +int +xencomm_hypercall_multicall(void *call_list, int nr_calls) +{ + int rc; + int i; + struct multicall_entry *mce; + struct xencomm_handle *desc; + XENCOMM_MINI_ALIGNED(xc_area, nr_calls * 2); + + for (i = 0; i < nr_calls; i++) { + mce = (struct multicall_entry *)call_list + i; + + switch (mce->op) { + case __HYPERVISOR_update_va_mapping: + case __HYPERVISOR_mmu_update: + /* No-op on ia64. */ + break; + case __HYPERVISOR_grant_table_op: + rc = xencommize_grant_table_op + (&xc_area, + mce->args[0], (void *)mce->args[1], + mce->args[2], &desc); + if (rc) + return rc; + mce->args[1] = (unsigned long)desc; + break; + case __HYPERVISOR_memory_op: + default: + printk(KERN_DEBUG + "%s: unhandled multicall op entry op %lu\n", + __func__, mce->op); + return -ENOSYS; + } + } + + desc = xencomm_map_no_alloc(call_list, + nr_calls * sizeof(struct multicall_entry)); + if (desc == NULL) + return -EINVAL; + + return xencomm_arch_hypercall_multicall(desc, nr_calls); +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_multicall); + +int +xencomm_hypercall_callback_op(int cmd, void *arg) +{ + unsigned int argsize; + switch (cmd) { + case CALLBACKOP_register: + argsize = sizeof(struct callback_register); + break; + case CALLBACKOP_unregister: + argsize = sizeof(struct callback_unregister); + break; + default: + printk(KERN_DEBUG + "%s: unknown callback op %d\n", __func__, cmd); + return -ENOSYS; + } + + return xencomm_arch_hypercall_callback_op + (cmd, xencomm_map_no_alloc(arg, argsize)); +} + +static int +xencommize_memory_reservation(struct xencomm_mini *xc_area, + struct xen_memory_reservation *mop) +{ + struct xencomm_handle *desc; + + desc = __xencomm_map_no_alloc(xen_guest_handle(mop->extent_start), + mop->nr_extents * + sizeof(*xen_guest_handle(mop->extent_start)), + xc_area); + if (desc == NULL) + return -EINVAL; + + set_xen_guest_handle(mop->extent_start, (void *)desc); + return 0; +} + +int +xencomm_hypercall_memory_op(unsigned int cmd, void *arg) +{ + GUEST_HANDLE(xen_pfn_t) extent_start_va[2] = { {NULL}, {NULL} }; + struct xen_memory_reservation *xmr = NULL; + int rc; + struct xencomm_handle *desc; + unsigned int argsize; + XENCOMM_MINI_ALIGNED(xc_area, 2); + + switch (cmd) { + case XENMEM_increase_reservation: + case XENMEM_decrease_reservation: + case XENMEM_populate_physmap: + xmr = (struct xen_memory_reservation *)arg; + set_xen_guest_handle(extent_start_va[0], + xen_guest_handle(xmr->extent_start)); + + argsize = sizeof(*xmr); + rc = xencommize_memory_reservation(xc_area, xmr); + if (rc) + return rc; + xc_area++; + break; + + case XENMEM_maximum_ram_page: + argsize = 0; + break; + + case XENMEM_add_to_physmap: + argsize = sizeof(struct xen_add_to_physmap); + break; + + default: + printk(KERN_DEBUG "%s: unknown memory op %d\n", __func__, cmd); + return -ENOSYS; + } + + desc = xencomm_map_no_alloc(arg, argsize); + if (desc == NULL) + return -EINVAL; + + rc = xencomm_arch_hypercall_memory_op(cmd, desc); + + switch (cmd) { + case XENMEM_increase_reservation: + case XENMEM_decrease_reservation: + case XENMEM_populate_physmap: + set_xen_guest_handle(xmr->extent_start, + xen_guest_handle(extent_start_va[0])); + break; + } + + return rc; +} +EXPORT_SYMBOL_GPL(xencomm_hypercall_memory_op); + +int +xencomm_hypercall_suspend(unsigned long srec) +{ + struct sched_shutdown arg; + + arg.reason = SHUTDOWN_suspend; + + return xencomm_arch_hypercall_sched_op( + SCHEDOP_shutdown, xencomm_map_no_alloc(&arg, sizeof(arg))); +} + +long +xencomm_hypercall_vcpu_op(int cmd, int cpu, void *arg) +{ + unsigned int argsize; + switch (cmd) { + case VCPUOP_register_runstate_memory_area: { + struct vcpu_register_runstate_memory_area *area = + (struct vcpu_register_runstate_memory_area *)arg; + argsize = sizeof(*arg); + set_xen_guest_handle(area->addr.h, + (void *)xencomm_map_no_alloc(area->addr.v, + sizeof(area->addr.v))); + break; + } + + default: + printk(KERN_DEBUG "%s: unknown vcpu op %d\n", __func__, cmd); + return -ENOSYS; + } + + return xencomm_arch_hypercall_vcpu_op(cmd, cpu, + xencomm_map_no_alloc(arg, argsize)); +} + +long +xencomm_hypercall_opt_feature(void *arg) +{ + return xencomm_arch_hypercall_opt_feature( + xencomm_map_no_alloc(arg, + sizeof(struct xen_ia64_opt_feature))); +} diff --git a/arch/ia64/xen/xen_pv_ops.c b/arch/ia64/xen/xen_pv_ops.c new file mode 100644 index 0000000..04cd123 --- /dev/null +++ b/arch/ia64/xen/xen_pv_ops.c @@ -0,0 +1,364 @@ +/****************************************************************************** + * arch/ia64/xen/xen_pv_ops.c + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <linux/console.h> +#include <linux/irq.h> +#include <linux/kernel.h> +#include <linux/pm.h> + +#include <asm/xen/hypervisor.h> +#include <asm/xen/xencomm.h> +#include <asm/xen/privop.h> + +#include "irq_xen.h" +#include "time.h" + +/*************************************************************************** + * general info + */ +static struct pv_info xen_info __initdata = { + .kernel_rpl = 2, /* or 1: determin at runtime */ + .paravirt_enabled = 1, + .name = "Xen/ia64", +}; + +#define IA64_RSC_PL_SHIFT 2 +#define IA64_RSC_PL_BIT_SIZE 2 +#define IA64_RSC_PL_MASK \ + (((1UL << IA64_RSC_PL_BIT_SIZE) - 1) << IA64_RSC_PL_SHIFT) + +static void __init +xen_info_init(void) +{ + /* Xenified Linux/ia64 may run on pl = 1 or 2. + * determin at run time. */ + unsigned long rsc = ia64_getreg(_IA64_REG_AR_RSC); + unsigned int rpl = (rsc & IA64_RSC_PL_MASK) >> IA64_RSC_PL_SHIFT; + xen_info.kernel_rpl = rpl; +} + +/*************************************************************************** + * pv_init_ops + * initialization hooks. + */ + +static void +xen_panic_hypercall(struct unw_frame_info *info, void *arg) +{ + current->thread.ksp = (__u64)info->sw - 16; + HYPERVISOR_shutdown(SHUTDOWN_crash); + /* we're never actually going to get here... */ +} + +static int +xen_panic_event(struct notifier_block *this, unsigned long event, void *ptr) +{ + unw_init_running(xen_panic_hypercall, NULL); + /* we're never actually going to get here... */ + return NOTIFY_DONE; +} + +static struct notifier_block xen_panic_block = { + xen_panic_event, NULL, 0 /* try to go last */ +}; + +static void xen_pm_power_off(void) +{ + local_irq_disable(); + HYPERVISOR_shutdown(SHUTDOWN_poweroff); +} + +static void __init +xen_banner(void) +{ + printk(KERN_INFO + "Running on Xen! pl = %d start_info_pfn=0x%lx nr_pages=%ld " + "flags=0x%x\n", + xen_info.kernel_rpl, + HYPERVISOR_shared_info->arch.start_info_pfn, + xen_start_info->nr_pages, xen_start_info->flags); +} + +static int __init +xen_reserve_memory(struct rsvd_region *region) +{ + region->start = (unsigned long)__va( + (HYPERVISOR_shared_info->arch.start_info_pfn << PAGE_SHIFT)); + region->end = region->start + PAGE_SIZE; + return 1; +} + +static void __init +xen_arch_setup_early(void) +{ + struct shared_info *s; + BUG_ON(!xen_pv_domain()); + + s = HYPERVISOR_shared_info; + xen_start_info = __va(s->arch.start_info_pfn << PAGE_SHIFT); + + /* Must be done before any hypercall. */ + xencomm_initialize(); + + xen_setup_features(); + /* Register a call for panic conditions. */ + atomic_notifier_chain_register(&panic_notifier_list, + &xen_panic_block); + pm_power_off = xen_pm_power_off; + + xen_ia64_enable_opt_feature(); +} + +static void __init +xen_arch_setup_console(char **cmdline_p) +{ + add_preferred_console("xenboot", 0, NULL); + add_preferred_console("tty", 0, NULL); + /* use hvc_xen */ + add_preferred_console("hvc", 0, NULL); + +#if !defined(CONFIG_VT) || !defined(CONFIG_DUMMY_CONSOLE) + conswitchp = NULL; +#endif +} + +static int __init +xen_arch_setup_nomca(void) +{ + return 1; +} + +static void __init +xen_post_smp_prepare_boot_cpu(void) +{ + xen_setup_vcpu_info_placement(); +} + +static const struct pv_init_ops xen_init_ops __initdata = { + .banner = xen_banner, + + .reserve_memory = xen_reserve_memory, + + .arch_setup_early = xen_arch_setup_early, + .arch_setup_console = xen_arch_setup_console, + .arch_setup_nomca = xen_arch_setup_nomca, + + .post_smp_prepare_boot_cpu = xen_post_smp_prepare_boot_cpu, +}; + +/*************************************************************************** + * pv_cpu_ops + * intrinsics hooks. + */ + +static void xen_setreg(int regnum, unsigned long val) +{ + switch (regnum) { + case _IA64_REG_AR_KR0 ... _IA64_REG_AR_KR7: + xen_set_kr(regnum - _IA64_REG_AR_KR0, val); + break; +#ifdef CONFIG_IA32_SUPPORT + case _IA64_REG_AR_EFLAG: + xen_set_eflag(val); + break; +#endif + case _IA64_REG_CR_TPR: + xen_set_tpr(val); + break; + case _IA64_REG_CR_ITM: + xen_set_itm(val); + break; + case _IA64_REG_CR_EOI: + xen_eoi(val); + break; + default: + ia64_native_setreg_func(regnum, val); + break; + } +} + +static unsigned long xen_getreg(int regnum) +{ + unsigned long res; + + switch (regnum) { + case _IA64_REG_PSR: + res = xen_get_psr(); + break; +#ifdef CONFIG_IA32_SUPPORT + case _IA64_REG_AR_EFLAG: + res = xen_get_eflag(); + break; +#endif + case _IA64_REG_CR_IVR: + res = xen_get_ivr(); + break; + case _IA64_REG_CR_TPR: + res = xen_get_tpr(); + break; + default: + res = ia64_native_getreg_func(regnum); + break; + } + return res; +} + +/* turning on interrupts is a bit more complicated.. write to the + * memory-mapped virtual psr.i bit first (to avoid race condition), + * then if any interrupts were pending, we have to execute a hyperprivop + * to ensure the pending interrupt gets delivered; else we're done! */ +static void +xen_ssm_i(void) +{ + int old = xen_get_virtual_psr_i(); + xen_set_virtual_psr_i(1); + barrier(); + if (!old && xen_get_virtual_pend()) + xen_hyper_ssm_i(); +} + +/* turning off interrupts can be paravirtualized simply by writing + * to a memory-mapped virtual psr.i bit (implemented as a 16-bit bool) */ +static void +xen_rsm_i(void) +{ + xen_set_virtual_psr_i(0); + barrier(); +} + +static unsigned long +xen_get_psr_i(void) +{ + return xen_get_virtual_psr_i() ? IA64_PSR_I : 0; +} + +static void +xen_intrin_local_irq_restore(unsigned long mask) +{ + if (mask & IA64_PSR_I) + xen_ssm_i(); + else + xen_rsm_i(); +} + +static const struct pv_cpu_ops xen_cpu_ops __initdata = { + .fc = xen_fc, + .thash = xen_thash, + .get_cpuid = xen_get_cpuid, + .get_pmd = xen_get_pmd, + .getreg = xen_getreg, + .setreg = xen_setreg, + .ptcga = xen_ptcga, + .get_rr = xen_get_rr, + .set_rr = xen_set_rr, + .set_rr0_to_rr4 = xen_set_rr0_to_rr4, + .ssm_i = xen_ssm_i, + .rsm_i = xen_rsm_i, + .get_psr_i = xen_get_psr_i, + .intrin_local_irq_restore + = xen_intrin_local_irq_restore, +}; + +/****************************************************************************** + * replacement of hand written assembly codes. + */ + +extern char xen_switch_to; +extern char xen_leave_syscall; +extern char xen_work_processed_syscall; +extern char xen_leave_kernel; + +const struct pv_cpu_asm_switch xen_cpu_asm_switch = { + .switch_to = (unsigned long)&xen_switch_to, + .leave_syscall = (unsigned long)&xen_leave_syscall, + .work_processed_syscall = (unsigned long)&xen_work_processed_syscall, + .leave_kernel = (unsigned long)&xen_leave_kernel, +}; + +/*************************************************************************** + * pv_iosapic_ops + * iosapic read/write hooks. + */ +static void +xen_pcat_compat_init(void) +{ + /* nothing */ +} + +static struct irq_chip* +xen_iosapic_get_irq_chip(unsigned long trigger) +{ + return NULL; +} + +static unsigned int +xen_iosapic_read(char __iomem *iosapic, unsigned int reg) +{ + struct physdev_apic apic_op; + int ret; + + apic_op.apic_physbase = (unsigned long)iosapic - + __IA64_UNCACHED_OFFSET; + apic_op.reg = reg; + ret = HYPERVISOR_physdev_op(PHYSDEVOP_apic_read, &apic_op); + if (ret) + return ret; + return apic_op.value; +} + +static void +xen_iosapic_write(char __iomem *iosapic, unsigned int reg, u32 val) +{ + struct physdev_apic apic_op; + + apic_op.apic_physbase = (unsigned long)iosapic - + __IA64_UNCACHED_OFFSET; + apic_op.reg = reg; + apic_op.value = val; + HYPERVISOR_physdev_op(PHYSDEVOP_apic_write, &apic_op); +} + +static const struct pv_iosapic_ops xen_iosapic_ops __initdata = { + .pcat_compat_init = xen_pcat_compat_init, + .__get_irq_chip = xen_iosapic_get_irq_chip, + + .__read = xen_iosapic_read, + .__write = xen_iosapic_write, +}; + +/*************************************************************************** + * pv_ops initialization + */ + +void __init +xen_setup_pv_ops(void) +{ + xen_info_init(); + pv_info = xen_info; + pv_init_ops = xen_init_ops; + pv_cpu_ops = xen_cpu_ops; + pv_iosapic_ops = xen_iosapic_ops; + pv_irq_ops = xen_irq_ops; + pv_time_ops = xen_time_ops; + + paravirt_cpu_asm_init(&xen_cpu_asm_switch); +} diff --git a/arch/ia64/xen/xencomm.c b/arch/ia64/xen/xencomm.c new file mode 100644 index 0000000..1f5d7ac --- /dev/null +++ b/arch/ia64/xen/xencomm.c @@ -0,0 +1,105 @@ +/* + * Copyright (C) 2006 Hollis Blanchard <hollisb@us.ibm.com>, IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#include <linux/mm.h> + +static unsigned long kernel_virtual_offset; +static int is_xencomm_initialized; + +/* for xen early printk. It uses console io hypercall which uses xencomm. + * However early printk may use it before xencomm initialization. + */ +int +xencomm_is_initialized(void) +{ + return is_xencomm_initialized; +} + +void +xencomm_initialize(void) +{ + kernel_virtual_offset = KERNEL_START - ia64_tpa(KERNEL_START); + is_xencomm_initialized = 1; +} + +/* Translate virtual address to physical address. */ +unsigned long +xencomm_vtop(unsigned long vaddr) +{ + struct page *page; + struct vm_area_struct *vma; + + if (vaddr == 0) + return 0UL; + + if (REGION_NUMBER(vaddr) == 5) { + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + pte_t *ptep; + + /* On ia64, TASK_SIZE refers to current. It is not initialized + during boot. + Furthermore the kernel is relocatable and __pa() doesn't + work on addresses. */ + if (vaddr >= KERNEL_START + && vaddr < (KERNEL_START + KERNEL_TR_PAGE_SIZE)) + return vaddr - kernel_virtual_offset; + + /* In kernel area -- virtually mapped. */ + pgd = pgd_offset_k(vaddr); + if (pgd_none(*pgd) || pgd_bad(*pgd)) + return ~0UL; + + pud = pud_offset(pgd, vaddr); + if (pud_none(*pud) || pud_bad(*pud)) + return ~0UL; + + pmd = pmd_offset(pud, vaddr); + if (pmd_none(*pmd) || pmd_bad(*pmd)) + return ~0UL; + + ptep = pte_offset_kernel(pmd, vaddr); + if (!ptep) + return ~0UL; + + return (pte_val(*ptep) & _PFN_MASK) | (vaddr & ~PAGE_MASK); + } + + if (vaddr > TASK_SIZE) { + /* percpu variables */ + if (REGION_NUMBER(vaddr) == 7 && + REGION_OFFSET(vaddr) >= (1ULL << IA64_MAX_PHYS_BITS)) + ia64_tpa(vaddr); + + /* kernel address */ + return __pa(vaddr); + } + + /* XXX double-check (lack of) locking */ + vma = find_extend_vma(current->mm, vaddr); + if (!vma) + return ~0UL; + + /* We assume the page is modified. */ + page = follow_page(vma, vaddr, FOLL_WRITE | FOLL_TOUCH); + if (!page) + return ~0UL; + + return (page_to_pfn(page) << PAGE_SHIFT) | (vaddr & ~PAGE_MASK); +} diff --git a/arch/ia64/xen/xenivt.S b/arch/ia64/xen/xenivt.S new file mode 100644 index 0000000..3e71d50 --- /dev/null +++ b/arch/ia64/xen/xenivt.S @@ -0,0 +1,52 @@ +/* + * arch/ia64/xen/ivt.S + * + * Copyright (C) 2005 Hewlett-Packard Co + * Dan Magenheimer <dan.magenheimer@hp.com> + * + * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp> + * VA Linux Systems Japan K.K. + * pv_ops. + */ + +#include <asm/asmmacro.h> +#include <asm/kregs.h> +#include <asm/pgtable.h> + +#include "../kernel/minstate.h" + + .section .text,"ax" +GLOBAL_ENTRY(xen_event_callback) + mov r31=pr // prepare to save predicates + ;; + SAVE_MIN_WITH_COVER // uses r31; defines r2 and r3 + ;; + movl r3=XSI_PSR_IC + mov r14=1 + ;; + st4 [r3]=r14 + ;; + adds r3=8,r2 // set up second base pointer for SAVE_REST + srlz.i // ensure everybody knows psr.ic is back on + ;; + SAVE_REST + ;; +1: + alloc r14=ar.pfs,0,0,1,0 // must be first in an insn group + add out0=16,sp // pass pointer to pt_regs as first arg + ;; + br.call.sptk.many b0=xen_evtchn_do_upcall + ;; + movl r20=XSI_PSR_I_ADDR + ;; + ld8 r20=[r20] + ;; + adds r20=-1,r20 // vcpu_info->evtchn_upcall_pending + ;; + ld1 r20=[r20] + ;; + cmp.ne p6,p0=r20,r0 // if there are pending events, + (p6) br.spnt.few 1b // call evtchn_do_upcall again. + br.sptk.many xen_leave_kernel // we know ia64_leave_kernel is + // paravirtualized as xen_leave_kernel +END(xen_event_callback) diff --git a/arch/ia64/xen/xensetup.S b/arch/ia64/xen/xensetup.S new file mode 100644 index 0000000..28fed1f --- /dev/null +++ b/arch/ia64/xen/xensetup.S @@ -0,0 +1,83 @@ +/* + * Support routines for Xen + * + * Copyright (C) 2005 Dan Magenheimer <dan.magenheimer@hp.com> + */ + +#include <asm/processor.h> +#include <asm/asmmacro.h> +#include <asm/pgtable.h> +#include <asm/system.h> +#include <asm/paravirt.h> +#include <asm/xen/privop.h> +#include <linux/elfnote.h> +#include <linux/init.h> +#include <xen/interface/elfnote.h> + + .section .data.read_mostly + .align 8 + .global xen_domain_type +xen_domain_type: + data4 XEN_NATIVE_ASM + .previous + + __INIT +ENTRY(startup_xen) + // Calculate load offset. + // The constant, LOAD_OFFSET, can't be used because the boot + // loader doesn't always load to the LMA specified by the vmlinux.lds. + mov r9=ip // must be the first instruction to make sure + // that r9 = the physical address of startup_xen. + // Usually r9 = startup_xen - LOAD_OFFSET + movl r8=startup_xen + ;; + sub r9=r9,r8 // Usually r9 = -LOAD_OFFSET. + + mov r10=PARAVIRT_HYPERVISOR_TYPE_XEN + movl r11=_start + ;; + add r11=r11,r9 + movl r8=hypervisor_type + ;; + add r8=r8,r9 + mov b0=r11 + ;; + st8 [r8]=r10 + br.cond.sptk.many b0 + ;; +END(startup_xen) + + ELFNOTE(Xen, XEN_ELFNOTE_GUEST_OS, .asciz "linux") + ELFNOTE(Xen, XEN_ELFNOTE_GUEST_VERSION, .asciz "2.6") + ELFNOTE(Xen, XEN_ELFNOTE_XEN_VERSION, .asciz "xen-3.0") + ELFNOTE(Xen, XEN_ELFNOTE_ENTRY, data8.ua startup_xen - LOAD_OFFSET) + +#define isBP p3 // are we the Bootstrap Processor? + + .text + +GLOBAL_ENTRY(xen_setup_hook) + mov r8=XEN_PV_DOMAIN_ASM +(isBP) movl r9=xen_domain_type;; +(isBP) st4 [r9]=r8 + movl r10=xen_ivt;; + + mov cr.iva=r10 + + /* Set xsi base. */ +#define FW_HYPERCALL_SET_SHARED_INFO_VA 0x600 +(isBP) mov r2=FW_HYPERCALL_SET_SHARED_INFO_VA +(isBP) movl r28=XSI_BASE;; +(isBP) break 0x1000;; + + /* setup pv_ops */ +(isBP) mov r4=rp + ;; +(isBP) br.call.sptk.many rp=xen_setup_pv_ops + ;; +(isBP) mov rp=r4 + ;; + + br.ret.sptk.many rp + ;; +END(xen_setup_hook) diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c index 1ee9499..bbb3cae 100644 --- a/drivers/ata/libata-core.c +++ b/drivers/ata/libata-core.c @@ -5373,6 +5373,8 @@ struct ata_port *ata_port_alloc(struct ata_host *host) #ifdef CONFIG_ATA_SFF INIT_DELAYED_WORK(&ap->port_task, ata_pio_task); +#else + INIT_DELAYED_WORK(&ap->port_task, NULL); #endif INIT_DELAYED_WORK(&ap->hotplug_task, ata_scsi_hotplug); INIT_WORK(&ap->scsi_rescan_task, ata_scsi_dev_rescan); diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c index a93247c..5d687d7 100644 --- a/drivers/ata/libata-eh.c +++ b/drivers/ata/libata-eh.c @@ -1206,7 +1206,10 @@ void ata_eh_about_to_do(struct ata_link *link, struct ata_device *dev, ata_eh_clear_action(link, dev, ehi, action); - if (!(ehc->i.flags & ATA_EHI_QUIET)) + /* About to take EH action, set RECOVERED. Ignore actions on + * slave links as master will do them again. + */ + if (!(ehc->i.flags & ATA_EHI_QUIET) && link != ap->slave_link) ap->pflags |= ATA_PFLAG_RECOVERED; spin_unlock_irqrestore(ap->lock, flags); @@ -2010,8 +2013,13 @@ void ata_eh_autopsy(struct ata_port *ap) struct ata_eh_context *mehc = &ap->link.eh_context; struct ata_eh_context *sehc = &ap->slave_link->eh_context; + /* transfer control flags from master to slave */ + sehc->i.flags |= mehc->i.flags & ATA_EHI_TO_SLAVE_MASK; + + /* perform autopsy on the slave link */ ata_eh_link_autopsy(ap->slave_link); + /* transfer actions from slave to master and clear slave */ ata_eh_about_to_do(ap->slave_link, NULL, ATA_EH_ALL_ACTIONS); mehc->i.action |= sehc->i.action; mehc->i.dev_action[1] |= sehc->i.dev_action[1]; @@ -2447,14 +2455,14 @@ int ata_eh_reset(struct ata_link *link, int classify, dev->pio_mode = XFER_PIO_0; dev->flags &= ~ATA_DFLAG_SLEEPING; - if (ata_phys_link_offline(ata_dev_phys_link(dev))) - continue; - - /* apply class override */ - if (lflags & ATA_LFLAG_ASSUME_ATA) - classes[dev->devno] = ATA_DEV_ATA; - else if (lflags & ATA_LFLAG_ASSUME_SEMB) - classes[dev->devno] = ATA_DEV_SEMB_UNSUP; /* not yet */ + if (!ata_phys_link_offline(ata_dev_phys_link(dev))) { + /* apply class override */ + if (lflags & ATA_LFLAG_ASSUME_ATA) + classes[dev->devno] = ATA_DEV_ATA; + else if (lflags & ATA_LFLAG_ASSUME_SEMB) + classes[dev->devno] = ATA_DEV_SEMB_UNSUP; + } else + classes[dev->devno] = ATA_DEV_NONE; } /* record current link speed */ diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c index 2a4c516..4b47394 100644 --- a/drivers/ata/libata-sff.c +++ b/drivers/ata/libata-sff.c @@ -2153,8 +2153,17 @@ void ata_sff_error_handler(struct ata_port *ap) */ void ata_sff_post_internal_cmd(struct ata_queued_cmd *qc) { - if (qc->ap->ioaddr.bmdma_addr) + struct ata_port *ap = qc->ap; + unsigned long flags; + + spin_lock_irqsave(ap->lock, flags); + + ap->hsm_task_state = HSM_ST_IDLE; + + if (ap->ioaddr.bmdma_addr) ata_bmdma_stop(qc); + + spin_unlock_irqrestore(ap->lock, flags); } /** diff --git a/drivers/ata/sata_via.c b/drivers/ata/sata_via.c index 1cfa745..5b72e73 100644 --- a/drivers/ata/sata_via.c +++ b/drivers/ata/sata_via.c @@ -70,6 +70,7 @@ enum { static int svia_init_one(struct pci_dev *pdev, const struct pci_device_id *ent); static int svia_scr_read(struct ata_link *link, unsigned int sc_reg, u32 *val); static int svia_scr_write(struct ata_link *link, unsigned int sc_reg, u32 val); +static void svia_tf_load(struct ata_port *ap, const struct ata_taskfile *tf); static void svia_noop_freeze(struct ata_port *ap); static int vt6420_prereset(struct ata_link *link, unsigned long deadline); static int vt6421_pata_cable_detect(struct ata_port *ap); @@ -103,21 +104,26 @@ static struct scsi_host_template svia_sht = { ATA_BMDMA_SHT(DRV_NAME), }; -static struct ata_port_operations vt6420_sata_ops = { +static struct ata_port_operations svia_base_ops = { .inherits = &ata_bmdma_port_ops, + .sff_tf_load = svia_tf_load, +}; + +static struct ata_port_operations vt6420_sata_ops = { + .inherits = &svia_base_ops, .freeze = svia_noop_freeze, .prereset = vt6420_prereset, }; static struct ata_port_operations vt6421_pata_ops = { - .inherits = &ata_bmdma_port_ops, + .inherits = &svia_base_ops, .cable_detect = vt6421_pata_cable_detect, .set_piomode = vt6421_set_pio_mode, .set_dmamode = vt6421_set_dma_mode, }; static struct ata_port_operations vt6421_sata_ops = { - .inherits = &ata_bmdma_port_ops, + .inherits = &svia_base_ops, .scr_read = svia_scr_read, .scr_write = svia_scr_write, }; @@ -168,6 +174,29 @@ static int svia_scr_write(struct ata_link *link, unsigned int sc_reg, u32 val) return 0; } +/** + * svia_tf_load - send taskfile registers to host controller + * @ap: Port to which output is sent + * @tf: ATA taskfile register set + * + * Outputs ATA taskfile to standard ATA host controller. + * + * This is to fix the internal bug of via chipsets, which will + * reset the device register after changing the IEN bit on ctl + * register. + */ +static void svia_tf_load(struct ata_port *ap, const struct ata_taskfile *tf) +{ + struct ata_taskfile ttf; + + if (tf->ctl != ap->last_ctl) { + ttf = *tf; + ttf.flags |= ATA_TFLAG_DEVICE; + tf = &ttf; + } + ata_sff_tf_load(ap, tf); +} + static void svia_noop_freeze(struct ata_port *ap) { /* Some VIA controllers choke if ATA_NIEN is manipulated in diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 49c45fe..5c54fc2 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -406,19 +406,15 @@ static int register_snoop_agent(struct ib_mad_qp_info *qp_info, if (i == qp_info->snoop_table_size) { /* Grow table. */ - new_snoop_table = kmalloc(sizeof mad_snoop_priv * - qp_info->snoop_table_size + 1, - GFP_ATOMIC); + new_snoop_table = krealloc(qp_info->snoop_table, + sizeof mad_snoop_priv * + (qp_info->snoop_table_size + 1), + GFP_ATOMIC); if (!new_snoop_table) { i = -ENOMEM; goto out; } - if (qp_info->snoop_table) { - memcpy(new_snoop_table, qp_info->snoop_table, - sizeof mad_snoop_priv * - qp_info->snoop_table_size); - kfree(qp_info->snoop_table); - } + qp_info->snoop_table = new_snoop_table; qp_info->snoop_table_size++; } diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c index 3ddacf3..4346a24 100644 --- a/drivers/infiniband/core/ucma.c +++ b/drivers/infiniband/core/ucma.c @@ -904,8 +904,8 @@ static ssize_t ucma_join_multicast(struct ucma_file *file, mutex_lock(&file->mut); mc = ucma_alloc_multicast(ctx); - if (IS_ERR(mc)) { - ret = PTR_ERR(mc); + if (!mc) { + ret = -ENOMEM; goto err1; } diff --git a/drivers/infiniband/hw/cxgb3/iwch_cm.c b/drivers/infiniband/hw/cxgb3/iwch_cm.c index c325c44..44e936e 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_cm.c +++ b/drivers/infiniband/hw/cxgb3/iwch_cm.c @@ -1942,6 +1942,7 @@ fail4: fail3: cxgb3_free_atid(ep->com.tdev, ep->atid); fail2: + cm_id->rem_ref(cm_id); put_ep(&ep->com); out: return err; diff --git a/drivers/infiniband/hw/ehca/ehca_classes.h b/drivers/infiniband/hw/ehca/ehca_classes.h index 5d7b785..4df887a 100644 --- a/drivers/infiniband/hw/ehca/ehca_classes.h +++ b/drivers/infiniband/hw/ehca/ehca_classes.h @@ -128,6 +128,8 @@ struct ehca_shca { /* MR pgsize: bit 0-3 means 4K, 64K, 1M, 16M respectively */ u32 hca_cap_mr_pgsize; int max_mtu; + int max_num_qps; + int max_num_cqs; atomic_t num_cqs; atomic_t num_qps; }; diff --git a/drivers/infiniband/hw/ehca/ehca_cq.c b/drivers/infiniband/hw/ehca/ehca_cq.c index 33647a9..2f4c28a 100644 --- a/drivers/infiniband/hw/ehca/ehca_cq.c +++ b/drivers/infiniband/hw/ehca/ehca_cq.c @@ -132,9 +132,9 @@ struct ib_cq *ehca_create_cq(struct ib_device *device, int cqe, int comp_vector, if (cqe >= 0xFFFFFFFF - 64 - additional_cqe) return ERR_PTR(-EINVAL); - if (!atomic_add_unless(&shca->num_cqs, 1, ehca_max_cq)) { + if (!atomic_add_unless(&shca->num_cqs, 1, shca->max_num_cqs)) { ehca_err(device, "Unable to create CQ, max number of %i " - "CQs reached.", ehca_max_cq); + "CQs reached.", shca->max_num_cqs); ehca_err(device, "To increase the maximum number of CQs " "use the number_of_cqs module parameter.\n"); return ERR_PTR(-ENOSPC); diff --git a/drivers/infiniband/hw/ehca/ehca_main.c b/drivers/infiniband/hw/ehca/ehca_main.c index 598844d..bb02a86 100644 --- a/drivers/infiniband/hw/ehca/ehca_main.c +++ b/drivers/infiniband/hw/ehca/ehca_main.c @@ -44,6 +44,8 @@ #include <linux/slab.h> #endif +#include <linux/notifier.h> +#include <linux/memory.h> #include "ehca_classes.h" #include "ehca_iverbs.h" #include "ehca_mrmw.h" @@ -366,22 +368,23 @@ static int ehca_sense_attributes(struct ehca_shca *shca) shca->hca_cap_mr_pgsize |= pgsize_map[i + 1]; /* Set maximum number of CQs and QPs to calculate EQ size */ - if (ehca_max_qp == -1) - ehca_max_qp = min_t(int, rblock->max_qp, EHCA_MAX_NUM_QUEUES); - else if (ehca_max_qp < 1 || ehca_max_qp > rblock->max_qp) { - ehca_gen_err("Requested number of QPs is out of range (1 - %i) " - "specified by HW", rblock->max_qp); - ret = -EINVAL; - goto sense_attributes1; + if (shca->max_num_qps == -1) + shca->max_num_qps = min_t(int, rblock->max_qp, + EHCA_MAX_NUM_QUEUES); + else if (shca->max_num_qps < 1 || shca->max_num_qps > rblock->max_qp) { + ehca_gen_warn("The requested number of QPs is out of range " + "(1 - %i) specified by HW. Value is set to %i", + rblock->max_qp, rblock->max_qp); + shca->max_num_qps = rblock->max_qp; } - if (ehca_max_cq == -1) - ehca_max_cq = min_t(int, rblock->max_cq, EHCA_MAX_NUM_QUEUES); - else if (ehca_max_cq < 1 || ehca_max_cq > rblock->max_cq) { - ehca_gen_err("Requested number of CQs is out of range (1 - %i) " - "specified by HW", rblock->max_cq); - ret = -EINVAL; - goto sense_attributes1; + if (shca->max_num_cqs == -1) + shca->max_num_cqs = min_t(int, rblock->max_cq, + EHCA_MAX_NUM_QUEUES); + else if (shca->max_num_cqs < 1 || shca->max_num_cqs > rblock->max_cq) { + ehca_gen_warn("The requested number of CQs is out of range " + "(1 - %i) specified by HW. Value is set to %i", + rblock->max_cq, rblock->max_cq); } /* query max MTU from first port -- it's the same for all ports */ @@ -733,9 +736,13 @@ static int __devinit ehca_probe(struct of_device *dev, ehca_gen_err("Cannot allocate shca memory."); return -ENOMEM; } + mutex_init(&shca->modify_mutex); atomic_set(&shca->num_cqs, 0); atomic_set(&shca->num_qps, 0); + shca->max_num_qps = ehca_max_qp; + shca->max_num_cqs = ehca_max_cq; + for (i = 0; i < ARRAY_SIZE(shca->sport); i++) spin_lock_init(&shca->sport[i].mod_sqp_lock); @@ -755,7 +762,7 @@ static int __devinit ehca_probe(struct of_device *dev, goto probe1; } - eq_size = 2 * ehca_max_cq + 4 * ehca_max_qp; + eq_size = 2 * shca->max_num_cqs + 4 * shca->max_num_qps; /* create event queues */ ret = ehca_create_eq(shca, &shca->eq, EHCA_EQ, eq_size); if (ret) { @@ -964,6 +971,41 @@ void ehca_poll_eqs(unsigned long data) spin_unlock(&shca_list_lock); } +static int ehca_mem_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + static unsigned long ehca_dmem_warn_time; + + switch (action) { + case MEM_CANCEL_OFFLINE: + case MEM_CANCEL_ONLINE: + case MEM_ONLINE: + case MEM_OFFLINE: + return NOTIFY_OK; + case MEM_GOING_ONLINE: + case MEM_GOING_OFFLINE: + /* only ok if no hca is attached to the lpar */ + spin_lock(&shca_list_lock); + if (list_empty(&shca_list)) { + spin_unlock(&shca_list_lock); + return NOTIFY_OK; + } else { + spin_unlock(&shca_list_lock); + if (printk_timed_ratelimit(&ehca_dmem_warn_time, + 30 * 1000)) + ehca_gen_err("DMEM operations are not allowed" + "as long as an ehca adapter is" + "attached to the LPAR"); + return NOTIFY_BAD; + } + } + return NOTIFY_OK; +} + +static struct notifier_block ehca_mem_nb = { + .notifier_call = ehca_mem_notifier, +}; + static int __init ehca_module_init(void) { int ret; @@ -991,6 +1033,12 @@ static int __init ehca_module_init(void) goto module_init2; } + ret = register_memory_notifier(&ehca_mem_nb); + if (ret) { + ehca_gen_err("Failed registering memory add/remove notifier"); + goto module_init3; + } + if (ehca_poll_all_eqs != 1) { ehca_gen_err("WARNING!!!"); ehca_gen_err("It is possible to lose interrupts."); @@ -1003,6 +1051,9 @@ static int __init ehca_module_init(void) return 0; +module_init3: + ibmebus_unregister_driver(&ehca_driver); + module_init2: ehca_destroy_slab_caches(); @@ -1018,6 +1069,8 @@ static void __exit ehca_module_exit(void) ibmebus_unregister_driver(&ehca_driver); + unregister_memory_notifier(&ehca_mem_nb); + ehca_destroy_slab_caches(); ehca_destroy_comp_pool(); diff --git a/drivers/infiniband/hw/ehca/ehca_qp.c b/drivers/infiniband/hw/ehca/ehca_qp.c index 4dbe287..4d54b9f 100644 --- a/drivers/infiniband/hw/ehca/ehca_qp.c +++ b/drivers/infiniband/hw/ehca/ehca_qp.c @@ -465,9 +465,9 @@ static struct ehca_qp *internal_create_qp( u32 swqe_size = 0, rwqe_size = 0, ib_qp_num; unsigned long flags; - if (!atomic_add_unless(&shca->num_qps, 1, ehca_max_qp)) { + if (!atomic_add_unless(&shca->num_qps, 1, shca->max_num_qps)) { ehca_err(pd->device, "Unable to create QP, max number of %i " - "QPs reached.", ehca_max_qp); + "QPs reached.", shca->max_num_qps); ehca_err(pd->device, "To increase the maximum number of QPs " "use the number_of_qps module parameter.\n"); return ERR_PTR(-ENOSPC); @@ -502,6 +502,12 @@ static struct ehca_qp *internal_create_qp( if (init_attr->srq) { my_srq = container_of(init_attr->srq, struct ehca_qp, ib_srq); + if (qp_type == IB_QPT_UC) { + ehca_err(pd->device, "UC with SRQ not supported"); + atomic_dec(&shca->num_qps); + return ERR_PTR(-EINVAL); + } + has_srq = 1; parms.ext_type = EQPT_SRQBASE; parms.srq_qpn = my_srq->real_qp_num; diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c index cdca3a5..606f1e2 100644 --- a/drivers/infiniband/hw/mlx4/mad.c +++ b/drivers/infiniband/hw/mlx4/mad.c @@ -298,7 +298,7 @@ int mlx4_ib_mad_init(struct mlx4_ib_dev *dev) int p, q; int ret; - for (p = 0; p < dev->dev->caps.num_ports; ++p) + for (p = 0; p < dev->num_ports; ++p) for (q = 0; q <= 1; ++q) { agent = ib_register_mad_agent(&dev->ib_dev, p + 1, q ? IB_QPT_GSI : IB_QPT_SMI, @@ -314,7 +314,7 @@ int mlx4_ib_mad_init(struct mlx4_ib_dev *dev) return 0; err: - for (p = 0; p < dev->dev->caps.num_ports; ++p) + for (p = 0; p < dev->num_ports; ++p) for (q = 0; q <= 1; ++q) if (dev->send_agent[p][q]) ib_unregister_mad_agent(dev->send_agent[p][q]); @@ -327,7 +327,7 @@ void mlx4_ib_mad_cleanup(struct mlx4_ib_dev *dev) struct ib_mad_agent *agent; int p, q; - for (p = 0; p < dev->dev->caps.num_ports; ++p) { + for (p = 0; p < dev->num_ports; ++p) { for (q = 0; q <= 1; ++q) { agent = dev->send_agent[p][q]; dev->send_agent[p][q] = NULL; diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c index a3c2851..2e80f8f 100644 --- a/drivers/infiniband/hw/mlx4/main.c +++ b/drivers/infiniband/hw/mlx4/main.c @@ -574,7 +574,10 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) ibdev->ib_dev.owner = THIS_MODULE; ibdev->ib_dev.node_type = RDMA_NODE_IB_CA; ibdev->ib_dev.local_dma_lkey = dev->caps.reserved_lkey; - ibdev->ib_dev.phys_port_cnt = dev->caps.num_ports; + ibdev->num_ports = 0; + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB) + ibdev->num_ports++; + ibdev->ib_dev.phys_port_cnt = ibdev->num_ports; ibdev->ib_dev.num_comp_vectors = 1; ibdev->ib_dev.dma_device = &dev->pdev->dev; @@ -691,7 +694,7 @@ static void mlx4_ib_remove(struct mlx4_dev *dev, void *ibdev_ptr) struct mlx4_ib_dev *ibdev = ibdev_ptr; int p; - for (p = 1; p <= dev->caps.num_ports; ++p) + for (p = 1; p <= ibdev->num_ports; ++p) mlx4_CLOSE_PORT(dev, p); mlx4_ib_mad_cleanup(ibdev); @@ -706,6 +709,10 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, enum mlx4_dev_event event, int port) { struct ib_event ibev; + struct mlx4_ib_dev *ibdev = to_mdev((struct ib_device *) ibdev_ptr); + + if (port > ibdev->num_ports) + return; switch (event) { case MLX4_DEV_EVENT_PORT_UP: diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h index 6e2b0dc..9974e88 100644 --- a/drivers/infiniband/hw/mlx4/mlx4_ib.h +++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h @@ -162,6 +162,7 @@ struct mlx4_ib_ah { struct mlx4_ib_dev { struct ib_device ib_dev; struct mlx4_dev *dev; + int num_ports; void __iomem *uar_map; struct mlx4_uar priv_uar; diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c index baa01de..39167a7 100644 --- a/drivers/infiniband/hw/mlx4/qp.c +++ b/drivers/infiniband/hw/mlx4/qp.c @@ -451,6 +451,7 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd, struct ib_qp_init_attr *init_attr, struct ib_udata *udata, int sqpn, struct mlx4_ib_qp *qp) { + int qpn; int err; mutex_init(&qp->mutex); @@ -545,9 +546,17 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd, } } - err = mlx4_qp_alloc(dev->dev, sqpn, &qp->mqp); + if (sqpn) { + qpn = sqpn; + } else { + err = mlx4_qp_reserve_range(dev->dev, 1, 1, &qpn); + if (err) + goto err_wrid; + } + + err = mlx4_qp_alloc(dev->dev, qpn, &qp->mqp); if (err) - goto err_wrid; + goto err_qpn; /* * Hardware wants QPN written in big-endian order (after @@ -560,6 +569,10 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd, return 0; +err_qpn: + if (!sqpn) + mlx4_qp_release_range(dev->dev, qpn, 1); + err_wrid: if (pd->uobject) { if (!init_attr->srq) @@ -655,6 +668,10 @@ static void destroy_qp_common(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp, mlx4_ib_unlock_cqs(send_cq, recv_cq); mlx4_qp_free(dev->dev, &qp->mqp); + + if (!is_sqp(dev, qp)) + mlx4_qp_release_range(dev->dev, qp->mqp.qpn, 1); + mlx4_mtt_cleanup(dev->dev, &qp->mtt); if (is_user) { diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index 68ba5c3..e0c7dfa 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -507,6 +507,7 @@ int ipoib_pkey_dev_delay_open(struct net_device *dev); void ipoib_drain_cq(struct net_device *dev); void ipoib_set_ethtool_ops(struct net_device *dev); +int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca); #ifdef CONFIG_INFINIBAND_IPOIB_CM diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c index 66af5c1..e9795f6 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c @@ -42,6 +42,13 @@ static void ipoib_get_drvinfo(struct net_device *netdev, strncpy(drvinfo->driver, "ipoib", sizeof(drvinfo->driver) - 1); } +static u32 ipoib_get_rx_csum(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + return test_bit(IPOIB_FLAG_CSUM, &priv->flags) && + !test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); +} + static int ipoib_get_coalesce(struct net_device *dev, struct ethtool_coalesce *coal) { @@ -129,7 +136,7 @@ static void ipoib_get_ethtool_stats(struct net_device *dev, static const struct ethtool_ops ipoib_ethtool_ops = { .get_drvinfo = ipoib_get_drvinfo, - .get_tso = ethtool_op_get_tso, + .get_rx_csum = ipoib_get_rx_csum, .get_coalesce = ipoib_get_coalesce, .set_coalesce = ipoib_set_coalesce, .get_flags = ethtool_op_get_flags, diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 0e748ae..28eb6f0 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -685,10 +685,6 @@ int ipoib_ib_dev_open(struct net_device *dev) queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, round_jiffies_relative(HZ)); - init_timer(&priv->poll_timer); - priv->poll_timer.function = ipoib_ib_tx_timer_func; - priv->poll_timer.data = (unsigned long)dev; - set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags); return 0; @@ -906,6 +902,9 @@ int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port) return -ENODEV; } + setup_timer(&priv->poll_timer, ipoib_ib_tx_timer_func, + (unsigned long) dev); + if (dev->flags & IFF_UP) { if (ipoib_ib_dev_open(dev)) { ipoib_transport_dev_cleanup(dev); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index c0ee5143..fddded7 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -1173,11 +1173,48 @@ int ipoib_add_pkey_attr(struct net_device *dev) return device_create_file(&dev->dev, &dev_attr_pkey); } +int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca) +{ + struct ib_device_attr *device_attr; + int result = -ENOMEM; + + device_attr = kmalloc(sizeof *device_attr, GFP_KERNEL); + if (!device_attr) { + printk(KERN_WARNING "%s: allocation of %zu bytes failed\n", + hca->name, sizeof *device_attr); + return result; + } + + result = ib_query_device(hca, device_attr); + if (result) { + printk(KERN_WARNING "%s: ib_query_device failed (ret = %d)\n", + hca->name, result); + kfree(device_attr); + return result; + } + priv->hca_caps = device_attr->device_cap_flags; + + kfree(device_attr); + + if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) { + set_bit(IPOIB_FLAG_CSUM, &priv->flags); + priv->dev->features |= NETIF_F_SG | NETIF_F_IP_CSUM; + } + + if (lro) + priv->dev->features |= NETIF_F_LRO; + + if (priv->dev->features & NETIF_F_SG && priv->hca_caps & IB_DEVICE_UD_TSO) + priv->dev->features |= NETIF_F_TSO; + + return 0; +} + + static struct net_device *ipoib_add_port(const char *format, struct ib_device *hca, u8 port) { struct ipoib_dev_priv *priv; - struct ib_device_attr *device_attr; struct ib_port_attr attr; int result = -ENOMEM; @@ -1206,31 +1243,8 @@ static struct net_device *ipoib_add_port(const char *format, goto device_init_failed; } - device_attr = kmalloc(sizeof *device_attr, GFP_KERNEL); - if (!device_attr) { - printk(KERN_WARNING "%s: allocation of %zu bytes failed\n", - hca->name, sizeof *device_attr); + if (ipoib_set_dev_features(priv, hca)) goto device_init_failed; - } - - result = ib_query_device(hca, device_attr); - if (result) { - printk(KERN_WARNING "%s: ib_query_device failed (ret = %d)\n", - hca->name, result); - kfree(device_attr); - goto device_init_failed; - } - priv->hca_caps = device_attr->device_cap_flags; - - kfree(device_attr); - - if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) { - set_bit(IPOIB_FLAG_CSUM, &priv->flags); - priv->dev->features |= NETIF_F_SG | NETIF_F_IP_CSUM; - } - - if (lro) - priv->dev->features |= NETIF_F_LRO; /* * Set the full membership bit, so that we join the right @@ -1266,9 +1280,6 @@ static struct net_device *ipoib_add_port(const char *format, goto event_failed; } - if (priv->dev->features & NETIF_F_SG && priv->hca_caps & IB_DEVICE_UD_TSO) - priv->dev->features |= NETIF_F_TSO; - result = register_netdev(priv->dev); if (result) { printk(KERN_WARNING "%s: couldn't register ipoib port %d; error %d\n", diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c index b08eb56..2cf1a40 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c @@ -93,6 +93,10 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) priv->mcast_mtu = priv->admin_mtu = priv->dev->mtu; set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags); + result = ipoib_set_dev_features(priv, ppriv->ca); + if (result) + goto device_init_failed; + priv->pkey = pkey; memcpy(priv->dev->dev_addr, ppriv->dev->dev_addr, INFINIBAND_ALEN); diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index ad301ac..a5a9c6d 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -2504,6 +2504,15 @@ config PASEMI_MAC This driver supports the on-chip 1/10Gbit Ethernet controller on PA Semi's PWRficient line of chips. +config MLX4_EN + tristate "Mellanox Technologies 10Gbit Ethernet support" + depends on PCI && INET + select MLX4_CORE + select INET_LRO + help + This driver supports Mellanox Technologies ConnectX Ethernet + devices. + config MLX4_CORE tristate depends on PCI diff --git a/drivers/net/mlx4/Makefile b/drivers/net/mlx4/Makefile index 0952a65..a7a97bf 100644 --- a/drivers/net/mlx4/Makefile +++ b/drivers/net/mlx4/Makefile @@ -1,4 +1,9 @@ obj-$(CONFIG_MLX4_CORE) += mlx4_core.o mlx4_core-y := alloc.o catas.o cmd.o cq.o eq.o fw.o icm.o intf.o main.o mcg.o \ - mr.o pd.o profile.o qp.o reset.o srq.o + mr.o pd.o port.o profile.o qp.o reset.o srq.o + +obj-$(CONFIG_MLX4_EN) += mlx4_en.o + +mlx4_en-y := en_main.o en_tx.o en_rx.o en_params.o en_port.o en_cq.o \ + en_resources.o en_netdev.o diff --git a/drivers/net/mlx4/alloc.c b/drivers/net/mlx4/alloc.c index b411b79..ad95d5f 100644 --- a/drivers/net/mlx4/alloc.c +++ b/drivers/net/mlx4/alloc.c @@ -48,13 +48,16 @@ u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap) obj = find_next_zero_bit(bitmap->table, bitmap->max, bitmap->last); if (obj >= bitmap->max) { - bitmap->top = (bitmap->top + bitmap->max) & bitmap->mask; + bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top) + & bitmap->mask; obj = find_first_zero_bit(bitmap->table, bitmap->max); } if (obj < bitmap->max) { set_bit(obj, bitmap->table); - bitmap->last = (obj + 1) & (bitmap->max - 1); + bitmap->last = (obj + 1); + if (bitmap->last == bitmap->max) + bitmap->last = 0; obj |= bitmap->top; } else obj = -1; @@ -66,16 +69,90 @@ u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap) void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj) { - obj &= bitmap->max - 1; + mlx4_bitmap_free_range(bitmap, obj, 1); +} + +static unsigned long find_aligned_range(unsigned long *bitmap, + u32 start, u32 nbits, + int len, int align) +{ + unsigned long end, i; + +again: + start = ALIGN(start, align); + + while ((start < nbits) && test_bit(start, bitmap)) + start += align; + + if (start >= nbits) + return -1; + + end = start+len; + if (end > nbits) + return -1; + + for (i = start + 1; i < end; i++) { + if (test_bit(i, bitmap)) { + start = i + 1; + goto again; + } + } + + return start; +} + +u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt, int align) +{ + u32 obj, i; + + if (likely(cnt == 1 && align == 1)) + return mlx4_bitmap_alloc(bitmap); + + spin_lock(&bitmap->lock); + + obj = find_aligned_range(bitmap->table, bitmap->last, + bitmap->max, cnt, align); + if (obj >= bitmap->max) { + bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top) + & bitmap->mask; + obj = find_aligned_range(bitmap->table, 0, bitmap->max, + cnt, align); + } + + if (obj < bitmap->max) { + for (i = 0; i < cnt; i++) + set_bit(obj + i, bitmap->table); + if (obj == bitmap->last) { + bitmap->last = (obj + cnt); + if (bitmap->last >= bitmap->max) + bitmap->last = 0; + } + obj |= bitmap->top; + } else + obj = -1; + + spin_unlock(&bitmap->lock); + + return obj; +} + +void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt) +{ + u32 i; + + obj &= bitmap->max + bitmap->reserved_top - 1; spin_lock(&bitmap->lock); - clear_bit(obj, bitmap->table); + for (i = 0; i < cnt; i++) + clear_bit(obj + i, bitmap->table); bitmap->last = min(bitmap->last, obj); - bitmap->top = (bitmap->top + bitmap->max) & bitmap->mask; + bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top) + & bitmap->mask; spin_unlock(&bitmap->lock); } -int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, u32 reserved) +int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, + u32 reserved_bot, u32 reserved_top) { int i; @@ -85,14 +162,16 @@ int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, u32 reserved bitmap->last = 0; bitmap->top = 0; - bitmap->max = num; + bitmap->max = num - reserved_top; bitmap->mask = mask; + bitmap->reserved_top = reserved_top; spin_lock_init(&bitmap->lock); - bitmap->table = kzalloc(BITS_TO_LONGS(num) * sizeof (long), GFP_KERNEL); + bitmap->table = kzalloc(BITS_TO_LONGS(bitmap->max) * + sizeof (long), GFP_KERNEL); if (!bitmap->table) return -ENOMEM; - for (i = 0; i < reserved; ++i) + for (i = 0; i < reserved_bot; ++i) set_bit(i, bitmap->table); return 0; diff --git a/drivers/net/mlx4/cq.c b/drivers/net/mlx4/cq.c index 9bb50e3..b7ad282 100644 --- a/drivers/net/mlx4/cq.c +++ b/drivers/net/mlx4/cq.c @@ -300,7 +300,7 @@ int mlx4_init_cq_table(struct mlx4_dev *dev) INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC); err = mlx4_bitmap_init(&cq_table->bitmap, dev->caps.num_cqs, - dev->caps.num_cqs - 1, dev->caps.reserved_cqs); + dev->caps.num_cqs - 1, dev->caps.reserved_cqs, 0); if (err) return err; diff --git a/drivers/net/mlx4/en_cq.c b/drivers/net/mlx4/en_cq.c new file mode 100644 index 0000000..1368a80 --- /dev/null +++ b/drivers/net/mlx4/en_cq.c @@ -0,0 +1,146 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/mlx4/cq.h> +#include <linux/mlx4/qp.h> +#include <linux/mlx4/cmd.h> + +#include "mlx4_en.h" + +static void mlx4_en_cq_event(struct mlx4_cq *cq, enum mlx4_event event) +{ + return; +} + + +int mlx4_en_create_cq(struct mlx4_en_priv *priv, + struct mlx4_en_cq *cq, + int entries, int ring, enum cq_type mode) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + cq->size = entries; + if (mode == RX) + cq->buf_size = cq->size * sizeof(struct mlx4_cqe); + else + cq->buf_size = sizeof(struct mlx4_cqe); + + cq->ring = ring; + cq->is_tx = mode; + spin_lock_init(&cq->lock); + + err = mlx4_alloc_hwq_res(mdev->dev, &cq->wqres, + cq->buf_size, 2 * PAGE_SIZE); + if (err) + return err; + + err = mlx4_en_map_buffer(&cq->wqres.buf); + if (err) + mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size); + + return err; +} + +int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + cq->dev = mdev->pndev[priv->port]; + cq->mcq.set_ci_db = cq->wqres.db.db; + cq->mcq.arm_db = cq->wqres.db.db + 1; + *cq->mcq.set_ci_db = 0; + *cq->mcq.arm_db = 0; + cq->buf = (struct mlx4_cqe *) cq->wqres.buf.direct.buf; + memset(cq->buf, 0, cq->buf_size); + + err = mlx4_cq_alloc(mdev->dev, cq->size, &cq->wqres.mtt, &mdev->priv_uar, + cq->wqres.db.dma, &cq->mcq, cq->is_tx); + if (err) + return err; + + cq->mcq.comp = cq->is_tx ? mlx4_en_tx_irq : mlx4_en_rx_irq; + cq->mcq.event = mlx4_en_cq_event; + + if (cq->is_tx) { + init_timer(&cq->timer); + cq->timer.function = mlx4_en_poll_tx_cq; + cq->timer.data = (unsigned long) cq; + } else { + netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_rx_cq, 64); + napi_enable(&cq->napi); + } + + return 0; +} + +void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + mlx4_en_unmap_buffer(&cq->wqres.buf); + mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size); + cq->buf_size = 0; + cq->buf = NULL; +} + +void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + if (cq->is_tx) + del_timer(&cq->timer); + else + napi_disable(&cq->napi); + + mlx4_cq_free(mdev->dev, &cq->mcq); +} + +/* Set rx cq moderation parameters */ +int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq) +{ + return mlx4_cq_modify(priv->mdev->dev, &cq->mcq, + cq->moder_cnt, cq->moder_time); +} + +int mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq) +{ + cq->armed = 1; + mlx4_cq_arm(&cq->mcq, MLX4_CQ_DB_REQ_NOT, priv->mdev->uar_map, + &priv->mdev->uar_lock); + + return 0; +} + + diff --git a/drivers/net/mlx4/en_main.c b/drivers/net/mlx4/en_main.c new file mode 100644 index 0000000..1b0eebf --- /dev/null +++ b/drivers/net/mlx4/en_main.c @@ -0,0 +1,254 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/cpumask.h> +#include <linux/module.h> +#include <linux/delay.h> +#include <linux/netdevice.h> +#include <linux/cpumask.h> + +#include <linux/mlx4/driver.h> +#include <linux/mlx4/device.h> +#include <linux/mlx4/cmd.h> + +#include "mlx4_en.h" + +MODULE_AUTHOR("Liran Liss, Yevgeny Petrilin"); +MODULE_DESCRIPTION("Mellanox ConnectX HCA Ethernet driver"); +MODULE_LICENSE("Dual BSD/GPL"); +MODULE_VERSION(DRV_VERSION " ("DRV_RELDATE")"); + +static const char mlx4_en_version[] = + DRV_NAME ": Mellanox ConnectX HCA Ethernet driver v" + DRV_VERSION " (" DRV_RELDATE ")\n"; + +static void mlx4_en_event(struct mlx4_dev *dev, void *endev_ptr, + enum mlx4_dev_event event, int port) +{ + struct mlx4_en_dev *mdev = (struct mlx4_en_dev *) endev_ptr; + struct mlx4_en_priv *priv; + + if (!mdev->pndev[port]) + return; + + priv = netdev_priv(mdev->pndev[port]); + switch (event) { + case MLX4_DEV_EVENT_PORT_UP: + case MLX4_DEV_EVENT_PORT_DOWN: + /* To prevent races, we poll the link state in a separate + task rather than changing it here */ + priv->link_state = event; + queue_work(mdev->workqueue, &priv->linkstate_task); + break; + + case MLX4_DEV_EVENT_CATASTROPHIC_ERROR: + mlx4_err(mdev, "Internal error detected, restarting device\n"); + break; + + default: + mlx4_warn(mdev, "Unhandled event: %d\n", event); + } +} + +static void mlx4_en_remove(struct mlx4_dev *dev, void *endev_ptr) +{ + struct mlx4_en_dev *mdev = endev_ptr; + int i; + + mutex_lock(&mdev->state_lock); + mdev->device_up = false; + mutex_unlock(&mdev->state_lock); + + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) + if (mdev->pndev[i]) + mlx4_en_destroy_netdev(mdev->pndev[i]); + + flush_workqueue(mdev->workqueue); + destroy_workqueue(mdev->workqueue); + mlx4_mr_free(dev, &mdev->mr); + mlx4_uar_free(dev, &mdev->priv_uar); + mlx4_pd_free(dev, mdev->priv_pdn); + kfree(mdev); +} + +static void *mlx4_en_add(struct mlx4_dev *dev) +{ + static int mlx4_en_version_printed; + struct mlx4_en_dev *mdev; + int i; + int err; + + if (!mlx4_en_version_printed) { + printk(KERN_INFO "%s", mlx4_en_version); + mlx4_en_version_printed++; + } + + mdev = kzalloc(sizeof *mdev, GFP_KERNEL); + if (!mdev) { + dev_err(&dev->pdev->dev, "Device struct alloc failed, " + "aborting.\n"); + err = -ENOMEM; + goto err_free_res; + } + + if (mlx4_pd_alloc(dev, &mdev->priv_pdn)) + goto err_free_dev; + + if (mlx4_uar_alloc(dev, &mdev->priv_uar)) + goto err_pd; + + mdev->uar_map = ioremap(mdev->priv_uar.pfn << PAGE_SHIFT, PAGE_SIZE); + if (!mdev->uar_map) + goto err_uar; + spin_lock_init(&mdev->uar_lock); + + mdev->dev = dev; + mdev->dma_device = &(dev->pdev->dev); + mdev->pdev = dev->pdev; + mdev->device_up = false; + + mdev->LSO_support = !!(dev->caps.flags & (1 << 15)); + if (!mdev->LSO_support) + mlx4_warn(mdev, "LSO not supported, please upgrade to later " + "FW version to enable LSO\n"); + + if (mlx4_mr_alloc(mdev->dev, mdev->priv_pdn, 0, ~0ull, + MLX4_PERM_LOCAL_WRITE | MLX4_PERM_LOCAL_READ, + 0, 0, &mdev->mr)) { + mlx4_err(mdev, "Failed allocating memory region\n"); + goto err_uar; + } + if (mlx4_mr_enable(mdev->dev, &mdev->mr)) { + mlx4_err(mdev, "Failed enabling memory region\n"); + goto err_mr; + } + + /* Build device profile according to supplied module parameters */ + err = mlx4_en_get_profile(mdev); + if (err) { + mlx4_err(mdev, "Bad module parameters, aborting.\n"); + goto err_mr; + } + + /* Configure wich ports to start according to module parameters */ + mdev->port_cnt = 0; + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) + mdev->port_cnt++; + + /* If we did not receive an explicit number of Rx rings, default to + * the number of completion vectors populated by the mlx4_core */ + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) { + mlx4_info(mdev, "Using %d tx rings for port:%d\n", + mdev->profile.prof[i].tx_ring_num, i); + if (!mdev->profile.prof[i].rx_ring_num) { + mdev->profile.prof[i].rx_ring_num = 1; + mlx4_info(mdev, "Defaulting to %d rx rings for port:%d\n", + 1, i); + } else + mlx4_info(mdev, "Using %d rx rings for port:%d\n", + mdev->profile.prof[i].rx_ring_num, i); + } + + /* Create our own workqueue for reset/multicast tasks + * Note: we cannot use the shared workqueue because of deadlocks caused + * by the rtnl lock */ + mdev->workqueue = create_singlethread_workqueue("mlx4_en"); + if (!mdev->workqueue) { + err = -ENOMEM; + goto err_close_nic; + } + + /* At this stage all non-port specific tasks are complete: + * mark the card state as up */ + mutex_init(&mdev->state_lock); + mdev->device_up = true; + + /* Setup ports */ + + /* Create a netdev for each port */ + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) { + mlx4_info(mdev, "Activating port:%d\n", i); + if (mlx4_en_init_netdev(mdev, i, &mdev->profile.prof[i])) { + mdev->pndev[i] = NULL; + goto err_free_netdev; + } + } + return mdev; + + +err_free_netdev: + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) { + if (mdev->pndev[i]) + mlx4_en_destroy_netdev(mdev->pndev[i]); + } + + mutex_lock(&mdev->state_lock); + mdev->device_up = false; + mutex_unlock(&mdev->state_lock); + flush_workqueue(mdev->workqueue); + + /* Stop event queue before we drop down to release shared SW state */ + +err_close_nic: + destroy_workqueue(mdev->workqueue); +err_mr: + mlx4_mr_free(dev, &mdev->mr); +err_uar: + mlx4_uar_free(dev, &mdev->priv_uar); +err_pd: + mlx4_pd_free(dev, mdev->priv_pdn); +err_free_dev: + kfree(mdev); +err_free_res: + return NULL; +} + +static struct mlx4_interface mlx4_en_interface = { + .add = mlx4_en_add, + .remove = mlx4_en_remove, + .event = mlx4_en_event, +}; + +static int __init mlx4_en_init(void) +{ + return mlx4_register_interface(&mlx4_en_interface); +} + +static void __exit mlx4_en_cleanup(void) +{ + mlx4_unregister_interface(&mlx4_en_interface); +} + +module_init(mlx4_en_init); +module_exit(mlx4_en_cleanup); + diff --git a/drivers/net/mlx4/en_netdev.c b/drivers/net/mlx4/en_netdev.c new file mode 100644 index 0000000..a339afb --- /dev/null +++ b/drivers/net/mlx4/en_netdev.c @@ -0,0 +1,1088 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/etherdevice.h> +#include <linux/tcp.h> +#include <linux/if_vlan.h> +#include <linux/delay.h> + +#include <linux/mlx4/driver.h> +#include <linux/mlx4/device.h> +#include <linux/mlx4/cmd.h> +#include <linux/mlx4/cq.h> + +#include "mlx4_en.h" +#include "en_port.h" + + +static void mlx4_en_vlan_rx_register(struct net_device *dev, struct vlan_group *grp) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + mlx4_dbg(HW, priv, "Registering VLAN group:%p\n", grp); + priv->vlgrp = grp; + + mutex_lock(&mdev->state_lock); + if (mdev->device_up && priv->port_up) { + err = mlx4_SET_VLAN_FLTR(mdev->dev, priv->port, grp); + if (err) + mlx4_err(mdev, "Failed configuring VLAN filter\n"); + } + mutex_unlock(&mdev->state_lock); +} + +static void mlx4_en_vlan_rx_add_vid(struct net_device *dev, unsigned short vid) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + if (!priv->vlgrp) + return; + + mlx4_dbg(HW, priv, "adding VLAN:%d (vlgrp entry:%p)\n", + vid, vlan_group_get_device(priv->vlgrp, vid)); + + /* Add VID to port VLAN filter */ + mutex_lock(&mdev->state_lock); + if (mdev->device_up && priv->port_up) { + err = mlx4_SET_VLAN_FLTR(mdev->dev, priv->port, priv->vlgrp); + if (err) + mlx4_err(mdev, "Failed configuring VLAN filter\n"); + } + mutex_unlock(&mdev->state_lock); +} + +static void mlx4_en_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + if (!priv->vlgrp) + return; + + mlx4_dbg(HW, priv, "Killing VID:%d (vlgrp:%p vlgrp " + "entry:%p)\n", vid, priv->vlgrp, + vlan_group_get_device(priv->vlgrp, vid)); + vlan_group_set_device(priv->vlgrp, vid, NULL); + + /* Remove VID from port VLAN filter */ + mutex_lock(&mdev->state_lock); + if (mdev->device_up && priv->port_up) { + err = mlx4_SET_VLAN_FLTR(mdev->dev, priv->port, priv->vlgrp); + if (err) + mlx4_err(mdev, "Failed configuring VLAN filter\n"); + } + mutex_unlock(&mdev->state_lock); +} + +static u64 mlx4_en_mac_to_u64(u8 *addr) +{ + u64 mac = 0; + int i; + + for (i = 0; i < ETH_ALEN; i++) { + mac <<= 8; + mac |= addr[i]; + } + return mac; +} + +static int mlx4_en_set_mac(struct net_device *dev, void *addr) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + struct sockaddr *saddr = addr; + + if (!is_valid_ether_addr(saddr->sa_data)) + return -EADDRNOTAVAIL; + + memcpy(dev->dev_addr, saddr->sa_data, ETH_ALEN); + priv->mac = mlx4_en_mac_to_u64(dev->dev_addr); + queue_work(mdev->workqueue, &priv->mac_task); + return 0; +} + +static void mlx4_en_do_set_mac(struct work_struct *work) +{ + struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv, + mac_task); + struct mlx4_en_dev *mdev = priv->mdev; + int err = 0; + + mutex_lock(&mdev->state_lock); + if (priv->port_up) { + /* Remove old MAC and insert the new one */ + mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index); + err = mlx4_register_mac(mdev->dev, priv->port, + priv->mac, &priv->mac_index); + if (err) + mlx4_err(mdev, "Failed changing HW MAC address\n"); + } else + mlx4_dbg(HW, priv, "Port is down, exiting...\n"); + + mutex_unlock(&mdev->state_lock); +} + +static void mlx4_en_clear_list(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct dev_mc_list *plist = priv->mc_list; + struct dev_mc_list *next; + + while (plist) { + next = plist->next; + kfree(plist); + plist = next; + } + priv->mc_list = NULL; +} + +static void mlx4_en_cache_mclist(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + struct dev_mc_list *mclist; + struct dev_mc_list *tmp; + struct dev_mc_list *plist = NULL; + + for (mclist = dev->mc_list; mclist; mclist = mclist->next) { + tmp = kmalloc(sizeof(struct dev_mc_list), GFP_ATOMIC); + if (!tmp) { + mlx4_err(mdev, "failed to allocate multicast list\n"); + mlx4_en_clear_list(dev); + return; + } + memcpy(tmp, mclist, sizeof(struct dev_mc_list)); + tmp->next = NULL; + if (plist) + plist->next = tmp; + else + priv->mc_list = tmp; + plist = tmp; + } +} + + +static void mlx4_en_set_multicast(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + + if (!priv->port_up) + return; + + queue_work(priv->mdev->workqueue, &priv->mcast_task); +} + +static void mlx4_en_do_set_multicast(struct work_struct *work) +{ + struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv, + mcast_task); + struct mlx4_en_dev *mdev = priv->mdev; + struct net_device *dev = priv->dev; + struct dev_mc_list *mclist; + u64 mcast_addr = 0; + int err; + + mutex_lock(&mdev->state_lock); + if (!mdev->device_up) { + mlx4_dbg(HW, priv, "Card is not up, ignoring " + "multicast change.\n"); + goto out; + } + if (!priv->port_up) { + mlx4_dbg(HW, priv, "Port is down, ignoring " + "multicast change.\n"); + goto out; + } + + /* + * Promsicuous mode: disable all filters + */ + + if (dev->flags & IFF_PROMISC) { + if (!(priv->flags & MLX4_EN_FLAG_PROMISC)) { + if (netif_msg_rx_status(priv)) + mlx4_warn(mdev, "Port:%d entering promiscuous mode\n", + priv->port); + priv->flags |= MLX4_EN_FLAG_PROMISC; + + /* Enable promiscouos mode */ + err = mlx4_SET_PORT_qpn_calc(mdev->dev, priv->port, + priv->base_qpn, 1); + if (err) + mlx4_err(mdev, "Failed enabling " + "promiscous mode\n"); + + /* Disable port multicast filter (unconditionally) */ + err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0, + 0, MLX4_MCAST_DISABLE); + if (err) + mlx4_err(mdev, "Failed disabling " + "multicast filter\n"); + + /* Disable port VLAN filter */ + err = mlx4_SET_VLAN_FLTR(mdev->dev, priv->port, NULL); + if (err) + mlx4_err(mdev, "Failed disabling " + "VLAN filter\n"); + } + goto out; + } + + /* + * Not in promiscous mode + */ + + if (priv->flags & MLX4_EN_FLAG_PROMISC) { + if (netif_msg_rx_status(priv)) + mlx4_warn(mdev, "Port:%d leaving promiscuous mode\n", + priv->port); + priv->flags &= ~MLX4_EN_FLAG_PROMISC; + + /* Disable promiscouos mode */ + err = mlx4_SET_PORT_qpn_calc(mdev->dev, priv->port, + priv->base_qpn, 0); + if (err) + mlx4_err(mdev, "Failed disabling promiscous mode\n"); + + /* Enable port VLAN filter */ + err = mlx4_SET_VLAN_FLTR(mdev->dev, priv->port, priv->vlgrp); + if (err) + mlx4_err(mdev, "Failed enabling VLAN filter\n"); + } + + /* Enable/disable the multicast filter according to IFF_ALLMULTI */ + if (dev->flags & IFF_ALLMULTI) { + err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0, + 0, MLX4_MCAST_DISABLE); + if (err) + mlx4_err(mdev, "Failed disabling multicast filter\n"); + } else { + err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0, + 0, MLX4_MCAST_DISABLE); + if (err) + mlx4_err(mdev, "Failed disabling multicast filter\n"); + + /* Flush mcast filter and init it with broadcast address */ + mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, ETH_BCAST, + 1, MLX4_MCAST_CONFIG); + + /* Update multicast list - we cache all addresses so they won't + * change while HW is updated holding the command semaphor */ + netif_tx_lock_bh(dev); + mlx4_en_cache_mclist(dev); + netif_tx_unlock_bh(dev); + for (mclist = priv->mc_list; mclist; mclist = mclist->next) { + mcast_addr = mlx4_en_mac_to_u64(mclist->dmi_addr); + mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, + mcast_addr, 0, MLX4_MCAST_CONFIG); + } + err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0, + 0, MLX4_MCAST_ENABLE); + if (err) + mlx4_err(mdev, "Failed enabling multicast filter\n"); + + mlx4_en_clear_list(dev); + } +out: + mutex_unlock(&mdev->state_lock); +} + +#ifdef CONFIG_NET_POLL_CONTROLLER +static void mlx4_en_netpoll(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_cq *cq; + unsigned long flags; + int i; + + for (i = 0; i < priv->rx_ring_num; i++) { + cq = &priv->rx_cq[i]; + spin_lock_irqsave(&cq->lock, flags); + napi_synchronize(&cq->napi); + mlx4_en_process_rx_cq(dev, cq, 0); + spin_unlock_irqrestore(&cq->lock, flags); + } +} +#endif + +static void mlx4_en_tx_timeout(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + if (netif_msg_timer(priv)) + mlx4_warn(mdev, "Tx timeout called on port:%d\n", priv->port); + + if (netif_carrier_ok(dev)) { + priv->port_stats.tx_timeout++; + mlx4_dbg(DRV, priv, "Scheduling watchdog\n"); + queue_work(mdev->workqueue, &priv->watchdog_task); + } +} + + +static struct net_device_stats *mlx4_en_get_stats(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + + spin_lock_bh(&priv->stats_lock); + memcpy(&priv->ret_stats, &priv->stats, sizeof(priv->stats)); + spin_unlock_bh(&priv->stats_lock); + + return &priv->ret_stats; +} + +static void mlx4_en_set_default_moderation(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_cq *cq; + int i; + + /* If we haven't received a specific coalescing setting + * (module param), we set the moderation paramters as follows: + * - moder_cnt is set to the number of mtu sized packets to + * satisfy our coelsing target. + * - moder_time is set to a fixed value. + */ + priv->rx_frames = (mdev->profile.rx_moder_cnt == + MLX4_EN_AUTO_CONF) ? + MLX4_EN_RX_COAL_TARGET / + priv->dev->mtu + 1 : + mdev->profile.rx_moder_cnt; + priv->rx_usecs = (mdev->profile.rx_moder_time == + MLX4_EN_AUTO_CONF) ? + MLX4_EN_RX_COAL_TIME : + mdev->profile.rx_moder_time; + mlx4_dbg(INTR, priv, "Default coalesing params for mtu:%d - " + "rx_frames:%d rx_usecs:%d\n", + priv->dev->mtu, priv->rx_frames, priv->rx_usecs); + + /* Setup cq moderation params */ + for (i = 0; i < priv->rx_ring_num; i++) { + cq = &priv->rx_cq[i]; + cq->moder_cnt = priv->rx_frames; + cq->moder_time = priv->rx_usecs; + } + + for (i = 0; i < priv->tx_ring_num; i++) { + cq = &priv->tx_cq[i]; + cq->moder_cnt = MLX4_EN_TX_COAL_PKTS; + cq->moder_time = MLX4_EN_TX_COAL_TIME; + } + + /* Reset auto-moderation params */ + priv->pkt_rate_low = MLX4_EN_RX_RATE_LOW; + priv->rx_usecs_low = MLX4_EN_RX_COAL_TIME_LOW; + priv->pkt_rate_high = MLX4_EN_RX_RATE_HIGH; + priv->rx_usecs_high = MLX4_EN_RX_COAL_TIME_HIGH; + priv->sample_interval = MLX4_EN_SAMPLE_INTERVAL; + priv->adaptive_rx_coal = mdev->profile.auto_moder; + priv->last_moder_time = MLX4_EN_AUTO_CONF; + priv->last_moder_jiffies = 0; + priv->last_moder_packets = 0; + priv->last_moder_tx_packets = 0; + priv->last_moder_bytes = 0; +} + +static void mlx4_en_auto_moderation(struct mlx4_en_priv *priv) +{ + unsigned long period = (unsigned long) (jiffies - priv->last_moder_jiffies); + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_cq *cq; + unsigned long packets; + unsigned long rate; + unsigned long avg_pkt_size; + unsigned long rx_packets; + unsigned long rx_bytes; + unsigned long tx_packets; + unsigned long tx_pkt_diff; + unsigned long rx_pkt_diff; + int moder_time; + int i, err; + + if (!priv->adaptive_rx_coal || period < priv->sample_interval * HZ) + return; + + spin_lock_bh(&priv->stats_lock); + rx_packets = priv->stats.rx_packets; + rx_bytes = priv->stats.rx_bytes; + tx_packets = priv->stats.tx_packets; + spin_unlock_bh(&priv->stats_lock); + + if (!priv->last_moder_jiffies || !period) + goto out; + + tx_pkt_diff = ((unsigned long) (tx_packets - + priv->last_moder_tx_packets)); + rx_pkt_diff = ((unsigned long) (rx_packets - + priv->last_moder_packets)); + packets = max(tx_pkt_diff, rx_pkt_diff); + rate = packets * HZ / period; + avg_pkt_size = packets ? ((unsigned long) (rx_bytes - + priv->last_moder_bytes)) / packets : 0; + + /* Apply auto-moderation only when packet rate exceeds a rate that + * it matters */ + if (rate > MLX4_EN_RX_RATE_THRESH) { + /* If tx and rx packet rates are not balanced, assume that + * traffic is mainly BW bound and apply maximum moderation. + * Otherwise, moderate according to packet rate */ + if (2 * tx_pkt_diff > 3 * rx_pkt_diff || + 2 * rx_pkt_diff > 3 * tx_pkt_diff) { + moder_time = priv->rx_usecs_high; + } else { + if (rate < priv->pkt_rate_low) + moder_time = priv->rx_usecs_low; + else if (rate > priv->pkt_rate_high) + moder_time = priv->rx_usecs_high; + else + moder_time = (rate - priv->pkt_rate_low) * + (priv->rx_usecs_high - priv->rx_usecs_low) / + (priv->pkt_rate_high - priv->pkt_rate_low) + + priv->rx_usecs_low; + } + } else { + /* When packet rate is low, use default moderation rather than + * 0 to prevent interrupt storms if traffic suddenly increases */ + moder_time = priv->rx_usecs; + } + + mlx4_dbg(INTR, priv, "tx rate:%lu rx_rate:%lu\n", + tx_pkt_diff * HZ / period, rx_pkt_diff * HZ / period); + + mlx4_dbg(INTR, priv, "Rx moder_time changed from:%d to %d period:%lu " + "[jiff] packets:%lu avg_pkt_size:%lu rate:%lu [p/s])\n", + priv->last_moder_time, moder_time, period, packets, + avg_pkt_size, rate); + + if (moder_time != priv->last_moder_time) { + priv->last_moder_time = moder_time; + for (i = 0; i < priv->rx_ring_num; i++) { + cq = &priv->rx_cq[i]; + cq->moder_time = moder_time; + err = mlx4_en_set_cq_moder(priv, cq); + if (err) { + mlx4_err(mdev, "Failed modifying moderation for cq:%d " + "on port:%d\n", i, priv->port); + break; + } + } + } + +out: + priv->last_moder_packets = rx_packets; + priv->last_moder_tx_packets = tx_packets; + priv->last_moder_bytes = rx_bytes; + priv->last_moder_jiffies = jiffies; +} + +static void mlx4_en_do_get_stats(struct work_struct *work) +{ + struct delayed_work *delay = container_of(work, struct delayed_work, work); + struct mlx4_en_priv *priv = container_of(delay, struct mlx4_en_priv, + stats_task); + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + err = mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 0); + if (err) + mlx4_dbg(HW, priv, "Could not update stats for " + "port:%d\n", priv->port); + + mutex_lock(&mdev->state_lock); + if (mdev->device_up) { + if (priv->port_up) + mlx4_en_auto_moderation(priv); + + queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY); + } + mutex_unlock(&mdev->state_lock); +} + +static void mlx4_en_linkstate(struct work_struct *work) +{ + struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv, + linkstate_task); + struct mlx4_en_dev *mdev = priv->mdev; + int linkstate = priv->link_state; + + mutex_lock(&mdev->state_lock); + /* If observable port state changed set carrier state and + * report to system log */ + if (priv->last_link_state != linkstate) { + if (linkstate == MLX4_DEV_EVENT_PORT_DOWN) { + if (netif_msg_link(priv)) + mlx4_info(mdev, "Port %d - link down\n", priv->port); + netif_carrier_off(priv->dev); + } else { + if (netif_msg_link(priv)) + mlx4_info(mdev, "Port %d - link up\n", priv->port); + netif_carrier_on(priv->dev); + } + } + priv->last_link_state = linkstate; + mutex_unlock(&mdev->state_lock); +} + + +static int mlx4_en_start_port(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_cq *cq; + struct mlx4_en_tx_ring *tx_ring; + struct mlx4_en_rx_ring *rx_ring; + int rx_index = 0; + int tx_index = 0; + u16 stride; + int err = 0; + int i; + int j; + + if (priv->port_up) { + mlx4_dbg(DRV, priv, "start port called while port already up\n"); + return 0; + } + + /* Calculate Rx buf size */ + dev->mtu = min(dev->mtu, priv->max_mtu); + mlx4_en_calc_rx_buf(dev); + mlx4_dbg(DRV, priv, "Rx buf size:%d\n", priv->rx_skb_size); + stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) + + DS_SIZE * priv->num_frags); + /* Configure rx cq's and rings */ + for (i = 0; i < priv->rx_ring_num; i++) { + cq = &priv->rx_cq[i]; + rx_ring = &priv->rx_ring[i]; + + err = mlx4_en_activate_cq(priv, cq); + if (err) { + mlx4_err(mdev, "Failed activating Rx CQ\n"); + goto rx_err; + } + for (j = 0; j < cq->size; j++) + cq->buf[j].owner_sr_opcode = MLX4_CQE_OWNER_MASK; + err = mlx4_en_set_cq_moder(priv, cq); + if (err) { + mlx4_err(mdev, "Failed setting cq moderation parameters"); + mlx4_en_deactivate_cq(priv, cq); + goto cq_err; + } + mlx4_en_arm_cq(priv, cq); + + ++rx_index; + } + + err = mlx4_en_activate_rx_rings(priv); + if (err) { + mlx4_err(mdev, "Failed to activate RX rings\n"); + goto cq_err; + } + + err = mlx4_en_config_rss_steer(priv); + if (err) { + mlx4_err(mdev, "Failed configuring rss steering\n"); + goto rx_err; + } + + /* Configure tx cq's and rings */ + for (i = 0; i < priv->tx_ring_num; i++) { + /* Configure cq */ + cq = &priv->tx_cq[i]; + err = mlx4_en_activate_cq(priv, cq); + if (err) { + mlx4_err(mdev, "Failed allocating Tx CQ\n"); + goto tx_err; + } + err = mlx4_en_set_cq_moder(priv, cq); + if (err) { + mlx4_err(mdev, "Failed setting cq moderation parameters"); + mlx4_en_deactivate_cq(priv, cq); + goto tx_err; + } + mlx4_dbg(DRV, priv, "Resetting index of collapsed CQ:%d to -1\n", i); + cq->buf->wqe_index = cpu_to_be16(0xffff); + + /* Configure ring */ + tx_ring = &priv->tx_ring[i]; + err = mlx4_en_activate_tx_ring(priv, tx_ring, cq->mcq.cqn, + priv->rx_ring[0].srq.srqn); + if (err) { + mlx4_err(mdev, "Failed allocating Tx ring\n"); + mlx4_en_deactivate_cq(priv, cq); + goto tx_err; + } + /* Set initial ownership of all Tx TXBBs to SW (1) */ + for (j = 0; j < tx_ring->buf_size; j += STAMP_STRIDE) + *((u32 *) (tx_ring->buf + j)) = 0xffffffff; + ++tx_index; + } + + /* Configure port */ + err = mlx4_SET_PORT_general(mdev->dev, priv->port, + priv->rx_skb_size + ETH_FCS_LEN, + mdev->profile.tx_pause, + mdev->profile.tx_ppp, + mdev->profile.rx_pause, + mdev->profile.rx_ppp); + if (err) { + mlx4_err(mdev, "Failed setting port general configurations" + " for port %d, with error %d\n", priv->port, err); + goto tx_err; + } + /* Set default qp number */ + err = mlx4_SET_PORT_qpn_calc(mdev->dev, priv->port, priv->base_qpn, 0); + if (err) { + mlx4_err(mdev, "Failed setting default qp numbers\n"); + goto tx_err; + } + /* Set port mac number */ + mlx4_dbg(DRV, priv, "Setting mac for port %d\n", priv->port); + err = mlx4_register_mac(mdev->dev, priv->port, + priv->mac, &priv->mac_index); + if (err) { + mlx4_err(mdev, "Failed setting port mac\n"); + goto tx_err; + } + + /* Init port */ + mlx4_dbg(HW, priv, "Initializing port\n"); + err = mlx4_INIT_PORT(mdev->dev, priv->port); + if (err) { + mlx4_err(mdev, "Failed Initializing port\n"); + goto mac_err; + } + + /* Schedule multicast task to populate multicast list */ + queue_work(mdev->workqueue, &priv->mcast_task); + + priv->port_up = true; + netif_start_queue(dev); + return 0; + +mac_err: + mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index); +tx_err: + while (tx_index--) { + mlx4_en_deactivate_tx_ring(priv, &priv->tx_ring[tx_index]); + mlx4_en_deactivate_cq(priv, &priv->tx_cq[tx_index]); + } + + mlx4_en_release_rss_steer(priv); +rx_err: + for (i = 0; i < priv->rx_ring_num; i++) + mlx4_en_deactivate_rx_ring(priv, &priv->rx_ring[rx_index]); +cq_err: + while (rx_index--) + mlx4_en_deactivate_cq(priv, &priv->rx_cq[rx_index]); + + return err; /* need to close devices */ +} + + +static void mlx4_en_stop_port(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int i; + + if (!priv->port_up) { + mlx4_dbg(DRV, priv, "stop port (%d) called while port already down\n", + priv->port); + return; + } + netif_stop_queue(dev); + + /* Synchronize with tx routine */ + netif_tx_lock_bh(dev); + priv->port_up = false; + netif_tx_unlock_bh(dev); + + /* close port*/ + mlx4_CLOSE_PORT(mdev->dev, priv->port); + + /* Unregister Mac address for the port */ + mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index); + + /* Free TX Rings */ + for (i = 0; i < priv->tx_ring_num; i++) { + mlx4_en_deactivate_tx_ring(priv, &priv->tx_ring[i]); + mlx4_en_deactivate_cq(priv, &priv->tx_cq[i]); + } + msleep(10); + + for (i = 0; i < priv->tx_ring_num; i++) + mlx4_en_free_tx_buf(dev, &priv->tx_ring[i]); + + /* Free RSS qps */ + mlx4_en_release_rss_steer(priv); + + /* Free RX Rings */ + for (i = 0; i < priv->rx_ring_num; i++) { + mlx4_en_deactivate_rx_ring(priv, &priv->rx_ring[i]); + while (test_bit(NAPI_STATE_SCHED, &priv->rx_cq[i].napi.state)) + msleep(1); + mlx4_en_deactivate_cq(priv, &priv->rx_cq[i]); + } +} + +static void mlx4_en_restart(struct work_struct *work) +{ + struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv, + watchdog_task); + struct mlx4_en_dev *mdev = priv->mdev; + struct net_device *dev = priv->dev; + + mlx4_dbg(DRV, priv, "Watchdog task called for port %d\n", priv->port); + mlx4_en_stop_port(dev); + if (mlx4_en_start_port(dev)) + mlx4_err(mdev, "Failed restarting port %d\n", priv->port); +} + + +static int mlx4_en_open(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int i; + int err = 0; + + mutex_lock(&mdev->state_lock); + + if (!mdev->device_up) { + mlx4_err(mdev, "Cannot open - device down/disabled\n"); + err = -EBUSY; + goto out; + } + + /* Reset HW statistics and performance counters */ + if (mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 1)) + mlx4_dbg(HW, priv, "Failed dumping statistics\n"); + + memset(&priv->stats, 0, sizeof(priv->stats)); + memset(&priv->pstats, 0, sizeof(priv->pstats)); + + for (i = 0; i < priv->tx_ring_num; i++) { + priv->tx_ring[i].bytes = 0; + priv->tx_ring[i].packets = 0; + } + for (i = 0; i < priv->rx_ring_num; i++) { + priv->rx_ring[i].bytes = 0; + priv->rx_ring[i].packets = 0; + } + + mlx4_en_set_default_moderation(priv); + err = mlx4_en_start_port(dev); + if (err) + mlx4_err(mdev, "Failed starting port:%d\n", priv->port); + +out: + mutex_unlock(&mdev->state_lock); + return err; +} + + +static int mlx4_en_close(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + if (netif_msg_ifdown(priv)) + mlx4_info(mdev, "Close called for port:%d\n", priv->port); + + mutex_lock(&mdev->state_lock); + + mlx4_en_stop_port(dev); + netif_carrier_off(dev); + + mutex_unlock(&mdev->state_lock); + return 0; +} + +static void mlx4_en_free_resources(struct mlx4_en_priv *priv) +{ + int i; + + for (i = 0; i < priv->tx_ring_num; i++) { + if (priv->tx_ring[i].tx_info) + mlx4_en_destroy_tx_ring(priv, &priv->tx_ring[i]); + if (priv->tx_cq[i].buf) + mlx4_en_destroy_cq(priv, &priv->tx_cq[i]); + } + + for (i = 0; i < priv->rx_ring_num; i++) { + if (priv->rx_ring[i].rx_info) + mlx4_en_destroy_rx_ring(priv, &priv->rx_ring[i]); + if (priv->rx_cq[i].buf) + mlx4_en_destroy_cq(priv, &priv->rx_cq[i]); + } +} + +static int mlx4_en_alloc_resources(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_port_profile *prof = priv->prof; + int i; + + /* Create tx Rings */ + for (i = 0; i < priv->tx_ring_num; i++) { + if (mlx4_en_create_cq(priv, &priv->tx_cq[i], + prof->tx_ring_size, i, TX)) + goto err; + + if (mlx4_en_create_tx_ring(priv, &priv->tx_ring[i], + prof->tx_ring_size, TXBB_SIZE)) + goto err; + } + + /* Create rx Rings */ + for (i = 0; i < priv->rx_ring_num; i++) { + if (mlx4_en_create_cq(priv, &priv->rx_cq[i], + prof->rx_ring_size, i, RX)) + goto err; + + if (mlx4_en_create_rx_ring(priv, &priv->rx_ring[i], + prof->rx_ring_size, priv->stride)) + goto err; + } + + return 0; + +err: + mlx4_err(mdev, "Failed to allocate NIC resources\n"); + return -ENOMEM; +} + + +void mlx4_en_destroy_netdev(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + mlx4_dbg(DRV, priv, "Destroying netdev on port:%d\n", priv->port); + + /* Unregister device - this will close the port if it was up */ + if (priv->registered) + unregister_netdev(dev); + + if (priv->allocated) + mlx4_free_hwq_res(mdev->dev, &priv->res, MLX4_EN_PAGE_SIZE); + + cancel_delayed_work(&priv->stats_task); + cancel_delayed_work(&priv->refill_task); + /* flush any pending task for this netdev */ + flush_workqueue(mdev->workqueue); + + /* Detach the netdev so tasks would not attempt to access it */ + mutex_lock(&mdev->state_lock); + mdev->pndev[priv->port] = NULL; + mutex_unlock(&mdev->state_lock); + + mlx4_en_free_resources(priv); + free_netdev(dev); +} + +static int mlx4_en_change_mtu(struct net_device *dev, int new_mtu) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int err = 0; + + mlx4_dbg(DRV, priv, "Change MTU called - current:%d new:%d\n", + dev->mtu, new_mtu); + + if ((new_mtu < MLX4_EN_MIN_MTU) || (new_mtu > priv->max_mtu)) { + mlx4_err(mdev, "Bad MTU size:%d.\n", new_mtu); + return -EPERM; + } + dev->mtu = new_mtu; + + if (netif_running(dev)) { + mutex_lock(&mdev->state_lock); + if (!mdev->device_up) { + /* NIC is probably restarting - let watchdog task reset + * the port */ + mlx4_dbg(DRV, priv, "Change MTU called with card down!?\n"); + } else { + mlx4_en_stop_port(dev); + mlx4_en_set_default_moderation(priv); + err = mlx4_en_start_port(dev); + if (err) { + mlx4_err(mdev, "Failed restarting port:%d\n", + priv->port); + queue_work(mdev->workqueue, &priv->watchdog_task); + } + } + mutex_unlock(&mdev->state_lock); + } + return 0; +} + +int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port, + struct mlx4_en_port_profile *prof) +{ + struct net_device *dev; + struct mlx4_en_priv *priv; + int i; + int err; + + dev = alloc_etherdev(sizeof(struct mlx4_en_priv)); + if (dev == NULL) { + mlx4_err(mdev, "Net device allocation failed\n"); + return -ENOMEM; + } + + SET_NETDEV_DEV(dev, &mdev->dev->pdev->dev); + + /* + * Initialize driver private data + */ + + priv = netdev_priv(dev); + memset(priv, 0, sizeof(struct mlx4_en_priv)); + priv->dev = dev; + priv->mdev = mdev; + priv->prof = prof; + priv->port = port; + priv->port_up = false; + priv->rx_csum = 1; + priv->flags = prof->flags; + priv->tx_ring_num = prof->tx_ring_num; + priv->rx_ring_num = prof->rx_ring_num; + priv->mc_list = NULL; + priv->mac_index = -1; + priv->msg_enable = MLX4_EN_MSG_LEVEL; + spin_lock_init(&priv->stats_lock); + INIT_WORK(&priv->mcast_task, mlx4_en_do_set_multicast); + INIT_WORK(&priv->mac_task, mlx4_en_do_set_mac); + INIT_DELAYED_WORK(&priv->refill_task, mlx4_en_rx_refill); + INIT_WORK(&priv->watchdog_task, mlx4_en_restart); + INIT_WORK(&priv->linkstate_task, mlx4_en_linkstate); + INIT_DELAYED_WORK(&priv->stats_task, mlx4_en_do_get_stats); + + /* Query for default mac and max mtu */ + priv->max_mtu = mdev->dev->caps.eth_mtu_cap[priv->port]; + priv->mac = mdev->dev->caps.def_mac[priv->port]; + if (ILLEGAL_MAC(priv->mac)) { + mlx4_err(mdev, "Port: %d, invalid mac burned: 0x%llx, quiting\n", + priv->port, priv->mac); + err = -EINVAL; + goto out; + } + + priv->stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) + + DS_SIZE * MLX4_EN_MAX_RX_FRAGS); + err = mlx4_en_alloc_resources(priv); + if (err) + goto out; + + /* Populate Rx default RSS mappings */ + mlx4_en_set_default_rss_map(priv, &priv->rss_map, priv->rx_ring_num * + RSS_FACTOR, priv->rx_ring_num); + /* Allocate page for receive rings */ + err = mlx4_alloc_hwq_res(mdev->dev, &priv->res, + MLX4_EN_PAGE_SIZE, MLX4_EN_PAGE_SIZE); + if (err) { + mlx4_err(mdev, "Failed to allocate page for rx qps\n"); + goto out; + } + priv->allocated = 1; + + /* Populate Tx priority mappings */ + mlx4_en_set_prio_map(priv, priv->tx_prio_map, prof->tx_ring_num); + + /* + * Initialize netdev entry points + */ + + dev->open = &mlx4_en_open; + dev->stop = &mlx4_en_close; + dev->hard_start_xmit = &mlx4_en_xmit; + dev->get_stats = &mlx4_en_get_stats; + dev->set_multicast_list = &mlx4_en_set_multicast; + dev->set_mac_address = &mlx4_en_set_mac; + dev->change_mtu = &mlx4_en_change_mtu; + dev->tx_timeout = &mlx4_en_tx_timeout; + dev->watchdog_timeo = MLX4_EN_WATCHDOG_TIMEOUT; + dev->vlan_rx_register = mlx4_en_vlan_rx_register; + dev->vlan_rx_add_vid = mlx4_en_vlan_rx_add_vid; + dev->vlan_rx_kill_vid = mlx4_en_vlan_rx_kill_vid; +#ifdef CONFIG_NET_POLL_CONTROLLER + dev->poll_controller = mlx4_en_netpoll; +#endif + SET_ETHTOOL_OPS(dev, &mlx4_en_ethtool_ops); + + /* Set defualt MAC */ + dev->addr_len = ETH_ALEN; + for (i = 0; i < ETH_ALEN; i++) + dev->dev_addr[ETH_ALEN - 1 - i] = + (u8) (priv->mac >> (8 * i)); + + /* + * Set driver features + */ + dev->features |= NETIF_F_SG; + dev->features |= NETIF_F_HW_CSUM; + dev->features |= NETIF_F_HIGHDMA; + dev->features |= NETIF_F_HW_VLAN_TX | + NETIF_F_HW_VLAN_RX | + NETIF_F_HW_VLAN_FILTER; + if (mdev->profile.num_lro) + dev->features |= NETIF_F_LRO; + if (mdev->LSO_support) { + dev->features |= NETIF_F_TSO; + dev->features |= NETIF_F_TSO6; + } + + mdev->pndev[port] = dev; + + netif_carrier_off(dev); + err = register_netdev(dev); + if (err) { + mlx4_err(mdev, "Netdev registration failed\n"); + goto out; + } + priv->registered = 1; + queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY); + return 0; + +out: + mlx4_en_destroy_netdev(dev); + return err; +} + diff --git a/drivers/net/mlx4/en_params.c b/drivers/net/mlx4/en_params.c new file mode 100644 index 0000000..c2e69b1 --- /dev/null +++ b/drivers/net/mlx4/en_params.c @@ -0,0 +1,480 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/kernel.h> +#include <linux/ethtool.h> +#include <linux/netdevice.h> + +#include "mlx4_en.h" +#include "en_port.h" + +#define MLX4_EN_PARM_INT(X, def_val, desc) \ + static unsigned int X = def_val;\ + module_param(X , uint, 0444); \ + MODULE_PARM_DESC(X, desc); + + +/* + * Device scope module parameters + */ + + +/* Use a XOR rathern than Toeplitz hash function for RSS */ +MLX4_EN_PARM_INT(rss_xor, 0, "Use XOR hash function for RSS"); + +/* RSS hash type mask - default to <saddr, daddr, sport, dport> */ +MLX4_EN_PARM_INT(rss_mask, 0xf, "RSS hash type bitmask"); + +/* Number of LRO sessions per Rx ring (rounded up to a power of two) */ +MLX4_EN_PARM_INT(num_lro, MLX4_EN_MAX_LRO_DESCRIPTORS, + "Number of LRO sessions per ring or disabled (0)"); + +/* Priority pausing */ +MLX4_EN_PARM_INT(pptx, MLX4_EN_DEF_TX_PAUSE, + "Pause policy on TX: 0 never generate pause frames " + "1 generate pause frames according to RX buffer threshold"); +MLX4_EN_PARM_INT(pprx, MLX4_EN_DEF_RX_PAUSE, + "Pause policy on RX: 0 ignore received pause frames " + "1 respect received pause frames"); +MLX4_EN_PARM_INT(pfctx, 0, "Priority based Flow Control policy on TX[7:0]." + " Per priority bit mask"); +MLX4_EN_PARM_INT(pfcrx, 0, "Priority based Flow Control policy on RX[7:0]." + " Per priority bit mask"); + +/* Interrupt moderation tunning */ +MLX4_EN_PARM_INT(rx_moder_cnt, MLX4_EN_AUTO_CONF, + "Max coalesced descriptors for Rx interrupt moderation"); +MLX4_EN_PARM_INT(rx_moder_time, MLX4_EN_AUTO_CONF, + "Timeout following last packet for Rx interrupt moderation"); +MLX4_EN_PARM_INT(auto_moder, 1, "Enable dynamic interrupt moderation"); + +MLX4_EN_PARM_INT(rx_ring_num1, 0, "Number or Rx rings for port 1 (0 = #cores)"); +MLX4_EN_PARM_INT(rx_ring_num2, 0, "Number or Rx rings for port 2 (0 = #cores)"); + +MLX4_EN_PARM_INT(tx_ring_size1, MLX4_EN_AUTO_CONF, "Tx ring size for port 1"); +MLX4_EN_PARM_INT(tx_ring_size2, MLX4_EN_AUTO_CONF, "Tx ring size for port 2"); +MLX4_EN_PARM_INT(rx_ring_size1, MLX4_EN_AUTO_CONF, "Rx ring size for port 1"); +MLX4_EN_PARM_INT(rx_ring_size2, MLX4_EN_AUTO_CONF, "Rx ring size for port 2"); + + +int mlx4_en_get_profile(struct mlx4_en_dev *mdev) +{ + struct mlx4_en_profile *params = &mdev->profile; + + params->rx_moder_cnt = min_t(int, rx_moder_cnt, MLX4_EN_AUTO_CONF); + params->rx_moder_time = min_t(int, rx_moder_time, MLX4_EN_AUTO_CONF); + params->auto_moder = auto_moder; + params->rss_xor = (rss_xor != 0); + params->rss_mask = rss_mask & 0x1f; + params->num_lro = min_t(int, num_lro , MLX4_EN_MAX_LRO_DESCRIPTORS); + params->rx_pause = pprx; + params->rx_ppp = pfcrx; + params->tx_pause = pptx; + params->tx_ppp = pfctx; + if (params->rx_ppp || params->tx_ppp) { + params->prof[1].tx_ring_num = MLX4_EN_TX_RING_NUM; + params->prof[2].tx_ring_num = MLX4_EN_TX_RING_NUM; + } else { + params->prof[1].tx_ring_num = 1; + params->prof[2].tx_ring_num = 1; + } + params->prof[1].rx_ring_num = min_t(int, rx_ring_num1, MAX_RX_RINGS); + params->prof[2].rx_ring_num = min_t(int, rx_ring_num2, MAX_RX_RINGS); + + if (tx_ring_size1 == MLX4_EN_AUTO_CONF) + tx_ring_size1 = MLX4_EN_DEF_TX_RING_SIZE; + params->prof[1].tx_ring_size = + (tx_ring_size1 < MLX4_EN_MIN_TX_SIZE) ? + MLX4_EN_MIN_TX_SIZE : roundup_pow_of_two(tx_ring_size1); + + if (tx_ring_size2 == MLX4_EN_AUTO_CONF) + tx_ring_size2 = MLX4_EN_DEF_TX_RING_SIZE; + params->prof[2].tx_ring_size = + (tx_ring_size2 < MLX4_EN_MIN_TX_SIZE) ? + MLX4_EN_MIN_TX_SIZE : roundup_pow_of_two(tx_ring_size2); + + if (rx_ring_size1 == MLX4_EN_AUTO_CONF) + rx_ring_size1 = MLX4_EN_DEF_RX_RING_SIZE; + params->prof[1].rx_ring_size = + (rx_ring_size1 < MLX4_EN_MIN_RX_SIZE) ? + MLX4_EN_MIN_RX_SIZE : roundup_pow_of_two(rx_ring_size1); + + if (rx_ring_size2 == MLX4_EN_AUTO_CONF) + rx_ring_size2 = MLX4_EN_DEF_RX_RING_SIZE; + params->prof[2].rx_ring_size = + (rx_ring_size2 < MLX4_EN_MIN_RX_SIZE) ? + MLX4_EN_MIN_RX_SIZE : roundup_pow_of_two(rx_ring_size2); + return 0; +} + + +/* + * Ethtool support + */ + +static void mlx4_en_update_lro_stats(struct mlx4_en_priv *priv) +{ + int i; + + priv->port_stats.lro_aggregated = 0; + priv->port_stats.lro_flushed = 0; + priv->port_stats.lro_no_desc = 0; + + for (i = 0; i < priv->rx_ring_num; i++) { + priv->port_stats.lro_aggregated += priv->rx_ring[i].lro.stats.aggregated; + priv->port_stats.lro_flushed += priv->rx_ring[i].lro.stats.flushed; + priv->port_stats.lro_no_desc += priv->rx_ring[i].lro.stats.no_desc; + } +} + +static void +mlx4_en_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *drvinfo) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + sprintf(drvinfo->driver, DRV_NAME " (%s)", mdev->dev->board_id); + strncpy(drvinfo->version, DRV_VERSION " (" DRV_RELDATE ")", 32); + sprintf(drvinfo->fw_version, "%d.%d.%d", + (u16) (mdev->dev->caps.fw_ver >> 32), + (u16) ((mdev->dev->caps.fw_ver >> 16) & 0xffff), + (u16) (mdev->dev->caps.fw_ver & 0xffff)); + strncpy(drvinfo->bus_info, pci_name(mdev->dev->pdev), 32); + drvinfo->n_stats = 0; + drvinfo->regdump_len = 0; + drvinfo->eedump_len = 0; +} + +static u32 mlx4_en_get_tso(struct net_device *dev) +{ + return (dev->features & NETIF_F_TSO) != 0; +} + +static int mlx4_en_set_tso(struct net_device *dev, u32 data) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + + if (data) { + if (!priv->mdev->LSO_support) + return -EPERM; + dev->features |= (NETIF_F_TSO | NETIF_F_TSO6); + } else + dev->features &= ~(NETIF_F_TSO | NETIF_F_TSO6); + return 0; +} + +static u32 mlx4_en_get_rx_csum(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + return priv->rx_csum; +} + +static int mlx4_en_set_rx_csum(struct net_device *dev, u32 data) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + priv->rx_csum = (data != 0); + return 0; +} + +static const char main_strings[][ETH_GSTRING_LEN] = { + "rx_packets", "tx_packets", "rx_bytes", "tx_bytes", "rx_errors", + "tx_errors", "rx_dropped", "tx_dropped", "multicast", "collisions", + "rx_length_errors", "rx_over_errors", "rx_crc_errors", + "rx_frame_errors", "rx_fifo_errors", "rx_missed_errors", + "tx_aborted_errors", "tx_carrier_errors", "tx_fifo_errors", + "tx_heartbeat_errors", "tx_window_errors", + + /* port statistics */ + "lro_aggregated", "lro_flushed", "lro_no_desc", "tso_packets", + "queue_stopped", "wake_queue", "tx_timeout", "rx_alloc_failed", + "rx_csum_good", "rx_csum_none", "tx_chksum_offload", + + /* packet statistics */ + "broadcast", "rx_prio_0", "rx_prio_1", "rx_prio_2", "rx_prio_3", + "rx_prio_4", "rx_prio_5", "rx_prio_6", "rx_prio_7", "tx_prio_0", + "tx_prio_1", "tx_prio_2", "tx_prio_3", "tx_prio_4", "tx_prio_5", + "tx_prio_6", "tx_prio_7", +}; +#define NUM_MAIN_STATS 21 +#define NUM_ALL_STATS (NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PKT_STATS + NUM_PERF_STATS) + +static u32 mlx4_en_get_msglevel(struct net_device *dev) +{ + return ((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable; +} + +static void mlx4_en_set_msglevel(struct net_device *dev, u32 val) +{ + ((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable = val; +} + +static void mlx4_en_get_wol(struct net_device *netdev, + struct ethtool_wolinfo *wol) +{ + wol->supported = 0; + wol->wolopts = 0; + + return; +} + +static int mlx4_en_get_sset_count(struct net_device *dev, int sset) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + + if (sset != ETH_SS_STATS) + return -EOPNOTSUPP; + + return NUM_ALL_STATS + (priv->tx_ring_num + priv->rx_ring_num) * 2; +} + +static void mlx4_en_get_ethtool_stats(struct net_device *dev, + struct ethtool_stats *stats, uint64_t *data) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int index = 0; + int i; + + spin_lock_bh(&priv->stats_lock); + + mlx4_en_update_lro_stats(priv); + + for (i = 0; i < NUM_MAIN_STATS; i++) + data[index++] = ((unsigned long *) &priv->stats)[i]; + for (i = 0; i < NUM_PORT_STATS; i++) + data[index++] = ((unsigned long *) &priv->port_stats)[i]; + for (i = 0; i < priv->tx_ring_num; i++) { + data[index++] = priv->tx_ring[i].packets; + data[index++] = priv->tx_ring[i].bytes; + } + for (i = 0; i < priv->rx_ring_num; i++) { + data[index++] = priv->rx_ring[i].packets; + data[index++] = priv->rx_ring[i].bytes; + } + for (i = 0; i < NUM_PKT_STATS; i++) + data[index++] = ((unsigned long *) &priv->pkstats)[i]; + spin_unlock_bh(&priv->stats_lock); + +} + +static void mlx4_en_get_strings(struct net_device *dev, + uint32_t stringset, uint8_t *data) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int index = 0; + int i; + + if (stringset != ETH_SS_STATS) + return; + + /* Add main counters */ + for (i = 0; i < NUM_MAIN_STATS; i++) + strcpy(data + (index++) * ETH_GSTRING_LEN, main_strings[i]); + for (i = 0; i < NUM_PORT_STATS; i++) + strcpy(data + (index++) * ETH_GSTRING_LEN, + main_strings[i + NUM_MAIN_STATS]); + for (i = 0; i < priv->tx_ring_num; i++) { + sprintf(data + (index++) * ETH_GSTRING_LEN, + "tx%d_packets", i); + sprintf(data + (index++) * ETH_GSTRING_LEN, + "tx%d_bytes", i); + } + for (i = 0; i < priv->rx_ring_num; i++) { + sprintf(data + (index++) * ETH_GSTRING_LEN, + "rx%d_packets", i); + sprintf(data + (index++) * ETH_GSTRING_LEN, + "rx%d_bytes", i); + } + for (i = 0; i < NUM_PKT_STATS; i++) + strcpy(data + (index++) * ETH_GSTRING_LEN, + main_strings[i + NUM_MAIN_STATS + NUM_PORT_STATS]); +} + +static int mlx4_en_get_settings(struct net_device *dev, struct ethtool_cmd *cmd) +{ + cmd->autoneg = AUTONEG_DISABLE; + cmd->supported = SUPPORTED_10000baseT_Full; + cmd->advertising = SUPPORTED_10000baseT_Full; + if (netif_carrier_ok(dev)) { + cmd->speed = SPEED_10000; + cmd->duplex = DUPLEX_FULL; + } else { + cmd->speed = -1; + cmd->duplex = -1; + } + return 0; +} + +static int mlx4_en_set_settings(struct net_device *dev, struct ethtool_cmd *cmd) +{ + if ((cmd->autoneg == AUTONEG_ENABLE) || + (cmd->speed != SPEED_10000) || (cmd->duplex != DUPLEX_FULL)) + return -EINVAL; + + /* Nothing to change */ + return 0; +} + +static int mlx4_en_get_coalesce(struct net_device *dev, + struct ethtool_coalesce *coal) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + + coal->tx_coalesce_usecs = 0; + coal->tx_max_coalesced_frames = 0; + coal->rx_coalesce_usecs = priv->rx_usecs; + coal->rx_max_coalesced_frames = priv->rx_frames; + + coal->pkt_rate_low = priv->pkt_rate_low; + coal->rx_coalesce_usecs_low = priv->rx_usecs_low; + coal->pkt_rate_high = priv->pkt_rate_high; + coal->rx_coalesce_usecs_high = priv->rx_usecs_high; + coal->rate_sample_interval = priv->sample_interval; + coal->use_adaptive_rx_coalesce = priv->adaptive_rx_coal; + return 0; +} + +static int mlx4_en_set_coalesce(struct net_device *dev, + struct ethtool_coalesce *coal) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int err, i; + + priv->rx_frames = (coal->rx_max_coalesced_frames == + MLX4_EN_AUTO_CONF) ? + MLX4_EN_RX_COAL_TARGET / + priv->dev->mtu + 1 : + coal->rx_max_coalesced_frames; + priv->rx_usecs = (coal->rx_coalesce_usecs == + MLX4_EN_AUTO_CONF) ? + MLX4_EN_RX_COAL_TIME : + coal->rx_coalesce_usecs; + + /* Set adaptive coalescing params */ + priv->pkt_rate_low = coal->pkt_rate_low; + priv->rx_usecs_low = coal->rx_coalesce_usecs_low; + priv->pkt_rate_high = coal->pkt_rate_high; + priv->rx_usecs_high = coal->rx_coalesce_usecs_high; + priv->sample_interval = coal->rate_sample_interval; + priv->adaptive_rx_coal = coal->use_adaptive_rx_coalesce; + priv->last_moder_time = MLX4_EN_AUTO_CONF; + if (priv->adaptive_rx_coal) + return 0; + + for (i = 0; i < priv->rx_ring_num; i++) { + priv->rx_cq[i].moder_cnt = priv->rx_frames; + priv->rx_cq[i].moder_time = priv->rx_usecs; + err = mlx4_en_set_cq_moder(priv, &priv->rx_cq[i]); + if (err) + return err; + } + return 0; +} + +static int mlx4_en_set_pauseparam(struct net_device *dev, + struct ethtool_pauseparam *pause) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + mdev->profile.tx_pause = pause->tx_pause != 0; + mdev->profile.rx_pause = pause->rx_pause != 0; + err = mlx4_SET_PORT_general(mdev->dev, priv->port, + priv->rx_skb_size + ETH_FCS_LEN, + mdev->profile.tx_pause, + mdev->profile.tx_ppp, + mdev->profile.rx_pause, + mdev->profile.rx_ppp); + if (err) + mlx4_err(mdev, "Failed setting pause params to\n"); + + return err; +} + +static void mlx4_en_get_pauseparam(struct net_device *dev, + struct ethtool_pauseparam *pause) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + pause->tx_pause = mdev->profile.tx_pause; + pause->rx_pause = mdev->profile.rx_pause; +} + +static void mlx4_en_get_ringparam(struct net_device *dev, + struct ethtool_ringparam *param) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + + memset(param, 0, sizeof(*param)); + param->rx_max_pending = mdev->dev->caps.max_rq_sg; + param->tx_max_pending = mdev->dev->caps.max_sq_sg; + param->rx_pending = mdev->profile.prof[priv->port].rx_ring_size; + param->tx_pending = mdev->profile.prof[priv->port].tx_ring_size; +} + +const struct ethtool_ops mlx4_en_ethtool_ops = { + .get_drvinfo = mlx4_en_get_drvinfo, + .get_settings = mlx4_en_get_settings, + .set_settings = mlx4_en_set_settings, +#ifdef NETIF_F_TSO + .get_tso = mlx4_en_get_tso, + .set_tso = mlx4_en_set_tso, +#endif + .get_sg = ethtool_op_get_sg, + .set_sg = ethtool_op_set_sg, + .get_link = ethtool_op_get_link, + .get_rx_csum = mlx4_en_get_rx_csum, + .set_rx_csum = mlx4_en_set_rx_csum, + .get_tx_csum = ethtool_op_get_tx_csum, + .set_tx_csum = ethtool_op_set_tx_ipv6_csum, + .get_strings = mlx4_en_get_strings, + .get_sset_count = mlx4_en_get_sset_count, + .get_ethtool_stats = mlx4_en_get_ethtool_stats, + .get_wol = mlx4_en_get_wol, + .get_msglevel = mlx4_en_get_msglevel, + .set_msglevel = mlx4_en_set_msglevel, + .get_coalesce = mlx4_en_get_coalesce, + .set_coalesce = mlx4_en_set_coalesce, + .get_pauseparam = mlx4_en_get_pauseparam, + .set_pauseparam = mlx4_en_set_pauseparam, + .get_ringparam = mlx4_en_get_ringparam, + .get_flags = ethtool_op_get_flags, + .set_flags = ethtool_op_set_flags, +}; + + + + + diff --git a/drivers/net/mlx4/en_port.c b/drivers/net/mlx4/en_port.c new file mode 100644 index 0000000..c5a4c03 --- /dev/null +++ b/drivers/net/mlx4/en_port.c @@ -0,0 +1,261 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + + +#include <linux/if_vlan.h> + +#include <linux/mlx4/device.h> +#include <linux/mlx4/cmd.h> + +#include "en_port.h" +#include "mlx4_en.h" + + +int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port, + u64 mac, u64 clear, u8 mode) +{ + return mlx4_cmd(dev, (mac | (clear << 63)), port, mode, + MLX4_CMD_SET_MCAST_FLTR, MLX4_CMD_TIME_CLASS_B); +} + +int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, u8 port, struct vlan_group *grp) +{ + struct mlx4_cmd_mailbox *mailbox; + struct mlx4_set_vlan_fltr_mbox *filter; + int i; + int j; + int index = 0; + u32 entry; + int err = 0; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + + filter = mailbox->buf; + if (grp) { + memset(filter, 0, sizeof *filter); + for (i = VLAN_FLTR_SIZE - 1; i >= 0; i--) { + entry = 0; + for (j = 0; j < 32; j++) + if (vlan_group_get_device(grp, index++)) + entry |= 1 << j; + filter->entry[i] = cpu_to_be32(entry); + } + } else { + /* When no vlans are configured we block all vlans */ + memset(filter, 0, sizeof(*filter)); + } + err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_SET_VLAN_FLTR, + MLX4_CMD_TIME_CLASS_B); + mlx4_free_cmd_mailbox(dev, mailbox); + return err; +} + + +int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu, + u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx) +{ + struct mlx4_cmd_mailbox *mailbox; + struct mlx4_set_port_general_context *context; + int err; + u32 in_mod; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + context = mailbox->buf; + memset(context, 0, sizeof *context); + + context->flags = SET_PORT_GEN_ALL_VALID; + context->mtu = cpu_to_be16(mtu); + context->pptx = (pptx * (!pfctx)) << 7; + context->pfctx = pfctx; + context->pprx = (pprx * (!pfcrx)) << 7; + context->pfcrx = pfcrx; + + in_mod = MLX4_SET_PORT_GENERAL << 8 | port; + err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT, + MLX4_CMD_TIME_CLASS_B); + + mlx4_free_cmd_mailbox(dev, mailbox); + return err; +} + +int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn, + u8 promisc) +{ + struct mlx4_cmd_mailbox *mailbox; + struct mlx4_set_port_rqp_calc_context *context; + int err; + u32 in_mod; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + context = mailbox->buf; + memset(context, 0, sizeof *context); + + context->base_qpn = cpu_to_be32(base_qpn); + context->promisc = cpu_to_be32(promisc << SET_PORT_PROMISC_SHIFT | base_qpn); + context->mcast = cpu_to_be32(1 << SET_PORT_PROMISC_SHIFT | base_qpn); + context->intra_no_vlan = 0; + context->no_vlan = MLX4_NO_VLAN_IDX; + context->intra_vlan_miss = 0; + context->vlan_miss = MLX4_VLAN_MISS_IDX; + + in_mod = MLX4_SET_PORT_RQP_CALC << 8 | port; + err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT, + MLX4_CMD_TIME_CLASS_B); + + mlx4_free_cmd_mailbox(dev, mailbox); + return err; +} + + +int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset) +{ + struct mlx4_en_stat_out_mbox *mlx4_en_stats; + struct mlx4_en_priv *priv = netdev_priv(mdev->pndev[port]); + struct net_device_stats *stats = &priv->stats; + struct mlx4_cmd_mailbox *mailbox; + u64 in_mod = reset << 8 | port; + int err; + + mailbox = mlx4_alloc_cmd_mailbox(mdev->dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + memset(mailbox->buf, 0, sizeof(*mlx4_en_stats)); + err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, in_mod, 0, + MLX4_CMD_DUMP_ETH_STATS, MLX4_CMD_TIME_CLASS_B); + if (err) + goto out; + + mlx4_en_stats = mailbox->buf; + + spin_lock_bh(&priv->stats_lock); + + stats->rx_packets = be32_to_cpu(mlx4_en_stats->RTOTFRMS) - + be32_to_cpu(mlx4_en_stats->RDROP); + stats->tx_packets = be64_to_cpu(mlx4_en_stats->TTOT_prio_0) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_1) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_2) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_3) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_4) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_5) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_6) + + be64_to_cpu(mlx4_en_stats->TTOT_prio_7) + + be64_to_cpu(mlx4_en_stats->TTOT_novlan) + + be64_to_cpu(mlx4_en_stats->TTOT_loopbk); + stats->rx_bytes = be64_to_cpu(mlx4_en_stats->ROCT_prio_0) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_1) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_2) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_3) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_4) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_5) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_6) + + be64_to_cpu(mlx4_en_stats->ROCT_prio_7) + + be64_to_cpu(mlx4_en_stats->ROCT_novlan); + + stats->tx_bytes = be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_0) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_1) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_2) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_3) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_4) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_5) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_6) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_prio_7) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_novlan) + + be64_to_cpu(mlx4_en_stats->TTTLOCT_loopbk); + + stats->rx_errors = be64_to_cpu(mlx4_en_stats->PCS) + + be32_to_cpu(mlx4_en_stats->RdropLength) + + be32_to_cpu(mlx4_en_stats->RJBBR) + + be32_to_cpu(mlx4_en_stats->RCRC) + + be32_to_cpu(mlx4_en_stats->RRUNT); + stats->tx_errors = be32_to_cpu(mlx4_en_stats->TDROP); + stats->multicast = be64_to_cpu(mlx4_en_stats->MCAST_prio_0) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_1) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_2) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_3) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_4) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_5) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_6) + + be64_to_cpu(mlx4_en_stats->MCAST_prio_7) + + be64_to_cpu(mlx4_en_stats->MCAST_novlan); + stats->collisions = 0; + stats->rx_length_errors = be32_to_cpu(mlx4_en_stats->RdropLength); + stats->rx_over_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw); + stats->rx_crc_errors = be32_to_cpu(mlx4_en_stats->RCRC); + stats->rx_frame_errors = 0; + stats->rx_fifo_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw); + stats->rx_missed_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw); + stats->tx_aborted_errors = 0; + stats->tx_carrier_errors = 0; + stats->tx_fifo_errors = 0; + stats->tx_heartbeat_errors = 0; + stats->tx_window_errors = 0; + + priv->pkstats.broadcast = + be64_to_cpu(mlx4_en_stats->RBCAST_prio_0) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_1) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_2) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_3) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_4) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_5) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_6) + + be64_to_cpu(mlx4_en_stats->RBCAST_prio_7) + + be64_to_cpu(mlx4_en_stats->RBCAST_novlan); + priv->pkstats.rx_prio[0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_0); + priv->pkstats.rx_prio[1] = be64_to_cpu(mlx4_en_stats->RTOT_prio_1); + priv->pkstats.rx_prio[2] = be64_to_cpu(mlx4_en_stats->RTOT_prio_2); + priv->pkstats.rx_prio[3] = be64_to_cpu(mlx4_en_stats->RTOT_prio_3); + priv->pkstats.rx_prio[4] = be64_to_cpu(mlx4_en_stats->RTOT_prio_4); + priv->pkstats.rx_prio[5] = be64_to_cpu(mlx4_en_stats->RTOT_prio_5); + priv->pkstats.rx_prio[6] = be64_to_cpu(mlx4_en_stats->RTOT_prio_6); + priv->pkstats.rx_prio[7] = be64_to_cpu(mlx4_en_stats->RTOT_prio_7); + priv->pkstats.tx_prio[0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_0); + priv->pkstats.tx_prio[1] = be64_to_cpu(mlx4_en_stats->TTOT_prio_1); + priv->pkstats.tx_prio[2] = be64_to_cpu(mlx4_en_stats->TTOT_prio_2); + priv->pkstats.tx_prio[3] = be64_to_cpu(mlx4_en_stats->TTOT_prio_3); + priv->pkstats.tx_prio[4] = be64_to_cpu(mlx4_en_stats->TTOT_prio_4); + priv->pkstats.tx_prio[5] = be64_to_cpu(mlx4_en_stats->TTOT_prio_5); + priv->pkstats.tx_prio[6] = be64_to_cpu(mlx4_en_stats->TTOT_prio_6); + priv->pkstats.tx_prio[7] = be64_to_cpu(mlx4_en_stats->TTOT_prio_7); + spin_unlock_bh(&priv->stats_lock); + +out: + mlx4_free_cmd_mailbox(mdev->dev, mailbox); + return err; +} + diff --git a/drivers/net/mlx4/en_port.h b/drivers/net/mlx4/en_port.h new file mode 100644 index 0000000..e6477f1 --- /dev/null +++ b/drivers/net/mlx4/en_port.h @@ -0,0 +1,570 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef _MLX4_EN_PORT_H_ +#define _MLX4_EN_PORT_H_ + + +#define SET_PORT_GEN_ALL_VALID 0x7 +#define SET_PORT_PROMISC_SHIFT 31 + +enum { + MLX4_CMD_SET_VLAN_FLTR = 0x47, + MLX4_CMD_SET_MCAST_FLTR = 0x48, + MLX4_CMD_DUMP_ETH_STATS = 0x49, +}; + +struct mlx4_set_port_general_context { + u8 reserved[3]; + u8 flags; + u16 reserved2; + __be16 mtu; + u8 pptx; + u8 pfctx; + u16 reserved3; + u8 pprx; + u8 pfcrx; + u16 reserved4; +}; + +struct mlx4_set_port_rqp_calc_context { + __be32 base_qpn; + __be32 flags; + u8 reserved[3]; + u8 mac_miss; + u8 intra_no_vlan; + u8 no_vlan; + u8 intra_vlan_miss; + u8 vlan_miss; + u8 reserved2[3]; + u8 no_vlan_prio; + __be32 promisc; + __be32 mcast; +}; + +#define VLAN_FLTR_SIZE 128 +struct mlx4_set_vlan_fltr_mbox { + __be32 entry[VLAN_FLTR_SIZE]; +}; + + +enum { + MLX4_MCAST_CONFIG = 0, + MLX4_MCAST_DISABLE = 1, + MLX4_MCAST_ENABLE = 2, +}; + + +struct mlx4_en_stat_out_mbox { + /* Received frames with a length of 64 octets */ + __be64 R64_prio_0; + __be64 R64_prio_1; + __be64 R64_prio_2; + __be64 R64_prio_3; + __be64 R64_prio_4; + __be64 R64_prio_5; + __be64 R64_prio_6; + __be64 R64_prio_7; + __be64 R64_novlan; + /* Received frames with a length of 127 octets */ + __be64 R127_prio_0; + __be64 R127_prio_1; + __be64 R127_prio_2; + __be64 R127_prio_3; + __be64 R127_prio_4; + __be64 R127_prio_5; + __be64 R127_prio_6; + __be64 R127_prio_7; + __be64 R127_novlan; + /* Received frames with a length of 255 octets */ + __be64 R255_prio_0; + __be64 R255_prio_1; + __be64 R255_prio_2; + __be64 R255_prio_3; + __be64 R255_prio_4; + __be64 R255_prio_5; + __be64 R255_prio_6; + __be64 R255_prio_7; + __be64 R255_novlan; + /* Received frames with a length of 511 octets */ + __be64 R511_prio_0; + __be64 R511_prio_1; + __be64 R511_prio_2; + __be64 R511_prio_3; + __be64 R511_prio_4; + __be64 R511_prio_5; + __be64 R511_prio_6; + __be64 R511_prio_7; + __be64 R511_novlan; + /* Received frames with a length of 1023 octets */ + __be64 R1023_prio_0; + __be64 R1023_prio_1; + __be64 R1023_prio_2; + __be64 R1023_prio_3; + __be64 R1023_prio_4; + __be64 R1023_prio_5; + __be64 R1023_prio_6; + __be64 R1023_prio_7; + __be64 R1023_novlan; + /* Received frames with a length of 1518 octets */ + __be64 R1518_prio_0; + __be64 R1518_prio_1; + __be64 R1518_prio_2; + __be64 R1518_prio_3; + __be64 R1518_prio_4; + __be64 R1518_prio_5; + __be64 R1518_prio_6; + __be64 R1518_prio_7; + __be64 R1518_novlan; + /* Received frames with a length of 1522 octets */ + __be64 R1522_prio_0; + __be64 R1522_prio_1; + __be64 R1522_prio_2; + __be64 R1522_prio_3; + __be64 R1522_prio_4; + __be64 R1522_prio_5; + __be64 R1522_prio_6; + __be64 R1522_prio_7; + __be64 R1522_novlan; + /* Received frames with a length of 1548 octets */ + __be64 R1548_prio_0; + __be64 R1548_prio_1; + __be64 R1548_prio_2; + __be64 R1548_prio_3; + __be64 R1548_prio_4; + __be64 R1548_prio_5; + __be64 R1548_prio_6; + __be64 R1548_prio_7; + __be64 R1548_novlan; + /* Received frames with a length of 1548 < octets < MTU */ + __be64 R2MTU_prio_0; + __be64 R2MTU_prio_1; + __be64 R2MTU_prio_2; + __be64 R2MTU_prio_3; + __be64 R2MTU_prio_4; + __be64 R2MTU_prio_5; + __be64 R2MTU_prio_6; + __be64 R2MTU_prio_7; + __be64 R2MTU_novlan; + /* Received frames with a length of MTU< octets and good CRC */ + __be64 RGIANT_prio_0; + __be64 RGIANT_prio_1; + __be64 RGIANT_prio_2; + __be64 RGIANT_prio_3; + __be64 RGIANT_prio_4; + __be64 RGIANT_prio_5; + __be64 RGIANT_prio_6; + __be64 RGIANT_prio_7; + __be64 RGIANT_novlan; + /* Received broadcast frames with good CRC */ + __be64 RBCAST_prio_0; + __be64 RBCAST_prio_1; + __be64 RBCAST_prio_2; + __be64 RBCAST_prio_3; + __be64 RBCAST_prio_4; + __be64 RBCAST_prio_5; + __be64 RBCAST_prio_6; + __be64 RBCAST_prio_7; + __be64 RBCAST_novlan; + /* Received multicast frames with good CRC */ + __be64 MCAST_prio_0; + __be64 MCAST_prio_1; + __be64 MCAST_prio_2; + __be64 MCAST_prio_3; + __be64 MCAST_prio_4; + __be64 MCAST_prio_5; + __be64 MCAST_prio_6; + __be64 MCAST_prio_7; + __be64 MCAST_novlan; + /* Received unicast not short or GIANT frames with good CRC */ + __be64 RTOTG_prio_0; + __be64 RTOTG_prio_1; + __be64 RTOTG_prio_2; + __be64 RTOTG_prio_3; + __be64 RTOTG_prio_4; + __be64 RTOTG_prio_5; + __be64 RTOTG_prio_6; + __be64 RTOTG_prio_7; + __be64 RTOTG_novlan; + + /* Count of total octets of received frames, includes framing characters */ + __be64 RTTLOCT_prio_0; + /* Count of total octets of received frames, not including framing + characters */ + __be64 RTTLOCT_NOFRM_prio_0; + /* Count of Total number of octets received + (only for frames without errors) */ + __be64 ROCT_prio_0; + + __be64 RTTLOCT_prio_1; + __be64 RTTLOCT_NOFRM_prio_1; + __be64 ROCT_prio_1; + + __be64 RTTLOCT_prio_2; + __be64 RTTLOCT_NOFRM_prio_2; + __be64 ROCT_prio_2; + + __be64 RTTLOCT_prio_3; + __be64 RTTLOCT_NOFRM_prio_3; + __be64 ROCT_prio_3; + + __be64 RTTLOCT_prio_4; + __be64 RTTLOCT_NOFRM_prio_4; + __be64 ROCT_prio_4; + + __be64 RTTLOCT_prio_5; + __be64 RTTLOCT_NOFRM_prio_5; + __be64 ROCT_prio_5; + + __be64 RTTLOCT_prio_6; + __be64 RTTLOCT_NOFRM_prio_6; + __be64 ROCT_prio_6; + + __be64 RTTLOCT_prio_7; + __be64 RTTLOCT_NOFRM_prio_7; + __be64 ROCT_prio_7; + + __be64 RTTLOCT_novlan; + __be64 RTTLOCT_NOFRM_novlan; + __be64 ROCT_novlan; + + /* Count of Total received frames including bad frames */ + __be64 RTOT_prio_0; + /* Count of Total number of received frames with 802.1Q encapsulation */ + __be64 R1Q_prio_0; + __be64 reserved1; + + __be64 RTOT_prio_1; + __be64 R1Q_prio_1; + __be64 reserved2; + + __be64 RTOT_prio_2; + __be64 R1Q_prio_2; + __be64 reserved3; + + __be64 RTOT_prio_3; + __be64 R1Q_prio_3; + __be64 reserved4; + + __be64 RTOT_prio_4; + __be64 R1Q_prio_4; + __be64 reserved5; + + __be64 RTOT_prio_5; + __be64 R1Q_prio_5; + __be64 reserved6; + + __be64 RTOT_prio_6; + __be64 R1Q_prio_6; + __be64 reserved7; + + __be64 RTOT_prio_7; + __be64 R1Q_prio_7; + __be64 reserved8; + + __be64 RTOT_novlan; + __be64 R1Q_novlan; + __be64 reserved9; + + /* Total number of Successfully Received Control Frames */ + __be64 RCNTL; + __be64 reserved10; + __be64 reserved11; + __be64 reserved12; + /* Count of received frames with a length/type field value between 46 + (42 for VLANtagged frames) and 1500 (also 1500 for VLAN-tagged frames), + inclusive */ + __be64 RInRangeLengthErr; + /* Count of received frames with length/type field between 1501 and 1535 + decimal, inclusive */ + __be64 ROutRangeLengthErr; + /* Count of received frames that are longer than max allowed size for + 802.3 frames (1518/1522) */ + __be64 RFrmTooLong; + /* Count frames received with PCS error */ + __be64 PCS; + + /* Transmit frames with a length of 64 octets */ + __be64 T64_prio_0; + __be64 T64_prio_1; + __be64 T64_prio_2; + __be64 T64_prio_3; + __be64 T64_prio_4; + __be64 T64_prio_5; + __be64 T64_prio_6; + __be64 T64_prio_7; + __be64 T64_novlan; + __be64 T64_loopbk; + /* Transmit frames with a length of 65 to 127 octets. */ + __be64 T127_prio_0; + __be64 T127_prio_1; + __be64 T127_prio_2; + __be64 T127_prio_3; + __be64 T127_prio_4; + __be64 T127_prio_5; + __be64 T127_prio_6; + __be64 T127_prio_7; + __be64 T127_novlan; + __be64 T127_loopbk; + /* Transmit frames with a length of 128 to 255 octets */ + __be64 T255_prio_0; + __be64 T255_prio_1; + __be64 T255_prio_2; + __be64 T255_prio_3; + __be64 T255_prio_4; + __be64 T255_prio_5; + __be64 T255_prio_6; + __be64 T255_prio_7; + __be64 T255_novlan; + __be64 T255_loopbk; + /* Transmit frames with a length of 256 to 511 octets */ + __be64 T511_prio_0; + __be64 T511_prio_1; + __be64 T511_prio_2; + __be64 T511_prio_3; + __be64 T511_prio_4; + __be64 T511_prio_5; + __be64 T511_prio_6; + __be64 T511_prio_7; + __be64 T511_novlan; + __be64 T511_loopbk; + /* Transmit frames with a length of 512 to 1023 octets */ + __be64 T1023_prio_0; + __be64 T1023_prio_1; + __be64 T1023_prio_2; + __be64 T1023_prio_3; + __be64 T1023_prio_4; + __be64 T1023_prio_5; + __be64 T1023_prio_6; + __be64 T1023_prio_7; + __be64 T1023_novlan; + __be64 T1023_loopbk; + /* Transmit frames with a length of 1024 to 1518 octets */ + __be64 T1518_prio_0; + __be64 T1518_prio_1; + __be64 T1518_prio_2; + __be64 T1518_prio_3; + __be64 T1518_prio_4; + __be64 T1518_prio_5; + __be64 T1518_prio_6; + __be64 T1518_prio_7; + __be64 T1518_novlan; + __be64 T1518_loopbk; + /* Counts transmit frames with a length of 1519 to 1522 bytes */ + __be64 T1522_prio_0; + __be64 T1522_prio_1; + __be64 T1522_prio_2; + __be64 T1522_prio_3; + __be64 T1522_prio_4; + __be64 T1522_prio_5; + __be64 T1522_prio_6; + __be64 T1522_prio_7; + __be64 T1522_novlan; + __be64 T1522_loopbk; + /* Transmit frames with a length of 1523 to 1548 octets */ + __be64 T1548_prio_0; + __be64 T1548_prio_1; + __be64 T1548_prio_2; + __be64 T1548_prio_3; + __be64 T1548_prio_4; + __be64 T1548_prio_5; + __be64 T1548_prio_6; + __be64 T1548_prio_7; + __be64 T1548_novlan; + __be64 T1548_loopbk; + /* Counts transmit frames with a length of 1549 to MTU bytes */ + __be64 T2MTU_prio_0; + __be64 T2MTU_prio_1; + __be64 T2MTU_prio_2; + __be64 T2MTU_prio_3; + __be64 T2MTU_prio_4; + __be64 T2MTU_prio_5; + __be64 T2MTU_prio_6; + __be64 T2MTU_prio_7; + __be64 T2MTU_novlan; + __be64 T2MTU_loopbk; + /* Transmit frames with a length greater than MTU octets and a good CRC. */ + __be64 TGIANT_prio_0; + __be64 TGIANT_prio_1; + __be64 TGIANT_prio_2; + __be64 TGIANT_prio_3; + __be64 TGIANT_prio_4; + __be64 TGIANT_prio_5; + __be64 TGIANT_prio_6; + __be64 TGIANT_prio_7; + __be64 TGIANT_novlan; + __be64 TGIANT_loopbk; + /* Transmit broadcast frames with a good CRC */ + __be64 TBCAST_prio_0; + __be64 TBCAST_prio_1; + __be64 TBCAST_prio_2; + __be64 TBCAST_prio_3; + __be64 TBCAST_prio_4; + __be64 TBCAST_prio_5; + __be64 TBCAST_prio_6; + __be64 TBCAST_prio_7; + __be64 TBCAST_novlan; + __be64 TBCAST_loopbk; + /* Transmit multicast frames with a good CRC */ + __be64 TMCAST_prio_0; + __be64 TMCAST_prio_1; + __be64 TMCAST_prio_2; + __be64 TMCAST_prio_3; + __be64 TMCAST_prio_4; + __be64 TMCAST_prio_5; + __be64 TMCAST_prio_6; + __be64 TMCAST_prio_7; + __be64 TMCAST_novlan; + __be64 TMCAST_loopbk; + /* Transmit good frames that are neither broadcast nor multicast */ + __be64 TTOTG_prio_0; + __be64 TTOTG_prio_1; + __be64 TTOTG_prio_2; + __be64 TTOTG_prio_3; + __be64 TTOTG_prio_4; + __be64 TTOTG_prio_5; + __be64 TTOTG_prio_6; + __be64 TTOTG_prio_7; + __be64 TTOTG_novlan; + __be64 TTOTG_loopbk; + + /* total octets of transmitted frames, including framing characters */ + __be64 TTTLOCT_prio_0; + /* total octets of transmitted frames, not including framing characters */ + __be64 TTTLOCT_NOFRM_prio_0; + /* ifOutOctets */ + __be64 TOCT_prio_0; + + __be64 TTTLOCT_prio_1; + __be64 TTTLOCT_NOFRM_prio_1; + __be64 TOCT_prio_1; + + __be64 TTTLOCT_prio_2; + __be64 TTTLOCT_NOFRM_prio_2; + __be64 TOCT_prio_2; + + __be64 TTTLOCT_prio_3; + __be64 TTTLOCT_NOFRM_prio_3; + __be64 TOCT_prio_3; + + __be64 TTTLOCT_prio_4; + __be64 TTTLOCT_NOFRM_prio_4; + __be64 TOCT_prio_4; + + __be64 TTTLOCT_prio_5; + __be64 TTTLOCT_NOFRM_prio_5; + __be64 TOCT_prio_5; + + __be64 TTTLOCT_prio_6; + __be64 TTTLOCT_NOFRM_prio_6; + __be64 TOCT_prio_6; + + __be64 TTTLOCT_prio_7; + __be64 TTTLOCT_NOFRM_prio_7; + __be64 TOCT_prio_7; + + __be64 TTTLOCT_novlan; + __be64 TTTLOCT_NOFRM_novlan; + __be64 TOCT_novlan; + + __be64 TTTLOCT_loopbk; + __be64 TTTLOCT_NOFRM_loopbk; + __be64 TOCT_loopbk; + + /* Total frames transmitted with a good CRC that are not aborted */ + __be64 TTOT_prio_0; + /* Total number of frames transmitted with 802.1Q encapsulation */ + __be64 T1Q_prio_0; + __be64 reserved13; + + __be64 TTOT_prio_1; + __be64 T1Q_prio_1; + __be64 reserved14; + + __be64 TTOT_prio_2; + __be64 T1Q_prio_2; + __be64 reserved15; + + __be64 TTOT_prio_3; + __be64 T1Q_prio_3; + __be64 reserved16; + + __be64 TTOT_prio_4; + __be64 T1Q_prio_4; + __be64 reserved17; + + __be64 TTOT_prio_5; + __be64 T1Q_prio_5; + __be64 reserved18; + + __be64 TTOT_prio_6; + __be64 T1Q_prio_6; + __be64 reserved19; + + __be64 TTOT_prio_7; + __be64 T1Q_prio_7; + __be64 reserved20; + + __be64 TTOT_novlan; + __be64 T1Q_novlan; + __be64 reserved21; + + __be64 TTOT_loopbk; + __be64 T1Q_loopbk; + __be64 reserved22; + + /* Received frames with a length greater than MTU octets and a bad CRC */ + __be32 RJBBR; + /* Received frames with a bad CRC that are not runts, jabbers, + or alignment errors */ + __be32 RCRC; + /* Received frames with SFD with a length of less than 64 octets and a + bad CRC */ + __be32 RRUNT; + /* Received frames with a length less than 64 octets and a good CRC */ + __be32 RSHORT; + /* Total Number of Received Packets Dropped */ + __be32 RDROP; + /* Drop due to overflow */ + __be32 RdropOvflw; + /* Drop due to overflow */ + __be32 RdropLength; + /* Total of good frames. Does not include frames received with + frame-too-long, FCS, or length errors */ + __be32 RTOTFRMS; + /* Total dropped Xmited packets */ + __be32 TDROP; +}; + + +#endif diff --git a/drivers/net/mlx4/en_resources.c b/drivers/net/mlx4/en_resources.c new file mode 100644 index 0000000..a054520 --- /dev/null +++ b/drivers/net/mlx4/en_resources.c @@ -0,0 +1,96 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/vmalloc.h> +#include <linux/mlx4/qp.h> + +#include "mlx4_en.h" + +void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride, + int is_tx, int rss, int qpn, int cqn, int srqn, + struct mlx4_qp_context *context) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + memset(context, 0, sizeof *context); + context->flags = cpu_to_be32(7 << 16 | rss << 13); + context->pd = cpu_to_be32(mdev->priv_pdn); + context->mtu_msgmax = 0xff; + context->rq_size_stride = 0; + if (is_tx) + context->sq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4); + else + context->sq_size_stride = 1; + context->usr_page = cpu_to_be32(mdev->priv_uar.index); + context->local_qpn = cpu_to_be32(qpn); + context->pri_path.ackto = 1 & 0x07; + context->pri_path.sched_queue = 0x83 | (priv->port - 1) << 6; + context->pri_path.counter_index = 0xff; + context->cqn_send = cpu_to_be32(cqn); + context->cqn_recv = cpu_to_be32(cqn); + context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2); + if (!rss) + context->srqn = cpu_to_be32(MLX4_EN_USE_SRQ | srqn); +} + + +int mlx4_en_map_buffer(struct mlx4_buf *buf) +{ + struct page **pages; + int i; + + if (BITS_PER_LONG == 64 || buf->nbufs == 1) + return 0; + + pages = kmalloc(sizeof *pages * buf->nbufs, GFP_KERNEL); + if (!pages) + return -ENOMEM; + + for (i = 0; i < buf->nbufs; ++i) + pages[i] = virt_to_page(buf->page_list[i].buf); + + buf->direct.buf = vmap(pages, buf->nbufs, VM_MAP, PAGE_KERNEL); + kfree(pages); + if (!buf->direct.buf) + return -ENOMEM; + + return 0; +} + +void mlx4_en_unmap_buffer(struct mlx4_buf *buf) +{ + if (BITS_PER_LONG == 64 || buf->nbufs == 1) + return; + + vunmap(buf->direct.buf); +} diff --git a/drivers/net/mlx4/en_rx.c b/drivers/net/mlx4/en_rx.c new file mode 100644 index 0000000..6232227 --- /dev/null +++ b/drivers/net/mlx4/en_rx.c @@ -0,0 +1,1080 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/mlx4/cq.h> +#include <linux/mlx4/qp.h> +#include <linux/skbuff.h> +#include <linux/if_ether.h> +#include <linux/if_vlan.h> +#include <linux/vmalloc.h> + +#include "mlx4_en.h" + +static void *get_wqe(struct mlx4_en_rx_ring *ring, int n) +{ + int offset = n << ring->srq.wqe_shift; + return ring->buf + offset; +} + +static void mlx4_en_srq_event(struct mlx4_srq *srq, enum mlx4_event type) +{ + return; +} + +static int mlx4_en_get_frag_header(struct skb_frag_struct *frags, void **mac_hdr, + void **ip_hdr, void **tcpudp_hdr, + u64 *hdr_flags, void *priv) +{ + *mac_hdr = page_address(frags->page) + frags->page_offset; + *ip_hdr = *mac_hdr + ETH_HLEN; + *tcpudp_hdr = (struct tcphdr *)(*ip_hdr + sizeof(struct iphdr)); + *hdr_flags = LRO_IPV4 | LRO_TCP; + + return 0; +} + +static int mlx4_en_alloc_frag(struct mlx4_en_priv *priv, + struct mlx4_en_rx_desc *rx_desc, + struct skb_frag_struct *skb_frags, + struct mlx4_en_rx_alloc *ring_alloc, + int i) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_frag_info *frag_info = &priv->frag_info[i]; + struct mlx4_en_rx_alloc *page_alloc = &ring_alloc[i]; + struct page *page; + dma_addr_t dma; + + if (page_alloc->offset == frag_info->last_offset) { + /* Allocate new page */ + page = alloc_pages(GFP_ATOMIC | __GFP_COMP, MLX4_EN_ALLOC_ORDER); + if (!page) + return -ENOMEM; + + skb_frags[i].page = page_alloc->page; + skb_frags[i].page_offset = page_alloc->offset; + page_alloc->page = page; + page_alloc->offset = frag_info->frag_align; + } else { + page = page_alloc->page; + get_page(page); + + skb_frags[i].page = page; + skb_frags[i].page_offset = page_alloc->offset; + page_alloc->offset += frag_info->frag_stride; + } + dma = pci_map_single(mdev->pdev, page_address(skb_frags[i].page) + + skb_frags[i].page_offset, frag_info->frag_size, + PCI_DMA_FROMDEVICE); + rx_desc->data[i].addr = cpu_to_be64(dma); + return 0; +} + +static int mlx4_en_init_allocator(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_rx_alloc *page_alloc; + int i; + + for (i = 0; i < priv->num_frags; i++) { + page_alloc = &ring->page_alloc[i]; + page_alloc->page = alloc_pages(GFP_ATOMIC | __GFP_COMP, + MLX4_EN_ALLOC_ORDER); + if (!page_alloc->page) + goto out; + + page_alloc->offset = priv->frag_info[i].frag_align; + mlx4_dbg(DRV, priv, "Initialized allocator:%d with page:%p\n", + i, page_alloc->page); + } + return 0; + +out: + while (i--) { + page_alloc = &ring->page_alloc[i]; + put_page(page_alloc->page); + page_alloc->page = NULL; + } + return -ENOMEM; +} + +static void mlx4_en_destroy_allocator(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_rx_alloc *page_alloc; + int i; + + for (i = 0; i < priv->num_frags; i++) { + page_alloc = &ring->page_alloc[i]; + mlx4_dbg(DRV, priv, "Freeing allocator:%d count:%d\n", + i, page_count(page_alloc->page)); + + put_page(page_alloc->page); + page_alloc->page = NULL; + } +} + + +static void mlx4_en_init_rx_desc(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring, int index) +{ + struct mlx4_en_rx_desc *rx_desc = ring->buf + ring->stride * index; + struct skb_frag_struct *skb_frags = ring->rx_info + + (index << priv->log_rx_info); + int possible_frags; + int i; + + /* Pre-link descriptor */ + rx_desc->next.next_wqe_index = cpu_to_be16((index + 1) & ring->size_mask); + + /* Set size and memtype fields */ + for (i = 0; i < priv->num_frags; i++) { + skb_frags[i].size = priv->frag_info[i].frag_size; + rx_desc->data[i].byte_count = + cpu_to_be32(priv->frag_info[i].frag_size); + rx_desc->data[i].lkey = cpu_to_be32(priv->mdev->mr.key); + } + + /* If the number of used fragments does not fill up the ring stride, + * remaining (unused) fragments must be padded with null address/size + * and a special memory key */ + possible_frags = (ring->stride - sizeof(struct mlx4_en_rx_desc)) / DS_SIZE; + for (i = priv->num_frags; i < possible_frags; i++) { + rx_desc->data[i].byte_count = 0; + rx_desc->data[i].lkey = cpu_to_be32(MLX4_EN_MEMTYPE_PAD); + rx_desc->data[i].addr = 0; + } +} + + +static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring, int index) +{ + struct mlx4_en_rx_desc *rx_desc = ring->buf + (index * ring->stride); + struct skb_frag_struct *skb_frags = ring->rx_info + + (index << priv->log_rx_info); + int i; + + for (i = 0; i < priv->num_frags; i++) + if (mlx4_en_alloc_frag(priv, rx_desc, skb_frags, ring->page_alloc, i)) + goto err; + + return 0; + +err: + while (i--) + put_page(skb_frags[i].page); + return -ENOMEM; +} + +static inline void mlx4_en_update_rx_prod_db(struct mlx4_en_rx_ring *ring) +{ + *ring->wqres.db.db = cpu_to_be32(ring->prod & 0xffff); +} + +static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_rx_ring *ring; + int ring_ind; + int buf_ind; + + for (buf_ind = 0; buf_ind < priv->prof->rx_ring_size; buf_ind++) { + for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) { + ring = &priv->rx_ring[ring_ind]; + + if (mlx4_en_prepare_rx_desc(priv, ring, + ring->actual_size)) { + if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) { + mlx4_err(mdev, "Failed to allocate " + "enough rx buffers\n"); + return -ENOMEM; + } else { + if (netif_msg_rx_err(priv)) + mlx4_warn(mdev, + "Only %d buffers allocated\n", + ring->actual_size); + goto out; + } + } + ring->actual_size++; + ring->prod++; + } + } +out: + return 0; +} + +static int mlx4_en_fill_rx_buf(struct net_device *dev, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int num = 0; + int err; + + while ((u32) (ring->prod - ring->cons) < ring->actual_size) { + err = mlx4_en_prepare_rx_desc(priv, ring, ring->prod & + ring->size_mask); + if (err) { + if (netif_msg_rx_err(priv)) + mlx4_warn(priv->mdev, + "Failed preparing rx descriptor\n"); + priv->port_stats.rx_alloc_failed++; + break; + } + ++num; + ++ring->prod; + } + if ((u32) (ring->prod - ring->cons) == ring->size) + ring->full = 1; + + return num; +} + +static void mlx4_en_free_rx_buf(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct skb_frag_struct *skb_frags; + struct mlx4_en_rx_desc *rx_desc; + dma_addr_t dma; + int index; + int nr; + + mlx4_dbg(DRV, priv, "Freeing Rx buf - cons:%d prod:%d\n", + ring->cons, ring->prod); + + /* Unmap and free Rx buffers */ + BUG_ON((u32) (ring->prod - ring->cons) > ring->size); + while (ring->cons != ring->prod) { + index = ring->cons & ring->size_mask; + rx_desc = ring->buf + (index << ring->log_stride); + skb_frags = ring->rx_info + (index << priv->log_rx_info); + mlx4_dbg(DRV, priv, "Processing descriptor:%d\n", index); + + for (nr = 0; nr < priv->num_frags; nr++) { + mlx4_dbg(DRV, priv, "Freeing fragment:%d\n", nr); + dma = be64_to_cpu(rx_desc->data[nr].addr); + + mlx4_dbg(DRV, priv, "Unmaping buffer at dma:0x%llx\n", (u64) dma); + pci_unmap_single(mdev->pdev, dma, skb_frags[nr].size, + PCI_DMA_FROMDEVICE); + put_page(skb_frags[nr].page); + } + ++ring->cons; + } +} + + +void mlx4_en_rx_refill(struct work_struct *work) +{ + struct delayed_work *delay = container_of(work, struct delayed_work, work); + struct mlx4_en_priv *priv = container_of(delay, struct mlx4_en_priv, + refill_task); + struct mlx4_en_dev *mdev = priv->mdev; + struct net_device *dev = priv->dev; + struct mlx4_en_rx_ring *ring; + int need_refill = 0; + int i; + + mutex_lock(&mdev->state_lock); + if (!mdev->device_up || !priv->port_up) + goto out; + + /* We only get here if there are no receive buffers, so we can't race + * with Rx interrupts while filling buffers */ + for (i = 0; i < priv->rx_ring_num; i++) { + ring = &priv->rx_ring[i]; + if (ring->need_refill) { + if (mlx4_en_fill_rx_buf(dev, ring)) { + ring->need_refill = 0; + mlx4_en_update_rx_prod_db(ring); + } else + need_refill = 1; + } + } + if (need_refill) + queue_delayed_work(mdev->workqueue, &priv->refill_task, HZ); + +out: + mutex_unlock(&mdev->state_lock); +} + + +int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring, u32 size, u16 stride) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int err; + int tmp; + + /* Sanity check SRQ size before proceeding */ + if (size >= mdev->dev->caps.max_srq_wqes) + return -EINVAL; + + ring->prod = 0; + ring->cons = 0; + ring->size = size; + ring->size_mask = size - 1; + ring->stride = stride; + ring->log_stride = ffs(ring->stride) - 1; + ring->buf_size = ring->size * ring->stride; + + tmp = size * roundup_pow_of_two(MLX4_EN_MAX_RX_FRAGS * + sizeof(struct skb_frag_struct)); + ring->rx_info = vmalloc(tmp); + if (!ring->rx_info) { + mlx4_err(mdev, "Failed allocating rx_info ring\n"); + return -ENOMEM; + } + mlx4_dbg(DRV, priv, "Allocated rx_info ring at addr:%p size:%d\n", + ring->rx_info, tmp); + + err = mlx4_alloc_hwq_res(mdev->dev, &ring->wqres, + ring->buf_size, 2 * PAGE_SIZE); + if (err) + goto err_ring; + + err = mlx4_en_map_buffer(&ring->wqres.buf); + if (err) { + mlx4_err(mdev, "Failed to map RX buffer\n"); + goto err_hwq; + } + ring->buf = ring->wqres.buf.direct.buf; + + /* Configure lro mngr */ + memset(&ring->lro, 0, sizeof(struct net_lro_mgr)); + ring->lro.dev = priv->dev; + ring->lro.features = LRO_F_NAPI; + ring->lro.frag_align_pad = NET_IP_ALIGN; + ring->lro.ip_summed = CHECKSUM_UNNECESSARY; + ring->lro.ip_summed_aggr = CHECKSUM_UNNECESSARY; + ring->lro.max_desc = mdev->profile.num_lro; + ring->lro.max_aggr = MAX_SKB_FRAGS; + ring->lro.lro_arr = kzalloc(mdev->profile.num_lro * + sizeof(struct net_lro_desc), + GFP_KERNEL); + if (!ring->lro.lro_arr) { + mlx4_err(mdev, "Failed to allocate lro array\n"); + goto err_map; + } + ring->lro.get_frag_header = mlx4_en_get_frag_header; + + return 0; + +err_map: + mlx4_en_unmap_buffer(&ring->wqres.buf); +err_hwq: + mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size); +err_ring: + vfree(ring->rx_info); + ring->rx_info = NULL; + return err; +} + +int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_wqe_srq_next_seg *next; + struct mlx4_en_rx_ring *ring; + int i; + int ring_ind; + int err; + int stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) + + DS_SIZE * priv->num_frags); + int max_gs = (stride - sizeof(struct mlx4_wqe_srq_next_seg)) / DS_SIZE; + + for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) { + ring = &priv->rx_ring[ring_ind]; + + ring->prod = 0; + ring->cons = 0; + ring->actual_size = 0; + ring->cqn = priv->rx_cq[ring_ind].mcq.cqn; + + ring->stride = stride; + ring->log_stride = ffs(ring->stride) - 1; + ring->buf_size = ring->size * ring->stride; + + memset(ring->buf, 0, ring->buf_size); + mlx4_en_update_rx_prod_db(ring); + + /* Initailize all descriptors */ + for (i = 0; i < ring->size; i++) + mlx4_en_init_rx_desc(priv, ring, i); + + /* Initialize page allocators */ + err = mlx4_en_init_allocator(priv, ring); + if (err) { + mlx4_err(mdev, "Failed initializing ring allocator\n"); + goto err_allocator; + } + + /* Fill Rx buffers */ + ring->full = 0; + } + if (mlx4_en_fill_rx_buffers(priv)) + goto err_buffers; + + for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) { + ring = &priv->rx_ring[ring_ind]; + + mlx4_en_update_rx_prod_db(ring); + + /* Configure SRQ representing the ring */ + ring->srq.max = ring->size; + ring->srq.max_gs = max_gs; + ring->srq.wqe_shift = ilog2(ring->stride); + + for (i = 0; i < ring->srq.max; ++i) { + next = get_wqe(ring, i); + next->next_wqe_index = + cpu_to_be16((i + 1) & (ring->srq.max - 1)); + } + + err = mlx4_srq_alloc(mdev->dev, mdev->priv_pdn, &ring->wqres.mtt, + ring->wqres.db.dma, &ring->srq); + if (err){ + mlx4_err(mdev, "Failed to allocate srq\n"); + goto err_srq; + } + ring->srq.event = mlx4_en_srq_event; + } + + return 0; + +err_srq: + while (ring_ind >= 0) { + ring = &priv->rx_ring[ring_ind]; + mlx4_srq_free(mdev->dev, &ring->srq); + ring_ind--; + } + +err_buffers: + for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) + mlx4_en_free_rx_buf(priv, &priv->rx_ring[ring_ind]); + + ring_ind = priv->rx_ring_num - 1; +err_allocator: + while (ring_ind >= 0) { + mlx4_en_destroy_allocator(priv, &priv->rx_ring[ring_ind]); + ring_ind--; + } + return err; +} + +void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + kfree(ring->lro.lro_arr); + mlx4_en_unmap_buffer(&ring->wqres.buf); + mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size); + vfree(ring->rx_info); + ring->rx_info = NULL; +} + +void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + mlx4_srq_free(mdev->dev, &ring->srq); + mlx4_en_free_rx_buf(priv, ring); + mlx4_en_destroy_allocator(priv, ring); +} + + +/* Unmap a completed descriptor and free unused pages */ +static int mlx4_en_complete_rx_desc(struct mlx4_en_priv *priv, + struct mlx4_en_rx_desc *rx_desc, + struct skb_frag_struct *skb_frags, + struct skb_frag_struct *skb_frags_rx, + struct mlx4_en_rx_alloc *page_alloc, + int length) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_frag_info *frag_info; + int nr; + dma_addr_t dma; + + /* Collect used fragments while replacing them in the HW descirptors */ + for (nr = 0; nr < priv->num_frags; nr++) { + frag_info = &priv->frag_info[nr]; + if (length <= frag_info->frag_prefix_size) + break; + + /* Save page reference in skb */ + skb_frags_rx[nr].page = skb_frags[nr].page; + skb_frags_rx[nr].size = skb_frags[nr].size; + skb_frags_rx[nr].page_offset = skb_frags[nr].page_offset; + dma = be64_to_cpu(rx_desc->data[nr].addr); + + /* Allocate a replacement page */ + if (mlx4_en_alloc_frag(priv, rx_desc, skb_frags, page_alloc, nr)) + goto fail; + + /* Unmap buffer */ + pci_unmap_single(mdev->pdev, dma, skb_frags[nr].size, + PCI_DMA_FROMDEVICE); + } + /* Adjust size of last fragment to match actual length */ + skb_frags_rx[nr - 1].size = length - + priv->frag_info[nr - 1].frag_prefix_size; + return nr; + +fail: + /* Drop all accumulated fragments (which have already been replaced in + * the descriptor) of this packet; remaining fragments are reused... */ + while (nr > 0) { + nr--; + put_page(skb_frags_rx[nr].page); + } + return 0; +} + + +static struct sk_buff *mlx4_en_rx_skb(struct mlx4_en_priv *priv, + struct mlx4_en_rx_desc *rx_desc, + struct skb_frag_struct *skb_frags, + struct mlx4_en_rx_alloc *page_alloc, + unsigned int length) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct sk_buff *skb; + void *va; + int used_frags; + dma_addr_t dma; + + skb = dev_alloc_skb(SMALL_PACKET_SIZE + NET_IP_ALIGN); + if (!skb) { + mlx4_dbg(RX_ERR, priv, "Failed allocating skb\n"); + return NULL; + } + skb->dev = priv->dev; + skb_reserve(skb, NET_IP_ALIGN); + skb->len = length; + skb->truesize = length + sizeof(struct sk_buff); + + /* Get pointer to first fragment so we could copy the headers into the + * (linear part of the) skb */ + va = page_address(skb_frags[0].page) + skb_frags[0].page_offset; + + if (length <= SMALL_PACKET_SIZE) { + /* We are copying all relevant data to the skb - temporarily + * synch buffers for the copy */ + dma = be64_to_cpu(rx_desc->data[0].addr); + dma_sync_single_range_for_cpu(&mdev->pdev->dev, dma, 0, + length, DMA_FROM_DEVICE); + skb_copy_to_linear_data(skb, va, length); + dma_sync_single_range_for_device(&mdev->pdev->dev, dma, 0, + length, DMA_FROM_DEVICE); + skb->tail += length; + } else { + + /* Move relevant fragments to skb */ + used_frags = mlx4_en_complete_rx_desc(priv, rx_desc, skb_frags, + skb_shinfo(skb)->frags, + page_alloc, length); + skb_shinfo(skb)->nr_frags = used_frags; + + /* Copy headers into the skb linear buffer */ + memcpy(skb->data, va, HEADER_COPY_SIZE); + skb->tail += HEADER_COPY_SIZE; + + /* Skip headers in first fragment */ + skb_shinfo(skb)->frags[0].page_offset += HEADER_COPY_SIZE; + + /* Adjust size of first fragment */ + skb_shinfo(skb)->frags[0].size -= HEADER_COPY_SIZE; + skb->data_len = length - HEADER_COPY_SIZE; + } + return skb; +} + +static void mlx4_en_copy_desc(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring, + int from, int to, int num) +{ + struct skb_frag_struct *skb_frags_from; + struct skb_frag_struct *skb_frags_to; + struct mlx4_en_rx_desc *rx_desc_from; + struct mlx4_en_rx_desc *rx_desc_to; + int from_index, to_index; + int nr, i; + + for (i = 0; i < num; i++) { + from_index = (from + i) & ring->size_mask; + to_index = (to + i) & ring->size_mask; + skb_frags_from = ring->rx_info + (from_index << priv->log_rx_info); + skb_frags_to = ring->rx_info + (to_index << priv->log_rx_info); + rx_desc_from = ring->buf + (from_index << ring->log_stride); + rx_desc_to = ring->buf + (to_index << ring->log_stride); + + for (nr = 0; nr < priv->num_frags; nr++) { + skb_frags_to[nr].page = skb_frags_from[nr].page; + skb_frags_to[nr].page_offset = skb_frags_from[nr].page_offset; + rx_desc_to->data[nr].addr = rx_desc_from->data[nr].addr; + } + } +} + + +int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_cqe *cqe; + struct mlx4_en_rx_ring *ring = &priv->rx_ring[cq->ring]; + struct skb_frag_struct *skb_frags; + struct skb_frag_struct lro_frags[MLX4_EN_MAX_RX_FRAGS]; + struct mlx4_en_rx_desc *rx_desc; + struct sk_buff *skb; + int index; + int nr; + unsigned int length; + int polled = 0; + int ip_summed; + + if (!priv->port_up) + return 0; + + /* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx + * descriptor offset can be deduced from the CQE index instead of + * reading 'cqe->index' */ + index = cq->mcq.cons_index & ring->size_mask; + cqe = &cq->buf[index]; + + /* Process all completed CQEs */ + while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK, + cq->mcq.cons_index & cq->size)) { + + skb_frags = ring->rx_info + (index << priv->log_rx_info); + rx_desc = ring->buf + (index << ring->log_stride); + + /* + * make sure we read the CQE after we read the ownership bit + */ + rmb(); + + /* Drop packet on bad receive or bad checksum */ + if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == + MLX4_CQE_OPCODE_ERROR)) { + mlx4_err(mdev, "CQE completed in error - vendor " + "syndrom:%d syndrom:%d\n", + ((struct mlx4_err_cqe *) cqe)->vendor_err_syndrome, + ((struct mlx4_err_cqe *) cqe)->syndrome); + goto next; + } + if (unlikely(cqe->badfcs_enc & MLX4_CQE_BAD_FCS)) { + mlx4_dbg(RX_ERR, priv, "Accepted frame with bad FCS\n"); + goto next; + } + + /* + * Packet is OK - process it. + */ + length = be32_to_cpu(cqe->byte_cnt); + ring->bytes += length; + ring->packets++; + + if (likely(priv->rx_csum)) { + if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) && + (cqe->checksum == cpu_to_be16(0xffff))) { + priv->port_stats.rx_chksum_good++; + /* This packet is eligible for LRO if it is: + * - DIX Ethernet (type interpretation) + * - TCP/IP (v4) + * - without IP options + * - not an IP fragment */ + if (mlx4_en_can_lro(cqe->status) && + dev->features & NETIF_F_LRO) { + + nr = mlx4_en_complete_rx_desc( + priv, rx_desc, + skb_frags, lro_frags, + ring->page_alloc, length); + if (!nr) + goto next; + + if (priv->vlgrp && (cqe->vlan_my_qpn & + cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK))) { + lro_vlan_hwaccel_receive_frags( + &ring->lro, lro_frags, + length, length, + priv->vlgrp, + be16_to_cpu(cqe->sl_vid), + NULL, 0); + } else + lro_receive_frags(&ring->lro, + lro_frags, + length, + length, + NULL, 0); + + goto next; + } + + /* LRO not possible, complete processing here */ + ip_summed = CHECKSUM_UNNECESSARY; + INC_PERF_COUNTER(priv->pstats.lro_misses); + } else { + ip_summed = CHECKSUM_NONE; + priv->port_stats.rx_chksum_none++; + } + } else { + ip_summed = CHECKSUM_NONE; + priv->port_stats.rx_chksum_none++; + } + + skb = mlx4_en_rx_skb(priv, rx_desc, skb_frags, + ring->page_alloc, length); + if (!skb) { + priv->stats.rx_dropped++; + goto next; + } + + skb->ip_summed = ip_summed; + skb->protocol = eth_type_trans(skb, dev); + + /* Push it up the stack */ + if (priv->vlgrp && (be32_to_cpu(cqe->vlan_my_qpn) & + MLX4_CQE_VLAN_PRESENT_MASK)) { + vlan_hwaccel_receive_skb(skb, priv->vlgrp, + be16_to_cpu(cqe->sl_vid)); + } else + netif_receive_skb(skb); + + dev->last_rx = jiffies; + +next: + ++cq->mcq.cons_index; + index = (cq->mcq.cons_index) & ring->size_mask; + cqe = &cq->buf[index]; + if (++polled == budget) { + /* We are here because we reached the NAPI budget - + * flush only pending LRO sessions */ + lro_flush_all(&ring->lro); + goto out; + } + } + + /* If CQ is empty flush all LRO sessions unconditionally */ + lro_flush_all(&ring->lro); + +out: + AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled); + mlx4_cq_set_ci(&cq->mcq); + wmb(); /* ensure HW sees CQ consumer before we post new buffers */ + ring->cons = cq->mcq.cons_index; + ring->prod += polled; /* Polled descriptors were realocated in place */ + if (unlikely(!ring->full)) { + mlx4_en_copy_desc(priv, ring, ring->cons - polled, + ring->prod - polled, polled); + mlx4_en_fill_rx_buf(dev, ring); + } + mlx4_en_update_rx_prod_db(ring); + return polled; +} + + +void mlx4_en_rx_irq(struct mlx4_cq *mcq) +{ + struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq); + struct mlx4_en_priv *priv = netdev_priv(cq->dev); + + if (priv->port_up) + netif_rx_schedule(cq->dev, &cq->napi); + else + mlx4_en_arm_cq(priv, cq); +} + +/* Rx CQ polling - called by NAPI */ +int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget) +{ + struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi); + struct net_device *dev = cq->dev; + struct mlx4_en_priv *priv = netdev_priv(dev); + int done; + + done = mlx4_en_process_rx_cq(dev, cq, budget); + + /* If we used up all the quota - we're probably not done yet... */ + if (done == budget) + INC_PERF_COUNTER(priv->pstats.napi_quota); + else { + /* Done for now */ + netif_rx_complete(dev, napi); + mlx4_en_arm_cq(priv, cq); + } + return done; +} + + +/* Calculate the last offset position that accomodates a full fragment + * (assuming fagment size = stride-align) */ +static int mlx4_en_last_alloc_offset(struct mlx4_en_priv *priv, u16 stride, u16 align) +{ + u16 res = MLX4_EN_ALLOC_SIZE % stride; + u16 offset = MLX4_EN_ALLOC_SIZE - stride - res + align; + + mlx4_dbg(DRV, priv, "Calculated last offset for stride:%d align:%d " + "res:%d offset:%d\n", stride, align, res, offset); + return offset; +} + + +static int frag_sizes[] = { + FRAG_SZ0, + FRAG_SZ1, + FRAG_SZ2, + FRAG_SZ3 +}; + +void mlx4_en_calc_rx_buf(struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int eff_mtu = dev->mtu + ETH_HLEN + VLAN_HLEN + ETH_LLC_SNAP_SIZE; + int buf_size = 0; + int i = 0; + + while (buf_size < eff_mtu) { + priv->frag_info[i].frag_size = + (eff_mtu > buf_size + frag_sizes[i]) ? + frag_sizes[i] : eff_mtu - buf_size; + priv->frag_info[i].frag_prefix_size = buf_size; + if (!i) { + priv->frag_info[i].frag_align = NET_IP_ALIGN; + priv->frag_info[i].frag_stride = + ALIGN(frag_sizes[i] + NET_IP_ALIGN, SMP_CACHE_BYTES); + } else { + priv->frag_info[i].frag_align = 0; + priv->frag_info[i].frag_stride = + ALIGN(frag_sizes[i], SMP_CACHE_BYTES); + } + priv->frag_info[i].last_offset = mlx4_en_last_alloc_offset( + priv, priv->frag_info[i].frag_stride, + priv->frag_info[i].frag_align); + buf_size += priv->frag_info[i].frag_size; + i++; + } + + priv->num_frags = i; + priv->rx_skb_size = eff_mtu; + priv->log_rx_info = ROUNDUP_LOG2(i * sizeof(struct skb_frag_struct)); + + mlx4_dbg(DRV, priv, "Rx buffer scatter-list (effective-mtu:%d " + "num_frags:%d):\n", eff_mtu, priv->num_frags); + for (i = 0; i < priv->num_frags; i++) { + mlx4_dbg(DRV, priv, " frag:%d - size:%d prefix:%d align:%d " + "stride:%d last_offset:%d\n", i, + priv->frag_info[i].frag_size, + priv->frag_info[i].frag_prefix_size, + priv->frag_info[i].frag_align, + priv->frag_info[i].frag_stride, + priv->frag_info[i].last_offset); + } +} + +/* RSS related functions */ + +/* Calculate rss size and map each entry in rss table to rx ring */ +void mlx4_en_set_default_rss_map(struct mlx4_en_priv *priv, + struct mlx4_en_rss_map *rss_map, + int num_entries, int num_rings) +{ + int i; + + rss_map->size = roundup_pow_of_two(num_entries); + mlx4_dbg(DRV, priv, "Setting default RSS map of %d entires\n", + rss_map->size); + + for (i = 0; i < rss_map->size; i++) { + rss_map->map[i] = i % num_rings; + mlx4_dbg(DRV, priv, "Entry %d ---> ring %d\n", i, rss_map->map[i]); + } +} + +static void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event) +{ + return; +} + + +static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, + int qpn, int srqn, int cqn, + enum mlx4_qp_state *state, + struct mlx4_qp *qp) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_qp_context *context; + int err = 0; + + context = kmalloc(sizeof *context , GFP_KERNEL); + if (!context) { + mlx4_err(mdev, "Failed to allocate qp context\n"); + return -ENOMEM; + } + + err = mlx4_qp_alloc(mdev->dev, qpn, qp); + if (err) { + mlx4_err(mdev, "Failed to allocate qp #%d\n", qpn); + goto out; + return err; + } + qp->event = mlx4_en_sqp_event; + + memset(context, 0, sizeof *context); + mlx4_en_fill_qp_context(priv, 0, 0, 0, 0, qpn, cqn, srqn, context); + + err = mlx4_qp_to_ready(mdev->dev, &priv->res.mtt, context, qp, state); + if (err) { + mlx4_qp_remove(mdev->dev, qp); + mlx4_qp_free(mdev->dev, qp); + } +out: + kfree(context); + return err; +} + +/* Allocate rx qp's and configure them according to rss map */ +int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_rss_map *rss_map = &priv->rss_map; + struct mlx4_qp_context context; + struct mlx4_en_rss_context *rss_context; + void *ptr; + int rss_xor = mdev->profile.rss_xor; + u8 rss_mask = mdev->profile.rss_mask; + int i, srqn, qpn, cqn; + int err = 0; + int good_qps = 0; + + mlx4_dbg(DRV, priv, "Configuring rss steering for port %u\n", priv->port); + err = mlx4_qp_reserve_range(mdev->dev, rss_map->size, + rss_map->size, &rss_map->base_qpn); + if (err) { + mlx4_err(mdev, "Failed reserving %d qps for port %u\n", + rss_map->size, priv->port); + return err; + } + + for (i = 0; i < rss_map->size; i++) { + cqn = priv->rx_ring[rss_map->map[i]].cqn; + srqn = priv->rx_ring[rss_map->map[i]].srq.srqn; + qpn = rss_map->base_qpn + i; + err = mlx4_en_config_rss_qp(priv, qpn, srqn, cqn, + &rss_map->state[i], + &rss_map->qps[i]); + if (err) + goto rss_err; + + ++good_qps; + } + + /* Configure RSS indirection qp */ + err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &priv->base_qpn); + if (err) { + mlx4_err(mdev, "Failed to reserve range for RSS " + "indirection qp\n"); + goto rss_err; + } + err = mlx4_qp_alloc(mdev->dev, priv->base_qpn, &rss_map->indir_qp); + if (err) { + mlx4_err(mdev, "Failed to allocate RSS indirection QP\n"); + goto reserve_err; + } + rss_map->indir_qp.event = mlx4_en_sqp_event; + mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn, + priv->rx_ring[0].cqn, 0, &context); + + ptr = ((void *) &context) + 0x3c; + rss_context = (struct mlx4_en_rss_context *) ptr; + rss_context->base_qpn = cpu_to_be32(ilog2(rss_map->size) << 24 | + (rss_map->base_qpn)); + rss_context->default_qpn = cpu_to_be32(rss_map->base_qpn); + rss_context->hash_fn = rss_xor & 0x3; + rss_context->flags = rss_mask << 2; + + err = mlx4_qp_to_ready(mdev->dev, &priv->res.mtt, &context, + &rss_map->indir_qp, &rss_map->indir_state); + if (err) + goto indir_err; + + return 0; + +indir_err: + mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state, + MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp); + mlx4_qp_remove(mdev->dev, &rss_map->indir_qp); + mlx4_qp_free(mdev->dev, &rss_map->indir_qp); +reserve_err: + mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1); +rss_err: + for (i = 0; i < good_qps; i++) { + mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i], + MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]); + mlx4_qp_remove(mdev->dev, &rss_map->qps[i]); + mlx4_qp_free(mdev->dev, &rss_map->qps[i]); + } + mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, rss_map->size); + return err; +} + +void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_rss_map *rss_map = &priv->rss_map; + int i; + + mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state, + MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp); + mlx4_qp_remove(mdev->dev, &rss_map->indir_qp); + mlx4_qp_free(mdev->dev, &rss_map->indir_qp); + mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1); + + for (i = 0; i < rss_map->size; i++) { + mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i], + MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]); + mlx4_qp_remove(mdev->dev, &rss_map->qps[i]); + mlx4_qp_free(mdev->dev, &rss_map->qps[i]); + } + mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, rss_map->size); +} + + + + + diff --git a/drivers/net/mlx4/en_tx.c b/drivers/net/mlx4/en_tx.c new file mode 100644 index 0000000..8592f8f --- /dev/null +++ b/drivers/net/mlx4/en_tx.c @@ -0,0 +1,820 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <asm/page.h> +#include <linux/mlx4/cq.h> +#include <linux/mlx4/qp.h> +#include <linux/skbuff.h> +#include <linux/if_vlan.h> +#include <linux/vmalloc.h> + +#include "mlx4_en.h" + +enum { + MAX_INLINE = 104, /* 128 - 16 - 4 - 4 */ +}; + +static int inline_thold __read_mostly = MAX_INLINE; + +module_param_named(inline_thold, inline_thold, int, 0444); +MODULE_PARM_DESC(inline_thold, "treshold for using inline data"); + +int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring, u32 size, + u16 stride) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int tmp; + int err; + + ring->size = size; + ring->size_mask = size - 1; + ring->stride = stride; + + inline_thold = min(inline_thold, MAX_INLINE); + + spin_lock_init(&ring->comp_lock); + + tmp = size * sizeof(struct mlx4_en_tx_info); + ring->tx_info = vmalloc(tmp); + if (!ring->tx_info) { + mlx4_err(mdev, "Failed allocating tx_info ring\n"); + return -ENOMEM; + } + mlx4_dbg(DRV, priv, "Allocated tx_info ring at addr:%p size:%d\n", + ring->tx_info, tmp); + + ring->bounce_buf = kmalloc(MAX_DESC_SIZE, GFP_KERNEL); + if (!ring->bounce_buf) { + mlx4_err(mdev, "Failed allocating bounce buffer\n"); + err = -ENOMEM; + goto err_tx; + } + ring->buf_size = ALIGN(size * ring->stride, MLX4_EN_PAGE_SIZE); + + err = mlx4_alloc_hwq_res(mdev->dev, &ring->wqres, ring->buf_size, + 2 * PAGE_SIZE); + if (err) { + mlx4_err(mdev, "Failed allocating hwq resources\n"); + goto err_bounce; + } + + err = mlx4_en_map_buffer(&ring->wqres.buf); + if (err) { + mlx4_err(mdev, "Failed to map TX buffer\n"); + goto err_hwq_res; + } + + ring->buf = ring->wqres.buf.direct.buf; + + mlx4_dbg(DRV, priv, "Allocated TX ring (addr:%p) - buf:%p size:%d " + "buf_size:%d dma:%llx\n", ring, ring->buf, ring->size, + ring->buf_size, (unsigned long long) ring->wqres.buf.direct.map); + + err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &ring->qpn); + if (err) { + mlx4_err(mdev, "Failed reserving qp for tx ring.\n"); + goto err_map; + } + + err = mlx4_qp_alloc(mdev->dev, ring->qpn, &ring->qp); + if (err) { + mlx4_err(mdev, "Failed allocating qp %d\n", ring->qpn); + goto err_reserve; + } + + return 0; + +err_reserve: + mlx4_qp_release_range(mdev->dev, ring->qpn, 1); +err_map: + mlx4_en_unmap_buffer(&ring->wqres.buf); +err_hwq_res: + mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size); +err_bounce: + kfree(ring->bounce_buf); + ring->bounce_buf = NULL; +err_tx: + vfree(ring->tx_info); + ring->tx_info = NULL; + return err; +} + +void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring) +{ + struct mlx4_en_dev *mdev = priv->mdev; + mlx4_dbg(DRV, priv, "Destroying tx ring, qpn: %d\n", ring->qpn); + + mlx4_qp_remove(mdev->dev, &ring->qp); + mlx4_qp_free(mdev->dev, &ring->qp); + mlx4_qp_release_range(mdev->dev, ring->qpn, 1); + mlx4_en_unmap_buffer(&ring->wqres.buf); + mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size); + kfree(ring->bounce_buf); + ring->bounce_buf = NULL; + vfree(ring->tx_info); + ring->tx_info = NULL; +} + +int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring, + int cq, int srqn) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int err; + + ring->cqn = cq; + ring->prod = 0; + ring->cons = 0xffffffff; + ring->last_nr_txbb = 1; + ring->poll_cnt = 0; + ring->blocked = 0; + memset(ring->tx_info, 0, ring->size * sizeof(struct mlx4_en_tx_info)); + memset(ring->buf, 0, ring->buf_size); + + ring->qp_state = MLX4_QP_STATE_RST; + ring->doorbell_qpn = swab32(ring->qp.qpn << 8); + + mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 1, 0, ring->qpn, + ring->cqn, srqn, &ring->context); + + err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, &ring->context, + &ring->qp, &ring->qp_state); + + return err; +} + +void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring) +{ + struct mlx4_en_dev *mdev = priv->mdev; + + mlx4_qp_modify(mdev->dev, NULL, ring->qp_state, + MLX4_QP_STATE_RST, NULL, 0, 0, &ring->qp); +} + + +static u32 mlx4_en_free_tx_desc(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring, + int index, u8 owner) +{ + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_tx_info *tx_info = &ring->tx_info[index]; + struct mlx4_en_tx_desc *tx_desc = ring->buf + index * TXBB_SIZE; + struct mlx4_wqe_data_seg *data = (void *) tx_desc + tx_info->data_offset; + struct sk_buff *skb = tx_info->skb; + struct skb_frag_struct *frag; + void *end = ring->buf + ring->buf_size; + int frags = skb_shinfo(skb)->nr_frags; + int i; + __be32 *ptr = (__be32 *)tx_desc; + __be32 stamp = cpu_to_be32(STAMP_VAL | (!!owner << STAMP_SHIFT)); + + /* Optimize the common case when there are no wraparounds */ + if (likely((void *) tx_desc + tx_info->nr_txbb * TXBB_SIZE <= end)) { + if (tx_info->linear) { + pci_unmap_single(mdev->pdev, + (dma_addr_t) be64_to_cpu(data->addr), + be32_to_cpu(data->byte_count), + PCI_DMA_TODEVICE); + ++data; + } + + for (i = 0; i < frags; i++) { + frag = &skb_shinfo(skb)->frags[i]; + pci_unmap_page(mdev->pdev, + (dma_addr_t) be64_to_cpu(data[i].addr), + frag->size, PCI_DMA_TODEVICE); + } + /* Stamp the freed descriptor */ + for (i = 0; i < tx_info->nr_txbb * TXBB_SIZE; i += STAMP_STRIDE) { + *ptr = stamp; + ptr += STAMP_DWORDS; + } + + } else { + if ((void *) data >= end) { + data = (struct mlx4_wqe_data_seg *) + (ring->buf + ((void *) data - end)); + } + + if (tx_info->linear) { + pci_unmap_single(mdev->pdev, + (dma_addr_t) be64_to_cpu(data->addr), + be32_to_cpu(data->byte_count), + PCI_DMA_TODEVICE); + ++data; + } + + for (i = 0; i < frags; i++) { + /* Check for wraparound before unmapping */ + if ((void *) data >= end) + data = (struct mlx4_wqe_data_seg *) ring->buf; + frag = &skb_shinfo(skb)->frags[i]; + pci_unmap_page(mdev->pdev, + (dma_addr_t) be64_to_cpu(data->addr), + frag->size, PCI_DMA_TODEVICE); + } + /* Stamp the freed descriptor */ + for (i = 0; i < tx_info->nr_txbb * TXBB_SIZE; i += STAMP_STRIDE) { + *ptr = stamp; + ptr += STAMP_DWORDS; + if ((void *) ptr >= end) { + ptr = ring->buf; + stamp ^= cpu_to_be32(0x80000000); + } + } + + } + dev_kfree_skb_any(skb); + return tx_info->nr_txbb; +} + + +int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + int cnt = 0; + + /* Skip last polled descriptor */ + ring->cons += ring->last_nr_txbb; + mlx4_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n", + ring->cons, ring->prod); + + if ((u32) (ring->prod - ring->cons) > ring->size) { + if (netif_msg_tx_err(priv)) + mlx4_warn(priv->mdev, "Tx consumer passed producer!\n"); + return 0; + } + + while (ring->cons != ring->prod) { + ring->last_nr_txbb = mlx4_en_free_tx_desc(priv, ring, + ring->cons & ring->size_mask, + !!(ring->cons & ring->size)); + ring->cons += ring->last_nr_txbb; + cnt++; + } + + if (cnt) + mlx4_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt); + + return cnt; +} + +void mlx4_en_set_prio_map(struct mlx4_en_priv *priv, u16 *prio_map, u32 ring_num) +{ + int block = 8 / ring_num; + int extra = 8 - (block * ring_num); + int num = 0; + u16 ring = 1; + int prio; + + if (ring_num == 1) { + for (prio = 0; prio < 8; prio++) + prio_map[prio] = 0; + return; + } + + for (prio = 0; prio < 8; prio++) { + if (extra && (num == block + 1)) { + ring++; + num = 0; + extra--; + } else if (!extra && (num == block)) { + ring++; + num = 0; + } + prio_map[prio] = ring; + mlx4_dbg(DRV, priv, " prio:%d --> ring:%d\n", prio, ring); + num++; + } +} + +static void mlx4_en_process_tx_cq(struct net_device *dev, struct mlx4_en_cq *cq) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_cq *mcq = &cq->mcq; + struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring]; + struct mlx4_cqe *cqe = cq->buf; + u16 index; + u16 new_index; + u32 txbbs_skipped = 0; + u32 cq_last_sav; + + /* index always points to the first TXBB of the last polled descriptor */ + index = ring->cons & ring->size_mask; + new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask; + if (index == new_index) + return; + + if (!priv->port_up) + return; + + /* + * We use a two-stage loop: + * - the first samples the HW-updated CQE + * - the second frees TXBBs until the last sample + * This lets us amortize CQE cache misses, while still polling the CQ + * until is quiescent. + */ + cq_last_sav = mcq->cons_index; + do { + do { + /* Skip over last polled CQE */ + index = (index + ring->last_nr_txbb) & ring->size_mask; + txbbs_skipped += ring->last_nr_txbb; + + /* Poll next CQE */ + ring->last_nr_txbb = mlx4_en_free_tx_desc( + priv, ring, index, + !!((ring->cons + txbbs_skipped) & + ring->size)); + ++mcq->cons_index; + + } while (index != new_index); + + new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask; + } while (index != new_index); + AVG_PERF_COUNTER(priv->pstats.tx_coal_avg, + (u32) (mcq->cons_index - cq_last_sav)); + + /* + * To prevent CQ overflow we first update CQ consumer and only then + * the ring consumer. + */ + mlx4_cq_set_ci(mcq); + wmb(); + ring->cons += txbbs_skipped; + + /* Wakeup Tx queue if this ring stopped it */ + if (unlikely(ring->blocked)) { + if (((u32) (ring->prod - ring->cons) <= + ring->size - HEADROOM - MAX_DESC_TXBBS) && !cq->armed) { + + /* TODO: support multiqueue netdevs. Currently, we block + * when *any* ring is full. Note that: + * - 2 Tx rings can unblock at the same time and call + * netif_wake_queue(), which is OK since this + * operation is idempotent. + * - We might wake the queue just after another ring + * stopped it. This is no big deal because the next + * transmission on that ring would stop the queue. + */ + ring->blocked = 0; + netif_wake_queue(dev); + priv->port_stats.wake_queue++; + } + } +} + +void mlx4_en_tx_irq(struct mlx4_cq *mcq) +{ + struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq); + struct mlx4_en_priv *priv = netdev_priv(cq->dev); + struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring]; + + spin_lock_irq(&ring->comp_lock); + cq->armed = 0; + mlx4_en_process_tx_cq(cq->dev, cq); + if (ring->blocked) + mlx4_en_arm_cq(priv, cq); + else + mod_timer(&cq->timer, jiffies + 1); + spin_unlock_irq(&ring->comp_lock); +} + + +void mlx4_en_poll_tx_cq(unsigned long data) +{ + struct mlx4_en_cq *cq = (struct mlx4_en_cq *) data; + struct mlx4_en_priv *priv = netdev_priv(cq->dev); + struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring]; + u32 inflight; + + INC_PERF_COUNTER(priv->pstats.tx_poll); + + netif_tx_lock(priv->dev); + spin_lock_irq(&ring->comp_lock); + mlx4_en_process_tx_cq(cq->dev, cq); + inflight = (u32) (ring->prod - ring->cons - ring->last_nr_txbb); + + /* If there are still packets in flight and the timer has not already + * been scheduled by the Tx routine then schedule it here to guarantee + * completion processing of these packets */ + if (inflight && priv->port_up) + mod_timer(&cq->timer, jiffies + MLX4_EN_TX_POLL_TIMEOUT); + + spin_unlock_irq(&ring->comp_lock); + netif_tx_unlock(priv->dev); +} + +static struct mlx4_en_tx_desc *mlx4_en_bounce_to_desc(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring, + u32 index, + unsigned int desc_size) +{ + u32 copy = (ring->size - index) * TXBB_SIZE; + int i; + + for (i = desc_size - copy - 4; i >= 0; i -= 4) { + if ((i & (TXBB_SIZE - 1)) == 0) + wmb(); + + *((u32 *) (ring->buf + i)) = + *((u32 *) (ring->bounce_buf + copy + i)); + } + + for (i = copy - 4; i >= 4 ; i -= 4) { + if ((i & (TXBB_SIZE - 1)) == 0) + wmb(); + + *((u32 *) (ring->buf + index * TXBB_SIZE + i)) = + *((u32 *) (ring->bounce_buf + i)); + } + + /* Return real descriptor location */ + return ring->buf + index * TXBB_SIZE; +} + +static inline void mlx4_en_xmit_poll(struct mlx4_en_priv *priv, int tx_ind) +{ + struct mlx4_en_cq *cq = &priv->tx_cq[tx_ind]; + struct mlx4_en_tx_ring *ring = &priv->tx_ring[tx_ind]; + + /* If we don't have a pending timer, set one up to catch our recent + post in case the interface becomes idle */ + if (!timer_pending(&cq->timer)) + mod_timer(&cq->timer, jiffies + MLX4_EN_TX_POLL_TIMEOUT); + + /* Poll the CQ every mlx4_en_TX_MODER_POLL packets */ + if ((++ring->poll_cnt & (MLX4_EN_TX_POLL_MODER - 1)) == 0) + mlx4_en_process_tx_cq(priv->dev, cq); +} + +static void *get_frag_ptr(struct sk_buff *skb) +{ + struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0]; + struct page *page = frag->page; + void *ptr; + + ptr = page_address(page); + if (unlikely(!ptr)) + return NULL; + + return ptr + frag->page_offset; +} + +static int is_inline(struct sk_buff *skb, void **pfrag) +{ + void *ptr; + + if (inline_thold && !skb_is_gso(skb) && skb->len <= inline_thold) { + if (skb_shinfo(skb)->nr_frags == 1) { + ptr = get_frag_ptr(skb); + if (unlikely(!ptr)) + return 0; + + if (pfrag) + *pfrag = ptr; + + return 1; + } else if (unlikely(skb_shinfo(skb)->nr_frags)) + return 0; + else + return 1; + } + + return 0; +} + +static int inline_size(struct sk_buff *skb) +{ + if (skb->len + CTRL_SIZE + sizeof(struct mlx4_wqe_inline_seg) + <= MLX4_INLINE_ALIGN) + return ALIGN(skb->len + CTRL_SIZE + + sizeof(struct mlx4_wqe_inline_seg), 16); + else + return ALIGN(skb->len + CTRL_SIZE + 2 * + sizeof(struct mlx4_wqe_inline_seg), 16); +} + +static int get_real_size(struct sk_buff *skb, struct net_device *dev, + int *lso_header_size) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + int real_size; + + if (skb_is_gso(skb)) { + *lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb); + real_size = CTRL_SIZE + skb_shinfo(skb)->nr_frags * DS_SIZE + + ALIGN(*lso_header_size + 4, DS_SIZE); + if (unlikely(*lso_header_size != skb_headlen(skb))) { + /* We add a segment for the skb linear buffer only if + * it contains data */ + if (*lso_header_size < skb_headlen(skb)) + real_size += DS_SIZE; + else { + if (netif_msg_tx_err(priv)) + mlx4_warn(mdev, "Non-linear headers\n"); + dev_kfree_skb_any(skb); + return 0; + } + } + if (unlikely(*lso_header_size > MAX_LSO_HDR_SIZE)) { + if (netif_msg_tx_err(priv)) + mlx4_warn(mdev, "LSO header size too big\n"); + dev_kfree_skb_any(skb); + return 0; + } + } else { + *lso_header_size = 0; + if (!is_inline(skb, NULL)) + real_size = CTRL_SIZE + (skb_shinfo(skb)->nr_frags + 1) * DS_SIZE; + else + real_size = inline_size(skb); + } + + return real_size; +} + +static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc, struct sk_buff *skb, + int real_size, u16 *vlan_tag, int tx_ind, void *fragptr) +{ + struct mlx4_wqe_inline_seg *inl = &tx_desc->inl; + int spc = MLX4_INLINE_ALIGN - CTRL_SIZE - sizeof *inl; + + if (skb->len <= spc) { + inl->byte_count = cpu_to_be32(1 << 31 | skb->len); + skb_copy_from_linear_data(skb, inl + 1, skb_headlen(skb)); + if (skb_shinfo(skb)->nr_frags) + memcpy(((void *)(inl + 1)) + skb_headlen(skb), fragptr, + skb_shinfo(skb)->frags[0].size); + + } else { + inl->byte_count = cpu_to_be32(1 << 31 | spc); + if (skb_headlen(skb) <= spc) { + skb_copy_from_linear_data(skb, inl + 1, skb_headlen(skb)); + if (skb_headlen(skb) < spc) { + memcpy(((void *)(inl + 1)) + skb_headlen(skb), + fragptr, spc - skb_headlen(skb)); + fragptr += spc - skb_headlen(skb); + } + inl = (void *) (inl + 1) + spc; + memcpy(((void *)(inl + 1)), fragptr, skb->len - spc); + } else { + skb_copy_from_linear_data(skb, inl + 1, spc); + inl = (void *) (inl + 1) + spc; + skb_copy_from_linear_data_offset(skb, spc, inl + 1, + skb_headlen(skb) - spc); + if (skb_shinfo(skb)->nr_frags) + memcpy(((void *)(inl + 1)) + skb_headlen(skb) - spc, + fragptr, skb_shinfo(skb)->frags[0].size); + } + + wmb(); + inl->byte_count = cpu_to_be32(1 << 31 | (skb->len - spc)); + } + tx_desc->ctrl.vlan_tag = cpu_to_be16(*vlan_tag); + tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN * !!(*vlan_tag); + tx_desc->ctrl.fence_size = (real_size / 16) & 0x3f; +} + +static int get_vlan_info(struct mlx4_en_priv *priv, struct sk_buff *skb, + u16 *vlan_tag) +{ + int tx_ind; + + /* Obtain VLAN information if present */ + if (priv->vlgrp && vlan_tx_tag_present(skb)) { + *vlan_tag = vlan_tx_tag_get(skb); + /* Set the Tx ring to use according to vlan priority */ + tx_ind = priv->tx_prio_map[*vlan_tag >> 13]; + } else { + *vlan_tag = 0; + tx_ind = 0; + } + return tx_ind; +} + +int mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev) +{ + struct mlx4_en_priv *priv = netdev_priv(dev); + struct mlx4_en_dev *mdev = priv->mdev; + struct mlx4_en_tx_ring *ring; + struct mlx4_en_cq *cq; + struct mlx4_en_tx_desc *tx_desc; + struct mlx4_wqe_data_seg *data; + struct skb_frag_struct *frag; + struct mlx4_en_tx_info *tx_info; + int tx_ind = 0; + int nr_txbb; + int desc_size; + int real_size; + dma_addr_t dma; + u32 index; + __be32 op_own; + u16 vlan_tag; + int i; + int lso_header_size; + void *fragptr; + + if (unlikely(!skb->len)) { + dev_kfree_skb_any(skb); + return NETDEV_TX_OK; + } + real_size = get_real_size(skb, dev, &lso_header_size); + if (unlikely(!real_size)) + return NETDEV_TX_OK; + + /* Allign descriptor to TXBB size */ + desc_size = ALIGN(real_size, TXBB_SIZE); + nr_txbb = desc_size / TXBB_SIZE; + if (unlikely(nr_txbb > MAX_DESC_TXBBS)) { + if (netif_msg_tx_err(priv)) + mlx4_warn(mdev, "Oversized header or SG list\n"); + dev_kfree_skb_any(skb); + return NETDEV_TX_OK; + } + + tx_ind = get_vlan_info(priv, skb, &vlan_tag); + ring = &priv->tx_ring[tx_ind]; + + /* Check available TXBBs And 2K spare for prefetch */ + if (unlikely(((int)(ring->prod - ring->cons)) > + ring->size - HEADROOM - MAX_DESC_TXBBS)) { + /* every full Tx ring stops queue. + * TODO: implement multi-queue support (per-queue stop) */ + netif_stop_queue(dev); + ring->blocked = 1; + priv->port_stats.queue_stopped++; + + /* Use interrupts to find out when queue opened */ + cq = &priv->tx_cq[tx_ind]; + mlx4_en_arm_cq(priv, cq); + return NETDEV_TX_BUSY; + } + + /* Now that we know what Tx ring to use */ + if (unlikely(!priv->port_up)) { + if (netif_msg_tx_err(priv)) + mlx4_warn(mdev, "xmit: port down!\n"); + dev_kfree_skb_any(skb); + return NETDEV_TX_OK; + } + + /* Track current inflight packets for performance analysis */ + AVG_PERF_COUNTER(priv->pstats.inflight_avg, + (u32) (ring->prod - ring->cons - 1)); + + /* Packet is good - grab an index and transmit it */ + index = ring->prod & ring->size_mask; + + /* See if we have enough space for whole descriptor TXBB for setting + * SW ownership on next descriptor; if not, use a bounce buffer. */ + if (likely(index + nr_txbb <= ring->size)) + tx_desc = ring->buf + index * TXBB_SIZE; + else + tx_desc = (struct mlx4_en_tx_desc *) ring->bounce_buf; + + /* Save skb in tx_info ring */ + tx_info = &ring->tx_info[index]; + tx_info->skb = skb; + tx_info->nr_txbb = nr_txbb; + + /* Prepare ctrl segement apart opcode+ownership, which depends on + * whether LSO is used */ + tx_desc->ctrl.vlan_tag = cpu_to_be16(vlan_tag); + tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN * !!vlan_tag; + tx_desc->ctrl.fence_size = (real_size / 16) & 0x3f; + tx_desc->ctrl.srcrb_flags = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE | + MLX4_WQE_CTRL_SOLICITED); + if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) { + tx_desc->ctrl.srcrb_flags |= cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM | + MLX4_WQE_CTRL_TCP_UDP_CSUM); + priv->port_stats.tx_chksum_offload++; + } + + /* Handle LSO (TSO) packets */ + if (lso_header_size) { + /* Mark opcode as LSO */ + op_own = cpu_to_be32(MLX4_OPCODE_LSO | (1 << 6)) | + ((ring->prod & ring->size) ? + cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0); + + /* Fill in the LSO prefix */ + tx_desc->lso.mss_hdr_size = cpu_to_be32( + skb_shinfo(skb)->gso_size << 16 | lso_header_size); + + /* Copy headers; + * note that we already verified that it is linear */ + memcpy(tx_desc->lso.header, skb->data, lso_header_size); + data = ((void *) &tx_desc->lso + + ALIGN(lso_header_size + 4, DS_SIZE)); + + priv->port_stats.tso_packets++; + i = ((skb->len - lso_header_size) / skb_shinfo(skb)->gso_size) + + !!((skb->len - lso_header_size) % skb_shinfo(skb)->gso_size); + ring->bytes += skb->len + (i - 1) * lso_header_size; + ring->packets += i; + } else { + /* Normal (Non LSO) packet */ + op_own = cpu_to_be32(MLX4_OPCODE_SEND) | + ((ring->prod & ring->size) ? + cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0); + data = &tx_desc->data; + ring->bytes += max(skb->len, (unsigned int) ETH_ZLEN); + ring->packets++; + + } + AVG_PERF_COUNTER(priv->pstats.tx_pktsz_avg, skb->len); + + + /* valid only for none inline segments */ + tx_info->data_offset = (void *) data - (void *) tx_desc; + + tx_info->linear = (lso_header_size < skb_headlen(skb) && !is_inline(skb, NULL)) ? 1 : 0; + data += skb_shinfo(skb)->nr_frags + tx_info->linear - 1; + + if (!is_inline(skb, &fragptr)) { + /* Map fragments */ + for (i = skb_shinfo(skb)->nr_frags - 1; i >= 0; i--) { + frag = &skb_shinfo(skb)->frags[i]; + dma = pci_map_page(mdev->dev->pdev, frag->page, frag->page_offset, + frag->size, PCI_DMA_TODEVICE); + data->addr = cpu_to_be64(dma); + data->lkey = cpu_to_be32(mdev->mr.key); + wmb(); + data->byte_count = cpu_to_be32(frag->size); + --data; + } + + /* Map linear part */ + if (tx_info->linear) { + dma = pci_map_single(mdev->dev->pdev, skb->data + lso_header_size, + skb_headlen(skb) - lso_header_size, PCI_DMA_TODEVICE); + data->addr = cpu_to_be64(dma); + data->lkey = cpu_to_be32(mdev->mr.key); + wmb(); + data->byte_count = cpu_to_be32(skb_headlen(skb) - lso_header_size); + } + } else + build_inline_wqe(tx_desc, skb, real_size, &vlan_tag, tx_ind, fragptr); + + ring->prod += nr_txbb; + + /* If we used a bounce buffer then copy descriptor back into place */ + if (tx_desc == (struct mlx4_en_tx_desc *) ring->bounce_buf) + tx_desc = mlx4_en_bounce_to_desc(priv, ring, index, desc_size); + + /* Run destructor before passing skb to HW */ + if (likely(!skb_shared(skb))) + skb_orphan(skb); + + /* Ensure new descirptor hits memory + * before setting ownership of this descriptor to HW */ + wmb(); + tx_desc->ctrl.owner_opcode = op_own; + + /* Ring doorbell! */ + wmb(); + writel(ring->doorbell_qpn, mdev->uar_map + MLX4_SEND_DOORBELL); + dev->trans_start = jiffies; + + /* Poll CQ here */ + mlx4_en_xmit_poll(priv, tx_ind); + + return 0; +} + diff --git a/drivers/net/mlx4/eq.c b/drivers/net/mlx4/eq.c index 8a8b561..de16933 100644 --- a/drivers/net/mlx4/eq.c +++ b/drivers/net/mlx4/eq.c @@ -558,7 +558,7 @@ int mlx4_init_eq_table(struct mlx4_dev *dev) int i; err = mlx4_bitmap_init(&priv->eq_table.bitmap, dev->caps.num_eqs, - dev->caps.num_eqs - 1, dev->caps.reserved_eqs); + dev->caps.num_eqs - 1, dev->caps.reserved_eqs, 0); if (err) return err; diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c index 7e32955..be09fdb 100644 --- a/drivers/net/mlx4/fw.c +++ b/drivers/net/mlx4/fw.c @@ -88,6 +88,7 @@ static void dump_dev_cap_flags(struct mlx4_dev *dev, u32 flags) [ 8] = "P_Key violation counter", [ 9] = "Q_Key violation counter", [10] = "VMM", + [12] = "DPDP", [16] = "MW support", [17] = "APM support", [18] = "Atomic ops support", @@ -346,7 +347,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) MLX4_GET(field, outbox, QUERY_DEV_CAP_VL_PORT_OFFSET); dev_cap->max_vl[i] = field >> 4; MLX4_GET(field, outbox, QUERY_DEV_CAP_MTU_WIDTH_OFFSET); - dev_cap->max_mtu[i] = field >> 4; + dev_cap->ib_mtu[i] = field >> 4; dev_cap->max_port_width[i] = field & 0xf; MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GID_OFFSET); dev_cap->max_gids[i] = 1 << (field & 0xf); @@ -354,9 +355,13 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) dev_cap->max_pkeys[i] = 1 << (field & 0xf); } } else { +#define QUERY_PORT_SUPPORTED_TYPE_OFFSET 0x00 #define QUERY_PORT_MTU_OFFSET 0x01 +#define QUERY_PORT_ETH_MTU_OFFSET 0x02 #define QUERY_PORT_WIDTH_OFFSET 0x06 #define QUERY_PORT_MAX_GID_PKEY_OFFSET 0x07 +#define QUERY_PORT_MAC_OFFSET 0x08 +#define QUERY_PORT_MAX_MACVLAN_OFFSET 0x0a #define QUERY_PORT_MAX_VL_OFFSET 0x0b for (i = 1; i <= dev_cap->num_ports; ++i) { @@ -365,8 +370,10 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) if (err) goto out; + MLX4_GET(field, outbox, QUERY_PORT_SUPPORTED_TYPE_OFFSET); + dev_cap->supported_port_types[i] = field & 3; MLX4_GET(field, outbox, QUERY_PORT_MTU_OFFSET); - dev_cap->max_mtu[i] = field & 0xf; + dev_cap->ib_mtu[i] = field & 0xf; MLX4_GET(field, outbox, QUERY_PORT_WIDTH_OFFSET); dev_cap->max_port_width[i] = field & 0xf; MLX4_GET(field, outbox, QUERY_PORT_MAX_GID_PKEY_OFFSET); @@ -374,6 +381,11 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) dev_cap->max_pkeys[i] = 1 << (field & 0xf); MLX4_GET(field, outbox, QUERY_PORT_MAX_VL_OFFSET); dev_cap->max_vl[i] = field & 0xf; + MLX4_GET(field, outbox, QUERY_PORT_MAX_MACVLAN_OFFSET); + dev_cap->log_max_macs[i] = field & 0xf; + dev_cap->log_max_vlans[i] = field >> 4; + MLX4_GET(dev_cap->eth_mtu[i], outbox, QUERY_PORT_ETH_MTU_OFFSET); + MLX4_GET(dev_cap->def_mac[i], outbox, QUERY_PORT_MAC_OFFSET); } } @@ -407,7 +419,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) mlx4_dbg(dev, "Max CQEs: %d, max WQEs: %d, max SRQ WQEs: %d\n", dev_cap->max_cq_sz, dev_cap->max_qp_sz, dev_cap->max_srq_sz); mlx4_dbg(dev, "Local CA ACK delay: %d, max MTU: %d, port width cap: %d\n", - dev_cap->local_ca_ack_delay, 128 << dev_cap->max_mtu[1], + dev_cap->local_ca_ack_delay, 128 << dev_cap->ib_mtu[1], dev_cap->max_port_width[1]); mlx4_dbg(dev, "Max SQ desc size: %d, max SQ S/G: %d\n", dev_cap->max_sq_desc_sz, dev_cap->max_sq_sg); @@ -819,7 +831,7 @@ int mlx4_INIT_PORT(struct mlx4_dev *dev, int port) flags |= (dev->caps.port_width_cap[port] & 0xf) << INIT_PORT_PORT_WIDTH_SHIFT; MLX4_PUT(inbox, flags, INIT_PORT_FLAGS_OFFSET); - field = 128 << dev->caps.mtu_cap[port]; + field = 128 << dev->caps.ib_mtu_cap[port]; MLX4_PUT(inbox, field, INIT_PORT_MTU_OFFSET); field = dev->caps.gid_table_len[port]; MLX4_PUT(inbox, field, INIT_PORT_MAX_GID_OFFSET); diff --git a/drivers/net/mlx4/fw.h b/drivers/net/mlx4/fw.h index decbb5c..526d7f3 100644 --- a/drivers/net/mlx4/fw.h +++ b/drivers/net/mlx4/fw.h @@ -66,11 +66,13 @@ struct mlx4_dev_cap { int local_ca_ack_delay; int num_ports; u32 max_msg_sz; - int max_mtu[MLX4_MAX_PORTS + 1]; + int ib_mtu[MLX4_MAX_PORTS + 1]; int max_port_width[MLX4_MAX_PORTS + 1]; int max_vl[MLX4_MAX_PORTS + 1]; int max_gids[MLX4_MAX_PORTS + 1]; int max_pkeys[MLX4_MAX_PORTS + 1]; + u64 def_mac[MLX4_MAX_PORTS + 1]; + u16 eth_mtu[MLX4_MAX_PORTS + 1]; u16 stat_rate_support; u32 flags; int reserved_uars; @@ -102,6 +104,9 @@ struct mlx4_dev_cap { u32 reserved_lkey; u64 max_icm_sz; int max_gso_sz; + u8 supported_port_types[MLX4_MAX_PORTS + 1]; + u8 log_max_macs[MLX4_MAX_PORTS + 1]; + u8 log_max_vlans[MLX4_MAX_PORTS + 1]; }; struct mlx4_adapter { diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c index 1252a919..468921b8 100644 --- a/drivers/net/mlx4/main.c +++ b/drivers/net/mlx4/main.c @@ -85,6 +85,57 @@ static struct mlx4_profile default_profile = { .num_mtt = 1 << 20, }; +static int log_num_mac = 2; +module_param_named(log_num_mac, log_num_mac, int, 0444); +MODULE_PARM_DESC(log_num_mac, "Log2 max number of MACs per ETH port (1-7)"); + +static int log_num_vlan; +module_param_named(log_num_vlan, log_num_vlan, int, 0444); +MODULE_PARM_DESC(log_num_vlan, "Log2 max number of VLANs per ETH port (0-7)"); + +static int use_prio; +module_param_named(use_prio, use_prio, bool, 0444); +MODULE_PARM_DESC(use_prio, "Enable steering by VLAN priority on ETH ports " + "(0/1, default 0)"); + +static int mlx4_check_port_params(struct mlx4_dev *dev, + enum mlx4_port_type *port_type) +{ + int i; + + for (i = 0; i < dev->caps.num_ports - 1; i++) { + if (port_type[i] != port_type[i+1] && + !(dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP)) { + mlx4_err(dev, "Only same port types supported " + "on this HCA, aborting.\n"); + return -EINVAL; + } + } + if ((port_type[0] == MLX4_PORT_TYPE_ETH) && + (port_type[1] == MLX4_PORT_TYPE_IB)) { + mlx4_err(dev, "eth-ib configuration is not supported.\n"); + return -EINVAL; + } + + for (i = 0; i < dev->caps.num_ports; i++) { + if (!(port_type[i] & dev->caps.supported_type[i+1])) { + mlx4_err(dev, "Requested port type for port %d is not " + "supported on this HCA\n", i + 1); + return -EINVAL; + } + } + return 0; +} + +static void mlx4_set_port_mask(struct mlx4_dev *dev) +{ + int i; + + dev->caps.port_mask = 0; + for (i = 1; i <= dev->caps.num_ports; ++i) + if (dev->caps.port_type[i] == MLX4_PORT_TYPE_IB) + dev->caps.port_mask |= 1 << (i - 1); +} static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) { int err; @@ -120,10 +171,13 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) dev->caps.num_ports = dev_cap->num_ports; for (i = 1; i <= dev->caps.num_ports; ++i) { dev->caps.vl_cap[i] = dev_cap->max_vl[i]; - dev->caps.mtu_cap[i] = dev_cap->max_mtu[i]; + dev->caps.ib_mtu_cap[i] = dev_cap->ib_mtu[i]; dev->caps.gid_table_len[i] = dev_cap->max_gids[i]; dev->caps.pkey_table_len[i] = dev_cap->max_pkeys[i]; dev->caps.port_width_cap[i] = dev_cap->max_port_width[i]; + dev->caps.eth_mtu_cap[i] = dev_cap->eth_mtu[i]; + dev->caps.def_mac[i] = dev_cap->def_mac[i]; + dev->caps.supported_type[i] = dev_cap->supported_port_types[i]; } dev->caps.num_uars = dev_cap->uar_size / PAGE_SIZE; @@ -134,7 +188,6 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) dev->caps.max_rq_sg = dev_cap->max_rq_sg; dev->caps.max_wqes = dev_cap->max_qp_sz; dev->caps.max_qp_init_rdma = dev_cap->max_requester_per_qp; - dev->caps.reserved_qps = dev_cap->reserved_qps; dev->caps.max_srq_wqes = dev_cap->max_srq_sz; dev->caps.max_srq_sge = dev_cap->max_rq_sg - 1; dev->caps.reserved_srqs = dev_cap->reserved_srqs; @@ -163,9 +216,138 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap) dev->caps.stat_rate_support = dev_cap->stat_rate_support; dev->caps.max_gso_sz = dev_cap->max_gso_sz; + dev->caps.log_num_macs = log_num_mac; + dev->caps.log_num_vlans = log_num_vlan; + dev->caps.log_num_prios = use_prio ? 3 : 0; + + for (i = 1; i <= dev->caps.num_ports; ++i) { + if (dev->caps.supported_type[i] != MLX4_PORT_TYPE_ETH) + dev->caps.port_type[i] = MLX4_PORT_TYPE_IB; + else + dev->caps.port_type[i] = MLX4_PORT_TYPE_ETH; + + if (dev->caps.log_num_macs > dev_cap->log_max_macs[i]) { + dev->caps.log_num_macs = dev_cap->log_max_macs[i]; + mlx4_warn(dev, "Requested number of MACs is too much " + "for port %d, reducing to %d.\n", + i, 1 << dev->caps.log_num_macs); + } + if (dev->caps.log_num_vlans > dev_cap->log_max_vlans[i]) { + dev->caps.log_num_vlans = dev_cap->log_max_vlans[i]; + mlx4_warn(dev, "Requested number of VLANs is too much " + "for port %d, reducing to %d.\n", + i, 1 << dev->caps.log_num_vlans); + } + } + + mlx4_set_port_mask(dev); + + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] = dev_cap->reserved_qps; + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] = + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] = + (1 << dev->caps.log_num_macs) * + (1 << dev->caps.log_num_vlans) * + (1 << dev->caps.log_num_prios) * + dev->caps.num_ports; + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH] = MLX4_NUM_FEXCH; + + dev->caps.reserved_qps = dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] + + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] + + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] + + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH]; + return 0; } +/* + * Change the port configuration of the device. + * Every user of this function must hold the port mutex. + */ +static int mlx4_change_port_types(struct mlx4_dev *dev, + enum mlx4_port_type *port_types) +{ + int err = 0; + int change = 0; + int port; + + for (port = 0; port < dev->caps.num_ports; port++) { + if (port_types[port] != dev->caps.port_type[port + 1]) { + change = 1; + dev->caps.port_type[port + 1] = port_types[port]; + } + } + if (change) { + mlx4_unregister_device(dev); + for (port = 1; port <= dev->caps.num_ports; port++) { + mlx4_CLOSE_PORT(dev, port); + err = mlx4_SET_PORT(dev, port); + if (err) { + mlx4_err(dev, "Failed to set port %d, " + "aborting\n", port); + goto out; + } + } + mlx4_set_port_mask(dev); + err = mlx4_register_device(dev); + } + +out: + return err; +} + +static ssize_t show_port_type(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info, + port_attr); + struct mlx4_dev *mdev = info->dev; + + return sprintf(buf, "%s\n", + mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_IB ? + "ib" : "eth"); +} + +static ssize_t set_port_type(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info, + port_attr); + struct mlx4_dev *mdev = info->dev; + struct mlx4_priv *priv = mlx4_priv(mdev); + enum mlx4_port_type types[MLX4_MAX_PORTS]; + int i; + int err = 0; + + if (!strcmp(buf, "ib\n")) + info->tmp_type = MLX4_PORT_TYPE_IB; + else if (!strcmp(buf, "eth\n")) + info->tmp_type = MLX4_PORT_TYPE_ETH; + else { + mlx4_err(mdev, "%s is not supported port type\n", buf); + return -EINVAL; + } + + mutex_lock(&priv->port_mutex); + for (i = 0; i < mdev->caps.num_ports; i++) + types[i] = priv->port[i+1].tmp_type ? priv->port[i+1].tmp_type : + mdev->caps.port_type[i+1]; + + err = mlx4_check_port_params(mdev, types); + if (err) + goto out; + + for (i = 1; i <= mdev->caps.num_ports; i++) + priv->port[i].tmp_type = 0; + + err = mlx4_change_port_types(mdev, types); + +out: + mutex_unlock(&priv->port_mutex); + return err ? err : count; +} + static int mlx4_load_fw(struct mlx4_dev *dev) { struct mlx4_priv *priv = mlx4_priv(dev); @@ -211,7 +393,8 @@ static int mlx4_init_cmpt_table(struct mlx4_dev *dev, u64 cmpt_base, ((u64) (MLX4_CMPT_TYPE_QP * cmpt_entry_sz) << MLX4_CMPT_SHIFT), cmpt_entry_sz, dev->caps.num_qps, - dev->caps.reserved_qps, 0, 0); + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], + 0, 0); if (err) goto err; @@ -336,7 +519,8 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap, init_hca->qpc_base, dev_cap->qpc_entry_sz, dev->caps.num_qps, - dev->caps.reserved_qps, 0, 0); + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], + 0, 0); if (err) { mlx4_err(dev, "Failed to map QP context memory, aborting.\n"); goto err_unmap_dmpt; @@ -346,7 +530,8 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap, init_hca->auxc_base, dev_cap->aux_entry_sz, dev->caps.num_qps, - dev->caps.reserved_qps, 0, 0); + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], + 0, 0); if (err) { mlx4_err(dev, "Failed to map AUXC context memory, aborting.\n"); goto err_unmap_qp; @@ -356,7 +541,8 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap, init_hca->altc_base, dev_cap->altc_entry_sz, dev->caps.num_qps, - dev->caps.reserved_qps, 0, 0); + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], + 0, 0); if (err) { mlx4_err(dev, "Failed to map ALTC context memory, aborting.\n"); goto err_unmap_auxc; @@ -366,7 +552,8 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap, init_hca->rdmarc_base, dev_cap->rdmarc_entry_sz << priv->qp_table.rdmarc_shift, dev->caps.num_qps, - dev->caps.reserved_qps, 0, 0); + dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], + 0, 0); if (err) { mlx4_err(dev, "Failed to map RDMARC context memory, aborting\n"); goto err_unmap_altc; @@ -565,6 +752,7 @@ static int mlx4_setup_hca(struct mlx4_dev *dev) { struct mlx4_priv *priv = mlx4_priv(dev); int err; + int port; err = mlx4_init_uar_table(dev); if (err) { @@ -663,8 +851,20 @@ static int mlx4_setup_hca(struct mlx4_dev *dev) goto err_qp_table_free; } + for (port = 1; port <= dev->caps.num_ports; port++) { + err = mlx4_SET_PORT(dev, port); + if (err) { + mlx4_err(dev, "Failed to set port %d, aborting\n", + port); + goto err_mcg_table_free; + } + } + return 0; +err_mcg_table_free: + mlx4_cleanup_mcg_table(dev); + err_qp_table_free: mlx4_cleanup_qp_table(dev); @@ -728,11 +928,45 @@ no_msi: priv->eq_table.eq[i].irq = dev->pdev->irq; } +static int mlx4_init_port_info(struct mlx4_dev *dev, int port) +{ + struct mlx4_port_info *info = &mlx4_priv(dev)->port[port]; + int err = 0; + + info->dev = dev; + info->port = port; + mlx4_init_mac_table(dev, &info->mac_table); + mlx4_init_vlan_table(dev, &info->vlan_table); + + sprintf(info->dev_name, "mlx4_port%d", port); + info->port_attr.attr.name = info->dev_name; + info->port_attr.attr.mode = S_IRUGO | S_IWUSR; + info->port_attr.show = show_port_type; + info->port_attr.store = set_port_type; + + err = device_create_file(&dev->pdev->dev, &info->port_attr); + if (err) { + mlx4_err(dev, "Failed to create file for port %d\n", port); + info->port = -1; + } + + return err; +} + +static void mlx4_cleanup_port_info(struct mlx4_port_info *info) +{ + if (info->port < 0) + return; + + device_remove_file(&info->dev->pdev->dev, &info->port_attr); +} + static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id) { struct mlx4_priv *priv; struct mlx4_dev *dev; int err; + int port; printk(KERN_INFO PFX "Initializing %s\n", pci_name(pdev)); @@ -807,6 +1041,8 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id) INIT_LIST_HEAD(&priv->ctx_list); spin_lock_init(&priv->ctx_lock); + mutex_init(&priv->port_mutex); + INIT_LIST_HEAD(&priv->pgdir_list); mutex_init(&priv->pgdir_mutex); @@ -842,15 +1078,24 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id) if (err) goto err_close; + for (port = 1; port <= dev->caps.num_ports; port++) { + err = mlx4_init_port_info(dev, port); + if (err) + goto err_port; + } + err = mlx4_register_device(dev); if (err) - goto err_cleanup; + goto err_port; pci_set_drvdata(pdev, dev); return 0; -err_cleanup: +err_port: + for (port = 1; port <= dev->caps.num_ports; port++) + mlx4_cleanup_port_info(&priv->port[port]); + mlx4_cleanup_mcg_table(dev); mlx4_cleanup_qp_table(dev); mlx4_cleanup_srq_table(dev); @@ -907,8 +1152,10 @@ static void mlx4_remove_one(struct pci_dev *pdev) if (dev) { mlx4_unregister_device(dev); - for (p = 1; p <= dev->caps.num_ports; ++p) + for (p = 1; p <= dev->caps.num_ports; p++) { + mlx4_cleanup_port_info(&priv->port[p]); mlx4_CLOSE_PORT(dev, p); + } mlx4_cleanup_mcg_table(dev); mlx4_cleanup_qp_table(dev); @@ -948,6 +1195,8 @@ static struct pci_device_id mlx4_pci_table[] = { { PCI_VDEVICE(MELLANOX, 0x6354) }, /* MT25408 "Hermon" QDR */ { PCI_VDEVICE(MELLANOX, 0x6732) }, /* MT25408 "Hermon" DDR PCIe gen2 */ { PCI_VDEVICE(MELLANOX, 0x673c) }, /* MT25408 "Hermon" QDR PCIe gen2 */ + { PCI_VDEVICE(MELLANOX, 0x6368) }, /* MT25408 "Hermon" EN 10GigE */ + { PCI_VDEVICE(MELLANOX, 0x6750) }, /* MT25408 "Hermon" EN 10GigE PCIe gen2 */ { 0, } }; @@ -960,10 +1209,28 @@ static struct pci_driver mlx4_driver = { .remove = __devexit_p(mlx4_remove_one) }; +static int __init mlx4_verify_params(void) +{ + if ((log_num_mac < 0) || (log_num_mac > 7)) { + printk(KERN_WARNING "mlx4_core: bad num_mac: %d\n", log_num_mac); + return -1; + } + + if ((log_num_vlan < 0) || (log_num_vlan > 7)) { + printk(KERN_WARNING "mlx4_core: bad num_vlan: %d\n", log_num_vlan); + return -1; + } + + return 0; +} + static int __init mlx4_init(void) { int ret; + if (mlx4_verify_params()) + return -EINVAL; + ret = mlx4_catas_init(); if (ret) return ret; diff --git a/drivers/net/mlx4/mcg.c b/drivers/net/mlx4/mcg.c index c83f88c..592c01a 100644 --- a/drivers/net/mlx4/mcg.c +++ b/drivers/net/mlx4/mcg.c @@ -368,8 +368,8 @@ int mlx4_init_mcg_table(struct mlx4_dev *dev) struct mlx4_priv *priv = mlx4_priv(dev); int err; - err = mlx4_bitmap_init(&priv->mcg_table.bitmap, - dev->caps.num_amgms, dev->caps.num_amgms - 1, 0); + err = mlx4_bitmap_init(&priv->mcg_table.bitmap, dev->caps.num_amgms, + dev->caps.num_amgms - 1, 0, 0); if (err) return err; diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h index 5337e3a..fa431fa 100644 --- a/drivers/net/mlx4/mlx4.h +++ b/drivers/net/mlx4/mlx4.h @@ -111,6 +111,7 @@ struct mlx4_bitmap { u32 last; u32 top; u32 max; + u32 reserved_top; u32 mask; spinlock_t lock; unsigned long *table; @@ -251,6 +252,38 @@ struct mlx4_catas_err { struct list_head list; }; +#define MLX4_MAX_MAC_NUM 128 +#define MLX4_MAC_TABLE_SIZE (MLX4_MAX_MAC_NUM << 3) + +struct mlx4_mac_table { + __be64 entries[MLX4_MAX_MAC_NUM]; + int refs[MLX4_MAX_MAC_NUM]; + struct mutex mutex; + int total; + int max; +}; + +#define MLX4_MAX_VLAN_NUM 128 +#define MLX4_VLAN_TABLE_SIZE (MLX4_MAX_VLAN_NUM << 2) + +struct mlx4_vlan_table { + __be32 entries[MLX4_MAX_VLAN_NUM]; + int refs[MLX4_MAX_VLAN_NUM]; + struct mutex mutex; + int total; + int max; +}; + +struct mlx4_port_info { + struct mlx4_dev *dev; + int port; + char dev_name[16]; + struct device_attribute port_attr; + enum mlx4_port_type tmp_type; + struct mlx4_mac_table mac_table; + struct mlx4_vlan_table vlan_table; +}; + struct mlx4_priv { struct mlx4_dev dev; @@ -279,6 +312,8 @@ struct mlx4_priv { struct mlx4_uar driver_uar; void __iomem *kar; + struct mlx4_port_info port[MLX4_MAX_PORTS + 1]; + struct mutex port_mutex; }; static inline struct mlx4_priv *mlx4_priv(struct mlx4_dev *dev) @@ -288,7 +323,10 @@ static inline struct mlx4_priv *mlx4_priv(struct mlx4_dev *dev) u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap); void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj); -int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, u32 reserved); +u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt, int align); +void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt); +int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, + u32 reserved_bot, u32 resetrved_top); void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap); int mlx4_reset(struct mlx4_dev *dev); @@ -346,4 +384,9 @@ void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type); void mlx4_handle_catas_err(struct mlx4_dev *dev); +void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table); +void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table); + +int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port); + #endif /* MLX4_H */ diff --git a/drivers/net/mlx4/mlx4_en.h b/drivers/net/mlx4/mlx4_en.h new file mode 100644 index 0000000..11fb17c --- /dev/null +++ b/drivers/net/mlx4/mlx4_en.h @@ -0,0 +1,561 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef _MLX4_EN_H_ +#define _MLX4_EN_H_ + +#include <linux/compiler.h> +#include <linux/list.h> +#include <linux/mutex.h> +#include <linux/netdevice.h> +#include <linux/inet_lro.h> + +#include <linux/mlx4/device.h> +#include <linux/mlx4/qp.h> +#include <linux/mlx4/cq.h> +#include <linux/mlx4/srq.h> +#include <linux/mlx4/doorbell.h> + +#include "en_port.h" + +#define DRV_NAME "mlx4_en" +#define DRV_VERSION "1.4.0" +#define DRV_RELDATE "Sep 2008" + + +#define MLX4_EN_MSG_LEVEL (NETIF_MSG_LINK | NETIF_MSG_IFDOWN) + +#define mlx4_dbg(mlevel, priv, format, arg...) \ + if (NETIF_MSG_##mlevel & priv->msg_enable) \ + printk(KERN_DEBUG "%s %s: " format , DRV_NAME ,\ + (&priv->mdev->pdev->dev)->bus_id , ## arg) + +#define mlx4_err(mdev, format, arg...) \ + printk(KERN_ERR "%s %s: " format , DRV_NAME ,\ + (&mdev->pdev->dev)->bus_id , ## arg) +#define mlx4_info(mdev, format, arg...) \ + printk(KERN_INFO "%s %s: " format , DRV_NAME ,\ + (&mdev->pdev->dev)->bus_id , ## arg) +#define mlx4_warn(mdev, format, arg...) \ + printk(KERN_WARNING "%s %s: " format , DRV_NAME ,\ + (&mdev->pdev->dev)->bus_id , ## arg) + +/* + * Device constants + */ + + +#define MLX4_EN_PAGE_SHIFT 12 +#define MLX4_EN_PAGE_SIZE (1 << MLX4_EN_PAGE_SHIFT) +#define MAX_TX_RINGS 16 +#define MAX_RX_RINGS 16 +#define MAX_RSS_MAP_SIZE 64 +#define RSS_FACTOR 2 +#define TXBB_SIZE 64 +#define HEADROOM (2048 / TXBB_SIZE + 1) +#define MAX_LSO_HDR_SIZE 92 +#define STAMP_STRIDE 64 +#define STAMP_DWORDS (STAMP_STRIDE / 4) +#define STAMP_SHIFT 31 +#define STAMP_VAL 0x7fffffff +#define STATS_DELAY (HZ / 4) + +/* Typical TSO descriptor with 16 gather entries is 352 bytes... */ +#define MAX_DESC_SIZE 512 +#define MAX_DESC_TXBBS (MAX_DESC_SIZE / TXBB_SIZE) + +/* + * OS related constants and tunables + */ + +#define MLX4_EN_WATCHDOG_TIMEOUT (15 * HZ) + +#define MLX4_EN_ALLOC_ORDER 2 +#define MLX4_EN_ALLOC_SIZE (PAGE_SIZE << MLX4_EN_ALLOC_ORDER) + +#define MLX4_EN_MAX_LRO_DESCRIPTORS 32 + +/* Receive fragment sizes; we use at most 4 fragments (for 9600 byte MTU + * and 4K allocations) */ +enum { + FRAG_SZ0 = 512 - NET_IP_ALIGN, + FRAG_SZ1 = 1024, + FRAG_SZ2 = 4096, + FRAG_SZ3 = MLX4_EN_ALLOC_SIZE +}; +#define MLX4_EN_MAX_RX_FRAGS 4 + +/* Minimum ring size for our page-allocation sceme to work */ +#define MLX4_EN_MIN_RX_SIZE (MLX4_EN_ALLOC_SIZE / SMP_CACHE_BYTES) +#define MLX4_EN_MIN_TX_SIZE (4096 / TXBB_SIZE) + +#define MLX4_EN_TX_RING_NUM 9 +#define MLX4_EN_DEF_TX_RING_SIZE 1024 +#define MLX4_EN_DEF_RX_RING_SIZE 1024 + +/* Target number of bytes to coalesce with interrupt moderation */ +#define MLX4_EN_RX_COAL_TARGET 0x20000 +#define MLX4_EN_RX_COAL_TIME 0x10 + +#define MLX4_EN_TX_COAL_PKTS 5 +#define MLX4_EN_TX_COAL_TIME 0x80 + +#define MLX4_EN_RX_RATE_LOW 400000 +#define MLX4_EN_RX_COAL_TIME_LOW 0 +#define MLX4_EN_RX_RATE_HIGH 450000 +#define MLX4_EN_RX_COAL_TIME_HIGH 128 +#define MLX4_EN_RX_SIZE_THRESH 1024 +#define MLX4_EN_RX_RATE_THRESH (1000000 / MLX4_EN_RX_COAL_TIME_HIGH) +#define MLX4_EN_SAMPLE_INTERVAL 0 + +#define MLX4_EN_AUTO_CONF 0xffff + +#define MLX4_EN_DEF_RX_PAUSE 1 +#define MLX4_EN_DEF_TX_PAUSE 1 + +/* Interval between sucessive polls in the Tx routine when polling is used + instead of interrupts (in per-core Tx rings) - should be power of 2 */ +#define MLX4_EN_TX_POLL_MODER 16 +#define MLX4_EN_TX_POLL_TIMEOUT (HZ / 4) + +#define ETH_LLC_SNAP_SIZE 8 + +#define SMALL_PACKET_SIZE (256 - NET_IP_ALIGN) +#define HEADER_COPY_SIZE (128 - NET_IP_ALIGN) + +#define MLX4_EN_MIN_MTU 46 +#define ETH_BCAST 0xffffffffffffULL + +#ifdef MLX4_EN_PERF_STAT +/* Number of samples to 'average' */ +#define AVG_SIZE 128 +#define AVG_FACTOR 1024 +#define NUM_PERF_STATS NUM_PERF_COUNTERS + +#define INC_PERF_COUNTER(cnt) (++(cnt)) +#define ADD_PERF_COUNTER(cnt, add) ((cnt) += (add)) +#define AVG_PERF_COUNTER(cnt, sample) \ + ((cnt) = ((cnt) * (AVG_SIZE - 1) + (sample) * AVG_FACTOR) / AVG_SIZE) +#define GET_PERF_COUNTER(cnt) (cnt) +#define GET_AVG_PERF_COUNTER(cnt) ((cnt) / AVG_FACTOR) + +#else + +#define NUM_PERF_STATS 0 +#define INC_PERF_COUNTER(cnt) do {} while (0) +#define ADD_PERF_COUNTER(cnt, add) do {} while (0) +#define AVG_PERF_COUNTER(cnt, sample) do {} while (0) +#define GET_PERF_COUNTER(cnt) (0) +#define GET_AVG_PERF_COUNTER(cnt) (0) +#endif /* MLX4_EN_PERF_STAT */ + +/* + * Configurables + */ + +enum cq_type { + RX = 0, + TX = 1, +}; + + +/* + * Useful macros + */ +#define ROUNDUP_LOG2(x) ilog2(roundup_pow_of_two(x)) +#define XNOR(x, y) (!(x) == !(y)) +#define ILLEGAL_MAC(addr) (addr == 0xffffffffffffULL || addr == 0x0) + + +struct mlx4_en_tx_info { + struct sk_buff *skb; + u32 nr_txbb; + u8 linear; + u8 data_offset; +}; + + +#define MLX4_EN_BIT_DESC_OWN 0x80000000 +#define CTRL_SIZE sizeof(struct mlx4_wqe_ctrl_seg) +#define MLX4_EN_MEMTYPE_PAD 0x100 +#define DS_SIZE sizeof(struct mlx4_wqe_data_seg) + + +struct mlx4_en_tx_desc { + struct mlx4_wqe_ctrl_seg ctrl; + union { + struct mlx4_wqe_data_seg data; /* at least one data segment */ + struct mlx4_wqe_lso_seg lso; + struct mlx4_wqe_inline_seg inl; + }; +}; + +#define MLX4_EN_USE_SRQ 0x01000000 + +struct mlx4_en_rx_alloc { + struct page *page; + u16 offset; +}; + +struct mlx4_en_tx_ring { + struct mlx4_hwq_resources wqres; + u32 size ; /* number of TXBBs */ + u32 size_mask; + u16 stride; + u16 cqn; /* index of port CQ associated with this ring */ + u32 prod; + u32 cons; + u32 buf_size; + u32 doorbell_qpn; + void *buf; + u16 poll_cnt; + int blocked; + struct mlx4_en_tx_info *tx_info; + u8 *bounce_buf; + u32 last_nr_txbb; + struct mlx4_qp qp; + struct mlx4_qp_context context; + int qpn; + enum mlx4_qp_state qp_state; + struct mlx4_srq dummy; + unsigned long bytes; + unsigned long packets; + spinlock_t comp_lock; +}; + +struct mlx4_en_rx_desc { + struct mlx4_wqe_srq_next_seg next; + /* actual number of entries depends on rx ring stride */ + struct mlx4_wqe_data_seg data[0]; +}; + +struct mlx4_en_rx_ring { + struct mlx4_srq srq; + struct mlx4_hwq_resources wqres; + struct mlx4_en_rx_alloc page_alloc[MLX4_EN_MAX_RX_FRAGS]; + struct net_lro_mgr lro; + u32 size ; /* number of Rx descs*/ + u32 actual_size; + u32 size_mask; + u16 stride; + u16 log_stride; + u16 cqn; /* index of port CQ associated with this ring */ + u32 prod; + u32 cons; + u32 buf_size; + int need_refill; + int full; + void *buf; + void *rx_info; + unsigned long bytes; + unsigned long packets; +}; + + +static inline int mlx4_en_can_lro(__be16 status) +{ + return (status & cpu_to_be16(MLX4_CQE_STATUS_IPV4 | + MLX4_CQE_STATUS_IPV4F | + MLX4_CQE_STATUS_IPV6 | + MLX4_CQE_STATUS_IPV4OPT | + MLX4_CQE_STATUS_TCP | + MLX4_CQE_STATUS_UDP | + MLX4_CQE_STATUS_IPOK)) == + cpu_to_be16(MLX4_CQE_STATUS_IPV4 | + MLX4_CQE_STATUS_IPOK | + MLX4_CQE_STATUS_TCP); +} + +struct mlx4_en_cq { + struct mlx4_cq mcq; + struct mlx4_hwq_resources wqres; + int ring; + spinlock_t lock; + struct net_device *dev; + struct napi_struct napi; + /* Per-core Tx cq processing support */ + struct timer_list timer; + int size; + int buf_size; + unsigned vector; + enum cq_type is_tx; + u16 moder_time; + u16 moder_cnt; + int armed; + struct mlx4_cqe *buf; +#define MLX4_EN_OPCODE_ERROR 0x1e +}; + +struct mlx4_en_port_profile { + u32 flags; + u32 tx_ring_num; + u32 rx_ring_num; + u32 tx_ring_size; + u32 rx_ring_size; +}; + +struct mlx4_en_profile { + int rss_xor; + int num_lro; + u8 rss_mask; + u32 active_ports; + u32 small_pkt_int; + int rx_moder_cnt; + int rx_moder_time; + int auto_moder; + u8 rx_pause; + u8 rx_ppp; + u8 tx_pause; + u8 tx_ppp; + u8 no_reset; + struct mlx4_en_port_profile prof[MLX4_MAX_PORTS + 1]; +}; + +struct mlx4_en_dev { + struct mlx4_dev *dev; + struct pci_dev *pdev; + struct mutex state_lock; + struct net_device *pndev[MLX4_MAX_PORTS + 1]; + u32 port_cnt; + bool device_up; + struct mlx4_en_profile profile; + u32 LSO_support; + struct workqueue_struct *workqueue; + struct device *dma_device; + void __iomem *uar_map; + struct mlx4_uar priv_uar; + struct mlx4_mr mr; + u32 priv_pdn; + spinlock_t uar_lock; +}; + + +struct mlx4_en_rss_map { + int size; + int base_qpn; + u16 map[MAX_RSS_MAP_SIZE]; + struct mlx4_qp qps[MAX_RSS_MAP_SIZE]; + enum mlx4_qp_state state[MAX_RSS_MAP_SIZE]; + struct mlx4_qp indir_qp; + enum mlx4_qp_state indir_state; +}; + +struct mlx4_en_rss_context { + __be32 base_qpn; + __be32 default_qpn; + u16 reserved; + u8 hash_fn; + u8 flags; + __be32 rss_key[10]; +}; + +struct mlx4_en_pkt_stats { + unsigned long broadcast; + unsigned long rx_prio[8]; + unsigned long tx_prio[8]; +#define NUM_PKT_STATS 17 +}; + +struct mlx4_en_port_stats { + unsigned long lro_aggregated; + unsigned long lro_flushed; + unsigned long lro_no_desc; + unsigned long tso_packets; + unsigned long queue_stopped; + unsigned long wake_queue; + unsigned long tx_timeout; + unsigned long rx_alloc_failed; + unsigned long rx_chksum_good; + unsigned long rx_chksum_none; + unsigned long tx_chksum_offload; +#define NUM_PORT_STATS 11 +}; + +struct mlx4_en_perf_stats { + u32 tx_poll; + u64 tx_pktsz_avg; + u32 inflight_avg; + u16 tx_coal_avg; + u16 rx_coal_avg; + u32 napi_quota; +#define NUM_PERF_COUNTERS 6 +}; + +struct mlx4_en_frag_info { + u16 frag_size; + u16 frag_prefix_size; + u16 frag_stride; + u16 frag_align; + u16 last_offset; + +}; + +struct mlx4_en_priv { + struct mlx4_en_dev *mdev; + struct mlx4_en_port_profile *prof; + struct net_device *dev; + struct vlan_group *vlgrp; + struct net_device_stats stats; + struct net_device_stats ret_stats; + spinlock_t stats_lock; + + unsigned long last_moder_packets; + unsigned long last_moder_tx_packets; + unsigned long last_moder_bytes; + unsigned long last_moder_jiffies; + int last_moder_time; + u16 rx_usecs; + u16 rx_frames; + u16 tx_usecs; + u16 tx_frames; + u32 pkt_rate_low; + u16 rx_usecs_low; + u32 pkt_rate_high; + u16 rx_usecs_high; + u16 sample_interval; + u16 adaptive_rx_coal; + u32 msg_enable; + + struct mlx4_hwq_resources res; + int link_state; + int last_link_state; + bool port_up; + int port; + int registered; + int allocated; + int stride; + int rx_csum; + u64 mac; + int mac_index; + unsigned max_mtu; + int base_qpn; + + struct mlx4_en_rss_map rss_map; + u16 tx_prio_map[8]; + u32 flags; +#define MLX4_EN_FLAG_PROMISC 0x1 + u32 tx_ring_num; + u32 rx_ring_num; + u32 rx_skb_size; + struct mlx4_en_frag_info frag_info[MLX4_EN_MAX_RX_FRAGS]; + u16 num_frags; + u16 log_rx_info; + + struct mlx4_en_tx_ring tx_ring[MAX_TX_RINGS]; + struct mlx4_en_rx_ring rx_ring[MAX_RX_RINGS]; + struct mlx4_en_cq tx_cq[MAX_TX_RINGS]; + struct mlx4_en_cq rx_cq[MAX_RX_RINGS]; + struct work_struct mcast_task; + struct work_struct mac_task; + struct delayed_work refill_task; + struct work_struct watchdog_task; + struct work_struct linkstate_task; + struct delayed_work stats_task; + struct mlx4_en_perf_stats pstats; + struct mlx4_en_pkt_stats pkstats; + struct mlx4_en_port_stats port_stats; + struct dev_mc_list *mc_list; + struct mlx4_en_stat_out_mbox hw_stats; +}; + + +void mlx4_en_destroy_netdev(struct net_device *dev); +int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port, + struct mlx4_en_port_profile *prof); + +int mlx4_en_get_profile(struct mlx4_en_dev *mdev); + +int mlx4_en_create_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq, + int entries, int ring, enum cq_type mode); +void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq); +int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq); +void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq); +int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq); +int mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq); + +void mlx4_en_poll_tx_cq(unsigned long data); +void mlx4_en_tx_irq(struct mlx4_cq *mcq); +int mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev); + +int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring, + u32 size, u16 stride); +void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring); +int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring, + int cq, int srqn); +void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_tx_ring *ring); + +int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring, + u32 size, u16 stride); +void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring); +int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv); +void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv, + struct mlx4_en_rx_ring *ring); +int mlx4_en_process_rx_cq(struct net_device *dev, + struct mlx4_en_cq *cq, + int budget); +int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget); +void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride, + int is_tx, int rss, int qpn, int cqn, int srqn, + struct mlx4_qp_context *context); +int mlx4_en_map_buffer(struct mlx4_buf *buf); +void mlx4_en_unmap_buffer(struct mlx4_buf *buf); + +void mlx4_en_calc_rx_buf(struct net_device *dev); +void mlx4_en_set_default_rss_map(struct mlx4_en_priv *priv, + struct mlx4_en_rss_map *rss_map, + int num_entries, int num_rings); +void mlx4_en_set_prio_map(struct mlx4_en_priv *priv, u16 *prio_map, u32 ring_num); +int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv); +void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv); +int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring); +void mlx4_en_rx_refill(struct work_struct *work); +void mlx4_en_rx_irq(struct mlx4_cq *mcq); + +int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port, u64 mac, u64 clear, u8 mode); +int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, u8 port, struct vlan_group *grp); +int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu, + u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx); +int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn, + u8 promisc); + +int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset); + +/* + * Globals + */ +extern const struct ethtool_ops mlx4_en_ethtool_ops; +#endif diff --git a/drivers/net/mlx4/mr.c b/drivers/net/mlx4/mr.c index d1dd5b4..0caf74c 100644 --- a/drivers/net/mlx4/mr.c +++ b/drivers/net/mlx4/mr.c @@ -461,7 +461,7 @@ int mlx4_init_mr_table(struct mlx4_dev *dev) int err; err = mlx4_bitmap_init(&mr_table->mpt_bitmap, dev->caps.num_mpts, - ~0, dev->caps.reserved_mrws); + ~0, dev->caps.reserved_mrws, 0); if (err) return err; diff --git a/drivers/net/mlx4/pd.c b/drivers/net/mlx4/pd.c index aa61689..26d1a7a 100644 --- a/drivers/net/mlx4/pd.c +++ b/drivers/net/mlx4/pd.c @@ -62,7 +62,7 @@ int mlx4_init_pd_table(struct mlx4_dev *dev) struct mlx4_priv *priv = mlx4_priv(dev); return mlx4_bitmap_init(&priv->pd_bitmap, dev->caps.num_pds, - (1 << 24) - 1, dev->caps.reserved_pds); + (1 << 24) - 1, dev->caps.reserved_pds, 0); } void mlx4_cleanup_pd_table(struct mlx4_dev *dev) @@ -100,7 +100,7 @@ int mlx4_init_uar_table(struct mlx4_dev *dev) return mlx4_bitmap_init(&mlx4_priv(dev)->uar_table.bitmap, dev->caps.num_uars, dev->caps.num_uars - 1, - max(128, dev->caps.reserved_uars)); + max(128, dev->caps.reserved_uars), 0); } void mlx4_cleanup_uar_table(struct mlx4_dev *dev) diff --git a/drivers/net/mlx4/port.c b/drivers/net/mlx4/port.c new file mode 100644 index 0000000..e2fdab4 --- /dev/null +++ b/drivers/net/mlx4/port.c @@ -0,0 +1,282 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/errno.h> +#include <linux/if_ether.h> + +#include <linux/mlx4/cmd.h> + +#include "mlx4.h" + +#define MLX4_MAC_VALID (1ull << 63) +#define MLX4_MAC_MASK 0xffffffffffffULL + +#define MLX4_VLAN_VALID (1u << 31) +#define MLX4_VLAN_MASK 0xfff + +void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table) +{ + int i; + + mutex_init(&table->mutex); + for (i = 0; i < MLX4_MAX_MAC_NUM; i++) { + table->entries[i] = 0; + table->refs[i] = 0; + } + table->max = 1 << dev->caps.log_num_macs; + table->total = 0; +} + +void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table) +{ + int i; + + mutex_init(&table->mutex); + for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) { + table->entries[i] = 0; + table->refs[i] = 0; + } + table->max = 1 << dev->caps.log_num_vlans; + table->total = 0; +} + +static int mlx4_set_port_mac_table(struct mlx4_dev *dev, u8 port, + __be64 *entries) +{ + struct mlx4_cmd_mailbox *mailbox; + u32 in_mod; + int err; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + + memcpy(mailbox->buf, entries, MLX4_MAC_TABLE_SIZE); + + in_mod = MLX4_SET_PORT_MAC_TABLE << 8 | port; + err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT, + MLX4_CMD_TIME_CLASS_B); + + mlx4_free_cmd_mailbox(dev, mailbox); + return err; +} + +int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index) +{ + struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table; + int i, err = 0; + int free = -1; + + mlx4_dbg(dev, "Registering MAC: 0x%llx\n", (unsigned long long) mac); + mutex_lock(&table->mutex); + for (i = 0; i < MLX4_MAX_MAC_NUM - 1; i++) { + if (free < 0 && !table->refs[i]) { + free = i; + continue; + } + + if (mac == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) { + /* MAC already registered, increase refernce count */ + *index = i; + ++table->refs[i]; + goto out; + } + } + mlx4_dbg(dev, "Free MAC index is %d\n", free); + + if (table->total == table->max) { + /* No free mac entries */ + err = -ENOSPC; + goto out; + } + + /* Register new MAC */ + table->refs[free] = 1; + table->entries[free] = cpu_to_be64(mac | MLX4_MAC_VALID); + + err = mlx4_set_port_mac_table(dev, port, table->entries); + if (unlikely(err)) { + mlx4_err(dev, "Failed adding MAC: 0x%llx\n", (unsigned long long) mac); + table->refs[free] = 0; + table->entries[free] = 0; + goto out; + } + + *index = free; + ++table->total; +out: + mutex_unlock(&table->mutex); + return err; +} +EXPORT_SYMBOL_GPL(mlx4_register_mac); + +void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int index) +{ + struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table; + + mutex_lock(&table->mutex); + if (!table->refs[index]) { + mlx4_warn(dev, "No MAC entry for index %d\n", index); + goto out; + } + if (--table->refs[index]) { + mlx4_warn(dev, "Have more references for index %d," + "no need to modify MAC table\n", index); + goto out; + } + table->entries[index] = 0; + mlx4_set_port_mac_table(dev, port, table->entries); + --table->total; +out: + mutex_unlock(&table->mutex); +} +EXPORT_SYMBOL_GPL(mlx4_unregister_mac); + +static int mlx4_set_port_vlan_table(struct mlx4_dev *dev, u8 port, + __be32 *entries) +{ + struct mlx4_cmd_mailbox *mailbox; + u32 in_mod; + int err; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + + memcpy(mailbox->buf, entries, MLX4_VLAN_TABLE_SIZE); + in_mod = MLX4_SET_PORT_VLAN_TABLE << 8 | port; + err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT, + MLX4_CMD_TIME_CLASS_B); + + mlx4_free_cmd_mailbox(dev, mailbox); + + return err; +} + +int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index) +{ + struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table; + int i, err = 0; + int free = -1; + + mutex_lock(&table->mutex); + for (i = MLX4_VLAN_REGULAR; i < MLX4_MAX_VLAN_NUM; i++) { + if (free < 0 && (table->refs[i] == 0)) { + free = i; + continue; + } + + if (table->refs[i] && + (vlan == (MLX4_VLAN_MASK & + be32_to_cpu(table->entries[i])))) { + /* Vlan already registered, increase refernce count */ + *index = i; + ++table->refs[i]; + goto out; + } + } + + if (table->total == table->max) { + /* No free vlan entries */ + err = -ENOSPC; + goto out; + } + + /* Register new MAC */ + table->refs[free] = 1; + table->entries[free] = cpu_to_be32(vlan | MLX4_VLAN_VALID); + + err = mlx4_set_port_vlan_table(dev, port, table->entries); + if (unlikely(err)) { + mlx4_warn(dev, "Failed adding vlan: %u\n", vlan); + table->refs[free] = 0; + table->entries[free] = 0; + goto out; + } + + *index = free; + ++table->total; +out: + mutex_unlock(&table->mutex); + return err; +} +EXPORT_SYMBOL_GPL(mlx4_register_vlan); + +void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index) +{ + struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table; + + if (index < MLX4_VLAN_REGULAR) { + mlx4_warn(dev, "Trying to free special vlan index %d\n", index); + return; + } + + mutex_lock(&table->mutex); + if (!table->refs[index]) { + mlx4_warn(dev, "No vlan entry for index %d\n", index); + goto out; + } + if (--table->refs[index]) { + mlx4_dbg(dev, "Have more references for index %d," + "no need to modify vlan table\n", index); + goto out; + } + table->entries[index] = 0; + mlx4_set_port_vlan_table(dev, port, table->entries); + --table->total; +out: + mutex_unlock(&table->mutex); +} +EXPORT_SYMBOL_GPL(mlx4_unregister_vlan); + +int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port) +{ + struct mlx4_cmd_mailbox *mailbox; + int err; + u8 is_eth = dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH; + + mailbox = mlx4_alloc_cmd_mailbox(dev); + if (IS_ERR(mailbox)) + return PTR_ERR(mailbox); + + memset(mailbox->buf, 0, 256); + if (is_eth) { + ((u8 *) mailbox->buf)[3] = 6; + ((__be16 *) mailbox->buf)[4] = cpu_to_be16(1 << 15); + ((__be16 *) mailbox->buf)[6] = cpu_to_be16(1 << 15); + } + err = mlx4_cmd(dev, mailbox->dma, port, is_eth, MLX4_CMD_SET_PORT, + MLX4_CMD_TIME_CLASS_B); + + mlx4_free_cmd_mailbox(dev, mailbox); + return err; +} diff --git a/drivers/net/mlx4/qp.c b/drivers/net/mlx4/qp.c index c49a860..1c565ef 100644 --- a/drivers/net/mlx4/qp.c +++ b/drivers/net/mlx4/qp.c @@ -147,19 +147,42 @@ int mlx4_qp_modify(struct mlx4_dev *dev, struct mlx4_mtt *mtt, } EXPORT_SYMBOL_GPL(mlx4_qp_modify); -int mlx4_qp_alloc(struct mlx4_dev *dev, int sqpn, struct mlx4_qp *qp) +int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align, int *base) +{ + struct mlx4_priv *priv = mlx4_priv(dev); + struct mlx4_qp_table *qp_table = &priv->qp_table; + int qpn; + + qpn = mlx4_bitmap_alloc_range(&qp_table->bitmap, cnt, align); + if (qpn == -1) + return -ENOMEM; + + *base = qpn; + return 0; +} +EXPORT_SYMBOL_GPL(mlx4_qp_reserve_range); + +void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt) +{ + struct mlx4_priv *priv = mlx4_priv(dev); + struct mlx4_qp_table *qp_table = &priv->qp_table; + if (base_qpn < dev->caps.sqp_start + 8) + return; + + mlx4_bitmap_free_range(&qp_table->bitmap, base_qpn, cnt); +} +EXPORT_SYMBOL_GPL(mlx4_qp_release_range); + +int mlx4_qp_alloc(struct mlx4_dev *dev, int qpn, struct mlx4_qp *qp) { struct mlx4_priv *priv = mlx4_priv(dev); struct mlx4_qp_table *qp_table = &priv->qp_table; int err; - if (sqpn) - qp->qpn = sqpn; - else { - qp->qpn = mlx4_bitmap_alloc(&qp_table->bitmap); - if (qp->qpn == -1) - return -ENOMEM; - } + if (!qpn) + return -EINVAL; + + qp->qpn = qpn; err = mlx4_table_get(dev, &qp_table->qp_table, qp->qpn); if (err) @@ -208,9 +231,6 @@ err_put_qp: mlx4_table_put(dev, &qp_table->qp_table, qp->qpn); err_out: - if (!sqpn) - mlx4_bitmap_free(&qp_table->bitmap, qp->qpn); - return err; } EXPORT_SYMBOL_GPL(mlx4_qp_alloc); @@ -239,9 +259,6 @@ void mlx4_qp_free(struct mlx4_dev *dev, struct mlx4_qp *qp) mlx4_table_put(dev, &qp_table->altc_table, qp->qpn); mlx4_table_put(dev, &qp_table->auxc_table, qp->qpn); mlx4_table_put(dev, &qp_table->qp_table, qp->qpn); - - if (qp->qpn >= dev->caps.sqp_start + 8) - mlx4_bitmap_free(&qp_table->bitmap, qp->qpn); } EXPORT_SYMBOL_GPL(mlx4_qp_free); @@ -255,6 +272,7 @@ int mlx4_init_qp_table(struct mlx4_dev *dev) { struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table; int err; + int reserved_from_top = 0; spin_lock_init(&qp_table->lock); INIT_RADIX_TREE(&dev->qp_table_tree, GFP_ATOMIC); @@ -264,9 +282,40 @@ int mlx4_init_qp_table(struct mlx4_dev *dev) * block of special QPs must be aligned to a multiple of 8, so * round up. */ - dev->caps.sqp_start = ALIGN(dev->caps.reserved_qps, 8); + dev->caps.sqp_start = + ALIGN(dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], 8); + + { + int sort[MLX4_NUM_QP_REGION]; + int i, j, tmp; + int last_base = dev->caps.num_qps; + + for (i = 1; i < MLX4_NUM_QP_REGION; ++i) + sort[i] = i; + + for (i = MLX4_NUM_QP_REGION; i > 0; --i) { + for (j = 2; j < i; ++j) { + if (dev->caps.reserved_qps_cnt[sort[j]] > + dev->caps.reserved_qps_cnt[sort[j - 1]]) { + tmp = sort[j]; + sort[j] = sort[j - 1]; + sort[j - 1] = tmp; + } + } + } + + for (i = 1; i < MLX4_NUM_QP_REGION; ++i) { + last_base -= dev->caps.reserved_qps_cnt[sort[i]]; + dev->caps.reserved_qps_base[sort[i]] = last_base; + reserved_from_top += + dev->caps.reserved_qps_cnt[sort[i]]; + } + + } + err = mlx4_bitmap_init(&qp_table->bitmap, dev->caps.num_qps, - (1 << 24) - 1, dev->caps.sqp_start + 8); + (1 << 23) - 1, dev->caps.sqp_start + 8, + reserved_from_top); if (err) return err; diff --git a/drivers/net/mlx4/srq.c b/drivers/net/mlx4/srq.c index 533eb6d..fe9f218 100644 --- a/drivers/net/mlx4/srq.c +++ b/drivers/net/mlx4/srq.c @@ -245,7 +245,7 @@ int mlx4_init_srq_table(struct mlx4_dev *dev) INIT_RADIX_TREE(&srq_table->tree, GFP_ATOMIC); err = mlx4_bitmap_init(&srq_table->bitmap, dev->caps.num_srqs, - dev->caps.num_srqs - 1, dev->caps.reserved_srqs); + dev->caps.num_srqs - 1, dev->caps.reserved_srqs, 0); if (err) return err; diff --git a/drivers/pcmcia/Makefile b/drivers/pcmcia/Makefile index b46c60b..23e492b 100644 --- a/drivers/pcmcia/Makefile +++ b/drivers/pcmcia/Makefile @@ -70,7 +70,7 @@ pxa2xx-obj-$(CONFIG_MACH_MAINSTONE) += pxa2xx_mainstone.o pxa2xx-obj-$(CONFIG_PXA_SHARPSL) += pxa2xx_sharpsl.o pxa2xx-obj-$(CONFIG_MACH_ARMCORE) += pxa2xx_cm_x2xx_cs.o pxa2xx-obj-$(CONFIG_ARCH_VIPER) += pxa2xx_viper.o -pxa2xx-obj-$(CONFIG_TRIZEPS_PCMCIA) += pxa2xx_trizeps.o +pxa2xx-obj-$(CONFIG_TRIZEPS_PCMCIA) += pxa2xx_trizeps4.o pxa2xx-obj-$(CONFIG_MACH_PALMTX) += pxa2xx_palmtx.o pxa2xx-obj-$(CONFIG_MACH_PALMLD) += pxa2xx_palmld.o diff --git a/drivers/serial/Kconfig b/drivers/serial/Kconfig index db783b7..c94d3c4 100644 --- a/drivers/serial/Kconfig +++ b/drivers/serial/Kconfig @@ -457,7 +457,7 @@ config SERIAL_SAMSUNG config SERIAL_SAMSUNG_DEBUG bool "Samsung SoC serial debug" - depends on SERIAL_SAMSUNG + depends on SERIAL_SAMSUNG && DEBUG_LL help Add support for debugging the serial driver. Since this is generally being used as a console, we use our own output diff --git a/include/linux/libata.h b/include/linux/libata.h index 947cf84..c261aa0 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h @@ -340,6 +340,9 @@ enum { ATA_EHI_DID_RESET = ATA_EHI_DID_SOFTRESET | ATA_EHI_DID_HARDRESET, + /* mask of flags to transfer *to* the slave link */ + ATA_EHI_TO_SLAVE_MASK = ATA_EHI_NO_AUTOPSY | ATA_EHI_QUIET, + /* max tries if error condition is still set after ->error_handler */ ATA_EH_MAX_TRIES = 5, diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h index 77323a7..cf9c679 100644 --- a/include/linux/mlx4/cmd.h +++ b/include/linux/mlx4/cmd.h @@ -132,6 +132,15 @@ enum { MLX4_MAILBOX_SIZE = 4096 }; +enum { + /* set port opcode modifiers */ + MLX4_SET_PORT_GENERAL = 0x0, + MLX4_SET_PORT_RQP_CALC = 0x1, + MLX4_SET_PORT_MAC_TABLE = 0x2, + MLX4_SET_PORT_VLAN_TABLE = 0x3, + MLX4_SET_PORT_PRIO_MAP = 0x4, +}; + struct mlx4_dev; struct mlx4_cmd_mailbox { diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h index b2f9444..bd9977b 100644 --- a/include/linux/mlx4/device.h +++ b/include/linux/mlx4/device.h @@ -60,6 +60,7 @@ enum { MLX4_DEV_CAP_FLAG_IPOIB_CSUM = 1 << 7, MLX4_DEV_CAP_FLAG_BAD_PKEY_CNTR = 1 << 8, MLX4_DEV_CAP_FLAG_BAD_QKEY_CNTR = 1 << 9, + MLX4_DEV_CAP_FLAG_DPDP = 1 << 12, MLX4_DEV_CAP_FLAG_MEM_WINDOW = 1 << 16, MLX4_DEV_CAP_FLAG_APM = 1 << 17, MLX4_DEV_CAP_FLAG_ATOMIC = 1 << 18, @@ -145,6 +146,29 @@ enum { MLX4_MTT_FLAG_PRESENT = 1 }; +enum mlx4_qp_region { + MLX4_QP_REGION_FW = 0, + MLX4_QP_REGION_ETH_ADDR, + MLX4_QP_REGION_FC_ADDR, + MLX4_QP_REGION_FC_EXCH, + MLX4_NUM_QP_REGION +}; + +enum mlx4_port_type { + MLX4_PORT_TYPE_IB = 1 << 0, + MLX4_PORT_TYPE_ETH = 1 << 1, +}; + +enum mlx4_special_vlan_idx { + MLX4_NO_VLAN_IDX = 0, + MLX4_VLAN_MISS_IDX, + MLX4_VLAN_REGULAR +}; + +enum { + MLX4_NUM_FEXCH = 64 * 1024, +}; + static inline u64 mlx4_fw_ver(u64 major, u64 minor, u64 subminor) { return (major << 32) | (minor << 16) | subminor; @@ -154,7 +178,9 @@ struct mlx4_caps { u64 fw_ver; int num_ports; int vl_cap[MLX4_MAX_PORTS + 1]; - int mtu_cap[MLX4_MAX_PORTS + 1]; + int ib_mtu_cap[MLX4_MAX_PORTS + 1]; + u64 def_mac[MLX4_MAX_PORTS + 1]; + int eth_mtu_cap[MLX4_MAX_PORTS + 1]; int gid_table_len[MLX4_MAX_PORTS + 1]; int pkey_table_len[MLX4_MAX_PORTS + 1]; int local_ca_ack_delay; @@ -169,7 +195,6 @@ struct mlx4_caps { int max_rq_desc_sz; int max_qp_init_rdma; int max_qp_dest_rdma; - int reserved_qps; int sqp_start; int num_srqs; int max_srq_wqes; @@ -201,6 +226,15 @@ struct mlx4_caps { u16 stat_rate_support; u8 port_width_cap[MLX4_MAX_PORTS + 1]; int max_gso_sz; + int reserved_qps_cnt[MLX4_NUM_QP_REGION]; + int reserved_qps; + int reserved_qps_base[MLX4_NUM_QP_REGION]; + int log_num_macs; + int log_num_vlans; + int log_num_prios; + enum mlx4_port_type port_type[MLX4_MAX_PORTS + 1]; + u8 supported_type[MLX4_MAX_PORTS + 1]; + u32 port_mask; }; struct mlx4_buf_list { @@ -355,6 +389,11 @@ struct mlx4_init_port_param { u64 si_guid; }; +#define mlx4_foreach_port(port, dev, type) \ + for ((port) = 1; (port) <= (dev)->caps.num_ports; (port)++) \ + if (((type) == MLX4_PORT_TYPE_IB ? (dev)->caps.port_mask : \ + ~(dev)->caps.port_mask) & 1 << ((port) - 1)) + int mlx4_buf_alloc(struct mlx4_dev *dev, int size, int max_direct, struct mlx4_buf *buf); void mlx4_buf_free(struct mlx4_dev *dev, int size, struct mlx4_buf *buf); @@ -400,7 +439,10 @@ int mlx4_cq_alloc(struct mlx4_dev *dev, int nent, struct mlx4_mtt *mtt, int collapsed); void mlx4_cq_free(struct mlx4_dev *dev, struct mlx4_cq *cq); -int mlx4_qp_alloc(struct mlx4_dev *dev, int sqpn, struct mlx4_qp *qp); +int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align, int *base); +void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt); + +int mlx4_qp_alloc(struct mlx4_dev *dev, int qpn, struct mlx4_qp *qp); void mlx4_qp_free(struct mlx4_dev *dev, struct mlx4_qp *qp); int mlx4_srq_alloc(struct mlx4_dev *dev, u32 pdn, struct mlx4_mtt *mtt, @@ -416,6 +458,12 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16], int block_mcast_loopback); int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16]); +int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index); +void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int index); + +int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index); +void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index); + int mlx4_map_phys_fmr(struct mlx4_dev *dev, struct mlx4_fmr *fmr, u64 *page_list, int npages, u64 iova, u32 *lkey, u32 *rkey); int mlx4_fmr_alloc(struct mlx4_dev *dev, u32 pd, u32 access, int max_pages, |