summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'master' of ↵David S. Miller2011-07-2115-112/+327
|\ | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/kaber/nf-next-2.6
| * netfilter: ipset: fix compiler warnings "'hash_ip4_data_next' declared ↵Chris Friesen2011-07-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | inline after being called" Some gcc versions warn about prototypes without "inline" when the declaration includes the "inline" keyword. The fix generates a false error message "marked inline, but without a definition" with sparse below 0.4.2. Signed-off-by: Chris Friesen <chris.friesen@genband.com> Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: ipset: hash:net,iface fixed to handle overlapping nets behind ↵Jozsef Kadlecsik2011-07-218-51/+117
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | different interfaces If overlapping networks with different interfaces was added to the set, the type did not handle it properly. Example ipset create test hash:net,iface ipset add test 192.168.0.0/16,eth0 ipset add test 192.168.0.0/24,eth1 Now, if a packet was sent from 192.168.0.0/24,eth0, the type returned a match. In the patch the algorithm is fixed in order to correctly handle overlapping networks. Limitation: the same network cannot be stored with more than 64 different interfaces in a single set. Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: ipset: make possible to hash some part of the data element onlyJozsef Kadlecsik2011-07-211-4/+10
| | | | | | | | | | Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: nfnetlink_queue: batch verdict supportFlorian Westphal2011-07-192-12/+104
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduces a new nfnetlink type that applies a given verdict to all queued packets with an id <= the id in the verdict message. If a mark is provided it is applied to all matched packets. This reduces the number of verdicts that have to be sent. Applications that make use of this feature need to maintain a timeout to send a batchverdict periodically to avoid starvation. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: nfnetlink_queue: assert monotonic packet idsEric Dumazet2011-07-191-11/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Packet identifier is currently setup in nfqnl_build_packet_message(), using one atomic_inc_return(). Problem is that since several cpus might concurrently call nfqnl_enqueue_packet() for the same queue, we can deliver packets to consumer in non monotonic way (packet N+1 being delivered after packet N) This patch moves the packet id setup from nfqnl_build_packet_message() to nfqnl_enqueue_packet() to guarantee correct delivery order. This also removes one atomic operation. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Florian Westphal <fw@strlen.de> CC: Pablo Neira Ayuso <pablo@netfilter.org> CC: Eric Leblond <eric@regit.org> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: nfnetlink_queue: provide rcu enabled callbacksEric Dumazet2011-07-181-29/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | nenetlink_queue operations on SMP are not efficent if several queues are used, because of nfnl_mutex contention when applications give packet verdict. Use new call_rcu field in struct nfnl_callback to advertize a callback that is called under rcu_read_lock instead of nfnl_mutex. On my 2x4x2 machine, I was able to reach 2.000.000 pps going through user land returning NF_ACCEPT verdicts without losses, instead of less than 500.000 pps before patch. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Florian Westphal <fw@strlen.de> CC: Eric Leblond <eric@regit.org> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: nfnetlink: add RCU in nfnetlink_rcv_msg()Eric Dumazet2011-07-182-10/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Goal of this patch is to permit nfnetlink providers not mandate nfnl_mutex being held while nfnetlink_rcv_msg() calls them. If struct nfnl_callback contains a non NULL call_rcu(), then nfnetlink_rcv_msg() will use it instead of call() field, holding rcu_read_lock instead of nfnl_mutex Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Florian Westphal <fw@strlen.de> CC: Eric Leblond <eric@regit.org> Signed-off-by: Patrick McHardy <kaber@trash.net>
| * netfilter: add SELinux context support to AUDIT targetMr Dash Four2011-06-303-0/+41
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In this revision the conversion of secid to SELinux context and adding it to the audit log is moved from xt_AUDIT.c to audit.c with the aid of a separate helper function - audit_log_secctx - which does both the conversion and logging of SELinux context, thus also preventing internal secid number being leaked to userspace. If conversion is not successful an error is raised. With the introduction of this helper function the work done in xt_AUDIT.c is much more simplified. It also opens the possibility of this helper function being used by other modules (including auditd itself), if desired. With this addition, typical (raw auditd) output after applying the patch would be: type=NETFILTER_PKT msg=audit(1305852240.082:31012): action=0 hook=1 len=52 inif=? outif=eth0 saddr=10.1.1.7 daddr=10.1.2.1 ipid=16312 proto=6 sport=56150 dport=22 obj=system_u:object_r:ssh_client_packet_t:s0 type=NETFILTER_PKT msg=audit(1306772064.079:56): action=0 hook=3 len=48 inif=eth0 outif=? smac=00:05:5d:7c:27:0b dmac=00:02:b3:0a:7f:81 macproto=0x0800 saddr=10.1.2.1 daddr=10.1.1.7 ipid=462 proto=6 sport=22 dport=3561 obj=system_u:object_r:ssh_server_packet_t:s0 Acked-by: Eric Paris <eparis@redhat.com> Signed-off-by: Mr Dash Four <mr.dash.four@googlemail.com> Signed-off-by: Patrick McHardy <kaber@trash.net>
* | net: filter: BPF 'JIT' compiler for PPC64Matt Evans2011-07-217-1/+1106
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | An implementation of a code generator for BPF programs to speed up packet filtering on PPC64, inspired by Eric Dumazet's x86-64 version. Filter code is generated as an ABI-compliant function in module_alloc()'d mem with stackframe & prologue/epilogue generated if required (simple filters don't need anything more than an li/blr). The filter's local variables, M[], live in registers. Supports all BPF opcodes, although "complicated" loads from negative packet offsets (e.g. SKF_LL_OFF) are not yet supported. There are a couple of further optimisations left for future work; many-pass assembly with branch-reach reduction and a register allocator to push M[] variables into volatile registers would improve the code quality further. This currently supports big-endian 64-bit PowerPC only (but is fairly simple to port to PPC32 or LE!). Enabled in the same way as x86-64: echo 1 > /proc/sys/net/core/bpf_jit_enable Or, enabled with extra debug output: echo 2 > /proc/sys/net/core/bpf_jit_enable Signed-off-by: Matt Evans <matt@ozlabs.org> Acked-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2: Fix endian swapping on firmware version stringMichael Chan2011-07-211-3/+3
| | | | | | | | | | | | | | | | so that ethtool -i will display it correctly on big endian systems. Signed-off-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | cnic: Add VLAN ID as a parameter during netevent upcallMichael Chan2011-07-214-20/+52
| | | | | | | | | | | | | | | | The bnx2fc driver needs to handle netdev events on VLAN devices. Signed-off-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | cnic: Wait for all Context IDs to be deleted before sending FCOE_DESTROY_FUNCMichael Chan2011-07-211-18/+27
| | | | | | | | | | | | | | | | | | | | Otherwise, the firmware will not respond and we'll have to wait for timeout. Refactor the wait loop we already have into a separate function for this purpose. Signed-off-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | cnic: Fix Context ID space calculationMichael Chan2011-07-211-9/+1
| | | | | | | | | | | | | | | | | | Include FCoE CID space only for E2_PLUS devices. Remove old CID offset adjustments that are no longer needed. Signed-off-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix NVRAM selftest failures for 5720 devsMatt Carlson2011-07-211-1/+1
| | | | | | | | | | | | | | | | | | This patch fixes NVRAM selftest failures for 5720 devices by fixing the checksum area size. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Return size from tg3_vpd_readblock()Matt Carlson2011-07-211-12/+13
| | | | | | | | | | | | | | | | | | | | | | Newer VPD datablocks can exceed the size the tg3 driver is traditionally used to. This can cause some of the routines that operate on the VPD data to fail when in-fact they could have succeeded had they known the correct size. This patch fixes the problem. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix int selftest for recent devices.Matt Carlson2011-07-211-2/+6
| | | | | | | | | | | | | | | | | | | | This patch fixes interrupt selftest failures for recent devices (57765, 5717, 5718. 5719, 5720) by disabling MSI one-shot mode and applying the status tag workaround to the selftest code. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix RSS indirection table distributionMatt Carlson2011-07-211-6/+15
| | | | | | | | | | | | | | | | | | | | The current RSS indirection table is populated such that more traffic will hit the first RSS ring. This patch adjusts the indirection table so that the load is more evenly distributed. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Add 5719 and 5720 to EEE_CAP listMatt Carlson2011-07-211-1/+3
| | | | | | | | | | | | | | | | | | This patch adds the 5719 and the 5720 to the list of devices that are EEE capable. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix link down notify failure when EEE disabledMatt Carlson2011-07-211-13/+25
| | | | | | | | | | | | | | | | | | | | | | | | Occasionally, when the network cable is removed after a successful autonegotiation, the device will not send a link down interrupt to the driver. This happens because of a bad interaction of an EEE workaround. The fix is to adjust the code so that the root cause condition does not happen. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix link flap at 100Mbps with EEE enabledMatt Carlson2011-07-212-5/+6
| | | | | | | | | | | | | | | | | | | | This patch increases the scope of the EEE interoperability workaround to include more asic revisions. The workarond value is tuned to workaround a link flap issue at 100Mbps. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tg3: Fix io failures after chip resetMatt Carlson2011-07-211-12/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | Commit f2096f94b514d88593355995d5dd276961e88af1, entitled "tg3: Add 5720 H2BMC support", needed to add code to preserve some bits set by firmware. Unfortunately the new code causes throughput to stop after a chip reset because it enables state machines before they are ready. This patch undoes the problematic code. The bits will be restored later in the init sequence. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2x: Broken self-test in SF mode on 578xxVladislav Zolotarov2011-07-213-3/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch fixes both the failure in the self-test on 578xx and a hole in a parity recovery flow that this failure has discovered: - internal 'pending' state in a VLAN_MAC object wasn't been cleared when the object state change was called with DRV_ONLY flag, which in particular happens when a parity error happens during the self-test. - bp->sp_state wasn't cleared in the similar circumstances as described above. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2x: Parity errors recovery for 578xxVladislav Zolotarov2011-07-214-71/+168
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix the parity errors recovery flow for 578xx: - Add a separate column for the 578xx in the parity mask registers DB. - Fix the bnx2x_process_kill_chip_reset() to handle the blocks newly introduced in the 578xx. Cover ATC and PGLUE_B blocks for 57712 and 578xx. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2x: Read FIP mac from SHMEM in single function modeVladislav Zolotarov2011-07-211-7/+9
| | | | | | | | | | | | | | | | | | | | Read FIP MAC address from SHMEM's "port" section similar to what we do in a MF mode when we read it from a "func" section of SHMEM. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2x: Fixed ethtool -d for 578xxVladislav Zolotarov2011-07-212-811/+1202
| | | | | | | | | | | | | | | | | | | | | | Registers dump code erroneously treated 578xx as 57712. This patch fixes the above and also removes unused data structures. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | bnx2x: Implementation for netdev->ndo_fcoe_get_wwnVladislav Zolotarov2011-07-214-3/+96
| | | | | | | | | | | | Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | slcan: remove unused 'leased', 'line' and 'pid' fields from the 'slcan' ↵Matvejchikov Ilya2011-07-191-9/+1
| | | | | | | | | | | | | | | | structure Signed-off-by: Matvejchikov Ilya <matvejchikov@gmail.com> Acked-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | slip: remove unused 'line' field from the 'slip' structureMatvejchikov Ilya2011-07-192-4/+0
| | | | | | | | | | Signed-off-by: Matvejchikov Ilya <matvejchikov@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge branch 'vhost-net-next' of ↵David S. Miller2011-07-194-33/+64
|\ \ | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
| * | vhost: optimize interrupt enable/disableMichael S. Tsirkin2011-07-191-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | As we now only update used ring after enabling the backend, we can write flags with __put_user: as that's done on data path, it matters. Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
| * | vhost: fix zcopy reference countingMichael S. Tsirkin2011-07-191-1/+0
| | | | | | | | | | | | | | | | | | | | | Fix get/put refcount imbalance with zero copy, which caused qemu to hang forever on guest driver unload. Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
| * | vhost: set log when updating used flags or avail eventJason Wang2011-07-191-30/+54
| | | | | | | | | | | | | | | | | | | | | | | | | | | We need to log writes when updating used flags and avail event fields. Otherwise the guest may see a stale value after migration and miss notifying the host. Signed-off-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
| * | vhost: init used ring after backend was setJason Wang2011-07-194-8/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Move the used ring initialization after backend was set. This makes it possible to disable the backend and tweak the used ring, then restart. This will also make it possible to log the used ring write correctly. Signed-off-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
* | | Merge branch 'davem-next.r8169' of ↵David S. Miller2011-07-191-5/+3
|\ \ \ | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/romieu/netdev-2.6
| * | | r8169: fix sticky accepts packet bits in RxConfig.Francois Romieu2011-07-191-5/+3
| |/ / | | | | | | | | | | | | | | | | | | | | | e542a2269f232d61270ceddd42b73a4348dee2bb (r8169: adjust the RxConfig settings) broke the return from promiscuous mode to physical address match mode. Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Cc: Signed-off-by: Hayes Wang <hayeswang@realtek.com>
* | | bnx2x: disable FCoE for 578xx devices since not yet supportedDmitry Kravkov2011-07-191-2/+2
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: fix memory barriersVladislav Zolotarov2011-07-192-19/+75
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: use BNX2X_Q_FLG_TPA_IPV6 for TPA queue configurationVladislav Zolotarov2011-07-193-2/+5
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: disable loacal BH when scheduling FCOE napiVladislav Zolotarov2011-07-191-1/+8
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: fix MB index for 4-port devicesDmitry Kravkov2011-07-191-1/+1
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: DCB reworkDmitry Kravkov2011-07-196-112/+268
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | create DCB related states in function state-machine allow handling of DCB errors from FW allow disablement of DCB in FW, when peer disappears or error clean up unused functions/variables as pointed by David Binderman <dcb314@hotmail.com> Reported-by: David Binderman <dcb314@hotmail.com> Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: remove unnecessary dma_syncVladislav Zolotarov2011-07-191-5/+0
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: stop tx before CNIC_STOPVladislav Zolotarov2011-07-191-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | It may take some time to cnic to respond, this prevents tx_timeout when it happens. Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: add missing command in error handling flowDmitry Kravkov2011-07-191-0/+1
| | | | | | | | | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | bnx2x: use correct dma_sync functionVladislav Zolotarov2011-07-192-2/+2
|/ / | | | | | | | | | | Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | net: can: remove custom hex_to_bin()Andy Shevchenko2011-07-181-21/+5
| | | | | | | | | | | | | | Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Wolfgang Grandegger <wg@grandegger.com> Acked-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | stmmac: Allow SOCs to use Store forward mode eventhough tx_coe is 0. (V2)Srinivas Kandagatla2011-07-182-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | This patch adds new field 'force_sf_dma_mode' to plat_stmmacenet_data struct to allow users to specify if they want to use force store forward eventhough tx_coe is not available in hw. without this flag stmmac driver will use cut-thru mode not use store-forward mode. Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@st.com> Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | stmmac: add memory barriers at appropriate placesShiraz Hashim2011-07-181-0/+5
| | | | | | | | | | | | | | | | | | | | | | | | This patch, provided by ST SPEAr developers, has fixed a problem raised on ARM CA9 where happened that the dma_transmission was enabled before the dma descriptors were properly filled. To guarantee this data memory barriers have been explicity used in the driver. Signed-off-by: Shiraz Hashim <shiraz.hashim@st.com> Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | vhost: vhost TX zero-copy supportMichael S. Tsirkin2011-07-183-16/+220
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | >From: Shirley Ma <mashirle@us.ibm.com> This adds experimental zero copy support in vhost-net, disabled by default. To enable, set experimental_zcopytx module option to 1. This patch maintains the outstanding userspace buffers in the sequence it is delivered to vhost. The outstanding userspace buffers will be marked as done once the lower device buffers DMA has finished. This is monitored through last reference of kfree_skb callback. Two buffer indices are used for this purpose. The vhost-net device passes the userspace buffers info to lower device skb through message control. DMA done status check and guest notification are handled by handle_tx: in the worst case is all buffers in the vq are in pending/done status, so we need to notify guest to release DMA done buffers first before we get any new buffers from the vq. One known problem is that if the guest stops submitting buffers, buffers might never get used until some further action, e.g. device reset. This does not seem to affect linux guests. Signed-off-by: Shirley <xma@us.ibm.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
OpenPOWER on IntegriCloud