| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The nand_chip::erase_cmd callback previously served a dual purpose; for
one, it allowed a per-flash-chip override, so that AG-AND devices could
use a different erase command than other NAND. These AND devices were
dropped in commit 14c6578683367b1e7af0c3c09e872b45a45183a7 (mtd: nand:
remove AG-AND support). On the other hand, some drivers (denali and
doc-g4) need to use this sort of callback to implement
controller-specific erase operations.
To make the latter operation easier for some drivers (e.g., ST's new BCH
NAND driver), it helps if the command dispatch and wait functions can be
lumped together, rather than called separately.
This patch does two things:
1. Pull the call to chip->waitfunc() into chip->erase_cmd(), and return
the status from this callback
2. Rename erase_cmd() to just erase(), since this callback does a
little more than just send a command
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Tested-by: Lee Jones <lee.jones@linaro.org>
|
|
|
|
|
| |
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
mtd_oobtest writes OOB, read it back and verify. The verification is
not correctly done if oobsize is not multiple of 4. Although the data
to be written and the data to be compared are generated by several
prandom_byte_state() calls starting with the same seed, these two are
generated with the different size and different number of calls.
Due to the implementation of prandom_byte_state() if the size on each
call is not multiple of 4, the resulting data is not always same.
This fixes it by just calling prandom_byte_state() once and using
correct range instead of calling it multiple times for each.
Reported-by: George Cherian <george.cherian@ti.com>
Reported-by: Lothar Waßmann <LW@KARO-electronics.de>
Tested-by: Lothar Waßmann <LW@KARO-electronics.de>
Cc: George Cherian <george.cherian@ti.com>
Cc: Lothar Waßmann <LW@KARO-electronics.de>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Lee Jones <lee.jones@linaro.org>
Cc: linux-mtd@lists.infradead.org
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
|\
| |
| |
| |
| | |
Addition of the spi-nor framework, plus updates to the ST SPI FSM
driver.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
COMPILE_TEST allows us to build this driver on other arch'es. But not
all arch'es have the right I/O accessors -- particularly, x86 is missing
readsl() and writesl().
So just restrict this driver to ARCH_STI. It's still buildable for a
multiplatform ARM kernel, so it can get decent compile coverage.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Compile-testing for a 64-bit arch uncovers several bad casts:
In file included from include/linux/linkage.h:4:0,
from include/linux/kernel.h:6,
from drivers/mtd/devices/st_spi_fsm.c:15:
drivers/mtd/devices/st_spi_fsm.c: In function ‘stfsm_read_fifo’:
drivers/mtd/devices/st_spi_fsm.c:758:11: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
BUG_ON((((uint32_t)buf) & 0x3) || (size & 0x3));
...
Use uintptr_t instead of uint32_t, since it's guaranteed to be
pointer-sized.
We also see this warning, if size_t is not 32 bits wide:
In file included from drivers/mtd/devices/st_spi_fsm.c:15:0:
drivers/mtd/devices/st_spi_fsm.c: In function ‘stfsm_mtd_write’:
include/linux/kernel.h:712:17: warning: comparison of distinct pointer types lacks a cast [enabled by default]
(void) (&_min1 == &_min2); \
^
drivers/mtd/devices/st_spi_fsm.c:1704:11: note: in expansion of macro ‘min’
bytes = min(FLASH_PAGESIZE - page_offs, len);
^
Just use min_t() to force the type conversion, since we don't really
want to upgrade 'page_offs' and 'bytes' to size_t; they only should be
handling <= 256 byte offsets.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This corrects some sparse warnings:
drivers/mtd/spi-nor/fsl-quadspi.c:281:31: warning: incorrect type in initializer (different address spaces) [sparse]
drivers/mtd/spi-nor/fsl-quadspi.c:281:31: expected void *[noderef] <asn:2>base [sparse]
drivers/mtd/spi-nor/fsl-quadspi.c:281:31: got void [noderef] <asn:2>*iobase [sparse]
(etc.)
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
There's no reason this can't be a module. Also, give SPI-NOR its own
submenu.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Many of the serial_flash_cmds.h opcodes are duplicated with spi-nor.h.
Let's begin to unify them.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Begin to unify the differences between serial_flash_cmds.h and
spi-nor.h.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
These are also in serial_flash_cmds.h. (FWIW, I didn't know the C
preprocessor allowed redefinitions without warning like this.)
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
I hear that this driver should depend on ARCH_STI, and that "SH" is not
actually a real symbol. At the same time, let's allow compile-testing on
other ARCH'es.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| | |
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Huang Shijie <b32955@freescale.com>
|
| |
| |
| |
| |
| | |
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
serial_flash_cmds.h defines our opcodes a little differently. Let's
borrow its naming, since it's borrowed from the SFDP standard, and it's
more extensible.
This prepares us for merging serial_flash_cmds.h and spi-nor.h opcode
listing.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Huang Shijie <b32955@freescale.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Qualify these with a better namespace, and prepare them for use in more
drivers.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Huang Shijie <b32955@freescale.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Spacing is a little non-standard here. Fix up tabs vs. spaces.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>
Reviewed-by: Marek Vasut <marex@denx.de>
|
| |
| |
| |
| |
| |
| |
| | |
Add the copyright information for spi-nor.c and spi-nor.h.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch adds support for the Macronix MX25L3255E device. Unlike the other
Macronix devices we have seen, this device supports WRITE_1_4_4 at reasonable
frequencies. Rather than masking out WRITE_1_4_4 support altogether, we now
rely on the table parameters to indicate whether or not WRITE_1_4_4 should be
used.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Add Spansion S25FL032P to the list of known devices.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch refactors the fsm_read_status() and fsm_write_status() code to
support 1 or 2 byte operations, with a specified command. This allows us to
remove device/register specific code, such as the N25Q fsm_wrvcr() function.
The 'QE' configuration code is updated accordingly, with minor tweaks to ensure
the register values are only written if actually required. One notable change
in this area is that the 'W25Q_STATUS_QE' bit-field is now defined with respect
to the 'SR2' register, rather than the combined 'SR1+SR2' register which is only
used for write operations.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Update the configuration of the Macronix 'QE' bit, such that
we only set or clear the bit if required.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Support for the Macronix 32-bit addressing scheme was originally developed using
the MX25L25635E device. As is often the case, it was found that the presence of
a "WAIT" instruction was required for the "EN4B/EX4B" FSM Sequence to complete.
(It is known that the SPI FSM Controller makes certain undocumented assumptions
regarding what constitutes a valid sequence.) However, further testing
suggested that a small delay was required after issuing the "EX4B" command;
without this delay, data corruptions were observed, consistent with the device
not being ready to retrieve data. Although the issue was not fully understood,
the workaround of adding a small delay was implemented, while awaiting
clarification from Macronix.
The same behaviour has now been found with a second Macronix device, the
MX25L25655E. However, with this device, it seems that the delay is also
required after the 'EN4B' commands. This discovery has prompted us to revisit
the issue.
Although still not conclusive, further tests have suggested that the issue is
down to the SPI FSM Controller, rather than the Macronix devices. Furthermore,
an alternative workaround has emerged which is to set the WAIT time to
0x00000001, rather then 0x00000000. (Note, the WAIT instruction is used purely
for the purpose of achieving "sequence validity", rather than actually
implementing a delay!)
The issue is now being investigated by the Design and Validation teams. In the
meantime, we implement the alternative workaround, which reduces the effective
delay from 1us to 1ns.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Add Macronix MX25L25655E to the list of known devices.
Signed-off-by: Angus Clark <angus.clark@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fix errors like this:
ERROR: "spi_nor_ids" [drivers/mtd/devices/m25p80.ko] undefined!
ERROR: "spi_nor_scan" [drivers/mtd/devices/m25p80.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
(0) What is the QuadSPI controller?
The QuadSPI(Quad Serial Peripheral Interface) acts as an interface to
one single or two external serial flash devices, each with up to 4
bidirectional data lines.
(1) The QuadSPI controller is driven by the LUT(Look-up Table) registers.
The LUT registers are a look-up-table for sequences of instructions.
A valid sequence consists of four LUT registers.
(2) The definition of the LUT register shows below:
---------------------------------------------------
| INSTR1 | PAD1 | OPRND1 | INSTR0 | PAD0 | OPRND0 |
---------------------------------------------------
There are several types of INSTRx, such as:
CMD : the SPI NOR command.
ADDR : the address for the SPI NOR command.
DUMMY : the dummy cycles needed by the SPI NOR command.
....
There are several types of PADx, such as:
PAD1 : use a singe I/O line.
PAD2 : use two I/O lines.
PAD4 : use quad I/O lines.
....
(3) Test this driver with the JFFS2 and UBIFS:
For jffs2:
-------------
#flash_eraseall /dev/mtd0
#mount -t jffs2 /dev/mtdblock0 tmp
#bonnie++ -d tmp -u 0 -s 10 -r 5
For ubifs:
-------------
#flash_eraseall /dev/mtd0
#ubiattach /dev/ubi_ctrl -m 0
#ubimkvol /dev/ubi0 -N test -m
#mount -t ubifs ubi0:test tmp
#bonnie++ -d tmp -u 0 -s 10 -r 5
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| | |
This patch adds the binding file for Freescale QuadSPI driver.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add the spi_nor_match_id() to find the proper spi_device_id with the
NOR flash's name in the spi_nor_ids table.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Use the new SPI nor framework, and rewrite the m25p80:
(0) remove all the NOR comands.
(1) change the m25p->command to an array.
(2) implement the necessary hooks, such as m25p80_read/m25p80_write.
Tested with the m25p32.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
[Brian: rebased]
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This patch adds the document for the SPI NOR framework.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch cloned most of the m25p80.c. In theory, it adds a new spi-nor layer.
Before this patch, the layer is like:
MTD
------------------------
m25p80
------------------------
spi bus driver
------------------------
SPI NOR chip
After this patch, the layer is like:
MTD
------------------------
spi-nor
------------------------
m25p80
------------------------
spi bus driver
------------------------
SPI NOR chip
With the spi-nor controller driver(Freescale Quadspi), it looks like:
MTD
------------------------
spi-nor
------------------------
fsl-quadspi
------------------------
SPI NOR chip
New APIs:
spi_nor_scan: used to scan a spi-nor flash.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
[Brian: rebased to include additional m25p_ids[] entry]
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The spi_nor{} is cloned from the m25p{}.
The spi_nor{} can be used by both the m25p80 and spi-nor controller.
We also add the spi_nor_xfer_cfg{} which can be used by the two
fundamental primitives: read_xfer/write_xfer.
1) the hooks for spi_nor{}:
@prepare/unpreare: used to do some work before or after the
read/write/erase/lock/unlock.
@read_xfer/write_xfer: We can use these two hooks to code all
the following hooks if the driver tries to implement them
by itself.
@read_reg: used to read the registers, such as read status register,
read configure register.
@write_reg: used to write the registers, such as write enable,
erase sector.
@read_id: read out the ID info.
@wait_till_ready: wait till the NOR becomes ready.
@read: read out the data from the NOR.
@write: write data to the NOR.
@erase: erase a sector of the NOR.
2) Add a new field sst_write_second for the SST NOR write.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch adds a new header :spi-nor.h,
and copies all the SPI NOR commands and relative macros into this new header.
This hearder can be used by the m25p80.c and other spi-nor controller,
such as Freescale's Quadspi.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
More and more chips use the GPMI controller, but these chips may use different
version of the IPs for GPMI and BCH. Different IPs have
different features, such as the BCH's maximum ECC strength:
imx23/imx28 -- the BCH's maximum ECC strength is 20
imx6q -- the BCH's maximum ECC strength is 40
imx6sx -- the BCH's maximum ECC strength is 62
This patch does the following things:
[1] add a new data structure, gpmi_devdata{}, to store the information for
each IP. Besides the IP version, we store the following information:
<1> BCH's maximum ECC strength.
<2> the maximum chain delay in ns used by the EDO mode.
but we may add more information in future.
[2] add the gpmi_devdata_imx{23|28|6q} to replace the gpmi_ids.
[3] simplify the code by using the ECC strength from gpmi_devdata, such as
gpmi_check_ecc() and legacy_set_geometry();
[4] use the maximum chain delay to initialize the EDO mode,
see gpmi_compute_edo_timing().
[5] rewrite the macros, such GPMI_IS_MX{23|28|6Q}.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| | |
Add line feed to pr_err.
Signed-off-by: Raphaël Poggi <poggi.raph@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Sub page write doesn't work because of hw issue in controller found on
Keystone SOCs. AEMIF controller is also used on DaVinci SOCs which
don't seems to have any issue. So add "ti,keysone-nand" compatible
to nand driver in order to set NAND_NO_SUBPAGE_WRITE option.
Cc: Warner Losh <imp@bsdimp.com>
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
writeback buffer
mtd_blkdevs is device with volatile cache (writeback buffer), so it should support
REQ_FLUSH to do explicit flush.
Without this patch 'sync' does not guarantee that writeback buffer will be flushed
on disk in case of power off, e.g.:
$ cp some_file /mnt
$ sync
### POWER OFF
In case of this sequence writeback buffer will not be flushed on disk.
This patch fixes this behaviour and explicitly reports to block layer that flush
requests are being supported.
Signed-off-by: Roman Peniaev <r.peniaev@gmail.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: linux-mtd@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Now that the index variable is correctly set earlier in this function
we can use it in other places that compute the same thing too.
Signed-off-by: Ron Lee <ron@debian.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 2c9f2365 (mtd: nand: omap: ecc.calculate: merge omap3_calculate_ecc_bch4
in omap_calculate_ecc_bch) introduced minor compile warning
"‘erased_sector_bitflips’ defined but not used [-Wunused-function]" when
compiling without CONFIG_MTD_NAND_OMAP_BCH. Move function
erased_sector_bitflips() into the same ifdef section as the only caller.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Pekon Gupta <pekon@ti.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some versions of gcc even warn about it:
mm/shmem.c: In function ‘shmem_file_aio_read’:
mm/shmem.c:1414: warning: ‘error’ may be used uninitialized in this function
If the loop is aborted during the first iteration by one of the two
first break statements, error will be uninitialized.
Introduced by commit 6e58e79db8a1 ("introduce copy_page_to_iter, kill
loop over iovec in generic_file_aio_read()").
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On 32 bit, size_t is "unsigned int", not "unsigned long", causing the
following warning when comparing with PAGE_SIZE, which is always "unsigned
long":
fs/cifs/file.c: In function ‘cifs_readdata_to_iov’:
fs/cifs/file.c:2757: warning: comparison of distinct pointer types lacks a cast
Introduced by commit 7f25bba819a3 ("cifs_iovec_read: keep iov_iter
between the calls of cifs_readdata_to_iov()"), which changed the
signedness of "remaining" and the code from min_t() to min().
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux
Pull slab changes from Pekka Enberg:
"The biggest change is byte-sized freelist indices which reduces slab
freelist memory usage:
https://lkml.org/lkml/2013/12/2/64"
* 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux:
mm: slab/slub: use page->list consistently instead of page->lru
mm/slab.c: cleanup outdated comments and unify variables naming
slab: fix wrongly used macro
slub: fix high order page allocation problem with __GFP_NOFAIL
slab: Make allocations with GFP_ZERO slightly more efficient
slab: make more slab management structure off the slab
slab: introduce byte sized index for the freelist of a slab
slab: restrict the number of objects in a slab
slab: introduce helper functions to get/set free object
slab: factor out calculate nr objects in cache_estimate
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
'struct page' has two list_head fields: 'lru' and 'list'. Conveniently,
they are unioned together. This means that code can use them
interchangably, which gets horribly confusing like with this nugget from
slab.c:
> list_del(&page->lru);
> if (page->active == cachep->num)
> list_add(&page->list, &n->slabs_full);
This patch makes the slab and slub code use page->lru universally instead
of mixing ->list and ->lru.
So, the new rule is: page->lru is what the you use if you want to keep
your page on a list. Don't like the fact that it's not called ->list?
Too bad.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Pekka Enberg <penberg@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
As time goes, the code changes a lot, and this leads to that
some old-days comments scatter around , which instead of faciliating
understanding, but make more confusion. So this patch cleans up them.
Also, this patch unifies some variables naming.
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Jianyu Zhan <nasa4836@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
commit 'slab: restrict the number of objects in a slab' uses
__builtin_constant_p() on #if macro. It is wrong usage of builtin
function, but it is compiled on x86 without any problem, so I can't
find it before 0 day build system find it.
This commit fixes the situation by using KMALLOC_MIN_SIZE, instead of
KMALLOC_SHIFT_LOW. KMALLOC_SHIFT_LOW is parsed to ilog2() on some
architecture and this ilog2() uses __builtin_constant_p() and results in
the problem. This problem would disappear by using KMALLOC_MIN_SIZE,
since it is just constant.
Tested-by: David Rientjes <rientjes@google.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
SLUB already try to allocate high order page with clearing __GFP_NOFAIL.
But, when allocating shadow page for kmemcheck, it missed clearing
the flag. This trigger WARN_ON_ONCE() reported by Christian Casteyde.
https://bugzilla.kernel.org/show_bug.cgi?id=65991
https://lkml.org/lkml/2013/12/3/764
This patch fix this situation by using same allocation flag as original
allocation.
Reported-by: Christian Casteyde <casteyde.christian@free.fr>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Use the likely mechanism already around valid
pointer tests to better choose when to memset
to 0 allocations with __GFP_ZERO
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Now, the size of the freelist for the slab management diminish,
so that the on-slab management structure can waste large space
if the object of the slab is large.
Consider a 128 byte sized slab. If on-slab is used, 31 objects can be
in the slab. The size of the freelist for this case would be 31 bytes
so that 97 bytes, that is, more than 75% of object size, are wasted.
In a 64 byte sized slab case, no space is wasted if we use on-slab.
So set off-slab determining constraint to 128 bytes.
Acked-by: Christoph Lameter <cl@linux.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently, the freelist of a slab consist of unsigned int sized indexes.
Since most of slabs have less number of objects than 256, large sized
indexes is needless. For example, consider the minimum kmalloc slab. It's
object size is 32 byte and it would consist of one page, so 256 indexes
through byte sized index are enough to contain all possible indexes.
There can be some slabs whose object size is 8 byte. We cannot handle
this case with byte sized index, so we need to restrict minimum
object size. Since these slabs are not major, wasted memory from these
slabs would be negligible.
Some architectures' page size isn't 4096 bytes and rather larger than
4096 bytes (One example is 64KB page size on PPC or IA64) so that
byte sized index doesn't fit to them. In this case, we will use
two bytes sized index.
Below is some number for this patch.
* Before *
kmalloc-512 525 640 512 8 1 : tunables 54 27 0 : slabdata 80 80 0
kmalloc-256 210 210 256 15 1 : tunables 120 60 0 : slabdata 14 14 0
kmalloc-192 1016 1040 192 20 1 : tunables 120 60 0 : slabdata 52 52 0
kmalloc-96 560 620 128 31 1 : tunables 120 60 0 : slabdata 20 20 0
kmalloc-64 2148 2280 64 60 1 : tunables 120 60 0 : slabdata 38 38 0
kmalloc-128 647 682 128 31 1 : tunables 120 60 0 : slabdata 22 22 0
kmalloc-32 11360 11413 32 113 1 : tunables 120 60 0 : slabdata 101 101 0
kmem_cache 197 200 192 20 1 : tunables 120 60 0 : slabdata 10 10 0
* After *
kmalloc-512 521 648 512 8 1 : tunables 54 27 0 : slabdata 81 81 0
kmalloc-256 208 208 256 16 1 : tunables 120 60 0 : slabdata 13 13 0
kmalloc-192 1029 1029 192 21 1 : tunables 120 60 0 : slabdata 49 49 0
kmalloc-96 529 589 128 31 1 : tunables 120 60 0 : slabdata 19 19 0
kmalloc-64 2142 2142 64 63 1 : tunables 120 60 0 : slabdata 34 34 0
kmalloc-128 660 682 128 31 1 : tunables 120 60 0 : slabdata 22 22 0
kmalloc-32 11716 11780 32 124 1 : tunables 120 60 0 : slabdata 95 95 0
kmem_cache 197 210 192 21 1 : tunables 120 60 0 : slabdata 10 10 0
kmem_caches consisting of objects less than or equal to 256 byte have
one or more objects than before. In the case of kmalloc-32, we have 11 more
objects, so 352 bytes (11 * 32) are saved and this is roughly 9% saving of
memory. Of couse, this percentage decreases as the number of objects
in a slab decreases.
Here are the performance results on my 4 cpus machine.
* Before *
Performance counter stats for 'perf bench sched messaging -g 50 -l 1000' (10 runs):
229,945,138 cache-misses ( +- 0.23% )
11.627897174 seconds time elapsed ( +- 0.14% )
* After *
Performance counter stats for 'perf bench sched messaging -g 50 -l 1000' (10 runs):
218,640,472 cache-misses ( +- 0.42% )
11.504999837 seconds time elapsed ( +- 0.21% )
cache-misses are reduced by this patchset, roughly 5%.
And elapsed times are improved by 1%.
Acked-by: Christoph Lameter <cl@linux.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
|