The driver uses plat_nand. As the platform_device is loaded from DT, we need
to lookup the node and attach our xway specific "struct platform_nand_data"
to it.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
If plat_nand loads and the platform_data is not properly set it will segfault.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
CONFIG_MTD_NAND_VERIFY_WRITE was killed recently, so remove it from
defconfigs as well.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Just as Artem suggested:
"Both UBI and JFFS2 are able to read verify what they wrote already.
There are also MTD tests which do this verification. So I think there
is no reason to keep this in the NAND layer, let alone wasting RAM in
the driver to support this feature. Besides, it does not work for sub-pages
and many drivers have it broken. It hurts more than it provides benefits."
So kill MTD_NAND_VERIFY_WRITE entirely.
Signed-off-by: Huang Shijie <shijie8@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The manufacturer datasheet can be found on the Micron website,
under the name n25q_256mb_3v_65nm.pdf:
http://www.micron.com/search?source=ps&q=n25q_256mb_3v_65nm
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Removes disabled printk (which should be dev_dbg these days) as well
as #if 0 blocks (which are trivial to reimplement if ever needed) to
meet basic CodingStyle guidelines.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
There is new format specified that helps to dump small buffers. It makes the
code simpler and nicer.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Jiandong Zheng <jdzheng@broadcom.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
add OF support for the davinci nand controller.
Signed-off-by: Heiko Schocher <hs@denx.de>
Acked-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch fixes regression introduced by
"8bdc81c jffs2: get rid of jffs2_sync_super". We submit a delayed work in order
to make sure the write-buffer is synchronized at some point. But we do not
flush it when we unmount, which causes an oops when we unmount the file-system
and then the delayed work is executed.
This patch fixes the issue by adding a "cancel_delayed_work_sync()" infocation
in the '->sync_fs()' handler. This will make sure the delayed work is canceled
on sync, unmount and re-mount. And because VFS always callse 'sync_fs()' before
unmounting or remounting, this fixes the issue.
Reported-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Cc: stable@vger.kernel.org [3.5+]
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Tested-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The uclinux.c map driver has traditionally been used only on non-MMU based
systems. But there is no fundamental reason it can't be used on systems
running with virtual memory.
Some ColdFire CPU based systems now have full paged MMU hardware and can use
the uclinux.c mapping driver, so making the uclinux.c driver configuration
depend on !CONFIG_MMU doesn't make sense now. Allow the CONFIG_MTD_UCLINUX
option to be enabled if CONFIG_COLDFIRE is enabled. (I have chosen not to
just more generally allow uclinux.c for any MMU type to keep this option
hidden for most systems that are not interested in setting it).
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The uclinux.c mapping driver uses ioremap_nocache() to map its physical
mapping address to a system virtual address. Problem is that the region
it is mapping is not device memory. It is ordinary system RAM. On most
non-MMU systems this doesn't matter, and the mapping is always a 1:1
translation of the address. On paged memory systems on some architectures
the page table mappings are not compatible between normal RAM and device
memory.
If we want to use the uclinux.c mapping driver on real MMU enabled systems
we should be using the kernel virtual address that the mapping is at. For
architectures that support the traditional initrd they use phys_to_virt or
__va to convert the physical start initrd address to a kernel usable virtual
address. The uclinux filesystem mapping is even more restrictive than the
typical initrd, it always follows the kernels own bss section (so always in
directly mapped memory). Therefore we can use the usual phys_to_virt to
translate the physical start address to a virtual address.
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fixes the following checkpatch warnings:
WARNING: line over 80 characters
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fixes checkpatch warnings and errors related to whitespaces.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fixes the following checkpatch errors:
ERROR: do not initialise statics to 0 or NULL
+static int hardware_ecc = 0;
ERROR: do not initialise statics to 0 or NULL
+static const int clock_stop = 0;
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fixes the following checkpatch warning:
WARNING: Use #include <linux/io.h> instead of <asm/io.h>
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This makes the code simpler by eliminating module_init() and
module_exit().
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Include IMX6 in the list of supported SoCs.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Because we can have a single kernel to support multiple machines, we
need to make loading specific drivers for the target platform only.
For this, driver is converted to the platform driver.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Update driver autcpu12-nvram.c so it compiles; map_read32/map_write32
no longer exist in the kernel so the driver is totally broken.
Additionally, map_info name passed to simple_map_init is incorrect.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Cc: stable@vger.kernel.org
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The NAND_CHIPOPTIONS_MSK has limited utility and is causing real bugs. It
silently masks off at least one flag that might be set by the driver
(NAND_NO_SUBPAGE_WRITE). This breaks the GPMI NAND driver and possibly
others.
Really, as long as driver writers exercise a small amount of care with
NAND_* options, this mask is not necessary at all; it was only here to
prevent certain options from accidentally being set by the driver. But the
original thought turns out to be a bad idea occasionally. Thus, kill it.
Note, this patch fixes some major gpmi-nand breakage.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Tested-by: Huang Shijie <shijie8@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
without this the gpio will not be muxed as a gpio by the current custom pinmux
or later by the pinctrl
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch does two closely related things:
(1) Currently the ecc.read_page() method does not fill the nand->oob_poi buffer
with the oob data, but instead reads oob into a local buffer. Fix this by
filling the oob_poi buffer instead of a local buffer. The 'oob_required'
argument is quietly ignored; the device must always read oob after the page
data, and it is presumed that there's no harm in filling oob_poi, even when not
explicitly requested.
(2) Always read oob from the device in ecc.read_oob(), instead of copying it
from a local buffer under some circumstances.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
drivers/mtd/devices/spear_smi.c: In function 'spear_smi_probe':
drivers/mtd/devices/spear_smi.c:984:2: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
dev_get_platdata returns a pointer, so the failure value would be NULL
rather than a negative integer.
The semantic match that finds this problem is: (http://coccinelle.lip6.fr/)
// <smpl>
@@
expression x,e;
statement S1,S2;
@@
*x = dev_get_platdata(...)
... when != x = e
*if (x < 0) S1 else S2
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Stefan Roese <sr@denx.de>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch makes the MLC NAND driver independent of the single AMBA DMA engine
driver by using the platform data provided dma_filter callback.
Signed-off-by: Roland Stigge <stigge@antcom.de>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch makes the SLC NAND driver independent of the single AMBA DMA engine
driver by using the platform data provided dma_filter callback.
Signed-off-by: Roland Stigge <stigge@antcom.de>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
None of these scanning functions use MTD_OPS_RAW mode any more, so there's
really nothing 'raw' about them. Rename them to (hopefully) make the code
a little clearer.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
scan_read_raw_oob() is used in only in places where the MTD_OPS_PLACE_OOB mode
is preferable to MTD_OPS_RAW mode, so use MTD_OPS_PLACE_OOB instead.
MTD_OPS_PLACE_OOB provides the same functionality with the potential[1] added
bonus of error correction.
This brings scan_block_full() in line with scan_block_fast() so that they
both read bad block markers with MTD_OPS_PLACE_OOB. This can help in
preventing 0xff markers (in good blocks) from being interpreted as bad
block indicators in the presence of a single bitflip.
Note that ECC error codes (EUCLEAN or EBADMSG) are already silently
ignored in all users of scan_read_raw_oob().
[1] Few drivers perform proper error correction on OOB data. In those
cases, the use of MTD_OPS_RAW vs. MTD_OPS_PLACE_OOB is not
significant.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Some nand_bbt code can be shortened by using memcmp() and memchr_inv().
As an added bonus, there is a possible performance benefit.
Borrowed some code from Akinobu Mita.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The return codes for read_abs_bbts() and search_read_bbts() are always
non-zero, and so don't have much meaning. Just remove them.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
mtd_read_oob() has some unexpected similarities to mtd_read(). For
instance, when ops->datbuf != NULL, nand_base.c might return max_bitflips;
however, when ops->datbuf == NULL, nand_base's code potentially could
return -EUCLEAN (no in-tree drivers do this yet). In any case where the
driver might return max_bitflips, we should translate this into an
appropriate return code using the bitflip_threshold.
Essentially, mtd_read_oob() duplicates the logic from mtd_read().
This prevents users of mtd_read_oob() from receiving a positive return
value (i.e., from max_bitflips) and interpreting it as an unknown error.
Artem: amend comments.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Due to the implementation of the following loop at the end
of jedec_probe():
776 for (tmp = 0; tmp < ARRAY_SIZE(m25p_ids) - 1; tmp++) {
777 info = (void *)m25p_ids[tmp].driver_data;
778 if (info->jedec_id == jedec) {
779 if (info->ext_id != 0 && info->ext_id != ext_jedec)
780 continue;
781 return &m25p_ids[tmp];
782 }
783 }
In particular line 779 in the above numbering, the chips with ext_id != 0 must
be ordered first in the list of chips (m25p_ids[]).
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The gpmi nand driver may needs several clocks(MX6Q needs five clocks).
In the old clock framework, all these clocks are chained together,
all you need is to manipulate the first clock.
But the kernel uses the common clk framework now, which forces us to
get the clocks one by one. When we use them, we have to enable them
one by one too.
Signed-off-by: Huang Shijie <shijie8@gmail.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Check the new oob_required flag and only copy the OOB data to the internal
buffer if needed.
Signed-off-by: Bastian Hecht <hechtb@gmail.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Elements have been copied "manually" in a loop. Better use memcpy().
Signed-off-by: Bastian Hecht <hechtb@gmail.com>
Reviewed-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
filter noisy error sources at a fine grained level based on platform
specific knowledge.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=n9qI
-----END PGP SIGNATURE-----
Merge tag 'please-pull-naveen' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras into x86/mce
Pull MCE updates from Tony Luck:
"Option to let the bios set per-bank CMCI thresholds so they can
filter noisy error sources at a fine grained level based on platform
specific knowledge."
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This bug was introduced in commit 4052147 ("mm, slab: Match SLAB and
SLUB kmem_cache_alloc_xxx_trace() prototype").
Cc: David Rientjes <rientjes@google.com>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Ezequiel Garcia <elezegarcia@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Jan Kara have spotted interesting issue:
There are potential data corruption issue with direct IO overwrites
racing with truncate:
Like:
dio write truncate_task
->ext4_ext_direct_IO
->overwrite == 1
->down_read(&EXT4_I(inode)->i_data_sem);
->mutex_unlock(&inode->i_mutex);
->ext4_setattr()
->inode_dio_wait()
->truncate_setsize()
->ext4_truncate()
->down_write(&EXT4_I(inode)->i_data_sem);
->__blockdev_direct_IO
->ext4_get_block
->submit_io()
->up_read(&EXT4_I(inode)->i_data_sem);
# truncate data blocks, allocate them to
# other inode - bad stuff happens because
# dio is still in flight.
In order to serialize with truncate dio worker should grab extra i_dio_count
reference before drop i_mutex.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If we have enough aggressive DIO readers, truncate and other dio
waiters will wait forever inside inode_dio_wait(). It is reasonable
to disable nonlock DIO read optimization during truncate.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Current serialization will works only for DIO which holds
i_mutex, but nonlocked DIO following race is possible:
dio_nolock_read_task truncate_task
->ext4_setattr()
->inode_dio_wait()
->ext4_ext_direct_IO
->ext4_ind_direct_IO
->__blockdev_direct_IO
->ext4_get_block
->truncate_setsize()
->ext4_truncate()
#alloc truncated blocks
#to other inode
->submit_io()
#INFORMATION LEAK
In order to serialize with unlocked DIO reads we have to
rearrange wait sequence
1) update i_size first
2) if i_size about to be reduced wait for outstanding DIO requests
3) and only after that truncate inode blocks
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Inode's block defrag and ext4_change_inode_journal_flag() may
affect nonlocked DIO reads result, so proper synchronization
required.
- Add missed inode_dio_wait() calls where appropriate
- Check inode state under extra i_dio_count reference.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Current unwritten extent conversion state-machine is very fuzzy.
- For unknown reason it performs conversion under i_mutex. What for?
My diagnosis:
We already protect extent tree with i_data_sem, truncate and punch_hole
should wait for DIO, so the only data we have to protect is end_io->flags
modification, but only flush_completed_IO and end_io_work modified this
flags and we can serialize them via i_completed_io_lock.
Currently all these games with mutex_trylock result in the following deadlock
truncate: kworker:
ext4_setattr ext4_end_io_work
mutex_lock(i_mutex)
inode_dio_wait(inode) ->BLOCK
DEADLOCK<- mutex_trylock()
inode_dio_done()
#TEST_CASE1_BEGIN
MNT=/mnt_scrach
unlink $MNT/file
fallocate -l $((1024*1024*1024)) $MNT/file
aio-stress -I 100000 -O -s 100m -n -t 1 -c 10 -o 2 -o 3 $MNT/file
sleep 2
truncate -s 0 $MNT/file
#TEST_CASE1_END
Or use 286's xfstests https://github.com/dmonakhov/xfstests/blob/devel/286
This patch makes state machine simple and clean:
(1) xxx_end_io schedule final extent conversion simply by calling
ext4_add_complete_io(), which append it to ei->i_completed_io_list
NOTE1: because of (2A) work should be queued only if
->i_completed_io_list was empty, otherwise the work is scheduled already.
(2) ext4_flush_completed_IO is responsible for handling all pending
end_io from ei->i_completed_io_list
Flushing sequence consists of following stages:
A) LOCKED: Atomically drain completed_io_list to local_list
B) Perform extents conversion
C) LOCKED: move converted io's to to_free list for final deletion
This logic depends on context which we was called from.
D) Final end_io context destruction
NOTE1: i_mutex is no longer required because end_io->flags modification
is protected by ei->ext4_complete_io_lock
Full list of changes:
- Move all completion end_io related routines to page-io.c in order to improve
logic locality
- Move open coded logic from various xx_end_xx routines to ext4_add_complete_io()
- remove EXT4_IO_END_FSYNC
- Improve SMP scalability by removing useless i_mutex which does not
protect io->flags anymore.
- Reduce lock contention on i_completed_io_lock by optimizing list walk.
- Rename ext4_end_io_nolock to end4_end_io and make it static
- Check flush completion status to ext4_ext_punch_hole(). Because it is
not good idea to punch blocks from corrupted inode.
Changes since V3 (in request to Jan's comments):
Fall back to active flush_completed_IO() approach in order to prevent
performance issues with nolocked DIO reads.
Changes since V2:
Fix use-after-free caused by race truncate vs end_io_work
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_set_io_unwritten_flag() will increment i_unwritten counter, so
once we mark end_io with EXT4_END_IO_UNWRITTEN we have to revert it back
on error path.
- add missed error checks to prevent counter leakage
- ext4_end_io_nolock() will clear EXT4_END_IO_UNWRITTEN flag to signal
that conversion finished.
- add BUG_ON to ext4_free_end_io() to prevent similar leakage in future.
Visible effect of this bug is that unaligned aio_stress may deadlock
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
AIO/DIO prefix is wrong because it account unwritten extents which
also may be scheduled from buffered write endio
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Generic inode has unused i_private pointer which may be used as cur_aio_dio
storage.
TODO: If cur_aio_dio will be passed as an argument to get_block_t this allow
to have concurent AIO_DIO requests.
Reviewed-by: Zheng Liu <wenqing.lz@taobao.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
