similar to: [PATCH] ocfs2: avoid direct write if we fall back to buffered v2

Signed-off-by: Dave Kleikamp <dave.kleikamp at oracle.com> Acked-by: Joel Becker <jlbec at evilplan.org> Cc: Zach Brown <zab at zabbo.net> Cc: Mark Fasheh <mfasheh at suse.com> Cc: ocfs2-devel at oss.oracle.com --- fs/ocfs2/aops.h | 2 +- fs/ocfs2/file.c | 55 ++++++++++++++++++++++---------------------------- fs/ocfs2/ocfs2_trace.h | 6 +++--- 3 files

From: Zach Brown <zab@zabbo.net> ocfs2''s .aio_read and .aio_write methods are changed to take iov_iter and pass it to generic functions. Wrappers are made to pack the iovecs into iters and call these new functions. Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com> Cc: Zach Brown <zab@zabbo.net> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker

generic_write_checks() expects count to be initialized to the size of the write. Writes to files open with O_DIRECT|O_LARGEFILE write 0 bytes because count in uninitialized. Signed-off-by: Goldwyn Rodrigues <rgoldwyn at suse.de> diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index a5887df..5b9c8af 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1841,6 +1841,7 @@ relock: if (ret)

V1->V2: Check to see if our current ppos is >= i_size after a short DIO read, just in case it was actually a short read and we need to just return. This is similar to what already happens in the write case. If we have a short read while doing O_DIRECT, instead of just returning, fallthrough and try to read the rest via buffered IO. BTRFS needs this because if we encounter a compressed or

I have the exact same issue on several Sunfire v20z (hardware LSI RAID controler). Under high IO load, the raid controler starts complaining with disk errors : [163442.483878] mptbase: ioc0: LogInfo(0x11070000): F/W: DMA Error [163445.172333] sd 2:0:0:0: [sda] Unhandled error code [163445.172338] sd 2:0:0:0: [sda] Result: hostbyte=DID_SOFT_ERROR driverbyte=DRIVER_OK [163445.172345] sd 2:0:0:0:

Adding linux-btrfs to Cc. Am Freitag, 27. Januar 2012 schrieb Jeff Moyer: > Hi, Hi, > As it stands, generic_file_aio_write will call into generic_write_sync > when -EIOCBQUEUED is returned from __generic_file_aio_write. > EIOCBQUEUED indicates that an I/O was submitted but NOT completed. > Thus, we will flush the disk cache, potentially before the write(s) > even make it to

Having done a current checkout, creating a new FS and running iozone [1] on it results in an oops [2]. remove_suid is called, accessing offset 14 of a NULL pointer. Let me know if you''d like me to test any fix, do further debugging or get more information. Thanks, Daniel --- [1] # mkfs.btrfs /dev/sda4 # mount /dev/sda4 /mnt /mnt# iozone -a . --- [2] [ 899.118926] BUG: unable to

V4 changes: add Acked-by: Joel Becker <jlbec at evilplan.org> V3 changes: - add Cc: stable at vger.kernel.org in the patch header to align with stable rules - add Acked-by: Jeff Moyer <jmoyer at redhat.com> V2 changes: - update the patch header of the first patch to make it more clear. This patch list fixes an issue about ocfs2 aio/dio write process hang. The call trace is like

V1->V2: -Changed dio_end_io to EXPORT_SYMBOL_GPL -Removed the own_submit blockdev dio helper -Removed the boundary change V2->V3 -Made it so we keep track of what the current logical offset in the file we have a BIO setup for so we can pass it into the submit_io hook. Because BTRFS can do RAID and such, we need our own submit hook so we can setup the bio''s in the correct fashion,

Found 15 patches (out of 162) that appeared relevant to ocfs2 1.4. Please review. Sunil

Add parenthesis to wrap the check for O_DIRECT. Signed-off-by: Tao Ma <tao.ma at oracle.com> --- fs/ocfs2/file.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 06ccf6a..b2ca980 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2013,8 +2013,8 @@ out_dio: /* buffered aio wouldn't have proper lock coverage today

use i_size_read() instead of accessing inode->i_size directly. Signed-off-by: Wengang Wang <wen.gang.wang at oracle.com> -- diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index b2c52b3..4491851 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -1920,7 +1920,7 @@ int ocfs2_write_end_nolock(struct address_space *mapping, out_write_size: pos += copied; - if (pos >

On Mon, Feb 06, 2023 at 07:00:35AM +0000, Arseniy Krasnov wrote: >This adds main logic of MSG_ZEROCOPY flag processing for packet >creation. When this flag is set and user's iov iterator fits for >zerocopy transmission, call 'get_user_pages()' and add returned >pages to the newly created skb. > >Signed-off-by: Arseniy Krasnov <AVKrasnov at sberdevices.ru>

Apparently it is ok to submit a read to an IDE device with the same target page for different offsets. This is what Windows does under qemu. The problem is under DIO we expect them to be different buffers for checksumming reasons, and so this sort of thing will result in checksum errors, when in reality the file is fine. So when reading, check to make sure that all iov bases are different, and

V1->V2: Check to see if our current ppos is >= i_size after a short DIO read, just in case it was actually a short read and we need to just return. This is similar to what already happens in the write case. If we have a short read while doing O_DIRECT, instead of just returning, fallthrough and try to read the rest via buffered IO. BTRFS needs this because if we encounter a compressed or

V1->V2: Check to see if our current ppos is >= i_size after a short DIO read, just in case it was actually a short read and we need to just return. This is similar to what already happens in the write case. If we have a short read while doing O_DIRECT, instead of just returning, fallthrough and try to read the rest via buffered IO. BTRFS needs this because if we encounter a compressed or

This is my still-working-on-it code for btrfs directIO read. I''m posting it so people can see the progress being made on the project and can take an early shot at telling me this is just a bad idea and I''m crazy if they want to, or point out where I made some stupid mistake with btrfs core functions. The code is not complete and *NOT* ready for review or testing. I looked at

This idea is from ext4. By this patch, we can make the dio write parallel, and improve the performance. We needn''t worry about the race between dio write and truncate, because the truncate need wait untill all the dio write end. And we also needn''t worry about the race between dio write and punch hole, because we have extent lock to protect our operation. I ran fio to test the

i_alloc_sem has always been a bit of an odd "lock". It''s the only remaining rw_semaphore that can be released by a different thread than the one that locked it, and it''s use case in the core direct I/O code is more like a counter given that the writers already have external serialization. This series removes it in favour of a simpler counter scheme, thus getting rid

Hi, this is about an issue newer kernels show, bysplitting direct I/O requests into 4k pieces to directly merge them in the Block Device Layer afterwards. If anyone is interested in own tests just use a simple command like dd if=/mnt/test/test-dd1 of=/dev/null iflag=direct bs=64k count=1 in combination with blktrace. The following patch is more a proposal for discussion than a solution, well