iov_iter retry fixes
Jens Axboe
maintainerHi Linus,
This adds a helper to save/restore iov_iter state, and modifies io_uring to use it. After that is done, we can kill the iter->truncated addition that we added for this release. The io_uring change is being overly cautious with the save/restore/advance, but better safe than sorry and we can always improve that and reduce the overhead if it proves to be of concern. The only case to be worried about in this regard is huge IO, where iteration can take a while to iterate segments.
I spent some time writing test cases, and expanded the coverage quite a bit from the last posting of this. liburing carries this regression test case now:
https://git.kernel.dk/cgit/liburing/tree/test/file-verify.c
which exercises all of this. It now also supports provided buffers, and explicitly tests for end-of-file/device truncation as well.
On top of that, Pavel sanitized the IOPOLL retry path to follow the exact same pattern as normal IO.
Please pull!
The following changes since commit d6c338a741295c04ed84679153448b2fffd2c9cf:
Merge tag 'for-linus-5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml (2021-09-09 13:45:26 -0700)
are available in the Git repository at:
git://git.kernel.dk/linux-block.git tags/iov_iter.3-5.15-2021-09-17
for you to fetch changes up to b66ceaf324b394428bb47054140ddf03d8172e64:
io_uring: move iopoll reissue into regular IO path (2021-09-15 09:22:35 -0600)
----------------------------------------------------------------
iov_iter.3-5.15-2021-09-17
----------------------------------------------------------------
Jens Axboe (3):
iov_iter: add helper to save iov_iter state
io_uring: use iov_iter state save/restore helpers Revert "iov_iter: track truncated size"
Pavel Begunkov (1):
io_uring: move iopoll reissue into regular IO path
fs/io_uring.c | 116 ++++++++++++++++++++++++++++++++++------------------ include/linux/uio.h | 21 +++++++--- lib/iov_iter.c | 36 ++++++++++++++++ 3 files changed, 128 insertions(+), 45 deletions(-)
Signature · 1 line
Jens Axboe
Quoted 14 lines · Jens Axboe
Hi Linus,
This adds a helper to save/restore
iov_iterstate, and modifiesio_uringto use it. After that is done, we can kill theiter->truncatedaddition that we added for this release. Theio_uringchange is being overly cautious with the save/restore/advance, but better safe than sorry and we can always improve that and reduce the overhead if it proves to be of concern. The only case to be worried about in this regard is huge IO, where iteration can take a while to iterate segments.I spent some time writing test cases, and expanded the coverage quite a bit from the last posting of this. liburing carries this regression test case now:
Quoted 7 lines
https://git.kernel.dk/cgit/liburing/tree/test/file-verify.c
which exercises all of this. It now also supports provided buffers, and explicitly tests for end-of-file/device truncation as well.
On top of that, Pavel sanitized the IOPOLL retry path to follow the exact same pattern as normal IO.
I can live with that; I do have problems with io-uring and its interactions with iov_iter, but those are mostly independent from the stuff you are doing here.
Still digging through your FSM from hell... ;-/
Jens Axboe
maintainerQuoted 20 lines · Al Viro
Quoted 19 lines · Jens Axboe
Hi Linus,
This adds a helper to save/restore
iov_iterstate, and modifiesio_uringto use it. After that is done, we can kill theiter->truncatedaddition that we added for this release. Theio_uringchange is being overly cautious with the save/restore/advance, but better safe than sorry and we can always improve that and reduce the overhead if it proves to be of concern. The only case to be worried about in this regard is huge IO, where iteration can take a while to iterate segments.I spent some time writing test cases, and expanded the coverage quite a bit from the last posting of this. liburing carries this regression test case now:
https://git.kernel.dk/cgit/liburing/tree/test/file-verify.c
which exercises all of this. It now also supports provided buffers, and explicitly tests for end-of-file/device truncation as well.
Quoted 6 lines
Quoted 2 lines
On top of that, Pavel sanitized the IOPOLL retry path to follow the exact same pattern as normal IO.
I can live with that; I do have problems with io-uring and its interactions with
iov_iter, but those are mostly independent from the stuff you are doing here.
I actually think we can end up cleaning up some of those bits post this, it'll definitely make it easier by having clean save/restore states.
Signature · 1 line
Jens Axboe
The pull request you sent on Fri, 17 Sep 2021 08:44:32 -0600:
The pull request you sent on Fri, 17 Sep 2021 08:44:32 -0600:
Quoted 1 line
git://git.kernel.dk/linux-block.gittags/iov_iter.3-5.15-2021-09-17
has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/ddf21bd8ab984ccaa924f090fc7f515bb6d51414
Thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/prtracker.html