This is a note to let you know that I've just added the patch titled iomap: Support partial direct I/O on user copy failures to the 5.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: iomap-support-partial-direct-i-o-on-user-copy-failures.patch and it can be found in the queue-5.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From foo@baz Fri Apr 29 11:07:48 AM CEST 2022 From: Anand Jain <anand.jain@xxxxxxxxxx> Date: Fri, 15 Apr 2022 06:28:49 +0800 Subject: iomap: Support partial direct I/O on user copy failures To: stable@xxxxxxxxxxxxxxx Cc: linux-btrfs@xxxxxxxxxxxxxxx, Andreas Gruenbacher <agruenba@xxxxxxxxxx>, "Darrick J . Wong" <djwong@xxxxxxxxxx>, Anand Jain <anand.jain@xxxxxxxxxx> Message-ID: <a85564f9b06b5bae198a27c7f60cd02b39c2ce79.1649951733.git.anand.jain@xxxxxxxxxx> From: Andreas Gruenbacher <agruenba@xxxxxxxxxx> commit 97308f8b0d867e9ef59528cd97f0db55ffdf5651 upstream In iomap_dio_rw, when iomap_apply returns an -EFAULT error and the IOMAP_DIO_PARTIAL flag is set, complete the request synchronously and return a partial result. This allows the caller to deal with the page fault and retry the remainder of the request. Signed-off-by: Andreas Gruenbacher <agruenba@xxxxxxxxxx> Reviewed-by: Darrick J. Wong <djwong@xxxxxxxxxx> Signed-off-by: Anand Jain <anand.jain@xxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- fs/iomap/direct-io.c | 6 ++++++ include/linux/iomap.h | 7 +++++++ 2 files changed, 13 insertions(+) --- a/fs/iomap/direct-io.c +++ b/fs/iomap/direct-io.c @@ -581,6 +581,12 @@ __iomap_dio_rw(struct kiocb *iocb, struc if (iov_iter_rw(iter) == READ && iomi.pos >= dio->i_size) iov_iter_revert(iter, iomi.pos - dio->i_size); + if (ret == -EFAULT && dio->size && (dio_flags & IOMAP_DIO_PARTIAL)) { + if (!(iocb->ki_flags & IOCB_NOWAIT)) + wait_for_completion = true; + ret = 0; + } + /* magic error code to fall back to buffered I/O */ if (ret == -ENOTBLK) { wait_for_completion = true; --- a/include/linux/iomap.h +++ b/include/linux/iomap.h @@ -330,6 +330,13 @@ struct iomap_dio_ops { */ #define IOMAP_DIO_OVERWRITE_ONLY (1 << 1) +/* + * When a page fault occurs, return a partial synchronous result and allow + * the caller to retry the rest of the operation after dealing with the page + * fault. + */ +#define IOMAP_DIO_PARTIAL (1 << 2) + ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, const struct iomap_ops *ops, const struct iomap_dio_ops *dops, unsigned int dio_flags); Patches currently in stable-queue which might be from anand.jain@xxxxxxxxxx are queue-5.15/gup-turn-fault_in_pages_-readable-writeable-into-fault_in_-readable-writeable.patch queue-5.15/iov_iter-introduce-nofault-flag-to-disable-page-faults.patch queue-5.15/gfs2-fix-mmap-page-fault-deadlocks-for-direct-i-o.patch queue-5.15/iomap-support-partial-direct-i-o-on-user-copy-failures.patch queue-5.15/mm-gup-make-fault_in_safe_writeable-use-fixup_user_fault.patch queue-5.15/gfs2-add-wrapper-for-iomap_file_buffered_write.patch queue-5.15/gup-introduce-foll_nofault-flag-to-disable-page-faults.patch queue-5.15/iov_iter-introduce-fault_in_iov_iter_writeable.patch queue-5.15/gfs2-fix-mmap-page-fault-deadlocks-for-buffered-i-o.patch queue-5.15/btrfs-fallback-to-blocking-mode-when-doing-async-dio-over-multiple-extents.patch queue-5.15/gfs2-clean-up-function-may_grant.patch queue-5.15/btrfs-fix-deadlock-due-to-page-faults-during-direct-io-reads-and-writes.patch queue-5.15/iov_iter-turn-iov_iter_fault_in_readable-into-fault_in_iov_iter_readable.patch queue-5.15/gfs2-move-the-inode-glock-locking-to-gfs2_file_buffered_write.patch queue-5.15/gfs2-introduce-flag-for-glock-holder-auto-demotion.patch queue-5.15/iomap-fix-iomap_dio_rw-return-value-for-user-copies.patch queue-5.15/gfs2-eliminate-ip-i_gh.patch queue-5.15/iomap-add-done_before-argument-to-iomap_dio_rw.patch