linux-toradex.git/fs/nfs, branch v6.17

NFSv4/flexfiles: Fix layout merge mirror check.

2025-09-08T18:37:55+00:00

Typo in ff_lseg_match_mirrors makes the diff ineffective. This results
in merge happening all the time. Merge happening all the time is
problematic because it marks lsegs invalid. Marking lsegs invalid
causes all outstanding IO to get restarted with EAGAIN and connections
to get closed.

Closing connections constantly triggers race conditions in the RDMA
implementation...

Fixes: 660d1eb22301c ("pNFS/flexfile: Don't merge layout segments if the mirrors don't match")
Signed-off-by: Jonathan Curley 
Signed-off-by: Trond Myklebust

NFS: Fix the marking of the folio as up to date

2025-09-06T20:51:26+00:00

Since all callers of nfs_page_group_covers_page() have already ensured
that there is only one group member, all that is required is to check
that the entire folio contains dirty data.

Signed-off-by: Trond Myklebust

NFS: nfs_invalidate_folio() must observe the offset and size arguments

2025-09-06T20:51:26+00:00

If we're truncating part of the folio, then we need to write out the
data on the part that is not covered by the cancellation.

Fixes: d47992f86b30 ("mm: change invalidatepage prototype to accept length")
Signed-off-by: Trond Myklebust

NFSv4.2: Serialise O_DIRECT i/o and copy range

2025-09-06T20:51:25+00:00

Ensure that all O_DIRECT reads and writes complete before copying a file
range, so that the destination is up to date.

Fixes: a5864c999de6 ("NFS: Do not serialise O_DIRECT reads and writes")
Signed-off-by: Trond Myklebust

NFSv4.2: Serialise O_DIRECT i/o and clone range

2025-09-06T20:51:25+00:00

Ensure that all O_DIRECT reads and writes complete before cloning a file
range, so that both the source and destination are up to date.

Fixes: a5864c999de6 ("NFS: Do not serialise O_DIRECT reads and writes")
Signed-off-by: Trond Myklebust

NFSv4.2: Serialise O_DIRECT i/o and fallocate()

2025-09-06T20:51:25+00:00

Ensure that all O_DIRECT reads and writes complete before calling
fallocate so that we don't race w.r.t. attribute updates.

Fixes: 99f237832243 ("NFSv4.2: Always flush out writes in nfs42_proc_fallocate()")
Signed-off-by: Trond Myklebust

NFS: Serialise O_DIRECT i/o and truncate()

2025-09-06T20:51:25+00:00

Ensure that all O_DIRECT reads and writes are complete, and prevent the
initiation of new i/o until the setattr operation that will truncate the
file is complete.

Fixes: a5864c999de6 ("NFS: Do not serialise O_DIRECT reads and writes")
Signed-off-by: Trond Myklebust

NFSv4.2: Protect copy offload and clone against 'eof page pollution'

2025-09-06T20:51:25+00:00

The NFSv4.2 copy offload and clone functions can also end up extending
the size of the destination file, so they too need to call
nfs_truncate_last_folio().

Reported-by: Olga Kornievskaia 
Signed-off-by: Trond Myklebust

NFS: Protect against 'eof page pollution'

2025-09-06T20:51:25+00:00

This commit fixes the failing xfstest 'generic/363'.

When the user mmaps() an area that extends beyond the end of file, and
proceeds to write data into the folio that straddles that eof, we're
required to discard that folio data if the user calls some function that
extends the file length.

Signed-off-by: Trond Myklebust

flexfiles/pNFS: fix NULL checks on result of ff_layout_choose_ds_for_read

2025-09-06T20:51:25+00:00

Recent commit f06bedfa62d5 ("pNFS/flexfiles: don't attempt pnfs on fatal DS
errors") has changed the error return type of ff_layout_choose_ds_for_read() from
NULL to an error pointer. However, not all code paths have been updated
to match the change. Thus, some non-NULL checks will accept error pointers
as a valid return value.

Reported-by: Dan Carpenter 
Suggested-by: Dan Carpenter 
Fixes: f06bedfa62d5 ("pNFS/flexfiles: don't attempt pnfs on fatal DS errors")
Signed-off-by: Tigran Mkrtchyan 
Signed-off-by: Trond Myklebust