Commit graph

1,426,898 commits

Author SHA1 Message Date
Christian Brauner
86ab0c1009
Merge patch series "netfs: Fix use of fscache with ceph"
David Howells <dhowells@redhat.com> says:

Here are a couple of patches that fix the use of fscaching with ceph:

 (1) Fix the read collector to mark the write request that it creates to copy
     data to the cache with NETFS_RREQ_OFFLOAD_COLLECTION so that it will run
     the write collector on a workqueue as it's meant to run in the background
     and the app isn't going to wait for it.

 (2) Fix the read collector to wake up the copy-to-cache write request after
     it sets NETFS_RREQ_ALL_QUEUED if the write request doesn't have any
     subrequests left on it.  ALL_QUEUED indicates that there won't be any
     more subreqs coming and the collector should clean up - except that an
     event is needed to trigger that, but it only gets events from subreq
     termination and so the last event can beat us to setting ALL_QUEUED.

* patches from https://lore.kernel.org/20250711151005.2956810-1-dhowells@redhat.com:
  netfs: Fix race between cache write completion and ALL_QUEUED being set
  netfs: Fix copy-to-cache so that it performs collection with ceph+fscache

Link: https://lore.kernel.org/20250711151005.2956810-1-dhowells@redhat.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 11:05:07 +02:00
David Howells
89635eae07
netfs: Fix race between cache write completion and ALL_QUEUED being set
When netfslib is issuing subrequests, the subrequests start processing
immediately and may complete before we reach the end of the issuing
function.  At the end of the issuing function we set NETFS_RREQ_ALL_QUEUED
to indicate to the collector that we aren't going to issue any more subreqs
and that it can do the final notifications and cleanup.

Now, this isn't a problem if the request is synchronous
(NETFS_RREQ_OFFLOAD_COLLECTION is unset) as the result collection will be
done in-thread and we're guaranteed an opportunity to run the collector.

However, if the request is asynchronous, collection is primarily triggered
by the termination of subrequests queuing it on a workqueue.  Now, a race
can occur here if the app thread sets ALL_QUEUED after the last subrequest
terminates.

This can happen most easily with the copy2cache code (as used by Ceph)
where, in the collection routine of a read request, an asynchronous write
request is spawned to copy data to the cache.  Folios are added to the
write request as they're unlocked, but there may be a delay before
ALL_QUEUED is set as the write subrequests may complete before we get
there.

If all the write subreqs have finished by the ALL_QUEUED point, no further
events happen and the collection never happens, leaving the request
hanging.

Fix this by queuing the collector after setting ALL_QUEUED.  This is a bit
heavy-handed and it may be sufficient to do it only if there are no extant
subreqs.

Also add a tracepoint to cross-reference both requests in a copy-to-request
operation and add a trace to the netfs_rreq tracepoint to indicate the
setting of ALL_QUEUED.

Fixes: e2d46f2ec3 ("netfs: Change the read result collector to only use one work item")
Reported-by: Max Kellermann <max.kellermann@ionos.com>
Link: https://lore.kernel.org/r/CAKPOu+8z_ijTLHdiCYGU_Uk7yYD=shxyGLwfe-L7AV3DhebS3w@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/20250711151005.2956810-3-dhowells@redhat.com
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
cc: Paulo Alcantara <pc@manguebit.org>
cc: Viacheslav Dubeyko <slava@dubeyko.com>
cc: Alex Markuze <amarkuze@redhat.com>
cc: Ilya Dryomov <idryomov@gmail.com>
cc: netfs@lists.linux.dev
cc: ceph-devel@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
cc: stable@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 11:05:02 +02:00
David Howells
4c238e3077
netfs: Fix copy-to-cache so that it performs collection with ceph+fscache
The netfs copy-to-cache that is used by Ceph with local caching sets up a
new request to write data just read to the cache.  The request is started
and then left to look after itself whilst the app continues.  The request
gets notified by the backing fs upon completion of the async DIO write, but
then tries to wake up the app because NETFS_RREQ_OFFLOAD_COLLECTION isn't
set - but the app isn't waiting there, and so the request just hangs.

Fix this by setting NETFS_RREQ_OFFLOAD_COLLECTION which causes the
notification from the backing filesystem to put the collection onto a work
queue instead.

Fixes: e2d46f2ec3 ("netfs: Change the read result collector to only use one work item")
Reported-by: Max Kellermann <max.kellermann@ionos.com>
Link: https://lore.kernel.org/r/CAKPOu+8z_ijTLHdiCYGU_Uk7yYD=shxyGLwfe-L7AV3DhebS3w@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/20250711151005.2956810-2-dhowells@redhat.com
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
cc: Paulo Alcantara <pc@manguebit.org>
cc: Viacheslav Dubeyko <slava@dubeyko.com>
cc: Alex Markuze <amarkuze@redhat.com>
cc: Ilya Dryomov <idryomov@gmail.com>
cc: netfs@lists.linux.dev
cc: ceph-devel@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
cc: stable@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 11:05:02 +02:00
K Prateek Nayak
1eec89a671 sched/topology: Remove sched_domain_topology_level::flags
Support for overlapping domains added in commit e3589f6c81 ("sched:
Allow for overlapping sched_domain spans") also allowed forcefully
setting SD_OVERLAP for !NUMA domains via FORCE_SD_OVERLAP sched_feat().

Since NUMA domains had to be presumed overlapping to ensure correct
behavior, "sched_domain_topology_level::flags" was introduced. NUMA
domains added the SDTL_OVERLAP flag would ensure SD_OVERLAP was always
added during build_sched_domains() for these domains, even when
FORCE_SD_OVERLAP was off.

Condition for adding the SD_OVERLAP flag at the aforementioned commit
was as follows:

    if (tl->flags & SDTL_OVERLAP || sched_feat(FORCE_SD_OVERLAP))
            sd->flags |= SD_OVERLAP;

The FORCE_SD_OVERLAP debug feature was removed in commit af85596c74
("sched/topology: Remove FORCE_SD_OVERLAP") which left the NUMA domains
as the exclusive users of SDTL_OVERLAP, SD_OVERLAP, and SD_NUMA flags.

Get rid of SDTL_OVERLAP and SD_OVERLAP as they have become redundant
and instead rely on SD_NUMA to detect the only overlapping domain
currently supported. Since SDTL_OVERLAP was the only user of
"tl->flags", get rid of "sched_domain_topology_level::flags" too.

Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/ba4dbdf8-bc37-493d-b2e0-2efb00ea3e19@amd.com
2025-07-14 10:59:35 +02:00
Li Chen
f79c9aa446 x86/smpboot: avoid SMT domain attach/destroy if SMT is not enabled
Currently, the SMT domain is added into sched_domain_topology by default.

If cpu_attach_domain() finds that the CPU SMT domain’s cpumask_weight
is just 1, it will destroy it.

On a large machine, such as one with 512 cores, this results in
512 redundant domain attach/destroy operations.

Avoid these unnecessary operations by simply checking
cpu_smt_num_threads and skip SMT domain if the SMT domain is not
enabled.

Suggested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20250710105715.66594-5-me@linux.beauty
2025-07-14 10:59:34 +02:00
Li Chen
fbc2010d92 x86/smpboot: moves x86_topology to static initialize and truncate
The #ifdeffery and the initializers in build_sched_topology() are just
disgusting.

Statically initialize the domain levels in the topology array and let
build_sched_topology() invalidate the package domain level when NUMA in
package is available.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20250710105715.66594-4-me@linux.beauty
2025-07-14 10:59:34 +02:00
Li Chen
992de2b025 x86/smpboot: remove redundant CONFIG_SCHED_SMT
On x86 CONFIG_SCHED_SMT is default y if SMP is enabled, so let's
simply drop CONFIG_SCHED_SMT.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20250710105715.66594-3-me@linux.beauty
2025-07-14 10:59:34 +02:00
Li Chen
e075f43609 smpboot: introduce SDTL_INIT() helper to tidy sched topology setup
Define a small SDTL_INIT(maskfn, flagsfn, name) macro and use it to build the
sched_domain_topology_level array. Purely a cleanup; behaviour is unchanged.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20250710105715.66594-2-me@linux.beauty
2025-07-14 10:59:34 +02:00
Juri Lelli
634c24068a tools/sched: Add dl_bw_dump.py for printing bandwidth accounting info
dl_rq bandwidth accounting information is crucial for the correct
functioning of SCHED_DEADLINE.

Add a drgn script for accessing that information at runtime, so that
it's easier to check and debug issues related to it.

Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk> # nuc & rock5b
Link: https://lore.kernel.org/r/20250627115118.438797-6-juri.lelli@redhat.com
2025-07-14 10:59:33 +02:00
Juri Lelli
9fdb12c88e tools/sched: Add root_domains_dump.py which dumps root domains info
Root domains information is somewhat hard to access at runtime. Even
with sched_debug and sched_verbose, such information is only printed
on kernel console when domains are modified.

Add a simple drgn script to more easily retrieve root domains
information at runtime.

Since tools/sched is a new directory, add it to MAINTAINERS as well.

Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk> # nuc & rock5b
Link: https://lore.kernel.org/r/20250627115118.438797-5-juri.lelli@redhat.com
2025-07-14 10:59:33 +02:00
Juri Lelli
440989c10f sched/deadline: Fix accounting after global limits change
A global limits change (sched_rt_handler() logic) currently leaves stale
and/or incorrect values in variables related to accounting (e.g.
extra_bw).

Properly clean up per runqueue variables before implementing the change
and rebuild scheduling domains (so that accounting is also properly
restored) after such a change is complete.

Reported-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk> # nuc & rock5b
Link: https://lore.kernel.org/r/20250627115118.438797-4-juri.lelli@redhat.com
2025-07-14 10:59:33 +02:00
Juri Lelli
fcc9276c4d sched/deadline: Reset extra_bw to max_bw when clearing root domains
dl_clear_root_domain() doesn't take into account the fact that per-rq
extra_bw variables retain values computed before root domain changes,
resulting in broken accounting.

Fix it by resetting extra_bw to max_bw before restoring back dl-servers
contributions.

Fixes: 2ff899e351 ("sched/deadline: Rebuild root domain accounting after every update")
Reported-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk> # nuc & rock5b
Link: https://lore.kernel.org/r/20250627115118.438797-3-juri.lelli@redhat.com
2025-07-14 10:59:32 +02:00
Juri Lelli
9f239df555 sched/deadline: Initialize dl_servers after SMP
dl-servers are currently initialized too early at boot when CPUs are not
fully up (only boot CPU is). This results in miscalculation of per
runqueue DEADLINE variables like extra_bw (which needs a stable CPU
count).

Move initialization of dl-servers later on after SMP has been
initialized and CPUs are all online, so that CPU count is stable and
DEADLINE variables can be computed correctly.

Fixes: d741f297bc ("sched/fair: Fair server interface")
Reported-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Waiman Long <longman@redhat.com>
Tested-by: Marcel Ziswiler <marcel.ziswiler@codethink.co.uk> # nuc & rock5b
Link: https://lore.kernel.org/r/20250627115118.438797-2-juri.lelli@redhat.com
2025-07-14 10:59:32 +02:00
Aruna Ramakrishna
36569780b0 sched: Change nr_uninterruptible type to unsigned long
The commit e6fe3f422b ("sched: Make multiple runqueue task counters
32-bit") changed nr_uninterruptible to an unsigned int. But the
nr_uninterruptible values for each of the CPU runqueues can grow to
large numbers, sometimes exceeding INT_MAX. This is valid, if, over
time, a large number of tasks are migrated off of one CPU after going
into an uninterruptible state. Only the sum of all nr_interruptible
values across all CPUs yields the correct result, as explained in a
comment in kernel/sched/loadavg.c.

Change the type of nr_uninterruptible back to unsigned long to prevent
overflows, and thus the miscalculation of load average.

Fixes: e6fe3f422b ("sched: Make multiple runqueue task counters 32-bit")

Signed-off-by: Aruna Ramakrishna <aruna.ramakrishna@oracle.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20250709173328.606794-1-aruna.ramakrishna@oracle.com
2025-07-14 10:59:31 +02:00
Christian Brauner
2f368b5f93
Merge patch series "refactor the iomap writeback code v5"
Christoph Hellwig <hch@lst.de> says:

This is an alternative approach to the writeback part of the
"fuse: use iomap for buffered writes + writeback" series from Joanne.

The big difference compared to Joanne's version is that I hope the
split between the generic and ioend/bio based writeback code is a bit
cleaner here.  We have two methods that define the split between the
generic writeback code, and the implemementation of it, and all knowledge
of ioends and bios now sits below that layer.

This version passes testing on xfs, and gets as far as mainline for
gfs2 (crashes in generic/361).

* patches from https://lore.kernel.org/20250710133343.399917-1-hch@lst.de:
  iomap: build the writeback code without CONFIG_BLOCK
  iomap: add read_folio_range() handler for buffered writes
  iomap: improve argument passing to iomap_read_folio_sync
  iomap: replace iomap_folio_ops with iomap_write_ops
  iomap: export iomap_writeback_folio
  iomap: move folio_unlock out of iomap_writeback_folio
  iomap: rename iomap_writepage_map to iomap_writeback_folio
  iomap: move all ioend handling to ioend.c
  iomap: add public helpers for uptodate state manipulation
  iomap: hide ioends from the generic writeback code
  iomap: refactor the writeback interface
  iomap: cleanup the pending writeback tracking in iomap_writepage_map_blocks
  iomap: pass more arguments using the iomap writeback context
  iomap: header diet

Link: https://lore.kernel.org/20250710133343.399917-1-hch@lst.de
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:38 +02:00
Christoph Hellwig
5699b7e21d
iomap: build the writeback code without CONFIG_BLOCK
Allow fuse to use the iomap writeback code even when CONFIG_BLOCK is
not enabled.  Do this with an ifdef instead of a separate file to keep
the iomap_folio_state local to buffered-io.c.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-15-hch@lst.de
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:33 +02:00
Christoph Hellwig
c5690dd019
iomap: add read_folio_range() handler for buffered writes
Add a read_folio_range() handler for buffered writes that filesystems
may pass in if they wish to provide a custom handler for synchronously
reading in the contents of a folio.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
[hch: renamed to read_folio_range, pass less arguments]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-14-hch@lst.de
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:33 +02:00
Christoph Hellwig
e6caf01d3f
iomap: improve argument passing to iomap_read_folio_sync
Pass the iomap_iter and derive the map inside iomap_read_folio_sync
instead of in the caller, and use the more descriptive srcmap name for
the source iomap.  Stop passing the offset into folio argument as it
can be derived from the folio and the file offset.  Rename the
variables for the offset into the file and the length to be more
descriptive and match the rest of the code.

Rename the function itself to iomap_read_folio_range to make the use
more clear.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-13-hch@lst.de
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:33 +02:00
Christoph Hellwig
2a5574fc57
iomap: replace iomap_folio_ops with iomap_write_ops
The iomap_folio_ops are only used for buffered writes, including the zero
and unshare variants.  Rename them to iomap_write_ops to better describe
the usage, and pass them through the call chain like the other operation
specific methods instead of through the iomap.

xfs_iomap_valid grows a IOMAP_HOLE check to keep the existing behavior
that never attached the folio_ops to a iomap representing a hole.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-12-hch@lst.de
Acked-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:33 +02:00
Christoph Hellwig
8b217cf779
iomap: export iomap_writeback_folio
Allow fuse to use iomap_writeback_folio for folio laundering.  Note
that the caller needs to manually submit the pending writeback context.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-11-hch@lst.de
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:32 +02:00
Joanne Koong
f8b6a94a4c
iomap: move folio_unlock out of iomap_writeback_folio
Move unlocking the folio out of iomap_writeback_folio into the caller.
This means the end writeback machinery is now run with the folio locked
when no writeback happened, or writeback completed extremely fast.

Note that having the folio locked over the call to folio_end_writeback in
iomap_writeback_folio means that the dropbehind handling there will never
run because the trylock fails.  The only way this can happen is if the
writepage either never wrote back any dirty data at all, in which case
the dropbehind handling isn't needed, or if all writeback finished
instantly, which is rather unlikely.  Even in the latter case the
dropbehind handling is an optional optimization so skipping it will not
cause correctness issues.

This prepares for exporting iomap_writeback_folio for use in folio
laundering.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
[hch: split from a larger patch]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-10-hch@lst.de
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:32 +02:00
Christoph Hellwig
58f0d5a304
iomap: rename iomap_writepage_map to iomap_writeback_folio
->writepage is gone, and our naming wasn't always that great to start
with.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-9-hch@lst.de
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:32 +02:00
Christoph Hellwig
8f02cecd80
iomap: move all ioend handling to ioend.c
Now that the writeback code has the proper abstractions, all the ioend
code can be self-contained in ioend.c.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-8-hch@lst.de
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:32 +02:00
Joanne Koong
9caf1ea80c
iomap: add public helpers for uptodate state manipulation
Add a new iomap_start_folio_write helper to abstract away the
write_bytes_pending handling, and export it and the existing
iomap_finish_folio_write for non-iomap writeback in fuse.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
[hch: split from a larger patch]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-7-hch@lst.de
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:32 +02:00
Christoph Hellwig
f4fa7981fa
iomap: hide ioends from the generic writeback code
Replace the ioend pointer in iomap_writeback_ctx with a void *wb_ctx
one to facilitate non-block, non-ioend writeback for use.  Rename
the submit_ioend method to writeback_submit and make it mandatory so
that the generic writeback code stops seeing ioends and bios.

Co-developed-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-6-hch@lst.de
Acked-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:31 +02:00
Christoph Hellwig
fb7399cf2d
iomap: refactor the writeback interface
Replace ->map_blocks with a new ->writeback_range, which differs in the
following ways:

 - it must also queue up the I/O for writeback, that is called into the
   slightly refactored and extended in scope iomap_add_to_ioend for
   each region
 - can handle only a part of the requested region, that is the retry
   loop for partial mappings moves to the caller
 - handles cleanup on failures as well, and thus also replaces the
   discard_folio method only implemented by XFS.

This will allow to use the iomap writeback code also for file systems
that are not block based like fuse.

Co-developed-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-5-hch@lst.de
Acked-by: Damien Le Moal <dlemoal@kernel.org>	# zonefs
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:31 +02:00
Joanne Koong
40368a6acb
iomap: cleanup the pending writeback tracking in iomap_writepage_map_blocks
We don't care about the count of outstanding ioends, just if there is one.
Replace the count variable passed to iomap_writepage_map_blocks with a
boolean to make that more clear.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
[hch: rename the variable, update the commit message]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-4-hch@lst.de
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:31 +02:00
Christoph Hellwig
67fd9615a7
iomap: pass more arguments using the iomap writeback context
Add inode and wpc fields to pass the inode and writeback context that
are needed in the entire writeback call chain, and let the callers
initialize all fields in the writeback context before calling
iomap_writepages to simplify the argument passing.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-3-hch@lst.de
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:31 +02:00
Christoph Hellwig
8cd0a39cab
iomap: header diet
Drop various unused #include statements.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-2-hch@lst.de
Reviewed-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:51:31 +02:00
Al Viro
1f531e35c1
don't bother with path_get()/path_put() in unix_open_file()
Once unix_sock ->path is set, we are guaranteed that its ->path will remain
unchanged (and pinned) until the socket is closed.  OTOH, dentry_open()
does not modify the path passed to it.

IOW, there's no need to copy unix_sk(sk)->path in unix_open_file() - we
can just pass it to dentry_open() and be done with that.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Link: https://lore.kernel.org/20250712054157.GZ1880847@ZenIV
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:22:47 +02:00
Al Viro
fdfe013347
fix a leak in fcntl_dirnotify()
[into #fixes, unless somebody objects]

Lifetime of new_dn_mark is controlled by that of its ->fsn_mark,
pointed to by new_fsn_mark.  Unfortunately, a failure exit had
been inserted between the allocation of new_dn_mark and the
call of fsnotify_init_mark(), ending up with a leak.

Fixes: 1934b21261 "file: reclaim 24 bytes from f_owner"
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Link: https://lore.kernel.org/20250712171843.GB1880847@ZenIV
Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-07-14 10:13:31 +02:00
Al Viro
532c8b51b3 xen: fix UAF in dmabuf_exp_from_pages()
[dma_buf_fd() fixes; no preferences regarding the tree it goes through -
up to xen folks]

As soon as we'd inserted a file reference into descriptor table, another
thread could close it.  That's fine for the case when all we are doing is
returning that descriptor to userland (it's a race, but it's a userland
race and there's nothing the kernel can do about it).  However, if we
follow fd_install() with any kind of access to objects that would be
destroyed on close (be it the struct file itself or anything destroyed
by its ->release()), we have a UAF.

dma_buf_fd() is a combination of reserving a descriptor and fd_install().
gntdev dmabuf_exp_from_pages() calls it and then proceeds to access the
objects destroyed on close - starting with gntdev_dmabuf itself.

Fix that by doing reserving descriptor before anything else and do
fd_install() only when everything had been set up.

Fixes: a240d6e42e ("xen/gntdev: Implement dma-buf export functionality")
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: Juergen Gross <jgross@suse.com>
Message-ID: <20250712050916.GY1880847@ZenIV>
Signed-off-by: Juergen Gross <jgross@suse.com>
2025-07-14 10:12:07 +02:00
Dr. David Alan Gilbert
0df1195009 xen: Remove some deadcode (x)
Remove three uncalled functions:

  xenbus_mkdir() was added in 2007 by
commit 4bac07c993 ("xen: add the Xenbus sysfs and virtual device hotplug
driver")
but has remained unused.

  xen_get_runstate_snapshot() last use was removed in 2016 by
commit 6ba286ad84 ("xen: support runqueue steal time on xen")
which replaces the use by the _cpu version.

  xen_resume_notifier_unregister() last use was removed in 2017 by
commit 1914f0cd20 ("xen/acpi: upload PM state from init-domain to Xen")

Remove them.

Signed-off-by: "Dr. David Alan Gilbert" <linux@treblig.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20250713132625.164728-1-linux@treblig.org>
2025-07-14 09:55:52 +02:00
Ryan Chung
c79626899d xen-pciback: Replace scnprintf() with sysfs_emit_at()
This is the third revision (v3) of this patch series.
No changes since v2—only adding Reviewed-by lines.

Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Ryan Chung <seokwoo.chung130@gmail.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20250708001444.86155-1-seokwoo.chung130@gmail.com>
2025-07-14 09:35:42 +02:00
Peng Jiang
b7efeb081e xen/xenbus: fix W=1 build warning in xenbus_va_dev_error function
This patch fixes a W=1 format-string warning reported by GCC 12.3.0
by annotating xenbus_switch_fatal() and xenbus_va_dev_error()
with the __printf attribute. The attribute enables compile-time
validation of printf-style format strings in these functions.

The original warning trace:
drivers/xen/xenbus/xenbus_client.c:304:9: warning: function 'xenbus_va_dev_error' might be
a candidate for 'gnu_printf' format attribute [-Wsuggest-attribute=format]

Signed-off-by: Peng Jiang <jiang.peng9@zte.com.cn>
Reviewed-by: Juergen Gross <jgross@suse.com>
Message-ID: <20250620084104786r5xoR16_AmYZMJLnno3_Q@zte.com.cn>
Signed-off-by: Juergen Gross <jgross@suse.com>
2025-07-14 09:34:54 +02:00
Andy Shevchenko
148fbaf571 ata: pata_rdc: Use registered definition for the RDC vendor
Convert to PCI_VDEVICE() and use registered definition for RDC vendor
from pci_ids.h.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20250711113650.1475307-1-andriy.shevchenko@linux.intel.com
[cassel: add ata: prefix to subject, fix typo in Damien's Rb tag]
Signed-off-by: Niklas Cassel <cassel@kernel.org>
2025-07-14 09:22:32 +02:00
I Viswanath
c3ff7f06c7 i2c: Clarify behavior of I2C_M_RD flag
Update the description of I2C_M_RD to clarify that not setting it
signals a write transaction

Signed-off-by: I Viswanath <viswanathiyyappan@gmail.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
2025-07-14 09:15:58 +02:00
Wolfram Sang
a8780906ca i2c-host-fixes for v6.16-rc6
omap: add missing error check and fix PM disable in probe error
 path.
 
 stm32: unmap DMA buffer on transfer failure and use correct
 device when mapping and unmapping during transfers.
 -----BEGIN PGP SIGNATURE-----
 
 iIwEABYKADQWIQScDfrjQa34uOld1VLaeAVmJtMtbgUCaHEyAhYcYW5kaS5zaHl0
 aUBrZXJuZWwub3JnAAoJENp4BWYm0y1u2bMA/i4VbCCQovRGxqgNvi/xbn6AbTeJ
 HOAMhCWma1GBztWdAP9iL50wLU0GSbLlMsgjQqxhTybA9ep5EHmQAKXNjvXwCQ==
 =vccm
 -----END PGP SIGNATURE-----

Merge tag 'i2c-host-fixes-6.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux into i2c/for-current

i2c-host-fixes for v6.16-rc6

omap: add missing error check and fix PM disable in probe error
path.

stm32: unmap DMA buffer on transfer failure and use correct
device when mapping and unmapping during transfers.
2025-07-14 09:02:25 +02:00
Steffen Klassert
28712d6ed3 Merge branch 'ipsec: fix splat due to ipcomp fallback tunnel'
Sabrina Dubroca says:

====================
IPcomp tunnel states have an associated fallback tunnel, a keep a
reference on the corresponding xfrm_state, to allow deleting that
extra state when it's not needed anymore. These states cause issues
during netns deletion.

Commit f75a2804da ("xfrm: destroy xfrm_state synchronously on net
exit path") tried to address these problems but doesn't fully solve
them, and slowed down netns deletion by adding one synchronize_rcu per
deleted state.

The first patch solves the problem by moving the fallback state
deletion earlier (when we delete the user state, rather than at
destruction), then we can revert the previous fix.
====================

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
2025-07-14 08:59:48 +02:00
Brian Masney
80395c3b47 clk: sunxi-ng: ccu_nm: convert from round_rate() to determine_rate()
The round_rate() clk ops is deprecated, so migrate this driver from
round_rate() to determine_rate() using the Coccinelle semantic patch
on the cover letter of this series.

I manually fixed up one minor formatting issue that occurred after
applying the semantic patch:

        req->rate = ccu_nm_find_best(&nm->common, req->best_parent_rate,
                                     req->rate,
                                     &_nm);

I manually changed it to:

        req->rate = ccu_nm_find_best(&nm->common, req->best_parent_rate,
                                     req->rate, &_nm);

Signed-off-by: Brian Masney <bmasney@redhat.com>
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://patch.msgid.link/20250703-clk-cocci-drop-round-rate-v1-10-3a8da898367e@redhat.com
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:56:18 +08:00
Brian Masney
8bc614c6ac clk: sunxi-ng: ccu_nkmp: convert from round_rate() to determine_rate()
The round_rate() clk ops is deprecated, so migrate this driver from
round_rate() to determine_rate() using the Coccinelle semantic patch
on the cover letter of this series.

Signed-off-by: Brian Masney <bmasney@redhat.com>
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://patch.msgid.link/20250703-clk-cocci-drop-round-rate-v1-9-3a8da898367e@redhat.com
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:56:17 +08:00
Brian Masney
2b0d4f1b3f clk: sunxi-ng: ccu_nk: convert from round_rate() to determine_rate()
The round_rate() clk ops is deprecated, so migrate this driver from
round_rate() to determine_rate() using the Coccinelle semantic patch
on the cover letter of this series.

Signed-off-by: Brian Masney <bmasney@redhat.com>
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://patch.msgid.link/20250703-clk-cocci-drop-round-rate-v1-8-3a8da898367e@redhat.com
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:56:17 +08:00
Brian Masney
ee9c15ca0f clk: sunxi-ng: ccu_gate: convert from round_rate() to determine_rate()
The round_rate() clk ops is deprecated, so migrate this driver from
round_rate() to determine_rate() using the Coccinelle semantic patch
on the cover letter of this series.

Signed-off-by: Brian Masney <bmasney@redhat.com>
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://patch.msgid.link/20250703-clk-cocci-drop-round-rate-v1-7-3a8da898367e@redhat.com
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:56:17 +08:00
Paul Kocialkowski
ea879ce83d clk: sunxi-ng: v3s: Assign the de and tcon clocks to the video pll
It appears (based on experimentation) that both the de and tcon clocks
need to have the same parent for the two units to work together.

Assign them both to the video pll by manually clearing the parent
selection bits (effectively setting index 0) and marking the clocks
with the CLK_SET_RATE_NO_REPARENT flag, which ensures that they will
never use a different parent.

The video pll is also a possible parent for the camera subsystem,
but it can use the dedicated isp pll if needed so there should be
no negative side-effect due to this change.

Note that ccu_mux_helper_set_parent cannot be used at this stage as
it requires the clock driver to be initialized and this configuration
is best done before the clock driver is available to consumers.

Signed-off-by: Paul Kocialkowski <paulk@sys-base.io>
Link: https://patch.msgid.link/20250704154008.3463257-2-paulk@sys-base.io
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:51:15 +08:00
Paul Kocialkowski
e8ab346f99 clk: sunxi-ng: v3s: Fix de clock definition
The de clock is marked with CLK_SET_RATE_PARENT, which is really not
necessary (as confirmed from experimentation) and significantly
restricts flexibility for other clocks using the same parent.

In addition the source selection (parent) field is marked as using
2 bits, when it the documentation reports that it uses 3.

Fix both issues in the de clock definition.

Fixes: d0f11d14b0 ("clk: sunxi-ng: add support for V3s CCU")
Signed-off-by: Paul Kocialkowski <paulk@sys-base.io>
Link: https://patch.msgid.link/20250704154008.3463257-1-paulk@sys-base.io
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
2025-07-14 11:51:14 +08:00
Zhang Yi
5137d6c890 ext4: fix insufficient credits calculation in ext4_meta_trans_blocks()
The calculation of journal credits in ext4_meta_trans_blocks() should
include pextents, as each extent separately may be allocated from a
different group and thus need to update different bitmap and group
descriptor block.

Fixes: 0e32d86170 ("ext4: correct the journal credits calculations of allocating blocks")
Reported-by: Jan Kara <jack@suse.cz>
Closes: https://lore.kernel.org/linux-ext4/nhxfuu53wyacsrq7xqgxvgzcggyscu2tbabginahcygvmc45hy@t4fvmyeky33e/
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Baokun Li <libaokun1@huawei.com>
Link: https://patch.msgid.link/20250707140814.542883-11-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-07-13 23:41:52 -04:00
Zhang Yi
57661f2875 ext4: replace ext4_writepage_trans_blocks()
After ext4 supports large folios, the semantics of reserving credits in
pages is no longer applicable. In most scenarios, reserving credits in
extents is sufficient. Therefore, introduce ext4_chunk_trans_extent()
to replace ext4_writepage_trans_blocks(). move_extent_per_page() is the
only remaining location where we are still processing extents in pages.

Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://patch.msgid.link/20250707140814.542883-10-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-07-13 23:41:52 -04:00
Zhang Yi
bbbf150f3f ext4: reserved credits for one extent during the folio writeback
After ext4 supports large folios, reserving journal credits for one
maximum-ordered folio based on the worst case cenario during the
writeback process can easily exceed the maximum transaction credits.
Additionally, reserving journal credits for one page is also no
longer appropriate.

Currently, the folio writeback process can either extend the journal
credits or initiate a new transaction if the currently reserved journal
credits are insufficient. Therefore, it can be modified to reserve
credits for only one extent at the outset. In most cases involving
continuous mapping, these credits are generally adequate, and we may
only need to perform some basic credit expansion. However, in extreme
cases where the block size and folio size differ significantly, or when
the folios are sufficiently discontinuous, it may be necessary to
restart a new transaction and resubmit the folios.

Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://patch.msgid.link/20250707140814.542883-9-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-07-13 23:41:52 -04:00
Zhang Yi
95ad8ee45c ext4: correct the reserved credits for extent conversion
Now, we reserve journal credits for converting extents in only one page
to written state when the I/O operation is complete. This is
insufficient when large folio is enabled.

Fix this by reserving credits for converting up to one extent per block in
the largest 2MB folio, this calculation should only involve extents index
and leaf blocks, so it should not estimate too many credits.

Fixes: 7ac67301e8 ("ext4: enable large folio for regular file")
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Baokun Li <libaokun1@huawei.com>
Link: https://patch.msgid.link/20250707140814.542883-8-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-07-13 23:41:52 -04:00
Zhang Yi
6b132759b0 ext4: enhance tracepoints during the folios writeback
After mpage_map_and_submit_extent() supports restarting handle if
credits are insufficient during allocating blocks, it is more likely to
exit the current mapping iteration and continue to process the current
processing partially mapped folio again. The existing tracepoints are
not sufficient to track this situation, so enhance the tracepoints to
track the writeback position and the return value before and after
submitting the folios.

Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://patch.msgid.link/20250707140814.542883-7-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-07-13 23:41:51 -04:00