Commit graph

1,309,163 commits

Author SHA1 Message Date
Linus Torvalds
cef27048e5 bcachefs fixes for 6.9-rc5
various recovery fixes:
 
 - fixes for the btree_insert_entry being resized on path allocation
   btree_path array recently became dynamically resizable, and
   btree_insert_entry along with it; this was being observed during
   journal replay, when write buffer btree updates don't use the write
   buffer and instead use the normal btree update path
 - multiple fixes for deadlock in recovery when we need to do lots of
   btree node merges; excessive merges were clocking up the whole
   pipeline
 - write buffer path now correctly does btree node merges when needed
 - fix failure to go RW when superblock indicates recovery passes needed
   (i.e. to complete an unfinished upgrade)
 
 various unsafety fixes - test case contributed by a user who had two
 drives out of a six drive array write out a whole bunch of garbage after
 power failure
 
 new (tiny) on disk format feature: since it appears the btree node scan
 tool will be a more regular thing (crappy hardware, user error) - this
 adds a 64 bit per-device bitmap of regions that have ever had btree
 nodes.
 
 a path->should_be_locked fix, from a larger patch series tightening up
 invariants and assertions around btree transaction and path locking
 state; this particular fix prevents us from keeping around btree_paths
 that are no longer needed.
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEEKnAFLkS8Qha+jvQrE6szbY3KbnYFAmYdaRIACgkQE6szbY3K
 bnbqcA/9ETT0Jekf/V4klQmoWj9GX5nQstUz+ENABNNPL+5hld62EojiRvOW2qwU
 zVs7O0M59B8/+4v4KJoW+RqnLFjAF4z/Gf+/Uw9WarsHAKIxxFFFARxG93JpGqOn
 nGa8RSw0BaYQIdbMR0Bdacc2f0N+JkJQx956/+JV7EG5MAJqXgz00AvIuLqMZ+2t
 0m9av3n0tVmstyvvGqk8pouvQjK0XUvIDYN3oiUDl7WXOAIKXDlp6yviiGnTbusq
 DssmIt5fdeVBq/DAk5PMNEKM9NUP+weIZW1UWPWINaicarqyV+pn2fhvLrBxVl7q
 zBSN3v28viaABKC8A15b2bqj3IT2WIBDoBCEi406akMao9eiVsE6is13rFkPQwQI
 Obhc7NNDyOPPTvX25M3tKXpr8rSGoD2qHIMMKMIBe1ZWscj6lMbmUBErwzTOAW4+
 pNTvzWT2XwcS7tE8Fx50ZxcehTQl6ir0hQvjJL5JV2po8XMbdGxcImBe6xPmAa3n
 /XIzyglL8IvW494wjCsHxtTeOt+f8nW7BXJCrWB71UQeXIXq4b9FADOwWtlGTnxJ
 6XNprfi8TSp+RsSRxav6DBw2ou5viGjAjP2ddrO6Lw37XUYV0igS+BeDNEPA4dwI
 ZlbCzNE7qSXK2rjmGjyu7GCJ3+NOxJDQ8GdxkTDtpPrBF2kCOkQ=
 =NAId
 -----END PGP SIGNATURE-----

Merge tag 'bcachefs-2024-04-15' of https://evilpiepirate.org/git/bcachefs

Pull yet more bcachefs fixes from Kent Overstreet:
 "This gets recovery working again for the affected user I've been
  working with, and I'm still waiting to hear back on other bug reports
  but should fix it for everyone else who's been having issues with
  recovery.

   - Various recovery fixes:

       - fixes for the btree_insert_entry being resized on path
         allocation btree_path array recently became dynamically
         resizable, and btree_insert_entry along with it; this was being
         observed during journal replay, when write buffer btree updates
         don't use the write buffer and instead use the normal btree
         update path

       - multiple fixes for deadlock in recovery when we need to do lots
         of btree node merges; excessive merges were clocking up the
         whole pipeline

       - write buffer path now correctly does btree node merges when
         needed

       - fix failure to go RW when superblock indicates recovery passes
         needed (i.e. to complete an unfinished upgrade)

   - Various unsafety fixes - test case contributed by a user who had
     two drives out of a six drive array write out a whole bunch of
     garbage after power failure

   - New (tiny) on disk format feature: since it appears the btree node
     scan tool will be a more regular thing (crappy hardware, user
     error) - this adds a 64 bit per-device bitmap of regions that have
     ever had btree nodes.

   - A path->should_be_locked fix, from a larger patch series tightening
     up invariants and assertions around btree transaction and path
     locking state.

     This particular fix prevents us from keeping around btree_paths
     that are no longer needed"

* tag 'bcachefs-2024-04-15' of https://evilpiepirate.org/git/bcachefs: (24 commits)
  bcachefs: set_btree_iter_dontneed also clears should_be_locked
  bcachefs: fix error path of __bch2_read_super()
  bcachefs: Check for backpointer bucket_offset >= bucket size
  bcachefs: bch_member.btree_allocated_bitmap
  bcachefs: sysfs internal/trigger_journal_flush
  bcachefs: Fix bch2_btree_node_fill() for !path
  bcachefs: add safety checks in bch2_btree_node_fill()
  bcachefs: Interior known are required to have known key types
  bcachefs: add missing bounds check in __bch2_bkey_val_invalid()
  bcachefs: Fix btree node merging on write buffer btrees
  bcachefs: Disable merges from interior update path
  bcachefs: Run merges at BCH_WATERMARK_btree
  bcachefs: Fix missing write refs in fs fio paths
  bcachefs: Fix deadlock in journal replay
  bcachefs: Go rw if running any explicit recovery passes
  bcachefs: Standardize helpers for printing enum strs with bounds checks
  bcachefs: don't queue btree nodes for rewrites during scan
  bcachefs: fix race in bch2_btree_node_evict()
  bcachefs: fix unsafety in bch2_stripe_to_text()
  bcachefs: fix unsafety in bch2_extent_ptr_to_text()
  ...
2024-04-15 11:01:11 -07:00
Maíra Canal
35f4f8c9fc
drm/v3d: Don't increment enabled_ns twice
The commit 509433d814 ("drm/v3d: Expose the total GPU usage stats on sysfs")
introduced the calculation of global GPU stats. For the regards, it used
the already existing infrastructure provided by commit 09a93cc4f7 ("drm/v3d:
Implement show_fdinfo() callback for GPU usage stats"). While adding
global GPU stats calculation ability, the author forgot to delete the
existing one.

Currently, the value of `enabled_ns` is incremented twice by the end of
the job, when it should be added just once. Therefore, delete the
leftovers from commit 509433d814 ("drm/v3d: Expose the total GPU usage
stats on sysfs").

Fixes: 509433d814 ("drm/v3d: Expose the total GPU usage stats on sysfs")
Reported-by: Tvrtko Ursulin <tursulin@igalia.com>
Signed-off-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240403203517.731876-2-mcanal@igalia.com
2024-04-15 15:00:19 -03:00
Uladzislau Rezki (Sony)
0fd210baa0 rcu: Allocate WQ with WQ_MEM_RECLAIM bit set
synchronize_rcu() users have to be processed regardless
of memory pressure so our private WQ needs to have at least
one execution context what WQ_MEM_RECLAIM flag guarantees.

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:51:26 +02:00
Uladzislau Rezki (Sony)
462df2f543 rcu: Support direct wake-up of synchronize_rcu() users
This patch introduces a small enhancement which allows to do a
direct wake-up of synchronize_rcu() callers. It occurs after a
completion of grace period, thus by the gp-kthread.

Number of clients is limited by the hard-coded maximum allowed
threshold. The remaining part, if still exists is deferred to
a main worker.

Link: https://lore.kernel.org/lkml/Zd0ZtNu+Rt0qXkfS@lothringen/

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:47:51 +02:00
Uladzislau Rezki (Sony)
2053937a31 rcu: Add a trace event for synchronize_rcu_normal()
Add an rcu_sr_normal() trace event. It takes three arguments
first one is the name of RCU flavour, second one is a user id
which triggeres synchronize_rcu_normal() and last one is an
event.

There are two traces in the synchronize_rcu_normal(). On entry,
when a new request is registered and on exit point when request
is completed.

Please note, CONFIG_RCU_TRACE=y is required to activate traces.

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:47:51 +02:00
Uladzislau Rezki (Sony)
988f569ae0 rcu: Reduce synchronize_rcu() latency
A call to a synchronize_rcu() can be optimized from a latency
point of view. Workloads which depend on this can benefit of it.

The delay of wakeme_after_rcu() callback, which unblocks a waiter,
depends on several factors:

- how fast a process of offloading is started. Combination of:
    - !CONFIG_RCU_NOCB_CPU/CONFIG_RCU_NOCB_CPU;
    - !CONFIG_RCU_LAZY/CONFIG_RCU_LAZY;
    - other.
- when started, invoking path is interrupted due to:
    - time limit;
    - need_resched();
    - if limit is reached.
- where in a nocb list it is located;
- how fast previous callbacks completed;

Example:

1. On our embedded devices i can easily trigger the scenario when
it is a last in the list out of ~3600 callbacks:

<snip>
  <...>-29      [001] d..1. 21950.145313: rcu_batch_start: rcu_preempt CBs=3613 bl=28
...
  <...>-29      [001] ..... 21950.152578: rcu_invoke_callback: rcu_preempt rhp=00000000b2d6dee8 func=__free_vm_area_struct.cfi_jt
  <...>-29      [001] ..... 21950.152579: rcu_invoke_callback: rcu_preempt rhp=00000000a446f607 func=__free_vm_area_struct.cfi_jt
  <...>-29      [001] ..... 21950.152580: rcu_invoke_callback: rcu_preempt rhp=00000000a5cab03b func=__free_vm_area_struct.cfi_jt
  <...>-29      [001] ..... 21950.152581: rcu_invoke_callback: rcu_preempt rhp=0000000013b7e5ee func=__free_vm_area_struct.cfi_jt
  <...>-29      [001] ..... 21950.152582: rcu_invoke_callback: rcu_preempt rhp=000000000a8ca6f9 func=__free_vm_area_struct.cfi_jt
  <...>-29      [001] ..... 21950.152583: rcu_invoke_callback: rcu_preempt rhp=000000008f162ca8 func=wakeme_after_rcu.cfi_jt
  <...>-29      [001] d..1. 21950.152625: rcu_batch_end: rcu_preempt CBs-invoked=3612 idle=....
<snip>

2. We use cpuset/cgroup to classify tasks and assign them into
different cgroups. For example "backgrond" group which binds tasks
only to little CPUs or "foreground" which makes use of all CPUs.
Tasks can be migrated between groups by a request if an acceleration
is needed.

See below an example how "surfaceflinger" task gets migrated.
Initially it is located in the "system-background" cgroup which
allows to run only on little cores. In order to speed it up it
can be temporary moved into "foreground" cgroup which allows
to use big/all CPUs:

cgroup_attach_task():
 -> cgroup_migrate_execute()
   -> cpuset_can_attach()
     -> percpu_down_write()
       -> rcu_sync_enter()
         -> synchronize_rcu()
   -> now move tasks to the new cgroup.
 -> cgroup_migrate_finish()

<snip>
         rcuop/1-29      [000] .....  7030.528570: rcu_invoke_callback: rcu_preempt rhp=00000000461605e0 func=wakeme_after_rcu.cfi_jt
    PERFD-SERVER-1855    [000] d..1.  7030.530293: cgroup_attach_task: dst_root=3 dst_id=22 dst_level=1 dst_path=/foreground pid=1900 comm=surfaceflinger
   TimerDispatch-2768    [002] d..5.  7030.537542: sched_migrate_task: comm=surfaceflinger pid=1900 prio=98 orig_cpu=0 dest_cpu=4
<snip>

"Boosting a task" depends on synchronize_rcu() latency:

- first trace shows a completion of synchronize_rcu();
- second shows attaching a task to a new group;
- last shows a final step when migration occurs.

3. To address this drawback, maintain a separate track that consists
of synchronize_rcu() callers only. After completion of a grace period
users are deferred to a dedicated worker to process requests.

4. This patch reduces the latency of synchronize_rcu() approximately
by ~30-40% on synthetic tests. The real test case, camera launch time,
shows(time is in milliseconds):

1-run 542 vs 489 improvement 9%
2-run 540 vs 466 improvement 13%
3-run 518 vs 468 improvement 9%
4-run 531 vs 457 improvement 13%
5-run 548 vs 475 improvement 13%
6-run 509 vs 484 improvement 4%

Synthetic test(no "noise" from other callbacks):
Hardware: x86_64 64 CPUs, 64GB of memory
Linux-6.6

- 10K tasks(simultaneous);
- each task does(1000 loops)
     synchronize_rcu();
     kfree(p);

default: CONFIG_RCU_NOCB_CPU: takes 54 seconds to complete all users;
patch: CONFIG_RCU_NOCB_CPU: takes 35 seconds to complete all users.

Running 60K gives approximately same results on my setup. Please note
it is without any interaction with another type of callbacks, otherwise
it will impact a lot a default case.

5. By default it is disabled. To enable this perform one of the
below sequence:

echo 1 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp
or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=1"

Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay (AMD) <neeraj.iitr10@gmail.com>
Signed-off-by: Neeraj Upadhyay (AMD) <neeraj.iitr10@gmail.com>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:47:49 +02:00
Nikita Kiryushin
3758f7d991 rcu: Fix buffer overflow in print_cpu_stall_info()
The rcuc-starvation output from print_cpu_stall_info() might overflow the
buffer if there is a huge difference in jiffies difference.  The situation
might seem improbable, but computers sometimes get very confused about
time, which can result in full-sized integers, and, in this case,
buffer overflow.

Also, the unsigned jiffies difference is printed using %ld, which is
normally for signed integers.  This is intentional for debugging purposes,
but it is not obvious from the code.

This commit therefore changes sprintf() to snprintf() and adds a
clarifying comment about intention of %ld format.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: 245a629825 ("rcu: Dump rcuc kthread status for CPUs not reporting quiescent state")
Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:43:50 +02:00
Johannes Berg
80cd613a9a rcu: Mollify sparse with RCU guard
When using "guard(rcu)();" sparse will complain, because even
though it now understands the cleanup attribute, it doesn't
evaluate the calls from it at function exit, and thus doesn't
count the context correctly.

Given that there's a conditional in the resulting code:

  static inline void class_rcu_destructor(class_rcu_t *_T)
  {
      if (_T->lock) {
          rcu_read_unlock();
      }
  }

it seems that even trying to teach sparse to evalulate the
cleanup attribute function it'd still be difficult to really
make it understand the full context here.

Suppress the sparse warning by just releasing the context in
the acquisition part of the function, after all we know it's
safe with the guard, that's the whole point of it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:43:48 +02:00
Nikita Kiryushin
cc5645fddb rcu-tasks: Fix show_rcu_tasks_trace_gp_kthread buffer overflow
There is a possibility of buffer overflow in
show_rcu_tasks_trace_gp_kthread() if counters, passed
to sprintf() are huge. Counter numbers, needed for this
are unrealistically high, but buffer overflow is still
possible.

Use snprintf() with buffer size instead of sprintf().

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: edf3775f0a ("rcu-tasks: Add count for idle tasks on offline CPUs")
Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:36:55 +02:00
Zqiang
5f48fa85fd rcu-tasks: Fix the comments for tasks_rcu_exit_srcu_stall_timer
The synchronize_srcu() has been removed by commit("rcu-tasks: Eliminate
deadlocks involving do_exit() and RCU tasks") in rcu_tasks_postscan.
This commit therefore fixes the tasks_rcu_exit_srcu_stall_timer comment.

Signed-off-by: Zqiang <qiang.zhang1211@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:36:55 +02:00
Paul E. McKenney
8db610c3bd rcu-tasks: Replace exit_tasks_rcu_start() initialization with WARN_ON_ONCE()
Because the Tasks RCU ->rtp_exit_list is initialized at rcu_init()
time while there is only one CPU running with interrupts disabled, it
is not possible for an exiting task to encounter an uninitialized list.
This commit therefore replaces the conditional initialization with
a WARN_ON_ONCE().

Reported-by: Frederic Weisbecker <frederic@kernel.org>
Closes: https://lore.kernel.org/all/ZdiNXmO3wRvmzPsr@lothringen/
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 19:36:41 +02:00
Zack Rusin
7b0062036c drm/vmwgfx: Implement virtual crc generation
crc checksums are used to validate the output. Normally they're part
of the actual display hardware but on virtual stack there's nothing
to automatically generate them.

Implement crc generation for the vmwgfx stack. This works only on
screen targets, where it's possibly to easily make sure that the
guest side contents of the surface matches the host sides output.

Just like the vblank support, crc generation can only be enabled via:
guestinfo.vmwgfx.vkms_enable = "TRUE"
option in the vmx file.

Makes IGT's kms_pipe_crc_basic pass and allows a huge number of other
IGT tests which require CRC generation of the output to actually run
on vmwgfx. Makes it possible to actually validate a lof of the kms and
drm functionality with vmwgfx.

Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Martin Krastev <martin.krastev@broadcom.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412025511.78553-3-zack.rusin@broadcom.com
2024-04-15 13:32:45 -04:00
Zack Rusin
cd2eb57df1 drm/vmwgfx: Implement virtual kms
By default vmwgfx doesn't support vblanking or crc generation which
makes it impossible to use various IGT tests to validate vmwgfx.
Implement virtual kernel mode setting, which is mainly related to
simulated vblank support.

Code is very similar to amd's vkms and the vkms module itself, except
that it's integrated with vmwgfx three different output technologies -
legacy, screen object and screen targets.

Make IGT's kms_vblank pass on vmwgfx and allows a lot of other IGT
tests to run with vmwgfx.

Support for vkms needs to be manually enabled by adding:
guestinfo.vmwgfx.vkms_enable = "TRUE"
somewhere in the vmx file, otherwise it's off by default.

Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Martin Krastev <martin.krastev@broadcom.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412025511.78553-2-zack.rusin@broadcom.com
2024-04-15 13:32:18 -04:00
Kent Overstreet
ad29cf999a bcachefs: set_btree_iter_dontneed also clears should_be_locked
This is part of a larger series cleaning up the semantics of
should_be_locked and adding assertions around it; if we don't need an
iterator/path anymore, it clearly doesn't need to be locked.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2024-04-15 13:31:15 -04:00
Chao Yu
3078e059a5 bcachefs: fix error path of __bch2_read_super()
In __bch2_read_super(), if kstrdup() fails, it needs to release memory
in sb->holder, fix to call bch2_free_super() in the error path.

Signed-off-by: Chao Yu <chao@kernel.org>
Reviewed-by: Hongbo Li <lihongbo22@huawei.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2024-04-15 13:31:15 -04:00
Linus Torvalds
3fdfcd98f0 linux_kselftest-fixes-6.9-rc5
This kselftest fixes update for Linux 6.9-rc5 consists of a fix to
 kselftest harness to prevent infinite loop triggered in an assert
 in FIXTURE_TEARDOWN and a fix to a problem seen in being able to stop
 subsystem-enable tests when sched events are being traced.
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEEPZKym/RZuOCGeA/kCwJExA0NQxwFAmYdUS0ACgkQCwJExA0N
 Qxy3oA/+JsYyCNBJigi83TxhoaMbDwTl05R+HzCYsLTHUHKCZaAq2Q5SDzF7f4xN
 3tIzP2GcpnjcspJRwB3+LYyTZgs4Ja1eHupkcIRL/n0RElLJtGpEDm4of0pdCZxO
 Z1V+CsFdW83HuVmHSKAZg6f5G7hkG1f49WtibKnuu4H2MYassjLjEhMf243ai8O/
 9M9HgbaCdDKhaXcJ8pqOXaumlgC0Oa8MffN8RTDPjtfJZU6MkZMSA5qmn6r6SNDd
 H9swYr8Q61w6cKCju6CIxowWCcbr1cKbE7RT4/tj+b1XAXkvfaKDwLDglf+brVK8
 s7LVxTIaat/ljhCX35HsyOxVcejS2p5ThDdTAnr/XU2tp/yhGp/rBTgMoPKZEHgq
 ak6D+LA2Zxj897mtq5pGddzD4aSMQh2Mik70Qdjg2OOxjqfl2Pmc9JSJkQlG4Fng
 Vy49rzZA2q0M3JUFVEJcLtkBKNCCWU0L6j4EU47jOwIHNx7oS9aB2gFfJUWVjSvg
 blnBEDCfMA/FuseRkyPb4BpKFjEmh0pOVpXcSI2irXvl5h9RLcA8g8eiU0ZWL5le
 I7msjswjL+88gNQz41aTZKfFFQIVsJjMa4UI058xNm7B8TfrXg0YKtsDaykc/iCO
 5k579aKcRx1SWOYsltTa6Qm6Sl+m9rNNzK8IrmjF4Rckfdk4cnM=
 =8nki
 -----END PGP SIGNATURE-----

Merge tag 'linux_kselftest-fixes-6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest

Pull kselftest fixes from Shuah Khan:
 "A fix to kselftest harness to prevent infinite loop triggered in an
  assert in FIXTURE_TEARDOWN and a fix to a problem seen in being able
  to stop subsystem-enable tests when sched events are being traced"

* tag 'linux_kselftest-fixes-6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
  selftests/harness: Prevent infinite loop due to Assert in FIXTURE_TEARDOWN
  selftests/ftrace: Limit length in subsystem-enable tests
2024-04-15 10:08:10 -07:00
Tvrtko Ursulin
bfc7bc5393
dma-buf: Do not build debugfs related code when !CONFIG_DEBUG_FS
There is no point in compiling in the list and mutex operations which are
only used from the dma-buf debugfs code, if debugfs is not compiled in.

Put the code in questions behind some kconfig guards and so save some text
and maybe even a pointer per object at runtime when not enabled.

Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Christian König <christian.koenig@amd.com>
Cc: linux-media@vger.kernel.org
Cc: dri-devel@lists.freedesktop.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: linux-kernel@vger.kernel.org
Cc: kernel-dev@igalia.com
Reviewed-by: T.J. Mercier <tjmercier@google.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maíra Canal <mcanal@igalia.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240328145323.68872-1-tursulin@igalia.com
2024-04-15 13:56:32 -03:00
Maíra Canal
d894ea562f
drm/fb_dma: s/drm_panic_gem_get_scanout_buffer/drm_fb_dma_get_scanout_buffer
On version 11 of the "drm/panic: Add a drm panic handler" series [1], Thomas
suggested to change the name of the function `drm_panic_gem_get_scanout_buffer`
to `drm_fb_dma_get_scanout_buffer` and this request was applied on version 12,
which is the version that landed [2]. Although the name of the function
changed on the C file, it didn't changed on the header file, leading to a
compilation error as such:

drivers/gpu/drm/imx/ipuv3/ipuv3-plane.c:780:24: error: use of undeclared
identifier 'drm_fb_dma_get_scanout_buffer'; did you mean 'drm_panic_gem_get_scanout_buffer'?
  780 |         .get_scanout_buffer = drm_fb_dma_get_scanout_buffer,
      |                               ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      |                               drm_panic_gem_get_scanout_buffer
./include/drm/drm_fb_dma_helper.h:23:5: note: 'drm_panic_gem_get_scanout_buffer'
declared here
   23 | int drm_panic_gem_get_scanout_buffer(struct drm_plane *plane,
      |     ^
1 error generated.

Fix the compilation error by changing `drm_panic_gem_get_scanout_buffer` to
`drm_fb_dma_get_scanout_buffer` on the header file.

Link: https://lore.kernel.org/dri-devel/20240328120638.468738-1-jfalempe@redhat.com/ [1]
Link: https://lore.kernel.org/dri-devel/aea2aa01-7f03-453b-8b30-8f4d90b1b47f@redhat.com/ [2]
Fixes: 879b3b6511 ("drm/fb_dma: Add generic get_scanout_buffer() for drm_panic")
Signed-off-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240415151013.3210278-1-mcanal@igalia.com
2024-04-15 13:50:40 -03:00
Anshuman Khandual
015a12a4a6 arm64/hugetlb: Fix page table walk in huge_pte_alloc()
Currently normal HugeTLB fault ends up crashing the kernel, as p4dp derived
from p4d_offset() is an invalid address when PGTABLE_LEVEL = 5. A p4d level
entry needs to be allocated when not available while walking the page table
during HugeTLB faults. Let's call p4d_alloc() to allocate such entries when
required instead of current p4d_offset().

 Unable to handle kernel paging request at virtual address ffffffff80000000
 Mem abort info:
   ESR = 0x0000000096000005
   EC = 0x25: DABT (current EL), IL = 32 bits
   SET = 0, FnV = 0
   EA = 0, S1PTW = 0
   FSC = 0x05: level 1 translation fault
 Data abort info:
   ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000
   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
 swapper pgtable: 4k pages, 52-bit VAs, pgdp=0000000081da9000
 [ffffffff80000000] pgd=1000000082cec003, p4d=0000000082c32003, pud=0000000000000000
 Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP
 Modules linked in:
 CPU: 1 PID: 108 Comm: high_addr_hugep Not tainted 6.9.0-rc4 #48
 Hardware name: Foundation-v8A (DT)
 pstate: 01402005 (nzcv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--)
 pc : huge_pte_alloc+0xd4/0x334
 lr : hugetlb_fault+0x1b8/0xc68
 sp : ffff8000833bbc20
 x29: ffff8000833bbc20 x28: fff000080080cb58 x27: ffff800082a7cc58
 x26: 0000000000000000 x25: fff0000800378e40 x24: fff00008008d6c60
 x23: 00000000de9dbf07 x22: fff0000800378e40 x21: 0004000000000000
 x20: 0004000000000000 x19: ffffffff80000000 x18: 1ffe00010011d7a1
 x17: 0000000000000001 x16: ffffffffffffffff x15: 0000000000000001
 x14: 0000000000000000 x13: ffff8000816120d0 x12: ffffffffffffffff
 x11: 0000000000000000 x10: fff00008008ebd0c x9 : 0004000000000000
 x8 : 0000000000001255 x7 : fff00008003e2000 x6 : 00000000061d54b0
 x5 : 0000000000001000 x4 : ffffffff80000000 x3 : 0000000000200000
 x2 : 0000000000000004 x1 : 0000000080000000 x0 : 0000000000000000
 Call trace:
 huge_pte_alloc+0xd4/0x334
 hugetlb_fault+0x1b8/0xc68
 handle_mm_fault+0x260/0x29c
 do_page_fault+0xfc/0x47c
 do_translation_fault+0x68/0x74
 do_mem_abort+0x44/0x94
 el0_da+0x2c/0x9c
 el0t_64_sync_handler+0x70/0xc4
 el0t_64_sync+0x190/0x194
 Code: aa000084 cb010084 b24c2c84 8b130c93 (f9400260)
 ---[ end trace 0000000000000000 ]---

Cc: Will Deacon <will@kernel.org>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Fixes: a6bbf5d4d9 ("arm64: mm: Add definitions to support 5 levels of paging")
Reported-by: Dev Jain <dev.jain@arm.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
Reviewed-by: Ryan Roberts <ryan.roberts@arm.com>
Link: https://lore.kernel.org/r/20240415094003.1812018-1-anshuman.khandual@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2024-04-15 17:44:38 +01:00
Paul E. McKenney
10c9e40f29 rcu: Remove redundant CONFIG_PROVE_RCU #if condition
The #if condition controlling the rcu_preempt_sleep_check() definition
has a redundant check for CONFIG_PREEMPT_RCU, which is already checked
for by an enclosing #ifndef.  This commit therefore removes this redundant
condition from the inner #if.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 18:12:20 +02:00
Paul E. McKenney
fc2897d2ab rcu: Inform KCSAN of one-byte cmpxchg() in rcu_trc_cmpxchg_need_qs()
Tasks Trace RCU needs a single-byte cmpxchg(), but no such thing exists.
Therefore, rcu_trc_cmpxchg_need_qs() emulates one using field substitution
and a four-byte cmpxchg(), such that the other three bytes are always
atomically updated to their old values.  This works, but results in
false-positive KCSAN failures because as far as KCSAN knows, this
cmpxchg() operation is updating all four bytes.

This commit therefore encloses the cmpxchg() in a data_race() and adds
a single-byte instrument_atomic_read_write(), thus telling KCSAN exactly
what is going on so as to avoid the false positives.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Marco Elver <elver@google.com>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 18:12:18 +02:00
Jens Wiklander
08530a2aa9 firmware: arm_ffa: Add support for handling notification pending interrupt(NPI)
The FF-A uses the notification pending interrupt to inform the receiver
that it has a pending notification. This is a virtual interrupt and is
used by the following type of receivers:
1. A guest/VM running under a hypervisor.
2. An S-EL1 SP running under a S-EL2 SPMC.

The rules that govern the properties of the NPI are the same as the
rules for the SRI with couple of exceptions. Both SRI and NPI can be
supported simultaneously.

The handling of NPI is also same as the handling of notification for the
self/primary VM with ID 0 except the absence of global notification.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Link: https://lore.kernel.org/r/20240411-ffa_npi_support-v2-3-927a670254e6@arm.com
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
2024-04-15 16:56:17 +01:00
Sudeep Holla
f936c24255 firmware: arm_ffa: Refactor SRI handling in prepartion to add NPI support
In preparation to support handling of Notification Pending Interrupt(NPI)
in addition to the existing support for Schedule Receiver Interrupt(SRI),
refactor the code around SRI handling so that NPI support can reuse some
of it. This change shouldn't have any functionality impact. It neither
adds the support for NPIs nor changes any SRI support.

Tested-by: Jens Wiklander <jens.wiklander@linaro.org>
Link: https://lore.kernel.org/r/20240411-ffa_npi_support-v2-2-927a670254e6@arm.com
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
2024-04-15 16:56:03 +01:00
Jens Wiklander
2b9c66d1ab firmware: arm_ffa: Skip creation of the notification bitmaps
When the FF-A driver is running inside a guest VM under an hypervisor,
the driver/guest VM doesn't have the permission/capability to request
the creation of notification bitmaps. For a VM, the hypervisor reserves
memory for its VM and hypervisor framework notification bitmaps and the
SPMC reserves memory for its SP and SPMC framework notification bitmaps
before the hypervisor initializes it.

The hypervisor does not initialize a VM if memory cannot be reserved
for all its notification bitmaps. So the creation of all the necessary
bitmaps are already done when the driver initialises and hence it can be
skipped. We rely on FFA_FEATURES(FFA_NOTIFICATION_BITMAP_CREATE) to fail
when running in the guest to handle this in the driver.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Link: https://lore.kernel.org/r/20240411-ffa_npi_support-v2-1-927a670254e6@arm.com
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
2024-04-15 16:55:54 +01:00
Zack Rusin
d4c972bff3 drm/vmwgfx: Sort primary plane formats by order of preference
The table of primary plane formats wasn't sorted at all, leading to
applications picking our least desirable formats by defaults.

Sort the primary plane formats according to our order of preference.

Nice side-effect of this change is that it makes IGT's kms_atomic
plane-invalid-params pass because the test picks the first format
which for vmwgfx was DRM_FORMAT_XRGB1555 and uses fb's with odd sizes
which make Pixman, which IGT depends on assert due to the fact that our
16bpp formats aren't 32 bit aligned like Pixman requires all formats
to be.

Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Fixes: 36cc79bc90 ("drm/vmwgfx: Add universal plane support")
Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list@broadcom.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v4.12+
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412025511.78553-6-zack.rusin@broadcom.com
2024-04-15 11:48:57 -04:00
Zack Rusin
a60ccade88 drm/vmwgfx: Fix crtc's atomic check conditional
The conditional was supposed to prevent enabling of a crtc state
without a set primary plane. Accidently it also prevented disabling
crtc state with a set primary plane. Neither is correct.

Fix the conditional and just driver-warn when a crtc state has been
enabled without a primary plane which will help debug broken userspace.

Fixes IGT's kms_atomic_interruptible and kms_atomic_transition tests.

Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Fixes: 06ec41909e ("drm/vmwgfx: Add and connect CRTC helper functions")
Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list@broadcom.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v4.12+
Reviewed-by: Ian Forbes <ian.forbes@broadcom.com>
Reviewed-by: Martin Krastev <martin.krastev@broadcom.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412025511.78553-5-zack.rusin@broadcom.com
2024-04-15 11:48:48 -04:00
Zack Rusin
b32233acce drm/vmwgfx: Fix prime import/export
vmwgfx never supported prime import of external buffers. Furthermore the
driver exposes two different objects to userspace: vmw_surface's and
gem buffers but prime import/export only worked with vmw_surfaces.

Because gem buffers are used through the dumb_buffer interface this meant
that the driver created buffers couldn't have been prime exported or
imported.

Fix prime import/export. Makes IGT's kms_prime pass.

Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Fixes: 8afa13a058 ("drm/vmwgfx: Implement DRIVER_GEM")
Cc: <stable@vger.kernel.org> # v6.6+
Reviewed-by: Martin Krastev <martin.krastev@broadcom.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412025511.78553-4-zack.rusin@broadcom.com
2024-04-15 11:48:36 -04:00
Thorsten Leemhuis
8d939ae349 docs: verify/bisect: stable regressions: first stable, then mainline
Rearrange the instructions so that readers facing a regression within a
stable or longterm series first test its latest release before testing
mainline. This is less scary for some people. It also reduces the chance
that something goes sideways for readers that compile their first
kernel, as mainline can cause slightly more trouble.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/efd3cb9c68db450091021326bf9c334553df0ec2.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:56 -06:00
Thorsten Leemhuis
2bcfd71e8d docs: verify/bisect: describe how to use a build host
Describe how to build kernels on another system (with and without
cross-compiling), as building locally can be quite painfully on some
slow systems. This is done in an add-on section, as it would make the
step-by-step guide to complicated if this special case would be
described there.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/288160cb4769e46a3280250ca71da0abc4aa002d.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:56 -06:00
Thorsten Leemhuis
a421835a2a docs: verify/bisect: explain testing reverts, patches and newer code
Rename 'Supplementary tasks' to 'Complementary tasks' while introducing
a section 'Optional tasks: test reverts, patches, or later versions':
the latter is something readers occasionally will have to do after
reporting a bug and thus is best covered here.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/dacf26a4c48e9e8f04ecbc77e0a74c9b2a6a1103.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:56 -06:00
Thorsten Leemhuis
453de3207f docs: verify/bisect: proper headlines and more spacing
Various small improvements and fixes:

* Separate ref links from their target with a space for better
  readability.

* Add a proper heading for the note at the end of the step-by-step
  guide.

* Use proper 3rd and 4th level headlines in the reference section and
  add short intros for the 2nd level headlines that lacked one.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/f59f0f235a2192ed93899a7338153e4cb71075f0.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:56 -06:00
Thorsten Leemhuis
932c9a5398 docs: verify/bisect: add and fetch stable branches ahead of time
Add and fetch all required stable branches ahead of time. This fixes a
bug, as readers that wanted to bisect a regression within a stable or
longterm series otherwise did not have them available at the right time.
This way also matches the flow somewhat better and avoids some "if you
haven't already added it" phrases that otherwise become necessary in
future changes.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/57dcf312959476abe6151bf3d35eb79e3e9a83d1.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:56 -06:00
Thorsten Leemhuis
abbb99301e docs: verify/bisect: use git switch, tag kernel, and various fixes
Various small improvements and fixes:

* Use the more modern 'git switch' instead of 'git checkout', which
  makes it more obvious what's happening (among others due to the
  --discard-changes parameter that is more clear than --force).

* Provide a hint how a mainline version number and one from a stable
  series look like.

* When trying to validate the bisection result with a revert, add a
  special tag to facilitate the identification.

* Sync version numbers used in various examples for consistency: stick
  to 6.0.13, 6.0.15, and 6.1.5.

* Fix a few typos and oddities.

Signed-off-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/85029aa004447b0eeb5043fb014630f2acafacec.1712647788.git.linux@leemhuis.info
2024-04-15 09:41:55 -06:00
AngeloGioacchino Del Regno
fb7c3d8ba0 dt-bindings: pwm: mediatek,pwm-disp: Document power-domains property
Allow the power-domains property to the PWM_DISP block as on some SoCs
this does need at most one power domain.

Fixes: b09b179bac ("dt-bindings: pwm: Convert pwm-mtk-disp.txt to mediatek,pwm-disp.yaml format")
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240404081808.92199-1-angelogioacchino.delregno@collabora.com
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
2024-04-15 17:28:13 +02:00
Raag Jadav
a2ac1cbc53 pwm: dwc: allow suspend/resume for 16 channels
With 16 channel pwm support, we're registering two instances of pwm_chip
with 8 channels each. We need to update PM functions to use both instances
of pwm_chip during power state transitions.

Introduce struct dwc_pwm_drvdata and use it as driver_data, which will
maintain both instances of pwm_chip along with dwc_pwm_info and allow us
to use them inside suspend/resume handles.

Fixes: ebf2c89eb9 ("pwm: dwc: Add 16 channel support for Intel Elkhart Lake")
Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20240415074051.14681-1-raag.jadav@intel.com
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
2024-04-15 17:28:13 +02:00
Hannah Peuckmann
e0503d47e9 riscv: dts: starfive: visionfive 2: Remove non-existing I2S hardware
This partially reverts
commit 92cfc35838 ("riscv: dts: starfive: Add the nodes and pins of I2Srx/I2Stx0/I2Stx1")

This added device tree nodes for I2S hardware that is not actually on the
VisionFive 2 board, but connected on the 40pin header. Many different extension
boards could be added on those pins, so this should be handled by overlays
instead.
This also conflicts with the TDM node which also attempts to grab GPIO 44:

  starfive-jh7110-sys-pinctrl 13040000.pinctrl: pin GPIO44 already requested by 10090000.tdm; cannot claim for 120c0000.i2s

Fixes: 92cfc35838 ("riscv: dts: starfive: Add the nodes and pins of I2Srx/I2Stx0/I2Stx1")
Signed-off-by: Hannah Peuckmann <hannah.peuckmann@canonical.com>
Reviewed-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
2024-04-15 16:16:11 +01:00
Hannah Peuckmann
dcde4e97b1 riscv: dts: starfive: visionfive 2: Remove non-existing TDM hardware
This partially reverts
commit e7c304c034 ("riscv: dts: starfive: jh7110: add the node and pins configuration for tdm")

This added device tree nodes for TDM hardware that is not actually on the
VisionFive 2 board, but connected on the 40pin header. Many different extension
boards could be added on those pins, so this should be handled by overlays
instead.
This also conflicts with the I2S node which also attempts to grab GPIO 44:

  starfive-jh7110-sys-pinctrl 13040000.pinctrl: pin GPIO44 already requested by 10090000.tdm; cannot claim for 120c0000.i2s

Fixes: e7c304c034 ("riscv: dts: starfive: jh7110: add the node and pins configuration for tdm")
Signed-off-by: Hannah Peuckmann <hannah.peuckmann@canonical.com>
Reviewed-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
2024-04-15 16:16:10 +01:00
Christian König
b6976f323a drm/ttm: stop pooling cached NUMA pages v2
We only pool write combined and uncached allocations because they
require extra overhead on allocation and release.

If we also pool cached NUMA it not only means some extra unnecessary
overhead, but also that under memory pressure it can happen that
pages from the wrong NUMA node enters the pool and are re-used
over and over again.

This can lead to performance reduction after running into memory
pressure.

v2: restructure and cleanup the code a bit from the internal hack to
    test this.

Signed-off-by: Christian König <christian.koenig@amd.com>
Fixes: 4482d3c94d ("drm/ttm: add NUMA node id to the pool")
CC: stable@vger.kernel.org
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240415134821.1919-1-christian.koenig@amd.com
2024-04-15 16:57:41 +02:00
Alexey Dobriyan
e964fc7757
vfs, swap: compile out IS_SWAPFILE() on swapless configs
No swap support -- no swapfiles possible.

Signed-off-by: Alexey Dobriyan (Yandex) <adobriyan@gmail.com>
Link: https://lore.kernel.org/r/2391c7f5-0f83-4188-ae56-4ec7ccbf2576@p183
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-04-15 16:56:14 +02:00
Haifeng Xu
5b15f3fb89 slub: Set __GFP_COMP in kmem_cache by default
Now the __GFP_COMP is set only if the higher-order is not 0. However,
__GFP_COMP flag can be set unconditionally because compound page can
not be created in the order-0 case. And this can also simplify the code
a bit (no need to check the order is 0 or not).

Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
2024-04-15 16:50:28 +02:00
Paul E. McKenney
ae2b217ab5 rcu: Make hotplug operations track GP state, not flags
Currently, there are rcu_data structure fields named ->rcu_onl_gp_seq
and ->rcu_ofl_gp_seq that track the rcu_state.gp_flags field at the
time of the corresponding CPU's last online or offline operation,
respectively.  However, this information is not particularly useful.
It would be better to instead track the grace period state kept
in rcu_state.gp_state.  This would also be consistent with the
initialization in rcu_boot_init_percpu_data(), which is to RCU_GP_CLEANED
(an rcu_state.gp_state value), and also with the diagnostics in
rcu_implicit_dynticks_qs(), whose format is consistent with an integer,
not a bitmask.

This commit therefore makes this change and changes the names to
->rcu_onl_gp_flags and ->rcu_ofl_gp_flags, respectively.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 16:48:28 +02:00
Paul E. McKenney
09e077cf22 rcu: Mark loads from rcu_state.n_online_cpus
The rcu_state.n_online_cpus value is only ever updated by CPU-hotplug
operations, which are serialized.  However, this value is read locklessly.
This commit therefore marks those reads.  While in the area, it also
adds ASSERT_EXCLUSIVE_WRITER() calls just in case parallel CPU hotplug
becomes a thing.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 16:48:26 +02:00
Paul E. McKenney
a542d116ba rcu: Mark writes to rcu_sync ->gp_count field
The rcu_sync structure's ->gp_count field is updated under the protection
of ->rss_lock, but read locklessly, and KCSAN noted the data race.
This commit therefore uses WRITE_ONCE() to do this update to clearly
document its racy nature.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 16:28:46 +02:00
Paul E. McKenney
c90b9e4978 rcu: Bring diagnostic read of rcu_state.gp_flags into alignment
This commit adds READ_ONCE() to a lockless diagnostic read from
rcu_state.gp_flags to avoid giving the compiler any chance whatsoever
of confusing the diagnostic state printed.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
2024-04-15 16:28:44 +02:00
Stafford Horne
4dc70e1aad openrisc: Move FPU state out of pt_regs
My original, naive, FPU support patch had the FPCSR register stored
during both the *mode switch* and *context switch*.  This is wasteful.

Also, the original patches did not save the FPU state when handling
signals during the system call fast path.

We fix this by moving the FPCSR state to thread_struct in task_struct.
We also introduce new helper functions save_fpu and restore_fpu which
can be used to sync the FPU with thread_struct.  These functions are now
called when needed:

 - Setting up and restoring sigcontext when handling signals
 - Before and after __switch_to during context switches
 - When handling FPU exceptions
 - When reading and writing FPU register sets

In the future we can further optimize this by doing lazy FPU save and
restore.  For example, FPU sync is not needed when switching to and from
kernel threads (x86 does this).  FPU save and restore does not need to
be done two times if we have both rescheduling and signal work to do.
However, since OpenRISC FPU state is a single register, I leave these
optimizations for future consideration.

Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00
Stafford Horne
1f33446d0e openrisc: Add FPU config
Allow disabling FPU related code sequences to save space.

Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00
Stafford Horne
c88cfb5cea openrisc: traps: Don't send signals to kernel mode threads
OpenRISC exception handling sends signals to user processes on floating
point exceptions and trap instructions (for debugging) among others.
There is a bug where the trap handling logic may send signals to kernel
threads, we should not send these signals to kernel threads, if that
happens we treat it as an error.

This patch adds conditions to die if the kernel receives these
exceptions in kernel mode code.

Fixes: 27267655c5 ("openrisc: Support floating point user api")
Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00
Stafford Horne
ee7e551d28 openrisc: traps: Remove calls to show_registers before die
The die function calls show_registers unconditionally.  Remove calls to
show_registers before calling die to avoid printing all registers and
stack status two times during a crash.

This was found when testing kernel trap and floating point exception
handling.

Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00
Stafford Horne
c8fdf82ee3 openrisc: traps: Convert printks to pr_<level> macros
The pr_* macros are the convention and my upcoming patches add even more
printk's.  Use this opportunity to convert the printks in this file to
the pr_* macros to avoid patch check warnings.

Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00
Stafford Horne
23c6e901c7 openrisc: Add support for more module relocations
When testing modules in OpenRISC I found R_OR1K_AHI16 (signed adjusted
high 16-bit) and R_OR1K_SLO16 (split low 16-bit) relocations are used in
modules but not implemented yet.

This patch implements the relocations. I have tested with a few modules.

Signed-off-by: Stafford Horne <shorne@gmail.com>
2024-04-15 15:20:39 +01:00