snd_timer_close_locked() marks an instance with SNDRV_TIMER_IFLG_DEAD
and returns early when the flag is already set, but the flag is never
cleared again. A completed close ends in remove_slave_links(), which
leaves timeri->timer NULL, so a second close is already harmless through
the timer == NULL path; the early return can only be reached by an
instance that was opened again in between. For such an instance the
close unlinks nothing, so snd_timer_instance_free() frees an object that
is still on timer->open_list_head, still on snd_timer_master_list if it
was opened with a slave key, still owns any adopted slaves, and still
holds its timer and module references.
snd_seq_timer_open() reopens an instance exactly like that: it retries
its fallback open on the same object after a failure that has already
run snd_timer_close_locked() internally. An unprivileged user with
access to /dev/snd/timer and /dev/snd/seq can force that failure, since
snd_timer_check_master() returns -EBUSY when a pending slave matches the
new master's (slave_class, slave_id) key and the target timer has
reached max_instances, and SNDRV_TIMER_IOCTL_SELECT with dev_class =
SNDRV_TIMER_CLASS_SLAVE keeps the caller-supplied dev_sclass, so a
sequencer queue's key can be forged. The freed instance is afterwards
dereferenced by any further snd_timer_open() on that timer, by
snd_timer_check_slave(), and by /proc/asound/timers, which faults on the
stale ti->owner pointer.
The flag only has to be visible while the close is in progress, which is
all its other users need. Clear it in remove_slave_links(), under the
same timer->lock that sets it, once the instance is off every list.
Fixes: da3039e91d ("ALSA: timer: Forcibly close timer instances at closing")
Cc: stable@vger.kernel.org
Assisted-by: Claude:claude-opus-5
Signed-off-by: Norbert Szetei <norbert@doyensec.com>
Link: https://patch.msgid.link/CA41AA48-75BF-45E9-A36D-3A5D2F124F60@doyensec.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_timer_close_locked() drains the closing instance's own in-flight
callback (IFLG_CALLBACK) before freeing it, but not its slaves'. When a
master instance is closed, remove_slave_links() clears each slave's
->timer; the slave's own close then reads timer == NULL and takes the
branch that skips the drain entirely (snd_timer_stop_slave() also no-ops
on a NULL timer). So a slave whose callback is still running when the
master is closed is freed underneath the live callback, leading to
use-after-free.
Drain the slaves too before remove_slave_links() severs them.
snd_timer_stop() has already taken this instance off the active list, so
no new slave callback can be queued. Take the slaves off the ack list so
a pending one can't fire either, then wait for any that is already in
flight.
Fixes: 37745918e0 ("ALSA: timer: Introduce virtual userspace-driven timers")
Cc: stable@vger.kernel.org
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Norbert Szetei <norbert@doyensec.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/D26598EB-DBF7-4D76-9F71-8E4BD59822D4@doyensec.com
The userspace-driven timer (utimer) TRIGGER ioctl calls
snd_timer_interrupt() directly with no serialization, so two threads
triggering the same utimer can run snd_timer_interrupt() on one
snd_timer concurrently.
snd_timer_process_callbacks() drops timer->lock around each instance
callback and marks the in-flight callback with the single
SNDRV_TIMER_IFLG_CALLBACK bit; snd_timer_close_locked() waits on that
bit to drain an in-flight callback before freeing the instance. The bit
cannot represent two concurrent callbacks: when a second interrupt
re-queues an instance whose callback is still running, both run at once,
the first to finish clears the bit, and the close-path drain then frees
the instance (and its callback_data) while the other callback is still
live - a use-after-free reachable by any user able to open
/dev/snd/timer, both via a user timer instance and via a sequencer queue
timer bound to the utimer.
snd_timer_interrupt() sets IFLG_CALLBACK before dropping timer->lock, so
a concurrent interrupt already observes it under the lock. Skip
re-queuing an instance (and its slaves) to the ack/sack list while its
callback is in flight; the accumulated pticks are delivered on the next
tick, so no event is lost.
Fixes: 37745918e0 ("ALSA: timer: Introduce virtual userspace-driven timers")
Cc: stable@vger.kernel.org
Suggested-by: Takashi Iwai <tiwai@suse.de>
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Norbert Szetei <norbert@doyensec.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/6F9B6501-8E65-4265-B02C-7EFB240D1664@doyensec.com
Although we've covered the races around the timer object assignment
and release for timer instances, there are still races at starting or
stopping the timer instance. They refer to timeri->timer without
lock, hence they can still trigger UAFs.
For addressing it, this patch changes the existing slave_active_lock
spinlock to timeri_lock rwlock. It's a global rwlock applied as
read-lock when snd_timer_start() & co are called as well as
snd_timeri_timer_get() is called. In turn, the places where
timeri->timer is assigned or released are covered by the write-lock.
The patch replaces spinlock_irqsave with spinlock in a couple of
spaces because they are now already protected by timeri_lock, too.
Reported-by: Kyle Zeng <kylebot@openai.com>
Link: https://patch.msgid.link/20260614090714.773216-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
So far we've tried to address UAFs in ALSA timer code by applying the
locks at various places, but the fundamental problem is that the timer
object may be released while the belonging timer instance objects are
still present and accessing to it. This patch is a more proper fix to
address that issue, namely, by refcounting and keeping the timer
object.
The basic implementation is to use kref for the refcount of the timer
object, and take/release the reference at assigning/releasing the
instance, as well as at referring from ioctls or ALSA sequencer code.
The reference from ioctl or ALSA sequencer is abstracted with
snd_timeri_timer auto-cleanup.
Note that this change assumes that the code already took the fix
commit da3039e91d ("ALSA: timer: Forcibly close timer instances at
closing"); otherwise the refcount may be unbalanced when the timer is
freed while slave instances are still present.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/20260609115100.806869-2-tiwai@suse.de
At releasing a timer object, e.g. when a userspace timer
(CONFIG_SND_UTIMER) gets closed and snd_timer_free() is called, it
tries to detach the timer instances and release the resources.
However, it's still possible that other in-flight tasks are holding
the timer instance where the to-be-deleted timer object is associated,
and this may lead to racy accesses.
Fortunately, most of ioctls dealing with the timer instance list
already have the protection with register_mutex, and this also avoids
such races. But, SNDRV_TIMER_IOCTL_PARAMS isn't protected, hence the
concurrent ioctl may lead to use-after-free.
This patch just adds the guard with register_mutex to protect
snd_timer_user_params() for covering the code path as a quick
workaround. It's no hot-path but rather a rarely issued ioctl, so the
performance penalty doesn't matter.
Reported-by: Kyle Zeng <kylebot@openai.com>
Tested-by: Kyle Zeng <kylebot@openai.com>
Cc: <stable@vger.kernel.org>
Link: https://patch.msgid.link/20260606161145.1933447-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When snd_timer object is freed via snd_timer_free() and still pending
snd_timer_instance objects are assigned to the timer object, it tries
to unlink all instances and just set NULL to each ti->timer, then
releases the resources immediately. The problem is, however, when
there are slave timer instances that are associated with a master
instance linked to this timer: namely, those slave instances still
point to the freed timer object although the master instance is
unlinked, which may lead to user-after-free. The bug can be easily
triggered particularly when a new userspace-driven timers
(CONFIG_SND_UTIMER) is involved, since it can create and delete the
timer object via a simple file open/close, while the other
applications may keep accessing to that timer.
This patch is an attempt to paper over the problem above: now instead
of just unlinking, call snd_timer_close[_locked]() forcibly for each
pending timer instance, so that all assigned slave timer instances are
properly detached, too. Since snd_timer_close() might be called later
by the driver that created that instance, the check of
SNDRV_TIMER_IFLG_DEAD is added at the beginning, too.
Reported-by: Kyle Zeng <kylebot@openai.com>
Tested-by: Kyle Zeng <kylebot@openai.com>
Fixes: 37745918e0 ("ALSA: timer: Introduce virtual userspace-driven timers")
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260606161145.1933447-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_timer_dev_register() walks snd_timer_list looking for the
ordered insertion point and on loop fall-through passes
&timer1->device_list to list_add_tail():
list_for_each_entry(timer1, &snd_timer_list, device_list) {
...
break; /* on found-position */
...
}
list_add_tail(&timer->device_list, &timer1->device_list);
When the loop walks all entries without break, timer1 is
past-the-end. &timer1->device_list aliases &snd_timer_list (the
list head) via container_of offset cancellation, so the insert
lands at the list tail. That is the intended behaviour, but the
access is undefined per C11 even though it works in practice.
Track an explicit insert_before pointer initialised to the list
head and overwritten to &timer1->device_list only when the loop
breaks early. The observable behaviour is unchanged.
Fixes: 9244b2c307 ("[ALSA] alsa core: convert to list_for_each_entry*")
Signed-off-by: Maoyi Xie <maoyixie.tju@gmail.com>
Link: https://patch.msgid.link/20260518194023.1667857-2-maoyixie.tju@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_timer_check_slave() still walks all registered timers and all open
timer instances to find a matching master for a newly opened slave.
Maintain a global list of open master instances that can accept slave
links and use it for the slave lookup path instead. This keeps the
existing matching semantics while avoiding the nested walk over
snd_timer_list and each timer open_list_head.
The reverse path in snd_timer_check_master() already scans only the
pending slave list, so this makes both lookup paths closer in shape
without changing the master/slave linking logic.
Signed-off-by: Cássio Gabriel <cassiogabrielcontato@gmail.com>
Link: https://patch.msgid.link/20260316-alsa-timer-master-list-v1-1-fb95e547110a@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
We used to have a variable declaration with __free() initialized with
NULL. This was to keep the old coding style rule, but recently it's
relaxed and rather recommends to follow the new rule to declare in
place of use for __free() -- which avoids potential deadlocks or UAFs
with nested cleanups.
Although the current code has no bug, per se, let's follow the new
standard and move the declaration to the place of assignment (or
directly assign the allocated result) instead of NULL initializations.
Fixes: ed96f6394e ("ALSA: timer: Use automatic cleanup of kfree()")
Fixes: 37745918e0 ("ALSA: timer: Introduce virtual userspace-driven timers")
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/20251216140634.171890-8-tiwai@suse.de
timer_delete[_sync]() replaces del_timer[_sync](). Convert the whole tree
over and remove the historical wrapper inlines.
Conversion was done with coccinelle plus manual fixups where necessary.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
no_llseek had been defined to NULL two years ago, in commit 868941b144
("fs: remove no_llseek")
To quote that commit,
At -rc1 we'll need do a mechanical removal of no_llseek -
git grep -l -w no_llseek | grep -v porting.rst | while read i; do
sed -i '/\<no_llseek\>/d' $i
done
would do it.
Unfortunately, that hadn't been done. Linus, could you do that now, so
that we could finally put that thing to rest? All instances are of the
form
.llseek = no_llseek,
so it's obviously safe.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We used to wrap with no_free_ptr() for the return value from
memdup_user() with errors where the auto cleanup is applied. This was
a workaround because the initial implementation of kfree auto-cleanup
checked only NULL pointers.
Since recently, though, the kfree auto-cleanup checks with
IS_ERR_OR_NULL() (by the commit cd7eb8f83f ("mm/slab: make
__free(kfree) accept error pointers")), hence those workarounds became
superfluous. Let's drop them now.
Link: https://patch.msgid.link/20240902075246.3743-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Implement two ioctl calls in order to support virtual userspace-driven
ALSA timers.
The first ioctl is SNDRV_TIMER_IOCTL_CREATE, which gets the
snd_timer_uinfo struct as a parameter and puts a file descriptor of a
virtual timer into the `fd` field of the snd_timer_unfo structure. It
also updates the `id` field of the snd_timer_uinfo struct, which
provides a unique identifier for the timer (basically, the subdevice
number which can be used when creating timer instances).
This patch also introduces a tiny id allocator for the userspace-driven
timers, which guarantees that we don't have more than 128 of them in the
system.
Another ioctl is SNDRV_TIMER_IOCTL_TRIGGER, which allows us to trigger
the virtual timer (and calls snd_timer_interrupt for the timer under
the hood), causing all of the timer instances binded to this timer to
execute their callbacks.
The maximum amount of ticks available for the timer is 1 for the sake of
simplicity of the userspace API. 'start', 'stop', 'open' and 'close'
callbacks for the userspace-driven timers are empty since we don't
really do any hardware initialization here.
Suggested-by: Axel Holzinger <aholzinger@gmx.de>
Signed-off-by: Ivan Orlov <ivan.orlov0322@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/20240813120701.171743-4-ivan.orlov0322@gmail.com
We can simplify the code gracefully with new guard() macro and co for
automatic cleanup of locks.
For making changes easier, some functions widen the application of
register_mutex, but those shouldn't influence on any actual
performance.
Also, one code block was factored out as a function so that guard()
can be applied cleanly without much indentation.
There are still a few remaining explicit spin_lock/unlock calls, and
those are for the places where we do temporary unlock/relock, which
doesn't fit well with the guard(), so far.
Only the code refactoring, and no functional changes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20240227085306.9764-4-tiwai@suse.de
There are common patterns where a temporary buffer is allocated and
freed at the exit, and those can be simplified with the recent cleanup
mechanism via __free(kfree).
No functional changes, only code refactoring.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20240222111509.28390-5-tiwai@suse.de
When the timer instance was add into ack_list but was not currently in
process, the user could stop it via snd_timer_stop1() without delete it
from the ack_list. Then the user could free the timer instance and when
it was actually processed UAF occurred.
This issue could be reproduced via testcase snd_timer01 in ltp - running
several instances of that testcase at the same time.
What I actually met was that the ack_list of the timer broken and the
kernel went into deadloop with irqoff. That could be detected by
hardlockup detector on board or when we run it on qemu, we could use gdb
to dump the ack_list when the console has no response.
To fix this issue, we delete the timer instance from ack_list and
active_list unconditionally in snd_timer_stop1().
Signed-off-by: Wang Wensheng <wangwensheng4@huawei.com>
Suggested-by: Takashi Iwai <tiwai@suse.de>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20211103033517.80531-1-wangwensheng4@huawei.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
strlcpy is deprecated. see: Documentation/process/deprecated.rst
Change the calls that do not use the strlcpy return value to the
preferred strscpy.
Done with cocci script:
@@
expression e1, e2, e3;
@@
- strlcpy(
+ strscpy(
e1, e2, e3);
This cocci script leaves the instances where the return value is
used unchanged.
After this patch, sound/ has 3 uses of strlcpy() that need to be
manually inspected for conversion and changed one day.
$ git grep -w strlcpy sound/
sound/usb/card.c: len = strlcpy(card->longname, s, sizeof(card->longname));
sound/usb/mixer.c: return strlcpy(buf, p->name, buflen);
sound/usb/mixer.c: return strlcpy(buf, p->names[index], buflen);
Miscellenea:
o Remove trailing whitespace in conversion of sound/core/hwdep.c
Link: https://lore.kernel.org/lkml/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/22b393d1790bb268769d0bab7bacf0866dcb0c14.camel@perches.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The tasklet is an old API that should be deprecated, usually can be
converted to another decent API. In ALSA core timer API, the
callbacks can be offlined to a tasklet when a flag is set in the timer
backend. It can be achieved gracefully with a work queued in the
high-prio system workqueue.
This patch replaces the usage of tasklet in ALSA timer API with a
simple work. Currently the tasklet feature is used only in the system
timer and hrtimer backends, so both are patched to use the new flag
name SNDRV_TIMER_HW_WORK, too.
Link: https://lore.kernel.org/r/20200903104131.21097-3-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Back-merge to apply the tasklet conversion patches that are based
on the already applied tasklet API changes on 5.9-rc4.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
GCC reports a warning with W=1:
sound/core/timer.c: In function ‘snd_timer_user_read’:
sound/core/timer.c:2219:19: warning: initialized field overwritten
[-Woverride-init]
2219 | .tstamp_sec = tread->tstamp_nsec,
| ^~~~~
sound/core/timer.c:2219:19: note: (near initialization for
‘(anonymous).tstamp_sec’)
Assigning nsec values to sec fields is problematic in general, even
more so when the initial goal was to survive the 2030 timer
armageddon.
Fix by using the proper field in the initialization
Cc: Baolin Wang <baolin.wang@linaro.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Fixes: 07094ae6f9 ("ALSA: Avoid using timespec for struct snd_timer_tread")
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200111203325.20498-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The struct snd_timer_tread will use 'timespec' type variables to record
timestamp, which is not year 2038 safe on 32bits system.
Since the struct snd_timer_tread is passed through read() rather than
ioctl(), and the read syscall has no command number that lets us pick
between the 32-bit or 64-bit version of this structure.
Thus we introduced one new command SNDRV_TIMER_IOCTL_TREAD64 and new
struct snd_timer_tread64 replacing timespec with s64 type to handle
64bit time_t. That means we will set tu->tread = TREAD_FORMAT_64BIT
when user space has a 64bit time_t, then we will copy to user with
struct snd_timer_tread64. Otherwise we will use 32bit time_t variables
when copying to user.
Moreover this patch replaces timespec type with timespec64 type and
related y2038 safe APIs.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
struct snd_timer_status uses 'timespec' type variables to record
timestamp, which will be changed to an incompatible layout with
updated user space using 64-bit time_t.
To handle both the old and the new layout on 32-bit architectures,
this patch introduces 'struct snd_timer_status32' and 'struct snd_timer_status64'
to handle 32bit time_t and 64bit time_t in native mode and compat mode,
which replaces timespec with s64 type.
When glibc changes time_t to 64-bit, any recompiled program will issue
ioctl commands that the kernel does not understand without this patch.
In the public uapi header, snd_timer_status is now guarded by
an #ifndef __KERNEL__ to avoid referencing 'struct timespec'.
The timespec definition will be removed from the kernel to prevent
new y2038 bugs and to avoid the conflict with an incompatible libc
type of the same name.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Since timespec is not year 2038 safe on 32bit system, and we need to
convert all timespec variables to timespec64 type for sound subsystem.
This patch is used to do preparation for following patches, that will
convert all structures defined in uapi/sound/asound.h to use 64-bit
time_t.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
When a new timer instance is created and assigned to the active link
in snd_timer_open(), the caller still doesn't (can't) set its callback
and callback data. In both the user-timer and the sequencer-timer
code, they do manually set up the callbacks after calling
snd_timer_open(). This has a potential risk of race when the timer
instance is added to the already running timer target, as the callback
might get triggered during setting up the callback itself.
This patch tries to address it by changing the API usage slightly:
- An empty timer instance is created at first via the new function
snd_timer_instance_new(). This object isn't linked to the timer
list yet.
- The caller sets up the callbacks and others stuff for the new timer
instance.
- The caller invokes snd_timer_open() with this instance, so that it's
linked to the target timer.
For closing, do similarly:
- Call snd_timer_close(). This unlinks the timer instance from the
timer list.
- Free the timer instance via snd_timer_instance_free() after that.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20191107192008.32331-4-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The code in both snd_timer_check_master() and snd_timer_check_slave()
are almost identical, both check whether the master/slave link and
does linkage. Factor out the common code and call it from both
functions for readability.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20191107192008.32331-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
