Use BUILD_BUG_ON_MSG(statically_true(ulo > uhi), ...) for the sanity check
of the bounds in clamp(). Gives better error coverage and one less
expansion of the arguments.
Link: https://lkml.kernel.org/r/34d53778977747f19cce2abb287bb3e6@AcuMS.aculab.com Signed-off-by: David Laight <david.laight@aculab.com> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Arnd Bergmann <arnd@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dan Carpenter <dan.carpenter@linaro.org> Cc: Jason A. Donenfeld <Jason@zx2c4.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Cc: Mateusz Guzik <mjguzik@gmail.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Pedro Falcato <pedro.falcato@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since the test for signed values being non-negative only relies on
__builtion_constant_p() (not is_constexpr()) it can use the 'ux' variable
instead of the caller supplied expression. This means that the #define
parameters are only expanded twice. Once in the code and once quoted in
the error message.
Link: https://lkml.kernel.org/r/051afc171806425da991908ed8688a98@AcuMS.aculab.com Signed-off-by: David Laight <david.laight@aculab.com> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Arnd Bergmann <arnd@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dan Carpenter <dan.carpenter@linaro.org> Cc: Jason A. Donenfeld <Jason@zx2c4.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Cc: Mateusz Guzik <mjguzik@gmail.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Pedro Falcato <pedro.falcato@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
- Change three to several.
- Remove the comment about retaining constant expressions, no longer true.
- Realign to nearer 80 columns and break on major punctiation.
- Add a leading comment to the block before __signed_type() and __is_nonneg()
Otherwise the block explaining the cast is a bit 'floating'.
Reword the rest of that comment to improve readability.
Link: https://lkml.kernel.org/r/85b050c81c1d4076aeb91a6cded45fee@AcuMS.aculab.com Signed-off-by: David Laight <david.laight@aculab.com> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Arnd Bergmann <arnd@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dan Carpenter <dan.carpenter@linaro.org> Cc: Jason A. Donenfeld <Jason@zx2c4.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Cc: Mateusz Guzik <mjguzik@gmail.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Pedro Falcato <pedro.falcato@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
David Laight pointed out that we should deal with the min3() and max3()
mess too, which still does excessive expansion.
And our current macros are actually rather broken.
In particular, the macros did this:
#define min3(x, y, z) min((typeof(x))min(x, y), z)
#define max3(x, y, z) max((typeof(x))max(x, y), z)
and that not only is a nested expansion of possibly very complex
arguments with all that involves, the typing with that "typeof()" cast
is completely wrong.
For example, imagine what happens in max3() if 'x' happens to be a
'unsigned char', but 'y' and 'z' are 'unsigned long'. The types are
compatible, and there's no warning - but the result is just random
garbage.
No, I don't think we've ever hit that issue in practice, but since we
now have sane infrastructure for doing this right, let's just use it.
It fixes any excessive expansion, and also avoids these kinds of broken
type issues.
This clarifies the rules for min()/max()/clamp() type checking and makes
them a much more efficient macro expansion.
In particular, we now look at the type and range of the inputs to see
whether they work together, generating a mask of acceptable comparisons,
and then just verifying that the inputs have a shared case:
- an expression with a signed type can be used for
(1) signed comparisons
(2) unsigned comparisons if it is statically known to have a
non-negative value
- an expression with an unsigned type can be used for
(3) unsigned comparison
(4) signed comparisons if the type is smaller than 'int' and thus
the C integer promotion rules will make it signed anyway
Here rule (1) and (3) are obvious, and rule (2) is important in order to
allow obvious trivial constants to be used together with unsigned
values.
Rule (4) is not necessarily a good idea, but matches what we used to do,
and we have extant cases of this situation in the kernel. Notably with
bcachefs having an expression like
where bch2_bucket_sectors_dirty() returns an 's64', and
'ca->mi.bucket_size' is of type 'u16'.
Technically that bcachefs comparison is clearly sensible on a C type
level, because the 'u16' will go through the normal C integer promotion,
and become 'int', and then we're comparing two signed values and
everything looks sane.
However, it's not entirely clear that a 'min(s64,u16)' operation makes a
lot of conceptual sense, and it's possible that we will remove rule (4).
After all, the _reason_ we have these complicated type checks is exactly
that the C type promotion rules are not very intuitive.
But at least for now the rule is in place for backwards compatibility.
Also note that rule (2) existed before, but is hugely relaxed by this
commit. It used to be true only for the simplest compile-time
non-negative integer constants. The new macro model will allow cases
where the compiler can trivially see that an expression is non-negative
even if it isn't necessarily a constant.
because our old 'min()' macro would see that 'pia[addr] & 0x3F' is of
type 'int' and clearly not a C constant expression, so doing a 'min()'
with a 'size_t' is a signedness violation.
Our new 'min()' macro still sees that 'pia[addr] & 0x3F' is of type
'int', but is smart enough to also see that it is clearly non-negative,
and thus would allow that case without any complaints.
Now that we no longer have any C constant expression contexts (ie array
size declarations or static initializers) that use min() or max(), we
can simpify the implementation by not having to worry about the result
staying as a C constant expression.
So now we can unconditionally just use temporary variables of the right
type, and get rid of the excessive expansion that used to come from the
use of
__builtin_choose_expr(__is_constexpr(...), ..
to pick the specialized code for constant expressions.
Another expansion simplification is to pass the temporary variables (in
addition to the original expression) to our __types_ok() macro. That
may superficially look like it complicates the macro, but when we only
want the type of the expression, expanding the temporary variable names
is much simpler and smaller than expanding the potentially complicated
original expression.
As a result, on my machine, doing a
$ time make drivers/staging/media/atomisp/pci/isp/kernels/ynr/ynr_1.0/ia_css_ynr.host.i
goes from
real 0m16.621s
user 0m15.360s
sys 0m1.221s
to
real 0m2.532s
user 0m2.091s
sys 0m0.452s
because the token expansion goes down dramatically.
In particular, the longest line expansion (which was line 71 of that
'ia_css_ynr.host.c' file) shrinks from 23,338kB (yes, 23MB for one
single line) to "just" 1,444kB (now "only" 1.4MB).
And yes, that line is still the line from hell, because it's doing
multiple levels of "min()/max()" expansion thanks to some of them being
hidden inside the uDIGIT_FITTING() macro.
Lorenzo has a nice cleanup patch that makes that driver use inline
functions instead of macros for sDIGIT_FITTING() and uDIGIT_FITTING(),
which will fix that line once and for all, but the 16-fold reduction in
this case does show why we need to simplify these helpers.
Cc: David Laight <David.Laight@aculab.com> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We only had a couple of array[] declarations, and changing them to just
use 'MAX()' instead of 'max()' fixes the issue.
This will allow us to simplify our min/max macros enormously, since they
can now unconditionally use temporary variables to avoid using the
argument values multiple times.
Cc: David Laight <David.Laight@aculab.com> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The state->timer is a cyclic timer that schedules work_i2c_poll and
delayed_work_enable_hotplug, while rearming itself. Using timer_delete()
fails to guarantee the timer isn't still running when destroyed, similarly
cancel_delayed_work() cannot ensure delayed_work_enable_hotplug has
terminated if already executing. During probe failure after timer
initialization, these may continue running as orphans and reference the
already-freed tc358743_state object through tc358743_irq_poll_timer.
Replace timer_delete() with timer_delete_sync() and cancel_delayed_work()
with cancel_delayed_work_sync() to ensure proper termination of timer and
work items before resource cleanup.
This bug was initially identified through static analysis. For reproduction
and testing, I created a functional emulation of the tc358743 device via a
kernel module and introduced faults through the debugfs interface.
Fixes: 869f38ae07f7 ("media: i2c: tc358743: Fix crash in the probe error path when using polling") Fixes: d32d98642de6 ("[media] Driver for Toshiba TC358743 HDMI to CSI-2 bridge") Cc: stable@vger.kernel.org Signed-off-by: Duoming Zhou <duoming@zju.edu.cn> Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
[ replaced del_timer() instead of timer_delete() ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The original code uses cancel_delayed_work() in xc5000_release(), which
does not guarantee that the delayed work item timer_sleep has fully
completed if it was already running. This leads to use-after-free scenarios
where xc5000_release() may free the xc5000_priv while timer_sleep is still
active and attempts to dereference the xc5000_priv.
A typical race condition is illustrated below:
CPU 0 (release thread) | CPU 1 (delayed work callback)
xc5000_release() | xc5000_do_timer_sleep()
cancel_delayed_work() |
hybrid_tuner_release_state(priv) |
kfree(priv) |
| priv = container_of() // UAF
Replace cancel_delayed_work() with cancel_delayed_work_sync() to ensure
that the timer_sleep is properly canceled before the xc5000_priv memory
is deallocated.
A deadlock concern was considered: xc5000_release() is called in a process
context and is not holding any locks that the timer_sleep work item might
also need. Therefore, the use of the _sync() variant is safe here.
This bug was initially identified through static analysis.
Make sure the firmware is released when we leave
xc_load_fw_and_init_tuner()
This change makes smatch happy:
drivers/media/tuners/xc5000.c:1213 xc_load_fw_and_init_tuner() warn: 'fw' from request_firmware() not released on lines: 1213.
Will Deacon [Fri, 3 Oct 2025 18:40:18 +0000 (19:40 +0100)]
KVM: arm64: Fix softirq masking in FPSIMD register saving sequence
Stable commit 8f4dc4e54eed ("KVM: arm64: Fix kernel BUG() due to bad
backport of FPSIMD/SVE/SME fix") fixed a kernel BUG() caused by a bad
backport of upstream commit fbc7e61195e2 ("KVM: arm64: Unconditionally
save+flush host FPSIMD/SVE/SME state") by ensuring that softirqs are
disabled/enabled across the fpsimd register save operation.
Unfortunately, although this fixes the original issue, it can now lead
to deadlock when re-enabling softirqs causes pending softirqs to be
handled with locks already held:
Take a tiny step towards the upstream fix in 9b19700e623f ("arm64:
fpsimd: Drop unneeded 'busy' flag") by additionally disabling hardirqs
while saving the fpsimd registers.
Cc: Ard Biesheuvel <ardb@kernel.org> Cc: Lee Jones <lee@kernel.org> Cc: Sasha Levin <sashal@kernel.org> Cc: <stable@vger.kernel.org> # 6.1.y Fixes: 8f4dc4e54eed ("KVM: arm64: Fix kernel BUG() due to bad backport of FPSIMD/SVE/SME fix") Reported-by: Kenneth Van Alstyne <kvanals@kvanals.org> Link: https://lore.kernel.org/r/010001999bae0958-4d80d25d-8dda-4006-a6b9-798f3e774f6c-000000@email.amazonses.com Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It is possible that the topology parsing function
audioreach_widget_load_module_common() could return NULL or an error
pointer. Add missing NULL check so that we do not dereference it.
Reported-by: Dan Carpenter <dan.carpenter@linaro.org> Cc: Stable@vger.kernel.org Fixes: 36ad9bf1d93d ("ASoC: qdsp6: audioreach: add topology support") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com> Link: https://patch.msgid.link/20250825101247.152619-2-srinivas.kandagatla@oss.qualcomm.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Syzbot reports a KASAN issue as below:
BUG: KASAN: use-after-free in __create_pipe include/linux/usb.h:1945 [inline]
BUG: KASAN: use-after-free in send_packet+0xa2d/0xbc0 drivers/media/rc/imon.c:627
Read of size 4 at addr ffff8880256fb000 by task syz-executor314/4465
The iMON driver improperly releases the usb_device reference in
imon_disconnect without coordinating with active users of the
device.
Specifically, the fields usbdev_intf0 and usbdev_intf1 are not
protected by the users counter (ictx->users). During probe,
imon_init_intf0 or imon_init_intf1 increments the usb_device
reference count depending on the interface. However, during
disconnect, usb_put_dev is called unconditionally, regardless of
actual usage.
As a result, if vfd_write or other operations are still in
progress after disconnect, this can lead to a use-after-free of
the usb_device pointer.
Thread 1 vfd_write Thread 2 imon_disconnect
...
if
usb_put_dev(ictx->usbdev_intf0)
else
usb_put_dev(ictx->usbdev_intf1)
...
while
send_packet
if
pipe = usb_sndintpipe(
ictx->usbdev_intf0) UAF
else
pipe = usb_sndctrlpipe(
ictx->usbdev_intf0, 0) UAF
Guard access to usbdev_intf0 and usbdev_intf1 after disconnect by
checking ictx->disconnected in all writer paths. Add early return
with -ENODEV in send_packet(), vfd_write(), lcd_write() and
display_open() if the device is no longer present.
Set and read ictx->disconnected under ictx->lock to ensure memory
synchronization. Acquire the lock in imon_disconnect() before setting
the flag to synchronize with any ongoing operations.
Ensure writers exit early and safely after disconnect before the USB
core proceeds with cleanup.
Found by Linux Verification Center (linuxtesting.org) with Syzkaller.
Reported-by: syzbot+f1a69784f6efe748c3bf@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=f1a69784f6efe748c3bf Fixes: 21677cfc562a ("V4L/DVB: ir-core: add imon driver") Cc: stable@vger.kernel.org Signed-off-by: Larshin Sergey <Sergey.Larshin@kaspersky.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The original code uses cancel_delayed_work() in flexcop_pci_remove(), which
does not guarantee that the delayed work item irq_check_work has fully
completed if it was already running. This leads to use-after-free scenarios
where flexcop_pci_remove() may free the flexcop_device while irq_check_work
is still active and attempts to dereference the device.
A typical race condition is illustrated below:
CPU 0 (remove) | CPU 1 (delayed work callback)
flexcop_pci_remove() | flexcop_pci_irq_check_work()
cancel_delayed_work() |
flexcop_device_kfree(fc_pci->fc_dev) |
| fc = fc_pci->fc_dev; // UAF
Replace cancel_delayed_work() with cancel_delayed_work_sync() to ensure
that the delayed work item is properly canceled and any executing delayed
work has finished before the device memory is deallocated.
This bug was initially identified through static analysis. To reproduce
and test it, I simulated the B2C2 FlexCop PCI device in QEMU and introduced
artificial delays within the flexcop_pci_irq_check_work() function to
increase the likelihood of triggering the bug.
A buffer overflow arises from the usage of snprintf to write into the
buffer "buf" in target_lu_gp_members_show function located in
/drivers/target/target_core_configfs.c. This buffer is allocated with
size LU_GROUP_NAME_BUF (256 bytes).
snprintf(...) formats multiple strings into buf with the HBA name
(hba->hba_group.cg_item), a slash character, a devicename (dev->
dev_group.cg_item) and a newline character, the total formatted string
length may exceed the buffer size of 256 bytes.
Since snprintf() returns the total number of bytes that would have been
written (the length of %s/%sn ), this value may exceed the buffer length
(256 bytes) passed to memcpy(), this will ultimately cause function
memcpy reporting a buffer overflow error.
An additional check of the return value of snprintf() can avoid this
buffer overflow.
Reported-by: Wang Haoran <haoranwangsec@gmail.com> Reported-by: ziiiro <yuanmingbuaa@gmail.com> Signed-off-by: Wang Haoran <haoranwangsec@gmail.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kenta Akagi [Thu, 2 Oct 2025 14:17:59 +0000 (23:17 +0900)]
selftests: mptcp: connect: fix build regression caused by backport
Since v6.1.154, mptcp selftests have failed to build with the following
errors:
mptcp_connect.c: In function ‘main_loop_s’:
mptcp_connect.c:1040:59: error: ‘winfo’ undeclared (first use in this function)
1040 | err = copyfd_io(fd, remotesock, 1, true, &winfo);
| ^~~~~
mptcp_connect.c:1040:59: note: each undeclared identifier is reported only once for each function it appears in
mptcp_connect.c:1040:23: error: too many arguments to function ‘copyfd_io’; expected 4, have 5
1040 | err = copyfd_io(fd, remotesock, 1, true, &winfo);
| ^~~~~~~~~ ~~~~~~
mptcp_connect.c:845:12: note: declared here
845 | static int copyfd_io(int infd, int peerfd, int outfd, bool close_peerfd)
| ^~~~~~~~~
This is caused by commit ff160500c499 ("selftests: mptcp: connect: catch
IO errors on listen side"), a backport of upstream 14e22b43df25,
which attempts to use the undeclared variable 'winfo' and passes too many
arguments to copyfd_io(). Both the winfo variable and the updated
copyfd_io() function were introduced in upstream
commit ca7ae8916043 ("selftests: mptcp: mptfo Initiator/Listener"),
which is not present in v6.1.y.
The goal of the backport is to stop on errors from copyfd_io.
Therefore, the backport does not depend on the changes in upstream
commit ca7ae8916043 ("selftests: mptcp: mptfo Initiator/Listener").
This commit simply removes ', &winfo' to fix a build failure.
UBSAN: signed-integer-overflow in ./include/crypto/sha256_base.h:64:19 34152083 * 64 cannot be represented in type 'int'
...
BUG: unable to handle page fault for address: ff9fffff83b624c0
sha256_update (lib/crypto/sha256.c:137)
crypto_sha256_update (crypto/sha256_generic.c:40)
kexec_calculate_store_digests (kernel/kexec_file.c:769)
__se_sys_kexec_file_load (kernel/kexec_file.c:397 kernel/kexec_file.c:332)
...
(Line numbers based on commit da274362a7bd9 ("Linux 6.12.49")
This started happening after commit f4da7afe07523f
("kexec_file: increase maximum file size to 4G") that landed in v6.0,
which increased the file size for kexec.
This is not happening upstream (v6.16+), given that `block` type was
upgraded from "int" to "size_t" in commit 74a43a2cf5e8 ("crypto:
lib/sha256 - Move partial block handling out")
Upgrade the block type similar to the commit above, avoiding hitting the
overflow.
This patch is only suitable for the stable tree, and before 6.16, which
got commit 74a43a2cf5e8 ("crypto: lib/sha256 - Move partial block
handling out"). This is not required before f4da7afe07523f ("kexec_file:
increase maximum file size to 4G"). In other words, this fix is required
between versions v6.0 and v6.16.
Signed-off-by: Breno Leitao <leitao@debian.org> Fixes: f4da7afe07523f ("kexec_file: increase maximum file size to 4G") # Before v6.16 Reported-by: Michael van der Westhuizen <rmikey@meta.com> Reported-by: Tobias Fleig <tfleig@meta.com> Reviewed-by: Eric Biggers <ebiggers@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The scale() functions detects invalid parameters, but continues
its calculations anyway. This causes bad results if negative values
are used for unsigned operations. Worst case, a division by 0 error
will be seen if source_min == source_max.
On top of that, after v6.13, the sequence of WARN_ON() followed by clamp()
may result in a build error with gcc 13.x.
drivers/gpu/drm/i915/display/intel_backlight.c: In function 'scale':
include/linux/compiler_types.h:542:45: error:
call to '__compiletime_assert_415' declared with attribute error:
clamp() low limit source_min greater than high limit source_max
This happens if the compiler decides to rearrange the code as follows.
if (source_min > source_max) {
WARN(..);
/* Do the clamp() knowing that source_min > source_max */
source_val = clamp(source_val, source_min, source_max);
} else {
/* Do the clamp knowing that source_min <= source_max */
source_val = clamp(source_val, source_min, source_max);
}
Fix the problem by evaluating the return values from WARN_ON and returning
immediately after a warning. While at it, fix divide by zero error seen
if source_min == source_max.
This simplifies the min_t() and max_t() macros by no longer making them
work in the context of a C constant expression.
That means that you can no longer use them for static initializers or
for array sizes in type definitions, but there were only a couple of
such uses, and all of them were converted (famous last words) to use
MIN_T/MAX_T instead.
Cc: David Laight <David.Laight@aculab.com> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 3a7e02c040b1 ("minmax: avoid overly complicated constant
expressions in VM code") added the simpler MIN_T/MAX_T macros in order
to avoid some excessive expansion from the rather complicated regular
min/max macros.
The complexity of those macros stems from two issues:
(a) trying to use them in situations that require a C constant
expression (in static initializers and for array sizes)
(b) the type sanity checking
and MIN_T/MAX_T avoids both of these issues.
Now, in the whole (long) discussion about all this, it was pointed out
that the whole type sanity checking is entirely unnecessary for
min_t/max_t which get a fixed type that the comparison is done in.
But that still leaves min_t/max_t unnecessarily complicated due to
worries about the C constant expression case.
However, it turns out that there really aren't very many cases that use
min_t/max_t for this, and we can just force-convert those.
This does exactly that.
Which in turn will then allow for much simpler implementations of
min_t()/max_t(). All the usual "macros in all upper case will evaluate
the arguments multiple times" rules apply.
We should do all the same things for the regular min/max() vs MIN/MAX()
cases, but that has the added complexity of various drivers defining
their own local versions of MIN/MAX, so that needs another level of
fixes first.
This just standardizes the use of MIN() and MAX() macros, with the very
traditional semantics. The goal is to use these for C constant
expressions and for top-level / static initializers, and so be able to
simplify the min()/max() macros.
These macro names were used by various kernel code - they are very
traditional, after all - and all such users have been fixed up, with a
few different approaches:
- trivial duplicated macro definitions have been removed
Note that 'trivial' here means that it's obviously kernel code that
already included all the major kernel headers, and thus gets the new
generic MIN/MAX macros automatically.
- non-trivial duplicated macro definitions are guarded with #ifndef
This is the "yes, they define their own versions, but no, the include
situation is not entirely obvious, and maybe they don't get the
generic version automatically" case.
- strange use case #1
A couple of drivers decided that the way they want to describe their
versioning is with
#define MAJ 1
#define MIN 2
#define DRV_VERSION __stringify(MAJ) "." __stringify(MIN)
which adds zero value and I just did my Alexander the Great
impersonation, and rewrote that pointless Gordian knot as
#define DRV_VERSION "1.2"
instead.
- strange use case #2
A couple of drivers thought that it's a good idea to have a random
'MIN' or 'MAX' define for a value or index into a table, rather than
the traditional macro that takes arguments.
These values were re-written as C enum's instead. The new
function-line macros only expand when followed by an open
parenthesis, and thus don't clash with enum use.
Happily, there weren't really all that many of these cases, and a lot of
users already had the pattern of using '#ifndef' guarding (or in one
case just using '#undef MIN') before defining their own private version
that does the same thing. I left such cases alone.
Cc: David Laight <David.Laight@aculab.com> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Eliav Farber <farbere@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This occurs when memset() is called on a buffer that is not 4-byte aligned
and extends to the end of a guard page, i.e. the next page is unmapped.
The bug is that the loop at the end of kmsan_internal_set_shadow_origin()
accesses the wrong shadow memory bytes when the address is not 4-byte
aligned. Since each 4 bytes are associated with an origin, it rounds the
address and size so that it can access all the origins that contain the
buffer. However, when it checks the corresponding shadow bytes for a
particular origin, it incorrectly uses the original unrounded shadow
address. This results in reads from shadow memory beyond the end of the
buffer's shadow memory, which crashes when that memory is not mapped.
To fix this, correctly align the shadow address before accessing the 4
shadow bytes corresponding to each origin.
Link: https://lkml.kernel.org/r/20250911195858.394235-1-ebiggers@kernel.org Fixes: 2ef3cec44c60 ("kmsan: do not wipe out origin when doing partial unpoisoning") Signed-off-by: Eric Biggers <ebiggers@kernel.org> Tested-by: Alexander Potapenko <glider@google.com> Reviewed-by: Alexander Potapenko <glider@google.com> Cc: Dmitriy Vyukov <dvyukov@google.com> Cc: Marco Elver <elver@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Adjust context in tests ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The `ring_len` parameter provided by the virtual function (VF)
is assigned directly to the hardware memory context (HMC) without
any validation.
To address this, introduce an upper boundary check for both Tx and Rx
queue lengths. The maximum number of descriptors supported by the
hardware is 8k-32.
Additionally, enforce alignment constraints: Tx rings must be a multiple
of 8, and Rx rings must be a multiple of 32.
In Tables 8-12 and 8-22 in the X710/XXV710/XL710 datasheet, the QLEN
description states that the maximum size of the descriptor queue is 8k
minus 32, or 8160.
Signed-off-by: Justin Bronder <jsbronder@cold-front.org> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel) Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Link: https://lore.kernel.org/r/20231113231047.548659-2-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 55d225670def ("i40e: add validation for ring_len param") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
VF state I40E_VF_STATE_ACTIVE is not the only state in which
VF is actually active so it should not be used to determine
if a VF is allowed to obtain resources.
Use I40E_VF_STATE_RESOURCES_LOADED that is set only in
i40e_vc_get_vf_resources_msg() and cleared during reset.
Fixes: 61125b8be85d ("i40e: Fix failed opcode appearing if handling messages from VF") Cc: stable@vger.kernel.org Signed-off-by: Lukasz Czapnik <lukasz.czapnik@intel.com> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Signed-off-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Simon Horman <horms@kernel.org> Tested-by: Rafal Romanowski <rafal.romanowski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
[ Adjust context ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The minmax infrastructure is overkill for simple constants, and can
cause huge expansions because those simple constants are then used by
other things.
For example, 'pageblock_order' is a core VM constant, but because it was
implemented using 'min_t()' and all the type-checking that involves, it
actually expanded to something like 2.5kB of preprocessor noise.
And when that simple constant was then used inside other expansions:
the end result was that one statement expanding to 253kB in size.
There are probably other cases of this, but this one case certainly
stood out.
I've added 'MIN_T()' and 'MAX_T()' macros for this kind of "core simple
constant with specific type" use. These macros skip the type checking,
and as such need to be very sparingly used only for obvious cases that
have active issues like this.
It appears that compiler_types.h already have an implementation of the
__unconst_integer_typeof() called __unqual_scalar_typeof(). Use it
instead of the copy.
flush_dcache_folio() isn't technically new, but no architecture
implemented it, so I've done that for them. The old APIs remain around
but are mostly implemented by calling the new interfaces.
The new APIs are based around setting up N page table entries at once.
The N entries belong to the same PMD, the same folio and the same VMA, so
ptep++ is a legitimate operation, and locking is taken care of for you.
Some architectures can do a better job of it than just a loop, but I have
hesitated to make too deep a change to architectures I don't understand
well.
One thing I have changed in every architecture is that PG_arch_1 is now a
per-folio bit instead of a per-page bit when used for dcache clean/dirty
tracking. This was something that would have to happen eventually, and it
makes sense to do it now rather than iterate over every page involved in a
cache flush and figure out if it needs to happen.
The point of all this is better performance, and Fengwei Yin has measured
improvement on x86. I suspect you'll see improvement on your architecture
too. Try the new will-it-scale test mentioned here:
https://lore.kernel.org/linux-mm/20230206140639.538867-5-fengwei.yin@intel.com/
You'll need to run it on an XFS filesystem and have
CONFIG_TRANSPARENT_HUGEPAGE set.
This patchset is the basis for much of the anonymous large folio work
being done by Ryan, so it's received quite a lot of testing over the last
few months.
This patch (of 38):
Determine if a value lies within a range more efficiently (subtraction +
comparison vs two comparisons and an AND). It also has useful (under some
circumstances) behaviour if the range exceeds the maximum value of the
type. Convert all the conflicting definitions of in_range() within the
kernel; some can use the generic definition while others need their own
definition.
If migration succeeded, we called
folio_migrate_flags()->mem_cgroup_migrate() to migrate the memcg from the
old to the new folio. This will set memcg_data of the old folio to 0.
Similarly, if migration failed, memcg_data of the dst folio is left unset.
If we call folio_putback_lru() on such folios (memcg_data == 0), we will
add the folio to be freed to the LRU, making memcg code unhappy. Running
the hmm selftests:
Likely, nothing else goes wrong: putting the last folio reference will
remove the folio from the LRU again. So besides memcg complaining, adding
the folio to be freed to the LRU is just an unnecessary step.
The new flow resembles what we have in migrate_folio_move(): add the dst
to the lru, remove migration ptes, unlock and unref dst.
Link: https://lkml.kernel.org/r/20250210161317.717936-1-david@redhat.com Fixes: 8763cb45ab96 ("mm/migrate: new memory migration helper for use with device memory") Signed-off-by: David Hildenbrand <david@redhat.com> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Alistair Popple <apopple@nvidia.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David Hildenbrand <david@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
s390/cpum_cf: Fix uninitialized warning after backport of ce971233242b
Upstream commit ce971233242b ("s390/cpum_cf: Deny all sampling events by
counter PMU"), backported to 6.6 as commit d660c8d8142e ("s390/cpum_cf:
Deny all sampling events by counter PMU"), implicitly depends on the
unconditional initialization of err to -ENOENT added by upstream
commit aa1ac98268cd ("s390/cpumf: Fix double free on error in
cpumf_pmu_event_init()"). The latter change is missing from 6.6,
resulting in an instance of -Wuninitialized, which is fairly obvious
from looking at the actual diff.
arch/s390/kernel/perf_cpum_cf.c:858:10: warning: variable 'err' is uninitialized when used here [-Wuninitialized]
858 | return err;
| ^~~
Commit aa1ac98268cd ("s390/cpumf: Fix double free on error in
cpumf_pmu_event_init()") depends on commit c70ca298036c ("perf/core:
Simplify the perf_event_alloc() error path"), which is a part of a much
larger series unsuitable for stable.
Extract the unconditional initialization of err to -ENOENT from
commit aa1ac98268cd ("s390/cpumf: Fix double free on error in
cpumf_pmu_event_init()") and apply it to 6.6 as a standalone change to
resolve the warning.
Fixes: d660c8d8142e ("s390/cpum_cf: Deny all sampling events by counter PMU") Signed-off-by: Nathan Chancellor <nathan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 1a194e6c8e1e ("fbcon: fix integer overflow in fbcon_do_set_font")
introduced an out-of-bounds access by storing data and allocation sizes
in the same variable. Restore the old size calculation and use the new
variable 'alloc_size' for the allocation.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Fixes: 1a194e6c8e1e ("fbcon: fix integer overflow in fbcon_do_set_font") Reported-by: Jani Nikula <jani.nikula@linux.intel.com> Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15020 Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/6201 Cc: Samasth Norway Ananda <samasth.norway.ananda@oracle.com> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: George Kennedy <george.kennedy@oracle.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Simona Vetter <simona@ffwll.ch> Cc: Helge Deller <deller@gmx.de> Cc: "Ville Syrjälä" <ville.syrjala@linux.intel.com> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: Qianqiang Liu <qianqiang.liu@163.com> Cc: Shixiong Ou <oushixiong@kylinos.cn> Cc: Kees Cook <kees@kernel.org> Cc: <stable@vger.kernel.org> # v5.9+ Cc: Zsolt Kajtar <soci@c64.rulez.org> Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com> Reviewed-by: Qianqiang Liu <qianqiang.liu@163.com> Link: https://lore.kernel.org/r/20250922134619.257684-1-tzimmermann@suse.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix integer overflow vulnerabilities in fbcon_do_set_font() where font
size calculations could overflow when handling user-controlled font
parameters.
The vulnerabilities occur when:
1. CALC_FONTSZ(h, pitch, charcount) performs h * pith * charcount
multiplication with user-controlled values that can overflow.
2. FONT_EXTRA_WORDS * sizeof(int) + size addition can also overflow
3. This results in smaller allocations than expected, leading to buffer
overflows during font data copying.
Add explicit overflow checking using check_mul_overflow() and
check_add_overflow() kernel helpers to safety validate all size
calculations before allocation.
Signed-off-by: Samasth Norway Ananda <samasth.norway.ananda@oracle.com> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Fixes: 39b3cffb8cf3 ("fbcon: prevent user font height or width change from causing potential out-of-bounds access") Cc: George Kennedy <george.kennedy@oracle.com> Cc: stable <stable@vger.kernel.org> Cc: syzbot+38a3699c7eaf165b97a6@syzkaller.appspotmail.com Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Simona Vetter <simona@ffwll.ch> Cc: Helge Deller <deller@gmx.de> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: "Ville Syrjälä" <ville.syrjala@linux.intel.com> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: Qianqiang Liu <qianqiang.liu@163.com> Cc: Shixiong Ou <oushixiong@kylinos.cn> Cc: Kees Cook <kees@kernel.org> Cc: <stable@vger.kernel.org> # v5.9+ Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://lore.kernel.org/r/20250912170023.3931881-1-samasth.norway.ananda@oracle.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Migration may be raced with fallocating hole. remove_inode_single_folio
will unmap the folio if the folio is still mapped. However, it's called
without folio lock. If the folio is migrated and the mapped pte has been
converted to migration entry, folio_mapped() returns false, and won't
unmap it. Due to extra refcount held by remove_inode_single_folio,
migration fails, restores migration entry to normal pte, and the folio is
mapped again. As a result, we triggered BUG in filemap_unaccount_folio.
The log is as follows:
BUG: Bad page cache in process hugetlb pfn:156c00
page: refcount:515 mapcount:0 mapping:0000000099fef6e1 index:0x0 pfn:0x156c00
head: order:9 mapcount:1 entire_mapcount:1 nr_pages_mapped:0 pincount:0
aops:hugetlbfs_aops ino:dcc dentry name(?):"my_hugepage_file"
flags: 0x17ffffc00000c1(locked|waiters|head|node=0|zone=2|lastcpupid=0x1fffff)
page_type: f4(hugetlb)
page dumped because: still mapped when deleted
CPU: 1 UID: 0 PID: 395 Comm: hugetlb Not tainted 6.17.0-rc5-00044-g7aac71907bde-dirty #484 NONE
Hardware name: QEMU Ubuntu 24.04 PC (i440FX + PIIX, 1996), BIOS 0.0.0 02/06/2015
Call Trace:
<TASK>
dump_stack_lvl+0x4f/0x70
filemap_unaccount_folio+0xc4/0x1c0
__filemap_remove_folio+0x38/0x1c0
filemap_remove_folio+0x41/0xd0
remove_inode_hugepages+0x142/0x250
hugetlbfs_fallocate+0x471/0x5a0
vfs_fallocate+0x149/0x380
Hold folio lock before checking if the folio is mapped to avold race with
migration.
Link: https://lkml.kernel.org/r/20250912074139.3575005-1-tujinjiang@huawei.com Fixes: 4aae8d1c051e ("mm/hugetlbfs: unmap pages if page fault raced with hole punch") Signed-off-by: Jinjiang Tu <tujinjiang@huawei.com> Cc: David Hildenbrand <david@redhat.com> Cc: Kefeng Wang <wangkefeng.wang@huawei.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Muchun Song <muchun.song@linux.dev> Cc: Oscar Salvador <osalvador@suse.de> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
afs_put_server() accessed server->debug_id before the NULL check, which
could lead to a null pointer dereference. Move the debug_id assignment,
ensuring we never dereference a NULL server pointer.
Fixes: 2757a4dc1849 ("afs: Fix access after dec in put functions") Cc: stable@vger.kernel.org Signed-off-by: Zhen Ni <zhen.ni@easystack.cn> Acked-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeffrey Altman <jaltman@auristor.com> Signed-off-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since dynamic_events interface on tracefs is compatible with
kprobe_events and uprobe_events, it should also check the lockdown
status and reject if it is set.
Link: https://lore.kernel.org/all/175824455687.45175.3734166065458520748.stgit@devnote2/ Fixes: 17911ff38aa5 ("tracing: Add locked_down checks to the open calls of files created for tracefs") Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 1b34cbbf4f01 ("crypto: af_alg - Disallow concurrent writes in
af_alg_sendmsg") changed some fields from bool to 1-bit bitfields of
type u32.
However, some assignments to these fields, specifically 'more' and
'merge', assign values greater than 1. These relied on C's implicit
conversion to bool, such that zero becomes false and nonzero becomes
true.
With a 1-bit bitfields of type u32 instead, mod 2 of the value is taken
instead, resulting in 0 being assigned in some cases when 1 was intended.
The ITR index (itr_idx) is only 2 bits wide. When constructing the
register value for QINT_RQCTL, all fields are ORed together. Without
masking, higher bits from itr_idx may overwrite adjacent fields in the
register.
Apply I40E_QINT_RQCTL_ITR_INDX_MASK to ensure only the intended bits are
set.
futex_requeue_pi_wakeup_sync() // Q_REQUEUE_PI_LOCKED
futex_hash_put()
// back to userland, on stack futex_q is garbage
/* back */
wake_up_state(q->task, TASK_NORMAL);
In this scenario futex_wait_requeue_pi() is able to leave without using
futex_q::lock_ptr for synchronization.
This can be prevented by reading futex_q::task before updating the
futex_q::requeue_state. A reference on the task_struct is not needed
because requeue_pi_wake_futex() is invoked with a spinlock_t held which
implies a RCU read section.
Even if T1 terminates immediately after, the task_struct will remain valid
during T2's wake_up_state(). A READ_ONCE on futex_q::task before
futex_requeue_pi_complete() is enough because it ensures that the variable
is read before the state is updated.
Read futex_q::task before updating the requeue state, use it for the
following wakeup.
Fixes: 07d91ef510fb1 ("futex: Prevent requeue_pi() lock nesting issue on RT") Reported-by: syzbot+034246a838a10d181e78@syzkaller.appspotmail.com Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Closes: https://lore.kernel.org/all/68b75989.050a0220.3db4df.01dd.GAE@google.com/ Signed-off-by: Sasha Levin <sashal@kernel.org>
pci_set_drvdata sets the value of pdev->driver_data to NULL,
after which the driver_data obtained from the same dev is
dereferenced in oaktrail_hdmi_i2c_exit, and the i2c_dev is
extracted from it. To prevent this, swap these calls.
Found by Linux Verification Center (linuxtesting.org) with Svacer.
This code calls kfree_rcu(new_node, rcu) and then dereferences "new_node"
and then dereferences it on the next line. Two lines later, we take
a mutex so I don't think this is an RCU safe region. Re-order it to do
the dereferences before queuing up the free.
Fixes: 68fbff68dbea ("octeontx2-pf: Add police action for TC flower") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev> Link: https://patch.msgid.link/aNKCL1jKwK8GRJHh@stanley.mountain Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The blamed commit and others in that patch set started the trend
of reusing existing DSA driver API for a new purpose: calling
ds->ops->port_fdb_add() on the CPU port.
The lantiq_gswip driver was not prepared to handle that, as can be seen
from the many errors that Daniel presents in the logs:
[ 174.050000] gswip 1e108000.switch: port 2 failed to add fa:aa:72:f4:8b:1e vid 1 to fdb: -22
[ 174.060000] gswip 1e108000.switch lan2: entered promiscuous mode
[ 174.070000] gswip 1e108000.switch: port 2 failed to add 00:01:02:03:04:02 vid 0 to fdb: -22
[ 174.090000] gswip 1e108000.switch: port 2 failed to add 00:01:02:03:04:02 vid 1 to fdb: -22
[ 174.090000] gswip 1e108000.switch: port 2 failed to delete fa:aa:72:f4:8b:1e vid 1 from fdb: -2
The errors are because gswip_port_fdb() wants to get a handle to the
bridge that originated these FDB events, to associate it with a FID.
Absolutely honourable purpose, however this only works for user ports.
To get the bridge that generated an FDB entry for the CPU port, one
would need to look at the db.bridge.dev argument. But this was
introduced in commit c26933639b54 ("net: dsa: request drivers to perform
FDB isolation"), first appeared in v5.18, and when the blamed commit was
introduced in v5.14, no such API existed.
So the core DSA feature was introduced way too soon for lantiq_gswip.
Not acting on these host FDB entries and suppressing any errors has no
other negative effect, and practically returns us to not supporting the
host filtering feature at all - peacefully, this time.
Fixes: 10fae4ac89ce ("net: dsa: include bridge addresses which are local in the host fdb list") Reported-by: Daniel Golle <daniel@makrotopia.org> Closes: https://lore.kernel.org/netdev/aJfNMLNoi1VOsPrN@pidgin.makrotopia.org/ Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20250918072142.894692-3-vladimir.oltean@nxp.com Tested-by: Daniel Golle <daniel@makrotopia.org> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
A port added to a "single port bridge" operates as standalone, and this
is mutually exclusive to being part of a Linux bridge. In fact,
gswip_port_bridge_join() calls gswip_add_single_port_br() with
add=false, i.e. removes the port from the "single port bridge" to enable
autonomous forwarding.
The blamed commit seems to have incorrectly thought that ds->ops->port_enable()
is called one time per port, during the setup phase of the switch.
However, it is actually called during the ndo_open() implementation of
DSA user ports, which is to say that this sequence of events:
1. ip link set swp0 down
2. ip link add br0 type bridge
3. ip link set swp0 master br0
4. ip link set swp0 up
would cause swp0 to join back the "single port bridge" which step 3 had
just removed it from.
The correct DSA hook for one-time actions per port at switch init time
is ds->ops->port_setup(). This is what seems to match the coder's
intention; also see the comment at the beginning of the file:
* At the initialization the driver allocates one bridge table entry for
~~~~~~~~~~~~~~~~~~~~~
* each switch port which is used when the port is used without an
* explicit bridge.
Fixes: 8206e0ce96b3 ("net: dsa: lantiq: Add VLAN unaware bridge offloading") Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20250918072142.894692-2-vladimir.oltean@nxp.com Tested-by: Daniel Golle <daniel@makrotopia.org> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Before commit 74be4babe72f ("net: dsa: do not enable or disable non user
ports"), gswip_port_enable/disable() were also executed for the cpu port
in gswip_setup() which disabled the cpu port during initialization.
Let's restore this by removing the dsa_is_user_port checks. Also, let's
clean up the gswip_port_enable() function so that we only have to check
for the cpu port once. The operation reordering done here is safe.
Signed-off-by: Martin Schiller <ms@dev.tdt.de> Acked-by: Hauke Mehrtens <hauke@hauke-m.de> Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Link: https://lore.kernel.org/r/20240611135434.3180973-7-ms@dev.tdt.de Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: c0054b25e2f1 ("net: dsa: lantiq_gswip: move gswip_add_single_port_br() call to port_setup()") Signed-off-by: Sasha Levin <sashal@kernel.org>
IPv6 fdb groups functional
--------------------------
[...]
COMMAND: ip -netns me-nRsN3E nexthop add id 63 via 2001:db8:91::4
Error: Device attribute required for non-blackhole and non-fdb nexthops.
COMMAND: ip -netns me-nRsN3E nexthop add id 64 via 2001:db8:91::5
Error: Device attribute required for non-blackhole and non-fdb nexthops.
COMMAND: ip -netns me-nRsN3E nexthop add id 103 group 63/64 fdb
Error: Invalid nexthop id.
TEST: Fdb Nexthop group with non-fdb nexthops [ OK ]
[...]
IPv4 fdb groups functional
--------------------------
[...]
COMMAND: ip -netns me-nRsN3E nexthop add id 14 via 172.16.1.2
Error: Device attribute required for non-blackhole and non-fdb nexthops.
COMMAND: ip -netns me-nRsN3E nexthop add id 15 via 172.16.1.3
Error: Device attribute required for non-blackhole and non-fdb nexthops.
COMMAND: ip -netns me-nRsN3E nexthop add id 103 group 14/15 fdb
Error: Invalid nexthop id.
TEST: Fdb Nexthop group with non-fdb nexthops [ OK ]
COMMAND: ip -netns me-nRsN3E nexthop add id 16 via 172.16.1.2 fdb
COMMAND: ip -netns me-nRsN3E nexthop add id 17 via 172.16.1.3 fdb
COMMAND: ip -netns me-nRsN3E nexthop add id 104 group 14/15
Error: Invalid nexthop id.
TEST: Non-Fdb Nexthop group with fdb nexthops [ OK ]
[...]
COMMAND: ip -netns me-0dlhyd ro add 172.16.0.0/22 nhid 15
Error: Nexthop id does not exist.
TEST: Route add with fdb nexthop [ OK ]
In addition, as can be seen in the above output, a couple of IPv4 test
cases used the non-FDB nexthops (14 and 15) when they intended to use
the FDB nexthops (16 and 17). These test cases only passed because
failure was expected, but they failed for the wrong reason.
Fix the test to create the non-FDB nexthops with a nexthop device and
adjust the IPv4 test cases to use the FDB nexthops instead of the
non-FDB nexthops.
IPv6 fdb groups functional
--------------------------
[...]
COMMAND: ip -netns me-lNzfHP nexthop add id 63 via 2001:db8:91::4 dev veth1
COMMAND: ip -netns me-lNzfHP nexthop add id 64 via 2001:db8:91::5 dev veth1
COMMAND: ip -netns me-lNzfHP nexthop add id 103 group 63/64 fdb
Error: FDB nexthop group can only have fdb nexthops.
TEST: Fdb Nexthop group with non-fdb nexthops [ OK ]
[...]
IPv4 fdb groups functional
--------------------------
[...]
COMMAND: ip -netns me-lNzfHP nexthop add id 14 via 172.16.1.2 dev veth1
COMMAND: ip -netns me-lNzfHP nexthop add id 15 via 172.16.1.3 dev veth1
COMMAND: ip -netns me-lNzfHP nexthop add id 103 group 14/15 fdb
Error: FDB nexthop group can only have fdb nexthops.
TEST: Fdb Nexthop group with non-fdb nexthops [ OK ]
COMMAND: ip -netns me-lNzfHP nexthop add id 16 via 172.16.1.2 fdb
COMMAND: ip -netns me-lNzfHP nexthop add id 17 via 172.16.1.3 fdb
COMMAND: ip -netns me-lNzfHP nexthop add id 104 group 16/17
Error: Non FDB nexthop group cannot have fdb nexthops.
TEST: Non-Fdb Nexthop group with fdb nexthops [ OK ]
[...]
COMMAND: ip -netns me-lNzfHP ro add 172.16.0.0/22 nhid 16
Error: Route cannot point to a fdb nexthop.
TEST: Route add with fdb nexthop [ OK ]
[...]
Tests passed: 30
Tests failed: 0
Tests skipped: 0
The kernel forbids the creation of non-FDB nexthop groups with FDB
nexthops:
# ip nexthop add id 1 via 192.0.2.1 fdb
# ip nexthop add id 2 group 1
Error: Non FDB nexthop group cannot have fdb nexthops.
And vice versa:
# ip nexthop add id 3 via 192.0.2.2 dev dummy1
# ip nexthop add id 4 group 3 fdb
Error: FDB nexthop group can only have fdb nexthops.
However, as long as no routes are pointing to a non-FDB nexthop group,
the kernel allows changing the type of a nexthop from FDB to non-FDB and
vice versa:
# ip nexthop add id 5 via 192.0.2.2 dev dummy1
# ip nexthop add id 6 group 5
# ip nexthop replace id 5 via 192.0.2.2 fdb
# echo $?
0
This configuration is invalid and can result in a NPD [1] since FDB
nexthops are not associated with a nexthop device:
# ip route add 198.51.100.1/32 nhid 6
# ping 198.51.100.1
Fix by preventing nexthop FDB status change while the nexthop is in a
group:
# ip nexthop add id 7 via 192.0.2.2 dev dummy1
# ip nexthop add id 8 group 7
# ip nexthop replace id 7 via 192.0.2.2 fdb
Error: Cannot change nexthop FDB status while in a group.
In bnxt_tc_parse_pedit(), the code incorrectly writes IPv6
destination values to the source address field (saddr) when
processing pedit offsets within the destination address range.
This patch corrects the assignment to use daddr instead of saddr,
ensuring that pedit operations on IPv6 destination addresses are
applied correctly.
Fixes: 9b9eb518e338 ("bnxt_en: Add support for NAT(L3/L4 rewrite)") Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Link: https://patch.msgid.link/20250920121157.351921-1-alok.a.tiwari@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This fixes the following UFA in hci_acl_create_conn_sync where a
connection still pending is command submission (conn->state == BT_OPEN)
maybe freed, also since this also can happen with the likes of
hci_le_create_conn_sync fix it as well:
BUG: KASAN: slab-use-after-free in hci_acl_create_conn_sync+0x5ef/0x790 net/bluetooth/hci_sync.c:6861
Write of size 2 at addr ffff88805ffcc038 by task kworker/u11:2/9541
hci_resume_advertising_sync is suppose to resume all instance paused by
hci_pause_advertising_sync, this logic is used for procedures are only
allowed when not advertising, but instance 0x00 was not being
re-enabled.
Fixes: ad383c2c65a5 ("Bluetooth: hci_sync: Enable advertising when LL privacy is enabled") Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Explicitly uses a 64-bit constant when the number of bits used for its
shifting is 32 (which is the case for PC CAN FD interfaces supported by
this driver).
Sending an PF_PACKET allows to bypass the CAN framework logic and to
directly reach the xmit() function of a CAN driver. The only check
which is performed by the PF_PACKET framework is to make sure that
skb->len fits the interface's MTU.
Unfortunately, because the mcba_usb driver does not populate its
net_device_ops->ndo_change_mtu(), it is possible for an attacker to
configure an invalid MTU by doing, for example:
$ ip link set can0 mtu 9999
After doing so, the attacker could open a PF_PACKET socket using the
ETH_P_CANXL protocol:
The CAN drivers' xmit() function are calling can_dev_dropped_skb() to
check that the skb is valid, unfortunately under above conditions, the
malicious packet is able to go through can_dev_dropped_skb() checks:
1. the skb->protocol is set to ETH_P_CANXL which is valid (the
function does not check the actual device capabilities).
2. the length is a valid CAN XL length.
And so, mcba_usb_start_xmit() receives a CAN XL frame which it is not
able to correctly handle and will thus misinterpret it as a CAN frame.
This can result in a buffer overflow. The driver will consume cf->len
as-is with no further checks on these lines:
usb_msg.dlc = cf->len;
memcpy(usb_msg.data, cf->data, usb_msg.dlc);
Here, cf->len corresponds to the flags field of the CAN XL frame. In
our previous example, we set canxl_frame->flags to 0xff. Because the
maximum expected length is 8, a buffer overflow of 247 bytes occurs!
Populate net_device_ops->ndo_change_mtu() to ensure that the
interface's MTU can not be set to anything bigger than CAN_MTU. By
fixing the root cause, this prevents the buffer overflow.
Fixes: 51f3baad7de9 ("can: mcba_usb: Add support for Microchip CAN BUS Analyzer") Signed-off-by: Vincent Mailhol <mailhol@kernel.org> Link: https://patch.msgid.link/20250918-can-fix-mtu-v1-4-0d1cada9393b@kernel.org Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Sending an PF_PACKET allows to bypass the CAN framework logic and to
directly reach the xmit() function of a CAN driver. The only check
which is performed by the PF_PACKET framework is to make sure that
skb->len fits the interface's MTU.
Unfortunately, because the sun4i_can driver does not populate its
net_device_ops->ndo_change_mtu(), it is possible for an attacker to
configure an invalid MTU by doing, for example:
$ ip link set can0 mtu 9999
After doing so, the attacker could open a PF_PACKET socket using the
ETH_P_CANXL protocol:
The CAN drivers' xmit() function are calling can_dev_dropped_skb() to
check that the skb is valid, unfortunately under above conditions, the
malicious packet is able to go through can_dev_dropped_skb() checks:
1. the skb->protocol is set to ETH_P_CANXL which is valid (the
function does not check the actual device capabilities).
2. the length is a valid CAN XL length.
And so, sun4ican_start_xmit() receives a CAN XL frame which it is not
able to correctly handle and will thus misinterpret it as a CAN frame.
This can result in a buffer overflow. The driver will consume cf->len
as-is with no further checks on this line:
dlc = cf->len;
Here, cf->len corresponds to the flags field of the CAN XL frame. In
our previous example, we set canxl_frame->flags to 0xff. Because the
maximum expected length is 8, a buffer overflow of 247 bytes occurs a
couple line below when doing:
for (i = 0; i < dlc; i++)
writel(cf->data[i], priv->base + (dreg + i * 4));
Populate net_device_ops->ndo_change_mtu() to ensure that the
interface's MTU can not be set to anything bigger than CAN_MTU. By
fixing the root cause, this prevents the buffer overflow.
Fixes: 0738eff14d81 ("can: Allwinner A10/A20 CAN Controller support - Kernel module") Signed-off-by: Vincent Mailhol <mailhol@kernel.org> Link: https://patch.msgid.link/20250918-can-fix-mtu-v1-3-0d1cada9393b@kernel.org Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Sending an PF_PACKET allows to bypass the CAN framework logic and to
directly reach the xmit() function of a CAN driver. The only check
which is performed by the PF_PACKET framework is to make sure that
skb->len fits the interface's MTU.
Unfortunately, because the sun4i_can driver does not populate its
net_device_ops->ndo_change_mtu(), it is possible for an attacker to
configure an invalid MTU by doing, for example:
$ ip link set can0 mtu 9999
After doing so, the attacker could open a PF_PACKET socket using the
ETH_P_CANXL protocol:
The CAN drivers' xmit() function are calling can_dev_dropped_skb() to
check that the skb is valid, unfortunately under above conditions, the
malicious packet is able to go through can_dev_dropped_skb() checks:
1. the skb->protocol is set to ETH_P_CANXL which is valid (the
function does not check the actual device capabilities).
2. the length is a valid CAN XL length.
And so, hi3110_hard_start_xmit() receives a CAN XL frame which it is
not able to correctly handle and will thus misinterpret it as a CAN
frame. The driver will consume frame->len as-is with no further
checks.
This can result in a buffer overflow later on in hi3110_hw_tx() on
this line:
Here, frame->len corresponds to the flags field of the CAN XL frame.
In our previous example, we set canxl_frame->flags to 0xff. Because
the maximum expected length is 8, a buffer overflow of 247 bytes
occurs!
Populate net_device_ops->ndo_change_mtu() to ensure that the
interface's MTU can not be set to anything bigger than CAN_MTU. By
fixing the root cause, this prevents the buffer overflow.
Fixes: 57e83fb9b746 ("can: hi311x: Add Holt HI-311x CAN driver") Signed-off-by: Vincent Mailhol <mailhol@kernel.org> Link: https://patch.msgid.link/20250918-can-fix-mtu-v1-2-0d1cada9393b@kernel.org Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Sending an PF_PACKET allows to bypass the CAN framework logic and to
directly reach the xmit() function of a CAN driver. The only check
which is performed by the PF_PACKET framework is to make sure that
skb->len fits the interface's MTU.
Unfortunately, because the etas_es58x driver does not populate its
net_device_ops->ndo_change_mtu(), it is possible for an attacker to
configure an invalid MTU by doing, for example:
$ ip link set can0 mtu 9999
After doing so, the attacker could open a PF_PACKET socket using the
ETH_P_CANXL protocol:
The CAN drivers' xmit() function are calling can_dev_dropped_skb() to
check that the skb is valid, unfortunately under above conditions, the
malicious packet is able to go through can_dev_dropped_skb() checks:
1. the skb->protocol is set to ETH_P_CANXL which is valid (the
function does not check the actual device capabilities).
2. the length is a valid CAN XL length.
And so, es58x_start_xmit() receives a CAN XL frame which it is not
able to correctly handle and will thus misinterpret it as a CAN(FD)
frame.
This can result in a buffer overflow. For example, using the es581.4
variant, the frame will be dispatched to es581_4_tx_can_msg(), go
through the last check at the beginning of this function:
if (can_is_canfd_skb(skb))
return -EMSGSIZE;
and reach this line:
memcpy(tx_can_msg->data, cf->data, cf->len);
Here, cf->len corresponds to the flags field of the CAN XL frame. In
our previous example, we set canxl_frame->flags to 0xff. Because the
maximum expected length is 8, a buffer overflow of 247 bytes occurs!
Populate net_device_ops->ndo_change_mtu() to ensure that the
interface's MTU can not be set to anything bigger than CAN_MTU or
CANFD_MTU (depending on the device capabilities). By fixing the root
cause, this prevents the buffer overflow.
Fixes: 8537257874e9 ("can: etas_es58x: add core support for ETAS ES58X CAN USB interfaces") Signed-off-by: Vincent Mailhol <mailhol@kernel.org> Link: https://patch.msgid.link/20250918-can-fix-mtu-v1-1-0d1cada9393b@kernel.org Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
If we are using a hardcoded delay of 0 there's no point in
using delayed_work it only adds confusion.
The client also uses a normal work_struct and now
it is easier to move it to the common smbdirect_socket.
Cc: Namjae Jeon <linkinjeon@kernel.org> Cc: Steve French <smfrench@gmail.com> Cc: Tom Talpey <tom@talpey.com> Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Fixes: 0626e6641f6b ("cifsd: add server handler for central processing and tranport layers") Signed-off-by: Stefan Metzmacher <metze@samba.org> Acked-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 2a6c72738706 ("cpufreq: Initialize cpufreq-based
frequency-invariance later") postponed the frequency invariance
initialization to avoid disabling it in the error case.
This isn't locking safe, instead move the initialization up before
the subsys interface is registered (which will rebuild the
sched_domains) and add the corresponding disable on the error path.
Observed lockdep without this patch:
[ 0.989686] ======================================================
[ 0.989688] WARNING: possible circular locking dependency detected
[ 0.989690] 6.17.0-rc4-cix-build+ #31 Tainted: G S
[ 0.989691] ------------------------------------------------------
[ 0.989692] swapper/0/1 is trying to acquire lock:
[ 0.989693] ffff800082ada7f8 (sched_energy_mutex){+.+.}-{4:4}, at: rebuild_sched_domains_energy+0x30/0x58
[ 0.989705]
but task is already holding lock:
[ 0.989706] ffff000088c89bc8 (&policy->rwsem){+.+.}-{4:4}, at: cpufreq_online+0x7f8/0xbe0
[ 0.989713]
which lock already depends on the new lock.
Fixes: 2a6c72738706 ("cpufreq: Initialize cpufreq-based frequency-invariance later") Signed-off-by: Christian Loehle <christian.loehle@arm.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The TMU has two temperature measurement sites located on the chip. The
probe 0 is located inside of the ANAMIX, while the probe 1 is located near
the ARM core. This has been confirmed by checking with HW design team and
checking RTL code.
So correct the {cpu,soc}-thermal sensor index.
Fixes: 30cdd62dce6b ("arm64: dts: imx8mp: Add thermal zones support") Signed-off-by: Peng Fan <peng.fan@nxp.com> Reviewed-by: Frank Li <Frank.Li@nxp.com> Signed-off-by: Shawn Guo <shawnguo@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
mm/swap.c and mm/mlock.c agree to drain any per-CPU batch as soon as a
large folio is added: so collect_longterm_unpinnable_folios() just wastes
effort when calling lru_add_drain[_all]() on a large folio.
But although there is good reason not to batch up PMD-sized folios, we
might well benefit from batching a small number of low-order mTHPs (though
unclear how that "small number" limitation will be implemented).
So ask if folio_may_be_lru_cached() rather than !folio_test_large(), to
insulate those particular checks from future change. Name preferred to
"folio_is_batchable" because large folios can well be put on a batch: it's
just the per-CPU LRU caches, drained much later, which need care.
Marked for stable, to counter the increase in lru_add_drain_all()s from
"mm/gup: check ref_count instead of lru before migration".
Link: https://lkml.kernel.org/r/57d2eaf8-3607-f318-e0c5-be02dce61ad0@google.com Fixes: 9a4e9f3b2d73 ("mm: update get_user_pages_longterm to migrate pages allocated from CMA region") Signed-off-by: Hugh Dickins <hughd@google.com> Suggested-by: David Hildenbrand <david@redhat.com> Acked-by: David Hildenbrand <david@redhat.com> Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: Chris Li <chrisl@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Keir Fraser <keirf@google.com> Cc: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Li Zhe <lizhe.67@bytedance.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Peter Xu <peterx@redhat.com> Cc: Rik van Riel <riel@surriel.com> Cc: Shivank Garg <shivankg@amd.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Wei Xu <weixugc@google.com> Cc: Will Deacon <will@kernel.org> Cc: yangge <yangge1116@126.com> Cc: Yuanchu Xie <yuanchu@google.com> Cc: Yu Zhao <yuzhao@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Resolved conflicts in mm/swap.c; left "page" parts of mm/mlock.c as is ] Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In many cases, if collect_longterm_unpinnable_folios() does need to drain
the LRU cache to release a reference, the cache in question is on this
same CPU, and much more efficiently drained by a preliminary local
lru_add_drain(), than the later cross-CPU lru_add_drain_all().
Marked for stable, to counter the increase in lru_add_drain_all()s from
"mm/gup: check ref_count instead of lru before migration". Note for clean
backports: can take 6.16 commit a03db236aebf ("gup: optimize longterm
pin_user_pages() for large folio") first.
Link: https://lkml.kernel.org/r/66f2751f-283e-816d-9530-765db7edc465@google.com Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: David Hildenbrand <david@redhat.com> Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: Chris Li <chrisl@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Keir Fraser <keirf@google.com> Cc: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Li Zhe <lizhe.67@bytedance.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Peter Xu <peterx@redhat.com> Cc: Rik van Riel <riel@surriel.com> Cc: Shivank Garg <shivankg@amd.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Wei Xu <weixugc@google.com> Cc: Will Deacon <will@kernel.org> Cc: yangge <yangge1116@126.com> Cc: Yuanchu Xie <yuanchu@google.com> Cc: Yu Zhao <yuzhao@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Resolved minor conflicts ] Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Patch series "mm: better GUP pin lru_add_drain_all()", v2.
Series of lru_add_drain_all()-related patches, arising from recent mm/gup
migration report from Will Deacon.
This patch (of 5):
Will Deacon reports:-
When taking a longterm GUP pin via pin_user_pages(),
__gup_longterm_locked() tries to migrate target folios that should not be
longterm pinned, for example because they reside in a CMA region or
movable zone. This is done by first pinning all of the target folios
anyway, collecting all of the longterm-unpinnable target folios into a
list, dropping the pins that were just taken and finally handing the list
off to migrate_pages() for the actual migration.
It is critically important that no unexpected references are held on the
folios being migrated, otherwise the migration will fail and
pin_user_pages() will return -ENOMEM to its caller. Unfortunately, it is
relatively easy to observe migration failures when running pKVM (which
uses pin_user_pages() on crosvm's virtual address space to resolve stage-2
page faults from the guest) on a 6.15-based Pixel 6 device and this
results in the VM terminating prematurely.
In the failure case, 'crosvm' has called mlock(MLOCK_ONFAULT) on its
mapping of guest memory prior to the pinning. Subsequently, when
pin_user_pages() walks the page-table, the relevant 'pte' is not present
and so the faulting logic allocates a new folio, mlocks it with
mlock_folio() and maps it in the page-table.
Since commit 2fbb0c10d1e8 ("mm/munlock: mlock_page() munlock_page() batch
by pagevec"), mlock/munlock operations on a folio (formerly page), are
deferred. For example, mlock_folio() takes an additional reference on the
target folio before placing it into a per-cpu 'folio_batch' for later
processing by mlock_folio_batch(), which drops the refcount once the
operation is complete. Processing of the batches is coupled with the LRU
batch logic and can be forcefully drained with lru_add_drain_all() but as
long as a folio remains unprocessed on the batch, its refcount will be
elevated.
This deferred batching therefore interacts poorly with the pKVM pinning
scenario as we can find ourselves in a situation where the migration code
fails to migrate a folio due to the elevated refcount from the pending
mlock operation.
Hugh Dickins adds:-
!folio_test_lru() has never been a very reliable way to tell if an
lru_add_drain_all() is worth calling, to remove LRU cache references to
make the folio migratable: the LRU flag may be set even while the folio is
held with an extra reference in a per-CPU LRU cache.
5.18 commit 2fbb0c10d1e8 may have made it more unreliable. Then 6.11
commit 33dfe9204f29 ("mm/gup: clear the LRU flag of a page before adding
to LRU batch") tried to make it reliable, by moving LRU flag clearing; but
missed the mlock/munlock batches, so still unreliable as reported.
And it turns out to be difficult to extend 33dfe9204f29's LRU flag
clearing to the mlock/munlock batches: if they do benefit from batching,
mlock/munlock cannot be so effective when easily suppressed while !LRU.
Instead, switch to an expected ref_count check, which was more reliable
all along: some more false positives (unhelpful drains) than before, and
never a guarantee that the folio will prove migratable, but better.
Note on PG_private_2: ceph and nfs are still using the deprecated
PG_private_2 flag, with the aid of netfs and filemap support functions.
Although it is consistently matched by an increment of folio ref_count,
folio_expected_ref_count() intentionally does not recognize it, and ceph
folio migration currently depends on that for PG_private_2 folios to be
rejected. New references to the deprecated flag are discouraged, so do
not add it into the collect_longterm_unpinnable_folios() calculation: but
longterm pinning of transiently PG_private_2 ceph and nfs folios (an
uncommon case) may invoke a redundant lru_add_drain_all(). And this makes
easy the backport to earlier releases: up to and including 6.12, btrfs
also used PG_private_2, but without a ref_count increment.
Note for stable backports: requires 6.16 commit 86ebd50224c0 ("mm:
add folio_expected_ref_count() for reference count calculation").
Link: https://lkml.kernel.org/r/41395944-b0e3-c3ac-d648-8ddd70451d28@google.com Link: https://lkml.kernel.org/r/bd1f314a-fca1-8f19-cac0-b936c9614557@google.com Fixes: 9a4e9f3b2d73 ("mm: update get_user_pages_longterm to migrate pages allocated from CMA region") Signed-off-by: Hugh Dickins <hughd@google.com> Reported-by: Will Deacon <will@kernel.org> Closes: https://lore.kernel.org/linux-mm/20250815101858.24352-1-will@kernel.org/ Acked-by: Kiryl Shutsemau <kas@kernel.org> Acked-by: David Hildenbrand <david@redhat.com> Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: Chris Li <chrisl@kernel.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Keir Fraser <keirf@google.com> Cc: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Li Zhe <lizhe.67@bytedance.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Peter Xu <peterx@redhat.com> Cc: Rik van Riel <riel@surriel.com> Cc: Shivank Garg <shivankg@amd.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Wei Xu <weixugc@google.com> Cc: yangge <yangge1116@126.com> Cc: Yuanchu Xie <yuanchu@google.com> Cc: Yu Zhao <yuzhao@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Clean cherry-pick now into this tree ] Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Patch series " JFS: Implement migrate_folio for jfs_metapage_aops" v5.
This patchset addresses a warning that occurs during memory compaction due
to JFS's missing migrate_folio operation. The warning was introduced by
commit 7ee3647243e5 ("migrate: Remove call to ->writepage") which added
explicit warnings when filesystem don't implement migrate_folio.
The syzbot reported following [1]:
jfs_metapage_aops does not implement migrate_folio
WARNING: CPU: 1 PID: 5861 at mm/migrate.c:955 fallback_migrate_folio mm/migrate.c:953 [inline]
WARNING: CPU: 1 PID: 5861 at mm/migrate.c:955 move_to_new_folio+0x70e/0x840 mm/migrate.c:1007
Modules linked in:
CPU: 1 UID: 0 PID: 5861 Comm: syz-executor280 Not tainted 6.15.0-rc1-next-20250411-syzkaller #0 PREEMPT(full)
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/12/2025
RIP: 0010:fallback_migrate_folio mm/migrate.c:953 [inline]
RIP: 0010:move_to_new_folio+0x70e/0x840 mm/migrate.c:1007
To fix this issue, this series implement metapage_migrate_folio() for JFS
which handles both single and multiple metapages per page configurations.
While most filesystems leverage existing migration implementations like
filemap_migrate_folio(), buffer_migrate_folio_norefs() or
buffer_migrate_folio() (which internally used folio_expected_refs()),
JFS's metapage architecture requires special handling of its private data
during migration. To support this, this series introduce the
folio_expected_ref_count(), which calculates external references to a
folio from page/swap cache, private data, and page table mappings.
This standardized implementation replaces the previous ad-hoc
folio_expected_refs() function and enables JFS to accurately determine
whether a folio has unexpected references before attempting migration.
Implement folio_expected_ref_count() to calculate expected folio reference
counts from:
- Page/swap cache (1 per page)
- Private data (1)
- Page table mappings (1 per map)
While originally needed for page migration operations, this improved
implementation standardizes reference counting by consolidating all
refcount contributors into a single, reusable function that can benefit
any subsystem needing to detect unexpected references to folios.
The folio_expected_ref_count() returns the sum of these external
references without including any reference the caller itself might hold.
Callers comparing against the actual folio_ref_count() must account for
their own references separately.
Link: https://syzkaller.appspot.com/bug?extid=8bb6fd945af4e0ad9299 Link: https://lkml.kernel.org/r/20250430100150.279751-1-shivankg@amd.com Link: https://lkml.kernel.org/r/20250430100150.279751-2-shivankg@amd.com Signed-off-by: David Hildenbrand <david@redhat.com> Signed-off-by: Shivank Garg <shivankg@amd.com> Suggested-by: Matthew Wilcox <willy@infradead.org> Co-developed-by: David Hildenbrand <david@redhat.com> Cc: Alistair Popple <apopple@nvidia.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Donet Tom <donettom@linux.ibm.com> Cc: Jane Chu <jane.chu@oracle.com> Cc: Kefeng Wang <wangkefeng.wang@huawei.com> Cc: Zi Yan <ziy@nvidia.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: 98c6d259319e ("mm/gup: check ref_count instead of lru before migration")
[ Take the new function in mm.h, removing "const" from its parameter to stop
build warnings; but avoid all the conflicts of using it in mm/migrate.c. ] Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
After commit 1aaf8c122918 ("mm: gup: fix infinite loop within
__get_longterm_locked") we are able to longterm pin folios that are not
supposed to get longterm pinned, simply because they temporarily have the
LRU flag cleared (esp. temporarily isolated).
For example, two __get_longterm_locked() callers can race, or
__get_longterm_locked() can race with anything else that temporarily
isolates folios.
The introducing commit mentions the use case of a driver that uses
vm_ops->fault to insert pages allocated through cma_alloc() into the page
tables, assuming they can later get longterm pinned. These pages/ folios
would never have the LRU flag set and consequently cannot get isolated.
There is no known in-tree user making use of that so far, fortunately.
To handle that in the future -- and avoid retrying forever to
isolate/migrate them -- we will need a different mechanism for the CMA
area *owner* to indicate that it actually already allocated the page and
is fine with longterm pinning it. The LRU flag is not suitable for that.
Probably we can lookup the relevant CMA area and query the bitmap; we only
have have to care about some races, probably. If already allocated, we
could just allow longterm pinning)
Anyhow, let's fix the "must not be longterm pinned" problem first by
reverting the original commit.
Link: https://lkml.kernel.org/r/20250611131314.594529-1-david@redhat.com Fixes: 1aaf8c122918 ("mm: gup: fix infinite loop within __get_longterm_locked") Signed-off-by: David Hildenbrand <david@redhat.com> Closes: https://lore.kernel.org/all/20250522092755.GA3277597@tiffany/ Reported-by: Hyesoo Yu <hyesoo.yu@samsung.com> Reviewed-by: John Hubbard <jhubbard@nvidia.com> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Peter Xu <peterx@redhat.com> Cc: Zhaoyang Huang <zhaoyang.huang@unisoc.com> Cc: Aijun Sun <aijun.sun@unisoc.com> Cc: Alistair Popple <apopple@nvidia.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Revert v6.1.129 commit c986a5fb15ed ] Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix a bug where the driver's event subscription logic for SRQ-related
events incorrectly sets obj_type for RMP objects.
When subscribing to SRQ events, get_legacy_obj_type() did not handle
the MLX5_CMD_OP_CREATE_RMP case, which caused obj_type to be 0
(default).
This led to a mismatch between the obj_type used during subscription
(0) and the value used during notification (1, taken from the event's
type field). As a result, event mapping for SRQ objects could fail and
event notification would not be delivered correctly.
This fix adds handling for MLX5_CMD_OP_CREATE_RMP in get_legacy_obj_type,
returning MLX5_EVENT_QUEUE_TYPE_RQ so obj_type is consistent between
subscription and notification.
Fixes: 759738537142 ("IB/mlx5: Enable subscription for device events over DEVX") Link: https://patch.msgid.link/r/8f1048e3fdd1fde6b90607ce0ed251afaf8a148c.1755088962.git.leon@kernel.org Signed-off-by: Or Har-Toov <ohartoov@nvidia.com> Reviewed-by: Edward Srouji <edwards@nvidia.com> Signed-off-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
We have found more and more devices that have the same problem, that
the mixer's minimum value is muted. Accroding to pipewire's MR[1]
and Arch Linux wiki[2], this should be a very common problem in USB
audio devices. Move the quirk into common quirk,as a preparation of
more devices' quirk's patch coming on the road[3].
1. https://gitlab.freedesktop.org/pipewire/pipewire/-/merge_requests/2514
2. https://wiki.archlinux.org/index.php?title=PipeWire&oldid=804138#No_sound_from_USB_DAC_until_30%_volume
3. On the road, in the physical sense. We have been buying ton of
these devices for testing the problem.
The vendor Comtrue Inc. (0x2fc6) produces USB audio chipsets like
the CT7601 which are capable of Native DSD playback.
This patch adds QUIRK_FLAG_DSD_RAW for Comtrue (VID 0x2fc6), which enables
native DSD playback (DSD_U32_LE) on their USB Audio device. This has been
verified under Ubuntu 25.04 with JRiver.
The recent addition of DualSense mixer quirk relies on the input
device handle, and the build can fail if CONFIG_INPUT isn't set.
Put (rather ugly) workarounds to wrap with IS_REACHABLE() for avoiding
the build error.
Replace comma between expressions with semicolons.
Using a ',' in place of a ';' can have unintended side effects.
Although that is not the case here, it is seems best to use ';'
unless ',' is intended.
Found by inspection.
No functional change intended.
Compile tested only.
Fixes: 79d561c4ec04 ("ALSA: usb-audio: Add mixer quirk for Sony DualSense PS5") Signed-off-by: Chen Ni <nichen@iscas.ac.cn> Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com> Link: https://patch.msgid.link/20250612060228.1518028-1-nichen@iscas.ac.cn Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
The Sony DualSense wireless controller (PS5) features an internal mono
speaker, but it also provides a 3.5mm jack socket for headphone output
and headset microphone input.
Since this is a UAC1 device, it doesn't advertise any jack detection
capability. However, the controller is able to report HP & MIC insert
events via HID, i.e. through a dedicated input device managed by the
hid-playstation driver.
Add a quirk to create the jack controls for headphone and headset mic,
respectively, and setup an input handler for each of them in order to
intercept the related hotplug events.
Adding a memory barrier before wake_up() in
snd_usb_soundblaster_remote_complete() is supposed to ensure the write
to mixer->rc_code is visible in wait_event_interruptible() from
snd_usb_sbrc_hwdep_read().
However, this is not really necessary, since wake_up() is just a wrapper
over __wake_up() which already executes a full memory barrier before
accessing the state of the task to be waken up.
Drop the redundant call to wmb() and implicitly fix the checkpatch
complaint:
Issuing two writes to the same af_alg socket is bogus as the
data will be interleaved in an unpredictable fashion. Furthermore,
concurrent writes may create inconsistencies in the internal
socket state.
Disallow this by adding a new ctx->write field that indiciates
exclusive ownership for writing.
Fixes: 8ff590903d5 ("crypto: algif_skcipher - User-space interface for skcipher operations") Reported-by: Muhammad Alifa Ramdhan <ramdhan@starlabs.sg> Reported-by: Bing-Jhong Billy Jheng <billy@starlabs.sg> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If earlier opening of source graph fails (e.g. ADSP rejects due to
incorrect audioreach topology), the graph is closed and
"dai_data->graph[dai->id]" is assigned NULL. Preparing the DAI for sink
graph continues though and next call to q6apm_lpass_dai_prepare()
receives dai_data->graph[dai->id]=NULL leading to NULL pointer
exception:
qcom-apm gprsvc:service:2:1: Error (1) Processing 0x01001002 cmd
qcom-apm gprsvc:service:2:1: DSP returned error[1001002] 1
q6apm-lpass-dais 30000000.remoteproc:glink-edge:gpr:service@1:bedais: fail to start APM port 78
q6apm-lpass-dais 30000000.remoteproc:glink-edge:gpr:service@1:bedais: ASoC: error at snd_soc_pcm_dai_prepare on TX_CODEC_DMA_TX_3: -22
Unable to handle kernel NULL pointer dereference at virtual address 00000000000000a8
...
Call trace:
q6apm_graph_media_format_pcm+0x48/0x120 (P)
q6apm_lpass_dai_prepare+0x110/0x1b4
snd_soc_pcm_dai_prepare+0x74/0x108
__soc_pcm_prepare+0x44/0x160
dpcm_be_dai_prepare+0x124/0x1c0
Fixes: 30ad723b93ad ("ASoC: qdsp6: audioreach: add q6apm lpass dai support") Cc: stable@vger.kernel.org Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>
Message-ID: <20250904101849.121503-2-krzysztof.kozlowski@linaro.org> Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is an issue around with error handling and graph management with
the exising code, none of the error paths close the graph, which result in
leaving the loaded graph in dsp, however the driver thinks otherwise.
This can have a nasty side effect specially when we try to load the same
graph to dsp, dsp returns error which leaves the board with no sound and
requires restart.
Fix this by properly closing the graph when we hit errors between
open and close.
On multiple prepare calls, its possible that the playback graphs are
not unloaded from the DSP, which can have some wierd side-effects,
one of them is that the data not consumed without any errors.
Fixes: c2ac3aec474d("ASoC: qcom: q6apm-lpass-dai: unprepare stream if its already prepared") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Link: https://lore.kernel.org/r/20230323164403.6654-2-srinivas.kandagatla@linaro.org Signed-off-by: Mark Brown <broonie@kernel.org>
Stable-dep-of: 68f27f7c7708 ("ASoC: qcom: q6apm-lpass-dais: Fix NULL pointer dereference if source graph failed") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commit 7d5e9737efda ("net: rfkill: gpio: get the name and type from
device property") rfkill_find_type() gets called with the possibly
uninitialized "const char *type_name;" local variable.
On x86 systems when rfkill-gpio binds to a "BCM4752" or "LNV4752"
acpi_device, the rfkill->type is set based on the ACPI acpi_device_id:
rfkill->type = (unsigned)id->driver_data;
and there is no "type" property so device_property_read_string() will fail
and leave type_name uninitialized, leading to a potential crash.
rfkill_find_type() does accept a NULL pointer, fix the potential crash
by initializing type_name to NULL.
Note likely sofar this has not been caught because:
1. Not many x86 machines actually have a "BCM4752"/"LNV4752" acpi_device
2. The stack happened to contain NULL where type_name is stored
Fixes: 7d5e9737efda ("net: rfkill: gpio: get the name and type from device property") Cc: stable@vger.kernel.org Cc: Heikki Krogerus <heikki.krogerus@linux.intel.com> Signed-off-by: Hans de Goede <hansg@kernel.org> Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com> Link: https://patch.msgid.link/20250913113515.21698-1-hansg@kernel.org Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Allow probing rfkill-gpio via device tree. This hooks up the already
existing support that was started in commit 262c91ee5e52 ("net:
rfkill: gpio: prepare for DT and ACPI support") via the "rfkill-gpio"
compatible, with the "name" and "type" properties renamed to "label"
and "radio-type", respectively, in the device tree case.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Link: https://lore.kernel.org/r/20230102-rfkill-gpio-dt-v2-2-d1b83758c16d@pengutronix.de Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Stable-dep-of: b6f56a44e4c1 ("net: rfkill: gpio: Fix crash due to dereferencering uninitialized pointer") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
IO errors were correctly printed to stderr, and propagated up to the
main loop for the server side, but the returned value was ignored. As a
consequence, the program for the listener side was no longer exiting
with an error code in case of IO issues.
Because of that, some issues might not have been seen. But very likely,
most issues either had an effect on the client side, or the file
transfer was not the expected one, e.g. the connection got reset before
the end. Still, it is better to fix this.
The main consequence of this issue is the error that was reported by the
selftests: the received and sent files were different, and the MIB
counters were not printed. Also, when such errors happened during the
'disconnect' tests, the program tried to continue until the timeout.
Now when an IO error is detected, the program exits directly with an
error.
When the MPTCP DATA FIN have been ACKed, there is no more MPTCP related
metadata to exchange, and all subflows can be safely shutdown.
Before this patch, the subflows were actually terminated at 'close()'
time. That's certainly fine most of the time, but not when the userspace
'shutdown()' a connection, without close()ing it. When doing so, the
subflows were staying in LAST_ACK state on one side -- and consequently
in FIN_WAIT2 on the other side -- until the 'close()' of the MPTCP
socket.
Now, when the DATA FIN have been ACKed, all subflows are shutdown. A
consequence of this is that the TCP 'FIN' flag can be set earlier now,
but the end result is the same. This affects the packetdrill tests
looking at the end of the MPTCP connections, but for a good reason.
Note that tcp_shutdown() will check the subflow state, so no need to do
that again before calling it.
Fixes: 3721b9b64676 ("mptcp: Track received DATA_FIN sequence number and add related helpers") Cc: stable@vger.kernel.org Fixes: 16a9a9da1723 ("mptcp: Add helper to process acks of DATA_FIN") Reviewed-by: Mat Martineau <martineau@kernel.org> Reviewed-by: Geliang Tang <geliang@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://patch.msgid.link/20250912-net-mptcp-fix-sft-connect-v1-1-d40e77cbbf02@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
[ Adjust context ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / thirdparty / kernel / stable.git / log
