eventfs uses the tracefs_inode and assumes that it's already initialized
to zero. That is, it doesn't set fields to zero (like ti->private) after
getting its tracefs_inode. This causes bugs due to stale values.
Just initialize the entire structure to zero on allocation so there isn't
any more surprises.
This is a partial fix to access to ti->private. The assignment still needs
to be made before the dentry is instantiated.
Link: https://lore.kernel.org/linux-trace-kernel/20240131185512.315825944@goodmis.org Cc: stable@vger.kernel.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Christian Brauner <brauner@kernel.org> Cc: Al Viro <viro@ZenIV.linux.org.uk> Cc: Ajay Kaher <ajay.kaher@broadcom.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Fixes: 5790b1fb3d672 ("eventfs: Remove eventfs_file and just use eventfs_inode") Reported-by: kernel test robot <oliver.sang@intel.com> Closes: https://lore.kernel.org/oe-lkp/202401291043.e62e89dc-oliver.sang@intel.com Suggested-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The eventfs inodes and directories are allocated when referenced. But this
leaves the issue of keeping consistent inode numbers and the number is
only saved in the inode structure itself. When the inode is no longer
referenced, it can be freed. When the file that the inode was representing
is referenced again, the inode is once again created, but the inode number
needs to be the same as it was before.
Just making the inode numbers the same for all files is fine, but that
does not work with directories. The find command will check for loops via
the inode number and having the same inode number for directories triggers:
# find /sys/kernel/tracing
find: File system loop detected;
'/sys/kernel/debug/tracing/events/initcall/initcall_finish' is part of the same file system loop as
'/sys/kernel/debug/tracing/events/initcall'.
[..]
Linus pointed out that the eventfs_inode structure ends with a single
32bit int, and on 64 bit machines, there's likely a 4 byte hole due to
alignment. We can use this hole to store the inode number for the
eventfs_inode. All directories in eventfs are represented by an
eventfs_inode and that data structure can hold its inode number.
That last int was also purposely placed at the end of the structure to
prevent holes from within. Now that there's a 4 byte number to hold the
inode, both the inode number and the last integer can be moved up in the
structure for better cache locality, where the llist and rcu fields can be
moved to the end as they are only used when the eventfs_inode is being
deleted.
As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.
So, use the purpose specific kcalloc() function instead of the argument
size * count in the kzalloc() function.
The original eventfs code added a wrapper around the dcache_readdir open
callback and created all the dentries and inodes at open, and increment
their ref count. A wrapper was added around the dcache_readdir release
function to decrement all the ref counts of those created inodes and
dentries. But this proved to be buggy[1] for when a kprobe was created
during a dir read, it would create a dentry between the open and the
release, and because the release would decrement all ref counts of all
files and directories, that would include the kprobe directory that was
not there to have its ref count incremented in open. This would cause the
ref count to go to negative and later crash the kernel.
To solve this, the dentries and inodes that were created and had their ref
count upped in open needed to be saved. That list needed to be passed from
the open to the release, so that the release would only decrement the ref
counts of the entries that were incremented in the open.
Unfortunately, the dcache_readdir logic was already using the
file->private_data, which is the only field that can be used to pass
information from the open to the release. What was done was the eventfs
created another descriptor that had a void pointer to save the
dcache_readdir pointer, and it wrapped all the callbacks, so that it could
save the list of entries that had their ref counts incremented in the
open, and pass it to the release. The wrapped callbacks would just put
back the dcache_readdir pointer and call the functions it used so it could
still use its data[2].
But Linus had an issue with the "hijacking" of the file->private_data
(unfortunately this discussion was on a security list, so no public link).
Which we finally agreed on doing everything within the iterate_shared
callback and leave the dcache_readdir out of it[3]. All the information
needed for the getents() could be created then.
But this ended up being buggy too[4]. The iterate_shared callback was not
the right place to create the dentries and inodes. Even Christian Brauner
had issues with that[5].
An attempt was to go back to creating the inodes and dentries at
the open, create an array to store the information in the
file->private_data, and pass that information to the other callbacks.[6]
The difference between that and the original method, is that it does not
use dcache_readdir. It also does not up the ref counts of the dentries and
pass them. Instead, it creates an array of a structure that saves the
dentry's name and inode number. That information is used in the
iterate_shared callback, and the array is freed in the dir release. The
dentries and inodes created in the open are not used for the iterate_share
or release callbacks. Just their names and inode numbers.
Linus did not like that either[7] and just wanted to remove the dentries
being created in iterate_shared and use the hard coded inode numbers.
[ All this while Linus enjoyed an unexpected vacation during the merge
window due to lack of power. ]
The dentries and inodes are created in the readdir for the sole purpose of
getting a consistent inode number. Linus stated that is unnecessary, and
that all inodes can have the same inode number. For a virtual file system
they are pretty meaningless.
Instead use a single unique inode number for all files and one for all
directories.
As the ei->entries array is fixed for the duration of the eventfs_inode,
it can be used to skip over already read entries in eventfs_iterate().
That is, if ctx->pos is greater than zero, there's no reason in doing the
loop across the ei->entries array for the entries less than ctx->pos.
Instead, start the lookup of the entries at the current ctx->pos.
In order to apply a shortcut to skip over the current ctx->pos
immediately, by using the ei->entries array, the reading of that array
should be first. Moving the array reading before the linked list reading
will make the shortcut change diff nicer to read.
The ctx->pos was only updated when it added an entry, but the "skip to
current pos" check (c--) happened for every loop regardless of if the
entry was added or not. This inconsistency caused readdir to be incorrect.
It was due to:
for (i = 0; i < ei->nr_entries; i++) {
if (c > 0) {
c--;
continue;
}
mutex_lock(&eventfs_mutex);
/* If ei->is_freed then just bail here, nothing more to do */
if (ei->is_freed) {
mutex_unlock(&eventfs_mutex);
goto out;
}
r = entry->callback(name, &mode, &cdata, &fops);
mutex_unlock(&eventfs_mutex);
[..]
ctx->pos++;
}
But this can cause the iterator to return a file that was already read.
That's because of the way the callback() works. Some events may not have
all files, and the callback can return 0 to tell eventfs to skip the file
for this directory.
for instance, we have:
# ls /sys/kernel/tracing/events/ftrace/function
format hist hist_debug id inject
and
# ls /sys/kernel/tracing/events/sched/sched_switch/
enable filter format hist hist_debug id inject trigger
Where the function directory is missing "enable", "filter" and
"trigger". That's because the callback() for events has:
/*
* Only event directories that can be enabled should have
* triggers or filters, with the exception of the "print"
* event that can have a "trigger" file.
*/
if (!(call->flags & TRACE_EVENT_FL_IGNORE_ENABLE)) {
if (call->class->reg && strcmp(name, "enable") == 0) {
*mode = TRACE_MODE_WRITE;
*fops = &ftrace_enable_fops;
return 1;
}
Where the function event has the TRACE_EVENT_FL_IGNORE_ENABLE set.
This means that the entries array elements for "enable", "filter" and
"trigger" when called on the function event will have the callback return
0 and not 1, to tell eventfs to skip these files for it.
Because the "skip to current ctx->pos" check happened for all entries, but
the ctx->pos++ only happened to entries that exist, it would confuse the
reading of a directory. Which would cause:
# ls /sys/kernel/tracing/events/ftrace/function/
format hist hist hist_debug hist_debug id inject inject
The missing "enable", "filter" and "trigger" caused ls to show "hist",
"hist_debug" and "inject" twice.
Update the ctx->pos for every iteration to keep its update and the "skip"
update consistent. This also means that on error, the ctx->pos needs to be
decremented if it was incremented without adding something.
Link: https://lore.kernel.org/all/20240104150500.38b15a62@gandalf.local.home/ Link: https://lore.kernel.org/linux-trace-kernel/20240104220048.172295263@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christian Brauner <brauner@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Fixes: 493ec81a8fb8e ("eventfs: Stop using dcache_readdir() for getdents()") Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If ei->is_freed is set in eventfs_iterate(), it means that the directory
that is being iterated on is in the process of being freed. Just exit the
loop immediately when that is ever detected, and separate out the return
of the entry->callback() from ei->is_freed.
Link: https://lore.kernel.org/linux-trace-kernel/20240104220048.016261289@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christian Brauner <brauner@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The eventfs creates dynamically allocated dentries and inodes. Using the
dcache_readdir() logic for its own directory lookups requires hiding the
cursor of the dcache logic and playing games to allow the dcache_readdir()
to still have access to the cursor while the eventfs saved what it created
and what it needs to release.
Instead, just have eventfs have its own iterate_shared callback function
that will fill in the dent entries. This simplifies the code quite a bit.
Link: https://lore.kernel.org/linux-trace-kernel/20240104015435.682218477@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ajay Kaher <akaher@vmware.com> Cc: Al Viro <viro@ZenIV.linux.org.uk> Cc: Christian Brauner <brauner@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The "lookup" parameter is a way to differentiate the call to
create_file/dir_dentry() from when it's just a lookup (no need to up the
dentry refcount) and accessed via a readdir (need to up the refcount).
But reality, it just makes the code more complex. Just up the refcount and
let the caller decide to dput() the result or not.
If the power domains are registered first with genpd and *after that*
the driver attempts to power them on in the probe sequence, then it is
possible that a race condition occurs if genpd tries to power them on
in the same time.
The same is valid for powering them off before unregistering them
from genpd.
Attempt to fix race conditions by first removing the domains from genpd
and *after that* powering down domains.
Also first power up the domains and *after that* register them
to genpd.
There are a ton of build errors when REGMAP is not set, so select
REGMAP to fix all of them.
Examples (not all of them):
../drivers/iio/imu/bno055/bno055_ser_core.c:495:15: error: variable 'bno055_ser_regmap_bus' has initializer but incomplete type
495 | static struct regmap_bus bno055_ser_regmap_bus = {
../drivers/iio/imu/bno055/bno055_ser_core.c:496:10: error: 'struct regmap_bus' has no member named 'write'
496 | .write = bno055_ser_write_reg,
../drivers/iio/imu/bno055/bno055_ser_core.c:497:10: error: 'struct regmap_bus' has no member named 'read'
497 | .read = bno055_ser_read_reg,
../drivers/iio/imu/bno055/bno055_ser_core.c: In function 'bno055_ser_probe':
../drivers/iio/imu/bno055/bno055_ser_core.c:532:18: error: implicit declaration of function 'devm_regmap_init'; did you mean 'vmem_map_init'? [-Werror=implicit-function-declaration]
532 | regmap = devm_regmap_init(&serdev->dev, &bno055_ser_regmap_bus,
../drivers/iio/imu/bno055/bno055_ser_core.c:532:16: warning: assignment to 'struct regmap *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
532 | regmap = devm_regmap_init(&serdev->dev, &bno055_ser_regmap_bus,
../drivers/iio/imu/bno055/bno055_ser_core.c: At top level:
../drivers/iio/imu/bno055/bno055_ser_core.c:495:26: error: storage size of 'bno055_ser_regmap_bus' isn't known
495 | static struct regmap_bus bno055_ser_regmap_bus = {
Fixes: 2eef5a9cc643 ("iio: imu: add BNO055 serdev driver") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Andrea Merello <andrea.merello@iit.it> Cc: Jonathan Cameron <jic23@kernel.org> Cc: Lars-Peter Clausen <lars@metafoo.de> Cc: linux-iio@vger.kernel.org Cc: <Stable@vger.kernel.org> Link: https://lore.kernel.org/r/20240110185611.19723-1-rdunlap@infradead.org Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Aligning the buffer to the L1 cache is not sufficient in some platforms
as they might have larger cacheline sizes for caches after L1 and thus,
we can't guarantee DMA safety.
That was the whole reason to introduce IIO_DMA_MINALIGN in [1]. Do the same
for the sigma_delta ADCs.
Aligning the buffer to the L1 cache is not sufficient in some platforms
as they might have larger cacheline sizes for caches after L1 and thus,
we can't guarantee DMA safety.
That was the whole reason to introduce IIO_DMA_MINALIGN in [1]. Do the same
for the sigma_delta ADCs.
Fixes: 0fb6ee8d0b5e ("iio: ad_sigma_delta: Don't put SPI transfer buffer on the stack") Signed-off-by: Nuno Sa <nuno.sa@analog.com> Link: https://lore.kernel.org/r/20240117-dev_sigma_delta_no_irq_flags-v1-1-db39261592cf@analog.com Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The kernel fails when compiling without `CONFIG_REGMAP_I2C` but with
`CONFIG_BMA400`.
```
ld: drivers/iio/accel/bma400_i2c.o: in function `bma400_i2c_probe':
bma400_i2c.c:(.text+0x23): undefined reference to `__devm_regmap_init_i2c'
```
Aligning the buffer to the L1 cache is not sufficient in some platforms
as they might have larger cacheline sizes for caches after L1 and thus,
we can't guarantee DMA safety.
That was the whole reason to introduce IIO_DMA_MINALIGN in [1]. Do the same
for st_sensors common buffer.
While at it, moved the odr_lock before buffer_data as we definitely
don't want any other data to share a cacheline with the buffer.
Recently, we encounter kernel crash in function rm3100_common_probe
caused by out of bound access of array rm3100_samp_rates (because of
underlying hardware failures). Add boundary check to prevent out of
bound access.
Fixes: 121354b2eceb ("iio: magnetometer: Add driver support for PNI RM3100") Suggested-by: Zhouyi Zhou <zhouzhouyi@gmail.com> Signed-off-by: zhili.liu <zhili.liu@ucas.com.cn> Link: https://lore.kernel.org/r/1704157631-3814-1-git-send-email-zhouzhouyi@gmail.com Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 4c3577db3e4f ("Staging: iio: impedance-analyzer: Fix sparse
warning") fixed a compiler warning, but introduced a bug that resulted
in one of the two 16 bit IIO channels always being zero (when both are
enabled).
This is because int is 32 bits wide on most architectures and in the
case of a little-endian machine the two most significant bytes would
occupy the buffer for the second channel as 'val' is being passed as a
void pointer to 'iio_push_to_buffers()'.
Fix by defining 'val' as u16. Tested working on ARM64.
The commit allowed workqueue_apply_unbound_cpumask() to clear __WQ_ORDERED
on now removed implicitly ordered workqueues. This was incorrect in that
system-wide config change shouldn't break ordering properties of all
workqueues. The reason why apply_workqueue_attrs() path was allowed to do so
was because it was targeting the specific workqueue - either the workqueue
had WQ_SYSFS set or the workqueue user specifically tried to change
max_active, both of which indicate that the workqueue doesn't need to be
ordered.
The implicitly ordered workqueue promotion was removed by the previous
commit 3bc1e711c26b ("workqueue: Don't implicitly make UNBOUND workqueues w/
@max_active==1 ordered"). However, it didn't update this path and broke
build. Let's revert the commit which was incorrect in the first place which
also fixes build.
Fix to search a field from the structure which has anonymous union
correctly.
Since the reference `type` pointer was updated in the loop, the search
loop suddenly aborted where it hits an anonymous union. Thus it can not
find the field after the anonymous union. This avoids updating the
cursor `type` pointer in the loop.
Since the BTF type setting updates probe_arg::type, the type size
calculation and setting print-fmt should be done after that.
Without this fix, the argument size and print-fmt can be wrong.
Fix to show a parse error for bad type (non-string) for $comm/$COMM and
immediate-string. With this fix, error_log file shows appropriate error
message as below.
[ 30.144183] trace_kprobe: error: $comm and immediate-string only accepts string type
Command: p vfs_read $comm:u32
^
[ 62.618500] trace_kprobe: error: $comm and immediate-string only accepts string type
Command: p vfs_read \"hoge":u32
^ Link: https://lore.kernel.org/all/170602215411.215583.2238016352271091852.stgit@devnote2/ Fixes: 3dd1f7f24f8c ("tracing: probeevent: Fix to make the type of $comm string") Cc: stable@vger.kernel.org Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix trace_string() by assigning the string length to the return variable
which got lost in commit ddeea494a16f ("tracing/synthetic: Use union
instead of casts") and caused trace_string() to always return 0.
Link: https://lore.kernel.org/linux-trace-kernel/20240214220555.711598-1-thorsten.blum@toblux.com Cc: stable@vger.kernel.org Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Fixes: ddeea494a16f ("tracing/synthetic: Use union instead of casts") Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org> Signed-off-by: Thorsten Blum <thorsten.blum@toblux.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While looking at improving the saved_cmdlines cache I found a huge amount
of wasted memory that should be used for the cmdlines.
The tracing data saves pids during the trace. At sched switch, if a trace
occurred, it will save the comm of the task that did the trace. This is
saved in a "cache" that maps pids to comms and exposed to user space via
the /sys/kernel/tracing/saved_cmdlines file. Currently it only caches by
default 128 comms.
The structure that uses this creates an array to store the pids using
PID_MAX_DEFAULT (which is usually set to 32768). This causes the structure
to be of the size of 131104 bytes on 64 bit machines.
In hex: 131104 = 0x20020, and since the kernel allocates generic memory in
powers of two, the kernel would allocate 0x40000 or 262144 bytes to store
this structure. That leaves 131040 bytes of wasted space.
Worse, the structure points to an allocated array to store the comm names,
which is 16 bytes times the amount of names to save (currently 128), which
is 2048 bytes. Instead of allocating a separate array, make the structure
end with a variable length string and use the extra space for that.
This is similar to a recommendation that Linus had made about eventfs_inode names:
Instead of allocating a separate string array to hold the saved comms,
have the structure end with: char saved_cmdlines[]; and round up to the
next power of two over sizeof(struct saved_cmdline_buffers) + num_cmdlines * TASK_COMM_LEN
It will use this extra space for the saved_cmdline portion.
Now, instead of saving only 128 comms by default, by using this wasted
space at the end of the structure it can save over 8000 comms and even
saves space by removing the need for allocating the other array.
Link: https://lore.kernel.org/linux-trace-kernel/20240209063622.1f7b6d5f@rorschach.local.home Cc: stable@vger.kernel.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Vincent Donnefort <vdonnefort@google.com> Cc: Sven Schnelle <svens@linux.ibm.com> Cc: Mete Durlu <meted@linux.ibm.com> Fixes: 939c7a4f04fcd ("tracing: Introduce saved_cmdlines_size file") Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently, the timerlat's hrtimer is initialized at the first read of
timerlat_fd, and destroyed at close(). It works, but it causes an error
if the user program open() and close() the file without reading.
After updating bb_free in mb_free_blocks, it is possible to return without
updating bb_fragments because the block being freed is found to have
already been freed, which leads to inconsistency between bb_free and
bb_fragments.
Since the group may be unlocked in ext4_grp_locked_error(), this can lead
to problems such as dividing by zero when calculating the average fragment
length. Hence move the update of bb_free to after the block double-free
check guarantees that the corresponding statistics are updated only after
the core block bitmap is modified.
Fixes: eabe0444df90 ("ext4: speed-up releasing blocks on commit") CC: <stable@vger.kernel.org> # 3.10 Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20240104142040.2835097-5-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In ext4_move_extents(), moved_len is only updated when all moves are
successfully executed, and only discards orig_inode and donor_inode
preallocations when moved_len is not zero. When the loop fails to exit
after successfully moving some extents, moved_len is not updated and
remains at 0, so it does not discard the preallocations.
If the moved extents overlap with the preallocated extents, the
overlapped extents are freed twice in ext4_mb_release_inode_pa() and
ext4_process_freed_data() (as described in commit 94d7c16cbbbd ("ext4:
Fix double-free of blocks with EXT4_IOC_MOVE_EXT")), and bb_free is
incremented twice. Hence when trim is executed, a zero-division bug is
triggered in mb_update_avg_fragment_size() because bb_free is not zero
and bb_fragments is zero.
Therefore, update move_len after each extent move to avoid the issue.
Reported-by: Wei Chen <harperchen1110@gmail.com> Reported-by: xingwei lee <xrivendell7@gmail.com> Closes: https://lore.kernel.org/r/CAO4mrferzqBUnCag8R3m2zf897ts9UEuhjFQGPtODT92rYyR2Q@mail.gmail.com Fixes: fcf6b1b729bc ("ext4: refactor ext4_move_extents code base") CC: <stable@vger.kernel.org> # 3.18 Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20240104142040.2835097-2-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In remoteproc shutdown sequence, rpmsg_remove will get called which
would depopulate all the child nodes that have been created during
rpmsg_probe. This would result in cb_remove call for all the context
banks for the remoteproc. In cb_remove function, session 0 is
getting skipped which is not correct as session 0 will never become
available again. Add changes to mark session 0 also as invalid.
In (e)poll mode, threads often depend on I/O events to determine when
data is ready for consumption. Within binder, a thread may initiate a
command via BINDER_WRITE_READ without a read buffer and then make use
of epoll_wait() or similar to consume any responses afterwards.
It is then crucial that epoll threads are signaled via wakeup when they
queue their own work. Otherwise, they risk waiting indefinitely for an
event leaving their work unhandled. What is worse, subsequent commands
won't trigger a wakeup either as the thread has pending work.
Customer has reported an issue with specific desktop platform
where two CS42L42 codecs are connected to CS8409 HDA bridge.
If "Master Volume Control" is created then on Ubuntu OS UCM
left/right balance slider in UI audio settings has no effect.
This patch will fix this issue for a target paltform.
WCD938x sound codec driver ignores return status of getting regulators
and returns EINVAL instead of EPROBE_DEFER. If regulator provider
probes after the codec, system is left without probed audio:
wcd938x_codec audio-codec: wcd938x_probe: Fail to obtain platform data
wcd938x_codec: probe of audio-codec failed with error -22
Fixes: 16572522aece ("ASoC: codecs: wcd938x-sdw: add SoundWire driver") Cc: <stable@vger.kernel.org> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Link: https://msgid.link/r/20240117151208.1219755-1-krzysztof.kozlowski@linaro.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After the linked LLVM change, building ARCH=um defconfig results in a
segmentation fault in modpost. Prior to commit a23e7584ecf3 ("modpost:
unify 'sym' and 'to' in default_mismatch_handler()"), there was a
warning:
WARNING: modpost: vmlinux.o(__ex_table+0x88): Section mismatch in reference to the .ltext:(unknown)
WARNING: modpost: The relocation at __ex_table+0x88 references
section ".ltext" which is not in the list of
authorized sections. If you're adding a new section
and/or if this reference is valid, add ".ltext" to the
list of authorized sections to jump to on fault.
This can be achieved by adding ".ltext" to
OTHER_TEXT_SECTIONS in scripts/mod/modpost.c.
The linked LLVM change moves global objects to the '.ltext' (and
'.ltext.*' with '-ffunction-sections') sections with '-mcmodel=large',
which ARCH=um uses. These sections should be handled just as '.text'
and '.text.*' are, so add them to TEXT_SECTIONS.
The kernel builds with -fno-PIE, so commit 883354afbc10 ("um: link
vmlinux with -no-pie") added the compiler linker flag '-no-pie' via
cc-option because '-no-pie' was only supported in GCC 6.1.0 and newer.
While this works for GCC, this does not work for clang because cc-option
uses '-c', which stops the pipeline right before linking, so '-no-pie'
is unconsumed and clang warns, causing cc-option to fail just as it
would if the option was entirely unsupported:
$ clang -Werror -no-pie -c -o /dev/null -x c /dev/null
clang-16: error: argument unused during compilation: '-no-pie' [-Werror,-Wunused-command-line-argument]
A recent version of clang exposes this because it generates a relocation
under '-mcmodel=large' that is not supported in PIE mode:
/usr/sbin/ld: init/main.o: relocation R_X86_64_32 against symbol `saved_command_line' can not be used when making a PIE object; recompile with -fPIE
/usr/sbin/ld: failed to set dynamic section sizes: bad value
clang: error: linker command failed with exit code 1 (use -v to see invocation)
Remove the cc-option check altogether. It is wasteful to invoke the
compiler to check for '-no-pie' because only one supported compiler
version does not support it, GCC 5.x (as it is supported with the
minimum version of clang and GCC 6.1.0+). Use a combination of the
gcc-min-version macro and CONFIG_CC_IS_CLANG to unconditionally add
'-no-pie' with CONFIG_LD_SCRIPT_DYN=y, so that it is enabled with all
compilers that support this. Furthermore, using gcc-min-version can help
turn this back into
LINK-$(CONFIG_LD_SCRIPT_DYN) += -no-pie
when the minimum version of GCC is bumped past 6.1.0.
Invoking the make_tx_response() / push_tx_responses() pair with no lock
held would be acceptable only if all such invocations happened from the
same context (NAPI instance or dealloc thread). Since this isn't the
case, and since the interface "spec" also doesn't demand that multicast
operations may only be performed with no in-flight transmits,
MCAST_{ADD,DEL} processing also needs to acquire the response lock
around the invocations.
To prevent similar mistakes going forward, "downgrade" the present
functions to private helpers of just the two remaining ones using them
directly, with no forward declarations anymore. This involves renaming
what so far was make_tx_response(), for the new function of that name
to serve the new (wrapper) purpose.
While there,
- constify the txp parameters,
- correct xenvif_idx_release()'s status parameter's type,
- rename {,_}make_tx_response()'s status parameters for consistency with
xenvif_idx_release()'s.
Fixes: 210c34dcd8d9 ("xen-netback: add support for multicast control") Cc: stable@vger.kernel.org Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Paul Durrant <paul@xen.org> Link: https://lore.kernel.org/r/980c6c3d-e10e-4459-8565-e8fbde122f00@suse.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When using hotplug and bringing up a 32-bit CPU, ask the firmware about the
BTLB information to set up the static (block) TLB entries.
For that write access to the static btlb_info struct is needed, but
since it is marked __ro_after_init the kernel segfaults with missing
write permissions.
Fix the crash by dropping the __ro_after_init annotation.
The current exception handler implementation, which assists when accessing
user space memory, may exhibit random data corruption if the compiler decides
to use a different register than the specified register %r29 (defined in
ASM_EXCEPTIONTABLE_REG) for the error code. If the compiler choose another
register, the fault handler will nevertheless store -EFAULT into %r29 and thus
trash whatever this register is used for.
Looking at the assembly I found that this happens sometimes in emulate_ldd().
To solve the issue, the easiest solution would be if it somehow is
possible to tell the fault handler which register is used to hold the error
code. Using %0 or %1 in the inline assembly is not posssible as it will show
up as e.g. %r29 (with the "%r" prefix), which the GNU assembler can not
convert to an integer.
This patch takes another, better and more flexible approach:
We extend the __ex_table (which is out of the execution path) by one 32-word.
In this word we tell the compiler to insert the assembler instruction
"or %r0,%r0,%reg", where %reg references the register which the compiler
choosed for the error return code.
In case of an access failure, the fault handler finds the __ex_table entry and
can examine the opcode. The used register is encoded in the lowest 5 bits, and
the fault handler can then store -EFAULT into this register.
Since we extend the __ex_table to 3 words we can't use the BUILDTIME_TABLE_SORT
config option any longer.
Syzkaller reported [1] hitting a warning after failing to allocate
resources for skb in hsr_init_skb(). Since a WARN_ONCE() call will
not help much in this case, it might be prudent to switch to
netdev_warn_once(). At the very least it will suppress syzkaller
reports such as [1].
Just in case, use netdev_warn_once() in send_prp_supervision_frame()
for similar reasons.
rx_data_reassembly skb is stored during NCI data exchange for processing
fragmented packets. It is dropped only when the last fragment is processed
or when an NTF packet with NCI_OP_RF_DEACTIVATE_NTF opcode is received.
However, the NCI device may be deallocated before that which leads to skb
leak.
As by design the rx_data_reassembly skb is bound to the NCI device and
nothing prevents the device to be freed before the skb is processed in
some way and cleaned, free it on the NCI device cleanup.
Found by Linux Verification Center (linuxtesting.org) with Syzkaller.
Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation") Cc: stable@vger.kernel.org Reported-by: syzbot+6b7c68d9c21e4ee4251b@syzkaller.appspotmail.com Closes: https://lore.kernel.org/lkml/000000000000f43987060043da7b@google.com/ Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 90ceddcb4950 ("bpf: Support llvm-objcopy for vmlinux BTF")
changed the ELF type of .btf.vmlinux.bin.o to ET_REL via dd, which works
fine for little endian platforms:
While in the area, update the comment to mention that binutils 2.35+
matches LLD's behavior of rejecting an ET_EXEC input, which occurred
after the comment was added.
There currently exists two thinkpad headset jack fixups:
ALC285_FIXUP_THINKPAD_NO_BASS_SPK_HEADSET_JACK
ALC285_FIXUP_THINKPAD_HEADSET_JACK
The latter is applied to alc285 and alc287 thinkpads which contain
bass speakers.
However, the former was only being applied to alc285 thinkpads,
leaving non-bass alc287 thinkpads with no headset button controls.
This patch fixes that by adding ALC285_FIXUP_THINKPAD_NO_BASS_SPK_HEADSET_JACK
to the alc287 chains, allowing the detection of headset buttons.
The inode_getsecctx LSM hook has previously been corrected to have
-EOPNOTSUPP instead of 0 as the default return value to fix BPF LSM
behavior. However, the call_int_hook()-generated loop in
security_inode_getsecctx() was left treating 0 as the neutral value, so
after an LSM returns 0, the loop continues to try other LSMs, and if one
of them returns a non-zero value, the function immediately returns with
said value. So in a situation where SELinux and the BPF LSMs registered
this hook, -EOPNOTSUPP would be incorrectly returned whenever SELinux
returned 0.
Fix this by open-coding the call_int_hook() loop and making it use the
correct LSM_RET_DEFAULT() value as the neutral one, similar to what
other hooks do.
For these hooks the true "neutral" value is -EOPNOTSUPP, which is
currently what is returned when no LSM provides this hook and what LSMs
return when there is no security context set on the socket. Correct the
value in <linux/lsm_hooks.h> and adjust the dispatch functions in
security/security.c to avoid issues when the BPF LSM is enabled.
Cc: stable@vger.kernel.org Fixes: 98e828a0650f ("security: Refactor declaration of LSM hooks") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
[PM: subject line tweak] Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[why]
MAX_SURFACES is per stream, while MAX_PLANES is per asic. The
mpc_combine is an array that records all the planes per asic. Therefore
MAX_PLANES should be used as the array size. Using MAX_SURFACES causes
array overflow when there are more than 3 planes.
[how]
Use the MAX_PLANES for the mpc_combine array size.
Cc: Mario Limonciello <mario.limonciello@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Reviewed-by: Rodrigo Siqueira <rodrigo.siqueira@amd.com> Reviewed-by: Nevenko Stupar <nevenko.stupar@amd.com> Reviewed-by: Chaitanya Dhere <chaitanya.dhere@amd.com> Acked-by: Tom Chung <chiahsuan.chung@amd.com> Signed-off-by: Wenjing Liu <wenjing.liu@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Set the segment size of the virtio_gpu device to the value
used by the drm helpers when allocating sg lists to fix the
following complaint from DMA_API debug code:
DMA-API: virtio-pci 0000:07:00.0: mapping sg segment longer than
device claims to support [len=262144] [max=65536]
MCSPI controller have few limitations regarding the transaction
size when the FIFO buffer is enabled and the WCNT feature is used
to find the end of word, in this case if WCNT is not a multiple of
the FIFO Almost Empty Level (AEL), then the FIFO empty event is not
generated correctly. In addition to this limitation, few other unknown
sequence of events that causes the FIFO empty status to not reflect the
exact status were found when FIFO is being used without DMA enabled
during extended testing in AM65x platform. Till the exact root cause
is found and fixed, revert the FIFO support without DMA.
See J721E Technical Reference Manual (SPRUI1C), section 12.1.5
for further details: http://www.ti.com/lit/pdf/spruil1
This reverts commit 75223bbea840e ("spi: omap2-mcspi: Add FIFO support
without DMA")
This reverts commit c46bfba1337d ("connector: Fix proc_event_num_listeners
count not cleared").
It is not accurate to reset proc_event_num_listeners according to
cn_netlink_send_mult() return value -ESRCH.
In the case of stress-ng netlink-proc, -ESRCH will always be returned,
because netlink_broadcast_filtered will return -ESRCH,
which may cause stress-ng netlink-proc performance degradation.
Reported-by: kernel test robot <oliver.sang@intel.com> Closes: https://lore.kernel.org/oe-lkp/202401112259.b23a1567-oliver.sang@intel.com Fixes: c46bfba1337d ("connector: Fix proc_event_num_listeners count not cleared") Signed-off-by: Keqi Wang <wangkeqi_chris@163.com> Link: https://lore.kernel.org/r/20240209091659.68723-1-wangkeqi_chris@163.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Changing the locking order means that scheduler/msm_job_run() can race
with the recovery kthread worker, with the result that the GPU gets an
extra runpm get when we are trying to power it off. Leaving the GPU in
an unrecovered state.
I'll need to come up with a different scheme for appeasing lockdep.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Patchwork: https://patchwork.freedesktop.org/patch/573835/ Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ab4750332dbe ("drm/amdgpu/sdma5.2: add begin/end_use ring
callbacks") caused GFXOFF control to be used more heavily and the
codepath that was removed from commit 0dee72639533 ("drm/amd: flush any
delayed gfxoff on suspend entry") now can be exercised at suspend again.
Users report that by using GNOME to suspend the lockscreen trigger will
cause SDMA traffic and the system can deadlock.
The rkisp1 does share interrupt lines on some platforms, after all. Thus
we need to revert this, and implement a fix for the rkisp1 shared irq
handling in a follow-up patch.
Fastopen and PM-trigger subflow shutdown can race, as reported by
syzkaller.
In my first attempt to close such race, I missed the fact that
the subflow status can change again before the subflow_state_change
callback is invoked.
Address the issue additionally copying with all the states directly
reachable from TCP_FIN_WAIT1.
Fixes: 1e777f39b4d7 ("mptcp: add MSG_FASTOPEN sendmsg flag support") Fixes: 4fd19a307016 ("mptcp: fix inconsistent state on fastopen race") Cc: stable@vger.kernel.org Reported-by: syzbot+c53d4d3ddb327e80bc51@syzkaller.appspotmail.com Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/458 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Before adding a new entry in mptcp_userspace_pm_get_local_id(), it's
better to check whether this address is already in userspace pm local
address list. If it's in the list, no need to add a new entry, just
return it's address ID and use this address.
Fixes: 8b20137012d9 ("mptcp: read attributes of addr entries managed by userspace PMs") Cc: stable@vger.kernel.org Signed-off-by: Geliang Tang <geliang.tang@linux.dev> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mptcp_rcv_space_init() is supposed to happen under the msk socket
lock, but active msk socket does that without such protection.
Leverage the existing mptcp_propagate_state() helper to that extent.
We need to ensure mptcp_rcv_space_init will happen before
mptcp_rcv_space_adjust(), and the release_cb does not assure that:
explicitly check for such condition.
While at it, move the wnd_end initialization out of mptcp_rcv_space_init(),
it never belonged there.
Note that the race does not produce ill effect in practice, but
change allows cleaning-up and defying better the locking model.
Fixes: a6b118febbab ("mptcp: add receive buffer auto-tuning") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Such field is there to avoid acquiring the data lock in a few spots,
but it adds complexity to the already non trivial locking schema.
All the relevant call sites (mptcp-level re-injection, set socket
options), are slow-path, drop such field in favor of 'cb_flags', adding
the relevant locking.
This patch could be seen as an improvement, instead of a fix. But it
simplifies the next patch. The 'Fixes' tag has been added to help having
this series backported to stable.
Fixes: e9d09baca676 ("mptcp: avoid atomic bit manipulation when possible") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To avoid duplicated code in different MPTCP selftests, we can add
and use helpers defined in mptcp_lib.sh.
Export kill_wait() helper in userspace_pm.sh into mptcp_lib.sh and
rename it as mptcp_lib_kill_wait(). It can be used to instead of
kill_wait() in mptcp_join.sh. Use the new helper in both scripts.
If a CI executes the same selftest multiple times with different
options, all results from the same subtests will have the same title,
which confuse the CI. With the same title printed in TAP, the tests are
considered as the same ones.
Now, it is possible to override this prefix by using MPTCP_LIB_KSFT_TEST
env var, and have a different title.
While at it, use 'basename' to remove the suffix as well instead of
using an extra 'sed'.
On very slow environments -- e.g. when QEmu is used without KVM --,
mptcp_join.sh selftest can take a bit more than 20 minutes. Bump the
default timeout by 50% as it seems normal to take that long on some
environments.
When a debug kernel config is used, this selftest will take even longer,
but that's certainly not a common test env to consider for the timeout.
The Fixes tag that has been picked here is there simply to help having
this patch backported to older stable versions. It is difficult to point
to the exact commit that made some env reaching the timeout from time to
time.
Fixes: d17b968b9876 ("selftests: mptcp: increase timeout to 20 minutes") Cc: stable@vger.kernel.org Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://lore.kernel.org/r/20240131-upstream-net-20240131-mptcp-ci-issues-v1-5-4c1c11e571ff@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the MPTCP PM detects that a subflow is stale, all the packet
scheduler must re-inject all the mptcp-level unacked data. To avoid
acquiring unneeded locks, it first try to check if any unacked data
is present at all in the RTX queue, but such check is currently
broken, as it uses TCP-specific helper on an MPTCP socket.
Funnily enough fuzzers and static checkers are happy, as the accessed
memory still belongs to the mptcp_sock struct, and even from a
functional perspective the recovery completed successfully, as
the short-cut test always failed.
A recent unrelated TCP change - commit d5fed5addb2b ("tcp: reorganize
tcp_sock fast path variables") - exposed the issue, as the tcp field
reorganization makes the mptcp code always skip the re-inection.
Fix the issue dropping the bogus call: we are on a slow path, the early
optimization proved once again to be evil.
Fixes: 1e1d9d6f119c ("mptcp: handle pending data on closed subflow") Cc: stable@vger.kernel.org Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/468 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://lore.kernel.org/r/20240131-upstream-net-20240131-mptcp-ci-issues-v1-1-4c1c11e571ff@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
lld is now able to build ARMv4 and ARMv4T kernels, which means it can
generate thunks for those (__ARMv4PILongThunk_*, __ARMv4PILongBXThunk_*)
that can interfere with kallsyms table generation since they do not get
ignore like the corresponding ARMv5+ ones are:
Inconsistent kallsyms data
Try "make KALLSYMS_EXTRA_PASS=1" as a workaround
Replace the hardcoded list of thunk symbols with a more general regex that
covers this one along with future symbols that follow the same pattern.
get_line() does not trim the leading spaces, but the
parse_source_files() expects to get lines with source files paths where
the first space occurs after the file path.
Fixes: 70f30cfe5b89 ("modpost: use read_text_file() and get_line() for reading text files") Signed-off-by: Radek Krejci <radek.krejci@oracle.com> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
According to the Intel datasheets, software must reset the block
buffer index twice for block process call transactions: once before
writing the outgoing data to the buffer, and once again before
reading the incoming data from the buffer.
The driver is currently missing the second reset, causing the wrong
portion of the block buffer to be read.
Signed-off-by: Jean Delvare <jdelvare@suse.de> Reported-by: Piotr Zakowski <piotr.zakowski@intel.com> Closes: https://lore.kernel.org/linux-i2c/20240213120553.7b0ab120@endymion.delvare/ Fixes: 315cd67c9453 ("i2c: i801: Add Block Write-Block Read Process Call support") Reviewed-by: Alexander Sverdlin <alexander.sverdlin@gmail.com> Signed-off-by: Andi Shyti <andi.shyti@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
On powerpc, it is possible to compile test both the new apple (arm) and
old pasemi (powerpc) drivers for the i2c hardware at the same time,
which leads to a warning about linking the same object file twice:
scripts/Makefile.build:244: drivers/i2c/busses/Makefile: i2c-pasemi-core.o is added to multiple modules: i2c-apple i2c-pasemi
Rework the driver to have an explicit helper module, letting Kbuild
take care of whether this should be built-in or a loadable driver.
Fixes: 9bc5f4f660ff ("i2c: pasemi: Split pci driver to its own file") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Sven Peter <sven@svenpeter.dev> Signed-off-by: Andi Shyti <andi.shyti@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The function spapr_tce_platform_iommu_attach_dev() is missing to call
iommu_group_put() when the domain is already set. This refcount leak
shows up with BUG_ON() during DLPAR remove operation as:
KernelBug: Kernel bug in state 'None': kernel BUG at arch/powerpc/platforms/pseries/iommu.c:100!
Oops: Exception in kernel mode, sig: 5 [#1]
LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=8192 NUMA pSeries
<snip>
Hardware name: IBM,9080-HEX POWER10 (raw) 0x800200 0xf000006 of:IBM,FW1060.00 (NH1060_016) hv:phyp pSeries
NIP: c0000000000ff4d4 LR: c0000000000ff4cc CTR: 0000000000000000
REGS: c0000013aed5f840 TRAP: 0700 Tainted: G I (6.8.0-rc3-autotest-g99bd3cb0d12e)
MSR: 8000000000029033 <SF,EE,ME,IR,DR,RI,LE> CR: 44002402 XER: 20040000
CFAR: c000000000a0d170 IRQMASK: 0
...
NIP iommu_reconfig_notifier+0x94/0x200
LR iommu_reconfig_notifier+0x8c/0x200
Call Trace:
iommu_reconfig_notifier+0x8c/0x200 (unreliable)
notifier_call_chain+0xb8/0x19c
blocking_notifier_call_chain+0x64/0x98
of_reconfig_notify+0x44/0xdc
of_detach_node+0x78/0xb0
ofdt_write.part.0+0x86c/0xbb8
proc_reg_write+0xf4/0x150
vfs_write+0xf8/0x488
ksys_write+0x84/0x140
system_call_exception+0x138/0x330
system_call_vectored_common+0x15c/0x2ec
The patch adds the missing iommu_group_put() call.
Fixes: a8ca9fc9134c ("powerpc/iommu: Do not do platform domain attach atctions after probe") Reported-by: Venkat Rao Bagalkote <venkat88@linux.vnet.ibm.com> Closes: https://lore.kernel.org/all/274e0d2b-b5cc-475e-94e6-8427e88e271d@linux.vnet.ibm.com/ Signed-off-by: Shivaprasad G Bhat <sbhat@linux.ibm.com> Tested-by: Venkat Rao Bagalkote <venkat88@linux.vnet.ibm.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/170784021983.6249.10039296655906636112.stgit@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
KASAN is seen to increase stack usage, to the point that it was reported
to lead to stack overflow on some 32-bit machines (see link).
To avoid overflows the stack size was doubled for KASAN builds in
commit 3e8635fb2e07 ("powerpc/kasan: Force thread size increase with
KASAN").
However with a 32KB stack size to begin with, the doubling leads to a
64KB stack, which causes build errors:
arch/powerpc/kernel/switch.S:249: Error: operand out of range (0x000000000000fe50 is not between 0xffffffffffff8000 and 0x0000000000007fff)
Although the asm could be reworked, in practice a 32KB stack seems
sufficient even for KASAN builds - the additional usage seems to be in
the 2-3KB range for a 64-bit KASAN build.
So only increase the stack for KASAN if the stack size is < 32KB.
Although the GICv3 code base has gained some handling of systems failing to
handle the shareability attributes, the GICv4 side of things has been
firmly ignored.
This is unfortunate, as the new recent addition of the "dma-noncoherent" is
supposed to apply to all of the GICR tables, and not just the ones that are
common to v3 and v4.
Add some checks to handle the VPROPBASE/VPENDBASE shareability and
cacheability attributes in the same way we deal with the other GICR_BASE
registers, wrapping the flag check in a helper for improved readability.
Note that this has been found by inspection only, as I don't have access to
HW that suffers from this particular issue.
Fixes: 3a0fff0fb6a3 ("irqchip/gic-v3: Enable non-coherent redistributors/ITSes DT probing") Signed-off-by: Marc Zyngier <maz@kernel.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Lorenzo Pieralisi <lpieralisi@kernel.org> Link: https://lore.kernel.org/r/20240213101206.2137483-2-maz@kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org>
eiointc_domain_alloc() uses struct eiointc, which is not defined, for a
pointer. Older compilers treat that as a forward declaration and due to
assignment of a void pointer there is no warning emitted. As the variable
is then handed in as a void pointer argument to irq_domain_set_info() the
code is functional.
For i2c read operation in GSI mode, we are getting timeout
due to malformed TRE basically incorrect TRE sequence
in gpi(drivers/dma/qcom/gpi.c) driver.
I2C driver has geni_i2c_gpi(I2C_WRITE) function which generates GO TRE and
geni_i2c_gpi(I2C_READ)generates DMA TRE. Hence to generate GO TRE before
DMA TRE, we should move geni_i2c_gpi(I2C_WRITE) before
geni_i2c_gpi(I2C_READ) inside the I2C GSI mode transfer function
i.e. geni_i2c_gpi_xfer().
TRE stands for Transfer Ring Element - which is basically an element with
size of 4 words. It contains all information like slave address,
clk divider, dma address value data size etc).
Mainly we have 3 TREs(Config, GO and DMA tre).
- CONFIG TRE : consists of internal register configuration which is
required before start of the transfer.
- DMA TRE : contains DDR/Memory address, called as DMA descriptor.
- GO TRE : contains Transfer directions, slave ID, Delay flags, Length
of the transfer.
I2c driver calls GPI driver API to config each TRE depending on the
protocol.
For read operation tre sequence will be as below which is not aligned
to hardware programming guide.
- CONFIG tre
- DMA tre
- GO tre
As per Qualcomm's internal Hardware Programming Guide, we should configure
TREs in below sequence for any RX only transfer.
In this loop, we step through the buffer and after each item we check
if the size_left is greater than the minimum size we need. However,
the problem is that "bytes_left" is type ssize_t while sizeof() is type
size_t. That means that because of type promotion, the comparison is
done as an unsigned and if we have negative bytes left the loop
continues instead of ending.
Fixes: fe856be475f7 ("CIFS: parse and store info on iface queries") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit noted in fixes added a bogus requirement that runtime PM managed
devices need to be in the RPM_ACTIVE state for PME polling. In fact, only
devices in low power states should be polled.
However there's still a requirement that the device config space must be
accessible, which has implications for both the current state of the polled
device and the parent bridge, when present. It's not sufficient to assume
the bridge remains in D0 and cases have been observed where the bridge
passes the D0 test, but the PM state indicates RPM_SUSPENDING and config
space of the polled device becomes inaccessible during pci_pme_wakeup().
Therefore, since the bridge is already effectively required to be in the
RPM_ACTIVE state, formalize this in the code and elevate the PM usage count
to maintain the state while polling the subordinate device.
This resolves a regression reported in the bugzilla below where a
Thunderbolt/USB4 hierarchy fails to scan for an attached NVMe endpoint
downstream of a bridge in a D3hot power state.
Link: https://lore.kernel.org/r/20240123185548.1040096-1-alex.williamson@redhat.com Fixes: d3fcd7360338 ("PCI: Fix runtime PM race with PME polling") Reported-by: Sanath S <sanath.s@amd.com> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218360 Signed-off-by: Alex Williamson <alex.williamson@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Tested-by: Sanath S <sanath.s@amd.com> Reviewed-by: Rafael J. Wysocki <rafael@kernel.org> Cc: Lukas Wunner <lukas@wunner.de> Cc: Mika Westerberg <mika.westerberg@linux.intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In kasan_init_region, when k_start is not page aligned, at the begin of
for loop, k_cur = k_start & PAGE_MASK is less than k_start, and then
`va = block + k_cur - k_start` is less than block, the addr va is invalid,
because the memory address space from va to block is not alloced by
memblock_alloc, which will not be reserved by memblock_reserve later, it
will be used by other places.
As a result, memory overwriting occurs.
for example:
int __init __weak kasan_init_region(void *start, size_t size)
{
[...]
/* if say block(dcd97000) k_start(feef7400) k_end(feeff3fe) */
block = memblock_alloc(k_end - k_start, PAGE_SIZE);
[...]
for (k_cur = k_start & PAGE_MASK; k_cur < k_end; k_cur += PAGE_SIZE) {
/* at the begin of for loop
* block(dcd97000) va(dcd96c00) k_cur(feef7000) k_start(feef7400)
* va(dcd96c00) is less than block(dcd97000), va is invalid
*/
void *va = block + k_cur - k_start;
[...]
}
[...]
}
Therefore, page alignment is performed on k_start before
memblock_alloc() to ensure the validity of the VA address.
Fixes: 663c0c9496a6 ("powerpc/kasan: Fix shadow area set up for modules.") Signed-off-by: Jiangfeng Xiao <xiaojiangfeng@huawei.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/1705974359-43790-1-git-send-email-xiaojiangfeng@huawei.com Signed-off-by: Sasha Levin <sashal@kernel.org>
MMU_FTR_USE_HIGH_BATS is set for G2_LE cores and derivatives like e300cX,
but the high BATs need to be enabled in HID2 to work. Add register
definitions and add the needed setup to __setup_cpu_603.
This fixes boot on CPUs like the MPC5200B with STRICT_KERNEL_RWX enabled
on systems where the flag has not been set by the bootloader already.
Commit a940904443e4 ("powerpc/iommu: Add iommu_ops to report capabilities
and allow blocking domains") broke DLPAR add of PCI devices.
The above added iommu_device structure to pci_controller. During
system boot, PCI devices are discovered and this newly added iommu_device
structure is initialized by a call to iommu_device_register().
During DLPAR add of a PCI device, a new pci_controller structure is
allocated but there are no calls made to iommu_device_register()
interface.
Fix is to register the iommu device during DLPAR add as well.
Fixes: a940904443e4 ("powerpc/iommu: Add iommu_ops to report capabilities and allow blocking domains") Signed-off-by: Gaurav Batra <gbatra@linux.ibm.com>
[mpe: Trim oops and tweak some change log wording] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240122222407.39603-1-gbatra@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
fw_devlink can detect most overlapping/intersecting cycles. However it was
missing a few corner cases because of an incorrect optimization logic that
tries to avoid repeating cycle detection for devices that are already
marked as part of a cycle.
Here's an example provided by Xu Yang (edited for clarity):
Node A (tcpc) will be populated as device 1-0050.
Node B (usb) will be populated as device 38100000.usb.
Node C (usb-phy) will be populated as device 381f0040.usb-phy.
The description below uses the notation:
consumer --> supplier
child ==> parent
1. Node C is populated as device C. No cycles detected because cycle
detection is only run when a fwnode link is converted to a device link.
2. Node B is populated as device B. As we convert B --> C into a device
link we run cycle detection and find and mark the device link/fwnode
link cycle:
C--> A --> B.EP ==> B --> C
3. Node A is populated as device A. As we convert C --> A into a device
link, we see it's already part of a cycle (from step 2) and don't run
cycle detection. Thus we miss detecting the cycle:
A --> B.EP ==> B --> A.EP ==> A
Looking at it another way, A depends on B in one way:
A --> B.EP ==> B
But B depends on A in two ways and we only detect the first:
B --> C --> A
B --> A.EP ==> A
To detect both of these, we remove the incorrect optimization attempt in
step 3 and run cycle detection even if the fwnode link from which the
device link is being created has already been marked as part of a cycle.
Reported-by: Xu Yang <xu.yang_2@nxp.com> Closes: https://lore.kernel.org/lkml/DU2PR04MB8822693748725F85DC0CB86C8C792@DU2PR04MB8822.eurprd04.prod.outlook.com/ Fixes: 3fb16866b51d ("driver core: fw_devlink: Make cycle detection more robust") Signed-off-by: Saravana Kannan <saravanak@google.com> Tested-by: Xu Yang <xu.yang_2@nxp.com> Link: https://lore.kernel.org/r/20240202095636.868578-3-saravanak@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The CO0 BCM needs to be up at all times, otherwise some hardware (like
the UFS controller) loses its connection to the rest of the SoC,
resulting in a hang of the platform, accompanied by a spectacular
logspam.
In current scenario if Plug-out and Plug-In performed continuously
there could be a chance while checking for dwc->gadget_driver in
dwc3_gadget_suspend, a NULL pointer dereference may occur.
CPU1 basically clears the variable and CPU2 checks the variable.
Consider CPU1 is running and right before gadget_driver is cleared
and in parallel CPU2 executes dwc3_gadget_suspend where it finds
dwc->gadget_driver which is not NULL and resumes execution and then
CPU1 completes execution. CPU2 executes dwc3_disconnect_gadget where
it checks dwc->gadget_driver is already NULL because of which the
NULL pointer deference occur.
Currently, the function update_port_device_state gets the usb_hub from
udev->parent by calling usb_hub_to_struct_hub.
However, in case the actconfig or the maxchild is 0, the usb_hub would
be NULL and upon further accessing to get port_dev would result in null
pointer dereference.
Fix this by introducing an if check after the usb_hub is populated.
Fixes: 83cb2604f641 ("usb: core: add sysfs entry for usb device state") Cc: stable@vger.kernel.org Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com> Reviewed-by: Alan Stern <stern@rowland.harvard.edu> Link: https://lore.kernel.org/r/20240110095814.7626-1-quic_ugoswami@quicinc.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When power is recycled in usb controller during system power management,
the controller will recognize it and switch role if role has been changed
during power lost. In current design, it will be completed in resume()
function. However, this may bring issues since usb class devices have
their pm operations too and these device's resume() functions are still
not being called at this point. When usb controller recognized host role
should be stopped, these usb class devices will be removed at this point.
But these usb class devices can't be removed in some cases, such as scsi
devices. Since scsi driver may sync data to U-disk, however it will block
there because scsi drvier can only handle pm request when is in suspended
state. Therefore, there may exist a dependency between ci_resume() and usb
class device's resume(). To break this potential dependency, we need to
handle power lost work in a workqueue.
Fixes: 74494b33211d ("usb: chipidea: core: add controller resume support when controller is powered off")
cc: stable@vger.kernel.org Signed-off-by: Xu Yang <xu.yang_2@nxp.com> Link: https://lore.kernel.org/r/20240119123537.3614838-1-xu.yang_2@nxp.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When write UDC to empty and unbind gadget driver from gadget device, it is
possible that there are many queue failures for mass storage function.
The root cause is mass storage main thread alaways try to queue request to
receive a command from host if running flag is on, on platform like dwc3,
if pull down called, it will not queue request again and return
-ESHUTDOWN, but it not affect running flag of mass storage function.
Check return code from mass storage function and clear running flag if it
is -ESHUTDOWN, also indicate start in/out transfer failure to break loops.
projects / thirdparty / kernel / stable.git / log
