In the previous commit, I renamed the variable to differentiate
mac80211/mvm link STA, but forgot to adjust the check. The one
from mac80211 is already non-NULL anyway, but the mvm one can
be NULL when the mac80211 isn't during link switch conditions.
Fix the check.
Fixes: 2783ab506eaa ("wifi: iwlwifi: mvm: select STA mask only for active links") Reviewed-by: Daniel Gabay <daniel.gabay@intel.com> Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240325180850.e95b442bafe9.I8c0119fce7b00cb4f65782930d2c167ed5dd0a6e@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since the HW restart flow with multi-link is very similar to
the initial association, we do need to reconfigure TLC there.
Remove the check that prevented that.
Fixes: d2d0468f60cd ("wifi: iwlwifi: mvm: configure TLC on link activation") Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240320232419.a00adcfe381a.Ic798beccbb7b7d852dc976d539205353588853b0@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
During reconfig, we might send keys, but those should be only
sent to already active link stations. Iterate only active ones
to fix that issue.
Fixes: aea99650f731 ("wifi: iwlwifi: mvm: set STA mask for keys in MLO") Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240320232419.c6818d1c6033.I6357f05c55ef111002ddc169287eb356ca0c1b21@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
During recovery, the chanctx_conf in mac80211 is still non-NULL even
though the channel context has not yet been assigned again. In that
case, the real count is actually lower.
Switch to instead count the phy_ctx assignment and ensure that the
assignment is cleared at the start of recovery.
Fixes: 12bacfc2c065 ("wifi: iwlwifi: handle eSR transitions") Signed-off-by: Benjamin Berg <benjamin.berg@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240320232419.55f37339e7d1.I57006568a90ffb7a1232def1b2f3264dea711ba6@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
If scan request doesn't include a link ID to be used for TSF
reporting, don't select it as it might become inactive before
scan is actually started by the driver.
Instead, let the driver select one of the active links.
Fixes: cbde0b49f276 ("wifi: mac80211: Extend support for scanning while MLO connected") Signed-off-by: Ayala Beker <ayala.beker@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240320091155.a6b643a15755.Ic28ed9a611432387b7f85e9ca9a97a4ce34a6e0f@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
For the mvm driver, data structures match what's in the firmware,
we allocate FW IDs for them already etc. During link switch we
already allocate/free the STA links appropriately, but initially
we'd allocate them always. Fix this to allocate memory, a STA ID,
etc. only for active links.
If there was a possibility of an MLE basic STA profile without
subelements, we might reject it because we account for the one
octet for sta_info_len twice (it's part of itself, and in the
fixed portion). Like in ieee80211_mle_reconf_sta_prof_size_ok,
subtract 1 to adjust that.
When reading the elements we did take this into account, and
since there are always elements, this never really mattered.
Fixes: 7b6f08771bf6 ("wifi: ieee80211: Support validating ML station profile length") Signed-off-by: Johannes Berg <johannes.berg@intel.com> Reviewed-by: Ilan Peer <ilan.peer@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://msgid.link/20240318184907.00bb0b20ed60.I8c41dd6fc14c4b187ab901dea15ade73c79fb98c@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
aaa8736370db ("x86, relocs: Ignore relocations in .notes section")
... only started ignoring the .notes sections in print_absolute_relocs(),
but the same logic should also by applied in walk_relocs() to avoid
such relocations.
[ mingo: Fixed various typos in the changelog, removed extra curly braces from the code. ]
drivers/net/wireless/mediatek/mt76/mt7915/debugfs.c:1133:29: error: too long token expansion
drivers/net/wireless/mediatek/mt76/mt7915/debugfs.c:1133:29: error: too long token expansion
drivers/net/wireless/mediatek/mt76/mt7915/debugfs.c:1133:29: error: too long token expansion
drivers/net/wireless/mediatek/mt76/mt7915/debugfs.c:1133:29: error: too long token expansion
Due to an error during rebasing the patchset 320 MHz channel support got
broken. ath12k was setting the QoS bit instead of the correct flag.
WMI_PEER_EXT_320MHZ (0x2) is defined as an extended flag, replace
peer_flags by peer_flags_ext while sending peer data.
This affected both QCN9274 and WCN7850 which use the same flag.
Fixes: 6734cf9b4cc7 ("wifi: ath12k: peer assoc for 320 MHz") Signed-off-by: Aloka Dixit <quic_alokad@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20240314204651.11075-1-quic_alokad@quicinc.com Signed-off-by: Sasha Levin <sashal@kernel.org>
In bpf_objec_load_prog(), there's no guarantee that obj->btf is non-NULL
when passing it to btf__fd(), and this function does not perform any
check before dereferencing its argument (as bpf_object__btf_fd() used to
do). As a consequence, we get segmentation fault errors in bpftool (for
example) when trying to load programs that come without BTF information.
v2: Keep btf__fd() in the fix instead of reverting to bpf_object__btf_fd().
Fixes: df7c3f7d3a3d ("libbpf: make uniform use of btf__fd() accessor inside libbpf") Suggested-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Quentin Monnet <qmo@kernel.org> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20240314150438.232462-1-qmo@kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Current 'bpftool link' command does not show pids, e.g.,
$ tools/build/bpftool/bpftool link
...
4: tracing prog 23
prog_type lsm attach_type lsm_mac
target_obj_id 1 target_btf_id 31320
Hack the following change to enable normal libbpf debug output,
# --- a/tools/bpf/bpftool/pids.c
# +++ b/tools/bpf/bpftool/pids.c
# @@ -121,9 +121,9 @@ int build_obj_refs_table(struct hashmap **map, enum bpf_obj_type type)
# /* we don't want output polluted with libbpf errors if bpf_iter is not
# * supported
# */
# - default_print = libbpf_set_print(libbpf_print_none);
# + /* default_print = libbpf_set_print(libbpf_print_none); */
# err = pid_iter_bpf__load(skel);
# - libbpf_set_print(default_print);
# + /* libbpf_set_print(default_print); */
The 'file->private_data' returns a 'void' type and this caused subsequent 'link->type'
(insn #81) failed in verification.
To fix the issue, restore the previous BPF_CORE_READ so old kernels can also work.
With this patch, the 'bpftool link' runs successfully with 'pids'.
$ tools/build/bpftool/bpftool link
...
4: tracing prog 23
prog_type lsm attach_type lsm_mac
target_obj_id 1 target_btf_id 31320
pids systemd(1)
Fixes: 44ba7b30e84f ("bpftool: Use a local copy of BPF_LINK_TYPE_PERF_EVENT in pid_iter.bpf.c") Signed-off-by: Yonghong Song <yonghong.song@linux.dev> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Tested-by: Quentin Monnet <quentin@isovalent.com> Reviewed-by: Quentin Monnet <quentin@isovalent.com> Link: https://lore.kernel.org/bpf/20240312023249.3776718-1-yonghong.song@linux.dev Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently we force enable power save on non-running vdevs, this results
in unexpected ping latency in below scenarios:
1. disable power save from userspace.
2. trigger suspend/resume.
With step 1 power save is disabled successfully and we get a good latency:
PING 192.168.1.1 (192.168.1.1) 56(84) bytes of data.
64 bytes from 192.168.1.1: icmp_seq=1 ttl=64 time=5.13 ms
64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=5.45 ms
64 bytes from 192.168.1.1: icmp_seq=3 ttl=64 time=5.99 ms
64 bytes from 192.168.1.1: icmp_seq=4 ttl=64 time=6.34 ms
64 bytes from 192.168.1.1: icmp_seq=5 ttl=64 time=4.47 ms
64 bytes from 192.168.1.1: icmp_seq=6 ttl=64 time=6.45 ms
While after step 2, the latency becomes much larger:
PING 192.168.1.1 (192.168.1.1) 56(84) bytes of data.
64 bytes from 192.168.1.1: icmp_seq=1 ttl=64 time=17.7 ms
64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=15.0 ms
64 bytes from 192.168.1.1: icmp_seq=3 ttl=64 time=14.3 ms
64 bytes from 192.168.1.1: icmp_seq=4 ttl=64 time=16.5 ms
64 bytes from 192.168.1.1: icmp_seq=5 ttl=64 time=20.1 ms
The reason is, with step 2, power save is force enabled due to vdev not
running, although mac80211 was trying to disable it to honor userspace
configuration:
ath11k_pci 0000:03:00.0: wmi cmd sta powersave mode psmode 1 vdev id 0
Call Trace:
ath11k_wmi_pdev_set_ps_mode
ath11k_mac_op_bss_info_changed
ieee80211_bss_info_change_notify
ieee80211_reconfig
ieee80211_resume
wiphy_resume
This logic is taken from ath10k where it was added due to below comment:
Firmware doesn't behave nicely and consumes more power than
necessary if PS is disabled on a non-started vdev.
However we don't know whether such an issue also occurs to ath11k firmware
or not. But even if it does, it's not appropriate because it goes against
userspace, even cfg/mac80211 don't know we have enabled it in fact.
Remove it to fix this issue. In this way we not only get a better latency,
but also, and the most important, keeps the consistency between userspace
and kernel/driver. The biggest price for that would be the power consumption,
which is not that important, compared with the consistency.
Fixes: b2beffa7d9a6 ("ath11k: enable 802.11 power save mode in station mode") Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20240309113115.11498-1-quic_bqiang@quicinc.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The kzalloc() in brcmf_pcie_download_fw_nvram() will return null
if the physical memory has run out. As a result, if we use
get_random_bytes() to generate random bytes in the randbuf, the
null pointer dereference bug will happen.
In order to prevent allocation failure, this patch adds a separate
function using buffer on kernel stack to generate random bytes in
the randbuf, which could prevent the kernel stack from overflow.
Fixes: 91918ce88d9f ("wifi: brcmfmac: pcie: Provide a buffer of random bytes to the device") Suggested-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Duoming Zhou <duoming@zju.edu.cn> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://msgid.link/20240306140437.18177-1-duoming@zju.edu.cn Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently host relies on CE interrupts to get notified that
the service ready message is ready. This results in timeout
issue if the interrupt is not fired, due to some unknown
reasons. See below logs:
[76321.937866] ath10k_pci 0000:02:00.0: wmi service ready event not received
...
[76322.016738] ath10k_pci 0000:02:00.0: Could not init core: -110
And finally it causes WLAN interface bring up failure.
Change to give it one more chance here by polling CE rings,
before failing directly.
Currently, io_ticks is accounted based on sampling, specifically
update_io_ticks() will always account io_ticks by 1 jiffies from
bdev_start_io_acct()/blk_account_io_start(), and the result can be
inaccurate, for example(HZ is 250):
Test script:
fio -filename=/dev/sda -bs=4k -rw=write -direct=1 -name=test -thinktime=4ms
Test result: util is about 90%, while the disk is really idle.
This behaviour is introduced by commit 5b18b5a73760 ("block: delete
part_round_stats and switch to less precise counting"), however, there
was a key point that is missed that this patch also improve performance
a lot:
Before the commit:
part_round_stats:
if (part->stamp != now)
stats |= 1;
part_in_flight()
-> there can be lots of task here in 1 jiffies.
part_round_stats_single()
__part_stat_add()
part->stamp = now;
After the commit:
update_io_ticks:
stamp = part->bd_stamp;
if (time_after(now, stamp))
if (try_cmpxchg())
__part_stat_add()
-> only one task can reach here in 1 jiffies.
Hence in order to account io_ticks precisely, we only need to know if
there are IO inflight at most once in one jiffies. Noted that for
rq-based device, iterating tags should not be used here because
'tags->lock' is grabbed in blk_mq_find_and_get_req(), hence
part_stat_lock_inc/dec() and part_in_flight() is used to trace inflight.
The additional overhead is quite little:
- per cpu add/dec for each IO for rq-based device;
- per cpu sum for each jiffies;
And it's verified by null-blk that there are no performance degration
under heavy IO pressure.
Fixes: 5b18b5a73760 ("block: delete part_round_stats and switch to less precise counting") Signed-off-by: Yu Kuai <yukuai3@huawei.com> Link: https://lore.kernel.org/r/20240509123717.3223892-2-yukuai1@huaweicloud.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix the cmdline parsing of the "blkdevparts=" parameter using strsep(),
which makes the code simpler.
Before commit 146afeb235cc ("block: use strscpy() to instead of
strncpy()"), we used a strncpy() to copy a block device name and partition
names. The commit simply replaced a strncpy() and NULL termination with
a strscpy(). It did not update calculations of length passed to strscpy().
While the length passed to strncpy() is just a length of valid characters
without NULL termination ('\0'), strscpy() takes it as a length of the
destination buffer, including a NULL termination.
Since the source buffer is not necessarily NULL terminated, the current
code copies "length - 1" characters and puts a NULL character in the
destination buffer. It replaces the last character with NULL and breaks
the parsing.
As an example, that buffer will be passed to parse_parts() and breaks
parsing sub-partitions due to the missing ')' at the end, like the
following.
1. "48M@10M(kernel-1),..." is passed to strscpy() with length=17
from parse_parts()
2. strscpy() returns -E2BIG and the destination buffer has
"48M@10M(kernel-1\0"
3. "48M@10M(kernel-1\0" is passed to parse_subpart()
4. parse_subpart() fails to find ')' when parsing a partition name,
and returns error
By the way, strscpy() takes a length of destination buffer and it is
often confusing when copying characters with a specified length. Using
strsep() helps to separate the string by the specified character. Then,
we can use strscpy() naturally with the size of the destination buffer.
Separating the string on the fly is also useful to omit the redundant
string copy, reducing memory usage and improve the code readability.
Fixes: 146afeb235cc ("block: use strscpy() to instead of strncpy()") Suggested-by: Naohiro Aota <naota@elisp.net> Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Link: https://lore.kernel.org/r/20240421074005.565-1-musashino.open@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
blkdev_iomap_begin rounds down the offset to the logical block size
before stashing it in iomap->offset and checking that it still is
inside the inode size.
Check the i_size check to the raw pos value so that we don't try a
zero size write if iter->pos is unaligned.
Fixes: 487c607df790 ("block: use iomap for writes to block devices") Reported-by: syzbot+0a3683a0a6fecf909244@syzkaller.appspotmail.com Signed-off-by: Christoph Hellwig <hch@lst.de> Tested-by: syzbot+0a3683a0a6fecf909244@syzkaller.appspotmail.com Link: https://lore.kernel.org/r/20240503081042.2078062-1-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
The 4xxx driver can probe 4xxx and 402xx devices. However, the driver
only specifies the firmware images required for 4xxx.
This might result in external tools missing these binaries, if required,
in the initramfs.
Specify the firmware image used by 402xx with the MODULE_FIRMWARE()
macros in the 4xxx driver.
Fixes: a3e8c919b993 ("crypto: qat - add support for 402xx devices") Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Reviewed-by: Damian Muszynski <damian.muszynski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
md_do_sync
j = mddev->resync_min
while (j < max_sectors)
sectors = raid10_sync_request(mddev, j, &skipped)
if (!md_bitmap_start_sync(..., &sync_blocks))
// md_bitmap_start_sync set sync_blocks to 0
return sync_blocks + sectors_skippe;
// sectors = 0;
j += sectors;
// j never change
Root cause is that commit 301867b1c168 ("md/raid10: check
slab-out-of-bounds in md_bitmap_get_counter") return early from
md_bitmap_get_counter(), without setting returned blocks.
Fix this problem by always set returned blocks from
md_bitmap_get_counter"(), as it used to be.
Noted that this patch just fix the softlockup problem in kernel, the
case that bitmap size doesn't match array size still need to be fixed.
Fixes: 301867b1c168 ("md/raid10: check slab-out-of-bounds in md_bitmap_get_counter") Reported-and-tested-by: Nigel Croxon <ncroxon@redhat.com> Closes: https://lore.kernel.org/all/71ba5272-ab07-43ba-8232-d2da642acb4e@redhat.com/ Signed-off-by: Yu Kuai <yukuai3@huawei.com> Link: https://lore.kernel.org/r/20240422065824.2516-1-yukuai1@huaweicloud.com Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The EXEC_RODATA test plays a lot of tricks to live in the .rodata section,
and once again ran into objtool's (completely reasonable) assumptions
that executable code should live in an executable section. However, this
manifested only under CONFIG_CFI_CLANG=y, as one of the .cfi_sites was
pointing into the .rodata section.
Since we're testing non-CFI execution properties in perms.c (and
rodata.c), we can disable CFI for the involved functions, and remove the
CFI arguments from rodata.c entirely.
The recently introduced commit '635ce0db8956 ("soc: qcom: pmic_glink:
don't traverse clients list without a lock")' ensured that the clients
list is not modified while traversed.
But the callback is made from the GLINK IRQ handler and as such this
mutual exclusion can not be provided by a (sleepable) mutex.
Replace the mutex with a spinlock.
Fixes: 635ce0db8956 ("soc: qcom: pmic_glink: don't traverse clients list without a lock") Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20240430-pmic-glink-sleep-while-atomic-v1-1-88fb493e8545@quicinc.com Signed-off-by: Bjorn Andersson <andersson@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
SEND[MSG]_ZC produces multiple CQEs via notifications, LAZY_WAKE doesn't
handle it and so disable LAZY_WAKE for sendzc polling. It should be
fine, sends are not likely to be polled in the first place.
Ensure that prep handlers always initialize sr->done_io before any
potential failure conditions, and with that, we now it's always been
set even for the failure case.
With that, we don't need to use the REQ_F_PARTIAL_IO flag to gate on that.
Additionally, we should not overwrite req->cqe.res unless sr->done_io is
actually positive.
The clock management in this driver does not seem to be correct. The
struct hwrng .init callback enables the clock, but there is no matching
.cleanup callback to disable the clock. The clock get disabled as some
later point by runtime PM suspend callback.
Furthermore, both runtime PM and sleep suspend callbacks access registers
first and disable clock which are used for register access second. If the
IP is already in RPM suspend and the system enters sleep state, the sleep
callback will attempt to access registers while the register clock are
already disabled. This bug has been fixed once before already in commit 9bae54942b13 ("hwrng: stm32 - fix pm_suspend issue"), and regressed in
commit ff4e46104f2e ("hwrng: stm32 - rework power management sequences") .
Fix this slightly differently, disable register clock at the end of .init
callback, this way the IP is disabled after .init. On every access to the
IP, which really is only stm32_rng_read(), do pm_runtime_get_sync() which
is already done in stm32_rng_read() to bring the IP from RPM suspend, and
pm_runtime_mark_last_busy()/pm_runtime_put_sync_autosuspend() to put it
back into RPM suspend.
Change sleep suspend/resume callbacks to enable and disable register clock
around register access, as those cannot use the RPM suspend/resume callbacks
due to slightly different initialization in those sleep callbacks. This way,
the register access should always be performed with clock surely enabled.
Fixes: ff4e46104f2e ("hwrng: stm32 - rework power management sequences") Signed-off-by: Marek Vasut <marex@denx.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
In case of an irrecoverable failure, put the IP into RPM suspend
to avoid RPM imbalance. I did not trigger this case, but it seems
it should be done based on reading the code.
Fixes: b17bc6eb7c2b ("hwrng: stm32 - rework error handling in stm32_rng_read()") Signed-off-by: Marek Vasut <marex@denx.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The conditional is used to check whether err is non-zero OR whether
reg variable is non-zero after clearing bits from it. This should be
done using logical OR, not bitwise OR, fix it.
Fixes: 6b85a7e141cb ("hwrng: stm32 - implement STM32MP13x support") Signed-off-by: Marek Vasut <marex@denx.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The function adf_send_admin_tl_start() enables the telemetry (TL)
feature on a QAT device by sending the ICP_QAT_FW_TL_START message to
the firmware. This triggers the FW to start writing TL data to a DMA
buffer in memory and returns an array containing the number of
accelerators of each type (slices) supported by this HW.
The pointer to this array is stored in the adf_tl_hw_data data
structure called slice_cnt.
The array slice_cnt is then used in the function tl_print_dev_data()
to report in debugfs only statistics about the supported accelerators.
An incorrect value of the elements in slice_cnt might lead to an out
of bounds memory read.
At the moment, there isn't an implementation of FW that returns a wrong
value, but for robustness validate the slice count array returned by FW.
Fixes: 69e7649f7cc2 ("crypto: qat - add support for device telemetry") Signed-off-by: Lucas Segarra Fernandez <lucas.segarra.fernandez@intel.com> Reviewed-by: Damian Muszynski <damian.muszynski@intel.com> Reviewed-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
For cmdq jump command, offset 0 means relative jump and offset 1
means absolute jump. cmdq_pkt_jump() is absolute jump, so fix the
typo of CMDQ_JUMP_RELATIVE in cmdq_pkt_jump().
Fixes: 946f1792d3d7 ("soc: mediatek: cmdq: add jump function") Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Link: https://lore.kernel.org/r/20240222154120.16959-2-chunkuang.hu@kernel.org Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
It is possible qcom_scm_is_available() gives wrong indication that
if __scm is initialized while __scm->dev is not and similar issue
is also possible with __scm->waitq_comp.
Fix this appropriately by the use of release barrier and read barrier
that will make sure if __scm is initialized so, is all of its field
variable.
In case the client is registered after the pmic-glink recived a response
from the Protection Domain mapper, it is going to miss the notification
about the state. Notify clients about the current state upon
registration.
Take the client_lock before traversing the clients list at the
pmic_glink_state_notify_clients() function. This is required to keep the
list traversal safe from concurrent modification.
commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.
Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.
Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.
This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.
There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.
Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)
PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.
What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.
[ agordeev: Fixed sparse and checkpatch complaints and error handling ]
Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com> Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com> Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs") Signed-off-by: David Hildenbrand <david@redhat.com> Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
s390x must disable shared zeropages for processes running VMs, because
the VMs could end up making use of "storage keys" or protected
virtualization, which are incompatible with shared zeropages.
Yet, with userfaultfd it is possible to insert shared zeropages into
such processes. Let's fallback to simply allocating a fresh zeroed
anonymous folio and insert that instead.
mm_forbids_zeropage() was introduced in commit 593befa6ab74 ("mm: introduce
mm_forbids_zeropage function"), briefly before userfaultfd went
upstream.
Note that we don't want to fail the UFFDIO_ZEROPAGE request like we do
for hugetlb, it would be rather unexpected. Further, we also
cannot really indicated "not supported" to user space ahead of time: it
could be that the MM disallows zeropages after userfaultfd was already
registered.
[ agordeev: Fixed checkpatch complaints ]
Fixes: c1a4de99fada ("userfaultfd: mcopy_atomic|mfill_zeropage: UFFDIO_COPY|UFFDIO_ZEROPAGE preparation") Reviewed-by: Peter Xu <peterx@redhat.com> Link: https://lore.kernel.org/r/20240411161441.910170-2-david@redhat.com Signed-off-by: David Hildenbrand <david@redhat.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 361aee450c6e ("io-wq: add intermediate work step between pending
list and active work") closed a race between a cancellation and the work
being removed from the wq for execution. To ensure the request is
always reachable by the cancellation, we need to move it within the wq
lock, which also synchronizes the cancellation. But commit 42abc95f05bf ("io-wq: decouple work_list protection from the big
wqe->lock") replaced the wq lock here and accidentally reintroduced the
race by releasing the acct_lock too early.
In other words:
worker | cancellation
work = io_get_next_work() |
raw_spin_unlock(&acct->lock); |
|
| io_acct_cancel_pending_work
| io_wq_worker_cancel()
worker->next_work = work
Using acct_lock is still enough since we synchronize on it on
io_acct_cancel_pending_work.
Fixes: 42abc95f05bf ("io-wq: decouple work_list protection from the big wqe->lock") Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de> Link: https://lore.kernel.org/r/20240416021054.3940-2-krisman@suse.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Test robot reports:
> kernel test robot noticed a -19.0% regression of aim9.disk_src.ops_per_sec on:
>
> commit: a2e459555c5f9da3e619b7e47a63f98574dc75f1 ("shmem: stable directory offsets")
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
Feng Tang further clarifies that:
> ... the new simple_offset_add()
> called by shmem_mknod() brings extra cost related with slab,
> specifically the 'radix_tree_node', which cause the regression.
Willy's analysis is that, over time, the test workload causes
xa_alloc_cyclic() to fragment the underlying SLAB cache.
This patch replaces the offset_ctx's xarray with a Maple Tree in the
hope that Maple Tree's dense node mode will handle this scenario
more scalably.
In addition, we can widen the simple directory offset maximum to
signed long (as loff_t is also signed).
Suggested-by: Matthew Wilcox <willy@infradead.org> Reported-by: kernel test robot <oliver.sang@intel.com> Closes: https://lore.kernel.org/oe-lkp/202309081306.3ecb3734-oliver.sang@intel.com Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Link: https://lore.kernel.org/r/170820145616.6328.12620992971699079156.stgit@91.116.238.104.host.secureserver.net Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org>
Stable-dep-of: 23cdd0eed3f1 ("libfs: Fix simple_offset_rename_exchange()") Signed-off-by: Sasha Levin <sashal@kernel.org>
Liam and Matthew say that once the RCU read lock is released,
xa_state is not safe to re-use for the next xas_find() call. But the
RCU read lock must be released on each loop iteration so that
dput(), which might_sleep(), can be called safely.
Thus we are forced to walk the offset tree with fresh state for each
directory entry. xa_find() can do this for us, though it might be a
little less efficient than maintaining xa_state locally.
We believe that in the current code base, inode->i_rwsem provides
protection for the xa_state maintained in
offset_iterate_dir(). However, there is no guarantee that will
continue to be the case in the future.
Since offset_iterate_dir() doesn't build xa_state locally any more,
there's no longer a strong need for offset_find_next(). Clean up by
rolling these two helpers together.
Add a check to make sure that the requested xattr node size is no larger
than the eraseblock minus the cleanmarker.
Unlike the usual inode nodes, the xattr nodes aren't split into parts
and spread across multiple eraseblocks, which means that a xattr node
must not occupy more than one eraseblock. If the requested xattr value is
too large, the xattr node can spill onto the next eraseblock, overwriting
the nodes and causing errors such as:
jffs2: argh. node added in wrong place at 0x0000b050(2)
jffs2: nextblock 0x0000a000, expected at 0000b00c
jffs2: error: (823) do_verify_xattr_datum: node CRC failed at 0x01e050,
read=0xfc892c93, calc=0x000000
jffs2: notice: (823) jffs2_get_inode_nodes: Node header CRC failed
at 0x01e00c. {848f,2fc4,0fef511f,59a3d171}
jffs2: Node at 0x0000000c with length 0x00001044 would run over the
end of the erase block
jffs2: Perhaps the file system was created with the wrong erase size?
jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found
at 0x00000010: 0x1044 instead
This breaks the filesystem and can lead to KASAN crashes such as:
BUG: KASAN: slab-out-of-bounds in jffs2_sum_add_kvec+0x125e/0x15d0
Read of size 4 at addr ffff88802c31e914 by task repro/830
CPU: 0 PID: 830 Comm: repro Not tainted 6.9.0-rc3+ #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS Arch Linux 1.16.3-1-1 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0xc6/0x120
print_report+0xc4/0x620
? __virt_addr_valid+0x308/0x5b0
kasan_report+0xc1/0xf0
? jffs2_sum_add_kvec+0x125e/0x15d0
? jffs2_sum_add_kvec+0x125e/0x15d0
jffs2_sum_add_kvec+0x125e/0x15d0
jffs2_flash_direct_writev+0xa8/0xd0
jffs2_flash_writev+0x9c9/0xef0
? __x64_sys_setxattr+0xc4/0x160
? do_syscall_64+0x69/0x140
? entry_SYSCALL_64_after_hwframe+0x76/0x7e
[...]
Found by Linux Verification Center (linuxtesting.org) with Syzkaller.
Fixes: aa98d7cf59b5 ("[JFFS2][XATTR] XATTR support on JFFS2 (version. 5)") Signed-off-by: Ilya Denisyev <dev@elkcl.ru> Link: https://lore.kernel.org/r/20240412155357.237803-1-dev@elkcl.ru Signed-off-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 4fc8cb47fcfd ("drm/display: Move HDMI helpers into display-helper
module") turned the DRM_DW_HDMI dependency of DRM_SUN8I_DW_HDMI into a
depends on which ended up disabling the driver in the defconfig. Make
sure it's still enabled.
Fixes: 4fc8cb47fcfd ("drm/display: Move HDMI helpers into display-helper module") Reported-by: Mark Brown <broonie@kernel.org> Reported-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Maxime Ripard <mripard@kernel.org> Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com> Link: https://lore.kernel.org/r/20240403-fix-dw-hdmi-kconfig-v1-5-afbc4a835c38@kernel.org Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The rcuc-starvation output from print_cpu_stall_info() might overflow the
buffer if there is a huge difference in jiffies difference. The situation
might seem improbable, but computers sometimes get very confused about
time, which can result in full-sized integers, and, in this case,
buffer overflow.
Also, the unsigned jiffies difference is printed using %ld, which is
normally for signed integers. This is intentional for debugging purposes,
but it is not obvious from the code.
This commit therefore changes sprintf() to snprintf() and adds a
clarifying comment about intention of %ld format.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 245a62982502 ("rcu: Dump rcuc kthread status for CPUs not reporting quiescent state") Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru> Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is a possibility of buffer overflow in
show_rcu_tasks_trace_gp_kthread() if counters, passed
to sprintf() are huge. Counter numbers, needed for this
are unrealistically high, but buffer overflow is still
possible.
Use snprintf() with buffer size instead of sprintf().
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: edf3775f0ad6 ("rcu-tasks: Add count for idle tasks on offline CPUs") Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru> Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
io_task_work_pending() uses wq_list_empty() on ctx->work_llist, but it's
not an io_wq_work_list, it's a struct llist_head. They both have
->first as head-of-list, and it turns out the checks are identical. But
be proper and use the right helper.
Fixes: dac6a0eae793 ("io_uring: ensure iopoll runs local task work as well") Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
The subchannel-type field "st" of s390_cio_stsch and s390_cio_msch
tracepoints is incorrectly filled with the subchannel-enabled SCHIB
value "ena". Fix this by assigning the correct value.
Fixes: d1de8633d96a ("s390 cio: Rewrite trace point class s390_class_schib") Reviewed-by: Heiko Carstens <hca@linux.ibm.com> Signed-off-by: Peter Oberparleiter <oberpar@linux.ibm.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since sha512_transform_rorx() uses ymm registers, execute vzeroupper
before returning from it. This is necessary to avoid reducing the
performance of SSE code.
Fixes: e01d69cb0195 ("crypto: sha512 - Optimized SHA512 x86_64 assembly routine using AVX instructions.") Signed-off-by: Eric Biggers <ebiggers@google.com> Acked-by: Tim Chen <tim.c.chen@linux.intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since sha256_transform_rorx() uses ymm registers, execute vzeroupper
before returning from it. This is necessary to avoid reducing the
performance of SSE code.
Fixes: d34a460092d8 ("crypto: sha256 - Optimized sha256 x86_64 routine using AVX2's RORX instructions") Signed-off-by: Eric Biggers <ebiggers@google.com> Acked-by: Tim Chen <tim.c.chen@linux.intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
This is not a useful configuration, and there is not much point in saving
a few bytes when only one of the two is enabled, so just remove all
these ifdef checks and rely on of_match_node() and acpi_match_device()
returning NULL when these subsystems are disabled.
__cmpxchg_u8() had been added (initially) for the sake of
drivers/phy/ti/phy-tusb1210.c; the thing is, that drivers is
modular, so we need an export
Fixes: b344d6a83d01 "parisc: add support for cmpxchg on u8 pointers" Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
clang-14 points out that v_size is always smaller than a 64KB
page size if that is configured by the CPU architecture:
fs/nilfs2/ioctl.c:63:19: error: result of comparison of constant 65536 with expression of type '__u16' (aka 'unsigned short') is always false [-Werror,-Wtautological-constant-out-of-range-compare]
if (argv->v_size > PAGE_SIZE)
~~~~~~~~~~~~ ^ ~~~~~~~~~
This is ok, so just shut up that warning with a cast.
The 'TAG 66 Packet Format' description is missing the cipher code and
checksum fields that are packed into the message packet. As a result,
the buffer allocated for the packet is 3 bytes too small and
write_tag_66_packet() will write up to 3 bytes past the end of the
buffer.
Fix this by increasing the size of the allocation so the whole packet
will always fit in the buffer.
This fixes the below kasan slab-out-of-bounds bug:
BUG: KASAN: slab-out-of-bounds in ecryptfs_generate_key_packet_set+0x7d6/0xde0
Write of size 1 at addr ffff88800afbb2a5 by task touch/181
The buffer used to transfer data over the mailbox interface is mapped
using the client's device. This is incorrect, as the device performing
the DMA transfer is the mailbox itself. Fix it by using the mailbox
controller device instead.
This requires including the mailbox_controller.h header to dereference
the mbox_chan and mbox_controller structures. The header is not meant to
be included by clients. This could be fixed by extending the client API
with a function to access the controller's device.
Fixes: 4e3d60656a72 ("ARM: bcm2835: Add the Raspberry Pi firmware driver") Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Stefan Wahren <wahrenst@gmx.net> Tested-by: Ivan T. Ivanov <iivanov@suse.de> Link: https://lore.kernel.org/r/20240326195807.15163-3-laurent.pinchart@ideasonboard.com Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In spu2_dump_omd() value of ptr is increased by ciph_key_len
instead of hash_iv_len which could lead to going beyond the
buffer boundaries.
Fix this bug by changing ciph_key_len to hash_iv_len.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
The original mount API conversion inexplicably left out the change
from ->remount_fs to ->reconfigure; do that now.
Fixes: 7ab2fa7693c3 ("vfs: Convert openpromfs to use the new mount API") Signed-off-by: Eric Sandeen <sandeen@redhat.com> Link: https://lore.kernel.org/r/90b968aa-c979-420f-ba37-5acc3391b28f@redhat.com Signed-off-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The nsid value is a u32 that comes from nvmet_req_find_ns(). It's
endian data and we're on an error path and both of those raise red
flags. So let's make this safer.
1) Make the buffer large enough for any u32.
2) Remove the unnecessary initialization.
3) Use snprintf() instead of sprintf() for even more safety.
4) The sprintf() function returns the number of bytes printed, not
counting the NUL terminator. It is impossible for the return value to
be <= 0 so delete that.
Fixes: 505363957fad ("nvmet: fix nvme status code when namespace is disabled") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
epoll can call out to vfs_poll() with a file pointer that may race with
the last 'fput()'. That would make f_count go down to zero, and while
the ep->mtx locking means that the resulting file pointer tear-down will
be blocked until the poll returns, it means that f_count is already
dead, and any use of it won't actually get a reference to the file any
more: it's dead regardless.
Make sure we have a valid ref on the file pointer before we call down to
vfs_poll() from the epoll routines.
TLS requires a strict pdu pacing via MSG_EOR to signal the end
of a record and subsequent encryption. If we do not set MSG_EOR
at the end of a sequence the record won't be closed, encryption
doesn't start, and we end up with a send stall as the message
will never be passed on to the TCP layer.
So do not check for the queue status when TLS is enabled but
rather make the MSG_MORE setting dependent on the current
request only.
Signed-off-by: Hannes Reinecke <hare@kernel.org> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If the user disabled a nvmet namespace, it is removed from the subsystem
namespaces list. When nvmet processes a command directed to an nsid that
was disabled, it cannot differentiate between a nsid that is disabled
vs. a non-existent namespace, and resorts to return NVME_SC_INVALID_NS
with the dnr bit set.
This translates to a non-retryable status for the host, which translates
to a user error. We should expect disabled namespaces to not cause an
I/O error in a multipath environment.
Address this by searching a configfs item for the namespace nvmet failed
to find, and if we found one, conclude that the namespace is disabled
(perhaps temporarily). Return NVME_SC_INTERNAL_PATH_ERROR in this case
and keep DNR bit cleared.
When we teardown the controller, we wait for pending I/Os to complete
(sq->ref on all queues to drop to zero) and then we go over the commands,
and free their command buffers in case they are still fetching data from
the host (e.g. processing nvme writes) and have yet to take a reference
on the sq.
However, we may miss the case where commands have failed before executing
and are queued for sending a response, but will never occur because the
queue socket is already down. In this case we may miss deallocating command
buffers.
Solve this by freeing all commands buffers as nvmet_tcp_free_cmd_buffers is
idempotent anyways.
Reported-by: Yi Zhang <yi.zhang@redhat.com> Tested-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
While I/O is running, if the pci bus error occurs then
in-flight I/O can not complete. Worst, if at this time,
user (logically) hot-unplug the nvme disk then the
nvme_remove() code path can't forward progress until
in-flight I/O is cancelled. So these sequence of events
may potentially hang hot-unplug code path indefinitely.
This patch helps cancel the pending/in-flight I/O from the
nvme request timeout handler in case the nvme controller
is in the terminal (DEAD/DELETING/DELETING_NOIO) state and
that helps nvme_remove() code path forward progress and
finish successfully.
On system where native nvme multipath is configured and iopolicy
is set to numa but the nvme controller numa node id is undefined
or -1 (NUMA_NO_NODE) then avoid calculating node distance for
finding optimal io path. In such case we may access numa distance
table with invalid index and that may potentially refer to incorrect
memory. So this patch ensures that if the nvme controller numa node
id is -1 then instead of calculating node distance for finding optimal
io path, we set the numa node distance of such controller to default 10
(LOCAL_DISTANCE).
... and I think that's actually the important thing here:
- the first page fault is from user space, and triggers the vsyscall
emulation.
- the second page fault is from __do_sys_gettimeofday(), and that should
just have caused the exception that then sets the return value to
-EFAULT
- the third nested page fault is due to _raw_spin_unlock_irqrestore() ->
preempt_schedule() -> trace_sched_switch(), which then causes a BPF
trace program to run, which does that bpf_probe_read_compat(), which
causes that page fault under pagefault_disable().
It's quite the nasty backtrace, and there's a lot going on.
The problem is literally the vsyscall emulation, which sets
current->thread.sig_on_uaccess_err = 1;
and that causes the fixup_exception() code to send the signal *despite* the
exception being caught.
And I think that is in fact completely bogus. It's completely bogus
exactly because it sends that signal even when it *shouldn't* be sent -
like for the BPF user mode trace gathering.
In other words, I think the whole "sig_on_uaccess_err" thing is entirely
broken, because it makes any nested page-faults do all the wrong things.
Now, arguably, I don't think anybody should enable vsyscall emulation any
more, but this test case clearly does.
I think we should just make the "send SIGSEGV" be something that the
vsyscall emulation does on its own, not this broken per-thread state for
something that isn't actually per thread.
The x86 page fault code actually tried to deal with the "incorrect nesting"
by having that:
if (in_interrupt())
return;
which ignores the sig_on_uaccess_err case when it happens in interrupts,
but as shown by this example, these nested page faults do not need to be
about interrupts at all.
IOW, I think the only right thing is to remove that horrendously broken
code.
The attached patch looks like the ObviouslyCorrect(tm) thing to do.
NOTE! This broken code goes back to this commit in 2011:
4fc3490114bb ("x86-64: Set siginfo and context on vsyscall emulation faults")
... and back then the reason was to get all the siginfo details right.
Honestly, I do not for a moment believe that it's worth getting the siginfo
details right here, but part of the commit says:
This fixes issues with UML when vsyscall=emulate.
... and so my patch to remove this garbage will probably break UML in this
situation.
I do not believe that anybody should be running with vsyscall=emulate in
2024 in the first place, much less if you are doing things like UML. But
let's see if somebody screams.
There is a race condition when re-creating a kfd_process for a process.
This has been observed when a process under the debugger executes
exec(3). In this scenario:
- The process executes exec.
- This will eventually release the process's mm, which will cause the
kfd_process object associated with the process to be freed
(kfd_process_free_notifier decrements the reference count to the
kfd_process to 0). This causes kfd_process_ref_release to enqueue
kfd_process_wq_release to the kfd_process_wq.
- The debugger receives the PTRACE_EVENT_EXEC notification, and tries to
re-enable AMDGPU traps (KFD_IOC_DBG_TRAP_ENABLE).
- When handling this request, KFD tries to re-create a kfd_process.
This eventually calls kfd_create_process and kobject_init_and_add.
At this point the call to kobject_init_and_add can fail because the
old kfd_process.kobj has not been freed yet by kfd_process_wq_release.
This patch proposes to avoid this race by making sure to drain
kfd_process_wq before creating a new kfd_process object. This way, we
know that any cleanup task is done executing when we reach
kobject_init_and_add.
Signed-off-by: Lancelot SIX <lancelot.six@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
[why]
preOS will not support display mode programming and link training
for UHBR rates.
[how]
If we detect a sink that's UHBR capable, disable seamless boot
Reviewed-by: Anthony Koo <anthony.koo@amd.com> Acked-by: Wayne Lin <wayne.lin@amd.com> Signed-off-by: Sung Joon Kim <sungjoon.kim@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
[Why && How]
Screen flickering saw on 4K@60 eDP with high refresh rate external
monitor when booting up in DC mode. DC Mode Capping is disabled
which caused wrong UCLK being used.
Reviewed-by: Alvin Lee <alvin.lee2@amd.com> Acked-by: Wayne Lin <wayne.lin@amd.com> Signed-off-by: Leo Ma <hanghong.ma@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add VCO speed parameters in the bounding box array.
Acked-by: Wayne Lin <wayne.lin@amd.com> Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
[Why]
During DP tunnel creation, CM preallocates BW and reduces
estimated BW of other DPIA. CM release preallocation only
when allocation is complete. Display mode validation logic
validates timings based on bw available per host router.
In multi display setup, this causes bw allocation failure
when allocation greater than estimated bw.
[How]
Do zero alloc to make the CM to release preallocation and
update estimated BW correctly for all DPIAs per host router.
Currently DCN315 clk manager is missing code to enable/disable dtbclk.
Because of this, "optimized_required" flag is constantly set
and this prevents FreeSync from engaging for certain high bandwidth
display Modes which require DTBCLK.
Subtract the VRAM pinned memory when checking for available memory
in amdgpu_amdkfd_reserve_mem_limit function since that memory is not
available for use.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As described in the added code comment, a reference to .exit.text is ok
for drivers registered via module_platform_driver_probe(). Make this
explicit to prevent the following section mismatch warning
The SOF driver is selected whenever specific I2C/I2S HIDs are reported
as 'present' in the ACPI DSDT. In some cases, an HID is reported but
the hardware does not actually rely on I2C/I2S. This false positive
leads to an invalid selection of the SOF driver and as a result an
invalid topology is loaded.
This patch hardens the detection with a check that the NHLT table is
consistent with the report of an I2S-based codec in DSDT. This table
should expose at least one SSP endpoint configured for an I2S-codec
connection.
Tested on Huawei Matebook D14 (NBLB-WAX9N) using an HDaudio codec with
an invalid ES8336 ACPI HID reported:
[ 7.858249] snd_hda_intel 0000:00:1f.3: DSP detected with PCI class/subclass/prog-if info 0x040380
[ 7.858312] snd_hda_intel 0000:00:1f.3: snd_intel_dsp_find_config: no valid SSP found for HID ESSX8336, skipped
The rcu quiescent state is reported in the rcu-read critical section, so
the lockdep warning is triggered.
Fix this by splitting out the inner working of __do_softirq() into a helper
function which takes an argument to distinguish between ksoftirqd task
context and interrupted context and invoke it from the relevant call sites
with the proper context information and use that for the conditional
invocation of rcu_softirq_qs().
We did not delay after the second strobe signal, so another immediately
following access could potentially corrupt the written value.
This is a purely speculative fix with no supporting evidence, but after
taking out the spinlocks around the writes, it seems plausible that a
modern processor could be actually too fast. Also, it's just cleaner to
be consistent.
When a load is marked PROBE_MEM - e.g. due to PTR_UNTRUSTED access - the
address being loaded from is not necessarily valid. The BPF jit sets up
exception handlers for each such load which catch page faults and 0 out
the destination register.
If the address for the load is outside kernel address space, the load
will escape the exception handling and crash the kernel. To prevent this
from happening, the emits some instruction to verify that addr is > end
of userspace addresses.
x86 has a legacy vsyscall ABI where a page at address 0xffffffffff600000
is mapped with user accessible permissions. The addresses in this page
are considered userspace addresses by the fault handler. Therefore, a
BPF program accessing this page will crash the kernel.
This patch fixes the runtime checks to also check that the PROBE_MEM
address is below VSYSCALL_ADDR.
Example BPF program:
SEC("fentry/tcp_v4_connect")
int BPF_PROG(fentry_tcp_v4_connect, struct sock *sk)
{
*(volatile unsigned long *)&sk->sk_tsq_flags;
return 0;
}
Add PCI subdevice ID for the Intel D5005 Stratix 10 FPGA card as
used with the Open FPGA Stack (OFS) FPGA Interface Manager (FIM).
Unlike the Intel D5005 PAC FIM which exposed a separate PCI device ID,
the OFS FIM reuses the same device ID for all DFL-based FPGA cards
and differentiates on the subdevice ID. The subdevice ID values were
chosen as the numeric part of the FPGA card names in hexadecimal.
This happens because when we call rename we already have the inode mutex
held, and then we acquire the subvol_sem if we are a subvolume. This
makes the dependency
inode lock -> subvol sem
When we're running data relocation we will preallocate space for the
data relocation inode, and we always run the relocation under the
->cleaner_mutex. This now creates the dependency of
cleaner_mutex -> inode lock (from the prealloc) -> subvol_sem
Qgroup delete is doing this in the opposite order, it is acquiring the
subvol_sem and then it is acquiring the cleaner_mutex, which results in
this lockdep splat. This deadlock can't happen in reality, because we
won't ever rename the data reloc inode, nor is the data reloc inode a
subvolume.
However this is fairly easy to fix, simply take the cleaner mutex in the
case where we are disabling qgroups before we take the subvol_sem. This
resolves the lockdep splat.
Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Josef Bacik <josef@toxicpanda.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Assert that accesses to a non-existent vgic-v2 CPU interface
consistently fail across the various KVM device attr ioctls. This also
serves as a regression test for a bug wherein KVM hits a NULL
dereference when the CPUID specified in the ioctl is invalid.
Note that there is no need to print the observed errno, as TEST_ASSERT()
will take care of it.
projects / thirdparty / kernel / stable.git / log
