wifi: mac80211: pass 'sta' to ieee80211_rx_data_set_sta()

[ Upstream commit 0d846bdc11101ac0ba4d89c2be359af08cb9379b ]

There's at least one case in ieee80211_rx_for_interface()
where we might pass &((struct sta_info *)NULL)->sta to it
only to then do container_of(), and then checking the
result for NULL, but checking the result of container_of()
for NULL looks really odd.

Fix this by just passing the struct sta_info * instead.

Fixes: e66b7920aa5a ("wifi: mac80211: fix initialization of rx->link and rx->link_sta")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Fix map_kptr test.

[ Upstream commit 62d101d5f422cde39b269f7eb4cbbe2f1e26f9d4 ]

The compiler is optimizing out majority of unref_ptr read/writes, so the test
wasn't testing much. For example, one could delete '__kptr' tag from
'struct prog_test_ref_kfunc __kptr *unref_ptr;' and the test would still "pass".

Convert it to volatile stores. Confirmed by comparing bpf asm before/after.

Fixes: 2cbc469a6fc3 ("selftests/bpf: Add C tests for kptr")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20230214235051.22938-1-alexei.starovoitov@gmail.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/hisi: Drop second sensor hi3660

[ Upstream commit 15cc25829a97c3957e520e971868aacc84341317 ]

The commit 74c8e6bffbe1 ("driver core: Add __alloc_size hint to devm
allocators") exposes a panic "BRK handler: Fatal exception" on the
hi3660_thermal_probe funciton.
This is because the function allocates memory for only one
sensors array entry, but tries to fill up a second one.

Fix this by removing the unneeded second access.

Fixes: 7d3a2a2bbadb ("thermal/drivers/hisi: Fix number of sensors on hi3660")
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Link: https://lore.kernel.org/linux-mm/20221101223321.1326815-5-keescook@chromium.org/
Link: https://lore.kernel.org/r/20230210141507.71014-1-yongqin.liu@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/lib/thermal: Fix thermal_sampling_exit()

[ Upstream commit a29cbd76aaf63f5493e962aa2fbaadcdc4615143 ]

thermal_sampling_init() suscribes to THERMAL_GENL_SAMPLING_GROUP_NAME group
so thermal_sampling_exit() should unsubscribe from the same group.

Fixes: 47c4b0de080a ("tools/lib/thermal: Add a thermal library")
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/20230202102812.453357-1-vincent.guittot@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: fix off-by-one link setting

[ Upstream commit cf08e29db760b144bde51e2444f3430c75763e26 ]

The convention for find_first_bit() is 0-based, while ffs()
is 1-based, so this is now off-by-one. I cannot reproduce the
gcc-9 problem, but since the -1 is now removed, I'm hoping it
will still avoid the original issue.

Reported-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Fixes: 1d8d4af43474 ("wifi: mac80211: avoid u32_encode_bits() warning")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: avoid u32_encode_bits() warning

[ Upstream commit 1d8d4af4347420d657be448f8be4c39c558f3b5d ]

gcc-9 triggers a false-postive warning in ieee80211_mlo_multicast_tx()
for u32_encode_bits(ffs(links) - 1, ...), since ffs() can return zero
on an empty bitmask, and the negative argument to u32_encode_bits()
is then out of range:

In file included from include/linux/ieee80211.h:21,
                 from include/net/cfg80211.h:23,
                 from net/mac80211/tx.c:23:
In function 'u32_encode_bits',
    inlined from 'ieee80211_mlo_multicast_tx' at net/mac80211/tx.c:4437:17,
    inlined from 'ieee80211_subif_start_xmit' at net/mac80211/tx.c:4485:3:
include/linux/bitfield.h:177:3: error: call to '__field_overflow' declared with attribute error: value doesn't fit into mask
  177 |   __field_overflow();     \
      |   ^~~~~~~~~~~~~~~~~~
include/linux/bitfield.h:197:2: note: in expansion of macro '____MAKE_OP'
  197 |  ____MAKE_OP(u##size,u##size,,)
      |  ^~~~~~~~~~~
include/linux/bitfield.h:200:1: note: in expansion of macro '__MAKE_OP'
  200 | __MAKE_OP(32)
      | ^~~~~~~~~

Newer compiler versions do not cause problems with the zero argument
because they do not consider this a __builtin_constant_p().
It's also harmless since the hweight16() check already guarantees
that this cannot be 0.

Replace the ffs() with an equivalent find_first_bit() check that
matches the later for_each_set_bit() style and avoids the warning.

Fixes: 963d0e8d08d9 ("wifi: mac80211: optionally implement MLO multicast TX")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230214132025.1532147-1-arnd@kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: Don't translate MLD addresses for multicast

[ Upstream commit daf8fb4295dccc032515cdc1bd3873370063542b ]

MLD address translation should be done only for individually addressed
frames. Otherwise, AAD calculation would be wrong and the decryption
would fail.

Fixes: e66b7920aa5ac ("wifi: mac80211: fix initialization of rx->link and rx->link_sta")
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Link: https://lore.kernel.org/r/20230214101048.792414-1-andrei.otcheretianski@intel.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: fix non-MLO station association

[ Upstream commit aaacf1740f2f95e0c5449ff3bbcff252d69cf952 ]

Non-MLO station frames are dropped in Rx path due to the condition
check in ieee80211_rx_is_valid_sta_link_id(). In multi-link AP scenario,
non-MLO stations try to connect in any of the valid links in the ML AP,
where the station valid_links and link_id params are valid in the
ieee80211_sta object. But ieee80211_rx_is_valid_sta_link_id() always
return false for the non-MLO stations by the assumption taken is
valid_links and link_id are not valid in non-MLO stations object
(ieee80211_sta), this assumption is wrong. Due to this assumption,
non-MLO station frames are dropped which leads to failure in association.

Fix it by removing the condition check and allow the link validation
check for the non-MLO stations.

Fixes: e66b7920aa5a ("wifi: mac80211: fix initialization of rx->link and rx->link_sta")
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Link: https://lore.kernel.org/r/20230206160330.1613-1-quic_periyasa@quicinc.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: make rate u32 in sta_set_rate_info_rx()

[ Upstream commit 59336e07b287d91dc4ec265e07724e8f7e3d0209 ]

The value of last_rate in ieee80211_sta_rx_stats is degraded from u32 to
u16 after being assigned to rate variable, which causes information loss
in STA_STATS_FIELD_TYPE and later bitfields.

Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Link: https://lore.kernel.org/r/20230209110659.25447-1-shayne.chen@mediatek.com
Fixes: 41cbb0f5a295 ("mac80211: add support for HE")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mac80211: move color collision detection report in a delayed work

[ Upstream commit 9288188438d85e22c23cfd6657ee8a801babc83c ]

Move color collision report in a dedicated delayed work and do not run
it in interrupt context in order to rate-limit the number of events
reported to userspace. Moreover grab wdev mutex in
ieee80211_color_collision_detection_work routine since it is required
by cfg80211_obss_color_collision_notify().

Tested-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Fixes: 5f9404abdf2a ("mac80211: add support for BSS color change")
Link: https://lore.kernel.org/r/3f6cf60c892ad40c1cca4a55d62b1224ef1c6ce9.1674644379.git.lorenzo@kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: crypto4xx - Call dma_unmap_page when done

[ Upstream commit bcdda4301bdc4955d45f7e1ffefb6207967b067e ]

In crypto4xx_cipher_done, we should be unmapping the dst page, not
mapping it.

This was flagged by a sparse warning about the unused addr variable.
While we're at it, also fix a sparse warning regarding the unused
ctx variable in crypto4xx_ahash_done (by actually using it).

Fixes: 049359d65527 ("crypto: amcc - Add crypt4xx driver")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tested-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: octeontx2 - Fix objects shared between several modules

[ Upstream commit 72bc4e71dbeedee0a446bcbc37c9bb25449072b7 ]

cn10k_cpt.o, otx2_cptlf.o and otx2_cpt_mbox_common.o are linked
into both rvu_cptpf and rvu_cptvf modules:

> scripts/Makefile.build:252: ./drivers/crypto/marvell/octeontx2/Makefile:
> cn10k_cpt.o is added to multiple modules: rvu_cptpf rvu_cptvf
> scripts/Makefile.build:252: ./drivers/crypto/marvell/octeontx2/Makefile:
> otx2_cptlf.o is added to multiple modules: rvu_cptpf rvu_cptvf
> scripts/Makefile.build:252: ./drivers/crypto/marvell/octeontx2/Makefile:
> otx2_cpt_mbox_common.o is added to multiple modules: rvu_cptpf rvu_cptvf

Despite they're build under the same Kconfig option
(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT), it's better do link the common
code into a standalone module and export the shared functions. Under
certain circumstances, this can lead to the same situation as fixed
by commit 637a642f5ca5 ("zstd: Fixing mixed module-builtin objects").
Plus, those three common object files are relatively big to duplicate
them several times.

Introduce the new module, rvu_cptcommon, to provide the common
functions to both modules.

Fixes: 19d8e8c7be15 ("crypto: octeontx2 - add virtual function driver support")
Suggested-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Reviewed-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: resource: Do IRQ override on all TongFang GMxRGxx

[ Upstream commit 17bb7046e7ce038a73ee97eaa804e0300c5199e2 ]

Apply commit 7592b79ba4a9 ("ACPI: resource: do IRQ override on XMG Core 15")
override for all vendors using this mainboard.

Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Fixes: 9946e39fe8d0 ("ACPI: resource: skip IRQ override on AMD Zen platforms")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: resource: Add IRQ overrides for MAINGEAR Vector Pro 2 models

[ Upstream commit cb18703c179713056bd7e3bdfc2260ab4e8658f0 ]

Fix a regression introduced by commit 9946e39fe8d0 ("ACPI: resource: skip
IRQ override on AMD Zen platforms") on MAINGEAR Vector Pro 2 systems, which
causes the built-in keyboard to not work. This restores the functionality
by adding an IRQ override.

No other IRQs were being overridden before, so this should be all that is
needed for these systems. I have personally tested this on the 15" model
(MG-VCP2-15A3070T), and I have confirmation that the issue is present on
the 17" model (MG-VCP2-17A3070T).

Fixes: 9946e39fe8d0 ("ACPI: resource: skip IRQ override on AMD Zen platforms")
Signed-off-by: Adam Niederer <adam.niederer@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Fix out-of-srctree build

[ Upstream commit 0b0757244754ea1d0721195c824770f5576e119e ]

Building BPF selftests out of srctree fails with:

  make: *** No rule to make target '/linux-build//ima_setup.sh', needed by 'ima_setup.sh'.  Stop.

The culprit is the rule that defines convenient shorthands like
"make test_progs", which builds $(OUTPUT)/test_progs. These shorthands
make sense only for binaries that are built though; scripts that live
in the source tree do not end up in $(OUTPUT).

Therefore drop $(TEST_PROGS) and $(TEST_PROGS_EXTENDED) from the rule.

The issue exists for a while, but it became a problem only after commit
d68ae4982cb7 ("selftests/bpf: Install all required files to run selftests"),
which added dependencies on these scripts.

Fixes: 03dcb78460c2 ("selftests/bpf: Add simple per-test targets to Makefile")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230208231211.283606-1-iii@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mwifiex: fix loop iterator in mwifiex_update_ampdu_txwinsize()

[ Upstream commit 3cfb7df24cee0f5fdc4cc5d3176cab9aadfcb430 ]

This code re-uses "i" to be the iterator for both the inside and outside
loops.  It means the outside loop will exit earlier than intended.

Fixes: d219b7eb3792 ("mwifiex: handle BT coex event to adjust Rx BA window size")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/Y+ERnaDaZD7RtLvX@kili
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwl4965: Add missing check for create_singlethread_workqueue()

[ Upstream commit 26e6775f75517ad6844fe5b79bc5f3fa8c22ee61 ]

Add the check for the return value of the create_singlethread_workqueue()
in order to avoid NULL pointer dereference.

Fixes: b481de9ca074 ("[IWLWIFI]: add iwlwifi wireless drivers")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Acked-by: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230209010748.45454-1-jiasheng@iscas.ac.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwl3945: Add missing check for create_singlethread_workqueue

[ Upstream commit 1fdeb8b9f29dfd64805bb49475ac7566a3cb06cb ]

Add the check for the return value of the create_singlethread_workqueue
in order to avoid NULL pointer dereference.

Fixes: b481de9ca074 ("[IWLWIFI]: add iwlwifi wireless drivers")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Acked-by: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230208063032.42763-2-jiasheng@iscas.ac.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

clocksource/drivers/riscv: Patch riscv_clock_next_event() jump before first use

[ Upstream commit 225b9596cb0227c1c1b1e4a836dad43595c3e61a ]

A static key is used to select between SBI and Sstc timer usage in
riscv_clock_next_event(), but currently the direction is resolved
after cpuhp_setup_state() is called (which sets the next event).  The
first event will therefore fall through the sbi_set_timer() path; this
breaks Sstc-only systems.  So, apply the jump patching before first
use.

Fixes: 9f7a8ff6391f ("RISC-V: Prefer sstc extension if available")
Signed-off-by: Matt Evans <mev@rivosinc.com>
Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/CDDAB2D0-264E-42F3-8E31-BA210BEB8EC1@rivosinc.com
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RISC-V: time: initialize hrtimer based broadcast clock event device

[ Upstream commit 8b3b8fbb4896984b5564789a42240e4b3caddb61 ]

Similarly to commit 022eb8ae8b5e ("ARM: 8938/1: kernel: initialize
broadcast hrtimer based clock event device"), RISC-V needs to initiate
hrtimer based broadcast clock event device before C3STOP can be used.
Otherwise, the introduction of C3STOP for the RISC-V arch timer in
commit 232ccac1bd9b ("clocksource/drivers/riscv: Events are stopped
during CPU suspend") leaves us without any broadcast timer registered.
This prevents the kernel from entering oneshot mode, which breaks timer
behaviour, for example clock_nanosleep().

A test app that sleeps each cpu for 6, 5, 4, 3 ms respectively, HZ=250
& C3STOP enabled, the sleep times are rounded up to the next jiffy:
== CPU: 1 ==      == CPU: 2 ==      == CPU: 3 ==      == CPU: 4 ==
Mean: 7.974992    Mean: 7.976534    Mean: 7.962591    Mean: 3.952179
Std Dev: 0.154374 Std Dev: 0.156082 Std Dev: 0.171018 Std Dev: 0.076193
Hi: 9.472000      Hi: 10.495000     Hi: 8.864000      Hi: 4.736000
Lo: 6.087000      Lo: 6.380000      Lo: 4.872000      Lo: 3.403000
Samples: 521      Samples: 521      Samples: 521      Samples: 521

Link: https://lore.kernel.org/linux-riscv/YzYTNQRxLr7Q9JR0@spud/
Fixes: 232ccac1bd9b ("clocksource/drivers/riscv: Events are stopped during CPU suspend")
Suggested-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Samuel Holland <samuel@sholland.org>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
Link: https://lore.kernel.org/r/20230103141102.772228-2-apatel@ventanamicro.com
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

m68k: /proc/hardware should depend on PROC_FS

[ Upstream commit 1e5b5df65af99013b4d31607ddb3ca5731dbe44d ]

When CONFIG_PROC_FS is not set, there is a build error for an unused
function. Make PROC_HARDWARE depend on PROC_FS to prevent this error.

In file included from ../arch/m68k/kernel/setup.c:3:
../arch/m68k/kernel/setup_mm.c:477:12: error: 'hardware_proc_show' defined but not used [-Werror=unused-function]
  477 | static int hardware_proc_show(struct seq_file *m, void *v)
      |            ^~~~~~~~~~~~~~~~~~

Fixes: 66d857b08b8c ("m68k: merge m68k and m68knommu arch directories") # v3.0
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Link: https://lore.kernel.org/r/20230209010825.24136-1-rdunlap@infradead.org
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: rsa-pkcs1pad - Use akcipher_request_complete

[ Upstream commit 564cabc0ca0bdfa8f0fc1ae74b24d0a7554522c5 ]

Use the akcipher_request_complete helper instead of calling the
completion function directly.  In fact the previous code was buggy
in that EINPROGRESS was never passed back to the original caller.

Fixes: 3d5b1ecdea6f ("crypto: rsa - RSA padding algorithm")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rds: rds_rm_zerocopy_callback() correct order for list_add_tail()

[ Upstream commit 68762148d1b011d47bc2ceed7321739b5aea1e63 ]

rds_rm_zerocopy_callback() uses list_add_tail() with swapped
arguments. This links the list head with the new entry, losing
the references to the remaining part of the list.

Fixes: 9426bbc6de99 ("rds: use list structure to track information for zerocopy completion notification")
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

xen/grant-dma-iommu: Implement a dummy probe_device() callback

[ Upstream commit 2062f9fb6445451b189595e295765c69f43bc12e ]

Update stub IOMMU driver (which main purpose is to reuse generic
IOMMU device-tree bindings by Xen grant DMA-mapping layer on Arm)
according to the recent changes done in the following
commit 57365a04c921 ("iommu: Move bus setup to IOMMU device registration").

With probe_device() callback being called during IOMMU device registration,
the uninitialized callback just leads to the "kernel NULL pointer
dereference" issue during boot. Fix that by adding a dummy callback.

Looks like the release_device() callback is not mandatory to be
implemented as IOMMU framework makes sure that callback is initialized
before dereferencing.

Reported-by: Viresh Kumar <viresh.kumar@linaro.org>
Fixes: 57365a04c921 ("iommu: Move bus setup to IOMMU device registration")
Signed-off-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Link: https://lore.kernel.org/r/20230208153649.3604857-1-olekstysh@gmail.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

libbpf: Fix alen calculation in libbpf_nla_dump_errormsg()

[ Upstream commit 17bcd27a08a21397698edf143084d7c87ce17946 ]

The code assumes that everything that comes after nlmsgerr are nlattrs.
When calculating their size, it does not account for the initial
nlmsghdr. This may lead to accessing uninitialized memory.

Fixes: bbf48c18ee0c ("libbpf: add error reporting in XDP")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20230210001210.395194-8-iii@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/ap: fix status returned by ap_qact()

[ Upstream commit a2522c80f074c35254974fec39fffe8b8d75befe ]

Since commit 159491f3b509 ("s390/ap: rework assembler functions to use
unions for in/out register variables") the  function ap_qact() tries to
grab the status from the wrong part of the register. Thus we always end
up with zeros. Which is wrong, among others, because we detect failures
via status.response_code.

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Reported-by: Harald Freudenberger <freude@linux.ibm.com>
Fixes: 159491f3b509 ("s390/ap: rework assembler functions to use unions for in/out register variables")
Reviewed-by: Harald Freudenberger <freude@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/ap: fix status returned by ap_aqic()

[ Upstream commit 394740d7645ea767795074287769dd26dbd4d782 ]

There function ap_aqic() tries to grab the status from the
wrong part of the register. Thus we always end up with
zeros. Which is wrong, among others, because we detect
failures via status.response_code.

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Reported-by: Janosch Frank <frankja@linux.ibm.com>
Fixes: 159491f3b509 ("s390/ap: rework assembler functions to use unions for in/out register variables")
Reviewed-by: Harald Freudenberger <freude@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390: vfio-ap: tighten the NIB validity check

[ Upstream commit a64a6d23874c574d30a9816124b2dc37467f3811 ]

The NIB is architecturally invalid if the address designates a
storage location that is not installed or if it is zero.

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Reported-by: Janosch Frank <frankja@linux.ibm.com>
Fixes: ec89b55e3bce ("s390: ap: implement PAPQ AQIC interception in kernel")
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Pierre Morel <pmorel@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: ipa: generic command param fix

[ Upstream commit 2df181f09c961377a55510a864216d48d787fe49 ]

Starting at IPA v4.11, the GSI_GENERIC_COMMAND GSI register got a
new PARAMS field.  The code that encodes a value into that field
sets it unconditionally, which is wrong.

We currently only provide 0 as the field's value, so this error has
no real effect.  Still, it's a bug, so let's fix it.

Fix an (unrelated) incorrect comment as well.  Fields in the
ERROR_LOG GSI register actually *are* defined for IPA versions
prior to v3.5.1.

Fixes: fe68c43ce388 ("net: ipa: support enhanced channel flow control")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: hci_qca: get wakeup status from serdev device handle

[ Upstream commit 03b0093f7b310493bc944a20f725228cfe0d3fea ]

Bluetooth controller attached via the UART is handled by the serdev driver.
Get the wakeup status from the device handle through serdev, instead of the
parent path.

Fixes: c1a74160eaf1 ("Bluetooth: hci_qca: Add device_may_wakeup support")
Signed-off-by: Zhengping Jiang <jiangzp@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: L2CAP: Fix potential user-after-free

[ Upstream commit df5703348813235874d851934e957c3723d71644 ]

This fixes all instances of which requires to allocate a buffer calling
alloc_skb which may release the chan lock and reacquire later which
makes it possible that the chan is disconnected in the meantime.

Fixes: a6a5568c03c4 ("Bluetooth: Lock the L2CAP channel when sending")
Reported-by: Alexander Coffin <alex.coffin@matician.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: hci_conn: Refactor hci_bind_bis() since it always succeeds

[ Upstream commit a00a29b0eeea6caaaf9edc3dd284f81b072ee343 ]

The compiler thinks "conn" might be NULL after a call to hci_bind_bis(),
which cannot happen. Avoid any confusion by just making it not return a
value since it cannot fail. Fixes the warnings seen with GCC 13:

In function 'arch_atomic_dec_and_test',
    inlined from 'atomic_dec_and_test' at ../include/linux/atomic/atomic-instrumented.h:576:9,
    inlined from 'hci_conn_drop' at ../include/net/bluetooth/hci_core.h:1391:6,
    inlined from 'hci_connect_bis' at ../net/bluetooth/hci_conn.c:2124:3:
../arch/x86/include/asm/rmwcc.h:37:9: warning: array subscript 0 is outside array bounds of 'atomic_t[0]' [-Warray-bounds=]
   37 |         asm volatile (fullop CC_SET(cc) \
      |         ^~~
...
In function 'hci_connect_bis':
cc1: note: source object is likely at address zero

Fixes: eca0ae4aea66 ("Bluetooth: Add initial implementation of BIS connections")
Cc: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Cc: Marcel Holtmann <marcel@holtmann.org>
Cc: Johan Hedberg <johan.hedberg@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: linux-bluetooth@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: davinci: Fix clk use after free

[ Upstream commit 5d8f384a9b4fc50f6a18405f1c08e5a87a77b5b3 ]

The remove function first frees the clks and only then calls
cpufreq_unregister_driver(). If one of the cpufreq callbacks is called
just before cpufreq_unregister_driver() is run, the freed clks might be
used.

Fixes: 6601b8030de3 ("davinci: add generic CPUFreq driver for DaVinci")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

OPP: fix error checking in opp_migrate_dentry()

[ Upstream commit eca4c0eea53432ec4b711b2a8ad282cbad231b4f ]

Since commit ff9fb72bc077 ("debugfs: return error values,
not NULL") changed return value of debugfs_rename() in
error cases from %NULL to %ERR_PTR(-ERROR), we should
also check error values instead of NULL.

Fixes: ff9fb72bc077 ("debugfs: return error values, not NULL")
Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tap: tap_open(): correctly initialize socket uid

[ Upstream commit 66b2c338adce580dfce2199591e65e2bab889cff ]

sock_init_data() assumes that the `struct socket` passed in input is
contained in a `struct socket_alloc` allocated with sock_alloc().
However, tap_open() passes a `struct socket` embedded in a `struct
tap_queue` allocated with sk_alloc().
This causes a type confusion when issuing a container_of() with
SOCK_INODE() in sock_init_data() which results in assigning a wrong
sk_uid to the `struct sock` in input.
On default configuration, the type confused field overlaps with
padding bytes between `int vnet_hdr_sz` and `struct tap_dev __rcu
*tap` in `struct tap_queue`, which makes the uid of all tap sockets 0,
i.e., the root one.
Fix the assignment by using sock_init_data_uid().

Fixes: 86741ec25462 ("net: core: Add a UID field to struct sock.")
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tun: tun_chr_open(): correctly initialize socket uid

[ Upstream commit a096ccca6e503a5c575717ff8a36ace27510ab0a ]

sock_init_data() assumes that the `struct socket` passed in input is
contained in a `struct socket_alloc` allocated with sock_alloc().
However, tun_chr_open() passes a `struct socket` embedded in a `struct
tun_file` allocated with sk_alloc().
This causes a type confusion when issuing a container_of() with
SOCK_INODE() in sock_init_data() which results in assigning a wrong
sk_uid to the `struct sock` in input.
On default configuration, the type confused field overlaps with the
high 4 bytes of `struct tun_struct __rcu *tun` of `struct tun_file`,
NULL at the time of call, which makes the uid of all tun sockets 0,
i.e., the root one.
Fix the assignment by using sock_init_data_uid().

Fixes: 86741ec25462 ("net: core: Add a UID field to struct sock.")
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: add sock_init_data_uid()

[ Upstream commit 584f3742890e966d2f0a1f3c418c9ead70b2d99e ]

Add sock_init_data_uid() to explicitly initialize the socket uid.
To initialise the socket uid, sock_init_data() assumes a the struct
socket* sock is always embedded in a struct socket_alloc, used to
access the corresponding inode uid. This may not be true.
Examples are sockets created in tun_chr_open() and tap_open().

Fixes: 86741ec25462 ("net: core: Add a UID field to struct sock.")
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/boot: fix mem_detect extended area allocation

[ Upstream commit 22476f47b6b7fb7d066c71f67ebc11892adb0849 ]

Allocation of mem_detect extended area was not considered neither
in commit 9641b8cc733f ("s390/ipl: read IPL report at early boot")
nor in commit b2d24b97b2a9 ("s390/kernel: add support for kernel address
space layout randomization (KASLR)"). As a result mem_detect extended
theoretically may overlap with ipl report or randomized kernel image
position. But as mem_detect code will allocate extended area only
upon exceeding 255 online regions (which should alternate with offline
memory regions) it is not seen in practice.

To make sure mem_detect extended area does not overlap with ipl report
or randomized kernel position extend usage of "safe_addr". Make initrd
handling and mem_detect extended area allocation code move it further
right and make KASLR takes in into consideration as well.

Fixes: 9641b8cc733f ("s390/ipl: read IPL report at early boot")
Fixes: b2d24b97b2a9 ("s390/kernel: add support for kernel address space layout randomization (KASLR)")
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/mem_detect: rely on diag260() if sclp_early_get_memsize() fails

[ Upstream commit eb33f9eb304a4c18beb5ba6362eaa5c4beaf40d8 ]

In case sclp_early_get_memsize() fails but diag260() succeeds make sure
some sane value is returned. This error scenario is highly unlikely,
but this change makes system able to boot in such case.

Suggested-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Stable-dep-of: 22476f47b6b7 ("s390/boot: fix mem_detect extended area allocation")
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/boot: cleanup decompressor header files

[ Upstream commit 9c3205b2b062420c26b33924b910880889acf832 ]

Move declarations to appropriate header files. Instead of cryptic
casting directly assign struct vmlinux_info type to _vmlinux_info
linker script variable - wich it actually is.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Stable-dep-of: 22476f47b6b7 ("s390/boot: fix mem_detect extended area allocation")
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/vmem: fix empty page tables cleanup under KASAN

[ Upstream commit 108303b0a2d27cb14eed565e33e64ad9eefe5d7e ]

Commit b9ff81003cf1 ("s390/vmem: cleanup empty page tables") introduced
empty page tables cleanup in vmem code, but when the kernel is built
with KASAN enabled the code has no effect due to wrong KASAN shadow
memory intersection condition, which effectively ignores any memory
range below KASAN shadow. Fix intersection condition to make code
work as anticipated.

Fixes: b9ff81003cf1 ("s390/vmem: cleanup empty page tables")
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/mem_detect: fix detect_memory() error handling

[ Upstream commit 3400c35a4090704e6c465449616ab7e67a9209e7 ]

Currently if for some reason sclp_early_read_info() fails,
sclp_early_get_memsize() will not set max_physmem_end and it
will stay uninitialized. Any garbage value other than 0 will lead
to detect_memory() taking wrong path or returning a garbage value
as max_physmem_end. To avoid that simply initialize max_physmem_end.

Fixes: 73045a08cf55 ("s390: unify identity mapping limits handling")
Reported-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

irqchip/ti-sci: Fix refcount leak in ti_sci_intr_irq_domain_probe

[ Upstream commit 02298b7bae12936ca313975b02e7f98b06670d37 ]

of_irq_find_parent() returns a node pointer with refcount incremented,
We should use of_node_put() on it when not needed anymore.
Add missing of_node_put() to avoid refcount leak.

Fixes: cd844b0715ce ("irqchip/ti-sci-intr: Add support for Interrupt Router driver")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20230102085611.3955984-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

irqchip/irq-mvebu-gicp: Fix refcount leak in mvebu_gicp_probe

[ Upstream commit 9419e700021a393f67be36abd0c4f3acc6139041 ]

of_irq_find_parent() returns a node pointer with refcount incremented,
We should use of_node_put() on it when not needed anymore.
Add missing of_node_put() to avoid refcount leak.

Fixes: a68a63cb4dfc ("irqchip/irq-mvebu-gicp: Add new driver for Marvell GICP")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20230102084208.3951758-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

irqchip/alpine-msi: Fix refcount leak in alpine_msix_init_domains

[ Upstream commit 071d068b89e95d1b078aa6bbcb9d0961b77d6aa1 ]

of_irq_find_parent() returns a node pointer with refcount incremented,
We should use of_node_put() on it when not needed anymore.
Add missing of_node_put() to avoid refcount leak.

Fixes: e6b78f2c3e14 ("irqchip: Add the Alpine MSIX interrupt controller")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20230102082811.3947760-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

irqchip: Fix refcount leak in platform_irqchip_probe

[ Upstream commit 6caa5a2b78f5f53c433d3a3781e53325da22f0ac ]

of_irq_find_parent() returns a node pointer with refcount incremented,
We should use of_node_put() on it when not needed anymore.
Add missing of_node_put() to avoid refcount leak.

Fixes: f8410e626569 ("irqchip: Add IRQCHIP_PLATFORM_DRIVER_BEGIN/END and IRQCHIP_MATCH helper macros")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20230102121318.3990586-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: Enhance debug print in page allocation failure

[ Upstream commit 7eef93003e5d20e1a6a6e59e12d914b5431cbda2 ]

Provide more details to aid debugging.

Fixes: bf0bf77f6519 ("mlx5: Support communicating arbitrary host page size to firmware")
Signed-off-by: Eran Ben Elisha <eranbe@nvidia.com>
Signed-off-by: Majd Dibbiny <majd@nvidia.com>
Signed-off-by: Jack Morgenstein <jackm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7921: fix error code of return in mt7921_acpi_read

[ Upstream commit 888d89034f9eaeab9b5b75f13dbe35376c7dd471 ]

Kernel NULL pointer dereference when ACPI SAR table isn't implemented well.
Fix the error code of return to mark the ACPI SAR table as invalid.

[    5.077128] mt7921e 0000:06:00.0: sar cnt = 0
[    5.077381] BUG: kernel NULL pointer dereference, address:
0000000000000004
[    5.077630] #PF: supervisor read access in kernel mode
[    5.077883] #PF: error_code(0x0000) - not-present page
[    5.078138] PGD 0 P4D 0
[    5.078398] Oops: 0000 [#1] PREEMPT SMP NOPTI
[    5.079202] RIP: 0010:mt7921_init_acpi_sar+0x106/0x220
[mt7921_common]
...
[    5.080786] Call Trace:
[    5.080786]  <TASK>
[    5.080786]  mt7921_register_device+0x37d/0x490 [mt7921_common]
[    5.080786]  mt7921_pci_probe.part.0+0x2ee/0x310 [mt7921e]
[    5.080786]  mt7921_pci_probe+0x52/0x70 [mt7921e]
[    5.080786]  local_pci_probe+0x47/0x90
[    5.080786]  pci_call_probe+0x55/0x190
[    5.080786]  pci_device_probe+0x84/0x120

Fixes: f965333e491e ("mt76: mt7921: introduce ACPI SAR support")
Signed-off-by: Aaron Ma <aaron.ma@canonical.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: add memory barrier to SDIO queue kick

[ Upstream commit 5f54237ad798f41cb6a503271aa9ca47188cfb9b ]

Ensure the entry has been fully updated before SDIO bus worker access
it. This patch would fix potential memory risk in both mt7663s and
mt7921s.

Fixes: 764dee47e2c1 ("mt76: sdio: move common code in mt76_sdio module")
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7915: fix WED TxS reporting

[ Upstream commit 0d7084e209a9e2c924cb0d6e7f1f978db2a54127 ]

The previous commit forgot to remove a leftover check.

Fixes: 43eaa3689507 ("wifi: mt76: add PPDU based TxS support for WED device")
Reported-By: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7915: fix memory leak in mt7915_mcu_exit

[ Upstream commit 49bd78282e79ad177d14f37f4049f0605bf92dad ]

Always purge mcu skb queues in mt7915_mcu_exit routine even if
mt7915_firmware_state fails.

Fixes: e57b7901469f ("mt76: add mac80211 driver for MT7915 PCIe-based chipsets")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7915: call mt7915_mcu_set_thermal_throttling() only after init_work

[ Upstream commit 7d12b38ab6f6b77198cd3a66db19587bbdd3308c ]

Enable thermal management by default shall not be executed during mcu
init. This causes thermal configuration being reset to the firmware
default settings.

Fixes: 0063b86c9120 ("mt76: mt7915e: Enable thermal management by default")
Reviewed-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Howard Hsu <howard-yh.hsu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpftool: profile online CPUs instead of possible

[ Upstream commit 377c16fa3f3c60d21e4b05314c8be034ce37f2eb ]

The number of online cpu may be not equal to possible cpu.
"bpftool prog profile" can not create pmu event on possible
but on online cpu.

$ dmidecode -s system-product-name
PowerEdge R620
$ cat /sys/devices/system/cpu/possible
0-47
$ cat /sys/devices/system/cpu/online
0-31

Disable cpu dynamically:
$ echo 0 > /sys/devices/system/cpu/cpuX/online

If one cpu is offline, perf_event_open will return ENODEV.
To fix this issue:
* check value returned and skip offline cpu.
* close pmu_fd immediately on error path, avoid fd leaking.

Fixes: 47c09d6a9f67 ("bpftool: Introduce "prog profile" command")
Signed-off-by: Tonghao Zhang <tong@infragraf.org>
Cc: Quentin Monnet <quentin@isovalent.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Martin KaFai Lau <martin.lau@linux.dev>
Cc: Song Liu <song@kernel.org>
Cc: Yonghong Song <yhs@fb.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Stanislav Fomichev <sdf@google.com>
Cc: Hao Luo <haoluo@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20230202131701.29519-1-tong@infragraf.org
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: ccp - Flush the SEV-ES TMR memory before giving it to firmware

[ Upstream commit 46a334a98f585ef78d51d8f5736596887bdd7f54 ]

Perform a cache flush on the SEV-ES TMR memory after allocation to prevent
any possibility of the firmware encountering an error should dirty cache
lines be present. Use clflush_cache_range() to flush the SEV-ES TMR memory.

Fixes: 97f9ac3db661 ("crypto: ccp - Add support for SEV-ES to the PSP driver")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Initialize tc in xdp_synproxy

[ Upstream commit 354bb4a0e0b6be8f55bacbe7f08c94b4741f5658 ]

xdp_synproxy/xdp fails in CI with:

    Error: bpf_tc_hook_create: File exists

The XDP version of the test should not be calling bpf_tc_hook_create();
the reason it's happening anyway is that if we don't specify --tc on the
command line, tc variable remains uninitialized.

Fixes: 784d5dc0efc2 ("selftests/bpf: Add selftests for raw syncookie helpers in TC mode")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Reported-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20230202235335.3403781-1-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

can: rcar_canfd: Fix R-Car V3U GAFLCFG field accesses

[ Upstream commit 9be8c5583581244d8a77e41afa16b8b0a5ddabc0 ]

Each Global Acceptance Filter List Configuration Register (GAFLCFG)
contains two fields, and stores the number of channel rules for one
channel pair.

As R-Car V3U and later can have more than 2 channels, the field
selection should be based on the LSB (even or odd) of the channel
number, instead of on the full channel number.

Fixes: 45721c406dcf50d4 ("can: rcar_canfd: Add support for r8a779a0 SoC")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/36bcf0ffb96d6aaed970751f9546b901af638bcf.1674499048.git.geert+renesas@glider.be
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

kselftest/arm64: Fix enumeration of systems without 128 bit SME

[ Upstream commit 5f389238534ac8ca4ee3ab12eeb89d3984d303a1 ]

The current signal handling tests for SME do not account for the fact that
unlike SVE all SME vector lengths are optional so we can't guarantee that
we will encounter the minimum possible VL, they will hang enumerating VLs
on such systems. Abort enumeration when we find the lowest VL.

Fixes: 4963aeb35a9e ("kselftest/arm64: signal: Add SME signal handling tests")
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230131-arm64-kselftest-sig-sme-no-128-v1-1-d47c13dc8e1e@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: mei: fix compilation errors in rfkill()

[ Upstream commit 9cbd5a8abca904441e36861e3a92961bec41d13f ]

The rfkill() callback was invoked with wrong parameters.
It was missed since MEI is defined now as depending on BROKEN.
Fix that.

Fixes: d288067ede4b ("wifi: iwlwifi: mei: avoid blocking sap messages handling due to rtnl lock")
Fixes: 5aa7ce31bd84 ("wifi: iwlwifi: mei: make sure ownership confirmed message is sent")
Fixes: 95170a46b7dd ("wifi: iwlwifi: mei: don't send SAP commands if AMT is disabled")
Link: https://lore.kernel.org/r/20230126222821.305122-2-gregory.greenman@intel.com
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/bpf: Add expoline to tail calls

[ Upstream commit bb4ef8fc3d193ed8d5583fb47cbeff5d8fb8302f ]

All the indirect jumps in the eBPF JIT already use expolines, except
for the tail call one.

Fixes: de5cb6eb514e ("s390: use expoline thunks in the BPF JIT")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20230129190501.1624747-3-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

leds: led-class: Add missing put_device() to led_put()

[ Upstream commit 445110941eb94709216363f9d807d2508e64abd7 ]

led_put() is used to "undo" a successful of_led_get() call,
of_led_get() uses class_find_device_by_of_node() which returns
a reference to the device which must be free-ed with put_device()
when the caller is done with it.

Add a put_device() call to led_put() to free the reference returned
by class_find_device_by_of_node().

And also add a put_device() in the error-exit case of try_module_get()
failing.

Fixes: 699a8c7c4bd3 ("leds: Add of_led_get() and led_put()")
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230120114524.408368-2-hdegoede@redhat.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: xts - Handle EBUSY correctly

[ Upstream commit 51c082514c2dedf2711c99d93c196cc4eedceb40 ]

As it is xts only handles the special return value of EINPROGRESS,
which means that in all other cases it will free data related to the
request.

However, as the caller of xts may specify MAY_BACKLOG, we also need
to expect EBUSY and treat it in the same way.  Otherwise backlogged
requests will trigger a use-after-free.

Fixes: 8083b1bf8163 ("crypto: xts - add support for ciphertext stealing")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Fix vmtest static compilation error

[ Upstream commit 2514a31241e1e9067d379e0fbdb60e4bc2bf4659 ]

As stated in README.rst, in order to resolve errors with linker errors,
'LDLIBS=-static' should be used. Most problems will be solved by this
option, but in the case of urandom_read, this won't fix the problem. So
the Makefile is currently implemented to strip the 'static' option when
compiling the urandom_read. However, stripping this static option isn't
configured properly on $(LDLIBS) correctly, which is now causing errors
on static compilation.

    # LDLIBS=-static ./vmtest.sh
    ld.lld: error: attempted static link of dynamic object liburandom_read.so
    clang: error: linker command failed with exit code 1 (use -v to see invocation)
    make: *** [Makefile:190: /linux/tools/testing/selftests/bpf/urandom_read] Error 1
    make: *** Waiting for unfinished jobs....

This commit fixes this problem by configuring the strip with $(LDLIBS).

Fixes: 68084a136420 ("selftests/bpf: Fix building bpf selftests statically")
Signed-off-by: Daniel T. Lee <danieltimlee@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230125100440.21734-1-danieltimlee@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Use consistent build-id type for liburandom_read.so

[ Upstream commit 61fc5e66f755db24d27ba37ce1ee4873def1a074 ]

lld produces "fast" style build-ids by default, which is inconsistent
with ld's "sha1" style. Explicitly specify build-id style to be "sha1"
when linking liburandom_read.so the same way it is already done for
urandom_read.

Signed-off-by: Artem Savkov <asavkov@redhat.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: KP Singh <kpsingh@kernel.org>
Link: https://lore.kernel.org/bpf/20221104094016.102049-1-asavkov@redhat.com
Stable-dep-of: 2514a31241e1 ("selftests/bpf: Fix vmtest static compilation error")
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/microcode: Adjust late loading result reporting message

[ Upstream commit 6eab3abac7043226e5375e9ead0c7607ced6767b ]

During late microcode loading, the "Reload completed" message is issued
unconditionally, regardless of success or failure.

Adjust the message to report the result of the update.

  [ bp: Massage. ]

Fixes: 9bd681251b7c ("x86/microcode: Announce reload operation's completion")
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Link: https://lore.kernel.org/lkml/874judpqqd.ffs@tglx/
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/microcode: Check CPU capabilities after late microcode update correctly

[ Upstream commit c0dd9245aa9e25a697181f6085692272c9ec61bc ]

The kernel caches each CPU's feature bits at boot in an x86_capability[]
structure. However, the capabilities in the BSP's copy can be turned off
as a result of certain command line parameters or configuration
restrictions, for example the SGX bit. This can cause a mismatch when
comparing the values before and after the microcode update.

Another example is X86_FEATURE_SRBDS_CTRL which gets added only after
microcode update:

#  --- cpuid.before 2023-01-21 14:54:15.652000747 +0100
#  +++ cpuid.after 2023-01-21 14:54:26.632001024 +0100
#  @@ -10,7 +10,7 @@ CPU:
#      0x00000004 0x04: eax=0x00000000 ebx=0x00000000 ecx=0x00000000 edx=0x00000000
#      0x00000005 0x00: eax=0x00000040 ebx=0x00000040 ecx=0x00000003 edx=0x11142120
#      0x00000006 0x00: eax=0x000027f7 ebx=0x00000002 ecx=0x00000001 edx=0x00000000
#  -   0x00000007 0x00: eax=0x00000000 ebx=0x029c6fbf ecx=0x40000000 edx=0xbc002400
#  +   0x00000007 0x00: eax=0x00000000 ebx=0x029c6fbf ecx=0x40000000 edx=0xbc002e00
        ^^^

and which proves for a gazillionth time that late loading is a bad bad
idea.

microcode_check() is called after an update to report any previously
cached CPUID bits which might have changed due to the update.

Therefore, store the cached CPU caps before the update and compare them
with the CPU caps after the microcode update has succeeded.

Thus, the comparison is done between the CPUID *hardware* bits before
and after the upgrade instead of using the cached, possibly runtime
modified values in BSP's boot_cpu_data copy.

As a result, false warnings about CPUID bits changes are avoided.

  [ bp:
   - Massage.
- Add SRBDS_CTRL example.
- Add kernel-doc.
- Incorporate forgotten review feedback from dhansen.
]

Fixes: 1008c52c09dc ("x86/CPU: Add a microcode loader callback")
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230109153555.4986-3-ashok.raj@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/microcode: Add a parameter to microcode_check() to store CPU capabilities

[ Upstream commit ab31c74455c64e69342ddab21fd9426fcbfefde7 ]

Add a parameter to store CPU capabilities before performing a microcode
update so that CPU capabilities can be compared before and after update.

  [ bp: Massage. ]

Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20230109153555.4986-2-ashok.raj@intel.com
Stable-dep-of: c0dd9245aa9e ("x86/microcode: Check CPU capabilities after late microcode update correctly")
Signed-off-by: Sasha Levin <sashal@kernel.org>

powercap: fix possible name leak in powercap_register_zone()

[ Upstream commit 1b6599f741a4525ca761ecde46e5885ff1e6ba58 ]

In the error path after calling dev_set_name(), the device
name is leaked. To fix this, calling dev_set_name() before
device_register(), and call put_device() if it returns error.

All the resources is released in powercap_release(), so it
can return from powercap_register_zone() directly.

Fixes: 75d2364ea0ca ("PowerCap: Add class driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: seqiv - Handle EBUSY correctly

[ Upstream commit 32e62025e5e52fbe4812ef044759de7010b15dbc ]

As it is seqiv only handles the special return value of EINPROGERSS,
which means that in all other cases it will free data related to the
request.

However, as the caller of seqiv may specify MAY_BACKLOG, we also need
to expect EBUSY and treat it in the same way.  Otherwise backlogged
requests will trigger a use-after-free.

Fixes: 0a270321dbf9 ("[CRYPTO] seqiv: Add Sequence Number IV Generator")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: essiv - Handle EBUSY correctly

[ Upstream commit b5a772adf45a32c68bef28e60621f12617161556 ]

As it is essiv only handles the special return value of EINPROGERSS,
which means that in all other cases it will free data related to the
request.

However, as the caller of essiv may specify MAY_BACKLOG, we also need
to expect EBUSY and treat it in the same way.  Otherwise backlogged
requests will trigger a use-after-free.

Fixes: be1eb7f78aa8 ("crypto: essiv - create wrapper template...")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: ccp - Failure on re-initialization due to duplicate sysfs filename

[ Upstream commit 299bf602b3f92f1456aef59c6413591fb02e762a ]

The following warning appears during the CCP module re-initialization:

[  140.965403] sysfs: cannot create duplicate filename
'/devices/pci0000:00/0000:00:07.1/0000:03:00.2/dma/dma0chan0'
[  140.975736] CPU: 0 PID: 388 Comm: kworker/0:2 Kdump: loaded Not
tainted 6.2.0-0.rc2.18.eln124.x86_64 #1
[  140.985185] Hardware name: HPE ProLiant DL325 Gen10/ProLiant DL325
Gen10, BIOS A41 07/17/2020
[  140.993761] Workqueue: events work_for_cpu_fn
[  140.998151] Call Trace:
[  141.000613]  <TASK>
[  141.002726]  dump_stack_lvl+0x33/0x46
[  141.006415]  sysfs_warn_dup.cold+0x17/0x23
[  141.010542]  sysfs_create_dir_ns+0xba/0xd0
[  141.014670]  kobject_add_internal+0xba/0x260
[  141.018970]  kobject_add+0x81/0xb0
[  141.022395]  device_add+0xdc/0x7e0
[  141.025822]  ? complete_all+0x20/0x90
[  141.029510]  __dma_async_device_channel_register+0xc9/0x130
[  141.035119]  dma_async_device_register+0x19e/0x3b0
[  141.039943]  ccp_dmaengine_register+0x334/0x3f0 [ccp]
[  141.045042]  ccp5_init+0x662/0x6a0 [ccp]
[  141.049000]  ? devm_kmalloc+0x40/0xd0
[  141.052688]  ccp_dev_init+0xbb/0xf0 [ccp]
[  141.056732]  ? __pci_set_master+0x56/0xd0
[  141.060768]  sp_init+0x70/0x90 [ccp]
[  141.064377]  sp_pci_probe+0x186/0x1b0 [ccp]
[  141.068596]  local_pci_probe+0x41/0x80
[  141.072374]  work_for_cpu_fn+0x16/0x20
[  141.076145]  process_one_work+0x1c8/0x380
[  141.080181]  worker_thread+0x1ab/0x380
[  141.083953]  ? __pfx_worker_thread+0x10/0x10
[  141.088250]  kthread+0xda/0x100
[  141.091413]  ? __pfx_kthread+0x10/0x10
[  141.095185]  ret_from_fork+0x2c/0x50
[  141.098788]  </TASK>
[  141.100996] kobject_add_internal failed for dma0chan0 with -EEXIST,
don't try to register things with the same name in the same directory.
[  141.113703] ccp 0000:03:00.2: ccp initialization failed

The /dma/dma0chan0 sysfs file is not removed since dma_chan object
has been released in ccp_dma_release() before releasing dma device.
A correct procedure would be: release dma channels first => unregister
dma device => release ccp dma object.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=216888
Fixes: 68dbe80f5b51 ("crypto: ccp - Release dma channels before dmaengine unrgister")
Tested-by: Vladis Dronov <vdronov@redhat.com>
Signed-off-by: Koba Ko <koba.ko@canonical.com>
Reviewed-by: Vladis Dronov <vdronov@redhat.com>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Fix build errors if CONFIG_NF_CONNTRACK=m

[ Upstream commit 92afc5329a5b23d876b215b783d200352d5aaea6 ]

If CONFIG_NF_CONNTRACK=m, there are no definitions of NF_NAT_MANIP_SRC
and NF_NAT_MANIP_DST in vmlinux.h, build test_bpf_nf.c failed.

$ make -C tools/testing/selftests/bpf/

  CLNG-BPF [test_maps] test_bpf_nf.bpf.o
progs/test_bpf_nf.c:160:42: error: use of undeclared identifier 'NF_NAT_MANIP_SRC'
                bpf_ct_set_nat_info(ct, &saddr, sport, NF_NAT_MANIP_SRC);
                                                       ^
progs/test_bpf_nf.c:163:42: error: use of undeclared identifier 'NF_NAT_MANIP_DST'
                bpf_ct_set_nat_info(ct, &daddr, dport, NF_NAT_MANIP_DST);
                                                       ^
2 errors generated.

Copy the definitions in include/net/netfilter/nf_nat.h to test_bpf_nf.c,
in order to avoid redefinitions if CONFIG_NF_CONNTRACK=y, rename them with
___local suffix. This is similar with commit 1058b6a78db2 ("selftests/bpf:
Do not fail build if CONFIG_NF_CONNTRACK=m/n").

Fixes: b06b45e82b59 ("selftests/bpf: add tests for bpf_ct_set_nat_info kfunc")
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Tested-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/1674028604-7113-1-git-send-email-yangtiezhu@loongson.cn
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: battery: Fix missing NUL-termination with large strings

[ Upstream commit f2ac14b5f197e4a2dec51e5ceaa56682ff1592bc ]

When encountering a string bigger than the destination buffer (32 bytes),
the string is not properly NUL-terminated, causing buffer overreads later.

This for example happens on the Inspiron 3505, where the battery
model name is larger than 32 bytes, which leads to sysfs showing
the model name together with the serial number string (which is
NUL-terminated and thus prevents worse).

Fix this by using strscpy() which ensures that the result is
always NUL-terminated.

Fixes: 106449e870b3 ("ACPI: Battery: Allow extract string from integer")
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: cfg80211: Fix extended KCK key length check in nl80211_set_rekey_data()

[ Upstream commit df4969ca135b9b3b2c38c07514aaa775112ac835 ]

The extended KCK key length check wrongly using the KEK key attribute
for validation. Due to this GTK rekey offload is failing when the KCK
key length is 24 bytes even though the driver advertising
WIPHY_FLAG_SUPPORTS_EXT_KEK_KCK flag. Use correct attribute to fix the
same.

Fixes: 093a48d2aa4b ("cfg80211: support bigger kek/kck key length")
Signed-off-by: Shivani Baranwal <quic_shivbara@quicinc.com>
Signed-off-by: Veerendranath Jakkam <quic_vjakkam@quicinc.com>
Link: https://lore.kernel.org/r/20221206143715.1802987-2-quic_vjakkam@quicinc.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath11k: Fix memory leak in ath11k_peer_rx_frag_setup

[ Upstream commit ed3f83b3459a67a3ab9d806490ac304b567b1c2d ]

crypto_alloc_shash() allocates resources, which should be released by
crypto_free_shash(). When ath11k_peer_find() fails, there has memory
leak. Add missing crypto_free_shash() to fix this.

Fixes: 243874c64c81 ("ath11k: handle RX fragments")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20230102081142.3937570-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath9k: Fix potential stack-out-of-bounds write in ath9k_wmi_rsp_callback()

[ Upstream commit 8a2f35b9830692f7a616f2f627f943bc748af13a ]

Fix a stack-out-of-bounds write that occurs in a WMI response callback
function that is called after a timeout occurs in ath9k_wmi_cmd().
The callback writes to wmi->cmd_rsp_buf, a stack-allocated buffer that
could no longer be valid when a timeout occurs. Set wmi->last_seq_id to
0 when a timeout occurred.

Found by a modified version of syzkaller.

BUG: KASAN: stack-out-of-bounds in ath9k_wmi_ctrl_rx
Write of size 4
Call Trace:
 memcpy
 ath9k_wmi_ctrl_rx
 ath9k_htc_rx_msg
 ath9k_hif_usb_reg_in_cb
 __usb_hcd_giveback_urb
 usb_hcd_giveback_urb
 dummy_timer
 call_timer_fn
 run_timer_softirq
 __do_softirq
 irq_exit_rcu
 sysvec_apic_timer_interrupt

Fixes: fb9987d0f748 ("ath9k_htc: Support for AR9271 chipset.")
Signed-off-by: Minsuk Kang <linuxlovemin@yonsei.ac.kr>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20230104124130.10996-1-linuxlovemin@yonsei.ac.kr
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath9k: hif_usb: clean up skbs if ath9k_hif_usb_rx_stream() fails

[ Upstream commit 0af54343a76263a12dbae7fafb64eb47c4a6ad38 ]

Syzkaller detected a memory leak of skbs in ath9k_hif_usb_rx_stream().
While processing skbs in ath9k_hif_usb_rx_stream(), the already allocated
skbs in skb_pool are not freed if ath9k_hif_usb_rx_stream() fails. If we
have an incorrect pkt_len or pkt_tag, the input skb is considered invalid
and dropped. All the associated packets already in skb_pool should be
dropped and freed. Added a comment describing this issue.

The patch also makes remain_skb NULL after being processed so that it
cannot be referenced after potential free. The initialization of hif_dev
fields which are associated with remain_skb (rx_remain_len,
rx_transfer_len and rx_pad_len) is moved after a new remain_skb is
allocated.

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

Fixes: 6ce708f54cc8 ("ath9k: Fix out-of-bound memcpy in ath9k_hif_usb_rx_stream")
Fixes: 44b23b488d44 ("ath9k: hif_usb: Reduce indent 1 column")
Reported-by: syzbot+e9632e3eb038d93d6bc6@syzkaller.appspotmail.com
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20230104123615.51511-1-pchelkin@ispras.ru
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath9k: htc_hst: free skb in ath9k_htc_rx_msg() if there is no callback function

[ Upstream commit 9b25e3985477ac3f02eca5fc1e0cc6850a3f7e69 ]

It is stated that ath9k_htc_rx_msg() either frees the provided skb or
passes its management to another callback function. However, the skb is
not freed in case there is no another callback function, and Syzkaller was
able to cause a memory leak. Also minor comment fix.

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

Fixes: fb9987d0f748 ("ath9k_htc: Support for AR9271 chipset.")
Reported-by: syzbot+e008dccab31bd3647609@syzkaller.appspotmail.com
Reported-by: syzbot+6692c72009680f7c4eb2@syzkaller.appspotmail.com
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20230104123546.51427-1-pchelkin@ispras.ru
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/imx_sc_thermal: Fix the loop condition

[ Upstream commit 4b26b7c9cdefdcb478047c30901d2379ef9e50fc ]

The minimal resource ID is 0: IMX_SC_R_AP_0=0, so fix
the loop condition. Aside of this - constify the array.

Fixes: 31fd4b9db13b ("thermal/drivers/imx_sc: Rely on the platform data to get the resource id")
Signed-off-by: Viorel Suman <viorel.suman@nxp.com>
Reviewed-by: Dong Aisheng <Aisheng.dong@nxp.com>
Link: https://lore.kernel.org/r/20230117091956.61729-1-viorel.suman@oss.nxp.com
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/imx_sc_thermal: Drop empty platform remove function

[ Upstream commit 5011a110295d25418f5918a6af7bfcdb00dd4e34 ]

A remove callback just returning 0 is equivalent to no remove callback
at all. So drop the useless function.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20221212220217.3777176-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Stable-dep-of: 4b26b7c9cdef ("thermal/drivers/imx_sc_thermal: Fix the loop condition")
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: orinoco: check return value of hermes_write_wordrec()

[ Upstream commit 1e346cbb096a5351a637ec1992beffbf330547f0 ]

There is currently no return check for writing an authentication
type (HERMES_AUTH_SHARED_KEY or HERMES_AUTH_OPEN). It looks like
it was accidentally skipped.

This patch adds a return check similar to the other checks in
__orinoco_hw_setup_enc() for hermes_write_wordrec().

Detected using the static analysis tool - Svace.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Alexey Kodanev <aleksei.kodanev@bell-sw.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20221227133306.201356-1-aleksei.kodanev@bell-sw.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rtl8xxxu: Fix memory leaks with RTL8723BU, RTL8192EU

[ Upstream commit b39f662ce1648db0b9de32e6a849b098480793cb ]

The wifi + bluetooth combo chip RTL8723BU can leak memory (especially?)
when it's connected to a bluetooth audio device. The busy bluetooth
traffic generates lots of C2H (card to host) messages, which are not
freed correctly.

To fix this, move the dev_kfree_skb() call in rtl8xxxu_c2hcmd_callback()
inside the loop where skb_dequeue() is called.

The RTL8192EU leaks memory because the C2H messages are added to the
queue and left there forever. (This was fine in the past because it
probably wasn't sending any C2H messages until commit e542e66b7c2e
("wifi: rtl8xxxu: gen2: Turn on the rate control"). Since that commit
it sends a C2H message when the TX rate changes.)

To fix this, delete the check for rf_paths > 1 and the goto. Let the
function process the C2H messages from RTL8192EU like the ones from
the other chips.

Theoretically the RTL8188FU could also leak like RTL8723BU, but it
most likely doesn't send C2H messages frequently enough.

This change was tested with RTL8723BU by Erhard F. I tested it with
RTL8188FU and RTL8192EU.

Reported-by: Erhard F. <erhard_f@mailbox.org>
Tested-by: Erhard F. <erhard_f@mailbox.org>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=215197
Fixes: e542e66b7c2e ("rtl8xxxu: add bluetooth co-existence support for single antenna")
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Reviewed-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/03b099c1-c671-d252-36f4-57b70d721f9d@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rtw89: Add missing check for alloc_workqueue

[ Upstream commit ed9e6166eb0984b718facb7ca59296098cc3aa64 ]

Add check for the return value of alloc_workqueue since it may return
NULL pointer.
Moreover, add destroy_workqueue when rtw89_load_firmware fails.

Fixes: e3ec7017f6a2 ("rtw89: add Realtek 802.11ax driver")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230104142901.1611-1-jiasheng@iscas.ac.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rtw89: fix potential leak in rtw89_append_probe_req_ie()

[ Upstream commit 4a0e218cc9c42d1903ade8b5a371dcf48cf918c5 ]

Do `kfree_skb(new)` before `goto out` to prevent potential leak.

Fixes: 895907779752 ("rtw89: 8852a: add ieee80211_ops::hw_scan")
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230103141054.17372-1-pkshih@realtek.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/tsens: limit num_sensors to 9 for msm8939

[ Upstream commit 903238a33c116edf5f64f7a3fd246e6169cccfa6 ]

On msm8939 last (hwid=10) sensor was added in the hw revision 3.0.
Calibration data for it was placed outside of the main calibration data
blob, so it is not accessible by the current blob-parsing code.

Moreover data for the sensor's p2 is not contiguous in the fuses. This
makes it hard to use nvmem_cell API to parse calibration data in a
generic way.

Since the sensor doesn't seem to be actually used by the existing
hardware, disable the sensor for now.

Fixes: 332bc8ebab2c ("thermal: qcom: tsens-v0_1: Add support for MSM8939")
Cc: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Cc: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Link: https://lore.kernel.org/r/20230101194034.831222-9-dmitry.baryshkov@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/tsens: fix slope values for msm8939

[ Upstream commit 5aec3b035e0cbf3f042c2a03d654e5ad6748feb7 ]

According to the vendor kernels (msm-3.10, 3.14 and 3.18), msm8939
uses non-standard slope values for calibrating the sensors. Fill them
accordingly.

Fixes: 332bc8ebab2c ("thermal: qcom: tsens-v0_1: Add support for MSM8939")
Cc: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Cc: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Link: https://lore.kernel.org/r/20230101194034.831222-8-dmitry.baryshkov@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/tsens: Sort out msm8976 vs msm8956 data

[ Upstream commit a7d3006be5ca7b04e4b84b5ceaae55a700e511bd ]

Tsens driver mentions that msm8976 data should be used for both msm8976
and msm8956 SoCs. This is not quite correct, as according to the
vendor kernels, msm8976 should use standard slope values (3200), while
msm8956 really uses the slope values found in the driver.

Add separate compatibility string for msm8956, move slope value
overrides to the corresponding init function and use the standard
compute_intercept_slope() function for both platforms.

Fixes: 0e580290170d ("thermal: qcom: tsens-v1: Add support for MSM8956 and MSM8976")
Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20230101194034.831222-7-dmitry.baryshkov@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/tsens: Drop msm8976-specific defines

[ Upstream commit 3bf0ea99e2e32b0335106b86d84404cc85bcd113 ]

Drop msm8976-specific defines, which duplicate generic ones.

Fixes: 0e580290170d ("thermal: qcom: tsens-v1: Add support for MSM8956 and MSM8976")
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20230101194034.831222-6-dmitry.baryshkov@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/signal: Fix the value returned by strict_sas_size()

[ Upstream commit ef6dfc4b238a435ab552207ec09e4a82b6426d31 ]

Functions used with __setup() return 1 when the argument has been
successfully parsed.

Reverse the returned value so that 1 is returned when kstrtobool() is
successful (i.e. returns 0).

My understanding of these __setup() functions is that returning 1 or 0
does not change much anyway - so this is more of a cleanup than a
functional fix.

I spot it and found it spurious while looking at something else.
Even if the output is not perfect, you'll get the idea with:

   $ git grep -B2 -A10 retu.*kstrtobool | grep __setup -B10

Fixes: 3aac3ebea08f ("x86/signal: Implement sigaltstack size validation")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/73882d43ebe420c9d8fb82d0560021722b243000.1673717552.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/vfio-ap: fix an error handling path in vfio_ap_mdev_probe_queue()

[ Upstream commit 08866d34c7099e981918d34aab5d6a437436628f ]

The commit in Fixes: has switch the order of a sysfs_create_group() and a
kzalloc().

It correctly removed the now useless kfree() but forgot to add a
sysfs_remove_group() in case of (unlikely) memory allocation failure.

Add it now.

Fixes: 260f3ea14138 ("s390/vfio-ap: move probe and remove callbacks to vfio_ap_ops.c")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Link: https://lore.kernel.org/r/d0c0a35eec4fa87cb7f3910d8ac4dc0f7dc9008a.1659283738.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

s390/early: fix sclp_early_sccb variable lifetime

[ Upstream commit 639886b71ddef085a0e7bb1f225b8ae3eda5c06f ]

Commit ada1da31ce34 ("s390/sclp: sort out physical vs
virtual pointers usage") fixed the notion of virtual
address for sclp_early_sccb pointer. However, it did
not take into account that kasan_early_init() can also
output messages and sclp_early_sccb should be adjusted
by the time kasan_early_init() is called.

Currently it is not a problem, since virtual and physical
addresses on s390 are the same. Nevertheless, should they
ever differ, this would cause an invalid pointer access.

Fixes: ada1da31ce34 ("s390/sclp: sort out physical vs virtual pointers usage")
Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

workqueue: Protects wq_unbound_cpumask with wq_pool_attach_mutex

[ Upstream commit 99c621ef243bda726fb8d982a274ded96570b410 ]

When unbind_workers() reads wq_unbound_cpumask to set the affinity of
freshly-unbound kworkers, it only holds wq_pool_attach_mutex. This isn't
sufficient as wq_unbound_cpumask is only protected by wq_pool_mutex.

Make wq_unbound_cpumask protected with wq_pool_attach_mutex and also
remove the need of temporary saved_cpumask.

Fixes: 10a5a651e3af ("workqueue: Restrict kworker in the offline CPU pool running on housekeeping CPUs")
Reported-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Lai Jiangshan <jiangshan.ljs@antgroup.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

kselftest/arm64: Fix syscall-abi for systems without 128 bit SME

[ Upstream commit 97ec597b26df774a257e3f8e97353fd1b4471615 ]

SME does not mandate any specific VL so we may not have 128 bit SME but
the algorithm used for enumerating VLs assumes that we will. Add the
required check to ensure that the algorithm terminates.

Fixes: 43e3f85523e4 ("kselftest/arm64: Add SME support to syscall ABI test")
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20221223-arm64-syscall-abi-sme-only-v1-1-4fabfbd62087@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64/cpufeature: Fix field sign for DIT hwcap detection

[ Upstream commit 50daf5b7c4ec4efcaf49a4128930f872bec7dbc0 ]

Since it was added our hwcap for DIT has specified that DIT is a signed
field but this appears to be incorrect, the two values for the enumeration
are:

0b0000 NI
0b0001 IMP

which look like a normal unsigned enumeration and the in-kernel DIT usage
added by 01ab991fc0ee ("arm64: Enable data independent timing (DIT) in the
kernel") detects the feature with an unsigned enum. Fix the hwcap to specify
the field as unsigned.

Fixes: 7206dc93a58f ("arm64: Expose Arm v8.4 features")
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20221207-arm64-sysreg-helpers-v3-1-0d71a7b174a8@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/xsk: print correct error codes when exiting

[ Upstream commit 085dcccfb7d3dc52ed708fc588587f319541bc83 ]

Print the correct error codes when exiting the test suite due to some
terminal error. Some of these had a switched sign and some of them
printed zero instead of errno.

Fixes: facb7cb2e909 ("selftests/bpf: Xsk selftests - SKB POLL, NOPOLL")
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20230111093526.11682-5-magnus.karlsson@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/xsk: print correct payload for packet dump

[ Upstream commit 2d0b2ae2871ae6d42a9f0a4280e0fb5bff8d38b8 ]

Print the correct payload when the packet dump option is selected. The
network to host conversion was forgotten and the payload was
erronously declared to be an int instead of an unsigned int.

Fixes: facb7cb2e909 ("selftests/bpf: Xsk selftests - SKB POLL, NOPOLL")
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20230111093526.11682-2-magnus.karlsson@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPICA: nsrepair: handle cases without a return value correctly

[ Upstream commit ca843a4c79486e99a19b859ef0b9887854afe146 ]

Previously acpi_ns_simple_repair() would crash if expected_btypes
contained any combination of ACPI_RTYPE_NONE with a different type,
e.g | ACPI_RTYPE_INTEGER because of slightly incorrect logic in the
!return_object branch, which wouldn't return AE_AML_NO_RETURN_VALUE
for such cases.

Found by Linux Verification Center (linuxtesting.org) with the SVACE
static analysis tool.

Link: https://github.com/acpica/acpica/pull/811
Fixes: 61db45ca2163 ("ACPICA: Restore code that repairs NULL package elements in return values.")
Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/chrome: cros_ec_typec: Update port DP VDO

[ Upstream commit 8d2b28df6c3dc1581d856f52d9f78059ef2a568f ]

The port advertising DP support is a Type-C receptacle. Fix the port's
DisplayPort VDO to reflect this.

Fixes: 1903adae0464 ("platform/chrome: cros_ec_typec: Add bit offset for DP VDO")
Signed-off-by: Prashant Malani <pmalani@chromium.org>
Reviewed-by: Benson Leung <bleung@chromium.org>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20221228004648.793339-6-pmalani@chromium.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: ccp - Avoid page allocation failure warning for SEV_GET_ID2

[ Upstream commit 91dfd98216d817ec5f1c55890bacb7b4fe9b068a ]

For SEV_GET_ID2, the user provided length does not have a specified
limitation because the length of the ID may change in the future.  The
kernel memory allocation, however, is implicitly limited to 4MB on x86 by
the page allocator, otherwise the kzalloc() will fail.

When this happens, it is best not to spam the kernel log with the warning.
Simply fail the allocation and return ENOMEM to the user.

Fixes: d6112ea0cb34 ("crypto: ccp - introduce SEV_GET_ID2 command")
Reported-by: Andy Nguyen <theflow@google.com>
Reported-by: Peter Gonda <pgonda@google.com>
Suggested-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David Rientjes <rientjes@google.com>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

lib/mpi: Fix buffer overrun when SG is too long

[ Upstream commit 7361d1bc307b926cbca214ab67b641123c2d6357 ]

The helper mpi_read_raw_from_sgl sets the number of entries in
the SG list according to nbytes.  However, if the last entry
in the SG list contains more data than nbytes, then it may overrun
the buffer because it only allocates enough memory for nbytes.

Fixes: 2d4d1eea540b ("lib/mpi: Add mpi sgl helpers")
Reported-by: Roberto Sassu <roberto.sassu@huaweicloud.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu-tasks: Fix synchronize_rcu_tasks() VS zap_pid_ns_processes()

[ Upstream commit 28319d6dc5e2ffefa452c2377dd0f71621b5bff0 ]

RCU Tasks and PID-namespace unshare can interact in do_exit() in a
complicated circular dependency:

1) TASK A calls unshare(CLONE_NEWPID), this creates a new PID namespace
   that every subsequent child of TASK A will belong to. But TASK A
   doesn't itself belong to that new PID namespace.

2) TASK A forks() and creates TASK B. TASK A stays attached to its PID
   namespace (let's say PID_NS1) and TASK B is the first task belonging
   to the new PID namespace created by unshare()  (let's call it PID_NS2).

3) Since TASK B is the first task attached to PID_NS2, it becomes the
   PID_NS2 child reaper.

4) TASK A forks() again and creates TASK C which get attached to PID_NS2.
   Note how TASK C has TASK A as a parent (belonging to PID_NS1) but has
   TASK B (belonging to PID_NS2) as a pid_namespace child_reaper.

5) TASK B exits and since it is the child reaper for PID_NS2, it has to
   kill all other tasks attached to PID_NS2, and wait for all of them to
   die before getting reaped itself (zap_pid_ns_process()).

6) TASK A calls synchronize_rcu_tasks() which leads to
   synchronize_srcu(&tasks_rcu_exit_srcu).

7) TASK B is waiting for TASK C to get reaped. But TASK B is under a
   tasks_rcu_exit_srcu SRCU critical section (exit_notify() is between
   exit_tasks_rcu_start() and exit_tasks_rcu_finish()), blocking TASK A.

8) TASK C exits and since TASK A is its parent, it waits for it to reap
   TASK C, but it can't because TASK A waits for TASK B that waits for
   TASK C.

Pid_namespace semantics can hardly be changed at this point. But the
coverage of tasks_rcu_exit_srcu can be reduced instead.

The current task is assumed not to be concurrently reapable at this
stage of exit_notify() and therefore tasks_rcu_exit_srcu can be
temporarily relaxed without breaking its constraints, providing a way
out of the deadlock scenario.

[ paulmck: Fix build failure by adding additional declaration. ]

Fixes: 3f95aa81d265 ("rcu: Make TASKS_RCU handle tasks that are almost done exiting")
Reported-by: Pengfei Xu <pengfei.xu@intel.com>
Suggested-by: Boqun Feng <boqun.feng@gmail.com>
Suggested-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Eric W . Biederman <ebiederm@xmission.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>