crypto: hisilicon/zip - save capability registers in probe process

[ Upstream commit 2ff0ad847951d61c2d8b309e1ccefb26c57dcc7b ]

Pre-store the valid value of the zip alg support related capability
register in hisi_zip_qm_init(), which will be called by hisi_zip_probe().
It can reduce the number of capability register queries and avoid
obtaining incorrect values in abnormal scenarios, such as reset failed
and the memory space disabled.

Fixes: db700974b69d ("crypto: hisilicon/zip - support zip capability")
Signed-off-by: Zhiqi Song <songzhiqi1@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/sec2 - save capability registers in probe process

[ Upstream commit f1115b0096c3163592e04e74f5a7548c25bda957 ]

Pre-store the valid value of the sec alg support related capability
register in sec_qm_init(), which will be called by probe process.
It can reduce the number of capability register queries and avoid
obtaining incorrect values in abnormal scenarios, such as reset
failed and the memory space disabled.

Fixes: 921715b6b782 ("crypto: hisilicon/sec - get algorithm bitmap from registers")
Signed-off-by: Zhiqi Song <songzhiqi1@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/hpre - save capability registers in probe process

[ Upstream commit cf8b5156bbc8c9376f699e8d35e9464b739e33ff ]

Pre-store the valid value of hpre alg support related capability
register in hpre_qm_init(), which will be called by hpre_probe().
It can reduce the number of capability register queries and avoid
obtaining incorrect values in abnormal scenarios, such as reset
failed and the memory space disabled.

Fixes: f214d59a0603 ("crypto: hisilicon/hpre - support hpre capability")
Signed-off-by: Zhiqi Song <songzhiqi1@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/qm - add a function to set qm algs

[ Upstream commit f76f0d7f20672611974d3cc705996751fc403734 ]

Extract a public function to set qm algs and remove
the similar code for setting qm algs in each module.

Signed-off-by: Wenkai Lin <linwenkai6@hisilicon.com>
Signed-off-by: Hao Fang <fanghao11@huawei.com>
Signed-off-by: Zhiqi Song <songzhiqi1@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Stable-dep-of: cf8b5156bbc8 ("crypto: hisilicon/hpre - save capability registers in probe process")
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/zip - add zip comp high perf mode configuration

[ Upstream commit a9864bae1806499ebf3757a9e71dddde5b9c48c6 ]

To meet specific application scenarios, the function of switching between
the high performance mode and the high compression mode is added.

Use the perf_mode=0/1 configuration to set the compression high perf mode,
0(default, high compression mode), 1(high performance mode). These two
modes only apply to the compression direction and are compatible with
software algorithm in both directions.

Signed-off-by: Chenghai Huang <huangchenghai2@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Stable-dep-of: cf8b5156bbc8 ("crypto: hisilicon/hpre - save capability registers in probe process")
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/qm - save capability registers in qm init process

[ Upstream commit cabe13d0bd2efb8dd50ed2310f57b33e1a69a0d4 ]

In previous capability register implementation, qm irq related values
were read from capability registers dynamically when needed. But in
abnormal scenario, e.g. the core is timeout and the device needs to
soft reset and reset failed after disabling the MSE, the device can
not be removed normally, causing the following call trace:

| Call trace:
        |  pci_irq_vector+0xfc/0x140
        |  hisi_qm_uninit+0x278/0x3b0 [hisi_qm]
        |  hpre_remove+0x16c/0x1c0 [hisi_hpre]
        |  pci_device_remove+0x6c/0x264
        |  device_release_driver_internal+0x1ec/0x3e0
        |  device_release_driver+0x3c/0x60
        |  pci_stop_bus_device+0xfc/0x22c
        |  pci_stop_and_remove_bus_device+0x38/0x70
        |  pci_iov_remove_virtfn+0x108/0x1c0
        |  sriov_disable+0x7c/0x1e4
        |  pci_disable_sriov+0x4c/0x6c
        |  hisi_qm_sriov_disable+0x90/0x160 [hisi_qm]
        |  hpre_remove+0x1a8/0x1c0 [hisi_hpre]
        |  pci_device_remove+0x6c/0x264
        |  device_release_driver_internal+0x1ec/0x3e0
        |  driver_detach+0x168/0x2d0
        |  bus_remove_driver+0xc0/0x230
        |  driver_unregister+0x58/0xdc
        |  pci_unregister_driver+0x40/0x220
        |  hpre_exit+0x34/0x64 [hisi_hpre]
        |  __arm64_sys_delete_module+0x374/0x620
        [...]

        | Call trace:
        |  free_msi_irqs+0x25c/0x300
        |  pci_disable_msi+0x19c/0x264
        |  pci_free_irq_vectors+0x4c/0x70
        |  hisi_qm_pci_uninit+0x44/0x90 [hisi_qm]
        |  hisi_qm_uninit+0x28c/0x3b0 [hisi_qm]
        |  hpre_remove+0x16c/0x1c0 [hisi_hpre]
        |  pci_device_remove+0x6c/0x264
        [...]

The reason for this call trace is that when the MSE is disabled, the value
of capability registers in the BAR space become invalid. This will make the
subsequent unregister process get the wrong irq vector through capability
registers and get the wrong irq number by pci_irq_vector().

So add a capability table structure to pre-store the valid value of the irq
information capability register in qm init process, avoid obtaining invalid
capability register value after the MSE is disabled.

Fixes: 3536cc55cada ("crypto: hisilicon/qm - support get device irq information from hardware registers")
Signed-off-by: Zhiqi Song <songzhiqi1@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sahara - fix error handling in sahara_hw_descriptor_create()

[ Upstream commit ee6e6f0a7f5b39d50a5ef5fcc006f4f693db18a7 ]

Do not call dma_unmap_sg() for scatterlists that were not mapped
successfully.

Fixes: 5de8875281e1 ("crypto: sahara - Add driver for SAHARA2 accelerator.")
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sahara - fix processing requests with cryptlen < sg->length

[ Upstream commit 5b8668ce3452827d27f8c34ff6ba080a8f983ed0 ]

It's not always the case that the entire sg entry needs to be processed.
Currently, when cryptlen is less than sg->legth, "Descriptor length" errors
are encountered.

The error was noticed when testing xts(sahara-ecb-aes) with arbitrary sized
input data. To fix this, take the actual request size into account when
populating the hw links.

Fixes: 5de8875281e1 ("crypto: sahara - Add driver for SAHARA2 accelerator.")
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sahara - fix ahash selftest failure

[ Upstream commit afffcf3db98b9495114b79d5381f8cc3f69476fb ]

update() calls should not modify the result buffer, so add an additional
check for "rctx->last" to make sure that only the final hash value is
copied into the buffer.

Fixes the following selftest failure:
alg: ahash: sahara-sha256 update() used result buffer on test vector 3,
cfg="init+update+final aligned buffer"

Fixes: 5a2bb93f5992 ("crypto: sahara - add support for SHA1/256")
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sahara - fix cbc selftest failure

[ Upstream commit 9f10bc28c0fb676ae58aa3bfa358db8f5de124bb ]

The kernel crypto API requires that all CBC implementations update the IV
buffer to contain the last ciphertext block.

This fixes the following cbc selftest error:
alg: skcipher: sahara-cbc-aes encryption test failed (wrong output IV) on
test vector 0, cfg="in-place (one sglist)"

Fixes: 5de8875281e1 ("crypto: sahara - Add driver for SAHARA2 accelerator.")
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sahara - remove FLAGS_NEW_KEY logic

[ Upstream commit 8fd183435728b139248a77978ea3732039341779 ]

Remove the FLAGS_NEW_KEY logic as it has the following issues:
- the wrong key may end up being used when there are multiple data streams:
       t1            t2
    setkey()
    encrypt()
                   setkey()
                   encrypt()

    encrypt() <--- key from t2 is used
- switching between encryption and decryption with the same key is not
  possible, as the hdr flags are only updated when a new setkey() is
  performed

With this change, the key is always sent along with the cryptdata when
performing encryption/decryption operations.

Fixes: 5de8875281e1 ("crypto: sahara - Add driver for SAHARA2 accelerator.")
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: safexcel - Add error handling for dma_map_sg() calls

[ Upstream commit 87e02063d07708cac5bfe9fd3a6a242898758ac8 ]

Macro dma_map_sg() may return 0 on error. This patch enables
checks in case of the macro failure and ensures unmapping of
previously mapped buffers with dma_unmap_sg().

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

Fixes: 49186a7d9e46 ("crypto: inside_secure - Avoid dma map if size is zero")
Signed-off-by: Nikita Zhandarovich <n.zhandarovich@fintech.ru>
Reviewed-by: Antoine Tenart <atenart@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: af_alg - Disallow multiple in-flight AIO requests

[ Upstream commit 67b164a871af1d736f131fd6fe78a610909f06f3 ]

Having multiple in-flight AIO requests results in unpredictable
output because they all share the same IV.  Fix this by only allowing
one request at a time.

Fixes: 83094e5e9e49 ("crypto: af_alg - add async support to algif_aead")
Fixes: a596999b7ddf ("crypto: algif - change algif_skcipher to be asynchronous")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: ccp - fix memleak in ccp_init_dm_workarea

[ Upstream commit a1c95dd5bc1d6a5d7a75a376c2107421b7d6240d ]

When dma_map_single() fails, wa->address is supposed to be freed
by the callers of ccp_init_dm_workarea() through ccp_dm_free().
However, many of the call spots don't expect to have to call
ccp_dm_free() on failure of ccp_init_dm_workarea(), which may
lead to a memleak. Let's free wa->address in ccp_init_dm_workarea()
when dma_map_single() fails.

Fixes: 63b945091a07 ("crypto: ccp - CCP device driver and interface support")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
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>

crypto: sa2ul - Return crypto_aead_setkey to transfer the error

[ Upstream commit ce852f1308ac738e61c5b2502517deea593a1554 ]

Return crypto_aead_setkey() in order to transfer the error if
it fails.

Fixes: d2c8ac187fc9 ("crypto: sa2ul - Add AEAD algorithm support")
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: virtio - Handle dataq logic with tasklet

[ Upstream commit fed93fb62e05c38152b0fc1dc9609639e63eed76 ]

Doing ipsec produces a spinlock recursion warning.
This is due to crypto_finalize_request() being called in the upper half.
Move virtual data queue processing of virtio-crypto driver to tasklet.

Fixes: dbaf0624ffa57 ("crypto: add virtio-crypto driver")
Reported-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: wangyangxin <wangyangxin1@huawei.com>
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: jh7110 - Correct deferred probe return

[ Upstream commit d57343022b71b9f41e731282dbe0baf0cff6ada8 ]

This fixes list_add corruption error when the driver is returned
with -EPROBE_DEFER. It is also required to roll back the previous
probe sequences in case of deferred_probe. So, this removes
'err_probe_defer" goto label and just use err_dma_init instead.

Fixes: 42ef0e944b01 ("crypto: starfive - Add crypto engine support")
Signed-off-by: Chanho Park <chanho61.park@samsung.com>
Reviewed-by: Jia Jie Ho <jiajie.ho@starfivetech.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: rsa - add a check for allocation failure

[ Upstream commit d872ca165cb67112f2841ef9c37d51ef7e63d1e4 ]

Static checkers insist that the mpi_alloc() allocation can fail so add
a check to prevent a NULL dereference.  Small allocations like this
can't actually fail in current kernels, but adding a check is very
simple and makes the static checkers happy.

Fixes: 6637e11e4ad2 ("crypto: rsa - allow only odd e and restrict value in FIPS mode")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selinux: Fix error priority for bind with AF_UNSPEC on PF_INET6 socket

[ Upstream commit bbf5a1d0e5d0fb3bdf90205aa872636122692a50 ]

The IPv6 network stack first checks the sockaddr length (-EINVAL error)
before checking the family (-EAFNOSUPPORT error).

This was discovered thanks to commit a549d055a22e ("selftests/landlock:
Add network tests").

Cc: Eric Paris <eparis@parisplace.org>
Cc: Konstantin Meskhidze <konstantin.meskhidze@huawei.com>
Cc: Paul Moore <paul@paul-moore.com>
Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
Reported-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Closes: https://lore.kernel.org/r/0584f91c-537c-4188-9e4f-04f192565667@collabora.com
Fixes: 0f8db8cc73df ("selinux: add AF_UNSPEC and INADDR_ANY checks to selinux_socket_bind()")
Signed-off-by: Mickaël Salaün <mic@digikod.net>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drivers/thermal/loongson2_thermal: Fix incorrect PTR_ERR() judgment

[ Upstream commit 15ef92e9c41124ee9d88b01208364f3fe1f45f84 ]

PTR_ERR() returns -ENODEV when thermal-zones are undefined, and we need
-ENODEV as the right value for comparison.

Otherwise, tz->type is NULL when thermal-zones is undefined, resulting
in the following error:

[   12.290030] CPU 1 Unable to handle kernel paging request at virtual address fffffffffffffff1, era == 900000000355f410, ra == 90000000031579b8
[   12.302877] Oops[#1]:
[   12.305190] CPU: 1 PID: 181 Comm: systemd-udevd Not tainted 6.6.0-rc7+ #5385
[   12.312304] pc 900000000355f410 ra 90000000031579b8 tp 90000001069e8000 sp 90000001069eba10
[   12.320739] a0 0000000000000000 a1 fffffffffffffff1 a2 0000000000000014 a3 0000000000000001
[   12.329173] a4 90000001069eb990 a5 0000000000000001 a6 0000000000001001 a7 900000010003431c
[   12.337606] t0 fffffffffffffff1 t1 54567fd5da9b4fd4 t2 900000010614ec40 t3 00000000000dc901
[   12.346041] t4 0000000000000000 t5 0000000000000004 t6 900000010614ee20 t7 900000000d00b790
[   12.354472] t8 00000000000dc901 u0 54567fd5da9b4fd4 s9 900000000402ae10 s0 900000010614ec40
[   12.362916] s1 90000000039fced0 s2 ffffffffffffffed s3 ffffffffffffffed s4 9000000003acc000
[   12.362931] s5 0000000000000004 s6 fffffffffffff000 s7 0000000000000490 s8 90000001028b2ec8
[   12.362938]    ra: 90000000031579b8 thermal_add_hwmon_sysfs+0x258/0x300
[   12.386411]   ERA: 900000000355f410 strscpy+0xf0/0x160
[   12.391626]  CRMD: 000000b0 (PLV0 -IE -DA +PG DACF=CC DACM=CC -WE)
[   12.397898]  PRMD: 00000004 (PPLV0 +PIE -PWE)
[   12.403678]  EUEN: 00000000 (-FPE -SXE -ASXE -BTE)
[   12.409859]  ECFG: 00071c1c (LIE=2-4,10-12 VS=7)
[   12.415882] ESTAT: 00010000 [PIL] (IS= ECode=1 EsubCode=0)
[   12.415907]  BADV: fffffffffffffff1
[   12.415911]  PRID: 0014a000 (Loongson-64bit, Loongson-2K1000)
[   12.415917] Modules linked in: loongson2_thermal(+) vfat fat uio_pdrv_genirq uio fuse zram zsmalloc
[   12.415950] Process systemd-udevd (pid: 181, threadinfo=00000000358b9718, task=00000000ace72fe3)
[   12.415961] Stack : 0000000000000dc0 54567fd5da9b4fd4 900000000402ae10 9000000002df9358
[   12.415982]         ffffffffffffffed 0000000000000004 9000000107a10aa8 90000001002a3410
[   12.415999]         ffffffffffffffed ffffffffffffffed 9000000107a11268 9000000003157ab0
[   12.416016]         9000000107a10aa8 ffffff80020fc0c8 90000001002a3410 ffffffffffffffed
[   12.416032]         0000000000000024 ffffff80020cc1e8 900000000402b2a0 9000000003acc000
[   12.416048]         90000001002a3410 0000000000000000 ffffff80020f4030 90000001002a3410
[   12.416065]         0000000000000000 9000000002df6808 90000001002a3410 0000000000000000
[   12.416081]         ffffff80020f4030 0000000000000000 90000001002a3410 9000000002df2ba8
[   12.416097]         00000000000000b4 90000001002a34f4 90000001002a3410 0000000000000002
[   12.416114]         ffffff80020f4030 fffffffffffffff0 90000001002a3410 9000000002df2f30
[   12.416131]         ...
[   12.416138] Call Trace:
[   12.416142] [<900000000355f410>] strscpy+0xf0/0x160
[   12.416167] [<90000000031579b8>] thermal_add_hwmon_sysfs+0x258/0x300
[   12.416183] [<9000000003157ab0>] devm_thermal_add_hwmon_sysfs+0x50/0xe0
[   12.416200] [<ffffff80020cc1e8>] loongson2_thermal_probe+0x128/0x200 [loongson2_thermal]
[   12.416232] [<9000000002df6808>] platform_probe+0x68/0x140
[   12.416249] [<9000000002df2ba8>] really_probe+0xc8/0x3c0
[   12.416269] [<9000000002df2f30>] __driver_probe_device+0x90/0x180
[   12.416286] [<9000000002df3058>] driver_probe_device+0x38/0x160
[   12.416302] [<9000000002df33a8>] __driver_attach+0xa8/0x200
[   12.416314] [<9000000002deffec>] bus_for_each_dev+0x8c/0x120
[   12.416330] [<9000000002df198c>] bus_add_driver+0x10c/0x2a0
[   12.416346] [<9000000002df46b4>] driver_register+0x74/0x160
[   12.416358] [<90000000022201a4>] do_one_initcall+0x84/0x220
[   12.416372] [<90000000022f3ab8>] do_init_module+0x58/0x2c0
[   12.416386] [<90000000022f6538>] init_module_from_file+0x98/0x100
[   12.416399] [<90000000022f67f0>] sys_finit_module+0x230/0x3c0
[   12.416412] [<900000000358f7c8>] do_syscall+0x88/0xc0
[   12.416431] [<900000000222137c>] handle_syscall+0xbc/0x158

Fixes: e7e3a7c35791 ("thermal/drivers/loongson-2: Add thermal management support")
Cc: Yinbo Zhu <zhuyinbo@loongson.cn>
Signed-off-by: Binbin Zhou <zhoubinbin@loongson.cn>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/343c14de98216636a47b43e8bfd47b70d0a8e068.1700817227.git.zhoubinbin@loongson.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpuidle: haltpoll: Do not enable interrupts when entering idle

[ Upstream commit c8f5caec3df84a02b937d6d9cda1f7ffa8dc443f ]

The cpuidle drivers' ->enter() methods are supposed to be IRQ invariant:

  5e26aa933911 ("cpuidle/poll: Ensure IRQs stay disabled after cpuidle_state::enter() calls")
  bb7b11258561 ("cpuidle: Move IRQ state validation")

Do that in the haltpoll driver too.

Fixes: 5e26aa933911 ("cpuidle/poll: Ensure IRQs stay disabled after cpuidle_state::enter() calls")
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218245
Reported-by: <forza@tnonline.net>
Tested-by: <forza@tnonline.net>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mtd: Fix gluebi NULL pointer dereference caused by ftl notifier

[ Upstream commit a43bdc376deab5fff1ceb93dca55bcab8dbdc1d6 ]

If both ftl.ko and gluebi.ko are loaded, the notifier of ftl
triggers NULL pointer dereference when trying to access
‘gluebi->desc’ in gluebi_read().

ubi_gluebi_init
  ubi_register_volume_notifier
    ubi_enumerate_volumes
      ubi_notify_all
        gluebi_notify    nb->notifier_call()
          gluebi_create
            mtd_device_register
              mtd_device_parse_register
                add_mtd_device
                  blktrans_notify_add   not->add()
                    ftl_add_mtd         tr->add_mtd()
                      scan_header
                        mtd_read
                          mtd_read_oob
                            mtd_read_oob_std
                              gluebi_read   mtd->read()
                                gluebi->desc - NULL

Detailed reproduction information available at the Link [1],

In the normal case, obtain gluebi->desc in the gluebi_get_device(),
and access gluebi->desc in the gluebi_read(). However,
gluebi_get_device() is not executed in advance in the
ftl_add_mtd() process, which leads to NULL pointer dereference.

The solution for the gluebi module is to run jffs2 on the UBI
volume without considering working with ftl or mtdblock [2].
Therefore, this problem can be avoided by preventing gluebi from
creating the mtdblock device after creating mtd partition of the
type MTD_UBIVOLUME.

Fixes: 2ba3d76a1e29 ("UBI: make gluebi a separate module")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217992
Link: https://lore.kernel.org/lkml/441107100.23734.1697904580252.JavaMail.zimbra@nod.at/
Signed-off-by: ZhaoLong Wang <wangzhaolong1@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Acked-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231220024619.2138625-1-wangzhaolong1@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

kunit: debugfs: Fix unchecked dereference in debugfs_print_results()

[ Upstream commit 34dfd5bb2e5507e69d9b6d6c90f546600c7a4977 ]

Move the call to kunit_suite_has_succeeded() after the check that
the kunit_suite pointer is valid.

This was found by smatch:

 lib/kunit/debugfs.c:66 debugfs_print_results() warn: variable
 dereferenced before check 'suite' (see line 63)

Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Fixes: 38289a26e1b8 ("kunit: fix debugfs code to use enum kunit_status, not bool")
Reviewed-by: Rae Moar <rmoar@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal: core: Fix NULL pointer dereference in zone registration error path

[ Upstream commit 04e6ccfc93c5a1aa1d75a537cf27e418895e20ea ]

If device_register() in thermal_zone_device_register_with_trips()
returns an error, the tz variable is set to NULL and subsequently
dereferenced in kfree(tz->tzp).

Commit adc8749b150c ("thermal/drivers/core: Use put_device() if
device_register() fails") added the tz = NULL assignment in question to
avoid a possible double-free after dropping the reference to the zone
device.  However, after commit 4649620d9404 ("thermal: core: Make
thermal_zone_device_unregister() return after freeing the zone"), that
assignment has become redundant, because dropping the reference to the
zone device does not cause the zone object to be freed any more.

Drop it to address the NULL pointer dereference.

Fixes: 3d439b1a2ad3 ("thermal/core: Alloc-copy-free the thermal zone parameters structure")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: extlog: Clear Extended Error Log status when RAS_CEC handled the error

[ Upstream commit 38c872a9e96f72f2947affc0526cc05659367d3d ]

When both CONFIG_RAS_CEC and CONFIG_ACPI_EXTLOG are enabled, Linux does
not clear the status word of the BIOS supplied error record for corrected
errors. This may prevent logging of subsequent uncorrected errors.

Fix by clearing the status.

Fixes: 23ba710a0864 ("x86/mce: Fix all mce notifiers to update the mce->kflags bitmask")
Reported-by: Erwin Tsaur <erwin.tsaur@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: LPSS: Fix the fractional clock divider flags

[ Upstream commit 3ebccf1d1ca74bbb78e6f8c38d1d172e468d91f8 ]

The conversion to CLK_FRAC_DIVIDER_POWER_OF_TWO_PS uses wrong flags
in the parameters and hence miscalculates the values in the clock
divider. Fix this by applying the flag to the proper parameter.

Fixes: 82f53f9ee577 ("clk: fractional-divider: Introduce POWER_OF_TWO_PS flag")
Reported-by: Alex Vinarskis <alex.vinarskis@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: sh-msiof: Enforce fixed DTDL for R-Car H3

[ Upstream commit e5c7bcb499840551cfbe85c6df177ebc50432bf0 ]

Documentation says only DTDL of 200 is allowed for this SoC.

Fixes: 4286db8456f4 ("spi: sh-msiof: Add R-Car Gen 2 and 3 fallback bindings")
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://msgid.link/r/20231212081239.14254-1-wsa+renesas@sang-engineering.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

efivarfs: Free s_fs_info on unmount

[ Upstream commit 547713d502f7b4b8efccd409cff84d731a23853b ]

Now that we allocate a s_fs_info struct on fs context creation, we
should ensure that we free it again when the superblock goes away.

Fixes: 5329aa5101f7 ("efivarfs: Add uid/gid mount options")
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

efivarfs: force RO when remounting if SetVariable is not supported

[ Upstream commit 0e8d2444168dd519fea501599d150e62718ed2fe ]

If SetVariable at runtime is not supported by the firmware we never assign
a callback for that function. At the same time mount the efivarfs as
RO so no one can call that.  However, we never check the permission flags
when someone remounts the filesystem as RW. As a result this leads to a
crash looking like this:

$ mount -o remount,rw /sys/firmware/efi/efivars
$ efi-updatevar -f PK.auth PK

[  303.279166] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
[  303.280482] Mem abort info:
[  303.280854]   ESR = 0x0000000086000004
[  303.281338]   EC = 0x21: IABT (current EL), IL = 32 bits
[  303.282016]   SET = 0, FnV = 0
[  303.282414]   EA = 0, S1PTW = 0
[  303.282821]   FSC = 0x04: level 0 translation fault
[  303.283771] user pgtable: 4k pages, 48-bit VAs, pgdp=000000004258c000
[  303.284913] [0000000000000000] pgd=0000000000000000, p4d=0000000000000000
[  303.286076] Internal error: Oops: 0000000086000004 [#1] PREEMPT SMP
[  303.286936] Modules linked in: qrtr tpm_tis tpm_tis_core crct10dif_ce arm_smccc_trng rng_core drm fuse ip_tables x_tables ipv6
[  303.288586] CPU: 1 PID: 755 Comm: efi-updatevar Not tainted 6.3.0-rc1-00108-gc7d0c4695c68 #1
[  303.289748] Hardware name: Unknown Unknown Product/Unknown Product, BIOS 2023.04-00627-g88336918701d 04/01/2023
[  303.291150] pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  303.292123] pc : 0x0
[  303.292443] lr : efivar_set_variable_locked+0x74/0xec
[  303.293156] sp : ffff800008673c10
[  303.293619] x29: ffff800008673c10 x28: ffff0000037e8000 x27: 0000000000000000
[  303.294592] x26: 0000000000000800 x25: ffff000002467400 x24: 0000000000000027
[  303.295572] x23: ffffd49ea9832000 x22: ffff0000020c9800 x21: ffff000002467000
[  303.296566] x20: 0000000000000001 x19: 00000000000007fc x18: 0000000000000000
[  303.297531] x17: 0000000000000000 x16: 0000000000000000 x15: 0000aaaac807ab54
[  303.298495] x14: ed37489f673633c0 x13: 71c45c606de13f80 x12: 47464259e219acf4
[  303.299453] x11: ffff000002af7b01 x10: 0000000000000003 x9 : 0000000000000002
[  303.300431] x8 : 0000000000000010 x7 : ffffd49ea8973230 x6 : 0000000000a85201
[  303.301412] x5 : 0000000000000000 x4 : ffff0000020c9800 x3 : 00000000000007fc
[  303.302370] x2 : 0000000000000027 x1 : ffff000002467400 x0 : ffff000002467000
[  303.303341] Call trace:
[  303.303679]  0x0
[  303.303938]  efivar_entry_set_get_size+0x98/0x16c
[  303.304585]  efivarfs_file_write+0xd0/0x1a4
[  303.305148]  vfs_write+0xc4/0x2e4
[  303.305601]  ksys_write+0x70/0x104
[  303.306073]  __arm64_sys_write+0x1c/0x28
[  303.306622]  invoke_syscall+0x48/0x114
[  303.307156]  el0_svc_common.constprop.0+0x44/0xec
[  303.307803]  do_el0_svc+0x38/0x98
[  303.308268]  el0_svc+0x2c/0x84
[  303.308702]  el0t_64_sync_handler+0xf4/0x120
[  303.309293]  el0t_64_sync+0x190/0x194
[  303.309794] Code: ???????? ???????? ???????? ???????? (????????)
[  303.310612] ---[ end trace 0000000000000000 ]---

Fix this by adding a .reconfigure() function to the fs operations which
we can use to check the requested flags and deny anything that's not RO
if the firmware doesn't implement SetVariable at runtime.

Fixes: f88814cc2578 ("efi/efivars: Expose RT service availability via efivars abstraction")
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

calipso: fix memory leak in netlbl_calipso_add_pass()

[ Upstream commit ec4e9d630a64df500641892f4e259e8149594a99 ]

If IPv6 support is disabled at boot (ipv6.disable=1),
the calipso_init() -> netlbl_calipso_ops_register() function isn't called,
and the netlbl_calipso_ops_get() function always returns NULL.
In this case, the netlbl_calipso_add_pass() function allocates memory
for the doi_def variable but doesn't free it with the calipso_doi_free().

BUG: memory leak
unreferenced object 0xffff888011d68180 (size 64):
  comm "syz-executor.1", pid 10746, jiffies 4295410986 (age 17.928s)
  hex dump (first 32 bytes):
    00 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<...>] kmalloc include/linux/slab.h:552 [inline]
    [<...>] netlbl_calipso_add_pass net/netlabel/netlabel_calipso.c:76 [inline]
    [<...>] netlbl_calipso_add+0x22e/0x4f0 net/netlabel/netlabel_calipso.c:111
    [<...>] genl_family_rcv_msg_doit+0x22f/0x330 net/netlink/genetlink.c:739
    [<...>] genl_family_rcv_msg net/netlink/genetlink.c:783 [inline]
    [<...>] genl_rcv_msg+0x341/0x5a0 net/netlink/genetlink.c:800
    [<...>] netlink_rcv_skb+0x14d/0x440 net/netlink/af_netlink.c:2515
    [<...>] genl_rcv+0x29/0x40 net/netlink/genetlink.c:811
    [<...>] netlink_unicast_kernel net/netlink/af_netlink.c:1313 [inline]
    [<...>] netlink_unicast+0x54b/0x800 net/netlink/af_netlink.c:1339
    [<...>] netlink_sendmsg+0x90a/0xdf0 net/netlink/af_netlink.c:1934
    [<...>] sock_sendmsg_nosec net/socket.c:651 [inline]
    [<...>] sock_sendmsg+0x157/0x190 net/socket.c:671
    [<...>] ____sys_sendmsg+0x712/0x870 net/socket.c:2342
    [<...>] ___sys_sendmsg+0xf8/0x170 net/socket.c:2396
    [<...>] __sys_sendmsg+0xea/0x1b0 net/socket.c:2429
    [<...>] do_syscall_64+0x30/0x40 arch/x86/entry/common.c:46
    [<...>] entry_SYSCALL_64_after_hwframe+0x61/0xc6

Found by InfoTeCS on behalf of Linux Verification Center
(linuxtesting.org) with Syzkaller

Fixes: cb72d38211ea ("netlabel: Initial support for the CALIPSO netlink protocol.")
Signed-off-by: Gavrilov Ilia <Ilia.Gavrilov@infotecs.ru>
[PM: merged via the LSM tree at Jakub Kicinski request]
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cpufreq: scmi: process the result of devm_of_clk_add_hw_provider()

[ Upstream commit c4a5118a3ae1eadc687d84eef9431f9e13eb015c ]

devm_of_clk_add_hw_provider() may return an errno, so
add a return value check

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

Fixes: 8410e7f3b31e ("cpufreq: scmi: Fix OPP addition failure with a dummy clock provider")
Signed-off-by: Alexandra Diupina <adiupina@astralinux.ru>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/x86/intel/vsec: Fix xa_alloc memory leak

[ Upstream commit 8cbcc1dbf8a62c730fadd60de761e0658547a589 ]

Commit 936874b77dd0 ("platform/x86/intel/vsec: Add PCI error recovery
support to Intel PMT") added an xarray to track the list of vsec devices to
be recovered after a PCI error. But it did not provide cleanup for the list
leading to a memory leak that was caught by kmemleak.  Do xa_alloc() before
devm_add_action_or_reset() so that the list may be cleaned up with
xa_erase() in the release function.

Fixes: 936874b77dd0 ("platform/x86/intel/vsec: Add PCI error recovery support to Intel PMT")
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/r/20231129222132.2331261-2-david.e.box@linux.intel.com
[hdegoede@redhat.com: Add missing xa_erase() on error-exit
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: cadence-quadspi: add missing clk_disable_unprepare() in cqspi_probe()

[ Upstream commit 5cb475174cce1bfedf1025b6e235e2c43d81144f ]

cqspi_jh7110_clk_init() is called after clk_prepare_enable(cqspi->clk),
if it fails, it should goto label 'probe_reset_failed' to disable
cqspi->clk.

In the error path after calling cqspi_jh7110_clk_init(),
cqspi_jh7110_disable_clk() need be called.

Fixes: 33f1ef6d4eb6 ("spi: cadence-quadspi: Add clock configuration for StarFive JH7110 QSPI")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20231129081147.628004-1-yangyingliang@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

KEYS: encrypted: Add check for strsep

[ Upstream commit b4af096b5df5dd131ab796c79cedc7069d8f4882 ]

Add check for strsep() in order to transfer the error.

Fixes: cd3bc044af48 ("KEYS: encrypted: Instantiate key with user-provided decrypted data")
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: LPIT: Avoid u32 multiplication overflow

[ Upstream commit 56d2eeda87995245300836ee4dbd13b002311782 ]

In lpit_update_residency() there is a possibility of overflow
in multiplication, if tsc_khz is large enough (> UINT_MAX/1000).

Change multiplication to mul_u32_u32().

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

Fixes: eeb2d80d502a ("ACPI / LPIT: Add Low Power Idle Table (LPIT) support")
Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: video: check for error while searching for backlight device parent

[ Upstream commit ccd45faf4973746c4f30ea41eec864e5cf191099 ]

If acpi_get_parent() called in acpi_video_dev_register_backlight()
fails, for example, because acpi_ut_acquire_mutex() fails inside
acpi_get_parent), this can lead to incorrect (uninitialized)
acpi_parent handle being passed to acpi_get_pci_dev() for detecting
the parent pci device.

Check acpi_get_parent() result and set parent device only in case of success.

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

Fixes: 9661e92c10a9 ("acpi: tie ACPI backlight devices to PCI devices if possible")
Signed-off-by: Nikita Kiryushin <kiryushin@ancud.ru>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mtd: rawnand: Increment IFC_TIMEOUT_MSECS for nand controller response

[ Upstream commit 923fb6238cb3ac529aa2bf13b3b1e53762186a8b ]

Under heavy load it is likely that the controller is done
with its own task but the thread unlocking the wait is not
scheduled in time. Increasing IFC_TIMEOUT_MSECS allows the
controller to respond within allowable timeslice of 1 sec.

fsl,ifc-nand 7e800000.nand: Controller is not responding

[<804b2047>] (nand_get_device) from [<804b5335>] (nand_write_oob+0x1b/0x4a)
[<804b5335>] (nand_write_oob) from [<804a3585>] (mtd_write+0x41/0x5c)
[<804a3585>] (mtd_write) from [<804c1d47>] (ubi_io_write+0x17f/0x22c)
[<804c1d47>] (ubi_io_write) from [<804c047b>] (ubi_eba_write_leb+0x5b/0x1d0)

Fixes: 82771882d960 ("NAND Machine support for Integrated Flash Controller")
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Ronald Monthero <debug.penguin32@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231118083156.776887-1-debug.penguin32@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: spi-zynqmp-gqspi: fix driver kconfig dependencies

[ Upstream commit 424a8166764e462258fdccaaefbdeb07517c8b21 ]

ZynqMP GQSPI driver no longer uses spi-master framework. It had been
converted to use spi-mem framework. So remove driver dependency from
spi-master and replace it with spi-mem.

Fixes: 1c26372e5aa9 ("spi: spi-zynqmp-gqspi: Update driver to use spi-mem framework")
Signed-off-by: Amit Kumar Mahapatra <amit.kumar-mahapatra@amd.com>
Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>
Link: https://lore.kernel.org/r/1699282435-884917-1-git-send-email-radhey.shyam.pandey@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/x86/intel/uncore: Fix NULL pointer dereference issue in upi_fill_topology()

[ Upstream commit 1692cf434ba13ee212495b5af795b6a07e986ce4 ]

Get logical socket id instead of physical id in discover_upi_topology()
to avoid out-of-bound access on 'upi = &type->topology[nid][idx];' line
that leads to NULL pointer dereference in upi_fill_topology()

Fixes: f680b6e6062e ("perf/x86/intel/uncore: Enable UPI topology discovery for Icelake Server")
Reported-by: Kyle Meyer <kyle.meyer@hpe.com>
Signed-off-by: Alexander Antonov <alexander.antonov@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Tested-by: Kyle Meyer <kyle.meyer@hpe.com>
Link: https://lore.kernel.org/r/20231127185246.2371939-2-alexander.antonov@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/fair: Update min_vruntime for reweight_entity() correctly

[ Upstream commit 5068d84054b766efe7c6202fc71b2350d1c326f1 ]

Since reweight_entity() may have chance to change the weight of
cfs_rq->curr entity, we should also update_min_vruntime() if
this is the case

Fixes: eab03c23c2a1 ("sched/eevdf: Fix vruntime adjustment on reweight")
Signed-off-by: Yiwei Lin <s921975628@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Abel Wu <wuyun.abel@bytedance.com>
Link: https://lore.kernel.org/r/20231117080106.12890-1-s921975628@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/imc-pmu: Add a null pointer check in update_events_in_group()

[ Upstream commit 0a233867a39078ebb0f575e2948593bbff5826b3 ]

kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure.

Fixes: 885dcd709ba9 ("powerpc/perf: Add nest IMC PMU support")
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231126093719.1440305-1-chentao@kylinos.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/powernv: Add a null pointer check in opal_powercap_init()

[ Upstream commit e123015c0ba859cf48aa7f89c5016cc6e98e018d ]

kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure.

Fixes: b9ef7b4b867f ("powerpc: Convert to using %pOFn instead of device_node.name")
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231126095739.1501990-1-chentao@kylinos.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/powernv: Add a null pointer check in opal_event_init()

[ Upstream commit 8649829a1dd25199bbf557b2621cedb4bf9b3050 ]

kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure.

Fixes: 2717a33d6074 ("powerpc/opal-irqchip: Use interrupt names if present")
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231127030755.1546750-1-chentao@kylinos.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/powernv: Add a null pointer check to scom_debug_init_one()

[ Upstream commit 9a260f2dd827bbc82cc60eb4f4d8c22707d80742 ]

kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure.
Add a null pointer check, and release 'ent' to avoid memory leaks.

Fixes: bfd2f0d49aef ("powerpc/powernv: Get rid of old scom_controller abstraction")
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231208085937.107210-1-chentao@kylinos.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/rtas: Avoid warning on invalid token argument to sys_rtas()

[ Upstream commit 01e346ffefda3a7088afebf02b940614179688e7 ]

rtas_token_to_function() WARNs when passed an invalid token; it's
meant to catch bugs in kernel-based users of RTAS functions. However,
user space controls the token value passed to rtas_token_to_function()
by block_rtas_call(), so user space with sufficient privilege to use
sys_rtas() can trigger the warnings at will:

  unexpected failed lookup for token 2048
  WARNING: CPU: 20 PID: 2247 at arch/powerpc/kernel/rtas.c:556
    rtas_token_to_function+0xfc/0x110
  ...
  NIP rtas_token_to_function+0xfc/0x110
  LR  rtas_token_to_function+0xf8/0x110
  Call Trace:
    rtas_token_to_function+0xf8/0x110 (unreliable)
    sys_rtas+0x188/0x880
    system_call_exception+0x268/0x530
    system_call_common+0x160/0x2c4

It's desirable to continue warning on bogus tokens in
rtas_token_to_function(). Currently it is used to look up RTAS
function descriptors when tracing, where we know there has to have
been a successful descriptor lookup by different means already, and it
would be a serious inconsistency for the reverse lookup to fail.

So instead of weakening rtas_token_to_function()'s contract by
removing the warnings, introduce rtas_token_to_function_untrusted(),
which has no opinion on failed lookups. Convert block_rtas_call() and
rtas_token_to_function() to use it.

Fixes: 8252b88294d2 ("powerpc/rtas: improve function information lookups")
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231212-papr-sys_rtas-vs-lockdown-v6-1-e9eafd0c8c6c@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/hv-gpci: Add return value check in affinity_domain_via_partition_show function

[ Upstream commit 070b71f428facd9130319707db854ed8bd24637a ]

To access hv-gpci kernel interface files data, the
"Enable Performance Information Collection" option has to be set
in hmc. Incase that option is not set and user try to read
the interface files, it should give error message as
operation not permitted.

Result of accessing added interface files with disabled
performance collection option:

[command]# cat processor_bus_topology
cat: processor_bus_topology: Operation not permitted

[command]# cat processor_config
cat: processor_config: Operation not permitted

[command]# cat affinity_domain_via_domain
cat: affinity_domain_via_domain: Operation not permitted

[command]# cat affinity_domain_via_virtual_processor
cat: affinity_domain_via_virtual_processor: Operation not permitted

[command]# cat affinity_domain_via_partition

Based on above result there is no error message when reading
affinity_domain_via_partition file because of missing
check for failed hcall. Fix this issue by adding
a check in the start of affinity_domain_via_partition_show
function, to return error incase hcall fails, with error type
other then H_PARAMETER.

Fixes: a15e0d6a6929 ("powerpc/hv_gpci: Add sysfs file inside hv_gpci device to show affinity domain via partition information")
Reported-by: Disha Goel <disgoel@linux.vnet.ibm.com>
Signed-off-by: Kajol Jain <kjain@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231116122033.160964-1-kjain@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/powerpc: Fix error handling in FPU/VMX preemption tests

[ Upstream commit 9dbd5927408c4a0707de73ae9dd9306b184e8fee ]

The FPU & VMX preemption tests do not check for errors returned by the
low-level asm routines, preempt_fpu() / preempt_vsx() respectively.
That means any register corruption detected by the asm routines does not
result in a test failure.

Fix it by returning the return value of the asm routines from the
pthread child routines.

Fixes: e5ab8be68e44 ("selftests/powerpc: Test preservation of FPU and VMX regs across preemption")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231128132748.1990179-1-mpe@ellerman.id.au
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: PPC: Book3S HV: Handle pending exceptions on guest entry with MSR_EE

[ Upstream commit ecd10702baae5c16a91d139bde7eff84ce55daee ]

Commit 026728dc5d41 ("KVM: PPC: Book3S HV P9: Inject pending xive
interrupts at guest entry") changed guest entry so that if external
interrupts are enabled, BOOK3S_IRQPRIO_EXTERNAL is not tested for. Test
for this regardless of MSR_EE.

For an L1 host, do not inject an interrupt, but always
use LPCR_MER. If the L0 desires it can inject an interrupt.

Fixes: 026728dc5d41 ("KVM: PPC: Book3S HV P9: Inject pending xive interrupts at guest entry")
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[jpn: use kvmpcc_get_msr(), write commit message]
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231201132618.555031-7-vaibhav@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: PPC: Book3S HV: Introduce low level MSR accessor

[ Upstream commit 6de2e837babb411cfb3cdb570581c3a65576ddaf ]

kvmppc_get_msr() and kvmppc_set_msr_fast() serve as accessors for the
MSR. However because the MSR is kept in the shared regs they include a
conditional check for kvmppc_shared_big_endian() and endian conversion.

Within the Book3S HV specific code there are direct reads and writes of
shregs::msr. In preparation for Nested APIv2 these accesses need to be
replaced with accessor functions so it is possible to extend their
behavior. However, using the kvmppc_get_msr() and kvmppc_set_msr_fast()
functions is undesirable because it would introduce a conditional branch
and endian conversion that is not currently present.

kvmppc_set_msr_hv() already exists, it is used for the
kvmppc_ops::set_msr callback.

Introduce a low level accessor __kvmppc_{s,g}et_msr_hv() that simply
gets and sets shregs::msr. This will be extend for Nested APIv2 support.

Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230914030600.16993-8-jniethe5@gmail.com
Stable-dep-of: ecd10702baae ("KVM: PPC: Book3S HV: Handle pending exceptions on guest entry with MSR_EE")
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: PPC: Book3S HV: Use accessors for VCPU registers

[ Upstream commit ebc88ea7a6ad0ea349df9c765357d3aa4e662aa9 ]

Introduce accessor generator macros for Book3S HV VCPU registers. Use
the accessor functions to replace direct accesses to this registers.

This will be important later for Nested APIv2 support which requires
additional functionality for accessing and modifying VCPU state.

Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230914030600.16993-7-jniethe5@gmail.com
Stable-dep-of: ecd10702baae ("KVM: PPC: Book3S HV: Handle pending exceptions on guest entry with MSR_EE")
Signed-off-by: Sasha Levin <sashal@kernel.org>

drivers/perf: hisi: Fix some event id for HiSilicon UC pmu

[ Upstream commit 38bbef7240b8c5f2dc4493eec356e2efbf2da5f4 ]

Some event id of HiSilicon uncore UC PMU driver is incorrect, fix them.

Fixes: 312eca95e28d ("drivers/perf: hisi: Add support for HiSilicon UC PMU driver")
Signed-off-by: Junhao He <hejunhao3@huawei.com>
Reviewed-by: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20231204110425.20354-1-hejunhao3@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/arm-cmn: Fix HN-F class_occup_id events

[ Upstream commit 590f23b092401f29e410fd4ca67128fcc45192fc ]

A subtle copy-paste error managed to slip through the reorganisation
of these patches in development, and not only give some HN-F events
the wrong type, but use that wrong type before the subsequent patch
defined it. Too late to fix history, but we can at least fix the bug.

Fixes: b1b7dc38e482 ("perf/arm-cmn: Refactor HN-F event selector macros")
Reported-by: Jing Zhang <renyu.zj@linux.alibaba.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/5a22439de84ff188ef76674798052448eb03a3e1.1700740693.git.robin.murphy@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/pseries/memhp: Fix access beyond end of drmem array

[ Upstream commit bd68ffce69f6cf8ddd3a3c32549d1d2275e49fc5 ]

dlpar_memory_remove_by_index() may access beyond the bounds of the
drmem lmb array when the LMB lookup fails to match an entry with the
given DRC index. When the search fails, the cursor is left pointing to
&drmem_info->lmbs[drmem_info->n_lmbs], which is one element past the
last valid entry in the array. The debug message at the end of the
function then dereferences this pointer:

        pr_debug("Failed to hot-remove memory at %llx\n",
                 lmb->base_addr);

This was found by inspection and confirmed with KASAN:

  pseries-hotplug-mem: Attempting to hot-remove LMB, drc index 1234
  ==================================================================
  BUG: KASAN: slab-out-of-bounds in dlpar_memory+0x298/0x1658
  Read of size 8 at addr c000000364e97fd0 by task bash/949

  dump_stack_lvl+0xa4/0xfc (unreliable)
  print_report+0x214/0x63c
  kasan_report+0x140/0x2e0
  __asan_load8+0xa8/0xe0
  dlpar_memory+0x298/0x1658
  handle_dlpar_errorlog+0x130/0x1d0
  dlpar_store+0x18c/0x3e0
  kobj_attr_store+0x68/0xa0
  sysfs_kf_write+0xc4/0x110
  kernfs_fop_write_iter+0x26c/0x390
  vfs_write+0x2d4/0x4e0
  ksys_write+0xac/0x1a0
  system_call_exception+0x268/0x530
  system_call_vectored_common+0x15c/0x2ec

  Allocated by task 1:
   kasan_save_stack+0x48/0x80
   kasan_set_track+0x34/0x50
   kasan_save_alloc_info+0x34/0x50
   __kasan_kmalloc+0xd0/0x120
   __kmalloc+0x8c/0x320
   kmalloc_array.constprop.0+0x48/0x5c
   drmem_init+0x2a0/0x41c
   do_one_initcall+0xe0/0x5c0
   kernel_init_freeable+0x4ec/0x5a0
   kernel_init+0x30/0x1e0
   ret_from_kernel_user_thread+0x14/0x1c

  The buggy address belongs to the object at c000000364e80000
   which belongs to the cache kmalloc-128k of size 131072
  The buggy address is located 0 bytes to the right of
   allocated 98256-byte region [c000000364e80000, c000000364e97fd0)

  ==================================================================
  pseries-hotplug-mem: Failed to hot-remove memory at 0

Log failed lookups with a separate message and dereference the
cursor only when it points to a valid entry.

Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Fixes: 51925fb3c5c9 ("powerpc/pseries: Implement memory hotplug remove in the kernel")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231114-pseries-memhp-fixes-v1-1-fb8f2bb7c557@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/44x: select I2C for CURRITUCK

[ Upstream commit 4a74197b65e69c46fe6e53f7df2f4d6ce9ffe012 ]

Fix build errors when CURRITUCK=y and I2C is not builtin (=m or is
not set). Fixes these build errors:

powerpc-linux-ld: arch/powerpc/platforms/44x/ppc476.o: in function `avr_halt_system':
ppc476.c:(.text+0x58): undefined reference to `i2c_smbus_write_byte_data'
powerpc-linux-ld: arch/powerpc/platforms/44x/ppc476.o: in function `ppc47x_device_probe':
ppc476.c:(.init.text+0x18): undefined reference to `i2c_register_driver'

Fixes: 2a2c74b2efcb ("IBM Akebono: Add the Akebono platform")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Closes: lore.kernel.org/r/202312010820.cmdwF5X9-lkp@intel.com
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231201055159.8371-1-rdunlap@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86: Fix CPUIDLE_FLAG_IRQ_ENABLE leaking timer reprogram

[ Upstream commit edc8fc01f608108b0b7580cb2c29dfb5135e5f0e ]

intel_idle_irq() re-enables IRQs very early. As a result, an interrupt
may fire before mwait() is eventually called. If such an interrupt queues
a timer, it may go unnoticed until mwait returns and the idle loop
handles the tick re-evaluation. And monitoring TIF_NEED_RESCHED doesn't
help because a local timer enqueue doesn't set that flag.

The issue is mitigated by the fact that this idle handler is only invoked
for shallow C-states when, presumably, the next tick is supposed to be
close enough. There may still be rare cases though when the next tick
is far away and the selected C-state is shallow, resulting in a timer
getting ignored for a while.

Fix this with using sti_mwait() whose IRQ-reenablement only triggers
upon calling mwait(), dealing with the race while keeping the interrupt
latency within acceptable bounds.

Fixes: c227233ad64c (intel_idle: enable interrupts before C1 on Xeons)
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Link: https://lkml.kernel.org/r/20231115151325.6262-3-frederic@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc: add crtsavres.o to always-y instead of extra-y

[ Upstream commit 1b1e38002648819c04773647d5242990e2824264 ]

crtsavres.o is linked to modules. However, as explained in commit
d0e628cd817f ("kbuild: doc: clarify the difference between extra-y
and always-y"), 'make modules' does not build extra-y.

For example, the following command fails:

  $ make ARCH=powerpc LLVM=1 KBUILD_MODPOST_WARN=1 mrproper ps3_defconfig modules
    [snip]
    LD [M]  arch/powerpc/platforms/cell/spufs/spufs.ko
  ld.lld: error: cannot open arch/powerpc/lib/crtsavres.o: No such file or directory
  make[3]: *** [scripts/Makefile.modfinal:56: arch/powerpc/platforms/cell/spufs/spufs.ko] Error 1
  make[2]: *** [Makefile:1844: modules] Error 2
  make[1]: *** [/home/masahiro/workspace/linux-kbuild/Makefile:350: __build_one_by_one] Error 2
  make: *** [Makefile:234: __sub-make] Error 2

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Fixes: baa25b571a16 ("powerpc/64: Do not link crtsavres.o in vmlinux")
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231120232332.4100288-1-masahiroy@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

EDAC/thunderx: Fix possible out-of-bounds string access

[ Upstream commit 475c58e1a471e9b873e3e39958c64a2d278275c8 ]

Enabling -Wstringop-overflow globally exposes a warning for a common bug
in the usage of strncat():

  drivers/edac/thunderx_edac.c: In function 'thunderx_ocx_com_threaded_isr':
  drivers/edac/thunderx_edac.c:1136:17: error: 'strncat' specified bound 1024 equals destination size [-Werror=stringop-overflow=]
   1136 |                 strncat(msg, other, OCX_MESSAGE_SIZE);
        |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   ...
   1145 |                                 strncat(msg, other, OCX_MESSAGE_SIZE);
   ...
   1150 |                                 strncat(msg, other, OCX_MESSAGE_SIZE);

   ...

Apparently the author of this driver expected strncat() to behave the
way that strlcat() does, which uses the size of the destination buffer
as its third argument rather than the length of the source buffer. The
result is that there is no check on the size of the allocated buffer.

Change it to strlcat().

  [ bp: Trim compiler output, fixup commit message. ]

Fixes: 41003396f932 ("EDAC, thunderx: Add Cavium ThunderX EDAC driver")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20231122222007.3199885-1-arnd@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/mce/inject: Clear test status value

[ Upstream commit 6175b407756b22e7fdc771181b7d832ebdedef5c ]

AMD systems generally allow MCA "simulation" where MCA registers can be
written with valid data and the full MCA handling flow can be tested by
software.

However, the platform on Scalable MCA systems, can prevent software from
writing data to the MCA registers. There is no architectural way to
determine this configuration. Therefore, the MCE injection module will
check for this behavior by writing and reading back a test status value.
This is done during module init, and the check can run on any CPU with
any valid MCA bank.

If MCA_STATUS writes are ignored by the platform, then there are no side
effects on the hardware state.

If the writes are not ignored, then the test status value will remain in
the hardware MCA_STATUS register. It is likely that the value will not
be overwritten by hardware or software, since the tested CPU and bank
are arbitrary. Therefore, the user may see a spurious, synthetic MCA
error reported whenever MCA is polled for this CPU.

Clear the test value immediately after writing it. It is very unlikely
that a valid MCA error is logged by hardware during the test. Errors
that cause an #MC won't be affected.

Fixes: 891e465a1bd8 ("x86/mce: Check whether writes to MCA_STATUS are getting ignored")
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20231118193248.1296798-2-yazen.ghannam@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/lib: Fix overflow when counting digits

[ Upstream commit a24d61c609813963aacc9f6ec8343f4fcaac7243 ]

tl;dr: The num_digits() function has a theoretical overflow issue.
But it doesn't affect any actual in-tree users.  Fix it by using
a larger type for one of the local variables.

Long version:

There is an overflow in variable m in function num_digits when val
is >= 1410065408 which leads to the digit calculation loop to
iterate more times than required. This results in either more
digits being counted or in some cases (for example where val is
1932683193) the value of m eventually overflows to zero and the
while loop spins forever).

Currently the function num_digits is currently only being used for
small values of val in the SMP boot stage for digit counting on the
number of cpus and NUMA nodes, so the overflow is never encountered.
However it is useful to fix the overflow issue in case the function
is used for other purposes in the future. (The issue was discovered
while investigating the digit counting performance in various
kernel helper functions rather than any real-world use-case).

The simplest fix is to make m a long long, the overhead in
multiplication speed for a long long is very minor for small values
of val less than 10000 on modern processors. The alternative
fix is to replace the multiplication with a constant division
by 10 loop (this compiles down to an multiplication and shift)
without needing to make m a long long, but this is slightly slower
than the fix in this commit when measured on a range of x86
processors).

[ dhansen: subject and changelog tweaks ]

Fixes: 646e29a1789a ("x86: Improve the printout of the SMP bootup CPU table")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231102174901.2590325-1-colin.i.king%40gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

Linux 6.6.13

Link: https://lore.kernel.org/r/20240118104320.029537060@linuxfoundation.org
Tested-by: Allen Pais <apais@linux.microsoft.com>
Tested-by: Salvatore Bonaccorso <carnil@debian.org>
Tested-by: Shuah Khan <skhan@linuxfoundation.org>
Tested-by: Bagas Sanjaya <bagasdotme@gmail.com>
Tested-by: Takeshi Ogasawara <takeshi.ogasawara@futuring-girl.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Ron Economos <re@w6rz.net>
Tested-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
Tested-by: Justin M. Forbes <jforbes@fedoraproject.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm/memory_hotplug: fix memmap_on_memory sysfs value retrieval

commit 11684134140bb708b6e6de969a060535630b1b53 upstream.

set_memmap_mode() stores the kernel parameter memmap mode as an integer.
However, the get_memmap_mode() function utilizes param_get_bool() to fetch
the value as a boolean, leading to potential endianness issue.  On
Big-endian architectures, the memmap_on_memory is consistently displayed
as 'N' regardless of its actual status.

To address this endianness problem, the solution involves obtaining the
mode as an integer.  This adjustment ensures the proper display of the
memmap_on_memory parameter, presenting it as one of the following options:
Force, Y, or N.

Link: https://lkml.kernel.org/r/20240110140127.241451-1-sumanthk@linux.ibm.com
Fixes: 2d1f649c7c08 ("mm/memory_hotplug: support memmap_on_memory when memmap is not aligned to pageblocks")
Signed-off-by: Sumanth Korikkar <sumanthk@linux.ibm.com>
Suggested-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: <stable@vger.kernel.org> [6.6+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scripts/decode_stacktrace.sh: optionally use LLVM utilities

commit efbd6398353315b7018e6943e41fee9ec35e875f upstream.

GNU's addr2line can have problems parsing a vmlinux built with LLVM,
particularly when LTO was used.  In order to decode the traces correctly
this patch adds the ability to switch to LLVM's utilities readelf and
addr2line.  The same approach is followed by Will in [1].

Before:
  $ scripts/decode_stacktrace.sh vmlinux < kernel.log
  [17716.240635] Call trace:
  [17716.240646] skb_cow_data (??:?)
  [17716.240654] esp6_input (ld-temp.o:?)
  [17716.240666] xfrm_input (ld-temp.o:?)
  [17716.240674] xfrm6_rcv (??:?)
  [...]

After:
  $ LLVM=1 scripts/decode_stacktrace.sh vmlinux < kernel.log
  [17716.240635] Call trace:
  [17716.240646] skb_cow_data (include/linux/skbuff.h:2172 net/core/skbuff.c:4503)
  [17716.240654] esp6_input (net/ipv6/esp6.c:977)
  [17716.240666] xfrm_input (net/xfrm/xfrm_input.c:659)
  [17716.240674] xfrm6_rcv (net/ipv6/xfrm6_input.c:172)
  [...]

Note that one could set CROSS_COMPILE=llvm- instead to hack around this
issue.  However, doing so can break the decodecode routine as it will
force the selection of other LLVM utilities down the line e.g.  llvm-as.

[1] https://lore.kernel.org/all/20230914131225.13415-3-will@kernel.org/

Link: https://lkml.kernel.org/r/20230929034836.403735-1-cmllamas@google.com
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Elliot Berman <quic_eberman@quicinc.com>
Tested-by: Justin Stitt <justinstitt@google.com>
Cc: Will Deacon <will@kernel.org>
Cc: John Stultz <jstultz@google.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Tom Rix <trix@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: etm4x: Fix width of CCITMIN field

commit cc0271a339cc70cae914c3ec20edc2a8058407da upstream.

CCITMIN is a 12 bit field and doesn't fit in a u8, so extend it to u16.
This probably wasn't an issue previously because values higher than 255
never occurred.

But since commit 4aff040bcc8d ("coresight: etm: Override TRCIDR3.CCITMIN
on errata affected cpus"), a comparison with 256 was done to enable the
errata, generating the following W=1 build error:

  coresight-etm4x-core.c:1188:24: error: result of comparison of
  constant 256 with expression of type 'u8' (aka 'unsigned char') is
  always false [-Werror,-Wtautological-constant-out-of-range-compare]

   if (drvdata->ccitmin == 256)

Cc: stable@vger.kernel.org
Fixes: 2e1cdfe184b5 ("coresight-etm4x: Adding CoreSight ETM4x driver")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202310302043.as36UFED-lkp@intel.com/
Reviewed-by: Mike Leach <mike.leach@linaro.org>
Signed-off-by: James Clark <james.clark@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Link: https://lore.kernel.org/r/20231101115206.70810-1-james.clark@arm.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

PCI: Add ACS quirk for more Zhaoxin Root Ports

commit e367e3c765f5477b2e79da0f1399aed49e2d1e37 upstream.

Add more Root Port Device IDs to pci_quirk_zhaoxin_pcie_ports_acs() for
some new Zhaoxin platforms.

Fixes: 299bd044a6f3 ("PCI: Add ACS quirk for Zhaoxin Root/Downstream Ports")
Link: https://lore.kernel.org/r/20231211091543.735903-1-LeoLiu-oc@zhaoxin.com
Signed-off-by: LeoLiuoc <LeoLiu-oc@zhaoxin.com>
[bhelgaas: update subject, drop changelog, add Fixes, add stable tag, fix
whitespace, wrap code comment]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: <stable@vger.kernel.org> # 5.7
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

leds: ledtrig-tty: Free allocated ttyname buffer on deactivate

commit 25054b232681c286fca9c678854f56494d1352cc upstream.

The ttyname buffer for the ledtrig_tty_data struct is allocated in the
sysfs ttyname_store() function. This buffer must be released on trigger
deactivation. This was missing and is thus a memory leak.

While we are at it, the TTY handler in the ledtrig_tty_data struct should
also be returned in case of the trigger deactivation call.

Cc: stable@vger.kernel.org
Fixes: fd4a641ac88f ("leds: trigger: implement a tty trigger")
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20231127081621.774866-1-fe@dev.tdt.de
Signed-off-by: Lee Jones <lee@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

parport: parport_serial: Add Brainboxes device IDs and geometry

commit 6aa1fc5a8085bbc01687aa708dcf2dbe637a5ee3 upstream.

Add device IDs for the Brainboxes UC-203, UC-257, UC-414, UC-475,
IS-300/IS-500 and PX-263/PX-295 and define the relevant "geometry"
for the cards.
This patch requires part 1 of this series.

Cc: <stable@vger.kernel.org>
Signed-off-by: Cameron Williams <cang1@live.co.uk>
Acked-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Link: https://lore.kernel.org/r/AS4PR02MB7903A4094564BE28F1F926A6C4A6A@AS4PR02MB7903.eurprd02.prod.outlook.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

parport: parport_serial: Add Brainboxes BAR details

commit 65fde134b0a4ffe838729f9ee11b459a2f6f2815 upstream.

Add BAR/enum entries for Brainboxes serial/parallel cards.

Cc: <stable@vger.kernel.org>
Signed-off-by: Cameron Williams <cang1@live.co.uk>
Acked-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Link: https://lore.kernel.org/r/AS4PR02MB79035155C2D5C3333AE6FA52C4A6A@AS4PR02MB7903.eurprd02.prod.outlook.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

uio: Fix use-after-free in uio_open

commit 0c9ae0b8605078eafc3bea053cc78791e97ba2e2 upstream.

core-1 core-2
-------------------------------------------------------
uio_unregister_device uio_open
idev = idr_find()
device_unregister(&idev->dev)
put_device(&idev->dev)
uio_device_release
get_device(&idev->dev)
kfree(idev)
uio_free_minor(minor)
uio_release
put_device(&idev->dev)
kfree(idev)
-------------------------------------------------------

In the core-1 uio_unregister_device(), the device_unregister will kfree
idev when the idev->dev kobject ref is 1. But after core-1
device_unregister, put_device and before doing kfree, the core-2 may
get_device. Then:
1. After core-1 kfree idev, the core-2 will do use-after-free for idev.
2. When core-2 do uio_release and put_device, the idev will be double
   freed.

To address this issue, we can get idev atomic & inc idev reference with
minor_lock.

Fixes: 57c5f4df0a5a ("uio: fix crash after the device is unregistered")
Cc: stable <stable@kernel.org>
Signed-off-by: Guanghui Feng <guanghuifeng@linux.alibaba.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Link: https://lore.kernel.org/r/1703152663-59949-1-git-send-email-guanghuifeng@linux.alibaba.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

binder: fix comment on binder_alloc_new_buf() return value

commit e1090371e02b601cbfcea175c2a6cc7c955fa830 upstream.

Update the comments of binder_alloc_new_buf() to reflect that the return
value of the function is now ERR_PTR(-errno) on failure.

No functional changes in this patch.

Cc: stable@vger.kernel.org
Fixes: 57ada2fb2250 ("binder: add log information for binder transaction failures")
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Link: https://lore.kernel.org/r/20231201172212.1813387-8-cmllamas@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

binder: fix trivial typo of binder_free_buf_locked()

commit 122a3c1cb0ff304c2b8934584fcfea4edb2fe5e3 upstream.

Fix minor misspelling of the function in the comment section.

No functional changes in this patch.

Cc: stable@vger.kernel.org
Fixes: 0f966cba95c7 ("binder: add flag to clear buffer on txn complete")
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Link: https://lore.kernel.org/r/20231201172212.1813387-7-cmllamas@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

binder: fix use-after-free in shinker's callback

commit 3f489c2067c5824528212b0fc18b28d51332d906 upstream.

The mmap read lock is used during the shrinker's callback, which means
that using alloc->vma pointer isn't safe as it can race with munmap().
As of commit dd2283f2605e ("mm: mmap: zap pages with read mmap_sem in
munmap") the mmap lock is downgraded after the vma has been isolated.

I was able to reproduce this issue by manually adding some delays and
triggering page reclaiming through the shrinker's debug sysfs. The
following KASAN report confirms the UAF:

  ==================================================================
  BUG: KASAN: slab-use-after-free in zap_page_range_single+0x470/0x4b8
  Read of size 8 at addr ffff356ed50e50f0 by task bash/478

  CPU: 1 PID: 478 Comm: bash Not tainted 6.6.0-rc5-00055-g1c8b86a3799f-dirty #70
  Hardware name: linux,dummy-virt (DT)
  Call trace:
   zap_page_range_single+0x470/0x4b8
   binder_alloc_free_page+0x608/0xadc
   __list_lru_walk_one+0x130/0x3b0
   list_lru_walk_node+0xc4/0x22c
   binder_shrink_scan+0x108/0x1dc
   shrinker_debugfs_scan_write+0x2b4/0x500
   full_proxy_write+0xd4/0x140
   vfs_write+0x1ac/0x758
   ksys_write+0xf0/0x1dc
   __arm64_sys_write+0x6c/0x9c

  Allocated by task 492:
   kmem_cache_alloc+0x130/0x368
   vm_area_alloc+0x2c/0x190
   mmap_region+0x258/0x18bc
   do_mmap+0x694/0xa60
   vm_mmap_pgoff+0x170/0x29c
   ksys_mmap_pgoff+0x290/0x3a0
   __arm64_sys_mmap+0xcc/0x144

  Freed by task 491:
   kmem_cache_free+0x17c/0x3c8
   vm_area_free_rcu_cb+0x74/0x98
   rcu_core+0xa38/0x26d4
   rcu_core_si+0x10/0x1c
   __do_softirq+0x2fc/0xd24

  Last potentially related work creation:
   __call_rcu_common.constprop.0+0x6c/0xba0
   call_rcu+0x10/0x1c
   vm_area_free+0x18/0x24
   remove_vma+0xe4/0x118
   do_vmi_align_munmap.isra.0+0x718/0xb5c
   do_vmi_munmap+0xdc/0x1fc
   __vm_munmap+0x10c/0x278
   __arm64_sys_munmap+0x58/0x7c

Fix this issue by performing instead a vma_lookup() which will fail to
find the vma that was isolated before the mmap lock downgrade. Note that
this option has better performance than upgrading to a mmap write lock
which would increase contention. Plus, mmap_write_trylock() has been
recently removed anyway.

Fixes: dd2283f2605e ("mm: mmap: zap pages with read mmap_sem in munmap")
Cc: stable@vger.kernel.org
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Link: https://lore.kernel.org/r/20231201172212.1813387-3-cmllamas@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

binder: use EPOLLERR from eventpoll.h

commit 6ac061db9c58ca5b9270b1b3940d2464fb3ff183 upstream.

Use EPOLLERR instead of POLLERR to make sure it is cast to the correct
__poll_t type. This fixes the following sparse issue:

  drivers/android/binder.c:5030:24: warning: incorrect type in return expression (different base types)
  drivers/android/binder.c:5030:24:    expected restricted __poll_t
  drivers/android/binder.c:5030:24:    got int

Fixes: f88982679f54 ("binder: check for binder_thread allocation failure in binder_poll()")
Cc: stable@vger.kernel.org
Cc: Eric Biggers <ebiggers@google.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Link: https://lore.kernel.org/r/20231201172212.1813387-2-cmllamas@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "md/raid5: Wait for MD_SB_CHANGE_PENDING in raid5d"

commit bed9e27baf52a09b7ba2a3714f1e24e17ced386d upstream.

This reverts commit 5e2cf333b7bd5d3e62595a44d598a254c697cd74.

That commit introduced the following race and can cause system hung.

 md_write_start:             raid5d:
 // mddev->in_sync == 1
 set "MD_SB_CHANGE_PENDING"
                            // running before md_write_start wakeup it
                             waiting "MD_SB_CHANGE_PENDING" cleared
                             >>>>>>>>> hung
 wakeup mddev->thread
 ...
 waiting "MD_SB_CHANGE_PENDING" cleared
 >>>> hung, raid5d should clear this flag
 but get hung by same flag.

The issue reverted commit fixing is fixed by last patch in a new way.

Fixes: 5e2cf333b7bd ("md/raid5: Wait for MD_SB_CHANGE_PENDING in raid5d")
Cc: stable@vger.kernel.org # v5.19+
Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20231108182216.73611-2-junxiao.bi@oracle.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ksmbd: free ppace array on error in parse_dacl

commit 8cf9bedfc3c47d24bb0de386f808f925dc52863e upstream.

The ppace array is not freed if one of the init_acl_state() calls inside
parse_dacl() fails. At the moment the function may fail only due to the
memory allocation errors so it's highly unlikely in this case but
nevertheless a fix is needed.

Move ppace allocation after the init_acl_state() calls with proper error
handling.

Found by Linux Verification Center (linuxtesting.org).

Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ksmbd: don't allow O_TRUNC open on read-only share

commit d592a9158a112d419f341f035d18d02f8d232def upstream.

When file is changed using notepad on read-only share(read_only = yes in
ksmbd.conf), There is a problem where existing data is truncated.
notepad in windows try to O_TRUNC open(FILE_OVERWRITE_IF) and all data
in file is truncated. This patch don't allow  O_TRUNC open on read-only
share and add KSMBD_TREE_CONN_FLAG_WRITABLE check in smb2_set_info().

Cc: stable@vger.kernel.org
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/amd/display: Pass pwrseq inst for backlight and ABM

commit b17ef04bf3a4346d66404454d6a646343ddc9749 upstream.

[Why]
OTG inst and pwrseq inst mapping is not align therefore we cannot use
otg_inst as pwrseq inst to get DCIO register.

[How]
1. Pass the correct pwrseq instance to dmub when set abm pipe.
2. LVTMA control index change from panel_inst to pwrseq_inst.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Phil Hsieh <phil.hsieh@amd.com>
Acked-by: Rodrigo Siqueira <rodrigo.siqueira@amd.com>
Signed-off-by: Lewis Huang <lewis.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ASoC: SOF: Intel: hda-codec: Delay the codec device registration

commit c344ef36dbc2fe920ec7291b68b11fe867a2c8f6 upstream.

The current code flow is:
1. snd_hdac_device_register()
2. set parameters needed by the hdac driver
3. request_codec_module()
   the hdac driver is probed at this point

During boot the codec drivers are not loaded when the hdac device is
registered, it is going to be probed later when loading the codec module,
which point the parameters are set.

On module remove/insert
rmmod snd_sof_pci_intel_tgl
modprobe snd_sof_pci_intel_tgl

The codec module remains loaded and the driver will be probed when the
hdac device is created right away, before the parameters for the driver
has been configured:

1. snd_hdac_device_register()
   the hdac driver is probed at this point
2. set parameters needed by the hdac driver
3. request_codec_module()
   will be a NOP as the module is already loaded

Move the snd_hdac_device_register() later, to be done right before
requesting the codec module to make sure that the parameters are all set
before the device is created:

1. set parameters needed by the hdac driver
2. snd_hdac_device_register()
3. request_codec_module()

This way at the hdac driver probe all parameters will be set in all cases.

Link: https://github.com/thesofproject/linux/issues/4731
Fixes: a0575b4add21 ("ASoC: hdac_hda: Conditionally register dais for HDMI and Analog")
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20231207095425.19597-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/ZYvUIxtrqBQZbNlC@shine.dominikbrodowski.net
Link: https://bugzilla.kernel.org/show_bug.cgi?id=218304
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

bus: moxtet: Add spi device table

commit aaafe88d5500ba18b33be72458439367ef878788 upstream.

The moxtet module fails to auto-load on. Add a SPI id table to
allow it to do so.

Signed-off-by: Sjoerd Simons <sjoerd@collabora.com>
Cc: <stable@vger.kernel.org>
Reviewed-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

bus: moxtet: Mark the irq as shared

commit e7830f5a83e96d8cb8efc0412902a03008f8fbe3 upstream.

The Turris Mox shares the moxtet IRQ with various devices on the board,
so mark the IRQ as shared in the driver as well.

Without this loading the module will fail with:
  genirq: Flags mismatch irq 40. 00002002 (moxtet) vs. 00002080 (mcp7940x)

Signed-off-by: Sjoerd Simons <sjoerd@collabora.com>
Cc: <stable@vger.kernel.org> # v6.2+
Reviewed-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ACPI: resource: Add another DMI match for the TongFang GMxXGxx

commit df0cced74159c79e36ce7971f0bf250673296d93 upstream.

The TongFang GMxXGxx, which needs IRQ overriding for the keyboard to work,
is also sold as the Eluktronics RP-15 which does not use the standard
TongFang GMxXGxx DMI board_name.

Add an entry for this laptop to the irq1_edge_low_force_override[] DMI
table to make the internal keyboard functional.

Reported-by: Luis Acuna <ldacuna@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ALSA: hda/realtek: Fix mute and mic-mute LEDs for HP Envy X360 13-ay0xxx

commit 6b3d14b7f9b1acaf7303d8499836bf78ee9c470c upstream.

This enables the mute and mic-mute LEDs on the HP Envy X360 13-ay0xxx
convertibles.
The quirk 'ALC245_FIXUP_HP_X360_MUTE_LEDS' already exists and is now
enabled for this device.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=216197
Signed-off-by: Tom Jason Schwanke <tom@catboys.cloud>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/651b26e9-e86b-45dd-aa90-3e43d6b99823@catboys.cloud
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/microcode: do not cache microcode if it will not be used

No relevant upstream kernel due to refactoring in 6.7

Builtin/initrd microcode will not be used the ucode loader is disabled.
But currently, save_microcode_in_initrd is always performed and it
accesses MSR_IA32_UCODE_REV even if dis_ucode_ldr is true, and in
particular even if X86_FEATURE_HYPERVISOR is set; the TDX module does not
implement the MSR and the result is a call trace at boot for TDX guests.

Mainline Linux fixed this as part of a more complex rework of microcode
caching that went into 6.7 (see in particular commits dd5e3e3ca6,
"x86/microcode/intel: Simplify early loading"; and a7939f0167203,
"x86/microcode/amd: Cache builtin/initrd microcode early").  Do the bare
minimum in stable kernels, setting initrd_gone just like mainline Linux
does in mark_initrd_gone().

Note that save_microcode_in_initrd() is not in the microcode application
path, which runs with paging disabled on 32-bit systems, so it can (and
has to) use dis_ucode_ldr instead of check_loader_disabled_ap().

Cc: stable@vger.kernel.org # v6.6+
Cc: x86@kernel.org # v6.6+
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Borislav Petkov (AMD) <bp@alien8.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/crtc: fix uninitialized variable use

[ Upstream commit 6e455f5dcdd15fa28edf0ffb5b44d3508512dccf ]

Commit 3823119b9c2b ("drm/crtc: Fix uninit-value bug in
drm_mode_setcrtc") was supposed to fix use of an uninitialized variable,
but introduced another.

num_connectors is only initialized if crtc_req->count_connectors > 0,
but it's used regardless. Fix it.

Fixes: 3823119b9c2b ("drm/crtc: Fix uninit-value bug in drm_mode_setcrtc")
Cc: syzbot+4fad2e57beb6397ab2fc@syzkaller.appspotmail.com
Cc: Ziqi Zhao <astrajoan@yahoo.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231208131238.2924571-1-jani.nikula@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/csum: clean up `csum_partial' further

[ Upstream commit a476aae3f1dc78a162a0d2e7945feea7d2b29401 ]

Commit 688eb8191b47 ("x86/csum: Improve performance of `csum_partial`")
ended up improving the code generation for the IP csum calculations, and
in particular special-casing the 40-byte case that is a hot case for
IPv6 headers.

It then had _another_ special case for the 64-byte unrolled loop, which
did two chains of 32-byte blocks, which allows modern CPU's to improve
performance by doing the chains in parallel thanks to renaming the carry
flag.

This just unifies the special cases and combines them into just one
single helper the 40-byte csum case, and replaces the 64-byte case by a
80-byte case that just does that single helper twice.  It avoids having
all these different versions of inline assembly, and actually improved
performance further in my tests.

There was never anything magical about the 64-byte unrolled case, even
though it happens to be a common size (and typically is the cacheline
size).

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/csum: Remove unnecessary odd handling

[ Upstream commit 5d4acb62853abac1da2deebcb1c1c5b79219bf3b ]

The special case for odd aligned buffers is unnecessary and mostly
just adds overhead. Aligned buffers is the expectations, and even for
unaligned buffer, the only case that was helped is if the buffer was
1-byte from word aligned which is ~1/7 of the cases. Overall it seems
highly unlikely to be worth to extra branch.

It was left in the previous perf improvement patch because I was
erroneously comparing the exact output of `csum_partial(...)`, but
really we only need `csum_fold(csum_partial(...))` to match so its
safe to remove.

All csum kunit tests pass.

Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Laight <david.laight@aculab.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: sun9i: smp: fix return code check of of_property_match_string

[ Upstream commit 643fe70e7bcdcc9e2d96952f7fc2bab56385cce5 ]

of_property_match_string returns an int; either an index from 0 or
greater if successful or negative on failure. Even it's very
unlikely that the DT CPU node contains multiple enable-methods
these checks should be fixed.

This patch was inspired by the work of Nick Desaulniers.

Link: https://lore.kernel.org/lkml/20230516-sunxi-v1-1-ac4b9651a8c1@google.com/T/
Cc: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
Link: https://lore.kernel.org/r/20231228193903.9078-2-wahrenst@gmx.net
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

connector: Fix proc_event_num_listeners count not cleared

[ Upstream commit c46bfba1337d301661dbb23cfd905d4cb51f27ca ]

When we register a cn_proc listening event, the proc_event_num_listener
variable will be incremented by one, but if PROC_CN_MCAST_IGNORE is
not called, the count will not decrease.
This will cause the proc_*_connector function to take the wrong path.
It will reappear when the forkstat tool exits via ctrl + c.
We solve this problem by determining whether
there are still listeners to clear proc_event_num_listener.

Signed-off-by: wangkeqi <wangkeqiwang@didiglobal.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: qrtr: ns: Return 0 if server port is not present

[ Upstream commit 9bf2e9165f90dc9f416af53c902be7e33930f728 ]

When a 'DEL_CLIENT' message is received from the remote, the corresponding
server port gets deleted. A DEL_SERVER message is then announced for this
server. As part of handling the subsequent DEL_SERVER message, the name-
server attempts to delete the server port which results in a '-ENOENT' error.
The return value from server_del() is then propagated back to qrtr_ns_worker,
causing excessive error prints.
To address this, return 0 from control_cmd_del_server() without checking the
return value of server_del(), since the above scenario is not an error case
and hence server_del() doesn't have any other error return value.

Signed-off-by: Sarannya Sasikumar <quic_sarannya@quicinc.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nfc: Do not send datagram if socket state isn't LLCP_BOUND

[ Upstream commit 6ec0d7527c4287369b52df3bcefd21a0c4fb2b7c ]

As we know we cannot send the datagram (state can be set to LLCP_CLOSED
by nfc_llcp_socket_release()), there is no need to proceed further.

Thus, bail out early from llcp_sock_sendmsg().

Signed-off-by: Siddh Raman Pant <code@siddh.me>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Suman Ghosh <sumang@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

virtio_blk: fix snprintf truncation compiler warning

[ Upstream commit b8e0792449928943c15d1af9f63816911d139267 ]

Commit 4e0400525691 ("virtio-blk: support polling I/O") triggers the
following gcc 13 W=1 warnings:

drivers/block/virtio_blk.c: In function ‘init_vq’:
drivers/block/virtio_blk.c:1077:68: warning: ‘%d’ directive output may be truncated writing between 1 and 11 bytes into a region of size 7 [-Wformat-truncation=]
 1077 |                 snprintf(vblk->vqs[i].name, VQ_NAME_LEN, "req_poll.%d", i);
      |                                                                    ^~
drivers/block/virtio_blk.c:1077:58: note: directive argument in the range [-2147483648, 65534]
 1077 |                 snprintf(vblk->vqs[i].name, VQ_NAME_LEN, "req_poll.%d", i);
      |                                                          ^~~~~~~~~~~~~
drivers/block/virtio_blk.c:1077:17: note: ‘snprintf’ output between 11 and 21 bytes into a destination of size 16
 1077 |                 snprintf(vblk->vqs[i].name, VQ_NAME_LEN, "req_poll.%d", i);
      |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

This is a false positive because the lower bound -2147483648 is
incorrect. The true range of i is [0, num_vqs - 1] where 0 < num_vqs <
65536.

The code mixes int, unsigned short, and unsigned int types in addition
to using "%d" for an unsigned value. Use unsigned short and "%u"
consistently to solve the compiler warning.

Cc: Suwan Kim <suwan.kim027@gmail.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202312041509.DIyvEt9h-lkp@intel.com/
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20231204140743.1487843-1-stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ida: Fix crash in ida_free when the bitmap is empty

[ Upstream commit af73483f4e8b6f5c68c9aa63257bdd929a9c194a ]

The IDA usually detects double-frees, but that detection failed to
consider the case when there are no nearby IDs allocated and so we have a
NULL bitmap rather than simply having a clear bit.  Add some tests to the
test-suite to be sure we don't inadvertently reintroduce this problem.
Unfortunately they're quite noisy so include a message to disregard
the warnings.

Reported-by: Zhenghan Wang <wzhmmmmm@gmail.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

posix-timers: Get rid of [COMPAT_]SYS_NI() uses

[ Upstream commit a4aebe936554dac6a91e5d091179c934f8325708 ]

Only the posix timer system calls use this (when the posix timer support
is disabled, which does not actually happen in any normal case), because
they had debug code to print out a warning about missing system calls.

Get rid of that special case, and just use the standard COND_SYSCALL
interface that creates weak system call stubs that return -ENOSYS for
when the system call does not exist.

This fixes a kCFI issue with the SYS_NI() hackery:

  CFI failure at int80_emulation+0x67/0xb0 (target: sys_ni_posix_timers+0x0/0x70; expected type: 0xb02b34d9)
  WARNING: CPU: 0 PID: 48 at int80_emulation+0x67/0xb0

Reported-by: kernel test robot <oliver.sang@intel.com>
Reviewed-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Sami Tolvanen <samitolvanen@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

pinctrl: cy8c95x0: Fix get_pincfg

[ Upstream commit 94c71705cc49092cef60ece13a28680809096fd4 ]

Invert the register value for PIN_CONFIG_OUTPUT_ENABLE to return
the opposite of PIN_CONFIG_INPUT_ENABLE.

Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Link: https://lore.kernel.org/r/20231219125120.4028862-3-patrick.rudolph@9elements.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

pinctrl: cy8c95x0: Fix regression

[ Upstream commit 04dfca968cf7349773126340991b68a83378aca2 ]

Commit 1fa3df901f2c ("pinctrl: cy8c95x0: Remove custom ->set_config()")
removed support for PIN_CONFIG_INPUT_ENABLE and
PIN_CONFIG_OUTPUT.

Add the following options to restore functionality:
- PIN_CONFIG_INPUT_ENABLE
- PIN_CONFIG_OUTPUT_ENABLE

Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Link: https://lore.kernel.org/r/20231219125120.4028862-2-patrick.rudolph@9elements.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

pinctrl: cy8c95x0: Fix typo

[ Upstream commit 47b1fa48116238208c1b1198dba10f56fc1b6eb2 ]

Fix typo to make pinctrl-cy8c95x compile again.

Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Link: https://lore.kernel.org/r/20231219125120.4028862-1-patrick.rudolph@9elements.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: get dprefclk ss info from integration info table

[ Upstream commit 51e7b64690776a9981355428b537af9048308a95 ]

[why & how]
we have two SSC_En:
we get ssc_info from dce_info for MPLL_SSC_EN.
we used to call VBIOS cmdtbl's smu_info's SS persentage for DPRECLK SS info,
is used for DP AUDIO and VBIOS' smu_info table was from systemIntegrationInfoTable.

since dcn35 VBIOS removed smu_info, driver need to use integrationInfotable directly.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Charlene Liu <charlene.liu@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>

drm/amd/display: Add case for dcn35 to support usb4 dmub hpd event

[ Upstream commit 989824589f793120833bef13aa4e21f5a836a707 ]

[Why & how]
Refactor dc_is_dmub_outbox_supported() a bit and add case for dcn35 to
register dmub outbox notification irq to handle usb4 relevant hpd event.

Reviewed-by: Roman Li <roman.li@amd.com>
Reviewed-by: Jun Lei <jun.lei@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@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>

drm/amdkfd: svm range always mapped flag not working on APU

[ Upstream commit ebab8c3eb6a6515dc14cd93fc29dd287709da6d3 ]

On gfx943 APU there is no VRAM and page migration, queue CWSR area, svm
range with always mapped flag, is not mapped to GPU correctly. This
works fine if retry fault on CWSR area can be recovered, but could cause
deadlock if there is another retry fault recover waiting for CWSR to
finish.

Fix this by mapping svm range with always mapped flag to GPU with ACCESS
attribute if XNACK ON.

There is side effect, because all GPUs have ACCESS attribute by default
on new svm range with XNACK on, the CWSR area will be mapped to all GPUs
after this change. This side effect will be fixed with Thunk change to
set CWSR svm range with ACCESS_IN_PLACE attribute on the GPU that user
queue is created.

Signed-off-by: Philip Yang <Philip.Yang@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>

i2c: rk3x: fix potential spinlock recursion on poll

[ Upstream commit 19cde9c92b8d3b7ee555d0da3bcb0232d3a784f4 ]

Possible deadlock scenario (on reboot):
rk3x_i2c_xfer_common(polling)
    -> rk3x_i2c_wait_xfer_poll()
        -> rk3x_i2c_irq(0, i2c);
            --> spin_lock(&i2c->lock);
            ...
        <rk3x i2c interrupt>
        -> rk3x_i2c_irq(0, i2c);
            --> spin_lock(&i2c->lock); (deadlock here)

Store the IRQ number and disable/enable it around the polling transfer.
This patch has been tested on NanoPC-T4.

Signed-off-by: Jensen Huang <jensenhuang@friendlyarm.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Andi Shyti <andi.shyti@kernel.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

smb: client: fix potential OOB in smb2_dump_detail()

[ Upstream commit 567320c46a60a3c39b69aa1df802d753817a3f86 ]

Validate SMB message with ->check_message() before calling
->calc_smb_size().

This fixes CVE-2023-6610.

Reported-by: j51569436@gmail.com
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218219
Cc; stable@vger.kernel.org
Signed-off-by: Paulo Alcantara <pc@manguebit.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>