xfrm: interface with if_id 0 should return error

[ Upstream commit 8dce43919566f06e865f7e8949f5c10d8c2493f5 ]

xfrm interface if_id = 0 would cause xfrm policy lookup errors since
Commit 9f8550e4bd9d.

Now explicitly fail to create an xfrm interface when if_id = 0

With this commit:
 ip link add ipsec0  type xfrm dev lo  if_id 0
 Error: if_id must be non zero.

v1->v2 change:
 - add Fixes: tag

Fixes: 9f8550e4bd9d ("xfrm: fix disable_xfrm sysctl when used on xfrm interfaces")
Signed-off-by: Antony Antony <antony.antony@secunet.com>
Reviewed-by: Eyal Birger <eyal.birger@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: hantro: Fix probe func error path

[ Upstream commit 37af43b250fda6162005d47bf7c959c70d52b107 ]

If clocks for some reason couldn't be enabled, probe function returns
immediately, without disabling PM. This obviously leaves PM ref counters
unbalanced.

Fix that by jumping to appropriate error path, so effects of PM functions
are reversed.

Fixes: 775fec69008d ("media: add Rockchip VPU JPEG encoder driver")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Acked-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/tegra: vic: Fix DMA API misuse

[ Upstream commit 5566174cb10a5167d59b0793871cab7990b149b8 ]

Upon failure, dma_alloc_coherent() returns NULL. If that does happen,
passing some uninitialised stack contents to dma_mapping_error() - which
belongs to a different API in the first place - has precious little
chance of detecting it.

Also include the correct header, because the fragile transitive
inclusion currently providing it is going to break soon.

Fixes: 20e7dce255e9 ("drm/tegra: Remove memory allocation from Falcon library")
CC: Thierry Reding <thierry.reding@gmail.com>
CC: Mikko Perttunen <mperttunen@nvidia.com>
CC: dri-devel@lists.freedesktop.org
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/tegra: gr2d: Explicitly control module reset

[ Upstream commit 271fca025a6d43f1c18a48543c5aaf31a31e4694 ]

As of commit 4782c0a5dd88 ("clk: tegra: Don't deassert reset on enabling
clocks"), module resets are no longer automatically deasserted when the
module clock is enabled. To make sure that the gr2d module continues to
work, we need to explicitly control the module reset.

Fixes: 4782c0a5dd88 ("clk: tegra: Don't deassert reset on enabling clocks")
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

gpu: host1x: select CONFIG_DMA_SHARED_BUFFER

[ Upstream commit 6c7a388b62366f0de9936db3c1921d7f4e0011bc ]

Linking fails when dma-buf is disabled:

ld.lld: error: undefined symbol: dma_fence_release
>>> referenced by fence.c
>>>               gpu/host1x/fence.o:(host1x_syncpt_fence_enable_signaling) in archive drivers/built-in.a
>>> referenced by fence.c
>>>               gpu/host1x/fence.o:(host1x_fence_signal) in archive drivers/built-in.a
>>> referenced by fence.c
>>>               gpu/host1x/fence.o:(do_fence_timeout) in archive drivers/built-in.a

Fixes: 687db2207b1b ("gpu: host1x: Add DMA fence implementation")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/bridge: ti-sn65dsi86: Set max register for regmap

[ Upstream commit 0b665d4af35837f0a0ae63135b84a3c187c1db3b ]

Set the maximum register to 0xff so we can dump the registers for this
device in debugfs.

Fixes: a095f15c00e2 ("drm/bridge: add support for sn65dsi86 bridge driver")
Cc: Rob Clark <robdclark@chromium.org>
Cc: Douglas Anderson <dianders@chromium.org>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20211215002529.382383-1-swboyd@chromium.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/dpu: fix safe status debugfs file

[ Upstream commit f31b0e24d31e18b4503eeaf0032baeacc0beaff6 ]

Make safe_status debugfs fs file actually return safe status rather than
danger status data.

Fixes: 25fdd5933e4c ("drm/msm: Add SDM845 DPU support")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Link: https://lore.kernel.org/r/20211201222633.2476780-3-dmitry.baryshkov@linaro.org
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: ipq6018: Fix gpio-ranges property

[ Upstream commit 72cb4c48a46a7cfa58eb5842c0d3672ddd5bd9ad ]

There must be three parameters in gpio-ranges property. Fixes this not
very helpful error message:

  OF: /soc/pinctrl@1000000: (null) = 3 found 3

Fixes: 1e8277854b49 ("arm64: dts: Add ipq6018 SoC and CP01 board support")
Cc: Sricharan R <sricharan@codeaurora.org>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/8a744cfd96aff5754bfdcf7298d208ddca5b319a.1638862030.git.baruch@tkos.co.il
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: c630: Fix soundcard setup

[ Upstream commit c02b360ca67ebeb9de07b47b2fe53f964c2561d1 ]

Currently Soundcard has 1 rx device for headset and SoundWire Speaker Playback.

This setup has issues, ex if we try to play on headset the audio stream is
also sent to SoundWire Speakers and we will hear sound in both headsets and speakers.

Make a separate device for Speakers and Headset so that the streams are
different and handled properly.

Fixes: 45021d35fcb2 ("arm64: dts: qcom: c630: Enable audio support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Tested-by: Steev Klimaszewski <steev@kali.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211209175342.20386-2-srinivas.kandagatla@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: hellcreek: Add missing PTP via UDP rules

[ Upstream commit 6cf01e451599da630ff1af529d61c5e4db4550ab ]

The switch supports PTP for UDP transport too. Therefore, add the missing static
FDB entries to ensure correct forwarding of these packets.

Fixes: ddd56dfe52c9 ("net: dsa: hellcreek: Add PTP clock support")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: hellcreek: Allow PTP P2P measurements on blocked ports

[ Upstream commit cad1798d2d0811ded37d1e946c6796102e58013b ]

Allow PTP peer delay measurements on blocked ports by STP. In case of topology
changes the PTP stack can directly start with the correct delays.

Fixes: ddd56dfe52c9 ("net: dsa: hellcreek: Add PTP clock support")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: hellcreek: Add STP forwarding rule

[ Upstream commit b7ade35eb53a2455f737a623c24e4b24455b2271 ]

Treat STP as management traffic. STP traffic is designated for the CPU port
only. In addition, STP traffic has to pass blocked ports.

Fixes: e4b27ebc780f ("net: dsa: Add DSA driver for Hirschmann Hellcreek switches")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: hellcreek: Fix insertion of static FDB entries

[ Upstream commit 4db4c3ea56978086ca367a355e440de17d534827 ]

The insertion of static FDB entries ignores the pass_blocked bit. That bit is
evaluated with regards to STP. Add the missing functionality.

Fixes: e4b27ebc780f ("net: dsa: Add DSA driver for Hirschmann Hellcreek switches")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ath11k: Fix a NULL pointer dereference in ath11k_mac_op_hw_scan()

[ Upstream commit eccd25136386a04ebf46a64f3a34e8e0fab6d9e1 ]

In ath11k_mac_op_hw_scan(), the return value of kzalloc() is directly
used in memcpy(), which may lead to a NULL pointer dereference on
failure of kzalloc().

Fix this bug by adding a check of arg.extraie.ptr.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_ATH11K=m show no new warnings, and our static
analyzer no longer warns about this code.

Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20211202155348.71315-1-zhou1615@umn.edu
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: coda/imx-vdoa: Handle dma_set_coherent_mask error codes

[ Upstream commit 43f0633f89947df57fe0b5025bdd741768007708 ]

The return value of dma_set_coherent_mask() is not always 0.
To catch the exception in case that dma is not support the mask.

Link: https://lore.kernel.org/linux-media/20211206022201.1639460-1-jiasheng@iscas.ac.cn
Fixes: b0444f18e0b1 ("[media] coda: add i.MX6 VDOA driver")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: msi001: fix possible null-ptr-deref in msi001_probe()

[ Upstream commit 3d5831a40d3464eea158180eb12cbd81c5edfb6a ]

I got a null-ptr-deref report:

BUG: kernel NULL pointer dereference, address: 0000000000000060
...
RIP: 0010:v4l2_ctrl_auto_cluster+0x57/0x270
...
Call Trace:
 msi001_probe+0x13b/0x24b [msi001]
 spi_probe+0xeb/0x130
...
 do_syscall_64+0x35/0xb0

In msi001_probe(), if the creation of control for bandwidth_auto
fails, there will be a null-ptr-deref issue when it is used in
v4l2_ctrl_auto_cluster().

Check dev->hdl.error before v4l2_ctrl_auto_cluster() to fix this bug.

Link: https://lore.kernel.org/linux-media/20211026112348.2878040-1-wanghai38@huawei.com
Fixes: 93203dd6c7c4 ("[media] msi001: Mirics MSi001 silicon tuner driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: dw2102: Fix use after free

[ Upstream commit 589a9f0eb799f77de2c09583bf5bad221fa5d685 ]

dvb_usb_device_init stores parts of properties at d->props
and d->desc and uses it on dvb_usb_device_exit.
Free of properties on module probe leads to use after free.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=204597

The patch makes properties static instead of allocated on heap to prevent
memleak and use after free.
Also fixes s421_properties.devices initialization to have 2 element
instead of 6 copied from p7500_properties.

[mchehab: fix function call alignments]
Link: https://lore.kernel.org/linux-media/20190822104147.4420-1-vasilyev@ispras.ru
Signed-off-by: Anton Vasilyev <vasilyev@ispras.ru>
Fixes: 299c7007e936 ("media: dw2102: Fix memleak on sequence of probes")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

perf/arm-cmn: Fix CPU hotplug unregistration

[ Upstream commit 56c7c6eaf3eb8ac1ec40d56096c0f2b27250da5f ]

Attempting to migrate the PMU context after we've unregistered the PMU
device, or especially if we never successfully registered it in the
first place, is a woefully bad idea. It's also fundamentally pointless
anyway. Make sure to unregister an instance from the hotplug handler
*without* invoking the teardown callback.

Fixes: 0ba64770a2f2 ("perf: Add Arm CMN-600 PMU driver")
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/2c221d745544774e4b07583b65b5d4d94f7e0fe4.1638530442.git.robin.murphy@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: gemini: NAS4220-B: fis-index-block with 128 KiB sectors

[ Upstream commit 4754eab7e5a78bdefe7a960c5c260c95ebbb5fa6 ]

Steven Maddox reported in the OpenWrt bugzilla, that his
RaidSonic IB-NAS4220-B was no longer booting with the new
OpenWrt 21.02 (uses linux 5.10's device-tree). However, it was
working with the previous OpenWrt 19.07 series (uses 4.14).

|[    5.548038] No RedBoot partition table detected in 30000000.flash
|[    5.618553] Searching for RedBoot partition table in 30000000.flash at offset 0x0
|[    5.739093] No RedBoot partition table detected in 30000000.flash
|...
|[    7.039504] Waiting for root device /dev/mtdblock3...

The provided bootlog shows that the RedBoot partition parser was
looking for the partition table "at offset 0x0". Which is strange
since the comment in the device-tree says it should be at 0xfe0000.

Further digging on the internet led to a review site that took
some useful PCB pictures of their review unit back in February 2009.
Their picture shows a Spansion S29GL128N11TFI01 flash chip.

>From Spansion's Datasheet:
"S29GL128N: One hundred twenty-eight 64 Kword (128 Kbyte) sectors"
Steven also provided a "cat /sys/class/mtd/mtd0/erasesize" from his
unit: "131072".

With the 128 KiB Sector/Erasesize in mind. This patch changes the
fis-index-block property to (0xfe0000 / 0x20000) = 0x7f.

Fixes: b5a923f8c739 ("ARM: dts: gemini: Switch to redboot partition parsing")
Reported-by: Steven Maddox <s.maddox@lantizia.me.uk>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Tested-by: Steven Maddox <s.maddox@lantizia.me.uk>
Link: https://lore.kernel.org/r/20211206004334.4169408-1-linus.walleij@linaro.org'
Bugzilla: https://bugs.openwrt.org/index.php?do=details&task_id=4137
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

spi: Fix incorrect cs_setup delay handling

[ Upstream commit 95c07247399536f83b89dc60cfe7b279d17e69f6 ]

Move the cs_setup delay to the end of spi_set_cs.

From include/linux/spi/spi.h:

 * @cs_setup: delay to be introduced by the controller after CS is
   asserted

The cs_setup delay needs to happen *after* CS is asserted, that is, at
the end of spi_set_cs, not at the beginning. Otherwise we're just
delaying before the SPI transaction starts at all, which isn't very
useful.

No drivers use this right now, but that is likely to change soon with an
upcoming Apple SPI HID transport driver.

Fixes: 25093bdeb6bc ("spi: implement SW control for CS times")
Signed-off-by: Hector Martin <marcan@marcan.st>
Link: https://lore.kernel.org/r/20211210170534.177139-1-marcan@marcan.st
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ath11k: Fix deleting uninitialized kernel timer during fragment cache flush

[ Upstream commit ba53ee7f7f38cf0592b8be1dcdabaf8f7535f8c1 ]

frag_timer will be created & initialized for stations when
they associate and will be deleted during every key installation
while flushing old fragments.

For AP interface self peer will be created and Group keys
will be installed for this peer, but there will be no real
Station entry & hence frag_timer won't be created and
initialized, deleting such uninitialized kernel timers causes below
warnings and backtraces printed with CONFIG_DEBUG_OBJECTS_TIMERS
enabled.

[ 177.828008] ODEBUG: assert_init not available (active state 0) object type: timer_list hint: 0x0
[ 177.836833] WARNING: CPU: 3 PID: 188 at lib/debugobjects.c:508 debug_print_object+0xb0/0xf0
[ 177.845185] Modules linked in: ath11k_pci ath11k qmi_helpers qrtr_mhi qrtr ns mhi
[ 177.852679] CPU: 3 PID: 188 Comm: hostapd Not tainted 5.14.0-rc3-32919-g4034139e1838-dirty #14
[ 177.865805] pstate: 60000005 (nZCv daif -PAN -UAO -TCO BTYPE=--)
[ 177.871804] pc : debug_print_object+0xb0/0xf0
[ 177.876155] lr : debug_print_object+0xb0/0xf0
[ 177.880505] sp : ffffffc01169b5a0
[ 177.883810] x29: ffffffc01169b5a0 x28: ffffff80081c2320 x27: ffffff80081c4078
[ 177.890942] x26: ffffff8003fe8f28 x25: ffffff8003de9890 x24: ffffffc01134d738
[ 177.898075] x23: ffffffc010948f20 x22: ffffffc010b2d2e0 x21: ffffffc01169b628
[ 177.905206] x20: ffffffc01134d700 x19: ffffffc010c80d98 x18: 00000000000003f6
[ 177.912339] x17: 203a657079742074 x16: 63656a626f202930 x15: 0000000000000152
[ 177.919471] x14: 0000000000000152 x13: 00000000ffffffea x12: ffffffc010d732e0
[ 177.926603] x11: 0000000000000003 x10: ffffffc010d432a0 x9 : ffffffc010d432f8
[ 177.933735] x8 : 000000000002ffe8 x7 : c0000000ffffdfff x6 : 0000000000000001
[ 177.940866] x5 : 0000000000000000 x4 : 0000000000000000 x3 : 00000000ffffffff
[ 177.947997] x2 : ffffffc010c93240 x1 : ffffff80023624c0 x0 : 0000000000000054
[ 177.955130] Call trace:
[ 177.957567] debug_print_object+0xb0/0xf0
[ 177.961570] debug_object_assert_init+0x124/0x178
[ 177.966269] try_to_del_timer_sync+0x1c/0x70
[ 177.970536] del_timer_sync+0x30/0x50
[ 177.974192] ath11k_peer_frags_flush+0x34/0x68 [ath11k]
[ 177.979439] ath11k_mac_op_set_key+0x1e4/0x338 [ath11k]
[ 177.984673] ieee80211_key_enable_hw_accel+0xc8/0x3d0
[ 177.989722] ieee80211_key_replace+0x360/0x740
[ 177.994160] ieee80211_key_link+0x16c/0x210
[ 177.998337] ieee80211_add_key+0x138/0x338
[ 178.002426] nl80211_new_key+0xfc/0x258
[ 178.006257] genl_family_rcv_msg_doit.isra.17+0xd8/0x120
[ 178.011565] genl_rcv_msg+0xd8/0x1c8
[ 178.015134] netlink_rcv_skb+0x38/0xf8
[ 178.018877] genl_rcv+0x34/0x48
[ 178.022012] netlink_unicast+0x174/0x230
[ 178.025928] netlink_sendmsg+0x188/0x388
[ 178.029845] ____sys_sendmsg+0x218/0x250
[ 178.033763] ___sys_sendmsg+0x68/0x90
[ 178.037418] __sys_sendmsg+0x44/0x88
[ 178.040988] __arm64_sys_sendmsg+0x20/0x28
[ 178.045077] invoke_syscall.constprop.5+0x54/0xe0
[ 178.049776] do_el0_svc+0x74/0xc0
[ 178.053084] el0_svc+0x10/0x18
[ 178.056133] el0t_64_sync_handler+0x88/0xb0
[ 178.060310] el0t_64_sync+0x148/0x14c
[ 178.063966] ---[ end trace 8a5cf0bf9d34a058 ]---

Add changes to not to delete frag timer for peers during
group key installation.

Tested on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-01092-QCAHKSWPL_SILICONZ-1

Fixes: c3944a562102 ("ath11k: Clear the fragment cache during key install")
Signed-off-by: Rameshkumar Sundaram <quic_ramess@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/1639071421-25078-1-git-send-email-quic_ramess@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: hisilicon/qm - fix incorrect return value of hisi_qm_resume()

[ Upstream commit 3f9dd4c802b96626e869b2d29c8e401dabadd23e ]

When hisi_qm_resume() returns 0, it indicates that the device has started
successfully.  If the device fails to start, hisi_qm_resume() needs to
return the actual error code to the caller instead of 0.

Fixes: d7ea53395b72 ("crypto: hisilicon - add runtime PM ops")
Signed-off-by: Weili Qian <qianweili@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32 - Revert broken pm_runtime_resume_and_get changes

[ Upstream commit 3d6b661330a7954d8136df98160d525eb04dcd6a ]

We should not call pm_runtime_resume_and_get where the reference
count is expected to be incremented unconditionally.  This patch
reverts these calls to the original unconditional get_sync call.

Reported-by: Heiner Kallweit <hkallweit1@gmail.com>
Fixes: 747bf30fd944 ("crypto: stm32/cryp - Fix PM reference leak...")
Fixes: 1cb3ad701970 ("crypto: stm32/hash - Fix PM reference leak...")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - fix bugs and crash in tests

[ Upstream commit 4b898d5cfa4d9a0ad5bc82cb5eafdc092394c6a9 ]

Extra crypto manager auto test were crashing or failling due
to 2 reasons:
- block in a dead loop (dues to issues in cipher end process management)
- crash due to read/write unmapped memory (this crash was also reported
when using openssl afalg engine)

Rework interrupt management, interrupts are masked as soon as they are
no more used: if input buffer is fully consumed, "Input FIFO not full"
interrupt is masked and if output buffer is full, "Output FIFO not
empty" interrupt is masked.
And crypto request finish when input *and* outpout buffer are fully
read/write.

About the crash due to unmapped memory, using scatterwalk_copychunks()
that will map and copy each block fix the issue.
Using this api and copying full block will also fix unaligned data
access, avoid early copy of in/out buffer, and make useless the extra
alignment constraint.

Fixes: 9e054ec21ef8 ("crypto: stm32 - Support for STM32 CRYP crypto module")

Reported-by: Marek Vasut <marex@denx.de>
Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - fix lrw chaining mode

[ Upstream commit fa97dc2d48b476ea98199d808d3248d285987e99 ]

This fixes the lrw autotest if lrw uses the CRYP as the AES block cipher
provider (as ecb(aes)). At end of request, CRYP should not update the IV
in case of ECB chaining mode. Indeed the ECB chaining mode never uses
the IV, but the software LRW chaining mode uses the IV field as
a counter and due to the (unexpected) update done by CRYP while the AES
block process, the counter get a wrong value when the IV overflow.

Fixes: 5f49f18d27cd ("crypto: stm32/cryp - update to return iv_out")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - fix double pm exit

[ Upstream commit 6c12e742785bf9333faf60bfb96575bdd763448e ]

Delete extraneous lines in probe error handling code: pm was
disabled twice.

Fixes: 65f9aa36ee47 ("crypto: stm32/cryp - Add power management support")

Reported-by: Marek Vasut <marex@denx.de>
Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - check early input data

[ Upstream commit 39e6e699c7fb92bdb2617b596ca4a4ea35c5d2a7 ]

Some auto tests failed because driver wasn't returning the expected
error with some input size/iv value/tag size.
Now:
 Return 0 early for empty buffer. (We don't need to start the engine for
 an empty input buffer).
 Accept any valid authsize for gcm(aes).
 Return -EINVAL if iv for ccm(aes) is invalid.
 Return -EINVAL if buffer size is a not a multiple of algorithm block size.

Fixes: 9e054ec21ef8 ("crypto: stm32 - Support for STM32 CRYP crypto module")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - fix xts and race condition in crypto_engine requests

[ Upstream commit d703c7a994ee34b7fa89baf21631fca0aa9f17fc ]

Don't erase key:
If key is erased before the crypto_finalize_.*_request() call, some
pending process will run with a key={ 0 }.
Moreover if the key is reset at end of request, it breaks xts chaining
mode, as for last xts block (in case input len is not a multiple of
block) a new AES request is started without calling again set_key().

Fixes: 9e054ec21ef8 ("crypto: stm32 - Support for STM32 CRYP crypto module")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: stm32/cryp - fix CTR counter carry

[ Upstream commit 41c76690b0990efacd15d35cfb4e77318cd80ebb ]

STM32 CRYP hardware doesn't manage CTR counter bigger than max U32, as
a workaround, at each block the current IV is saved, if the saved IV
lower u32 is 0xFFFFFFFF, the full IV is manually incremented, and set
in hardware.
Fixes: bbb2832620ac ("crypto: stm32 - Fix sparse warnings")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@foss.st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: harness: avoid false negatives if test has no ASSERTs

[ Upstream commit 3abedf4646fdc0036fcb8ebbc3b600667167fafe ]

Test can fail either immediately when ASSERT() failed or at the
end if one or more EXPECT() was not met. The exact return code
is decided based on the number of successful ASSERT()s.

If test has no ASSERT()s, however, the return code will be 0,
as if the test did not fail. Start counting ASSERT()s from 1.

Fixes: 369130b63178 ("selftests: Enhance kselftest_harness.h to print which assert failed")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: clone3: clone3: add case CLONE3_ARGS_NO_TEST

[ Upstream commit a531b0c23c0fc68ad758cc31a74cf612a4dafeb0 ]

Building selftests/clone3 with clang warns about enumeration not handled
in switch case:

clone3.c:54:10: warning: enumeration value 'CLONE3_ARGS_NO_TEST' not handled in switch [-Wswitch]
        switch (test_mode) {
                ^

Add the missing switch case with a comment.

Fixes: 17a810699c18 ("selftests: add tests for clone3()")
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/uaccess: Move variable into switch case statement

[ Upstream commit 61646ca83d3889696f2772edaff122dd96a2935e ]

When building with automatic stack variable initialization, GCC 12
complains about variables defined outside of switch case statements.
Move the variable into the case that uses it, which silences the warning:

./arch/x86/include/asm/uaccess.h:317:23: warning: statement will never be executed [-Wswitch-unreachable]
  317 |         unsigned char x_u8__; \
      |                       ^~~~~~

Fixes: 865c50e1d279 ("x86/uaccess: utilize CONFIG_CC_HAS_ASM_GOTO_OUTPUT")
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20211209043456.1377875-1-keescook@chromium.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

xfrm: fix a small bug in xfrm_sa_len()

[ Upstream commit 7770a39d7c63faec6c4f33666d49a8cb664d0482 ]

copy_user_offload() will actually push a struct struct xfrm_user_offload,
which is different than (struct xfrm_state *)->xso
(struct xfrm_state_offload)

Fixes: d77e38e612a01 ("xfrm: Add an IPsec hardware offloading API")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mwifiex: Fix possible ABBA deadlock

[ Upstream commit 1b8bb8919ef81bfc8873d223b9361f1685f2106d ]

Quoting Jia-Ju Bai <baijiaju1990@gmail.com>:

  mwifiex_dequeue_tx_packet()
     spin_lock_bh(&priv->wmm.ra_list_spinlock); --> Line 1432 (Lock A)
     mwifiex_send_addba()
       spin_lock_bh(&priv->sta_list_spinlock); --> Line 608 (Lock B)

  mwifiex_process_sta_tx_pause()
     spin_lock_bh(&priv->sta_list_spinlock); --> Line 398 (Lock B)
     mwifiex_update_ralist_tx_pause()
       spin_lock_bh(&priv->wmm.ra_list_spinlock); --> Line 941 (Lock A)

Similar report for mwifiex_process_uap_tx_pause().

While the locking expectations in this driver are a bit unclear, the
Fixed commit only intended to protect the sta_ptr, so we can drop the
lock as soon as we're done with it.

IIUC, this deadlock cannot actually happen, because command event
processing (which calls mwifiex_process_sta_tx_pause()) is
sequentialized with TX packet processing (e.g.,
mwifiex_dequeue_tx_packet()) via the main loop (mwifiex_main_process()).
But it's good not to leave this potential issue lurking.

Fixes: f0f7c2275fb9 ("mwifiex: minor cleanups w/ sta_list_spinlock in cfg80211.c")
Cc: Douglas Anderson <dianders@chromium.org>
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Link: https://lore.kernel.org/linux-wireless/0e495b14-efbb-e0da-37bd-af6bd677ee2c@gmail.com/
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/YaV0pllJ5p/EuUat@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/dsi: fix initialization in the bonded DSI case

[ Upstream commit 92cb1bedde9dba78d802fe2510949743a2581aed ]

Commit 739b4e7756d3 ("drm/msm/dsi: Fix an error code in
msm_dsi_modeset_init()") changed msm_dsi_modeset_init() to return an
error code in case msm_dsi_manager_validate_current_config() returns
false. However this is not an error case, but a slave DSI of the bonded
DSI link. In this case msm_dsi_modeset_init() should return 0, but just
skip connector and bridge initialization.

To reduce possible confusion, drop the
msm_dsi_manager_validate_current_config() function, and specif 'bonded
&& !master' condition directly in the msm_dsi_modeset_init().

Fixes: 739b4e7756d3 ("drm/msm/dsi: Fix an error code in msm_dsi_modeset_init()")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Link: https://lore.kernel.org/r/20211125180114.561278-1-dmitry.baryshkov@linaro.org
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wcn36xx: Fix max channels retrieval

[ Upstream commit 09cab4308bf9b8076ee4a3c56015daf9ef9cb23e ]

Kernel test robot reported:drivers/net/wireless/ath/wcn36xx/smd.c:943:33:
   sparse: sparse: cast truncates bits from constant value (780 becomes 80)

The 'channels' field is not a simple u8 array but an array of
channel_params. Using sizeof for retrieving the max number of
channels is then wrong.

In practice, it was not an issue, because the sizeof returned
value is 780, which is truncated in min_t (u8) to 80, which is
the value we expect...

Fix that properly using ARRAY_SIZE instead of sizeof.

Fixes: d707f812bb05 ("wcn36xx: Channel list update before hardware scan")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/1638435732-14657-1-git-send-email-loic.poulain@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

rcu/exp: Mark current CPU as exp-QS in IPI loop second pass

[ Upstream commit 81f6d49cce2d2fe507e3fddcc4a6db021d9c2e7b ]

Expedited RCU grace periods invoke sync_rcu_exp_select_node_cpus(), which
takes two passes over the leaf rcu_node structure's CPUs.  The first
pass gathers up the current CPU and CPUs that are in dynticks idle mode.
The workqueue will report a quiescent state on their behalf later.
The second pass sends IPIs to the rest of the CPUs, but excludes the
current CPU, incorrectly assuming it has been included in the first
pass's list of CPUs.

Unfortunately the current CPU may have changed between the first and
second pass, due to the fact that the various rcu_node structures'
->lock fields have been dropped, thus momentarily enabling preemption.
This means that if the second pass's CPU was not on the first pass's
list, it will be ignored completely.  There will be no IPI sent to
it, and there will be no reporting of quiescent states on its behalf.
Unfortunately, the expedited grace period will nevertheless be waiting
for that CPU to report a quiescent state, but with that CPU having no
reason to believe that such a report is needed.

The result will be an expedited grace period stall.

Fix this by no longer excluding the current CPU from consideration during
the second pass.

Fixes: b9ad4d6ed18e ("rcu: Avoid self-IPI in sync_rcu_exp_select_node_cpus()")
Reviewed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Uladzislau Rezki <urezki@gmail.com>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Joel Fernandes <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/dp: displayPort driver need algorithm rational

[ Upstream commit 53d22794711ad630f40d59dd726bd260d77d585f ]

Let's select RATIONAL with dp driver. avoid like:

[...]
x86_64-linux-gnu-ld: drivers/gpu/drm/msm/dp/dp_catalog.o: in function `dp_catalog_ctrl_config_msa':
dp_catalog.c:(.text+0x57e): undefined reference to `rational_best_approximation'

Fixes: c943b4948b58 ("drm/msm/dp: add displayPort driver support")
Reported-by: kernelbot <kernel-bot@kylinos.cn>
Signed-off-by: Jackie Liu <liuyun01@kylinos.cn>
Link: https://lore.kernel.org/r/20211110070950.3355597-2-liu.yun@linux.dev
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/msm/gpu: Don't allow zero fence_id

[ Upstream commit ca3ffcbeb0c866d9b0cb38eaa2bd4416597b5966 ]

Elsewhere we treat zero as "no fence" and __msm_gem_submit_destroy()
skips removal from fence_idr.  We could alternately change this to use
negative values for "no fence" but I think it is more clear to not allow
zero as a valid fence_id.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Fixes: a61acbbe9cf8 ("drm/msm: Track "seqno" fences by idr")
Link: https://lore.kernel.org/r/20211129182344.292609-1-robdclark@gmail.com
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Fix out of bounds access on DNC31 stream encoder regs

[ Upstream commit d374d3b493215d637b9e7be12a93f22caf4c1f97 ]

[Why]
During dcn31_stream_encoder_create, if PHYC/D get remapped to F/G on B0
then we'll index 5 or 6 into a array of length 5 - leading to an
access violation on some configs during device creation.

[How]
Software won't be touching PHYF/PHYG directly, so just extend the
array to cover all possible engine IDs.

Even if it does by try to access one of these registers by accident
the offset will be 0 and we'll get a warning during the access.

Fixes: 2fe9a0e1173f ("drm/amd/display: Fix DCN3 B0 DP Alt Mapping")
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amd/display: Fix bug in debugfs crc_win_update entry

[ Upstream commit 4bef85d4c9491415b7931407b07f24841c1e0390 ]

[Why]
crc_rd_wrk shouldn't be null in crc_win_update_set(). Current programming
logic is inconsistent in crc_win_update_set().

[How]
Initially, return if crc_rd_wrk is NULL. Later on, we can use member of
crc_rd_wrk safely.

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Fixes: 9a65df193108 ("drm/amd/display: Use PSP TA to read out crc")

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Pavle Kotarac <Pavle.Kotarac@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: btusb: Handle download_firmware failure cases

[ Upstream commit 00c0ee9850b7b0cb7c40b8daba806ae2245e59d4 ]

For Mediatek chipset, it can not enabled if there are something wrong
in btmtk_setup_firmware_79xx(). Thus, the process must be terminated
and returned error code.

Fixes: fc342c4dc4087 ("Bluetooth: btusb: Add protocol support for MediaTek MT7921U USB devices")
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Mark Chen <mark-yw.chen@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: MGMT: Use hci_dev_test_and_{set,clear}_flag

[ Upstream commit 6f59f991b4e735323f099ac6490e725ae8c750a5 ]

This make use of hci_dev_test_and_{set,clear}_flag instead of doing 2
operations in a row.

Fixes: cbbdfa6f33198 ("Bluetooth: Enable controller RPA resolution using Experimental feature")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: refactor set_exp_feature with a feature table

[ Upstream commit 93fb70bc112e922def6e50b37e20ccfce0c67c0a ]

This patch refactors the set_exp_feature with a feature table
consisting of UUIDs and the corresponding callback functions.
In this way, a new experimental feature setting function can be
simply added with its UUID and callback function.

Signed-off-by: Joseph Hwang <josephsih@chromium.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ath10k: Fix the MTU size on QCA9377 SDIO

[ Upstream commit 09b8cd69edcf2be04a781e1781e98e52a775c9ad ]

On an imx6dl-pico-pi board with a QCA9377 SDIO chip, simply trying to
connect via ssh to another machine causes:

[   55.824159] ath10k_sdio mmc1:0001:1: failed to transmit packet, dropping: -12
[   55.832169] ath10k_sdio mmc1:0001:1: failed to submit frame: -12
[   55.838529] ath10k_sdio mmc1:0001:1: failed to push frame: -12
[   55.905863] ath10k_sdio mmc1:0001:1: failed to transmit packet, dropping: -12
[   55.913650] ath10k_sdio mmc1:0001:1: failed to submit frame: -12
[   55.919887] ath10k_sdio mmc1:0001:1: failed to push frame: -12

, leading to an ssh connection failure.

One user inspected the size of frames on Wireshark and reported
the followig:

"I was able to narrow the issue down to the mtu. If I set the mtu for
the wlan0 device to 1486 instead of 1500, the issue does not happen.

The size of frames that I see on Wireshark is exactly 1500 after
setting it to 1486."

Clearing the HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE avoids the problem and
the ssh command works successfully after that.

Introduce a 'credit_size_workaround' field to ath10k_hw_params for
the QCA9377 SDIO, so that the HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE
is not set in this case.

Tested with QCA9377 SDIO with firmware WLAN.TF.1.1.1-00061-QCATFSWPZ-1.

Fixes: 2f918ea98606 ("ath10k: enable alt data of TX path for sdio")
Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20211124131047.713756-1-festevam@denx.de
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/rt: Try to restart rt period timer when rt runtime exceeded

[ Upstream commit 9b58e976b3b391c0cf02e038d53dd0478ed3013c ]

When rt_runtime is modified from -1 to a valid control value, it may
cause the task to be throttled all the time. Operations like the following
will trigger the bug. E.g:

  1. echo -1 > /proc/sys/kernel/sched_rt_runtime_us
  2. Run a FIFO task named A that executes while(1)
  3. echo 950000 > /proc/sys/kernel/sched_rt_runtime_us

When rt_runtime is -1, The rt period timer will not be activated when task
A enqueued. And then the task will be throttled after setting rt_runtime to
950,000. The task will always be throttled because the rt period timer is
not activated.

Fixes: d0b27fa77854 ("sched: rt-group: synchonised bandwidth period")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Li Hua <hucool.lihua@huawei.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20211203033618.11895-1-hucool.lihua@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wireless: iwlwifi: Fix a double free in iwl_txq_dyn_alloc_dma

[ Upstream commit f973795a8d19cbf3d03807704eb7c6ff65788d5a ]

In iwl_txq_dyn_alloc_dma, txq->tfds is freed at first time by:
iwl_txq_alloc()->goto err_free_tfds->dma_free_coherent(). But
it forgot to set txq->tfds to NULL.

Then the txq->tfds is freed again in iwl_txq_dyn_alloc_dma by:
goto error->iwl_txq_gen2_free_memory()->dma_free_coherent().

My patch sets txq->tfds to NULL after the first free to avoid the
double free.

Fixes: 0cd1ad2d7fd41 ("iwlwifi: move all bus-independent TX functions to common code")
Signed-off-by: Lv Yunlong <lyl2019@mail.ustc.edu.cn>
Link: https://lore.kernel.org/r/20210403054755.4781-1-lyl2019@mail.ustc.edu.cn
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: si2157: Fix "warm" tuner state detection

[ Upstream commit a6441ea29cb2c9314654e093a1cd8020b9b851c8 ]

Commit e955f959ac52 ("media: si2157: Better check for running tuner in
init") completely broke the "warm" tuner detection of the si2157 driver
due to a simple endian error: The Si2157 CRYSTAL_TRIM property code is
0x0402 and needs to be transmitted LSB first. However, it was inserted
MSB first, causing the warm detection to always fail and spam the kernel
log with tuner initialization messages each time the DVB frontend
device was closed and reopened:

[  312.215682] si2157 16-0060: found a 'Silicon Labs Si2157-A30'
[  312.264334] si2157 16-0060: firmware version: 3.0.5
[  342.248593] si2157 16-0060: found a 'Silicon Labs Si2157-A30'
[  342.295743] si2157 16-0060: firmware version: 3.0.5
[  372.328574] si2157 16-0060: found a 'Silicon Labs Si2157-A30'
[  372.385035] si2157 16-0060: firmware version: 3.0.5

Also, the reinitializations were observed disturb _other_ tuners on
multi-tuner cards such as the Hauppauge WinTV-QuadHD, leading to missed
or errored packets when one of the other DVB frontend devices on that
card was opened.

Fix the order of the property code bytes to make the warm detection work
again, also reducing the tuner initialization message in the kernel log
to once per power-on, as well as fixing the interference with other
tuners.

Link: https://lore.kernel.org/linux-media/trinity-2a86eb9d-6264-4387-95e1-ba7b79a4050f-1638392923493@3c-app-gmx-bap03
Fixes: e955f959ac52 ("media: si2157: Better check for running tuner in init")
Reported-by: Robert Schlabbach <robert_s@gmx.net>
Signed-off-by: Robert Schlabbach <robert_s@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: saa7146: mxb: Fix a NULL pointer dereference in mxb_attach()

[ Upstream commit 0407c49ebe330333478440157c640fffd986f41b ]

In mxb_attach(dev, info), saa7146_vv_init() is called to allocate a
new memory for dev->vv_data. saa7146_vv_release() will be called on
failure of mxb_probe(dev). There is a dereference of dev->vv_data
in saa7146_vv_release(), which could lead to a NULL pointer dereference
on failure of saa7146_vv_init().

Fix this bug by adding a check of saa7146_vv_init().

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_VIDEO_MXB=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: 03b1930efd3c ("V4L/DVB: saa7146: fix regression of the av7110/budget-av driver")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: dib8000: Fix a memleak in dib8000_init()

[ Upstream commit 8dbdcc7269a83305ee9d677b75064d3530a48ee2 ]

In dib8000_init(), the variable fe is not freed or passed out on the
failure of dib8000_identify(&state->i2c), which could lead to a memleak.

Fix this bug by adding a kfree of fe in the error path.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DVB_DIB8000=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: 77e2c0f5d471 ("V4L/DVB (12900): DiB8000: added support for DiBcom ISDB-T/ISDB-Tsb demodulator DiB8000")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

samples: bpf: Fix 'unknown warning group' build warning on Clang

[ Upstream commit 6f670d06e47c774bc065aaa84a527a4838f34bd8 ]

Clang doesn't have 'stringop-truncation' group like GCC does, and
complains about it when building samples which use xdp_sample_user
infra:

 samples/bpf/xdp_sample_user.h:48:32: warning: unknown warning group '-Wstringop-truncation', ignored [-Wunknown-warning-option]
 #pragma GCC diagnostic ignored "-Wstringop-truncation"
                                ^
[ repeat ]

Those are harmless, but avoidable when guarding it with ifdef.
I could guard push/pop as well, but this would require one more
ifdef cruft around a single line which I don't think is reasonable.

Fixes: 156f886cf697 ("samples: bpf: Add basic infrastructure for XDP samples")
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/bpf/20211203195004.5803-3-alexandr.lobakin@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

samples: bpf: Fix xdp_sample_user.o linking with Clang

[ Upstream commit e64fbcaa7a666f16329b1c67af15ea501bc84586 ]

Clang (13) doesn't get the jokes about specifying libraries to link in
cclags of individual .o objects:

clang-13: warning: -lm: 'linker' input unused [-Wunused-command-line-argument]
[ ... ]
  LD  samples/bpf/xdp_redirect_cpu
  LD  samples/bpf/xdp_redirect_map_multi
  LD  samples/bpf/xdp_redirect_map
  LD  samples/bpf/xdp_redirect
  LD  samples/bpf/xdp_monitor
/usr/bin/ld: samples/bpf/xdp_sample_user.o: in function `sample_summary_print':
xdp_sample_user.c:(.text+0x84c): undefined reference to `floor'
/usr/bin/ld: xdp_sample_user.c:(.text+0x870): undefined reference to `ceil'
/usr/bin/ld: xdp_sample_user.c:(.text+0x8cf): undefined reference to `floor'
/usr/bin/ld: xdp_sample_user.c:(.text+0x8f3): undefined reference to `ceil'
[ more ]

Specify '-lm' as ldflags for all xdp_sample_user.o users in the main
Makefile and remove it from ccflags of ^ in Makefile.target -- just
like it's done for all other samples. This works with all compilers.

Fixes: 6e1051a54e31 ("samples: bpf: Convert xdp_monitor to XDP samples helper")
Fixes: b926c55d856c ("samples: bpf: Convert xdp_redirect to XDP samples helper")
Fixes: e531a220cc59 ("samples: bpf: Convert xdp_redirect_cpu to XDP samples helper")
Fixes: bbe65865aa05 ("samples: bpf: Convert xdp_redirect_map to XDP samples helper")
Fixes: 594a116b2aa1 ("samples: bpf: Convert xdp_redirect_map_multi to XDP samples helper")
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/bpf/20211203195004.5803-2-alexandr.lobakin@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

samples/bpf: Clean up samples/bpf build failes

[ Upstream commit 527024f7aeb683ce7ef49b07ef7ce9ecf015288d ]

Remove xdp_samples_user.o rule redefinition which generates Makefile
warning and instead override TPROGS_CFLAGS. This seems to work fine when
building inside selftests/bpf.

That was one big head-scratcher before I found that generic
Makefile.target hid this surprising specialization for for xdp_samples_user.o.

Main change is to use actual locally installed libbpf headers.

Also drop printk macro re-definition (not even used!).

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20211201232824.3166325-8-andrii@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

samples/bpf: Install libbpf headers when building

[ Upstream commit 3f7a3318a7c60947e27be372950840a5eab976d0 ]

API headers from libbpf should not be accessed directly from the source
directory. Instead, they should be exported with "make install_headers".
Make sure that samples/bpf/Makefile installs the headers properly when
building.

The object compiled from and exported by libbpf are now placed into a
subdirectory of sample/bpf/ instead of remaining in tools/lib/bpf/. We
attempt to remove this directory on "make clean". However, the "clean"
target re-enters the samples/bpf/ directory from the root of the
repository ("$(MAKE) -C ../../ M=$(CURDIR) clean"), in such a way that
$(srctree) and $(src) are not defined, making it impossible to use
$(LIBBPF_OUTPUT) and $(LIBBPF_DESTDIR) in the recipe. So we only attempt
to clean $(CURDIR)/libbpf, which is the default value.

Add a dependency on libbpf's headers for the $(TRACE_HELPERS).

We also change the output directory for bpftool, to place the generated
objects under samples/bpf/bpftool/ instead of building in bpftool's
directory directly. Doing so, we make sure bpftool reuses the libbpf
library previously compiled and installed.

Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211007194438.34443-10-quentin@isovalent.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: mte: DC {GVA,GZVA} shouldn't be used when DCZID_EL0.DZP == 1

[ Upstream commit 685e2564daa1493053fcd7f1dbed38b35ee2f3cb ]

Currently, mte_set_mem_tag_range() and mte_zero_clear_page_tags() use
DC {GVA,GZVA} unconditionally.  But, they should make sure that
DCZID_EL0.DZP, which indicates whether or not use of those instructions
is prohibited, is zero when using those instructions.
Use ST{G,ZG,Z2G} instead when DCZID_EL0.DZP == 1.

Fixes: 013bb59dbb7c ("arm64: mte: handle tags zeroing at page allocation time")
Fixes: 3d0cca0b02ac ("kasan: speed up mte_set_mem_tag_range")
Signed-off-by: Reiji Watanabe <reijiw@google.com>
Link: https://lore.kernel.org/r/20211206004736.1520989-3-reijiw@google.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: clear_page() shouldn't use DC ZVA when DCZID_EL0.DZP == 1

[ Upstream commit f0616abd4e67143b45b04b565839148458857347 ]

Currently, clear_page() uses DC ZVA instruction unconditionally.  But it
should make sure that DCZID_EL0.DZP, which indicates whether or not use
of DC ZVA instruction is prohibited, is zero when using the instruction.
Use STNP instead when DCZID_EL0.DZP == 1.

Fixes: f27bb139c387 ("arm64: Miscellaneous library functions")
Signed-off-by: Reiji Watanabe <reijiw@google.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/20211206004736.1520989-2-reijiw@google.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Remove config check to enable bpf support for branch records

[ Upstream commit db52f57211b4e45f0ebb274e2c877b211dc18591 ]

Branch data available to BPF programs can be very useful to get stack traces
out of userspace application.

Commit fff7b64355ea ("bpf: Add bpf_read_branch_records() helper") added BPF
support to capture branch records in x86. Enable this feature also for other
architectures as well by removing checks specific to x86.

If an architecture doesn't support branch records, bpf_read_branch_records()
still has appropriate checks and it will return an -EINVAL in that scenario.
Based on UAPI helper doc in include/uapi/linux/bpf.h, unsupported architectures
should return -ENOENT in such case. Hence, update the appropriate check to
return -ENOENT instead.

Selftest 'perf_branches' result on power9 machine which has the branch stacks
support:

 - Before this patch:

  [command]# ./test_progs -t perf_branches
   #88/1 perf_branches/perf_branches_hw:FAIL
   #88/2 perf_branches/perf_branches_no_hw:OK
   #88 perf_branches:FAIL
  Summary: 0/1 PASSED, 0 SKIPPED, 1 FAILED

 - After this patch:

  [command]# ./test_progs -t perf_branches
   #88/1 perf_branches/perf_branches_hw:OK
   #88/2 perf_branches/perf_branches_no_hw:OK
   #88 perf_branches:OK
  Summary: 1/2 PASSED, 0 SKIPPED, 0 FAILED

Selftest 'perf_branches' result on power9 machine which doesn't have branch
stack report:

 - After this patch:

  [command]# ./test_progs -t perf_branches
   #88/1 perf_branches/perf_branches_hw:SKIP
   #88/2 perf_branches/perf_branches_no_hw:OK
   #88 perf_branches:OK
  Summary: 1/1 PASSED, 1 SKIPPED, 0 FAILED

Fixes: fff7b64355eac ("bpf: Add bpf_read_branch_records() helper")
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kajol Jain <kjain@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211206073315.77432-1-kjain@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Disallow BPF_LOG_KERNEL log level for bpf(BPF_BTF_LOAD)

[ Upstream commit 866de407444398bc8140ea70de1dba5f91cc34ac ]

BPF_LOG_KERNEL is only used internally, so disallow bpf_btf_load()
to set log level as BPF_LOG_KERNEL. The same checking has already
been done in bpf_check(), so factor out a helper to check the
validity of log attributes and use it in both places.

Fixes: 8580ac9404f6 ("bpf: Process in-kernel BTF")
Signed-off-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20211203053001.740945-1-houtao1@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Adjust BTF log size limit.

[ Upstream commit c5a2d43e998a821701029f23e25b62f9188e93ff ]

Make BTF log size limit to be the same as the verifier log size limit.
Otherwise tools that progressively increase log size and use the same log
for BTF loading and program loading will be hitting hard to debug EINVAL.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-7-alexei.starovoitov@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/fair: Fix per-CPU kthread and wakee stacking for asym CPU capacity

[ Upstream commit 014ba44e8184e1acf93e0cbb7089ee847802f8f0 ]

select_idle_sibling() has a special case for tasks woken up by a per-CPU
kthread where the selected CPU is the previous one. For asymmetric CPU
capacity systems, the assumption was that the wakee couldn't have a
bigger utilization during task placement than it used to have during the
last activation. That was not considering uclamp.min which can completely
change between two task activations and as a consequence mandates the
fitness criterion asym_fits_capacity(), even for the exit path described
above.

Fixes: b4c9c9f15649 ("sched/fair: Prefer prev cpu in asymmetric wakeup path")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lkml.kernel.org/r/20211129173115.4006346-1-vincent.donnefort@arm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/fair: Fix detection of per-CPU kthreads waking a task

[ Upstream commit 8b4e74ccb582797f6f0b0a50372ebd9fd2372a27 ]

select_idle_sibling() has a special case for tasks woken up by a per-CPU
kthread, where the selected CPU is the previous one. However, the current
condition for this exit path is incomplete. A task can wake up from an
interrupt context (e.g. hrtimer), while a per-CPU kthread is running. A
such scenario would spuriously trigger the special case described above.
Also, a recent change made the idle task like a regular per-CPU kthread,
hence making that situation more likely to happen
(is_per_cpu_kthread(swapper) being true now).

Checking for task context makes sure select_idle_sibling() will not
interpret a wake up from any other context as a wake up by a per-CPU
kthread.

Fixes: 52262ee567ad ("sched/fair: Allow a per-CPU kthread waking a task to stack on the same CPU, to fix XFS performance regression")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lore.kernel.org/r/20211201143450.479472-1-vincent.donnefort@arm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Fix the test_task_vma selftest to support output shorter than 1 kB

[ Upstream commit da54ab14953c38d98cb3e34c564c06c3739394b2 ]

The test for bpf_iter_task_vma assumes that the output will be longer
than 1 kB, as the comment above the loop says. Due to this assumption,
the loop becomes infinite if the output turns to be shorter than 1 kB.
The return value of read_fd_into_buffer is 0 when the end of file was
reached, and len isn't being increased any more.

This commit adds a break on EOF to handle short output correctly. For
the reference, this is the contents that I get when running test_progs
under vmtest.sh, and it's shorter than 1 kB:

00400000-00401000 r--p 00000000 fe:00 25867     /root/bpf/test_progs
00401000-00674000 r-xp 00001000 fe:00 25867     /root/bpf/test_progs
00674000-0095f000 r--p 00274000 fe:00 25867     /root/bpf/test_progs
0095f000-00983000 r--p 0055e000 fe:00 25867     /root/bpf/test_progs
00983000-00a8a000 rw-p 00582000 fe:00 25867     /root/bpf/test_progs
00a8a000-0484e000 rw-p 00000000 00:00 0
7f6c64000000-7f6c64021000 rw-p 00000000 00:00 0
7f6c64021000-7f6c68000000 ---p 00000000 00:00 0
7f6c6ac8f000-7f6c6ac90000 r--s 00000000 00:0d 8032
anon_inode:bpf-map
7f6c6ac90000-7f6c6ac91000 ---p 00000000 00:00 0
7f6c6ac91000-7f6c6b491000 rw-p 00000000 00:00 0
7f6c6b491000-7f6c6b492000 r--s 00000000 00:0d 8032
anon_inode:bpf-map
7f6c6b492000-7f6c6b493000 rw-s 00000000 00:0d 8032
anon_inode:bpf-map
7ffc1e23d000-7ffc1e25e000 rw-p 00000000 00:00 0
7ffc1e3b8000-7ffc1e3bc000 r--p 00000000 00:00 0
7ffc1e3bc000-7ffc1e3bd000 r-xp 00000000 00:00 0
7fffffffe000-7ffffffff000 --xp 00000000 00:00 0

Fixes: e8168840e16c ("selftests/bpf: Add test for bpf_iter_task_vma")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211130181811.594220-1-maximmi@nvidia.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

Bluetooth: btmtksdio: fix resume failure

[ Upstream commit 561ae1d46a8ddcbc13162d5771f5ed6c8249e730 ]

btmtksdio have to rely on MMC_PM_KEEP_POWER in pm_flags to avoid that
SDIO power is being shut off during the device is in suspend. That fixes
the SDIO command fails to access the bus after the device is resumed.

Fixes: 7f3c563c575e7 ("Bluetooth: btmtksdio: Add runtime PM support to SDIO based Bluetooth")
Co-developed-by: Mark-yw Chen <mark-yw.chen@mediatek.com>
Signed-off-by: Mark-yw Chen <mark-yw.chen@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

staging: rtl8192e: rtllib_module: fix error handle case in alloc_rtllib()

[ Upstream commit e730cd57ac2dfe94bca0f14a3be8e1b21de41a9c ]

Some variables are leaked in the error handling in alloc_rtllib(), free
the variables in the error path.

Fixes: 94a799425eee ("From: wlanfae <wlanfae@realtek.com>")
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20211202030704.2425621-3-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

staging: rtl8192e: return error code from rtllib_softmac_init()

[ Upstream commit 68bf78ff59a0891eb1239948e94ce10f73a9dd30 ]

If it fails to allocate 'dot11d_info', rtllib_softmac_init()
should return error code. And remove unneccessary error message.

Fixes: 94a799425eee ("From: wlanfae <wlanfae@realtek.com>")
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20211202030704.2425621-2-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

floppy: Fix hang in watchdog when disk is ejected

[ Upstream commit fb48febce7e30baed94dd791e19521abd2c3fd83 ]

When the watchdog detects a disk change, it calls cancel_activity(),
which in turn tries to cancel the fd_timer delayed work.

In the above scenario, fd_timer_fn is set to fd_watchdog(), meaning
it is trying to cancel its own work.
This results in a hang as cancel_delayed_work_sync() is waiting for the
watchdog (itself) to return, which never happens.

This can be reproduced relatively consistently by attempting to read a
broken floppy, and ejecting it while IO is being attempted and retried.

To resolve this, this patch calls cancel_delayed_work() instead, which
cancels the work without waiting for the watchdog to return and finish.

Before this regression was introduced, the code in this section used
del_timer(), and not del_timer_sync() to delete the watchdog timer.

Link: https://lore.kernel.org/r/399e486c-6540-db27-76aa-7a271b061f76@tasossah.com
Fixes: 070ad7e793dc ("floppy: convert to delayed work and single-thread wq")
Signed-off-by: Tasos Sahanidis <tasos@tasossah.com>
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mtd: core: provide unique name for nvmem device

[ Upstream commit c048b60d39e109c201d31ed5ad3a4f939064d6c4 ]

If there is more than one mtd device which supports OTP, there will
be a kernel warning about duplicated sysfs entries and the probing will
fail. This is because the nvmem device name is not unique. Make it
unique by prepending the name of the mtd. E.g. before the name was
"user-otp", now it will be "mtd0-user-otp".

For reference the kernel splash is:
[    4.665531] sysfs: cannot create duplicate filename '/bus/nvmem/devices/user-otp'
[    4.673056] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.15.0-next-20211101+ #1296
[    4.680565] Hardware name: Kontron SMARC-sAL28 (Single PHY) on SMARC Eval 2.0 carrier (DT)
[    4.688856] Call trace:
[    4.691303]  dump_backtrace+0x0/0x1bc
[    4.694984]  show_stack+0x24/0x30
[    4.698306]  dump_stack_lvl+0x68/0x84
[    4.701980]  dump_stack+0x18/0x34
[    4.705302]  sysfs_warn_dup+0x70/0x90
[    4.708973]  sysfs_do_create_link_sd+0x144/0x150
[    4.713603]  sysfs_create_link+0x2c/0x50
[    4.717535]  bus_add_device+0x74/0x120
[    4.721293]  device_add+0x330/0x890
[    4.724791]  device_register+0x2c/0x40
[    4.728550]  nvmem_register+0x240/0x9f0
[    4.732398]  mtd_otp_nvmem_register+0xb0/0x10c
[    4.736854]  mtd_device_parse_register+0x28c/0x2b4
[    4.741659]  spi_nor_probe+0x20c/0x2e0
[    4.745418]  spi_mem_probe+0x78/0xbc
[    4.749001]  spi_probe+0x90/0xf0
[    4.752237]  really_probe.part.0+0xa4/0x320
..
[    4.873936] mtd mtd1: Failed to register OTP NVMEM device
[    4.894468] spi-nor: probe of spi0.0 failed with error -17

Fixes: 4b361cfa8624 ("mtd: core: add OTP nvmem provider support")
Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20211104134843.2642800-1-michael@walle.cc
Signed-off-by: Sasha Levin <sashal@kernel.org>

serial: amba-pl011: do not request memory region twice

[ Upstream commit d1180405c7b5c7a1c6bde79d5fc24fe931430737 ]

With commit 3873e2d7f63a ("drivers: PL011: refactor pl011_probe()") the
function devm_ioremap() called from pl011_setup_port() was replaced with
devm_ioremap_resource(). Since this function not only remaps but also
requests the ports io memory region it now collides with the .config_port()
callback which requests the same region at uart port registration.

Since devm_ioremap_resource() already claims the memory successfully, the
request in .config_port() fails.

Later at uart port deregistration the attempt to release the unclaimed
memory also fails. The failure results in a “Trying to free nonexistent
resource" warning.

Fix these issues by removing the callbacks that implement the redundant
memory allocation/release. Also make sure that changing the drivers io
memory base address via TIOCSSERIAL is not allowed any more.

Fixes: 3873e2d7f63a ("drivers: PL011: refactor pl011_probe()")
Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Link: https://lore.kernel.org/r/20211129174238.8333-1-LinoSanfilippo@gmx.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tty: serial: uartlite: allow 64 bit address

[ Upstream commit 3672fb65155530b5eea6225685c75329b6debec3 ]

The base address of uartlite registers could be 64 bit address which is from
device resource. When ulite_probe() calls ulite_assign(), this 64 bit
address is casted to 32-bit. The fix is to replace "u32" type with
"phys_addr_t" type for the base address in ulite_assign() argument list.

Fixes: 8fa7b6100693 ("[POWERPC] Uartlite: Separate the bus binding from the driver proper")
Signed-off-by: Lizhi Hou <lizhi.hou@xilinx.com>
Link: https://lore.kernel.org/r/20211129202302.1319033-1-lizhi.hou@xilinx.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: k3-j7200: Correct the d-cache-sets info

[ Upstream commit a172c86931709d6663318609d71a811333bdf4b0 ]

A72 Cluster (chapter 1.3.1 [1]) has 48KB Icache, 32KB Dcache and 1MB L2 Cache
 - ICache is 3-way set-associative
 - Dcache is 2-way set-associative
 - Line size are 64bytes

32KB (Dcache)/64 (fixed line length of 64 bytes) = 512 ways
512 ways / 2 (Dcache is 2-way per set) = 256 sets.

So, correct the d-cache-sets info.

[1] https://www.ti.com/lit/pdf/spruiu1

Fixes: d361ed88455f ("arm64: dts: ti: Add support for J7200 SoC")
Reported-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Pratyush Yadav <p.yadav@ti.com>
Reviewed-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Link: https://lore.kernel.org/r/20211113042640.30955-1-nm@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: k3-j721e: Fix the L2 cache sets

[ Upstream commit e9ba3a5bc6fdc2c796c69fdaf5ed6c9957cf9f9d ]

A72's L2 cache[1] on J721e[2] is 1MB. A72's L2 is fixed line length of
64 bytes and 16-way set-associative cache structure.

1MB of L2 / 64 (line length) = 16384 ways
16384 ways / 16 = 1024 sets

Fix the l2 cache-sets.

[1] https://developer.arm.com/documentation/100095/0003/Level-2-Memory-System/About-the-L2-memory-system
[2] http://www.ti.com/lit/pdf/spruil1

Fixes: 2d87061e70de ("arm64: dts: ti: Add Support for J721E SoC")
Reported-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Pratyush Yadav <p.yadav@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Link: https://lore.kernel.org/r/20211113043639.4413-1-nm@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: k3-j7200: Fix the L2 cache sets

[ Upstream commit d0c826106f3fc11ff97285102b576b65576654ae ]

A72's L2 cache[1] on J7200[2] is 1MB. A72's L2 is fixed line length of
64 bytes and 16-way set-associative cache structure.

1MB of L2 / 64 (line length) = 16384 ways
16384 ways / 16 = 1024 sets

Fix the l2 cache-sets.

[1] https://developer.arm.com/documentation/100095/0003/Level-2-Memory-System/About-the-L2-memory-system
[2] https://www.ti.com/lit/pdf/spruiu1

Fixes: d361ed88455f ("arm64: dts: ti: Add support for J7200 SoC")
Reported-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Pratyush Yadav <p.yadav@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Link: https://lore.kernel.org/r/20211113043638.4358-1-nm@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: k3-am642: Fix the L2 cache sets

[ Upstream commit a27a93bf70045be54b594fa8482959ffb84166d7 ]

A53's L2 cache[1] on AM642[2] is 256KB. A53's L2 is fixed line length
of 64 bytes and 16-way set-associative cache structure.

256KB of L2 / 64 (line length) = 4096 ways
4096 ways / 16 = 256 sets

Fix the l2 cache-sets.

[1] https://developer.arm.com/documentation/ddi0500/j/Level-2-Memory-System/About-the-L2-memory-system?lang=en
[2] https://www.ti.com/lit/pdf/spruim2

Fixes: 8abae9389bdb ("arm64: dts: ti: Add support for AM642 SoC")
Reported-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Pratyush Yadav <p.yadav@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Link: https://lore.kernel.org/r/20211113043635.4296-1-nm@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: caam - save caam memory to support crypto engine retry mechanism.

[ Upstream commit 087e1d715bccf25dc0e83294576e416b0386ba20 ]

When caam queue is full (-ENOSPC), caam frees descriptor memory.
crypto-engine checks if retry support is true and h/w queue
is full(-ENOSPC), then requeue the crypto request.
During processing the requested descriptor again, caam gives below error.
(caam_jr 30902000.jr: 40000006: DECO: desc idx 0: Invalid KEY Command).

This patch adds a check to return when caam input ring is full
and retry support is true. so descriptor memory is not freed
and requeued request can be processed again.

Fixes: 2d653936eb2cf ("crypto: caam - enable crypto-engine retry mechanism")
Signed-off-by: Gaurav Jain <gaurav.jain@nxp.com>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

libbpf: Clean gen_loader's attach kind.

[ Upstream commit 19250f5fc0c283892a61f3abf9d65e6325f63897 ]

The gen_loader has to clear attach_kind otherwise the programs
without attach_btf_id will fail load if they follow programs
with attach_btf_id.

Fixes: 67234743736a ("libbpf: Generate loader program out of BPF ELF file.")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-12-alexei.starovoitov@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

[ Upstream commit ab50cb9df8896b39aae65c537a30de2c79c19735 ]

In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
which could lead to a NULL pointer dereference on failure of
radeon_vm_bo_add().

Fix this bug by adding a check of vm->ib_bo_va.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_RADEON=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/amdgpu: Fix a NULL pointer dereference in amdgpu_connector_lcd_native_mode()

[ Upstream commit b220110e4cd442156f36e1d9b4914bb9e87b0d00 ]

In amdgpu_connector_lcd_native_mode(), the return value of
drm_mode_duplicate() is assigned to mode, and there is a dereference
of it in amdgpu_connector_lcd_native_mode(), which will lead to a NULL
pointer dereference on failure of drm_mode_duplicate().

Fix this bug add a check of mode.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_AMDGPU=m show no new warnings, and
our static analyzer no longer warns about this code.

Fixes: d38ceaf99ed0 ("drm/amdgpu: add core driver (v4)")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/imx8mm: Enable ADC when enabling monitor

[ Upstream commit 3de89d8842a2b5d3dd22ebf97dd561ae0a330948 ]

The i.MX 8MP has a ADC_PD bit in the TMU_TER register that controls the
operating mode of the ADC:
* 0 means normal operating mode
* 1 means power down mode

When enabling/disabling the TMU, the ADC operating mode must be set
accordingly.

i.MX 8M Mini & Nano are lacking this bit.

Signed-off-by: Paul Gerber <Paul.Gerber@tq-group.com>
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Fixes: 2b8f1f0337c5 ("thermal: imx8mm: Add i.MX8MP support")
Link: https://lore.kernel.org/r/20211122114225.196280-1-alexander.stein@ew.tq-group.com
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ACPI: EC: Rework flushing of EC work while suspended to idle

[ Upstream commit 4a9af6cac050dce2e895ec3205c4615383ad9112 ]

The flushing of pending work in the EC driver uses drain_workqueue()
to flush the event handling work that can requeue itself via
advance_transaction(), but this is problematic, because that
work may also be requeued from the query workqueue.

Namely, if an EC transaction is carried out during the execution of
a query handler, it involves calling advance_transaction() which
may queue up the event handling work again.  This causes the kernel
to complain about attempts to add a work item to the EC event
workqueue while it is being drained and worst-case it may cause a
valid event to be skipped.

To avoid this problem, introduce two new counters, events_in_progress
and queries_in_progress, incremented when a work item is queued on
the event workqueue or the query workqueue, respectively, and
decremented at the end of the corresponding work function, and make
acpi_ec_dispatch_gpe() the workqueues in a loop until the both of
these counters are zero (or system wakeup is pending) instead of
calling acpi_ec_flush_work().

At the same time, change __acpi_ec_flush_work() to call
flush_workqueue() instead of drain_workqueue() to flush the event
workqueue.

While at it, use the observation that the work item queued in
acpi_ec_query() cannot be pending at that time, because it is used
only once, to simplify the code in there.

Additionally, clean up a comment in acpi_ec_query() and adjust white
space in acpi_ec_event_processor().

Fixes: f0ac20c3f613 ("ACPI: EC: Fix flushing of pending work")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cgroup: Trace event cgroup id fields should be u64

[ Upstream commit e14da77113bb890d7bf9e5d17031bdd476a7ce5e ]

Various trace event fields that store cgroup IDs were declared as
ints, but cgroup_id(() returns a u64 and the structures and associated
TP_printk() calls were not updated to reflect this.

Fixes: 743210386c03 ("cgroup: use cgrp->kn->id as the cgroup ID")
Signed-off-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/vmwgfx: Fail to initialize on broken configs

[ Upstream commit c451af78f301ff5156998d571c37cab329c10051 ]

Some of our hosts have a bug where rescaning a pci bus results in stale
fifo memory being mapped on the host. This makes any fifo communication
impossible resulting in various kernel crashes.

Instead of unexpectedly crashing, predictably fail to load the driver
which will preserve the system.

Fixes: fb1d9738ca05 ("drm/vmwgfx: Add DRM driver for VMware Virtual GPU")
Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211105193845.258816-4-zackr@vmware.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/vmwgfx: Remove the deprecated lower mem limit

[ Upstream commit 826c387d015247df396a91eadbaca94f0394853c ]

TTM during the transition to the new page allocator lost the ability
to constrain the allocations via the lower_mem_limit. The code has
been unused since the change:
256dd44bd897 ("drm/ttm: nuke old page allocator")
and there's no reason to keep it.

Fixes: 256dd44bd897 ("drm/ttm: nuke old page allocator")
Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211105193845.258816-2-zackr@vmware.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: msm8916: fix MMC controller aliases

[ Upstream commit b0293c19d42f6d6951c2fab9a47fed50baf2c14d ]

Change sdhcN aliases to mmcN to make them actually work. Currently the
board uses non-standard aliases sdhcN, which do not work, resulting in
mmc0 and mmc1 hosts randomly changing indices between boots.

Fixes: c4da5a561627 ("arm64: dts: qcom: Add msm8916 sdhci configuration nodes")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211201020559.1611890-1-dmitry.baryshkov@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc: Avoid discarding flags in system_call_exception()

[ Upstream commit 08b0af5b2affbe7419853e8dd1330e4b3fe27125 ]

Some thread flags can be set remotely, and so even when IRQs are disabled,
the flags can change under our feet. Thus, when setting flags we must use
an atomic operation rather than a plain read-modify-write sequence, as a
plain read-modify-write may discard flags which are concurrently set by a
remote thread, e.g.

// task A // task B
tmp = A->thread_info.flags;
set_tsk_thread_flag(A, NEWFLAG_B);
tmp |= NEWFLAG_A;
A->thread_info.flags = tmp;

arch/powerpc/kernel/interrupt.c's system_call_exception() sets
_TIF_RESTOREALL in the thread info flags with a read-modify-write, which
may result in other flags being discarded.

Elsewhere in the file it uses clear_bits() to atomically remove flag bits,
so use set_bits() here for consistency with those.

There may be reasons (e.g. instrumentation) that prevent the use of
set_thread_flag() and clear_thread_flag() here, which would otherwise be
preferable.

Fixes: ae7aaecc3f2f78b7 ("powerpc/64s: system call rfscv workaround for TM bugs")
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Eirik Fuller <efuller@redhat.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Link: https://lore.kernel.org/r/20211129130653.2037928-10-mark.rutland@arm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: bridge: add support for pppoe filtering

[ Upstream commit 28b78ecffea8078d81466b2e01bb5a154509f1ba ]

This makes 'bridge-nf-filter-pppoe-tagged' sysctl work for
bridged traffic.

Looking at the original commit it doesn't appear this ever worked:

 static unsigned int br_nf_post_routing(unsigned int hook, struct sk_buff **pskb,
[..]
        if (skb->protocol == htons(ETH_P_8021Q)) {
                skb_pull(skb, VLAN_HLEN);
                skb->network_header += VLAN_HLEN;
+       } else if (skb->protocol == htons(ETH_P_PPP_SES)) {
+               skb_pull(skb, PPPOE_SES_HLEN);
+               skb->network_header += PPPOE_SES_HLEN;
        }
 [..]
NF_HOOK(... POST_ROUTING, ...)

... but the adjusted offsets are never restored.

The alternative would be to rip this code out for good,
but otoh we'd have to keep this anyway for the vlan handling
(which works because vlan tag info is in the skb, not the packet
 payload).

Reported-and-tested-by: Amish Chana <amish@3g.co.za>
Fixes: 516299d2f5b6f97 ("[NETFILTER]: bridge-nf: filter bridged IPv4/IPv6 encapsulated in pppoe traffic")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

igc: AF_XDP zero-copy metadata adjust breaks SKBs on XDP_PASS

[ Upstream commit 4fa8fcd3440101dbacf4fae91de69877ef751977 ]

Driver already implicitly supports XDP metadata access in AF_XDP
zero-copy mode, as xsk_buff_pool's xp_alloc() naturally set xdp_buff
data_meta equal data.

This works fine for XDP and AF_XDP, but if a BPF-prog adjust via
bpf_xdp_adjust_meta() and choose to call XDP_PASS, then igc function
igc_construct_skb_zc() will construct an invalid SKB packet. The
function correctly include the xdp->data_meta area in the memcpy, but
forgot to pull header to take metasize into account.

Fixes: fc9df2a0b520 ("igc: Enable RX via AF_XDP zero-copy")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/imx: Implement runtime PM support

[ Upstream commit 4cf2ddf16e175ee18c5c29865c32da7d6269cf44 ]

Starting with commit d92ed2c9d3ff ("thermal: imx: Use driver's local
data to decide whether to run a measurement") this driver stared using
irq_enabled flag to make decision to power on/off the thermal
core. This triggered a regression, where after reaching critical
temperature, alarm IRQ handler set irq_enabled to false, disabled
thermal core and was not able read temperature and disable cooling
sequence.

In case the cooling device is "CPU/GPU freq", the system will run with
reduce performance until next reboot.

To solve this issue, we need to move all parts implementing hand made
runtime power management and let it handle actual runtime PM framework.

Fixes: d92ed2c9d3ff ("thermal: imx: Use driver's local data to decide whether to run a measurement")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Tested-by: Petr Beneš <petr.benes@ysoft.com>
Link: https://lore.kernel.org/r/20211117103426.81813-1-o.rempel@pengutronix.de
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: stmmac: Add platform level debug register dump feature

[ Upstream commit 4047b9db1aa7512a10ba3560a3f63821c8c40235 ]

dwmac-qcom-ethqos currently exposes a mechanism to dump rgmii registers
after the 'stmmac_dvr_probe()' returns. However with commit
5ec55823438e ("net: stmmac: add clocks management for gmac driver"),
we now let 'pm_runtime_put()' disable the clocks before returning from
'stmmac_dvr_probe()'.

This causes a crash when 'rgmii_dump()' register dumps are enabled,
as the clocks are already off.

Since other dwmac drivers (possible future users as well) might
require a similar register dump feature, introduce a platform level
callback to allow the same.

This fixes the crash noticed while enabling rgmii_dump() dumps in
dwmac-qcom-ethqos driver as well. It also allows future changes
to keep a invoking the register dump callback from the correct
place inside 'stmmac_dvr_probe()'.

Fixes: 5ec55823438e ("net: stmmac: add clocks management for gmac driver")
Cc: Joakim Zhang <qiangqing.zhang@nxp.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: venus: core: Fix a resource leak in the error handling path of 'venus_probe()'

[ Upstream commit 8cc7a1b2aca067397a016cdb971a5e6ad9b640c7 ]

A successful 'of_platform_populate()' call should be balanced by a
corresponding 'of_platform_depopulate()' call in the error handling path
of the probe, as already done in the remove function.

A successful 'venus_firmware_init()' call should be balanced by a
corresponding 'venus_firmware_deinit()' call in the error handling path
of the probe, as already done in the remove function.

Update the error handling path accordingly.

Fixes: f9799fcce4bb ("media: venus: firmware: register separate platform_device for firmware loader")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: venus: core: Fix a potential NULL pointer dereference in an error handling path

[ Upstream commit e4debea9be7d5db52bc6a565a4c02c3c6560d093 ]

The normal path of the function makes the assumption that
'pm_ops->core_power' may be NULL.
We should make the same assumption in the error handling path or a NULL
pointer dereference may occur.

Add the missing test before calling 'pm_ops->core_power'

Fixes: 9e8efdb57879 ("media: venus: core: vote for video-mem path")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: venus: correct low power frequency calculation for encoder

[ Upstream commit b1f9bb8020783a48151e3a2864fbdc70548566dd ]

In exististing implimentation, in min_loaded_core() for low_power
vpp frequency value is considering as vpp_freq instead of low_power_freq.
Fixed this by correcting vpp frequency calculation for encoder.

Fixes: 3cfe5815ce0e (media: venus: Enable low power setting for encoder)
Signed-off-by: Mansur Alisha Shaik <mansur@codeaurora.org>
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: coda: fix CODA960 JPEG encoder buffer overflow

[ Upstream commit 1a59cd88f55068710f6549bee548846661673780 ]

Stop the CODA960 JPEG encoder from overflowing capture buffers.
The bitstream buffer overflow interrupt doesn't seem to be connected,
so this has to be handled via timeout instead.

Reported-by: Martin Weber <martin.weber@br-automation.com>
Fixes: 96f6f62c4656 ("media: coda: jpeg: add CODA960 JPEG encoder support")
Tested-by: Martin Weber <martin.weber@br-automation.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: hantro: Hook up RK3399 JPEG encoder output

[ Upstream commit 230d683ae04894933720425c8dead9508a09ebc3 ]

The JPEG encoder found in the Hantro H1 encoder block only produces a
raw entropy-encoded scan. The driver is responsible for building a JPEG
compliant bitstream and placing the entropy-encoded scan in it. Right
now the driver uses a bounce buffer for the hardware to output the raw
scan to.

In commit e765dba11ec2 ("hantro: Move hantro_enc_buf_finish to JPEG
codec_ops.done"), the code that copies the raw scan from the bounce
buffer to the capture buffer was moved, but was only hooked up for the
Hantro H1 (then RK3288) variant. The RK3399 variant was broken,
producing a JPEG bitstream without the scan, and the capture buffer's
.bytesused field unset.

Fix this by duplicating the code that is executed when the JPEG encoder
finishes encoding a frame. As the encoded length is read back from
hardware, and the variants having different register layouts, the
code is duplicated rather than shared.

Fixes: e765dba11ec2 ("hantro: Move hantro_enc_buf_finish to JPEG codec_ops.done")
Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: mtk-vcodec: call v4l2_m2m_ctx_release first when file is released

[ Upstream commit 9f89c881bffbdffe4060ffaef3489a2830a6dd9c ]

The func v4l2_m2m_ctx_release waits for currently running jobs
to finish and then stop streaming both queues and frees the buffers.
All this should be done before the call to mtk_vcodec_enc_release
which frees the encoder handler. This fixes null-pointer dereference bug:

[  638.028076] Mem abort info:
[  638.030932]   ESR = 0x96000004
[  638.033978]   EC = 0x25: DABT (current EL), IL = 32 bits
[  638.039293]   SET = 0, FnV = 0
[  638.042338]   EA = 0, S1PTW = 0
[  638.045474]   FSC = 0x04: level 0 translation fault
[  638.050349] Data abort info:
[  638.053224]   ISV = 0, ISS = 0x00000004
[  638.057055]   CM = 0, WnR = 0
[  638.060018] user pgtable: 4k pages, 48-bit VAs, pgdp=000000012b6db000
[  638.066485] [00000000000001a0] pgd=0000000000000000, p4d=0000000000000000
[  638.073277] Internal error: Oops: 96000004 [#1] SMP
[  638.078145] Modules linked in: rfkill mtk_vcodec_dec mtk_vcodec_enc uvcvideo mtk_mdp mtk_vcodec_common videobuf2_dma_contig v4l2_h264 cdc_ether v4l2_mem2mem videobuf2_vmalloc usbnet videobuf2_memops videobuf2_v4l2 r8152 videobuf2_common videodev cros_ec_sensors cros_ec_sensors_core industrialio_triggered_buffer kfifo_buf elan_i2c elants_i2c sbs_battery mc cros_usbpd_charger cros_ec_chardev cros_usbpd_logger crct10dif_ce mtk_vpu fuse ip_tables x_tables ipv6
[  638.118583] CPU: 0 PID: 212 Comm: kworker/u8:5 Not tainted 5.15.0-06427-g58a1d4dcfc74-dirty #109
[  638.127357] Hardware name: Google Elm (DT)
[  638.131444] Workqueue: mtk-vcodec-enc mtk_venc_worker [mtk_vcodec_enc]
[  638.137974] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  638.144925] pc : vp8_enc_encode+0x34/0x2b0 [mtk_vcodec_enc]
[  638.150493] lr : venc_if_encode+0xac/0x1b0 [mtk_vcodec_enc]
[  638.156060] sp : ffff8000124d3c40
[  638.159364] x29: ffff8000124d3c40 x28: 0000000000000000 x27: 0000000000000000
[  638.166493] x26: 0000000000000000 x25: ffff0000e7f252d0 x24: ffff8000124d3d58
[  638.173621] x23: ffff8000124d3d58 x22: ffff8000124d3d60 x21: 0000000000000001
[  638.180750] x20: ffff80001137e000 x19: 0000000000000000 x18: 0000000000000001
[  638.187878] x17: 000000040044ffff x16: 00400032b5503510 x15: 0000000000000000
[  638.195006] x14: ffff8000118536c0 x13: ffff8000ee1da000 x12: 0000000030d4d91d
[  638.202134] x11: 0000000000000000 x10: 0000000000000980 x9 : ffff8000124d3b20
[  638.209262] x8 : ffff0000c18d4ea0 x7 : ffff0000c18d44c0 x6 : ffff0000c18d44c0
[  638.216391] x5 : ffff80000904a3b0 x4 : ffff8000124d3d58 x3 : ffff8000124d3d60
[  638.223519] x2 : ffff8000124d3d78 x1 : 0000000000000001 x0 : ffff80001137efb8
[  638.230648] Call trace:
[  638.233084]  vp8_enc_encode+0x34/0x2b0 [mtk_vcodec_enc]
[  638.238304]  venc_if_encode+0xac/0x1b0 [mtk_vcodec_enc]
[  638.243525]  mtk_venc_worker+0x110/0x250 [mtk_vcodec_enc]
[  638.248918]  process_one_work+0x1f8/0x498
[  638.252923]  worker_thread+0x140/0x538
[  638.256664]  kthread+0x148/0x158
[  638.259884]  ret_from_fork+0x10/0x20
[  638.263455] Code: f90023f9 2a0103f5 aa0303f6 aa0403f8 (f940d277)
[  638.269538] ---[ end trace e374fc10f8e181f5 ]---

[gst-master] root@debian:~/gst-build# [  638.019193] Unable to handle kernel NULL pointer dereference at virtual address 00000000000001a0
Fixes: 4e855a6efa547 ("[media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver")
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: si470x-i2c: fix possible memory leak in si470x_i2c_probe()

[ Upstream commit ef054e345ed8c79ce1121a3599b5a2dfd78e57a0 ]

n the 'radio->hdl.error' error handling, ctrl handler allocated by
v4l2_ctrl_new_std() does not released, and caused memory leak as
follows:

unreferenced object 0xffff888033d54200 (size 256):
  comm "i2c-si470x-19", pid 909, jiffies 4294914203 (age 8.072s)
  hex dump (first 32 bytes):
    e8 69 11 03 80 88 ff ff 00 46 d5 33 80 88 ff ff  .i.......F.3....
    10 42 d5 33 80 88 ff ff 10 42 d5 33 80 88 ff ff  .B.3.....B.3....
  backtrace:
    [<00000000086bd4ed>] __kmalloc_node+0x1eb/0x360
    [<00000000bdb68871>] kvmalloc_node+0x66/0x120
    [<00000000fac74e4c>] v4l2_ctrl_new+0x7b9/0x1c60 [videodev]
    [<00000000693bf940>] v4l2_ctrl_new_std+0x19b/0x270 [videodev]
    [<00000000c0cb91bc>] si470x_i2c_probe+0x2d3/0x9a0 [radio_si470x_i2c]
    [<0000000056a6f01f>] i2c_device_probe+0x4d8/0xbe0

Fix the error handling path to avoid memory leak.

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 8c081b6f9a9b ("media: radio: Critical v4l2 registration...")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: imx-pxp: Initialize the spinlock prior to using it

[ Upstream commit ed2f97ad4b21072f849cf4ae6645d1f2b1d3f550 ]

After devm_request_threaded_irq() is called there is a chance that an
interrupt may occur before the spinlock is initialized, which will trigger
a kernel oops.

To prevent that, move the initialization of the spinlock prior to
requesting the interrupts.

Fixes: 51abcf7fdb70 ("media: imx-pxp: add i.MX Pixel Pipeline driver")
Signed-off-by: Fabio Estevam <festevam@denx.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: rcar-csi2: Correct the selection of hsfreqrange

[ Upstream commit cee44d4fbacbbdfe62697ec94e76c6e4f726c5df ]

hsfreqrange should be chosen based on the calculated mbps which
is closer to the default bit rate  and within the range as per
table[1]. But current calculation always selects first value which
is greater than or equal to the calculated mbps which may lead
to chosing a wrong range in some cases.

For example for 360 mbps for H3/M3N
Existing logic selects
Calculated value 360Mbps : Default 400Mbps Range [368.125 -433.125 mbps]

This hsfreqrange is out of range.

The logic is changed to get the default value which is closest to the
calculated value [1]

Calculated value 360Mbps : Default 350Mbps  Range [320.625 -380.625 mpbs]

[1] specs r19uh0105ej0200-r-car-3rd-generation.pdf [Table 25.9]

Please note that According to Renesas in Table 25.9 the range for
220 default value is corrected as below

 |Range (Mbps)     |  Default  Bit rate (Mbps) |
 -----------------------------------------------
 | 197.125-244.125 |     220                   |
 -----------------------------------------------

Fixes: 769afd212b16 ("media: rcar-csi2: add Renesas R-Car MIPI CSI-2 receiver driver")
Signed-off-by: Suresh Udipi <sudipi@jp.adit-jv.com>
Signed-off-by: Kazuyoshi Akiyama <akiyama@nds-osk.co.jp>
Signed-off-by: Michael Rodin <mrodin@de.adit-jv.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: ov8865: Fix lockdep error

[ Upstream commit 6e1c9bc9ae96e57bcd8807174f2c0f44f9ef7938 ]

ov8865_state_init() calls ov8865_state_mipi_configure() which uses
__v4l2_ctrl_s_ctrl[_int64](). This means that sensor->mutex (which
is also sensor->ctrls.handler.lock) must be locked before calling
ov8865_state_init().

Note ov8865_state_mipi_configure() is also used in other places where
the lock is already held so it cannot be changed itself.

This fixes the following lockdep kernel WARN:

[   13.233421] WARNING: CPU: 0 PID: 8 at drivers/media/v4l2-core/v4l2-ctrls-api.c:833 __v4l2_ctrl_s_ctrl+0x4d/0x60 [videodev]
...
[   13.234063] Call Trace:
[   13.234074]  ov8865_state_configure+0x98b/0xc00 [ov8865]
[   13.234095]  ov8865_probe+0x4b1/0x54c [ov8865]
[   13.234117]  i2c_device_probe+0x13c/0x2d0

Fixes: 11c0d8fdccc5 ("media: i2c: Add support for the OV8865 image sensor")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: Re-order runtime pm initialisation

[ Upstream commit d2484fbf780762f6f9cc3abb7a07ee42dca2eaa3 ]

The kerneldoc for pm_runtime_set_suspended() says:

"It is not valid to call this function for devices with runtime PM
enabled"

To satisfy that requirement, re-order the calls so that
pm_runtime_enable() is the last one.

Fixes: 11c0d8fdccc5 ("media: i2c: Add support for the OV8865 image sensor")
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

media: i2c: imx274: fix s_frame_interval runtime resume not requested

[ Upstream commit da653498c20ba5b185214d8ae43b4e8e9594f520 ]

pm_runtime_resume_and_get should be called when the s_frame_interval
is called.

The driver will try to access device registers to configure VMAX, coarse
time and exposure.

Currently if the runtime is not resumed, this fails:
 # media-ctl -d /dev/media0 --set-v4l2 '"IMX274 1-001a":0[fmt:SRGGB10_1X10/3840x2
160@1/10]'

IMX274 1-001a: imx274_binning_goodness: ask 3840x2160, size 3840x2160, goodness 0
IMX274 1-001a: imx274_binning_goodness: ask 3840x2160, size 1920x1080, goodness -3000
IMX274 1-001a: imx274_binning_goodness: ask 3840x2160, size 1280x720, goodness -4000
IMX274 1-001a: imx274_binning_goodness: ask 3840x2160, size 1280x540, goodness -4180
IMX274 1-001a: __imx274_change_compose: selected 1x1 binning
IMX274 1-001a: imx274_set_frame_interval: input frame interval = 1 / 10
IMX274 1-001a: imx274_read_mbreg : addr 0x300e, val=0x1 (2 bytes)
IMX274 1-001a: imx274_set_frame_interval : register SVR = 1
IMX274 1-001a: imx274_read_mbreg : addr 0x30f6, val=0x6a8 (2 bytes)
IMX274 1-001a: imx274_set_frame_interval : register HMAX = 1704
IMX274 1-001a: imx274_set_frame_length : input length = 2112
IMX274 1-001a: imx274_write_mbreg : i2c bulk write failed, 30f8 = 884 (3 bytes)
IMX274 1-001a: imx274_set_frame_length error = -121
IMX274 1-001a: imx274_set_frame_interval error = -121
Unable to setup formats: Remote I/O error (121)

The device is not resumed thus the remote I/O error.

Setting the frame interval works at streaming time, because
pm_runtime_resume_and_get is called at s_stream time before sensor setup.
The failure happens when only the s_frame_interval is called separately
independently on streaming time.

Fixes: ad97bc37426c ("media: i2c: imx274: Add IMX274 power on and off sequence")
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>