Merge remote-tracking branch 'stable/linux-5.10.y' into rpi-5.10.y

Revert "drm/vc4: hdmi: Make sure the controller is powered up during bind"

This reverts commit 94bc403a9a7af3bd042ecdc160e6ab0b08331df8.

Revert "drm/vc4: hdmi: Switch to pm_runtime_resume_and_get"

This reverts commit a2224f9777778965ff76049548ba483a23d7933e.

configs: Add CONFIG_DRM_GUD=m

This enables support for the GUD USB Display DRM driver.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/gud: Add async_flush module parameter

[ drm-misc commit a0356899ebe8ecde0da9c5685cc47154db973a5e ]

Provide a way for userspace to choose synchronous flushing/pageflips.
This helps save CPU and power.

It is also useful for test scripts since userspace can know when a flush
has happended and wait before doing the next visual test.

Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Peter Stuge <peter@stuge.se>
Link: https://patchwork.freedesktop.org/patch/msgid/20210703141321.35494-2-noralf@tronnes.org

drm/gud: Add Raspberry Pi Pico ID

[ drm-misc commit b3f4ef669357d5b9a2c5e8c33e3967a2070db7f9 ]

Add VID/PID for the Raspberry Pi Pico implementation.
Source: https://github.com/notro/gud-pico

Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Peter Stuge <peter@stuge.se>
Link: https://patchwork.freedesktop.org/patch/msgid/20210703141321.35494-1-noralf@tronnes.org

drm/gud: Use scatter-gather USB bulk transfer

[ drm-misc commit 2eecd93b743b5611cd3654698794b4d0cefdc9ee ]

There'a limit to how big a kmalloc buffer can be, and as memory gets
fragmented it becomes more difficult to get big buffers. The downside of
smaller buffers is that the driver has to split the transfer up which
hampers performance. Compression might also take a hit because of the
splitting.

Solve this by allocating the transfer buffer using vmalloc and create a
SG table to be passed on to the USB subsystem. vmalloc_32() is used to
avoid DMA bounce buffers on USB controllers that can only access 32-bit
addresses.

This also solves the problem that split transfers can give host side
tearing since flushing is decoupled from rendering.

usb_sg_wait() doesn't have timeout handling builtin, so it is wrapped in
a timer like 4 out of 6 users in the kernel have done.

v2:
- Use DIV_ROUND_UP (Linus)
- Add timeout note to the commit log (Linus)
- Expand note about upper buffer limit (Linus)
- Change var name s/timer/ctx/ in gud_usb_bulk_timeout()

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210701170748.58009-2-noralf@tronnes.org

drm/gud: Free buffers on device removal

[ drm-misc commit f8ac863b6a93863334cefb94285daaa6617381b5 ]

Free transfer and compression buffers on device removal instead of at
DRM device removal time. This ensures that the usual 2x8MB buffers are
released when the device is unplugged and not kept around should
userspace keep the DRM device fd open.

At least Ubuntu 20.04 doesn't release the DRM device on unplug.

The damage_lock mutex is not destroyed because it is used outside the
drm_dev_enter/exit block in gud_pipe_update(). AFAICT it's possible for
an open fbdev descriptor to trigger a commit after the USB device is gone.

v2: Don't destroy damage_lock

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210701170748.58009-1-noralf@tronnes.org

drm/gud: cleanup coding style a bit

[ Upstream commit 6dd7efc437611db16d432e0030f72d0c7e890127 ]

Fix coccicheck warning:
drivers/gpu/drm/gud/gud_internal.h:89:2-3: Unneeded semicolon
drivers/gpu/drm/gud/gud_internal.h:107:2-3: Unneeded semicolon

Signed-off-by: Bernard Zhao <bernard@vivo.com>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210402085523.76928-1-bernard@vivo.com

drm/gud: Remove unneeded semicolon

[ Upstream commit 166633c85c21fbe0baec40c898ca5a4f6eee0a08 ]

drivers/gpu/drm/gud/gud_connector.c:658:2-3: Unneeded semicolon
drivers/gpu/drm/gud/gud_connector.c:186:2-3: Unneeded semicolon
drivers/gpu/drm/gud/gud_drv.c:511:3-4: Unneeded semicolon
drivers/gpu/drm/gud/gud_pipe.c:127:4-5: Unneeded semicolon

 Remove unneeded semicolon.

Generated by: scripts/coccinelle/misc/semicolon.cocci

Fixes: 40e1a70b4aed ("drm: Add GUD USB Display driver")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: kernel test robot <lkp@intel.com>
[fix subject and squash 3 per file patches]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210322174434.58849-2-noralf@tronnes.org

drm/gud: fix sizeof use

[ Upstream commit b91fbdc1df15befb3e51b9795b15d71fa4298602 ]

drivers/gpu/drm/gud/gud_connector.c:710:37-43: ERROR: application of sizeof to pointer

 sizeof when applied to a pointer typed expression gives the size of
 the pointer

Generated by: scripts/coccinelle/misc/noderef.cocci

Fixes: 40e1a70b4aed ("drm: Add GUD USB Display driver")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: kernel test robot <lkp@intel.com>
[fix subject]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210322174434.58849-1-noralf@tronnes.org

drm: Add GUD USB Display driver

[ Upstream commit 40e1a70b4aedf2859a1829991b48ef0ebe650bf2 ]

This adds a USB display driver with the intention that it can be
used with future USB interfaced low end displays/adapters. The Linux
gadget device driver will serve as the canonical device implementation.

The following DRM properties are supported:
- Plane rotation
- Connector TV properties

There is also support for backlight brightness exposed as a backlight
device.

Display modes can be made available to the host driver either as DRM
display modes or through EDID. If both are present, EDID is just passed
on to userspace.

Performance is preferred over color depth, so if the device supports
RGB565, DRM_CAP_DUMB_PREFERRED_DEPTH will return 16.

If the device transfer buffer can't fit an uncompressed framebuffer
update, the update is split up into parts that do fit.

Optimal user experience is achieved by providing damage reports either by
setting FB_DAMAGE_CLIPS on pageflips or calling DRM_IOCTL_MODE_DIRTYFB.

LZ4 compression is used if the device supports it.

The driver supports a one bit monochrome transfer format: R1. This is not
implemented in the gadget driver. It is added in preparation for future
monochrome e-ink displays.

The driver is MIT licensed to smooth the path for any BSD port of the
driver.

v2:
- Use devm_drm_dev_alloc() and drmm_mode_config_init()
- drm_fbdev_generic_setup: Use preferred_bpp=0, 16 was a copy paste error
- The drm_backlight_helper is dropped, copy in the code
- Support protocol version backwards compatibility for device

v3:
- Use donated Openmoko USB pid
- Use direct compression from framebuffer when pitch matches, not only on
  full frames, so split updates can benefit
- Use __le16 in struct gud_drm_req_get_connector_status
- Set edid property when the device only provides edid
- Clear compression fields in struct gud_drm_req_set_buffer
- Fix protocol version negotiation
- Remove mode->vrefresh, it's calculated

v4:
- Drop the status req polling which was a workaround for something that
  turned out to be a dwc2 udc driver problem
- Add a flag for the Linux gadget to require a status request on
  SET operations. Other devices will only get status req on STALL errors
- Use protocol specific error codes (Peter)
- Add a flag for devices that want to receive the entire framebuffer on
  each flush (Lubomir)
- Retry a failed framebuffer flush
- If mode has changed wait for worker and clear pending damage before
  queuing up new damage, fb width/height might have changed
- Increase error counter on bulk transfer failures
- Use DRM_MODE_CONNECTOR_USB
- Handle R1 kmalloc error (Peter)
- Don't try and replicate the USB get descriptor request standard for the
  display descriptor (Peter)
- Make max_buffer_size optional (Peter), drop the pow2 requirement since
  it's not necessary anymore.
- Don't pre-alloc a control request buffer, it was only 4k
- Let gud.h describe the whole protocol explicitly and don't let DRM
  leak into it (Peter)
- Drop display mode .hskew and .vscan from the protocol
- Shorten names: s/GUD_DRM_/GUD_/ s/gud_drm_/gud_/ (Peter)
- Fix gud_pipe_check() connector picking when switching connector
- Drop gud_drm_driver_gem_create_object() cached is default now
- Retrieve USB device from struct drm_device.dev instead of keeping a
  pointer
- Honour fb->offsets[0]
- Fix mode fetching when connector status is forced
- Check EDID length reported by the device
- Use drm_do_get_edid() so userspace can overrride EDID
- Set epoch counter to signal connector status change
- gud_drm_driver can be const now

v5:
- GUD_DRM_FORMAT_R1: Use non-human ascii values (Daniel)
- Change name to: GUD USB Display (Thomas, Simon)
- Change one __u32 -> __le32 in protocol header
- Always log fb flush errors, unless the previous one failed
- Run backlight update in a worker to avoid upsetting lockdep (Daniel)
- Drop backlight_ops.get_brightness, there's no readback from the device
  so it doesn't really add anything.
- Set dma mask, needed by dma-buf importers

v6:
- Use obj-y in Makefile (Peter)
- Fix missing le32_to_cpu() when using GUD_DISPLAY_MAGIC (Peter)
- Set initial brightness on backlight device

v7:
- LZ4_compress_default() can return zero, check for that
- Fix memory leak in gud_pipe_check() error path (Peter)
- Improve debug and error messages (Peter)
- Don't pass length in protocol structs (Peter)
- Pass USB interface to gud_usb_control_msg() et al. (Peter)
- Improve gud_connector_fill_properties() (Peter)
- Add GUD_PIXEL_FORMAT_RGB111 (Peter)
- Remove GUD_REQ_SET_VERSION (Peter)
- Fix DRM_IOCTL_MODE_OBJ_SETPROPERTY and the rotation property
- Fix dma-buf import (Thomas)

v8:
- Forgot to filter RGB111 from reaching userspace
- Handle a device that only returns unknown device properties (Peter)
- s/GUD_PIXEL_FORMAT_RGB111/GUD_PIXEL_FORMAT_XRGB1111/ (Peter)
- Fix R1 and XRGB1111 format conversion
- Add FIXME about Big Endian being broken (Peter, Ilia)

Cc: Lubomir Rintel <lkundrak@v3.sk>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Peter Stuge <peter@stuge.se>
Tested-by: Peter Stuge <peter@stuge.se>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210313112545.37527-4-noralf@tronnes.org
[ backport changes:
  - Remove include drm_gem_atomic_helper.h
  - s/drm_gem_simple_display_pipe_prepare_fb/drm_gem_fb_simple_display_pipe_prepare_fb/
  - Remove const from gud_drm_driver
  - Change drm_gem_shmem_{vmap,vunmap} signatures
  - Add gud_gem_create_object() to get cached memory mapping.
]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/uapi: Add USB connector type

[ Upstream commit 757e26712337be6be70f7405c38cc0639957eedc ]

Add a connector type for USB connected display panels.

Some examples of what current userspace will name the connector:
- Weston: "UNNAMED-%d"
- Mutter: "Unknown20-%d"
- X: "Unknown20-%d"

v2:
- Update drm_connector_enum_list
- Add examples to commit message

Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210313112545.37527-2-noralf@tronnes.org

media: i2c: imx477: Fix framerates for 1332x990 mode

The imx477 driver's line length for this mode had not been updated to
the value supplied to us by the sensor manufacturer. With this
correction the sensor delivers the framerates that are expected.

Signed-off-by: David Plowman <david.plowman@raspberrypi.com>

Linux 5.10.52

Link: https://lore.kernel.org/r/20210719144940.904087935@linuxfoundation.org
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210719184320.888029606@linuxfoundation.org
Tested-by: Salvatore Bonaccorso <carnil@debian.org>
Tested-by: Pavel Machek (CIP) <pavel@denx.de>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Shuah Khan <skhan@linuxfoundation.org>
Tested-by: Fox Chen <foxhlchen@gmail.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

seq_file: disallow extremely large seq buffer allocations

commit 8cae8cd89f05f6de223d63e6d15e31c8ba9cf53b upstream.

There is no reasonable need for a buffer larger than this, and it avoids
int overflow pitfalls.

Fixes: 058504edd026 ("fs/seq_file: fallback to vmalloc allocation")
Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Reported-by: Qualys Security Advisory <qsa@qualys.com>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: scsi_dh_alua: Fix signedness bug in alua_rtpg()

commit 80927822e8b6be46f488524cd7d5fe683de97fc4 upstream.

The "retval" variable needs to be signed for the error handling to work.

Link: https://lore.kernel.org/r/YLjMEAFNxOas1mIp@mwanda
Fixes: 7e26e3ea0287 ("scsi: scsi_dh_alua: Check for negative result value")
Reviewed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

MIPS: vdso: Invalid GIC access through VDSO

[ Upstream commit 47ce8527fbba145a7723685bc9a27d9855e06491 ]

Accessing raw timers (currently only CLOCK_MONOTONIC_RAW) through VDSO
doesn't return the correct time when using the GIC as clock source.
The address of the GIC mapped page is in this case not calculated
correctly. The GIC mapped page is calculated from the VDSO data by
subtracting PAGE_SIZE:

  void *get_gic(const struct vdso_data *data) {
    return (void __iomem *)data - PAGE_SIZE;
  }

However, the data pointer is not page aligned for raw clock sources.
This is because the VDSO data for raw clock sources (CS_RAW = 1) is
stored after the VDSO data for coarse clock sources (CS_HRES_COARSE = 0).
Therefore, only the VDSO data for CS_HRES_COARSE is page aligned:

  +--------------------+
  |                    |
  | vd[CS_RAW]         | ---+
  | vd[CS_HRES_COARSE] |    |
  +--------------------+    | -PAGE_SIZE
  |                    |    |
  |  GIC mapped page   | <--+
  |                    |
  +--------------------+

When __arch_get_hw_counter() is called with &vd[CS_RAW], get_gic returns
the wrong address (somewhere inside the GIC mapped page). The GIC counter
values are not returned which results in an invalid time.

Fixes: a7f4df4e21dd ("MIPS: VDSO: Add implementations of gettimeofday() and clock_gettime()")
Signed-off-by: Martin Fäcknitz <faecknitz@hotsplots.de>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mips: disable branch profiling in boot/decompress.o

[ Upstream commit 97e488073cfca0eea84450169ca4cbfcc64e33e3 ]

Use DISABLE_BRANCH_PROFILING for arch/mips/boot/compressed/decompress.o
to prevent linkage errors.

mips64-linux-ld: arch/mips/boot/compressed/decompress.o: in function `LZ4_decompress_fast_extDict':
decompress.c:(.text+0x8c): undefined reference to `ftrace_likely_update'
mips64-linux-ld: decompress.c:(.text+0xf4): undefined reference to `ftrace_likely_update'
mips64-linux-ld: decompress.c:(.text+0x200): undefined reference to `ftrace_likely_update'
mips64-linux-ld: decompress.c:(.text+0x230): undefined reference to `ftrace_likely_update'
mips64-linux-ld: decompress.c:(.text+0x320): undefined reference to `ftrace_likely_update'
mips64-linux-ld: arch/mips/boot/compressed/decompress.o:decompress.c:(.text+0x3f4): more undefined references to `ftrace_likely_update' follow

Fixes: e76e1fdfa8f8 ("lib: add support for LZ4-compressed kernel")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: linux-mips@vger.kernel.org
Cc: Kyungsik Lee <kyungsik.lee@lge.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mips: always link byteswap helpers into decompressor

[ Upstream commit cddc40f5617e53f97ef019d5b29c1bd6cbb031ec ]

My series to clean up the unaligned access implementation
across architectures caused some mips randconfig builds to
fail with:

   mips64-linux-ld: arch/mips/boot/compressed/decompress.o: in function `decompress_kernel':
   decompress.c:(.text.decompress_kernel+0x54): undefined reference to `__bswapsi2'

It turns out that this problem has already been fixed for the XZ
decompressor but now it also shows up in (at least) LZO and LZ4.  From my
analysis I concluded that the compiler could always have emitted those
calls, but the different implementation allowed it to make otherwise
better decisions about not inlining the byteswap, which results in the
link error when the out-of-line code is missing.

While it could be addressed by adding it to the two decompressor
implementations that are known to be affected, but as this only adds
112 bytes to the kernel, the safer choice is to always add them.

Fixes: c50ec6787536 ("MIPS: zboot: Fix the build with XZ compression on older GCC versions")
Fixes: 0652035a5794 ("asm-generic: unaligned: remove byteshift helpers")
Link: https://lore.kernel.org/linux-mm/202106301304.gz2wVY9w-lkp@intel.com/
Link: https://lore.kernel.org/linux-mm/202106260659.TyMe8mjr-lkp@intel.com/
Link: https://lore.kernel.org/linux-mm/202106172016.onWT6Tza-lkp@intel.com/
Link: https://lore.kernel.org/linux-mm/202105231743.JJcALnhS-lkp@intel.com/
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

static_call: Fix static_call_text_reserved() vs __init

[ Upstream commit 2bee6d16e4379326b1eea454e68c98b17456769e ]

It turns out that static_call_text_reserved() was reporting __init
text as being reserved past the time when the __init text was freed
and re-used.

This is mostly harmless and will at worst result in refusing a kprobe.

Fixes: 6333e8f73b83 ("static_call: Avoid kprobes on inline static_call()s")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Link: https://lore.kernel.org/r/20210628113045.106211657@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

jump_label: Fix jump_label_text_reserved() vs __init

[ Upstream commit 9e667624c291753b8a5128f620f493d0b5226063 ]

It turns out that jump_label_text_reserved() was reporting __init text
as being reserved past the time when the __init text was freed and
re-used.

For a long time, this resulted in, at worst, not being able to kprobe
text that happened to land at the re-used address. However a recent
commit e7bf1ba97afd ("jump_label, x86: Emit short JMP") made it a
fatal mistake because it now needs to read the instruction in order to
determine the conflict -- an instruction that's no longer there.

Fixes: 4c3ef6d79328 ("jump label: Add jump_label_text_reserved() to reserve jump points")
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Link: https://lore.kernel.org/r/20210628113045.045141693@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

sched/uclamp: Ignore max aggregation if rq is idle

[ Upstream commit 3e1493f46390618ea78607cb30c58fc19e2a5035 ]

When a task wakes up on an idle rq, uclamp_rq_util_with() would max
aggregate with rq value. But since there is no task enqueued yet, the
values are stale based on the last task that was running. When the new
task actually wakes up and enqueued, then the rq uclamp values should
reflect that of the newly woken up task effective uclamp values.

This is a problem particularly for uclamp_max because it default to
1024. If a task p with uclamp_max = 512 wakes up, then max aggregation
would ignore the capping that should apply when this task is enqueued,
which is wrong.

Fix that by ignoring max aggregation if the rq is idle since in that
case the effective uclamp value of the rq will be the ones of the task
that will wake up.

Fixes: 9d20ad7dfc9a ("sched/uclamp: Add uclamp_util_with()")
Signed-off-by: Xuewen Yan <xuewen.yan@unisoc.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
[qias: Changelog]
Reviewed-by: Qais Yousef <qais.yousef@arm.com>
Link: https://lore.kernel.org/r/20210630141204.8197-1-xuewen.yan94@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: be2iscsi: Fix an error handling path in beiscsi_dev_probe()

[ Upstream commit 030e4138d11fced3b831c2761e4cecf347bae99c ]

If an error occurs after a pci_enable_pcie_error_reporting() call, it must
be undone by a corresponding pci_disable_pcie_error_reporting() call, as
already done in the remove function.

Link: https://lore.kernel.org/r/77adb02cfea7f1364e5603ecf3930d8597ae356e.1623482155.git.christophe.jaillet@wanadoo.fr
Fixes: 3567f36a09d1 ("[SCSI] be2iscsi: Fix AER handling in driver")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: rockchip: Re-add regulator-always-on for vcc_sdio for rk3399-roc-pc

[ Upstream commit eb607cd4957fb0ef97beb2a8293478be6a54240a ]

Re-add the regulator-always-on property for vcc_sdio which supplies sdmmc,
since it gets disabled during reboot now and the bootrom expects it to be
enabled  when booting from SD card. This makes rebooting impossible in that
case and requires a hard reset to boot again.

Fixes: 04a0077fdb19 ("arm64: dts: rockchip: Remove always-on properties from regulator nodes on rk3399-roc-pc.")
Signed-off-by: Alex Bee <knaerzche@gmail.com>
Link: https://lore.kernel.org/r/20210619121306.7740-1-knaerzche@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: rockchip: Re-add regulator-boot-on, regulator-always-on for vdd_gpu on rk3399-roc-pc

[ Upstream commit 06b2818678d9b35102c9816ffaf6893caf306ed0 ]

This might be a limitation of either the current panfrost driver
devfreq implementation or how the gpu is implemented in RK3399 SoC.
The gpu regulator must never get disabled or the registers get
(randomly?) inaccessable by the driver. (see all other RK3399 boards)

Fixes: ec7d731d81e7 ("arm64: dts: rockchip: Add node for gpu on rk3399-roc-pc")
Signed-off-by: Alex Bee <knaerzche@gmail.com>
Link: https://lore.kernel.org/r/20210619121446.7802-1-knaerzche@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: turris-mox-rwtm: show message about HWRNG registration

[ Upstream commit fae20160992269431507708fb74c1fd9f3c309c1 ]

Currently it is hard to determinate if on Armada 3720 device is HWRNG
by running kernel accessible or not. So print information message into
dmesg when HWRNG is available and registration was successful.

Fixes: 389711b37493 ("firmware: Add Turris Mox rWTM firmware driver")
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: turris-mox-rwtm: fail probing when firmware does not support hwrng

[ Upstream commit 2eab59cf0d2036a5a9e264f719b71c21ccf679c2 ]

When Marvell's rWTM firmware, which does not support the GET_RANDOM
command, is used, kernel prints an error message
  hwrng: no data available
every 10 seconds.

Fail probing of this driver if the rWTM firmware does not support the
GET_RANDOM command.

Fixes: 389711b37493 ("firmware: Add Turris Mox rWTM firmware driver")
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: turris-mox-rwtm: report failures better

[ Upstream commit 72f99888944c44de1c899bbe44db1e53bdc9d994 ]

Report a notice level message if a command is not supported by the rWTM
firmware.

This should not be an error, merely a notice, because the firmware can
be used on boards that do not have manufacturing information burned.

Fixes: 389711b37493 ("firmware: Add Turris Mox rWTM firmware driver")
Signed-off-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: turris-mox-rwtm: fix reply status decoding function

[ Upstream commit e34e60253d9272311831daed8a2d967cf80ca3dc ]

The status decoding function mox_get_status() currently contains an
incorrect check: if the error status is not MBOX_STS_SUCCESS, it always
returns -EIO, so the comparison to MBOX_STS_FAIL is never executed and
we don't get the actual error code sent by the firmware.

Fix this.

Signed-off-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Fixes: 389711b37493 ("firmware: Add Turris Mox rWTM firmware driver")
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/rcar_gen3_thermal: Fix coefficient calculations

[ Upstream commit 8946187ab57ffd02088e50256c73dd31f49db06d ]

The fixed value of 157 used in the calculations are only correct for
M3-W, on other Gen3 SoC it should be 167. The constant can be derived
correctly from the static TJ_3 constant and the SoC specific TJ_1 value.
Update the calculation be correct on all Gen3 SoCs.

Fixes: 4eb39f79ef44 ("thermal: rcar_gen3_thermal: Update value of Tj_1")
Reported-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210605085211.564909-1-niklas.soderlund+renesas@ragnatech.se
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx6q-dhcom: Add gpios pinctrl for i2c bus recovery

[ Upstream commit ddc873cd3c0af4faad6a00bffda21c3f775126dd ]

The i2c bus can freeze at the end of transaction so the bus can no longer work.
This scenario is improved by adding scl/sda gpios definitions to implement the
i2c bus recovery mechanism.

Fixes: 52c7a088badd ("ARM: dts: imx6q: Add support for the DHCOM iMX6 SoM and PDK2")
Signed-off-by: Christoph Niedermaier <cniedermaier@dh-electronics.com>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Marek Vasut <marex@denx.de>
Cc: NXP Linux Team <linux-imx@nxp.com>
Cc: kernel@dh-electronics.com
To: linux-arm-kernel@lists.infradead.org
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx6q-dhcom: Fix ethernet plugin detection problems

[ Upstream commit e2bdd3484890441b9cc2560413a86e8f2aa04157 ]

To make the ethernet cable plugin detection reliable the
power detection of the smsc phy has been disabled.

Fixes: 52c7a088badd ("ARM: dts: imx6q: Add support for the DHCOM iMX6 SoM and PDK2")
Signed-off-by: Christoph Niedermaier <cniedermaier@dh-electronics.com>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Marek Vasut <marex@denx.de>
Cc: NXP Linux Team <linux-imx@nxp.com>
Cc: kernel@dh-electronics.com
To: linux-arm-kernel@lists.infradead.org
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: imx6q-dhcom: Fix ethernet reset time properties

[ Upstream commit c016c26c1631f539c652b5d82242a3ca402545c1 ]

Fix ethernet reset time properties as described in
Documentation/devicetree/bindings/net/ethernet-phy.yaml

Fixes: 52c7a088badd ("ARM: dts: imx6q: Add support for the DHCOM iMX6 SoM and PDK2")
Signed-off-by: Christoph Niedermaier <cniedermaier@dh-electronics.com>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Marek Vasut <marex@denx.de>
Cc: NXP Linux Team <linux-imx@nxp.com>
Cc: kernel@dh-electronics.com
To: linux-arm-kernel@lists.infradead.org
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

thermal/drivers/sprd: Add missing MODULE_DEVICE_TABLE

[ Upstream commit 4d57fd9aeaa013a245bf1fade81e2c30a5efd491 ]

MODULE_DEVICE_TABLE is used to extract the device information out of the
driver and builds a table when being compiled. If using this macro,
kernel can find the driver if available when the device is plugged in,
and then loads that driver and initializes the device.

Fixes: 554fdbaf19b18 ("thermal: sprd: Add Spreadtrum thermal driver support")
Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210512093752.243168-1-zhang.lyra@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: am437x: align ti,pindir-d0-out-d1-in property with dt-shema

[ Upstream commit 9b11fec7345f21995f4ea4bafb0e108b9a620238 ]

ti,pindir-d0-out-d1-in property is expected to be of type boolean.
Therefore, fix the property accordingly.

Fixes: b0b039515445 ("ARM: dts: am43x-epos-evm: set data pin directions for spi0 and spi1")
Signed-off-by: Aswath Govindraju <a-govindraju@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: am335x: align ti,pindir-d0-out-d1-in property with dt-shema

[ Upstream commit 414bfe1d26b60ef20b58e36efd5363188a694bab ]

ti,pindir-d0-out-d1-in property is expected to be of type boolean.
Therefore, fix the property accordingly.

Fixes: 444d66fafab8 ("ARM: dts: add spi wifi support to cm-t335")
Signed-off-by: Aswath Govindraju <a-govindraju@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: dra7: Fix duplicate USB4 target module node

[ Upstream commit 78b4b165280d3d70e7a217599f0c06a4c0bb11f9 ]

With [1] USB4 target-module node got defined in dra74x.dtsi file.
However, the earlier definition in [2] was not removed, and this
duplication of the target module is causing boot failure on dra74
variant boards - dra7-evm, dra76-evm.

USB4 is only present in DRA74x variants, so keeping the entry in
dra74x.dtsi and removing it from the top level interconnect hierarchy
dra7-l4.dtsi file. This change makes the USB4 target module no longer
visible to AM5718, DRA71x and DRA72x so removing references to it in
their respective dts files.

[1]: commit c7b72abca61ec ("ARM: OMAP2+: Drop legacy platform data for
dra7 dwc3")
[2]: commit 549fce068a311 ("ARM: dts: dra7: Add l4 interconnect
hierarchy and ti-sysc data")

Fixes: c7b72abca61ec ("ARM: OMAP2+: Drop legacy platform data for dra7 dwc3")
Signed-off-by: Gowtham Tammana <g-tammana@ti.com>
Reviewed-by: Grygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: allwinner: a64-sopine-baseboard: change RGMII mode to TXID

[ Upstream commit bd5431b2f9b30a70f6ed964dd5ee9a6d1c397c06 ]

Although the schematics of Pine A64-LTS and SoPine Baseboard shows both
the RX and TX internal delay are enabled, they're using the same broken
RTL8211E chip batch with Pine A64+, so they should use TXID instead, not
ID.

In addition, by checking the real components soldered on both a SoPine
Baseboard and a Pine A64-LTS, RX delay is not enabled (GR69 soldered and
GR70 NC) despite the schematics says it's enabled. It's a common
situation for Pine64 boards that the NC information on schematics is not
the same with the board.

So the RGMII delay mode should be TXID on these boards.

Fixes: c2b111e59a7b ("arm64: dts: allwinner: A64 Sopine: phy-mode rgmii-id")
Signed-off-by: Icenowy Zheng <icenowy@aosc.io>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20210609083843.463750-1-icenowy@aosc.io
Signed-off-by: Sasha Levin <sashal@kernel.org>

memory: fsl_ifc: fix leak of private memory on probe failure

[ Upstream commit 8e0d09b1232d0538066c40ed4c13086faccbdff6 ]

On probe error the driver should free the memory allocated for private
structure.  Fix this by using resource-managed allocation.

Fixes: a20cbdeffce2 ("powerpc/fsl: Add support for Integrated Flash Controller")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210527154322.81253-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

memory: fsl_ifc: fix leak of IO mapping on probe failure

[ Upstream commit 3b132ab67fc7a358fff35e808fa65d4bea452521 ]

On probe error the driver should unmap the IO memory.  Smatch reports:

  drivers/memory/fsl_ifc.c:298 fsl_ifc_ctrl_probe() warn: 'fsl_ifc_ctrl_dev->gregs' not released on lines: 298.

Fixes: a20cbdeffce2 ("powerpc/fsl: Add support for Integrated Flash Controller")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210527154322.81253-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: k3-j721e-main: Fix external refclk input to SERDES

[ Upstream commit 5c6d0b55b46aeb91355e6a9616decf50a3778c91 ]

Rename the external refclk inputs to the SERDES from
dummy_cmn_refclk/dummy_cmn_refclk1 to cmn_refclk/cmn_refclk1
respectively. Also move the external refclk DT nodes outside the
cbass_main DT node. Since in j721e common processor board, only the
cmn_refclk1 is connected to 100MHz clock, fix the clock frequency.

Fixes: afd094ebe69f ("arm64: dts: ti: k3-j721e-main: Add WIZ and SERDES PHY nodes")
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Reviewed-by: Aswath Govindraju <a-govindraju@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20210603143427.28735-2-kishon@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: renesas: r8a779a0: Drop power-domains property from GIC node

[ Upstream commit 1771a33b34421050c7b830f0a8af703178ba9d36 ]

"make dtbs_check":

    arm64/boot/dts/renesas/r8a779a0-falcon.dt.yaml: interrupt-controller@f1000000: 'power-domains' does not match any of the regexes: '^(msi-controller|gic-its|interrupt-controller)@[0-9a-f]+$', '^gic-its@', '^interrupt-controller@[0-9a-f]+$', 'pinctrl-[0-9]+'
    From schema: Documentation/devicetree/bindings/interrupt-controller/arm,gic-v3.yaml

Remove the "power-domains" property, as the GIC on R-Car V3U is
always-on, and not part of a clock domain.

Fixes: 834c310f541839b6 ("arm64: dts: renesas: Add Renesas R8A779A0 SoC support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/a9ae5cbc7c586bf2c6b18ddc665ad7051bd1d206.1622560236.git.geert+renesas@glider.be
Signed-off-by: Sasha Levin <sashal@kernel.org>

reset: bail if try_module_get() fails

[ Upstream commit 4fb26fb83f0def3d39c14e268bcd4003aae8fade ]

Abort instead of returning a new reset control for a reset controller
device that is going to have its module unloaded.

Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Fixes: 61fc41317666 ("reset: Add reset controller API")
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20210607082615.15160-1-p.zabel@pengutronix.de
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: BCM5301X: Fixup SPI binding

[ Upstream commit d5aede3e6dd1b8ca574600a1ecafe1e580c53f2f ]

1. Reorder interrupts
2. Fix typo: s/spi_lr_overhead/spi_lr_overread/
3. Rename node: s/spi-nor@0/flash@0/

This fixes:
arch/arm/boot/dts/bcm4709-buffalo-wxr-1900dhp.dt.yaml: spi@18029200: interrupt-names: 'oneOf' conditional failed, one must be fixed:
        ['spi_lr_fullness_reached', 'spi_lr_session_aborted', 'spi_lr_impatient', 'spi_lr_session_done', 'spi_lr_overhead', 'mspi_done', 'mspi_halted'] is too long
        Additional items are not allowed ('spi_lr_session_aborted', 'spi_lr_impatient', 'spi_lr_session_done', 'spi_lr_overhead', 'mspi_done', 'mspi_halted' were unexpected)
        'mspi_done' was expected
        'spi_l1_intr' was expected
        'mspi_halted' was expected
        'spi_lr_fullness_reached' was expected
        'spi_lr_session_aborted' was expected
        'spi_lr_impatient' was expected
        'spi_lr_session_done' was expected
        'spi_lr_overread' was expected
        From schema: Documentation/devicetree/bindings/spi/brcm,spi-bcm-qspi.yaml
arch/arm/boot/dts/bcm4709-buffalo-wxr-1900dhp.dt.yaml: spi-nor@0: $nodename:0: 'spi-nor@0' does not match '^flash(@.*)?$'
        From schema: Documentation/devicetree/bindings/mtd/jedec,spi-nor.yaml

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dt-bindings: i2c: at91: fix example for scl-gpios

[ Upstream commit 92e669017ff1616ba7d8ba3c65f5193bc2a7acbe ]

The SCL gpio pin used by I2C bus for recovery needs to be configured as
open drain, so fix the binding example accordingly.
In relation with fix c5a283802573 ("ARM: dts: at91: Configure I2C SCL
gpio as open drain").

Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Fixes: 19e5cef058a0 ("dt-bindings: i2c: at91: document optional bus recovery properties")
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: arm_scmi: Reset Rx buffer to max size during async commands

[ Upstream commit 0cb7af474e0dbb2f500c67aa62b6db9fafa74de2 ]

During an async commands execution the Rx buffer length is at first set
to max_msg_sz when the synchronous part of the command is first sent.
However once the synchronous part completes the transport layer waits
for the delayed response which will be processed using the same xfer
descriptor initially allocated. Since synchronous response received at
the end of the xfer will shrink the Rx buffer length to the effective
payload response length, it needs to be reset again.

Raise the Rx buffer length again to max_msg_sz before fetching the
delayed response to ensure full response is read correctly from the
shared memory.

Link: https://lore.kernel.org/r/20210601102421.26581-2-cristian.marussi@arm.com
Fixes: 58ecdf03dbb9 ("firmware: arm_scmi: Add support for asynchronous commands and delayed response")
Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
[sudeep.holla: moved reset to scmi_handle_response as it could race with
               do_xfer_with_response]
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

firmware: tegra: Fix error return code in tegra210_bpmp_init()

[ Upstream commit 7fea67710e9f6a111a2c9440576f2396ccd92d57 ]

When call irq_get_irq_data() to get the IRQ's irq_data failed, an
appropriate error code -ENOENT should be returned. However, we directly
return 'err', which records the IRQ number instead of the error code.

Fixes: 139251fc2208 ("firmware: tegra: add bpmp driver for Tegra210")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: qcom: trogdor: Add no-hpd to DSI bridge node

[ Upstream commit 5f551b5ce55575b14c26933fe9b49365ea246b3d ]

We should indicate that we're not using the HPD pin on this device, per
the binding document. Otherwise if code in the future wants to enable
HPD in the bridge when this property is absent we'll be wasting power
powering hpd when we don't use it on trogdor boards. We didn't notice
this before because the kernel driver blindly disables hpd, but that
won't be true for much longer.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Cc: Douglas Anderson <dianders@chromium.org>
Fixes: 7ec3e67307f8 ("arm64: dts: qcom: sc7180-trogdor: add initial trogdor and lazor dt")
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20210324025534.1837405-1-swboyd@chromium.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: stm32: Rework LAN8710Ai PHY reset on DHCOM SoM

[ Upstream commit 1cebcf9932ab76102e8cfc555879574693ba8956 ]

The Microchip LAN8710Ai PHY requires XTAL1/CLKIN external clock to be
enabled when the nRST is toggled according to datasheet Microchip
LAN8710A/LAN8710Ai DS00002164B page 35 section 3.8.5.1 Hardware Reset:
  "
  A Hardware reset is asserted by driving the nRST input pin low. When
  driven, nRST should be held low for the minimum time detailed in
  Section 5.5.3, "Power-On nRST & Configuration Strap Timing," on page
  59 to ensure a proper transceiver reset. During a Hardware reset, an
  external clock must be supplied to the XTAL1/CLKIN signal.
  "
This is accidentally fulfilled in the current setup, where ETHCK_K is used
to supply both PHY XTAL1/CLKIN and is also fed back through eth_clk_fb to
supply ETHRX clock of the DWMAC. Hence, the DWMAC enables ETHRX clock,
that has ETHCK_K as parent, so ETHCK_K clock are also enabled, and then
the PHY reset toggles.

However, this is not always the case, e.g. in case the PHY XTAL1/CLKIN
clock are supplied by some other clock source than ETHCK_K or in case
ETHRX clock are not supplied by ETHCK_K. In the later case, ETHCK_K would
be kept disabled, while ETHRX clock would be enabled, so the PHY would
not be receiving XTAL1/CLKIN clock and the reset would fail.

Improve the DT by adding the PHY clock phandle into the PHY node, which
then also requires moving the PHY reset GPIO specifier in the same place
and that then also requires correct PHY reset GPIO timing, so add that
too.

A brief note regarding the timing, the datasheet says the reset should
stay asserted for at least 100uS and software should wait at least 200nS
after deassertion. Set both delays to 500uS which should be plenty.

Fixes: 34e0c7847dcf ("ARM: dts: stm32: Add DH Electronics DHCOM STM32MP1 SoM and PDK2 board")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Alexandre Torgue <alexandre.torgue@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: linux-stm32@st-md-mailman.stormreply.com
To: linux-arm-kernel@lists.infradead.org
Signed-off-by: Alexandre Torgue <alexandre.torgue@foss.st.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: stm32: Connect PHY IRQ line on DH STM32MP1 SoM

[ Upstream commit 516728273ddfbf51b3d0fcaac05d26e299a7b456 ]

On the production DHCOM STM32MP15xx SoM, the PHY IRQ line is connected
to the PI11 pin. Describe it in the DT as well, so the PHY IRQ can be
used e.g. to detect cable insertion and removal.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Alexandre Torgue <alexandre.torgue@st.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: linux-stm32@st-md-mailman.stormreply.com
To: linux-arm-kernel@lists.infradead.org
Signed-off-by: Alexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: renesas: r8a7796[01]: Fix OPP table entry voltages

[ Upstream commit 659b38203f04f5c3d1dc60f1a3e54b582ad3841c ]

Correct the voltages in the "Power Optimized" (<= 1.5 GHz) Cortex-A57
operating point table entries for the R-Car M3-W and M3-W+ SoCs from
0.82V to 0.83V, as per the R-Car Gen3 EC Manual Errata for Revision
0.53.

Based on a patch for R-Car M3-W in the BSP by Takeshi Kihara
<takeshi.kihara.df@renesas.com>.

Fixes: da7e3113344fda50 ("arm64: dts: renesas: r8a7796: Add OPPs table for cpu devices")
Fixes: f51746ad7d1ff6b4 ("arm64: dts: renesas: Add Renesas R8A77961 SoC support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/b9e9db907514790574429b83d070c823b36085ef.1619699909.git.geert+renesas@glider.be
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: renesas: Add missing opp-suspend properties

[ Upstream commit 44b615ac9fab16d1552cd8360454077d411e3c35 ]

Tag the highest "Power Optimized" (1.5 GHz) Cortex-A57 operating point
table entries for the RZ/G2M, R-Car M3-W and M3-W+ SoCs with the
"opp-suspend" property.  This makes sure the system will enter suspend
in the same performance state as it will be resumed by the firmware
later, avoiding state inconsistencies after resume.

Based on a patch for R-Car M3-W in the BSP by Takeshi Kihara
<takeshi.kihara.df@renesas.com>.

Fixes: 800037e815b91d8c ("arm64: dts: renesas: r8a774a1: Add operating points")
Fixes: da7e3113344fda50 ("arm64: dts: renesas: r8a7796: Add OPPs table for cpu devices")
Fixes: f51746ad7d1ff6b4 ("arm64: dts: renesas: Add Renesas R8A77961 SoC support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Link: https://lore.kernel.org/r/45a061c3b0463aac7d10664f47c4afdd999da50d.1619699721.git.geert+renesas@glider.be
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: ti: j7200-main: Enable USB2 PHY RX sensitivity workaround

[ Upstream commit a2894d85f44ba3f2bdf5806c8dc62e2ec40c1c09 ]

Enable work around feature built into the controller to address issue with
RX Sensitivity for USB2 PHY.

Fixes: 6197d7139d12 ("arm64: dts: ti: k3-j7200-main: Add USB controller")
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Aswath Govindraju <a-govindraju@ti.com>
Reviewed-by: Vignesh Raghavendra <vigneshr@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20210512153308.5840-1-a-govindraju@ti.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: r8a7779, marzen: Fix DU clock names

[ Upstream commit 6ab8c23096a29b69044209a5925758a6f88bd450 ]

"make dtbs_check" complains:

    arch/arm/boot/dts/r8a7779-marzen.dt.yaml: display@fff80000: clock-names:0: 'du.0' was expected

Change the first clock name to match the DT bindings.
This has no effect on actual operation, as the Display Unit driver in
Linux does not use the first clock name on R-Car H1, but just grabs the
first clock.

Fixes: 665d79aa47cb3983 ("ARM: shmobile: marzen: Add DU external pixel clock to DT")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Link: https://lore.kernel.org/r/9d5e1b371121883b3b3e10a3df43802a29c6a9da.1619699965.git.geert+renesas@glider.be
Signed-off-by: Sasha Levin <sashal@kernel.org>

arm64: dts: renesas: v3msk: Fix memory size

[ Upstream commit a422ec20caef6a50cf3c1efa93538888ebd576a6 ]

The V3MSK board has 2 GiB RAM according to the datasheet and schematics.

Signed-off-by: Valentine Barshak <valentine.barshak@cogentembedded.com>
[geert: Verified schematics]
Fixes: cc3e267e9bb0ce7f ("arm64: dts: renesas: initial V3MSK board device tree")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20210326121050.1578460-1-geert+renesas@glider.be
Signed-off-by: Sasha Levin <sashal@kernel.org>

rtc: fix snprintf() checking in is_rtc_hctosys()

[ Upstream commit 54b909436ede47e0ee07f1765da27ec2efa41e84 ]

The scnprintf() function silently truncates the printf() and returns
the number bytes that it was able to copy (not counting the NUL
terminator).  Thus, the highest value it can return here is
"NAME_SIZE - 1" and the overflow check is dead code.  Fix this by
using the snprintf() function which returns the number of bytes that
would have been copied if there was enough space and changing the
condition from "> NAME_SIZE" to ">= NAME_SIZE".

Fixes: 92589c986b33 ("rtc-proc: permit the /proc/driver/rtc device to use other devices")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/YJov/pcGmhLi2pEl@mwanda
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: sun8i: h3: orangepi-plus: Fix ethernet phy-mode

[ Upstream commit b19d3479f25e8a0ff24df0b46c82e50ef0f900dd ]

Commit bbc4d71d6354 ("net: phy: realtek: fix rtl8211e rx/tx delay
config") sets the RX/TX delay according to the phy-mode property in the
device tree. For the Orange Pi Plus board this is "rgmii", which is the
wrong setting.

Following the example of a900cac3750b ("ARM: dts: sun7i: a20: bananapro:
Fix ethernet phy-mode") the phy-mode is changed to "rgmii-id" which gets
the Ethernet working again on this board.

Fixes: bbc4d71d6354 ("net: phy: realtek: fix rtl8211e rx/tx delay config")
Reported-by: "B.R. Oake" <broake@mailfence.com>
Reported-by: Vagrant Cascadian <vagrant@reproducible-builds.org>
Link: https://bugs.debian.org/988574
Signed-off-by: Salvatore Bonaccorso <carnil@debian.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20210524122111.416885-1-carnil@debian.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

memory: pl353: Fix error return code in pl353_smc_probe()

[ Upstream commit 76e5624f3f9343a621dd3f4006f4e4d9c3f91e33 ]

When no child nodes are matched, an appropriate error code -ENODEV should
be returned. However, we currently do not explicitly assign this error
code to 'err'. As a result, 0 was incorrectly returned.

Fixes: fee10bd22678 ("memory: pl353: Add driver for arm pl353 static memory controller")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Link: https://lore.kernel.org/r/20210515040004.6983-1-thunder.leizhen@huawei.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

reset: brcmstb: Add missing MODULE_DEVICE_TABLE

[ Upstream commit e207457f9045343a24d936fbb67eb4b412f1c6ad ]

This patch adds missing MODULE_DEVICE_TABLE definition which generates
correct modalias for automatic loading of this driver when it is built
as an external module.

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 77750bc089e4 ("reset: Add Broadcom STB SW_INIT reset controller driver")
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Link: https://lore.kernel.org/r/1620789283-15048-1-git-send-email-zou_wei@huawei.com
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

memory: atmel-ebi: add missing of_node_put for loop iteration

[ Upstream commit 907c5bbb514a4676160e79764522fff56ce3448e ]

Early exits from for_each_available_child_of_node() should decrement the
node reference counter.  Reported by Coccinelle:

  drivers/memory/atmel-ebi.c:593:1-33: WARNING:
    Function "for_each_available_child_of_node" should have of_node_put() before return around line 604.

Fixes: 6a4ec4cd0888 ("memory: add Atmel EBI (External Bus Interface) driver")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210423101815.119341-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

memory: stm32-fmc2-ebi: add missing of_node_put for loop iteration

[ Upstream commit 2f9dc6a357ff3b82c1e54d29fb5d52b8d4a0c587 ]

Early exits from for_each_available_child_of_node() should decrement the
node reference counter.  Reported by Coccinelle:

  drivers/memory/stm32-fmc2-ebi.c:1046:1-33: WARNING:
    Function "for_each_available_child_of_node" should have of_node_put() before return around line 1051.

Fixes: 66b8173a197f ("memory: stm32-fmc2-ebi: add STM32 FMC2 EBI controller driver")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Christophe Kerello <christophe.kerello@foss.st.com>
Link: https://lore.kernel.org/r/20210423101815.119341-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: exynos: fix PWM LED max brightness on Odroid XU4

[ Upstream commit fd2f1717966535b7d0b6fe45cf0d79e94330da5f ]

There is no "max_brightness" property as pointed out by dtschema:

  arch/arm/boot/dts/exynos5422-odroidxu4.dt.yaml: led-controller: led-1: 'max-brightness' is a required property

Fixes: 6658356014cb ("ARM: dts: Add support Odroid XU4 board for exynos5422-odroidxu4")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210505135941.59898-5-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: exynos: fix PWM LED max brightness on Odroid HC1

[ Upstream commit a7e59c84cf2055a1894f45855c8319191f2fa59e ]

There is no "max_brightness" property as pointed out by dtschema:

  arch/arm/boot/dts/exynos5422-odroidhc1.dt.yaml: led-controller: led-1: 'max-brightness' is a required property

Fixes: 1ac49427b566 ("ARM: dts: exynos: Add support for Hardkernel's Odroid HC1 board")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210505135941.59898-4-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: exynos: fix PWM LED max brightness on Odroid XU/XU3

[ Upstream commit 75121e1dc9fe4def41e63d57f6a53749b88006ed ]

There is no "max_brightness" property.  This brings the intentional
brightness reduce of green LED and dtschema checks as well:

  arch/arm/boot/dts/exynos5410-odroidxu.dt.yaml: led-controller-1: led-1: 'max-brightness' is a required property

Fixes: 719f39fec586 ("ARM: dts: exynos5422-odroidxu3: Hook up PWM and use it for LEDs")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210505135941.59898-3-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: exynos: add missing of_node_put for loop iteration

[ Upstream commit 48d551bf20858240f38a0276be3016ff379918ac ]

Early exits from for_each_compatible_node() should decrement the
node reference counter.  Reported by Coccinelle:

  arch/arm/mach-exynos/exynos.c:52:1-25: WARNING:
    Function "for_each_compatible_node" should have of_node_put() before break around line 58.

Fixes: b3205dea8fbf ("ARM: EXYNOS: Map SYSRAM through generic DT bindings")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20210425174945.164612-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

reset: a10sr: add missing of_match_table reference

[ Upstream commit 466ba3c8ff4fae39e455ff8d080b3d5503302765 ]

The driver defined of_device_id table but did not use it with
of_match_table.  This prevents usual matching via devicetree and causes
a W=1 warning:

  drivers/reset/reset-a10sr.c:111:34: warning:
    ‘a10sr_reset_of_match’ defined but not used [-Wunused-const-variable=]

Reported-by: kernel test robot <lkp@intel.com>
Fixes: 627006820268 ("reset: Add Altera Arria10 SR Reset Controller")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210507112803.20012-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

reset: RESET_INTEL_GW should depend on X86

[ Upstream commit 6ab9d6219f86f0db916105444813aafce626a2f4 ]

The Intel Gateway reset controller is only present on Intel Gateway
platforms.  Hence add a dependency on X86, to prevent asking the user
about this driver when configuring a kernel without Intel Gateway
support.

Fixes: c9aef213e38cde27 ("reset: intel: Add system reset controller driver")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

reset: RESET_BRCMSTB_RESCAL should depend on ARCH_BRCMSTB

[ Upstream commit 42f6a76fbe85e5243f83a3ed76809b1ebbb7087e ]

The Broadcom STB RESCAL reset controller is only present on Broadcom
BCM7216 platforms.  Hence add a dependency on ARCH_BRCMSTB, to prevent
asking the user about this driver when configuring a kernel without
BCM7216 support.

Also, merely enabling CONFIG_COMPILE_TEST should not enable additional
code, and thus should not enable this driver by default.

Fixes: 4cf176e52397853e ("reset: Add Broadcom STB RESCAL reset controller")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ARM: dts: gemini-rut1xx: remove duplicate ethernet node

[ Upstream commit 3d3bb3d27cd371d3edb43eeb1beb8ae4e92a356d ]

Two ethernet node was added by
commit 95220046a62c ("ARM: dts: Add ethernet to a bunch of platforms")
and commit d6d0cef55e5b ("ARM: dts: Add the FOTG210 USB host to Gemini boards")

This patch removes the duplicate one.

Fixes: d6d0cef55e5b ("ARM: dts: Add the FOTG210 USB host to Gemini boards")
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

hexagon: use common DISCARDS macro

[ Upstream commit 681ba73c72302214686401e707e2087ed11a6556 ]

ld.lld warns that the '.modinfo' section is not currently handled:

ld.lld: warning: kernel/built-in.a(workqueue.o):(.modinfo) is being placed in '.modinfo'
ld.lld: warning: kernel/built-in.a(printk/printk.o):(.modinfo) is being placed in '.modinfo'
ld.lld: warning: kernel/built-in.a(irq/spurious.o):(.modinfo) is being placed in '.modinfo'
ld.lld: warning: kernel/built-in.a(rcu/update.o):(.modinfo) is being placed in '.modinfo'

The '.modinfo' section was added in commit 898490c010b5 ("moduleparam:
Save information about built-in modules in separate file") to the DISCARDS
macro but Hexagon has never used that macro.  The unification of DISCARDS
happened in commit 023bf6f1b8bf ("linker script: unify usage of discard
definition") in 2009, prior to Hexagon being added in 2011.

Switch Hexagon over to the DISCARDS macro so that anything that is
expected to be discarded gets discarded.

Link: https://lkml.kernel.org/r/20210521011239.1332345-3-nathan@kernel.org
Fixes: e95bf452a9e2 ("Hexagon: Add configuration and makefiles for the Hexagon architecture.")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Acked-by: Brian Cain <bcain@codeaurora.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Oliver Glitta <glittao@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

hexagon: handle {,SOFT}IRQENTRY_TEXT in linker script

[ Upstream commit 6fef087d0d37ba7dba8f3d75566eb4c256cd6742 ]

Patch series "hexagon: Fix build error with CONFIG_STACKDEPOT and select CONFIG_ARCH_WANT_LD_ORPHAN_WARN".

This series fixes an error with ARCH=hexagon that was pointed out by the
patch "mm/slub: use stackdepot to save stack trace in objects".

The first patch fixes that error by handling the '.irqentry.text' and
'.softirqentry.text' sections.

The second patch switches Hexagon over to the common DISCARDS macro, which
should have been done when Hexagon was merged into the tree to match
commit 023bf6f1b8bf ("linker script: unify usage of discard definition").

The third patch selects CONFIG_ARCH_WANT_LD_ORPHAN_WARN so that something
like this does not happen again.

This patch (of 3):

Patch "mm/slub: use stackdepot to save stack trace in objects" in -mm
selects CONFIG_STACKDEPOT when CONFIG_STACKTRACE_SUPPORT is selected and
CONFIG_STACKDEPOT requires IRQENTRY_TEXT and SOFTIRQENTRY_TEXT to be
handled after commit 505a0ef15f96 ("kasan: stackdepot: move
filter_irq_stacks() to stackdepot.c") due to the use of the
__{,soft}irqentry_text_{start,end} section symbols.  If those sections are
not handled, the build is broken.

$ make ARCH=hexagon CROSS_COMPILE=hexagon-linux- LLVM=1 LLVM_IAS=1 defconfig all
...
ld.lld: error: undefined symbol: __irqentry_text_start
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a

ld.lld: error: undefined symbol: __irqentry_text_end
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a

ld.lld: error: undefined symbol: __softirqentry_text_start
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a

ld.lld: error: undefined symbol: __softirqentry_text_end
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a
>>> referenced by stackdepot.c
>>>               stackdepot.o:(filter_irq_stacks) in archive lib/built-in.a
...

Add these sections to the Hexagon linker script so the build continues to
work.  ld.lld's orphan section warning would have caught this prior to the
-mm commit mentioned above:

ld.lld: warning: kernel/built-in.a(softirq.o):(.softirqentry.text) is being placed in '.softirqentry.text'
ld.lld: warning: kernel/built-in.a(softirq.o):(.softirqentry.text) is being placed in '.softirqentry.text'
ld.lld: warning: kernel/built-in.a(softirq.o):(.softirqentry.text) is being placed in '.softirqentry.text'

Link: https://lkml.kernel.org/r/20210521011239.1332345-1-nathan@kernel.org
Link: https://lkml.kernel.org/r/20210521011239.1332345-2-nathan@kernel.org
Link: https://github.com/ClangBuiltLinux/linux/issues/1381
Fixes: 505a0ef15f96 ("kasan: stackdepot: move filter_irq_stacks() to stackdepot.c")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Acked-by: Brian Cain <bcain@codeaurora.org>
Cc: Oliver Glitta <glittao@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

NFSv4/pNFS: Don't call _nfs4_pnfs_v3_ds_connect multiple times

[ Upstream commit f46f84931a0aa344678efe412d4b071d84d8a805 ]

After we grab the lock in nfs4_pnfs_ds_connect(), there is no check for
whether or not ds->ds_clp has already been initialised, so we can end up
adding the same transports multiple times.

Fixes: fc821d59209d ("pnfs/NFSv4.1: Add multipath capabilities to pNFS flexfiles servers over NFSv3")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

NFSv4/pnfs: Fix layoutget behaviour after invalidation

[ Upstream commit 0b77f97a7e42adc72bd566ff8cb733ea426f74f6 ]

If the layout gets invalidated, we should wait for any outstanding
layoutget requests for that layout to complete, and we should resend
them only after re-establishing the layout stateid.

Fixes: d29b468da4f9 ("pNFS/NFSv4: Improve rejection of out-of-order layouts")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

NFSv4/pnfs: Fix the layout barrier update

[ Upstream commit aa95edf309ef31e2df4a37ebf0e5c2ca2a6772ab ]

If we have multiple outstanding layoutget requests, the current code to
update the layout barrier assumes that the outstanding layout stateids
are updated in order. That's not necessarily the case.

Instead of using the value of lo->plh_outstanding as a guesstimate for
the window of values we need to accept, just wait to update the window
until we're processing the last one. The intention here is just to
ensure that we don't process 2^31 seqid updates without also updating
the barrier.

Fixes: 1bcf34fdac5f ("pNFS/NFSv4: Update the layout barrier when we schedule a layoutreturn")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vdpa/mlx5: Clear vq ready indication upon device reset

[ Upstream commit e3aadf2e1614174dc81d52cbb9dabb77913b11c6 ]

After device reset, the virtqueues are not ready so clear the ready
field.

Failing to do so can result in virtio_vdpa failing to load if the device
was previously used by vhost_vdpa and the old values are ready.
virtio_vdpa expects to find VQs in "not ready" state.

Fixes: 1a86b377aa21 ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Link: https://lore.kernel.org/r/20210606053128.170399-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: isa: Fix error return code in snd_cmi8330_probe()

[ Upstream commit 31028cbed26a8afa25533a10425ffa2ab794c76c ]

When 'SB_HW_16' check fails, the error code -ENODEV instead of 0 should be
returned, which is the same as that returned when 'WSS_HW_CMI8330' check
fails.

Fixes: 43bcd973d6d0 ("[ALSA] Add snd_card_set_generic_dev() call to ISA drivers")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Link: https://lore.kernel.org/r/20210707074051.2663-1-thunder.leizhen@huawei.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nfsd: Reduce contention for the nfsd_file nf_rwsem

[ Upstream commit 474bc334698df98ce07c890f1898c7e7f389b0c7 ]

When flushing out the unstable file writes as part of a COMMIT call, try
to perform most of of the data writes and waits outside the semaphore.

This means that if the client is sending the COMMIT as part of a memory
reclaim operation, then it can continue performing I/O, with contention
for the lock occurring only once the data sync is finished.

Fixes: 5011af4c698a ("nfsd: Fix stable writes")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Tested-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvme-tcp: can't set sk_user_data without write_lock

[ Upstream commit 0755d3be2d9bb6ea38598ccd30d6bbaa1a5c3a50 ]

The sk_user_data pointer is supposed to be modified only while
holding the write_lock "sk_callback_lock", otherwise
we could race with other threads and crash the kernel.

we can't take the write_lock in nvmet_tcp_state_change()
because it would cause a deadlock, but the release_work queue
will set the pointer to NULL later so we can simply remove
the assignment.

Fixes: b5332a9f3f3d ("nvmet-tcp: fix incorrect locking in state_change sk callback")

Signed-off-by: Maurizio Lombardi <mlombard@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

virtio_net: move tx vq operation under tx queue lock

[ Upstream commit 5a2f966d0f3fa0ef6dada7ab9eda74cacee96b8a ]

It's unsafe to operate a vq from multiple threads.
Unfortunately this is exactly what we do when invoking
clean tx poll from rx napi.
Same happens with napi-tx even without the
opportunistic cleaning from the receive interrupt: that races
with processing the vq in start_xmit.

As a fix move everything that deals with the vq to under tx lock.

Fixes: b92f1e6751a6 ("virtio-net: transmit napi")
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vdpa/mlx5: Fix possible failure in umem size calculation

[ Upstream commit 71ab6a7cfbae27f86a3901daab10bfe13b3a1e3a ]

umem size is a 32 bit unsigned value so assigning it to an int could
cause false failures. Set the calculated value inside the function and
modify function name to reflect the fact it updates the size.

This bug was found during code review but never had real impact to this
date.

Fixes: 1a86b377aa21 ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Link: https://lore.kernel.org/r/20210530090349.8360-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vdpa/mlx5: Fix umem sizes assignments on VQ create

[ Upstream commit e3011776af16caf423f2c36d0047acd624c274fa ]

Fix copy paste bug assigning umem1 size to umem2 and umem3. The issue
was discovered when trying to use a 1:1 MR that covers the entire
address space where firmware complained that provided sizes are not
large enough. 1:1 MRs are required to support virtio_vdpa.

Fixes: 1a86b377aa21 ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Link: https://lore.kernel.org/r/20210530090317.8284-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: tegra194: Fix tegra_pcie_ep_raise_msi_irq() ill-defined shift

[ Upstream commit f67092eff2bd40650aad54a1a1910160f41d864a ]

tegra_pcie_ep_raise_msi_irq() shifted a signed 32-bit value left by 31
bits.  The behavior of this is implementation-defined.

Replace the shift by BIT(), which is well-defined.

Found by cppcheck:

  $ cppcheck --enable=all drivers/pci/controller/dwc/pcie-tegra194.c
  Checking drivers/pci/controller/dwc/pcie-tegra194.c ...

  drivers/pci/controller/dwc/pcie-tegra194.c:1829:23: portability: Shifting signed 32-bit value by 31 bits is implementation-defined behaviour. See condition at line 1826.  [shiftTooManyBitsSigned]

  appl_writel(pcie, (1 << irq), APPL_MSI_CTRL_1);
                     ^

[bhelgaas: commit log]
Link: https://lore.kernel.org/r/20210618160219.303092-1-jonathanh@nvidia.com
Fixes: c57247f940e8 ("PCI: tegra: Add support for PCIe endpoint mode in Tegra194")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

pwm: imx1: Don't disable clocks at device remove time

[ Upstream commit 1bc6ea31cb41d50302a3c9b401964cf0a88d41f9 ]

The .remove() callback disables clocks that were not enabled in
.probe(). So just probing and then unbinding the driver results in a clk
enable imbalance.

So just drop the call to disable the clocks. (Which BTW was also in the
wrong order because the call makes the PWM unfunctional and so should
have come only after pwmchip_remove()).

Fixes: 9f4c8f9607c3 ("pwm: imx: Add ipg clock operation")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: intel-gw: Fix INTx enable

[ Upstream commit 655832d12f2251e04031294f547c86935a0a126d ]

The legacy PCI interrupt lines need to be enabled using PCIE_APP_IRNEN bits
13 (INTA), 14 (INTB), 15 (INTC) and 16 (INTD). The old code however was
taking (for example) "13" as raw value instead of taking BIT(13).  Define
the legacy PCI interrupt bits using the BIT() macro and then use these in
PCIE_APP_IRN_INT.

Link: https://lore.kernel.org/r/20210106135540.48420-1-martin.blumenstingl@googlemail.com
Fixes: ed22aaaede44 ("PCI: dwc: intel: PCIe RC controller driver")
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rahul Tanwar <rtanwar@maxlinear.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/fpu: Limit xstate copy size in xstateregs_set()

[ Upstream commit 07d6688b22e09be465652cf2da0da6bf86154df6 ]

If the count argument is larger than the xstate size, this will happily
copy beyond the end of xstate.

Fixes: 91c3dba7dbc1 ("x86/fpu/xstate: Fix PTRACE frames for XSAVES")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Andy Lutomirski <luto@kernel.org>
Reviewed-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20210623121452.120741557@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>

x86/fpu: Fix copy_xstate_to_kernel() gap handling

[ Upstream commit 9625895011d130033d1bc7aac0d77a9bf68ff8a6 ]

The gap handling in copy_xstate_to_kernel() is wrong when XSAVES is in
use.

Using init_fpstate for copying the init state of features which are
not set in the xstate header is only correct for the legacy area, but
not for the extended features area because when XSAVES is in use then
init_fpstate is in compacted form which means the xstate offsets which
are used to copy from init_fpstate are not valid.

Fortunately, this is not a real problem today because all extended
features in use have an all-zeros init state, but it is wrong
nevertheless and with a potentially dynamically sized init_fpstate this
would result in an access outside of the init_fpstate.

Fix this by keeping track of the last copied state in the target buffer and
explicitly zero it when there is a feature or alignment gap.

Use the compacted offset when accessing the extended feature space in
init_fpstate.

As this is not a functional issue on older kernels this is intentionally
not tagged for stable.

Fixes: b8be15d58806 ("x86/fpu/xstate: Re-enable XSAVES")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20210623121451.294282032@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to avoid adding tab before doc section

[ Upstream commit 3c16dc40aab84bab9cf54c2b61a458bb86b180c3 ]

Otherwise whole section after tab will be invisible in compiled
html format document.

Cc: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Fixes: 89272ca1102e ("docs: filesystems: convert f2fs.txt to ReST")
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: iproc: Support multi-MSI only on uniprocessor kernel

[ Upstream commit 2dc0a201d0f59e6818ef443609f0850a32910844 ]

The interrupt affinity scheme used by this driver is incompatible with
multi-MSI as it implies moving the doorbell address to that of another MSI
group.  This isn't possible for multi-MSI, as all the MSIs must have the
same doorbell address. As such it is restricted to systems with a single
CPU.

Link: https://lore.kernel.org/r/20210622152630.40842-2-sbodomerle@gmail.com
Fixes: fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Sandor Bodo-Merle <sbodomerle@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Acked-by: Pali Rohár <pali@kernel.org>
Acked-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: iproc: Fix multi-MSI base vector number allocation

[ Upstream commit e673d697b9a234fc3544ac240e173cef8c82b349 ]

Commit fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
introduced multi-MSI support with a broken allocation mechanism (it failed
to reserve the proper number of bits from the inner domain).  Natural
alignment of the base vector number was also not guaranteed.

Link: https://lore.kernel.org/r/20210622152630.40842-1-sbodomerle@gmail.com
Fixes: fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
Reported-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Sandor Bodo-Merle <sbodomerle@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Acked-by: Pali Rohár <pali@kernel.org>
Acked-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ubifs: Set/Clear I_LINKABLE under i_lock for whiteout inode

[ Upstream commit a801fcfeef96702fa3f9b22ad56c5eb1989d9221 ]

xfstests-generic/476 reports a warning message as below:

WARNING: CPU: 2 PID: 30347 at fs/inode.c:361 inc_nlink+0x52/0x70
Call Trace:
  do_rename+0x502/0xd40 [ubifs]
  ubifs_rename+0x8b/0x180 [ubifs]
  vfs_rename+0x476/0x1080
  do_renameat2+0x67c/0x7b0
  __x64_sys_renameat2+0x6e/0x90
  do_syscall_64+0x66/0xe0
  entry_SYSCALL_64_after_hwframe+0x44/0xae

Following race case can cause this:
         rename_whiteout(Thread 1)             wb_workfn(Thread 2)
ubifs_rename
  do_rename
                                          __writeback_single_inode
    spin_lock(&inode->i_lock)
    whiteout->i_state |= I_LINKABLE
                                            inode->i_state &= ~dirty;
---- How race happens on i_state:
    (tmp = whiteout->i_state | I_LINKABLE)
                           (tmp = inode->i_state & ~dirty)
    (whiteout->i_state = tmp)
                           (inode->i_state = tmp)
----
    spin_unlock(&inode->i_lock)
    inc_nlink(whiteout)
    WARN_ON(!(inode->i_state & I_LINKABLE)) !!!

Fix to add i_lock to avoid i_state update race condition.

Fixes: 9e0a1fff8db56ea ("ubifs: Implement RENAME_WHITEOUT")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nfs: fix acl memory leak of posix_acl_create()

[ Upstream commit 1fcb6fcd74a222d9ead54d405842fc763bb86262 ]

When looking into another nfs xfstests report, I found acl and
default_acl in nfs3_proc_create() and nfs3_proc_mknod() error
paths are possibly leaked. Fix them in advance.

Fixes: 013cdf1088d7 ("nfs: use generic posix ACL infrastructure for v3 Posix ACLs")
Cc: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: Anna Schumaker <anna.schumaker@netapp.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

SUNRPC: prevent port reuse on transports which don't request it.

[ Upstream commit bc1c56e9bbe92766d017efb5f0a0c71f80da5570 ]

If an RPC client is created without RPC_CLNT_CREATE_REUSEPORT, it should
not reuse the source port when a TCP connection is re-established.
This is currently implemented by preventing the source port being
recorded after a successful connection (the call to xs_set_srcport()).

However the source port is also recorded after a successful bind in xs_bind().
This may not be needed at all and certainly is not wanted when
RPC_CLNT_CREATE_REUSEPORT wasn't requested.

So avoid that assignment when xprt.reuseport is not set.

With this change, NFSv4.1 and later mounts use a different port number on
each connection.  This is helpful with some firewalls which don't cope
well with port reuse.

Signed-off-by: NeilBrown <neilb@suse.de>
Fixes: e6237b6feb37 ("NFSv4.1: Don't rebind to the same source port when reconnecting to the server")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

watchdog: jz4740: Fix return value check in jz4740_wdt_probe()

[ Upstream commit 29e85f53fb58b45b9e9276dcdf1f1cb762dd1c9f ]

In case of error, the function device_node_to_regmap() returns
ERR_PTR() and never returns NULL. The NULL test in the return
value check should be replaced with IS_ERR().

Fixes: 6d532143c915 ("watchdog: jz4740: Use regmap provided by TCU driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Paul Cercueil <paul@crapouillou.net>
Link: https://lore.kernel.org/r/20210304045909.945799-1-weiyongjun1@huawei.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

watchdog: aspeed: fix hardware timeout calculation

[ Upstream commit e7dc481c92060f9ce872878b0b7a08c24713a7e5 ]

Fix hardware timeout calculation in aspeed_wdt_set_timeout function to
ensure the reload value does not exceed the hardware limit.

Fixes: efa859f7d786 ("watchdog: Add Aspeed watchdog driver")
Reported-by: Amithash Prasad <amithash@fb.com>
Signed-off-by: Tao Ren <rentao.bupt@gmail.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20210417034249.5978-1-rentao.bupt@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ubifs: journal: Fix error return code in ubifs_jnl_write_inode()

[ Upstream commit a2c2a622d41168f9fea2aa3f76b9fbaa88531aac ]

Fix to return a negative error code from the error handling case instead
of 0, as done elsewhere in this function.

Fixes: 9ca2d7326444 ("ubifs: Limit number of xattrs per inode")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ubifs: Fix off-by-one error

[ Upstream commit d984bcf5766dbdbe95d325bb8a1b49a996fecfd4 ]

An inode is allowed to have ubifs_xattr_max_cnt() xattrs, so we must
complain only when an inode has more xattrs, having exactly
ubifs_xattr_max_cnt() xattrs is fine.
With this the maximum number of xattrs can be created without hitting
the "has too many xattrs" warning when removing it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

um: fix error return code in winch_tramp()

[ Upstream commit ccf1236ecac476d9d2704866d9a476c86e387971 ]

Fix to return a negative error code from the error handling case instead
of 0, as done elsewhere in this function.

Fixes: 89df6bfc0405 ("uml: DEBUG_SHIRQ fixes")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Acked-By: anton.ivanov@cambridgegreys.com
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

um: fix error return code in slip_open()

[ Upstream commit b77e81fbe5f5fb4ad9a61ec80f6d1e30b6da093a ]

Fix to return a negative error code from the error handling case instead
of 0, as done elsewhere in this function.

Fixes: a3c77c67a443 ("[PATCH] uml: slirp and slip driver cleanups and fixes")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Acked-By: anton.ivanov@cambridgegreys.com
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

misc: alcor_pci: fix inverted branch condition

commit 281e468446994a7672733af2bf941f4110d4a895 upstream.

This patch fixes a trivial mistake that I made in the previous attempt
in fixing the null bridge issue. The branch condition is inverted and we
should call alcor_pci_find_cap_offset() only if bridge is not null.

Reported-by: Colin Ian King <colin.king@canonical.com>
Fixes: 3ce3e45cc333 ("misc: alcor_pci: fix null-ptr-deref when there is no PCI bridge")
Signed-off-by: Tong Zhang <ztong0001@gmail.com>
Link: https://lore.kernel.org/r/20210522043725.602179-1-ztong0001@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>