drm/nouveau/kms/nv50-: fix file release memory leak

When using single_open() for opening, single_release() should be
called, otherwise the 'op' allocated in single_open() will be leaked.

Fixes: 12885ecbfe62 ("drm/nouveau/kms/nvd9-: Add CRC support")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210911075023.3969054-1-yangyingliang@huawei.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/nouveau: avoid a use-after-free when BO init fails

nouveau_bo_init() is backed by ttm_bo_init() and ferries its return code
back to the caller. On failures, ttm_bo_init() invokes the provided
destructor which should de-initialize and free the memory.

Thus, when nouveau_bo_init() returns an error the gem object has already
been released and the memory freed by nouveau_bo_del_ttm().

Fixes: 019cbd4a4feb ("drm/nouveau: Initialize GEM object before TTM object")
Cc: Thierry Reding <treding@nvidia.com>
Signed-off-by: Jeremy Cline <jcline@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201203000220.18238-1-jcline@redhat.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

DRM: delete DRM IRQ legacy midlayer docs

Remove documentation associated with the removal of the DRM IRQ legacy
midlayer.

Eliminates these documentation warnings:

../drivers/gpu/drm/drm_irq.c:1: warning: 'irq helpers' not found
../drivers/gpu/drm/drm_irq.c:1: warning: no structured comments found

Fixes: c1736b9008cb ("drm: IRQ midlayer is now legacy")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: dri-devel@lists.freedesktop.org
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: linux-doc@vger.kernel.org
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20211005025312.20913-1-rdunlap@infradead.org
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

video: fbdev: gbefb: Only instantiate device when built for IP32

The gbefb driver not only registers a driver but also the device for that
driver. This is all well and good when run on the IP32 machines that are
supported by the driver but since the driver supports building with
COMPILE_TEST we might also be building on other platforms which do not have
this hardware and will crash instantiating the driver. Add an IS_ENABLED()
check so we compile out the device registration if we don't have the Kconfig
option for the machine enabled.

Fixes: 552ccf6b259d290c0c ("video: fbdev: gbefb: add COMPILE_TEST support")
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210921212102.30803-1-broonie@kernel.org
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

fbdev: simplefb: fix Kconfig dependencies

Configurations with both CONFIG_FB_SIMPLE=y and CONFIG_DRM_SIMPLEDRM=m
are allowed by Kconfig because the 'depends on !DRM_SIMPLEDRM' dependency
does not disallow FB_SIMPLE as long as SIMPLEDRM is not built-in. This
can however result in a build failure when cfb_fillrect() etc are then
also in loadable modules:

x86_64-linux-ld: drivers/video/fbdev/simplefb.o:(.rodata+0x1f8): undefined reference to `cfb_fillrect'
x86_64-linux-ld: drivers/video/fbdev/simplefb.o:(.rodata+0x200): undefined reference to `cfb_copyarea'
x86_64-linux-ld: drivers/video/fbdev/simplefb.o:(.rodata+0x208): undefined reference to `cfb_imageblit'

To work around this, change FB_SIMPLE to be a 'tristate' symbol,
which still allows both to be =m together, but not one of them to
be =y if the other one is =m. If a distro kernel picks this
configuration, it can be determined by local policy which of
the two modules gets loaded. The 'of_chosen' export is needed
as this is the first loadable module referencing it.

Alternatively, the Kconfig dependency could be changed to
'depends on DRM_SIMPLEDRM=n', which would forbid the configuration
with both drivers.

Fixes: 11e8f5fd223b ("drm: Add simpledrm driver")
Acked-by: Rob Herring <robh@kernel.org> # for drivers/of/
Link: https://lore.kernel.org/all/20210721151839.2484245-1-arnd@kernel.org/
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch> # fbdev support
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Borislav Petkov <bp@suse.de>
Cc: Javier Martinez Canillas <javierm@redhat.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Peter Collingbourne <pcc@google.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Cc: linux-kernel@vger.kernel.org
Cc: <stable@vger.kernel.org> # v5.14+
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210928145243.1098064-1-arnd@kernel.org
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/panel: abt-y030xx067a: yellow tint fix

The previous parameters caused an unbalanced yellow tint.

Fixes: 7467389bdafb ("drm/panel: Add ABT Y030XX067A 3.0" 320x480 panel")
Signed-off-by: Christophe Branchereau <cbranchereau@gmail.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
[Paul: Add Fixes: tag, and fix case and punctuation in commit message]
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20210914092716.2370039-1-cbranchereau@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

dt-bindings: panel: ili9341: correct indentation

Correct indentation warning:
  ilitek,ili9341.yaml:25:9: [warning] wrong indentation: expected 10 but found 8 (indentation)

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210819101020.26368-1-krzysztof.kozlowski@canonical.com
(cherry picked from commit 333ba0d9d5d5a2cf1f6bbb754045e4f2cb3ed22d)
Link: https://lore.kernel.org/dri-devel/CAL_JsqKcTfgnXNYzGDSFhKS2udhw2Dvk04ODwTxUdDRQjKdT0Q@mail.gmail.com/
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/nouveau/fifo/ga102: initialise chid on return from channel creation

Turns out caller isn't zero-initialised after-all.

Fixes: 49b2dfc08182 ("drm/nouveau/ga102-: support ttm buffer moves via copy engine")
Reported-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210921090735.247236-1-skeggsb@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/rockchip: Update crtc fixup to account for fractional clk change

After commit 928f9e268611 ("clk: fractional-divider: Hide
clk_fractional_divider_ops from wide audience") was merged it appears
that the DSI panel on my Odroid Go Advance stopped working. Upon closer
examination of the problem, it looks like it was the fixup in the
rockchip_drm_vop.c file was causing the issue. The changes made to the
clk driver appear to change some assumptions made in the fixup.

After debugging the working 5.14 kernel and the no-longer working
5.15 kernel, it looks like this was broken all along but still
worked, whereas after the fractional clock change it stopped
working despite the issue (it went from sort-of broken to very broken).

In the 5.14 kernel the dclk_vopb_frac was being requested to be set to
17000999 on my board. The clock driver was taking the value of the
parent clock and attempting to divide the requested value from it
(17000000/17000999 = 0), then subtracting 1 from it (making it -1),
and running it through fls_long to get 64. It would then subtract
the value of fd->mwidth from it to get 48, and then bit shift
17000999 to the left by 48, coming up with a very large number of
7649082492112076800. This resulted in a numerator of 65535 and a
denominator of 1 from the clk driver. The driver seemingly would
try again and get a correct 1:1 value later, and then move on.

Output from my 5.14 kernel (with some printfs for good measure):
[    2.830066] rockchip-drm display-subsystem: bound ff460000.vop (ops vop_component_ops)
[    2.839431] rockchip-drm display-subsystem: bound ff450000.dsi (ops dw_mipi_dsi_rockchip_ops)
[    2.855980] Clock is dclk_vopb_frac
[    2.856004] Scale 64, Rate 7649082492112076800, Oldrate 17000999, Parent Rate 17000000, Best Numerator 65535, Best Denominator 1, fd->mwidth 16
[    2.903529] Clock is dclk_vopb_frac
[    2.903556] Scale 0, Rate 17000000, Oldrate 17000000, Parent Rate 17000000, Best Numerator 1, Best Denominator 1, fd->mwidth 16
[    2.903579] Clock is dclk_vopb_frac
[    2.903583] Scale 0, Rate 17000000, Oldrate 17000000, Parent Rate 17000000, Best Numerator 1, Best Denominator 1, fd->mwidth 16

Contrast this with 5.15 after the clk change where the rate of 17000999
was getting passed and resulted in numerators/denomiators of 17001/
17000.

Output from my 5.15 kernel (with some printfs added for good measure):
[    2.817571] rockchip-drm display-subsystem: bound ff460000.vop (ops vop_component_ops)
[    2.826975] rockchip-drm display-subsystem: bound ff450000.dsi (ops dw_mipi_dsi_rockchip_ops)
[    2.843430] Rate 17000999, Parent Rate 17000000, Best Numerator 17018, Best Denominator 17017
[    2.891073] Rate 17001000, Parent Rate 17000000, Best Numerator 17001, Best Denominator 17000
[    2.891269] Rate 17001000, Parent Rate 17000000, Best Numerator 17001, Best Denominator 17000
[    2.891281] Rate 17001000, Parent Rate 17000000, Best Numerator 17001, Best Denominator 17000

I have tested the change extensively on my Odroid Go Advance (Rockchip
RK3326) and it appears to work well. However, this change will affect
all Rockchip SoCs that use this driver so I believe further testing
is warranted. Please note that without this change I can confirm
at least all PX30s with DSI panels will stop working with the 5.15
kernel.

Upon advice from Doug Anderson <dianders@chromium.org> it was decided
that we would first check if the clock rate can be set exactly as
requested, and only if it could not would we then add 999 to it and
attempt the process again. This way we can preserve the behavior for
clocks that still need it while resolving the specific issue for the
PX30 and DSI panels (since it is using a fractional clock).

Changes since v2:
 - Moved fixes to correct location.

Changes since v1:
 - Made the addition of 999 conditional based on whether the clock
   subsystem can set the actual clock rate as requested.
 - Updated the notes in the fixup routine to reflect this new behavior.
 - Added reference to original commit, as this has technically been
   broken since then however only now is it an issue due to the clock
   changes.

Fixes: 4e7cf74fa3b2 ("clk: fractional-divider: Export approximation algorithm to the CCF users")
Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210916202907.18394-1-macroalpha82@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/nouveau/ga102-: support ttm buffer moves via copy engine

We don't currently have any kind of real acceleration on Ampere GPUs,
but the TTM memcpy() fallback paths aren't really designed to handle
copies between different devices, such as on Optimus systems, and
result in a kernel OOPS.

A few options were investigated to try and fix this, but didn't work
out, and likely would have resulted in a very unpleasant experience
for users anyway.

This commit adds just enough support for setting up a single channel
connected to a copy engine, which the kernel can use to accelerate
the buffer copies between devices.  Userspace has no access to this
incomplete channel support, but it's suitable for TTM's needs.

A more complete implementation of host(fifo) for Ampere GPUs is in
the works, but the required changes are far too invasive that they
would be unsuitable to backport to fix this issue on current kernels.

v2: fix GPFIFO length in RAMFC (reported by Karol)

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Karol Herbst <kherbst@redhat.com>
Cc: <stable@vger.kernel.org> # v5.12+
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Tested-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210916220406.666454-1-skeggsb@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/nouveau/kms/tu102-: delay enabling cursor until after assign_windows

Prevent NVD core channel error code 67 occuring and hanging display,
managed to reproduce on GA102 while testing suspend/resume scenarios.

Required extension of earlier commit to fix interactions with EFI.

Fixes: e78b1b545c6c ("drm/nouveau/kms/nv50: workaround EFI GOP window channel format differences")
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Karol Herbst <kherbst@redhat.com>
Cc: <stable@vger.kernel.org> # v5.12+
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210906005628.11499-2-skeggsb@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/sun4i: dw-hdmi: Fix HDMI PHY clock setup

Recent rework, which made HDMI PHY driver a platform device, inadvertely
reversed clock setup order. HW is very touchy about it. Proper way is to
handle controllers resets and clocks first and HDMI PHYs second.

Currently, without this fix, first mode set completely fails (nothing on
HDMI monitor) on H3 era PHYs. On H6, it still somehow work.

Move HDMI PHY reset & clocks handling to sun8i_hdmi_phy_init() which
will assure that code is executed after controllers reset & clocks are
handled. Additionally, add sun8i_hdmi_phy_deinit() which will deinit
them at controllers driver unload.

Tested on A64, H3, H6 and R40.

Fixes: 9bf3797796f5 ("drm/sun4i: dw-hdmi: Make HDMI PHY into a platform device")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210915175836.3158839-1-jernej.skrabec@gmail.com
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/vc4: hdmi: Remove unused struct

Commitc7d30623540b ("drm/vc4: hdmi: Remove unused struct") removed the
references to the vc4_hdmi_audio_widgets and vc4_hdmi_audio_routes
structures, but not the structures themselves resulting in two warnings.
Remove it.

Fixes: c7d30623540b ("drm/vc4: hdmi: Remove unused struct")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210819140753.930751-1-maxime@cerno.tech
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

drm/kmb: Enable alpha blended second plane

Enable one additional plane that is alpha blended on top
of the primary plane.

This also fixes the below warnings when building with
-Warray-bounds:

drivers/gpu/drm/kmb/kmb_plane.c:135:20: warning: array subscript 3 is
above array bounds of 'struct layer_status[1]' [-Warray-bounds]
drivers/gpu/drm/kmb/kmb_plane.c:132:20: warning: array subscript 2 is
above array bounds of 'struct layer_status[1]' [-Warray-bounds]
drivers/gpu/drm/kmb/kmb_plane.c:129:20: warning: array subscript 1 is
above array bounds of 'struct layer_status[1]' [-Warray-bounds]

v2: corrected previous patch dependecies so it builds

Signed-off-by: Edmund Dea <edmund.j.dea@intel.com>
Signed-off-by: Anitha Chrisanthus <anitha.chrisanthus@intel.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.kernel.org/project/dri-devel/patch/20210728003126.1425028-13-anitha.chrisanthus@intel.com/
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

xen/balloon: fix cancelled balloon action

In case a ballooning action is cancelled the new kernel thread handling
the ballooning might end up in a busy loop.

Fix that by handling the cancelled action gracefully.

While at it introduce a short wait for the BP_WAIT case.

Cc: stable@vger.kernel.org
Fixes: 8480ed9c2bbd56 ("xen/balloon: use a kernel thread instead a workqueue")
Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Tested-by: Jason Andryuk <jandryuk@gmail.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/20211005133433.32008-1-jgross@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

ksmbd: add the check to vaildate if stream protocol length exceeds maximum value

This patch add MAX_STREAM_PROT_LEN macro and check if stream protocol
length exceeds maximum value. opencode pdu size check in
ksmbd_pdu_size_has_room().

Cc: Tom Talpey <tom@talpey.com>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Ralph Böhme <slow@samba.org>
Acked-by: Hyunchul Lee <hyc.lee@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>

ARM: dts: omap3430-sdp: Fix NAND device node

Nand is on CS1 so reg properties first field should be 1 not 0.

Fixes: 44e4716499b8 ("ARM: dts: omap3: Fix NAND device nodes")
Cc: stable@vger.kernel.org # v4.6+
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>

bus: ti-sysc: Use CLKDM_NOAUTO for dra7 dcan1 for errata i893

Commit 94f6345712b3 ("bus: ti-sysc: Implement quirk handling for
CLKDM_NOAUTO") should have also added the quirk for dra7 dcan1 in
addition to dcan2 for errata i893 handling.

Let's also pass the quirk flag for legacy mode booting for if "ti,hwmods"
dts property is used with related dcan hwmod data. This should be only
needed if anybody needs to git bisect earlier stable trees though.

Fixes: 94f6345712b3 ("bus: ti-sysc: Implement quirk handling for CLKDM_NOAUTO")
Signed-off-by: Tony Lindgren <tony@atomide.com>

Merge branch 'pruss-fix' into fixes

Merge in a fix for pruss reset issue caused by enabling pruss for am335x.

xtensa: xtfpga: Try software restart before simulating CPU reset

Rebooting xtensa images loaded with the '-kernel' option in qemu does
not work. When executing a reboot command, the qemu session either hangs
or experiences an endless sequence of error messages.

  Kernel panic - not syncing: Unrecoverable error in exception handler

Reset code jumps to the CPU restart address, but Linux can not recover
from there because code and data in the kernel init sections have been
discarded and overwritten at this point.

XTFPGA platforms have a means to reset the CPU by writing 0xdead into a
specific FPGA IO address. When used in QEMU the kernel image loaded with
the '-kernel' option gets restored to its original state allowing the
machine to boot successfully.

Use that mechanism to attempt a platform reset. If it does not work,
fall back to the existing mechanism.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>

xtensa: xtfpga: use CONFIG_USE_OF instead of CONFIG_OF

Use platform data to initialize xtfpga device drivers when CONFIG_USE_OF
is not selected. This fixes xtfpga networking when CONFIG_USE_OF is not
selected but CONFIG_OF is.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>

objtool: Remove redundant 'len' field from struct section

The section structure already contains sh_size, so just remove the extra
'len' member that requires extra mirroring and potential confusion.

Suggested-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reviewed-by: Miroslav Benes <mbenes@suse.cz>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Link: https://lore.kernel.org/r/20210822225037.54620-3-joe.lawrence@redhat.com
Cc: Andy Lavr <andy.lavr@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: x86@kernel.org
Cc: linux-kernel@vger.kernel.org

objtool: Make .altinstructions section entry size consistent

Commit e31694e0a7a7 ("objtool: Don't make .altinstructions writable")
aligned objtool-created and kernel-created .altinstructions section
flags, but there remains a minor discrepency in their use of a section
entry size: objtool sets one while the kernel build does not.

While sh_entsize of sizeof(struct alt_instr) seems intuitive, this small
deviation can cause failures with external tooling (kpatch-build).

Fix this by creating new .altinstructions sections with sh_entsize of 0
and then later updating sec->sh_size as alternatives are added to the
section.  An added benefit is avoiding the data descriptor and buffer
created by elf_create_section(), but previously unused by
elf_add_alternative().

Fixes: 9bc0bb50727c ("objtool/x86: Rewrite retpoline thunk calls")
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reviewed-by: Miroslav Benes <mbenes@suse.cz>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Link: https://lore.kernel.org/r/20210822225037.54620-2-joe.lawrence@redhat.com
Cc: Andy Lavr <andy.lavr@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: x86@kernel.org
Cc: linux-kernel@vger.kernel.org

objtool: Remove reloc symbol type checks in get_alt_entry()

Converting a special section's relocation reference to a symbol is
straightforward.  No need for objtool to complain that it doesn't know
how to handle it.  Just handle it.

This fixes the following warning:

  arch/x86/kvm/emulate.o: warning: objtool: __ex_table+0x4: don't know how to handle reloc symbol type: kvm_fastop_exception

Fixes: 24ff65257375 ("objtool: Teach get_alt_entry() about more relocation types")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Link: https://lore.kernel.org/r/feadbc3dfb3440d973580fad8d3db873cbfe1694.1633367242.git.jpoimboe@redhat.com
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: x86@kernel.org
Cc: Miroslav Benes <mbenes@suse.cz>
Cc: linux-kernel@vger.kernel.org

perf tests attr: Add missing topdown metrics events

The Topdown metrics events were added as 'perf stat' default events
since commit 42641d6f4d15e6db ("perf stat: Add Topdown metrics events as
default events").

However, the perf attr tests were not updated
accordingly.

The perf attr test fails on the platform which supports Topdown metrics.

  # perf test 17

    17: Setup struct perf_event_attr        :FAILED!

Add Topdown metrics events into perf attr test cases. Make them optional
since they are only available on newer platforms.

Fixes: 42641d6f4d15e6db ("perf stat: Add Topdown metrics events as default events")
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/1633031566-176517-1-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Merge tag 'warning-fixes-20211005' of git://git./linux/kernel/git/dhowells/linux-fs

Pull misc fs warning fixes from David Howells:
 "The first four patches fix kerneldoc warnings in fscache, afs, 9p and
  nfs - they're mostly just comment changes, though there's one place in
  9p where a comment got detached from the function it was attached to
  (v9fs_fid_add) and has to switch places with a function that got
  inserted between (__add_fid).

  The patch on the end removes an unused symbol in fscache"

* tag 'warning-fixes-20211005' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  fscache: Remove an unused static variable
  fscache: Fix some kerneldoc warnings shown up by W=1
  9p: Fix a bunch of kerneldoc warnings shown up by W=1
  afs: Fix kerneldoc warning shown up by W=1
  nfs: Fix kerneldoc warning shown up by W=1

tools include UAPI: Sync sound/asound.h copy with the kernel sources

Picking the changes from:

  09d23174402da0f1 ("ALSA: rawmidi: introduce SNDRV_RAWMIDI_IOCTL_USER_PVERSION")

Which entails no changes in the tooling side as it doesn't introduce new
SNDRV_PCM_IOCTL_ ioctls.

To silence this perf tools build warning:

  Warning: Kernel ABI header at 'tools/include/uapi/sound/asound.h' differs from latest version at 'include/uapi/sound/asound.h'
  diff -u tools/include/uapi/sound/asound.h include/uapi/sound/asound.h

Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf build: Fix plugin static linking with libopencsd on ARM and ARM64

Filter out -static flag when building plugins as they are always built
as dynamic libraries and -static and -dynamic don't work well together
on arm and arm64.

Signed-off-by: Branislav Rankov <branislav.rankov@arm.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Mark Brown <Mark.Brown@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: coresight@lists.linaro.org
Cc: nd@arm.com
Link: https://lore.kernel.org/r/e88952b3-2470-da96-dee9-e247a1759cd0@arm.com
Signed-off-by: Tamas Zsoldos <tamas.zsoldos@arm.com>
[ Split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf build: Add missing -lstdc++ when linking with libopencsd

Add -lstdc++ to perf when linking libopencsd as it is a dependency. It
does not hurt to add it when dynamic linking.

Signed-off-by: Branislav Rankov <branislav.rankov@arm.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Mark Brown <Mark.Brown@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: coresight@lists.linaro.org
Cc: nd@arm.com
Link: https://lore.kernel.org/r/e88952b3-2470-da96-dee9-e247a1759cd0@arm.com
Signed-off-by: Tamas Zsoldos <tamas.zsoldos@arm.com>
[ Split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf jevents: Free the sys_event_tables list after processing entries

The compiler reports that free_sys_event_tables() is dead code.

But according to the semantics, the "LIST_HEAD(sys_event_tables)" should
also be released, just like we do with 'arch_std_events' in main().

Fixes: e9d32c1bf0cd7a98 ("perf vendor events: Add support for arch standard events")
Signed-off-by: Like Xu <likexu@tencent.com>
Reviewed-by: John Garry <john.garry@huawei.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lore.kernel.org/lkml/20210928102938.69681-1-likexu@tencent.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

drm/amdgpu: handle the case of pci_channel_io_frozen only in amdgpu_pci_resume

In current code, when a PCI error state pci_channel_io_normal is detectd,
it will report PCI_ERS_RESULT_CAN_RECOVER status to PCI driver, and PCI
driver will continue the execution of PCI resume callback report_resume by
pci_walk_bridge, and the callback will go into amdgpu_pci_resume
finally, where write lock is releasd unconditionally without acquiring
such lock first. In this case, a deadlock will happen when other threads
start to acquire the read lock.

To fix this, add a member in amdgpu_device strucutre to cache
pci_channel_state, and only continue the execution in amdgpu_pci_resume
when it's pci_channel_io_frozen.

Fixes: c9a6b82f45e2 ("drm/amdgpu: Implement DPC recovery")
Suggested-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: init iommu after amdkfd device init

This patch is to fix clinfo failure in Raven/Picasso:

Number of platforms: 1
  Platform Profile: FULL_PROFILE
  Platform Version: OpenCL 2.2 AMD-APP (3364.0)
  Platform Name: AMD Accelerated Parallel Processing
  Platform Vendor: Advanced Micro Devices, Inc.
  Platform Extensions: cl_khr_icd cl_amd_event_callback

  Platform Name: AMD Accelerated Parallel Processing Number of devices: 0

Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Tested-by: James Zhu <James.Zhu@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: remove redundant iommu cleanup code

kfd_resume doesn't involve iommu operation, remove
redundant iommu cleanup code.

Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Tested-by: James Zhu <James.Zhu@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: fix dependencies for DRM_AMD_DC_SI

Depends on DRM_AMDGPU_SI and DRM_AMD_DC

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: During s0ix don't wait to signal GFXOFF

In the rare event when GFX IP suspend coincides with a s0ix entry, don't
schedule a delayed work, instead signal PMFW immediately to allow GFXOFF
entry. GFXOFF is a prerequisite for s0ix entry. PMFW needs to be
signaled about GFXOFF status before amd-pmc module passes OS HINT
to PMFW telling that everything is ready for a safe s0ix entry.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1712

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Documentation/gpu: remove spurious "+" in amdgpu.rst

Not sure why that was there.  Remove it.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: fix a potential ttm->sg memory leak

Memory is allocated for ttm->sg by kmalloc in kfd_mem_dmamap_userptr,
but isn't freed by kfree in kfd_mem_dmaunmap_userptr. Free it!

Fixes: 264fb4d332f5 ("drm/amdgpu: Add multi-GPU DMA mapping helpers")

Signed-off-by: Lang Yu <lang.yu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: USB4 bring up set correct address

[Why]
YELLOW_CARP_B0 address was not correct

[How]
Set YELLOW_CARP_B0 to 0x1A.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Jude Shih <shenshih@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amd/display: Fix DCN3 B0 DP Alt Mapping

[Why]
DCN3 B0 has a mux, which redirects PHYC and PHYD to PHYF and PHYG.

[How]
Fix DIG mapping.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Zhan Liu <Zhan.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
(cherry picked from commit 4b7786d87fb3adf3e534c4f1e4f824d8700b786b)

drm/amd/display: Fix B0 USB-C DP Alt mode

[Why]
Starting from B0, along with RDPCSTX, RDPCSPIPE registers are also used.

[How]
Make sure RDPCSPIPE registers are programmed correctly.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Zhan Liu <Zhan.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
(cherry picked from commit bdd1a21b52557ea8f61d0a5dc2f77151b576eb70)

ARM: sharpsl_param: work around -Wstringop-overread warning

gcc warns that accessing a pointer based on a numeric constant may
be an offset into a NULL pointer, and would therefore has zero
accessible bytes:

arch/arm/common/sharpsl_param.c: In function ‘sharpsl_save_param’:
arch/arm/common/sharpsl_param.c:43:9: error: ‘memcpy’ reading 64 bytes from a region of size 0 [-Werror=stringop-overread]
   43 |         memcpy(&sharpsl_param, param_start(PARAM_BASE), sizeof(struct sharpsl_param_info));
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

In this particular case, the warning is bogus since this is the actual
pointer, not an offset on a NULL pointer. Add a local variable to shut
up the warning and hope it doesn't come back.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Daniel Mack <daniel@zonque.org>
Cc: Daniel Mack <daniel@zonque.org>
Cc: Haojian Zhuang <haojian.zhuang@gmail.com>
Cc: Robert Jarzmik <robert.jarzmik@free.fr>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Linus Walleij <linus.walleij@linaro.org>
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578
Link: https://lore.kernel.org/r/20210927145332.2784005-1-arnd@kernel.org'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'imx-fixes-5.15' of git://git./linux/kernel/git/shawnguo/linux into arm/fixes

i.MX fixes for 5.15:

- Revert cc8870bf4c3a to fix the regression on i.MX6 that suspend
  support becomes broken.
- Add `qca,clk-out-frequency` property to fix Ethernet support on
  imx6qdl-pico board.
- Re-enable FB support in imx_v6_v7_defconfig.  It gets lost due to
  f611b1e7624c ("drm: Avoid circular dependencies for CONFIG_FB").
- Fix LP5562 LED support on imx6dl-yapp4 board.
- Add missing pinctrl-names for panel on M53Menlo board.
- Fix USB host power regulator polarity on M53Menlo board.

* tag 'imx-fixes-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
  ARM: imx_v6_v7_defconfig: enable fb
  ARM: dts: imx6qdl-pico: Fix Ethernet support
  ARM: dts: imx: Fix USB host power regulator polarity on M53Menlo
  ARM: dts: imx: Add missing pinctrl-names for panel on M53Menlo
  Revert "ARM: imx6q: drop of_platform_default_populate() from init_machine"
  ARM: dts: imx6dl-yapp4: Fix lp5562 LED driver probe

Link: https://lore.kernel.org/r/20210923063356.GK13480@dragon
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

ARM: defconfig: gemini: Restore framebuffer

The framebuffer is gone on the D-Link DIR-685, restore it.

Fixes: f611b1e7624c ("drm: Avoid circular dependencies for CONFIG_FB")
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20210922200933.1825752-1-linus.walleij@linaro.org'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

ARM: dove: mark 'putc' as inline

This can cause a randconfig warning without the 'inline' flag
that every other platform uses:

In file included from arch/arm/boot/compressed/misc.c:29:
arch/arm/mach-dove/include/mach/uncompress.h:14:13: error: 'putc' defined but not used [-Werror=unused-function]
   14 | static void putc(const char c)
      |             ^~~~

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20210927095343.1015422-1-arnd@kernel.org'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

ARM: omap1: move omap15xx local bus handling to usb.c

Commit 38225f2ef2f4 ("ARM/omap1: switch to use dma_direct_set_offset for
lbus DMA offsets") removed a lot of mach/memory.h, but left the USB
offset handling split into arch/arm/mach-omap1/usb.c and
drivers/usb/host/ohci-omap.c.

This can cause a randconfig build warning that now fails the build
with -Werror:

arch/arm/mach-omap1/usb.c:561:30: error: 'omap_1510_usb_ohci_nb' defined but not used [-Werror=unused-variable]
  561 | static struct notifier_block omap_1510_usb_ohci_nb = {
      |                              ^~~~~~~~~~~~~~~~~~~~~

Move it all into the platform file to get rid of the final
location that relies on mach/memory.h.

Acked-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20210927144118.2464881-1-arnd@kernel.org'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'qcom-drivers-fixes-for-5.15' of git://git./linux/kernel/git/qcom/linux into arm/fixes

Qualcomm driver fixes for v5.15

This restricts the QCOM_SCM driver to depend on ARCH_QCOM, to reduce
it's presence after becoming a loadable module.

It then fixes a regression in the mdt_loader, where firmware with the
hash segment marked as PT_LOAD would no longer be accepted, preventing
several MSM8974 and SDM660 devices from loading remoteproc firmware.

Lastly it corrects the drvdata associated with the socinfo device during
probe, to match that expected by the remove function.

* tag 'qcom-drivers-fixes-for-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
  firmware: qcom_scm: QCOM_SCM should depend on ARCH_QCOM
  soc: qcom: mdt_loader: Drop PT_LOAD check on hash segment
  soc: qcom: socinfo: Fixed argument passed to platform_set_data()

Link: https://lore.kernel.org/r/20210930025456.1035-1-bjorn.andersson@linaro.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'qcom-arm64-fixes-for-5.15' of git://git./linux/kernel/git/qcom/linux into arm/fixes

Qualcomm ARM64 DTS fixes for 5.15

This starts by reverting the SC7280 CPUfreq update, which was merged
before concensus about the associated drivers changes was reached.

It then moves the reserved-memory changes done to get IPA working on the
Lenovo Yoga C630 into the Yoga specific DTS, as changing the memory map
on the platform level did break a couple of the other boards.

It fixes the HDMI audio on Trogdor and add missing Aggre2 NOC qos clocks
on SDM6{30,36,60} which prevented some boards from booting.

Lastly it enables the PON module on SM8250/QRB5165, as the lack thereof
is blocking automated testing in LKFT.

* tag 'qcom-arm64-fixes-for-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
  arm64: dts: qcom: sdm630: Add missing a2noc qos clocks
  arm64: dts: qcom: qrb5165-rb5: enabled pwrkey and resin nodes
  arm64: dts: qcom: pm8150: specify reboot mode magics
  arm64: dts: qcom: pm8150: use qcom,pm8998-pon binding
  arm64: dts: qcom: sc7180-trogdor: Fix lpass dai link for HDMI
  arm64: dts: qcom: sdm850-yoga: Reshuffle IPA memory mappings
  Revert "arm64: dts: qcom: sc7280: Fixup the cpufreq node"

Link: https://lore.kernel.org/r/20210930025509.1091-1-bjorn.andersson@linaro.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'juno-fixes-5.15' of git://git./linux/kernel/git/sudeep.holla/linux into arm/fixes

 Juno/Vexpress fixes for v5.15

Bunch of DTS fixes to resolve addressing issues with some of the device
nodes, dropping unused/undocumented properties in various nodes, and
aligning node names with dtschema.

* tag 'juno-fixes-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
  arm: dts: vexpress: Fix motherboard bus 'interrupt-map'
  arm: dts: vexpress: Fix addressing issues with 'motherboard-bus' nodes
  arm: dts: vexpress-v2p-ca9: Fix the SMB unit-address
  arm: dts: vexpress: Drop unused properties from motherboard node
  arm64: dts: arm: drop unused interrupt-names in MHU
  ARM: dts: arm: align watchdog and mmc node names with dtschema
  arm64: dts: arm: align watchdog and mmc node names with dtschema
  arm64: dts: fvp: Remove panel timings

Link: https://lore.kernel.org/r/20210927105249.3583380-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'qcom-dts-fixes-for-5.15' of git://git./linux/kernel/git/qcom/linux into arm/fixes

Qualcomm DTS fixes for v5.15

This corrects the use of depricated chipid and clock names, for which
support was finally dropped from the driver. It also ensures that the
DSI PLL is fed by the correct clock, now that it's being migrated to not
rely on global clock names.

* tag 'qcom-dts-fixes-for-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
  ARM: dts: qcom: apq8064: update Adreno clock names
  ARM: dts: qcom: apq8064: Use 27MHz PXO clock as DSI PLL reference
  ARM: dts: qcom: apq8064: use compatible which contains chipid

Link: https://lore.kernel.org/r/20210930025526.1146-1-bjorn.andersson@linaro.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'at91-fixes-5.15-2' of git://git./linux/kernel/git/at91/linux into arm/fixes

AT91 fixes #2 for 5.15:

- More fixes for AT91 platform power management code related to the
  introduction of sama7g5:
  - management of DDR3L regulator rails for sama7g5ek
  - loading of TLB on different cores

- PIO controller slew-rate settings for sama7g5ek: be aligned with
  datasheet requirements.

* tag 'at91-fixes-5.15-2' of git://git.kernel.org/pub/scm/linux/kernel/git/at91/linux:
  ARM: dts: at91: sama7g5ek: to not touch slew-rate for SDMMC pins
  ARM: dts: at91: sama7g5ek: use proper slew-rate settings for GMACs
  ARM: at91: pm: preload base address of controllers in tlb
  ARM: at91: pm: group constants and addresses loading
  ARM: dts: at91: sama7g5ek: add suspend voltage for ddr3l rail

Link: https://lore.kernel.org/r/20211004114344.19304-1-nicolas.ferre@microchip.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

MAINTAINERS: Add Vignesh to TI K3 platform maintainership

Add Vignesh to TI K3 platform maintainership. We will rotate roles for
each kernel version to help spread the work load a little better.

Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20210915121308.26795-1-nm@ti.com'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'optee-fix-for-v5.15' of git://git.linaro.org/people/jens.wiklander/linux-tee into arm/fixes

Fix OP-TEE shm_pool lint warning

* tag 'optee-fix-for-v5.15' of git://git.linaro.org/people/jens.wiklander/linux-tee:
  tee/optee/shm_pool: fix application of sizeof to pointer

Link: https://lore.kernel.org/r/20210915113813.GA509196@jade
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

netlink: annotate data races around nlk->bound

While existing code is correct, KCSAN is reporting
a data-race in netlink_insert / netlink_sendmsg [1]

It is correct to read nlk->bound without a lock, as netlink_autobind()
will acquire all needed locks.

[1]
BUG: KCSAN: data-race in netlink_insert / netlink_sendmsg

write to 0xffff8881031c8b30 of 1 bytes by task 18752 on cpu 0:
 netlink_insert+0x5cc/0x7f0 net/netlink/af_netlink.c:597
 netlink_autobind+0xa9/0x150 net/netlink/af_netlink.c:842
 netlink_sendmsg+0x479/0x7c0 net/netlink/af_netlink.c:1892
 sock_sendmsg_nosec net/socket.c:703 [inline]
 sock_sendmsg net/socket.c:723 [inline]
 ____sys_sendmsg+0x360/0x4d0 net/socket.c:2392
 ___sys_sendmsg net/socket.c:2446 [inline]
 __sys_sendmsg+0x1ed/0x270 net/socket.c:2475
 __do_sys_sendmsg net/socket.c:2484 [inline]
 __se_sys_sendmsg net/socket.c:2482 [inline]
 __x64_sys_sendmsg+0x42/0x50 net/socket.c:2482
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

read to 0xffff8881031c8b30 of 1 bytes by task 18751 on cpu 1:
 netlink_sendmsg+0x270/0x7c0 net/netlink/af_netlink.c:1891
 sock_sendmsg_nosec net/socket.c:703 [inline]
 sock_sendmsg net/socket.c:723 [inline]
 __sys_sendto+0x2a8/0x370 net/socket.c:2019
 __do_sys_sendto net/socket.c:2031 [inline]
 __se_sys_sendto net/socket.c:2027 [inline]
 __x64_sys_sendto+0x74/0x90 net/socket.c:2027
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

value changed: 0x00 -> 0x01

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 18751 Comm: syz-executor.0 Not tainted 5.14.0-rc1-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011

Fixes: da314c9923fe ("netlink: Replace rhash_portid with bound")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: pcs: xpcs: fix incorrect CL37 AN sequence

According to Synopsys DesignWare Cores Ethernet PCS databook, it is
required to disable Clause 37 auto-negotiation by programming bit-12
(AN_ENABLE) to 0 if it is already enabled, before programming various
fields of VR_MII_AN_CTRL registers.

After all these programming are done, it is then required to enable
Clause 37 auto-negotiation by programming bit-12 (AN_ENABLE) to 1.

Fixes: b97b5331b8ab ("net: pcs: add C37 SGMII AN support for intel mGbE controller")
Cc: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: sfp: Fix typo in state machine debug string

The string should be "tx_disable" to match the state enum.

Fixes: 4005a7cb4f55 ("net: phy: sftp: print debug message with text, not numbers")
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

USB: cdc-acm: fix break reporting

A recent change that started reporting break events forgot to push the
event to the line discipline, which meant that a detected break would
not be reported until further characters had been receive (the port
could even have been closed and reopened in between).

Fixes: 08dff274edda ("cdc-acm: fix BREAK rx code path adding necessary calls")
Cc: stable@vger.kernel.org
Acked-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210929090937.7410-3-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

USB: cdc-acm: fix racy tty buffer accesses

A recent change that started reporting break events to the line
discipline caused the tty-buffer insertions to no longer be serialised
by inserting events also from the completion handler for the interrupt
endpoint.

Completion calls for distinct endpoints are not guaranteed to be
serialised. For example, in case a host-controller driver uses
bottom-half completion, the interrupt and bulk-in completion handlers
can end up running in parallel on two CPUs (high-and low-prio tasklets,
respectively) thereby breaking the tty layer's single producer
assumption.

Fix this by holding the read lock also when inserting characters from
the bulk endpoint.

Fixes: 08dff274edda ("cdc-acm: fix BREAK rx code path adding necessary calls")
Cc: stable@vger.kernel.org
Acked-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210929090937.7410-2-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: gadget: f_uac2: fixed EP-IN wMaxPacketSize

Async feedback patches broke enumeration on Windows 10 previously fixed
by commit 789ea77310f0 ("usb: gadget: f_uac2: always increase endpoint
max_packet_size by one audio slot").

While the existing calculation for EP OUT capture for async mode yields
size+1 frame due to uac2_opts->fb_max > 0, playback side lost the +1
feature.  Therefore the +1 frame addition must be re-introduced for
playback. Win10 enumerates the device only when both EP IN and EP OUT
max packet sizes are (at least) +1 frame.

Fixes: e89bb4288378 ("usb: gadget: u_audio: add real feedback implementation")
Cc: stable <stable@vger.kernel.org>
Tested-by: Henrik Enquist <henrik.enquist@gmail.com>
Tested-by: Jack Pham <jackp@codeaurora.org>
Signed-off-by: Pavel Hofman <pavel.hofman@ivitera.com>
Link: https://lore.kernel.org/r/20210924080027.5362-1-pavel.hofman@ivitera.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: cdc-wdm: Fix check for WWAN

CONFIG_WWAN_CORE was with CONFIG_WWAN in commit 89212e160b81 ("net: wwan:
Fix WWAN config symbols"), but did not update all users of it. Change it
back to use CONFIG_WWAN instead.

Fixes: 89212e160b81 ("net: wwan: Fix WWAN config symbols")
Cc: <stable@vger.kernel.org>
Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Acked-by: Oliver Neukum <oneukum@suse.com>
Link: https://lore.kernel.org/r/20210929194547.46954-2-rikard.falkeborn@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: chipidea: ci_hdrc_imx: Also search for 'phys' phandle

When passing 'phys' in the devicetree to describe the USB PHY phandle
(which is the recommended way according to
Documentation/devicetree/bindings/usb/ci-hdrc-usb2.txt) the
following NULL pointer dereference is observed on i.MX7 and i.MX8MM:

[    1.489344] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000098
[    1.498170] Mem abort info:
[    1.500966]   ESR = 0x96000044
[    1.504030]   EC = 0x25: DABT (current EL), IL = 32 bits
[    1.509356]   SET = 0, FnV = 0
[    1.512416]   EA = 0, S1PTW = 0
[    1.515569]   FSC = 0x04: level 0 translation fault
[    1.520458] Data abort info:
[    1.523349]   ISV = 0, ISS = 0x00000044
[    1.527196]   CM = 0, WnR = 1
[    1.530176] [0000000000000098] user address but active_mm is swapper
[    1.536544] Internal error: Oops: 96000044 [#1] PREEMPT SMP
[    1.542125] Modules linked in:
[    1.545190] CPU: 3 PID: 7 Comm: kworker/u8:0 Not tainted 5.14.0-dirty #3
[    1.551901] Hardware name: Kontron i.MX8MM N801X S (DT)
[    1.557133] Workqueue: events_unbound deferred_probe_work_func
[    1.562984] pstate: 80000005 (Nzcv daif -PAN -UAO -TCO BTYPE=--)
[    1.568998] pc : imx7d_charger_detection+0x3f0/0x510
[    1.573973] lr : imx7d_charger_detection+0x22c/0x510

This happens because the charger functions check for the phy presence
inside the imx_usbmisc_data structure (data->usb_phy), but the chipidea
core populates the usb_phy passed via 'phys' inside 'struct ci_hdrc'
(ci->usb_phy) instead.

This causes the NULL pointer dereference inside imx7d_charger_detection().

Fix it by also searching for 'phys' in case 'fsl,usbphy' is not found.

Tested on a imx7s-warp board.

Fixes: 746f316b753a ("usb: chipidea: introduce imx7d USB charger detection")
Cc: stable@vger.kernel.org
Reported-by: Heiko Thiery <heiko.thiery@gmail.com>
Tested-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Reviewed-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Acked-by: Peter Chen <peter.chen@kernel.org>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Link: https://lore.kernel.org/r/20210921113754.767631-1-festevam@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: typec: tcpm: handle SRC_STARTUP state if cc changes

TCPM for DRP should do the same action as SRC_ATTACHED when cc changes in
SRC_STARTUP state. Otherwise, TCPM will transition to SRC_UNATTACHED state
which is not satisfied with the Type-C spec.

Per Type-C spec:
DRP port should move to Unattached.SNK instead of Unattached.SRC if sink
removed.

Fixes: 4b4e02c83167 ("typec: tcpm: Move out of staging")
cc: <stable@vger.kernel.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20210928111639.3854174-1-xu.yang_2@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: typec: tcpci: don't handle vSafe0V event if it's not enabled

USB TCPCI Spec, 4.4.3 Mask Registers:
"A masked register will still indicate in the ALERT register, but shall
not set the Alert# pin low."

Thus, the Extended Status will still indicate in ALERT register if vSafe0V
is detected by TCPC even though being masked. In current code, howerer,
this event will not be handled in detection time. Rather it will be
handled when next ALERT event coming(CC evnet, PD event, etc).

Tcpm might transition to a wrong state in this situation. Thus, the vSafe0V
event should not be handled when it's masked.

Fixes: 766c485b86ef ("usb: typec: tcpci: Add support to report vSafe0V")
cc: <stable@vger.kernel.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20210926101415.3775058-1-xu.yang_2@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: typec: tipd: Remove dependency on "connector" child fwnode

There is no "connector" child node available on every
platform, so the driver can't fail to probe when it's
missing.

Fixes: 57560ee95cb7 ("usb: typec: tipd: Don't block probing of consumer of "connector" nodes")
Cc: stable@vger.kernel.org # 5.14+
Reported-by: "Regupathy, Rajaram" <rajaram.regupathy@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20210930124758.23233-1-heikki.krogerus@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

net/sched: sch_taprio: properly cancel timer from taprio_destroy()

There is a comment in qdisc_create() about us not calling ops->reset()
in some cases.

err_out4:
/*
 * Any broken qdiscs that would require a ops->reset() here?
 * The qdisc was never in action so it shouldn't be necessary.
 */

As taprio sets a timer before actually receiving a packet, we need
to cancel it from ops->destroy, just in case ops->reset has not
been called.

syzbot reported:

ODEBUG: free active (active state 0) object type: hrtimer hint: advance_sched+0x0/0x9a0 arch/x86/include/asm/atomic64_64.h:22
WARNING: CPU: 0 PID: 8441 at lib/debugobjects.c:505 debug_print_object+0x16e/0x250 lib/debugobjects.c:505
Modules linked in:
CPU: 0 PID: 8441 Comm: syz-executor813 Not tainted 5.14.0-rc6-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:debug_print_object+0x16e/0x250 lib/debugobjects.c:505
Code: ff df 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 af 00 00 00 48 8b 14 dd e0 d3 e3 89 4c 89 ee 48 c7 c7 e0 c7 e3 89 e8 5b 86 11 05 <0f> 0b 83 05 85 03 92 09 01 48 83 c4 18 5b 5d 41 5c 41 5d 41 5e c3
RSP: 0018:ffffc9000130f330 EFLAGS: 00010282
RAX: 0000000000000000 RBX: 0000000000000003 RCX: 0000000000000000
RDX: ffff88802baeb880 RSI: ffffffff815d87b5 RDI: fffff52000261e58
RBP: 0000000000000001 R08: 0000000000000000 R09: 0000000000000000
R10: ffffffff815d25ee R11: 0000000000000000 R12: ffffffff898dd020
R13: ffffffff89e3ce20 R14: ffffffff81653630 R15: dffffc0000000000
FS:  0000000000f0d300(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007ffb64b3e000 CR3: 0000000036557000 CR4: 00000000001506e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 __debug_check_no_obj_freed lib/debugobjects.c:987 [inline]
 debug_check_no_obj_freed+0x301/0x420 lib/debugobjects.c:1018
 slab_free_hook mm/slub.c:1603 [inline]
 slab_free_freelist_hook+0x171/0x240 mm/slub.c:1653
 slab_free mm/slub.c:3213 [inline]
 kfree+0xe4/0x540 mm/slub.c:4267
 qdisc_create+0xbcf/0x1320 net/sched/sch_api.c:1299
 tc_modify_qdisc+0x4c8/0x1a60 net/sched/sch_api.c:1663
 rtnetlink_rcv_msg+0x413/0xb80 net/core/rtnetlink.c:5571
 netlink_rcv_skb+0x153/0x420 net/netlink/af_netlink.c:2504
 netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline]
 netlink_unicast+0x533/0x7d0 net/netlink/af_netlink.c:1340
 netlink_sendmsg+0x86d/0xdb0 net/netlink/af_netlink.c:1929
 sock_sendmsg_nosec net/socket.c:704 [inline]
 sock_sendmsg+0xcf/0x120 net/socket.c:724
 ____sys_sendmsg+0x6e8/0x810 net/socket.c:2403
 ___sys_sendmsg+0xf3/0x170 net/socket.c:2457
 __sys_sendmsg+0xe5/0x1b0 net/socket.c:2486
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80

Fixes: 44d4775ca518 ("net/sched: sch_taprio: reset child qdiscs before freeing them")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Davide Caratti <dcaratti@redhat.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Acked-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Partially revert "usb: Kconfig: using select for USB_COMMON dependency"

This reverts commit cb9c1cfc86926d0e86d19c8e34f6c23458cd3478 for
USB_LED_TRIG.  This config symbol has bool type and enables extra code
in usb_common itself, not a separate driver.  Enabling it should not
force usb_common to be built-in!

Fixes: cb9c1cfc8692 ("usb: Kconfig: using select for USB_COMMON dependency")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Salvatore Bonaccorso <carnil@debian.org>
Link: https://lore.kernel.org/r/20210921143442.340087-1-carnil@debian.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: dwc3: gadget: Revert "set gadgets parent to the right controller"

The commit c6e23b89a95d ("usb: dwc3: gadget: set gadgets parent to the right
controller") changed the device for the UDC and broke the user space scripts
that instantiate the USB gadget(s) via ConfigFS.

Revert it for now until the better solution will be proposed.

Fixes: c6e23b89a95d ("usb: dwc3: gadget: set gadgets parent to the right controller")
Tested-by: Ferry Toth <fntoth@gmail.com>
Cc: Michael Grzeschik <m.grzeschik@pengutronix.de>
Cc: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20211004141839.49079-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge branch 'bridge-fixes'

Eric Dumazet says:

====================
net: bridge: br_get_linkxstats_size() fixes

This patch series attempts to fix the following syzbot report.

WARNING: CPU: 1 PID: 21425 at net/core/rtnetlink.c:5388 rtnl_stats_get+0x80f/0x8c0 net/core/rtnetlink.c:5388
Modules linked in:
CPU: 1 PID: 21425 Comm: syz-executor394 Not tainted 5.13.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:rtnl_stats_get+0x80f/0x8c0 net/core/rtnetlink.c:5388
Code: e9 9c fc ff ff 4c 89 e7 89 0c 24 e8 ab 8b a8 fa 8b 0c 24 e9 bc fc ff ff 4c 89 e7 e8 9b 8b a8 fa e9 df fe ff ff e8 61 85 63 fa <0f> 0b e9 f7 fc ff ff 41 be ea ff ff ff e9 f9 fc ff ff 41 be 97 ff
RSP: 0018:ffffc9000cf77688 EFLAGS: 00010293
RAX: 0000000000000000 RBX: 000000000000012c RCX: 0000000000000000
RDX: ffff8880211754c0 RSI: ffffffff8711571f RDI: 0000000000000003
RBP: ffff8880175aa780 R08: 00000000ffffffa6 R09: ffff88823bd5c04f
R10: ffffffff87115413 R11: 0000000000000001 R12: ffff8880175aab74
R13: ffff8880175aab40 R14: 00000000ffffffa6 R15: 0000000000000006
FS:  0000000001ff9300(0000) GS:ffff8880b9c00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000005cfd58 CR3: 000000002cd43000 CR4: 00000000001506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 rtnetlink_rcv_msg+0x44e/0xad0 net/core/rtnetlink.c:5562
 netlink_rcv_skb+0x153/0x420 net/netlink/af_netlink.c:2504
 netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline]
 netlink_unicast+0x533/0x7d0 net/netlink/af_netlink.c:1340
 netlink_sendmsg+0x856/0xd90 net/netlink/af_netlink.c:1929
 sock_sendmsg_nosec net/socket.c:654 [inline]
 sock_sendmsg+0xcf/0x120 net/socket.c:674
 ____sys_sendmsg+0x6e8/0x810 net/socket.c:2350
 ___sys_sendmsg+0xf3/0x170 net/socket.c:2404
 __sys_sendmsg+0xe5/0x1b0 net/socket.c:2433
 do_syscall_64+0x3a/0xb0 arch/x86/entry/common.c:47
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x4440d9
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

net: bridge: fix under estimation in br_get_linkxstats_size()

Commit de1799667b00 ("net: bridge: add STP xstats")
added an additional nla_reserve_64bit() in br_fill_linkxstats(),
but forgot to update br_get_linkxstats_size() accordingly.

This can trigger the following in rtnl_stats_get()

WARN_ON(err == -EMSGSIZE);

Fixes: de1799667b00 ("net: bridge: add STP xstats")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Nikolay Aleksandrov <nikolay@nvidia.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: bridge: use nla_total_size_64bit() in br_get_linkxstats_size()

bridge_fill_linkxstats() is using nla_reserve_64bit().

We must use nla_total_size_64bit() instead of nla_total_size()
for corresponding data structure.

Fixes: 1080ab95e3c7 ("net: bridge: add support for IGMP/MLD stats and export them via netlink")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Nikolay Aleksandrov <nikolay@nvidia.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

usb: xhci: tegra: mark PM functions as __maybe_unused

The added #ifdefs in the PM rework were almost correct, but still
cause warnings in some randconfig builds:

drivers/usb/host/xhci-tegra.c:2147:12: error: 'tegra_xusb_resume' defined but not used [-Werror=unused-function]
 2147 | static int tegra_xusb_resume(struct device *dev)
      |            ^~~~~~~~~~~~~~~~~
drivers/usb/host/xhci-tegra.c:2105:12: error: 'tegra_xusb_suspend' defined but not used [-Werror=unused-function]
 2105 | static int tegra_xusb_suspend(struct device *dev)

Replace the #ifdef checks with simpler __maybe_unused annotations to
reliably shut up these warnings.

Link: https://lore.kernel.org/all/20210421135613.3560777-2-arnd@kernel.org/
Fixes: 971ee247060d ("usb: xhci: tegra: Enable ELPG for runtime/system PM")
Reviewed-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20211005112057.2700888-1-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

r8152: avoid to resubmit rx immediately

For the situation that the disconnect event comes very late when the
device is unplugged, the driver would resubmit the RX bulk transfer
after getting the callback with -EPROTO immediately and continually.
Finally, soft lockup occurs.

This patch avoids to resubmit RX immediately. It uses a workqueue to
schedule the RX NAPI. And the NAPI would resubmit the RX. It let the
disconnect event have opportunity to stop the submission before soft
lockup.

Reported-by: Jason-ch Chen <jason-ch.chen@mediatek.com>
Tested-by: Jason-ch Chen <jason-ch.chen@mediatek.com>
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

etherdevice: use __dev_addr_set()

Andrew points out that eth_hw_addr_set() replaces memcpy()
calls so we can't use ether_addr_copy() which assumes
both arguments are 2-bytes aligned.

Reported-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

afs: Fix afs_launder_page() to set correct start file position

Fix afs_launder_page() to set the starting position of the StoreData RPC at
the offset into the page at which the modified data starts instead of at
the beginning of the page (the iov_iter is correctly offset).

The offset got lost during the conversion to passing an iov_iter into
afs_store_data().

Changes:
ver #2:
 - Use page_offset() rather than manually calculating it[1].

Fixes: bd80d8a80e12 ("afs: Use ITER_XARRAY for writing")
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeffrey Altman <jaltman@auristor.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/YST/0e92OdSH0zjg@casper.infradead.org/
Link: https://lore.kernel.org/r/162880783179.3421678.7795105718190440134.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/162937512409.1449272.18441473411207824084.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/162981148752.1901565.3663780601682206026.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/163005741670.2472992.2073548908229887941.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/163221839087.3143591.14278359695763025231.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/163292980654.4004896.7134735179887998551.stgit@warthog.procyon.org.uk/

netfs: Fix READ/WRITE confusion when calling iov_iter_xarray()

Fix netfs_clear_unread() to pass READ to iov_iter_xarray() instead of WRITE
(the flag is about the operation accessing the buffer, not what sort of
access it is doing to the buffer).

Fixes: 3d3c95046742 ("netfs: Provide readahead and readpage netfs helpers")
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
cc: linux-cachefs@redhat.com
cc: linux-afs@lists.infradead.org
cc: ceph-devel@vger.kernel.org
cc: linux-cifs@vger.kernel.org
cc: linux-nfs@vger.kernel.org
cc: v9fs-developer@lists.sourceforge.net
cc: linux-fsdevel@vger.kernel.org
cc: linux-mm@kvack.org
Link: https://lore.kernel.org/r/162729351325.813557.9242842205308443901.stgit@warthog.procyon.org.uk/
Link: https://lore.kernel.org/r/162886603464.3940407.3790841170414793899.stgit@warthog.procyon.org.uk
Link: https://lore.kernel.org/r/163239074602.1243337.14154704004485867017.stgit@warthog.procyon.org.uk

cachefiles: Fix oops with cachefiles_cull() due to NULL object

When cachefiles_cull() calls cachefiles_bury_object(), it passes
a NULL object.  When this occurs, either trace_cachefiles_unlink()
or trace_cachefiles_rename() may oops due to the NULL object.
Check for NULL object in the tracepoint and if so, set debug_id
to MAX_UINT as was done in 2908f5e101e3.

The following oops was seen with xfstests generic/100.
BUG: kernel NULL pointer dereference, address: 0000000000000010
...
RIP: 0010:trace_event_raw_event_cachefiles_unlink+0x4e/0xa0 [cachefiles]
...
 Call Trace:
   cachefiles_bury_object+0x242/0x430 [cachefiles]
   ? __vfs_removexattr_locked+0x10f/0x150
   ? vfs_removexattr+0x51/0xd0
   cachefiles_cull+0x84/0x120 [cachefiles]
   cachefiles_daemon_cull+0xd1/0x120 [cachefiles]
   cachefiles_daemon_write+0x158/0x190 [cachefiles]
   vfs_write+0xbc/0x260
   ksys_write+0x4f/0xc0
   do_syscall_64+0x3b/0x90

The following oops was seen with xfstests generic/290.
BUG: kernel NULL pointer dereference, address: 0000000000000010
...
RIP: 0010:trace_event_raw_event_cachefiles_rename+0x54/0xa0 [cachefiles]
...
Call Trace:
  cachefiles_bury_object+0x35c/0x430 [cachefiles]
  cachefiles_cull+0x84/0x120 [cachefiles]
  cachefiles_daemon_cull+0xd1/0x120 [cachefiles]
  cachefiles_daemon_write+0x158/0x190 [cachefiles]
  vfs_write+0xbc/0x260
  ksys_write+0x4f/0xc0
  do_syscall_64+0x3b/0x90

Fixes: 2908f5e101e3 ("fscache: Add a cookie debug ID and use that in traces")
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://listman.redhat.com/archives/linux-cachefs/2021-October/msg00009.html

drm/i915: Extend the async flip VT-d w/a to skl/bxt

Looks like skl/bxt/derivatives also need the plane stride
stretch w/a when using async flips and VT-d is enabled, or
else we get corruption on screen. To my surprise this was
even documented in bspec, but only as a note on the
CHICHKEN_PIPESL register description rather than on the
w/a list.

So very much the same thing as on HSW/BDW, except the bits
moved yet again.

Cc: stable@vger.kernel.org
Cc: Karthik B S <karthik.b.s@intel.com>
Fixes: 55ea1cb178ef ("drm/i915: Enable async flips in i915")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210930190943.17547-1-ville.syrjala@linux.intel.com
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
(cherry picked from commit d08df3b0bdb25546e86dc9a6c4e3ec0c43832299)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/bdb: Fix version check

With patch "drm/i915/vbt: Fix backlight parsing for VBT 234+"
the size of bdb_lfp_backlight_data structure has been increased,
causing if-statement in the parse_lfp_backlight function
that comapres this structure size to the one retrieved from BDB,
always to fail for older revisions.
This patch calculates expected size of the structure for a given
BDB version and compares it with the value gathered from BDB.
Tested on Chromebook Pixelbook (Nocturne) (reports bdb->version = 221)

Fixes: d381baad29b4 ("drm/i915/vbt: Fix backlight parsing for VBT 234+")

Tested-by: Lukasz Majczak <lma@semihalf.com>
Signed-off-by: Lukasz Majczak <lma@semihalf.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210930134606.227234-1-lma@semihalf.com
(cherry picked from commit 4378daf5d04eed59724e6d0e74755e17dce2e105)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/tc: Fix TypeC port init/resume time sanitization

Atm during driver loading and system resume TypeC ports are accessed
before their HW/SW state is synced. Move the TypeC port sanitization to
the encoder's sync_state hook to fix this.

v2: Handle the encoder disabled case in gen11_dsi_sync_state() as well
    (Jose, Jani)

Fixes: f9e76a6e68d3 ("drm/i915: Add an encoder hook to sanitize its state during init/resume")
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210929132833.2253961-1-imre.deak@intel.com
(cherry picked from commit 7194dc998dfffca096c30b3cd39625158608992d)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Fix runtime pm handling in i915_gem_shrink

We forgot to call intel_runtime_pm_put on error, fix it!

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Fixes: cf41a8f1dc1e ("drm/i915: Finally remove obj->mm.lock.")
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: <stable@vger.kernel.org> # v5.13+
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210830121006.2978297-9-maarten.lankhorst@linux.intel.com
(cherry picked from commit 239f3c2ee18376587026efecaea5250fa5926d20)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/audio: Use BIOS provided value for RKL HDA link

Commit 989634fb49ad ("drm/i915/audio: set HDA link parameters in
driver") makes HDMI audio on Lenovo P350 disappear.

So in addition to TGL, extend the logic to RKL to use BIOS provided
value to fix the regression.

Fixes: 989634fb49ad ("drm/i915/audio: set HDA link parameters in driver")
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210906041300.508458-1-kai.heng.feng@canonical.com
(cherry picked from commit c6b40ee330fe09b332715bb7ec1467e4fcbe2e65)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

xen/x86: adjust data placement

Both xen_pvh and xen_start_flags get written just once early during
init. Using the respective annotation then allows the open-coded placing
in .data to go away.

Additionally the former, like the latter, wants exporting, or else
xen_pvh_domain() can't be used from modules.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/8155ed26-5a1d-c06f-42d8-596d26e75849@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

x86/PVH: adjust function/data placement

Two of the variables can live in .init.data, allowing the open-coded
placing in .data to go away. Another "variable" is used to communicate a
size value only to very early assembly code, which hence can be both
const and live in .init.*. Additionally two functions were lacking
__init annotations.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/3b0bb22e-43f4-e459-c5cb-169f996b5669@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: hook up xen_banner() also for PVH

This was effectively lost while dropping PVHv1 code. Move the function
and arrange for it to be called the same way as done in PV mode. Clearly
this then needs re-introducing the XENFEAT_mmu_pt_update_preserve_ad
check that was recently removed, as that's a PV-only feature.

Since the string pointed at by pv_info.name describes the mode, drop
"paravirtualized" from the log message while moving the code.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/de03054d-a20d-2114-bb86-eec28e17b3b8@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: generalize preferred console model from PV to PVH Dom0

Without announcing hvc0 as preferred it won't get used as long as tty0
gets registered earlier. This is particularly problematic with there not
being any screen output for PVH Dom0 when the screen is in graphics
mode, as the necessary information doesn't get conveyed yet from the
hypervisor.

Follow PV's model, but be conservative and do this for Dom0 only for
now.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/582328b6-c86c-37f3-d802-5539b7a86736@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: make "earlyprintk=xen" work for HVM/PVH DomU

xenboot_write_console() is dealing with these quite fine so I don't see
why xenboot_console_setup() would return -ENOENT in this case.

Adjust documentation accordingly.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/3d212583-700e-8b2d-727a-845ef33ac265@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: allow "earlyprintk=xen" to work for PV Dom0

With preferred consoles "tty" and "hvc" announced as preferred,
registering "xenboot" early won't result in use of the console: It also
needs to be registered as preferred. Generalize this from being DomU-
only so far.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/d4a34540-a476-df2c-bca6-732d0d58c5f0@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: make "earlyprintk=xen" work better for PVH Dom0

The xen_hvm_early_write() path better wouldn't be taken in this case;
while port 0xE9 can be used, the hypercall path is quite a bit more
efficient. Put that first, as it may also work for DomU-s (see also
xen_raw_console_write()).

While there also bail from the function when the first
domU_write_console() failed - later ones aren't going to succeed.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/4fd89dcb-cfc5-c740-2e94-bb271e432d3e@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: allow PVH Dom0 without XEN_PV=y

Decouple XEN_DOM0 from XEN_PV, converting some existing uses of XEN_DOM0
to a new XEN_PV_DOM0. (I'm not convinced all are really / should really
be PV-specific, but for starters I've tried to be conservative.)

For PVH Dom0 the hypervisor populates MADT with only x2APIC entries, so
without x2APIC support enabled in the kernel things aren't going to work
very well. (As opposed, DomU-s would only ever see LAPIC entries in MADT
as of now.) Note that this then requires PVH Dom0 to be 64-bit, as
X86_X2APIC depends on X86_64.

In the course of this xen_running_on_version_or_later() needs to be
available more broadly. Move it from a PV-specific to a generic file,
considering that what it does isn't really PV-specific at all anyway.

Note that xen/interface/version.h cannot be included on its own; in
enlighten.c, which uses SCHEDOP_* anyway, include xen/interface/sched.h
first to resolve the apparently sole missing type (xen_ulong_t).

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/983bb72f-53df-b6af-14bd-5e088bd06a08@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/x86: prevent PVH type from getting clobbered

Like xen_start_flags, xen_domain_type gets set before .bss gets cleared.
Hence this variable also needs to be prevented from getting put in .bss,
which is possible because XEN_NATIVE is an enumerator evaluating to
zero. Any use prior to init_hvm_pv_info() setting the variable again
would lead to wrong decisions; one such case is xenboot_console_setup()
when called as a result of "earlyprintk=xen".

Use __ro_after_init as more applicable than either __section(".data") or
__read_mostly.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/d301677b-6f22-5ae6-bd36-458e1f323d0b@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

arm64: dts: imx8m*-venice-gw7902: fix M2_RST# gpio

Fix invalid M2_RST# gpio pinmux.

Fixes: ef484dfcf6f7 ("arm64: dts: imx: Add i.mx8mm/imx8mn Gateworks gw7902 dts support")
Cc: stable@vger.kernel.org
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>

xen/privcmd: drop "pages" parameter from xen_remap_pfn()

The function doesn't use it and all of its callers say in a comment that
their respective arguments are to be non-NULL only in auto-translated
mode. Since xen_remap_domain_mfn_array() isn't supposed to be used by
non-PV, drop the parameter there as well. It was bogusly passed as non-
NULL (PRIV_VMA_LOCKED) by its only caller anyway. For
xen_remap_domain_gfn_range(), otoh, it's not clear at all why this
wouldn't want / might not need to gain auto-translated support down the
road, so the parameter is retained there despite now remaining unused
(and the only caller passing NULL); correct a respective comment as
well.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/036ad8a2-46f9-ac3d-6219-bdc93ab9e10b@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/privcmd: fix error handling in mmap-resource processing

xen_pfn_t is the same size as int only on 32-bit builds (and not even
on Arm32). Hence pfns[] can't be used directly to read individual error
values returned from xen_remap_domain_mfn_array(); every other error
indicator would be skipped/ignored on 64-bit.

Fixes: 3ad0876554ca ("xen/privcmd: add IOCTL_PRIVCMD_MMAP_RESOURCE")
Cc: stable@vger.kernel.org
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/aa6d6a67-6889-338a-a910-51e889f792d5@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/privcmd: replace kcalloc() by kvcalloc() when allocating empty pages

Osstest has been suffering test failures for a little while from order-4
allocation failures, resulting from alloc_empty_pages() calling
kcalloc(). As there's no need for physically contiguous space here,
switch to kvcalloc().

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/6d698901-98a4-05be-c421-bcd0713f5335@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

ARM: imx6: disable the GIC CPU interface before calling stby-poweroff sequence

Any pending interrupt can prevent entering standby based power off state.
To avoid it, disable the GIC CPU interface.

Fixes: 8148d2136002 ("ARM: imx6: register pm_power_off handler if "fsl,pmic-stby-poweroff" is set")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>

scsi: iscsi: Fix iscsi_task use after free

Commit d39df158518c ("scsi: iscsi: Have abort handler get ref to conn")
added iscsi_get_conn()/iscsi_put_conn() calls during abort handling but
then also changed the handling of the case where we detect an already
completed task where we now end up doing a goto to the common put/cleanup
code. This results in a iscsi_task use after free, because the common
cleanup code will do a put on the iscsi_task.

This reverts the goto and moves the iscsi_get_conn() to after we've checked
if the iscsi_task is valid.

Link: https://lore.kernel.org/r/20211004210608.9962-1-michael.christie@oracle.com
Fixes: d39df158518c ("scsi: iscsi: Have abort handler get ref to conn")
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix memory overwrite during FC-GS I/O abort handling

When an FC-GS I/O is aborted by lpfc, the driver requires a node pointer
for a dereference operation.  In the abort I/O routine, the driver miscasts
a context pointer to the wrong data type and overwrites a single byte
outside of the allocated space.  This miscast is done in the abort I/O
function handler because the handler works on both FC-GS and FC-LS
commands. However, the code neglected to get the correct job location for
the node.

Fix this by acquiring the necessary node pointer from the correct job
structure depending on the I/O type.

Link: https://lore.kernel.org/r/20211004231210.35524-1-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: elx: efct: Delete stray unlock statement

It's not holding the lock at this stage and the IRQ "flags" are not correct
so it would restore something bogus. Delete the unlock statement.

Link: https://lore.kernel.org/r/20211004103851.GE25015@kili
Fixes: 3e6414003bf9 ("scsi: elx: efct: SCSI I/O handling routines")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Fix task management completion

The UFS driver uses blk_mq_tagset_busy_iter() when identifying task
management requests to complete, however blk_mq_tagset_busy_iter() doesn't
work.

blk_mq_tagset_busy_iter() only iterates requests dispatched by the block
layer. That appears as if it might have started since commit 37f4a24c2469
("blk-mq: centralise related handling into blk_mq_get_driver_tag") which
removed 'data->hctx->tags->rqs[rq->tag] = rq' from blk_mq_rq_ctx_init()
which gets called:

blk_get_request
blk_mq_alloc_request
__blk_mq_alloc_request
blk_mq_rq_ctx_init

Since UFS task management requests are not dispatched by the block layer,
hctx->tags->rqs[rq->tag] remains NULL, and since blk_mq_tagset_busy_iter()
relies on finding requests using hctx->tags->rqs[rq->tag], UFS task
management requests are never found by blk_mq_tagset_busy_iter().

By using blk_mq_tagset_busy_iter(), the UFS driver was relying on internal
details of the block layer, which was fragile and subsequently got
broken. Fix by removing the use of blk_mq_tagset_busy_iter() and having the
driver keep track of task management requests.

Link: https://lore.kernel.org/r/20210922091059.4040-1-adrian.hunter@intel.com
Fixes: 1235fc569e0b ("scsi: ufs: core: Fix task management request completion timeout")
Fixes: 69a6c269c097 ("scsi: ufs: Use blk_{get,put}_request() to allocate and free TMFs")
Cc: stable@vger.kernel.org
Tested-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: acornscsi: Remove scsi_cmd_to_tag() reference

Commit 756fb6a895af ("scsi: acornscsi: Remove tagged queuing vestiges")
mistakenly introduced a reference to function scsi_cmd_to_tag(). This
function does not exist as it was removed from an earlier series version
when I upstreamed the named commit - originally authored By Hannes - but
this reference still remained.

Fix by replacing the reference to scsi_cmd_to_tag() with
scsi_cmd_to_rq(scsi_scmd)->tag, which scsi_cmd_to_tag() was a wrapper for.

Link: https://lore.kernel.org/r/1633002717-79765-1-git-send-email-john.garry@huawei.com
Fixes: 756fb6a895af ("scsi: acornscsi: Remove tagged queuing vestiges")
Reported-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

riscv: Flush current cpu icache before other cpus

On SiFive Unmatched, I recently fell onto the following BUG when booting:

[    0.000000] ftrace: allocating 36610 entries in 144 pages
[    0.000000] Oops - illegal instruction [#1]
[    0.000000] Modules linked in:
[    0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 5.13.1+ #5
[    0.000000] Hardware name: SiFive HiFive Unmatched A00 (DT)
[    0.000000] epc : riscv_cpuid_to_hartid_mask+0x6/0xae
[    0.000000]  ra : __sbi_rfence_v02+0xc8/0x10a
[    0.000000] epc : ffffffff80007240 ra : ffffffff80009964 sp : ffffffff81803e10
[    0.000000]  gp : ffffffff81a1ea70 tp : ffffffff8180f500 t0 : ffffffe07fe30000
[    0.000000]  t1 : 0000000000000004 t2 : 0000000000000000 s0 : ffffffff81803e60
[    0.000000]  s1 : 0000000000000000 a0 : ffffffff81a22238 a1 : ffffffff81803e10
[    0.000000]  a2 : 0000000000000000 a3 : 0000000000000000 a4 : 0000000000000000
[    0.000000]  a5 : 0000000000000000 a6 : ffffffff8000989c a7 : 0000000052464e43
[    0.000000]  s2 : ffffffff81a220c8 s3 : 0000000000000000 s4 : 0000000000000000
[    0.000000]  s5 : 0000000000000000 s6 : 0000000200000100 s7 : 0000000000000001
[    0.000000]  s8 : ffffffe07fe04040 s9 : ffffffff81a22c80 s10: 0000000000001000
[    0.000000]  s11: 0000000000000004 t3 : 0000000000000001 t4 : 0000000000000008
[    0.000000]  t5 : ffffffcf04000808 t6 : ffffffe3ffddf188
[    0.000000] status: 0000000200000100 badaddr: 0000000000000000 cause: 0000000000000002
[    0.000000] [<ffffffff80007240>] riscv_cpuid_to_hartid_mask+0x6/0xae
[    0.000000] [<ffffffff80009474>] sbi_remote_fence_i+0x1e/0x26
[    0.000000] [<ffffffff8000b8f4>] flush_icache_all+0x12/0x1a
[    0.000000] [<ffffffff8000666c>] patch_text_nosync+0x26/0x32
[    0.000000] [<ffffffff8000884e>] ftrace_init_nop+0x52/0x8c
[    0.000000] [<ffffffff800f051e>] ftrace_process_locs.isra.0+0x29c/0x360
[    0.000000] [<ffffffff80a0e3c6>] ftrace_init+0x80/0x130
[    0.000000] [<ffffffff80a00f8c>] start_kernel+0x5c4/0x8f6
[    0.000000] ---[ end trace f67eb9af4d8d492b ]---
[    0.000000] Kernel panic - not syncing: Attempted to kill the idle task!
[    0.000000] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---

While ftrace is looping over a list of addresses to patch, it always failed
when patching the same function: riscv_cpuid_to_hartid_mask. Looking at the
backtrace, the illegal instruction is encountered in this same function.
However, patch_text_nosync, after patching the instructions, calls
flush_icache_range. But looking at what happens in this function:

flush_icache_range -> flush_icache_all
                   -> sbi_remote_fence_i
                   -> __sbi_rfence_v02
                   -> riscv_cpuid_to_hartid_mask

The icache and dcache of the current cpu are never synchronized between the
patching of riscv_cpuid_to_hartid_mask and calling this same function.

So fix this by flushing the current cpu's icache before asking for the other
cpus to do the same.

Signed-off-by: Alexandre Ghiti <alex@ghiti.fr>
Fixes: fab957c11efe ("RISC-V: Atomic and Locking Code")
Cc: stable@vger.kernel.org
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>