Pull EFI fixes from Ingo Molnar:
 "Various fixes all over the map: prevent boot crashes on HyperV,
  classify UEFI randomness as bootloader randomness, fix EFI boot for
  the Raspberry Pi2, fix efi_test permissions, etc"

* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi/efi_test: Lock down /dev/efi_test and require CAP_SYS_ADMIN
  x86, efi: Never relocate kernel below lowest acceptable address
  efi: libstub/arm: Account for firmware reserved memory at the base of RAM
  efi/random: Treat EFI_RNG_PROTOCOL output as bootloader randomness
  efi/tpm: Return -EINVAL when determining tpm final events log size fails
  efi: Make CONFIG_EFI_RCI2_TABLE selectable on x86 only

Pull arm64 fixes from Will Deacon:
 "These are almost exclusively related to CPU errata in CPUs from
  Broadcom and Qualcomm where the workarounds were either not being
  enabled when they should have been or enabled when they shouldn't have
  been.

  The only "interesting" fix is ensuring that writeable, shared mappings
  are initially mapped as clean since we inadvertently broke the logic
  back in v4.14 and then noticed the problem via code inspection the
  other day.

  The only critical issue we have outstanding is a sporadic NULL
  dereference in the scheduler, which doesn't appear to be
  arm64-specific and PeterZ is tearing his hair out over it at the
  moment.

  Summary:

   - Enable CPU errata workarounds for Broadcom Brahma-B53

   - Enable CPU errata workarounds for Qualcomm Hydra/Kryo CPUs

   - Fix initial dirty status of writeable, shared mappings"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: apply ARM64_ERRATUM_843419 workaround for Brahma-B53 core
  arm64: Brahma-B53 is SSB and spectre v2 safe
  arm64: apply ARM64_ERRATUM_845719 workaround for Brahma-B53 core
  arm64: cpufeature: Enable Qualcomm Falkor errata 1009 for Kryo
  arm64: cpufeature: Enable Qualcomm Falkor/Kryo errata 1003
  arm64: Ensure VM_WRITE|VM_SHARED ptes are clean by default

Pull kvm fixes from Paolo Bonzini:
 "generic:
   - fix memory leak on failure to create VM

  x86:
   - fix MMU corner case with AMD nested paging disabled"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: vmx, svm: always run with EFER.NXE=1 when shadow paging is active
  kvm: call kvm_arch_destroy_vm if vm creation fails
  kvm: Allocate memslots and buses before calling kvm_arch_init_vm

Pull drm fixes from Dave Airlie:
 "This is the regular drm fixes pull request for 5.4-rc6. It's a bit
  larger than I'd like but then last week was quieter than usual.

  The main fixes are amdgpu, and the two bigger area are navi fixes
  which are the newest GPU range so still getting actively fixed up, but
  also a bunch of clang stack alignment fixes (as amdgpu uses double in
  some places).

  Otherwise it's all fairly run of the mill fixes, i915, panfrost,
  etnaviv, v3d and radeon, along with a core scheduler fix.

  Summary:

  amdgpu:
   - clang alignment fixes
   - Updated golden settings
   - navi: gpuvm, sdma and display fixes
   - Freesync fix
   - Gamma fix for DCN
   - DP dongle detection fix
   - vega10: Fix for undervolting

  radeon:
   - reenable kexec fix for ppc

  scheduler:
   - set an error if hw job failed

  i915:
   - fix PCH reference clock for HSW/BDW
   - TGL display PLL doc fix

  panfrost:
   - warning fix
   - runtime pm fix
   - bad pointer dereference fix

  v3d:
   - memleak fix

  etnaviv:
   - memory corruption fix
   - deadlock fix
   - reintroduce lost debug message"

* tag 'drm-fixes-2019-11-01' of git://anongit.freedesktop.org/drm/drm: (29 commits)
  drm/amdgpu: enable -msse2 for GCC 7.1+ users
  drm/amdgpu: fix stack alignment ABI mismatch for GCC 7.1+
  drm/amdgpu: fix stack alignment ABI mismatch for Clang
  drm/radeon: Fix EEH during kexec
  drm/amdgpu/gmc10: properly set BANK_SELECT and FRAGMENT_SIZE
  drm/amdgpu/powerplay/vega10: allow undervolting in p7
  dc.c:use kzalloc without test
  drm/amd/display: setting the DIG_MODE to the correct value.
  drm/amd/display: Passive DP->HDMI dongle detection fix
  drm/amd/display: add 50us buffer as WA for pstate switch in active
  drm/amd/display: Allow inverted gamma
  drm/amd/display: do not synchronize "drr" displays
  drm/amdgpu: If amdgpu_ib_schedule fails return back the error.
  drm/sched: Set error to s_fence if HW job submission failed.
  drm/amdgpu/gfx10: update gfx golden settings for navi12
  drm/amdgpu/gfx10: update gfx golden settings for navi14
  drm/amdgpu/gfx10: update gfx golden settings
  drm/amd/display: Change Navi14's DWB flag to 1
  drm/amdgpu/sdma5: do not execute 0-sized IBs (v2)
  drm/amdgpu: Fix SDMA hang when performing VKexample test
  ...

Pull power management fix from Rafael Wysocki:
 "Fix a recently introduced (mostly theoretical) issue that the requests
  to confine the maximum CPU frequency coming from the platform firmware
  may not be taken into account if multiple CPUs are covered by one
  cpufreq policy on a system with ACPI"

* tag 'pm-5.4-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: processor: Add QoS requests for all CPUs

Pull rdma fixes from Jason Gunthorpe:
 "A number of bug fixes and a regression fix:

   - Various issues from static analysis in hfi1, uverbs, hns, and cxgb4

   - Fix for deadlock in a case when the new auto RDMA module loading is
     used

   - Missing _irq notation in a prior -rc patch found by lockdep

   - Fix a locking and lifetime issue in siw

   - Minor functional bug fixes in cxgb4, mlx5, qedr

   - Fix a regression where vlan interfaces no longer worked with RDMA
     CM in some cases"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/hns: Prevent memory leaks of eq->buf_list
  RDMA/iw_cxgb4: Avoid freeing skb twice in arp failure case
  RDMA/mlx5: Use irq xarray locking for mkey_table
  IB/core: Avoid deadlock during netlink message handling
  RDMA/nldev: Skip counter if port doesn't match
  RDMA/uverbs: Prevent potential underflow
  IB/core: Use rdma_read_gid_l2_fields to compare GID L2 fields
  RDMA/qedr: Fix reported firmware version
  RDMA/siw: free siw_base_qp in kref release routine
  RDMA/iwcm: move iw_rem_ref() calls out of spinlock
  iw_cxgb4: fix ECN check on the passive accept
  IB/hfi1: Use a common pad buffer for 9B and 16B packets
  IB/hfi1: Avoid excessive retry for TID RDMA READ request
  RDMA/mlx5: Clear old rate limit when closing QP

Pull sound fixes from Takashi Iwai:
 "A couple of regression fixes and a fix for mutex deadlock at
  hog-unplug, as well as other device-specific fixes:

   - A commit to avoid the spurious unsolicited interrupt on HD-audio
     bus caused a stall at shutdown, so it's reverted now.

   - The recent support of AMD/Nvidia audio component binding caused a
     mutex deadlock; fixed by splitting to another mutex

   - The device hot-unplug and the ALSA timer close combo may lead to
     another mutex deadlock; fixed by moving put_device() calls

   - Usual device-specific small quirks for HD- and USB-audio drivers

   - An old error check fix in FireWire driver"

* tag 'sound-5.4-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: timer: Fix mutex deadlock at releasing card
  ALSA: hda - Fix mutex deadlock in HDMI codec driver
  Revert "ALSA: hda: Flush interrupts on disabling"
  ALSA: bebob: Fix prototype of helper function to return negative value
  ALSA: hda/realtek - Fix 2 front mics of codec 0x623
  ALSA: hda/realtek - Add support for ALC623
  ALSA: usb-audio: Add DSD support for Gustard U16/X26 USB Interface

The Broadcom Brahma-B53 core is susceptible to the issue described by
ARM64_ERRATUM_843419 so this commit enables the workaround to be applied
when executing on that core.

Since there are now multiple entries to match, we must convert the
existing ARM64_ERRATUM_843419 into an erratum list and use
cpucap_multi_entry_cap_matches to match our entries.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Will Deacon <will@kernel.org>

Add the Brahma-B53 CPU (all versions) to the whitelists of CPUs for the
SSB and spectre v2 mitigations.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Will Deacon <will@kernel.org>

The Broadcom Brahma-B53 core is susceptible to the issue described by
ARM64_ERRATUM_845719 so this commit enables the workaround to be applied
when executing on that core.

Since there are now multiple entries to match, we must convert the
existing ARM64_ERRATUM_845719 into an erratum list.
Signed-off-by: Doug Berger <opendmb@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Will Deacon <will@kernel.org>

drm-fixes-5.4-2019-10-30:

amdgpu:
- clang fixes
- Updated golden settings
- GPUVM fixes for navi
- Navi sdma fix
- Navi display fixes
- Freesync fix
- Gamma fix for DCN
- DP dongle detection fix
- Fix for undervolting on vega10

radeon:
- enable kexec fix for PPC

scheduler:
- set an error on fence if hw job failed
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191030162339.44366-1-alexander.deucher@amd.com

Merge tag 'drm-intel-fixes-2019-10-31' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

- Fix PCH reference clock for FDI on HSW/BDW which was causing users blank screen
- Small documentation fix for TGL display PLLs
Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191031171209.GA6586@intel.com

Merge tag 'drm-misc-fixes-2019-10-30-1' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

 - three fixes for panfrost, one to silence a warning, one to fix
   runtime_pm and one to prevent bogus pointer dereferences
 - one fix for a memleak in v3d
Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Maxime Ripard <mripard@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20191030182207.evrscl7lnv42u5zu@hendrix

One memory corruption fix in the MMUv2 GPU coredump code, a deadlock
fix also in the coredump code and reintroduction of a helpful message,
which got dropped by accident in this cycle.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Lucas Stach <l.stach@pengutronix.de>
Link: https://patchwork.freedesktop.org/patch/msgid/b0d640267662e3ce5e0089d0afedc1baba55058d.camel@pengutronix.de

* pm-cpufreq:
  ACPI: processor: Add QoS requests for all CPUs

The Kryo cores share errata 1009 with Falkor, so add their model
definitions and enable it for them as well.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
[will: Update entry in silicon-errata.rst]
Signed-off-by: Will Deacon <will@kernel.org>

VMX already does so if the host has SMEP, in order to support the combination of
CR0.WP=1 and CR4.SMEP=1.  However, it is perfectly safe to always do so, and in
fact VMX already ends up running with EFER.NXE=1 on old processors that lack the
"load EFER" controls, because it may help avoiding a slow MSR write.  Removing
all the conditionals simplifies the code.

SVM does not have similar code, but it should since recent AMD processors do
support SMEP.  So this patch also makes the code for the two vendors more similar
while fixing NPT=0, CR0.WP=1 and CR4.SMEP=1 on AMD processors.

Cc: stable@vger.kernel.org
Cc: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

In kvm_create_vm(), if we've successfully called kvm_arch_init_vm(), but
then fail later in the function, we need to call kvm_arch_destroy_vm()
so that it can do any necessary cleanup (like freeing memory).

Fixes: 44a95dae ("KVM: x86: Detect and Initialize AVIC support")
Signed-off-by: John Sperbeck <jsperbeck@google.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Reviewed-by: Junaid Shahid <junaids@google.com>
[Remove dependency on "kvm: Don't clear reference count on
 kvm_create_vm() error path" which was not committed. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

The driver exposes EFI runtime services to user-space through an IOCTL
interface, calling the EFI services function pointers directly without
using the efivar API.

Disallow access to the /dev/efi_test character device when the kernel is
locked down to prevent arbitrary user-space to call EFI runtime services.

Also require CAP_SYS_ADMIN to open the chardev to prevent unprivileged
users to call the EFI runtime services, instead of just relying on the
chardev file mode bits for this.

The main user of this driver is the fwts [0] tool that already checks if
the effective user ID is 0 and fails otherwise. So this change shouldn't
cause any regression to this tool.

[0]: https://wiki.ubuntu.com/FirmwareTestSuite/Reference/uefivarinfoSigned-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Matthew Garrett <mjg59@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Link: https://lkml.kernel.org/r/20191029173755.27149-7-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

Currently, kernel fails to boot on some HyperV VMs when using EFI.
And it's a potential issue on all x86 platforms.

It's caused by broken kernel relocation on EFI systems, when below three
conditions are met:

1. Kernel image is not loaded to the default address (LOAD_PHYSICAL_ADDR)
   by the loader.
2. There isn't enough room to contain the kernel, starting from the
   default load address (eg. something else occupied part the region).
3. In the memmap provided by EFI firmware, there is a memory region
   starts below LOAD_PHYSICAL_ADDR, and suitable for containing the
   kernel.

EFI stub will perform a kernel relocation when condition 1 is met. But
due to condition 2, EFI stub can't relocate kernel to the preferred
address, so it fallback to ask EFI firmware to alloc lowest usable memory
region, got the low region mentioned in condition 3, and relocated
kernel there.

It's incorrect to relocate the kernel below LOAD_PHYSICAL_ADDR. This
is the lowest acceptable kernel relocation address.

The first thing goes wrong is in arch/x86/boot/compressed/head_64.S.
Kernel decompression will force use LOAD_PHYSICAL_ADDR as the output
address if kernel is located below it. Then the relocation before
decompression, which move kernel to the end of the decompression buffer,
will overwrite other memory region, as there is no enough memory there.

To fix it, just don't let EFI stub relocate the kernel to any address
lower than lowest acceptable address.

[ ardb: introduce efi_low_alloc_above() to reduce the scope of the change ]
Signed-off-by: Kairui Song <kasong@redhat.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Link: https://lkml.kernel.org/r/20191029173755.27149-6-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

The EFI stubloader for ARM starts out by allocating a 32 MB window
at the base of RAM, in order to ensure that the decompressor (which
blindly copies the uncompressed kernel into that window) does not
overwrite other allocations that are made while running in the context
of the EFI firmware.

In some cases, (e.g., U-Boot running on the Raspberry Pi 2), this is
causing boot failures because this initial allocation conflicts with
a page of reserved memory at the base of RAM that contains the SMP spin
tables and other pieces of firmware data and which was put there by
the bootloader under the assumption that the TEXT_OFFSET window right
below the kernel is only used partially during early boot, and will be
left alone once the memory reservations are processed and taken into
account.

So let's permit reserved memory regions to exist in the region starting
at the base of RAM, and ending at TEXT_OFFSET - 5 * PAGE_SIZE, which is
the window below the kernel that is not touched by the early boot code.
Tested-by: Guillaume Gardet <Guillaume.Gardet@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Chester Lin <clin@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Link: https://lkml.kernel.org/r/20191029173755.27149-5-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

Commit 428826f5 ("fdt: add support for rng-seed") introduced
add_bootloader_randomness(), permitting randomness provided by the
bootloader or firmware to be credited as entropy. However, the fact
that the UEFI support code was already wired into the RNG subsystem
via a call to add_device_randomness() was overlooked, and so it was
not converted at the same time.

Note that this UEFI (v2.4 or newer) feature is currently only
implemented for EFI stub booting on ARM, and further note that
CONFIG_RANDOM_TRUST_BOOTLOADER must be enabled, and this should be
done only if there indeed is sufficient trust in the bootloader
_and_ its source of randomness.

[ ardb: update commit log ]
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Link: https://lkml.kernel.org/r/20191029173755.27149-4-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

Currently nothing checks the return value of efi_tpm_eventlog_init(),
but in case that changes in the future make sure an error is
returned when it fails to determine the tpm final events log
size.
Suggested-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Jerry Snitselaar <jsnitsel@redhat.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Fixes: e658c82b ("efi/tpm: Only set 'efi_tpm_final_log_size' after ...")
Link: https://lkml.kernel.org/r/20191029173755.27149-3-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

For the EFI_RCI2_TABLE Kconfig option, 'make oldconfig' asks the user
for input on platforms where the option may not be applicable. This patch
modifies the Kconfig option to ask the user for input only when CONFIG_X86
or CONFIG_COMPILE_TEST is set to y.
Suggested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Narendra K <Narendra.K@dell.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-efi@vger.kernel.org
Link: https://lkml.kernel.org/r/20191029173755.27149-2-ardb@kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

Pull dmaengine fixes from Vinod Koul:
 "A few fixes to the dmaengine drivers:

   - fix in sprd driver for link list and potential memory leak

   - tegra transfer failure fix

   - imx size check fix for script_number

   - xilinx fix for 64bit AXIDMA and control reg update

   - qcom bam dma resource leak fix

   - cppi slave transfer fix when idle"

* tag 'dmaengine-fix-5.4-rc6' of git://git.infradead.org/users/vkoul/slave-dma:
  dmaengine: cppi41: Fix cppi41_dma_prep_slave_sg() when idle
  dmaengine: qcom: bam_dma: Fix resource leak
  dmaengine: sprd: Fix the possible memory leak issue
  dmaengine: xilinx_dma: Fix control reg update in vdma_channel_set_config
  dmaengine: xilinx_dma: Fix 64-bit simple AXIDMA transfer
  dmaengine: imx-sdma: fix size check for sdma script_number
  dmaengine: tegra210-adma: fix transfer failure
  dmaengine: sprd: Fix the link-list pointer register configuration issue

When a card is disconnected while in use, the system waits until all
opened files are closed then releases the card.  This is done via
put_device() of the card device in each device release code.

The recently reported mutex deadlock bug happens in this code path;
snd_timer_close() for the timer device deals with the global
register_mutex and it calls put_device() there.  When this timer
device is the last one, the card gets freed and it eventually calls
snd_timer_free(), which has again the protection with the global
register_mutex -- boom.

Basically put_device() call itself is race-free, so a relative simple
workaround is to move this put_device() call out of the mutex.  For
achieving that, in this patch, snd_timer_close_locked() got a new
argument to store the card device pointer in return, and each caller
invokes put_device() with the returned object after the mutex unlock.
Reported-and-tested-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>

A final attempt at enabling sse2 for GCC users.

Orininally attempted in:
commit 10117450 ("drm/amd/display: add -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines")

Reverted due to "reported instability" in:
commit 193392ed ("Revert "drm/amd/display: add -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines"")

Re-added just for Clang in:
commit 0f0727d9 ("drm/amd/display: readd -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines")

The original report didn't have enough information to know if the GPF
was due to misalignment, but I suspect that it was. (The missing
information was the disassembly of the function at the bottom of the
trace, to see if the instruction pointer pointed to an instruction with
16B alignment memory operand requirements.  The stack trace does show
the stack was only 8B but not 16B aligned though, which makes this a
strong possibility).

Now that the stack misalignment issue has been fixed for users of GCC
7.1+, reattempt adding -msse2. This matches Clang.

It will likely never be safe to enable this for pre-GCC 7.1 AND use a
16B aligned stack in these translation units.

This is only a functional change for GCC 7.1+ users, and should be boot
tested.

Link: https://bugs.freedesktop.org/show_bug.cgi?id=109487Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

GCC earlier than 7.1 errors when compiling code that makes use of
`double`s and sets a stack alignment outside of the range of [2^4-2^12]:

$ cat foo.c
double foo(double x, double y) {
  return x + y;
}
$ gcc-4.9 -mpreferred-stack-boundary=3 foo.c
error: -mpreferred-stack-boundary=3 is not between 4 and 12

This is likely why the AMDGPU driver was ever compiled with a different
stack alignment (and thus different ABI) than the rest of the x86
kernel. The kernel uses 8B stack alignment, while the driver was using
16B stack alignment in a few places.

Since GCC 7.1+ doesn't error, fix the ABI mismatch for users of newer
versions of GCC.

There was discussion about whether to mark the driver broken or not for
users of GCC earlier than 7.1, but since the driver currently is
working, don't explicitly break the driver for them here.

Relying on differing stack alignment is unspecified behavior, and
brittle, and may break in the future.

This patch is no functional change for GCC users earlier than 7.1. It's
been compile tested on GCC 4.9 and 8.3 to check the correct flags. It
should be boot tested when built with GCC 7.1+.

-mincoming-stack-boundary= or -mstackrealign may help keep this code
building for pre-GCC 7.1 users.

The version check for GCC is broken into two conditionals, both because
cc-ifversion is currently GCC specific, and it simplifies a subsequent
patch.
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

The x86 kernel is compiled with an 8B stack alignment via
`-mpreferred-stack-boundary=3` for GCC since 3.6-rc1 via
commit d9b0cde9 ("x86-64, gcc: Use -mpreferred-stack-boundary=3 if supported")
or `-mstack-alignment=8` for Clang. Parts of the AMDGPU driver are
compiled with 16B stack alignment.

Generally, the stack alignment is part of the ABI. Linking together two
different translation units with differing stack alignment is dangerous,
particularly when the translation unit with the smaller stack alignment
makes calls into the translation unit with the larger stack alignment.
While 8B aligned stacks are sometimes also 16B aligned, they are not
always.

Multiple users have reported General Protection Faults (GPF) when using
the AMDGPU driver compiled with Clang. Clang is placing objects in stack
slots assuming the stack is 16B aligned, and selecting instructions that
require 16B aligned memory operands.

At runtime, syscall handlers with 8B aligned stack call into code that
assumes 16B stack alignment.  When the stack is a multiple of 8B but not
16B, these instructions result in a GPF.

Remove the code that added compatibility between the differing compiler
flags, as it will result in runtime GPFs when built with Clang. Cleanups
for GCC will be sent in later patches in the series.

Link: https://github.com/ClangBuiltLinux/linux/issues/735Debugged-by: Yuxuan Shui <yshuiv7@gmail.com>
Reported-by: Shirish S <shirish.s@amd.com>
Reported-by: Yuxuan Shui <yshuiv7@gmail.com>
Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

During kexec some adapters hit an EEH since they are not properly
shut down in the radeon_pci_shutdown() function. Adding
radeon_suspend_kms() fixes this issue.
Enabled only on PPC because this patch causes issues on some other
boards.
Signed-off-by: Kyle Mahlkuch <kmahlkuc@linux.vnet.ibm.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

These were not aligned for optimal performance for GPUVM.
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tianci Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

The vega10_odn_update_soc_table() function does not allow the SCLK
dependent voltage to be set for power-state 7 to a value below the default
in pptable. Change the for-loop condition to allow undervolting in the
highest state.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=205277Signed-off-by: Pelle van Gils <pelle@vangils.xyz>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

dc.c:583:null check is needed after using kzalloc function
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: zhongshiqi <zhong.shiqi@zte.com.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

[Why]
This patch is for fixing Navi14 HDMI display pink screen issue.

[How]
Call stream->link->link_enc->funcs->setup twice. This is setting
the DIG_MODE to the correct value after having been overridden by
the call to transmitter control.
Signed-off-by: Zhan Liu <zhan.liu@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

[WHY]
i2c_read is called to differentiate passive DP->HDMI and DP->DVI-D dongles
The call is expected to fail in DVI-D case but pass in HDMI case
Some HDMI dongles have a chance to fail as well, causing misdetection as DVI-D

[HOW]
Retry i2c_read to ensure failed result is valid
Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

[why]
There's a use case for inverted gamma
and it's been confirmed that negative slopes are ok.

[how]
Remove code for blocking non-monotonically increasing gamma
Signed-off-by: Aidan Yang <Aidan.Yang@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Reza Amini <Reza.Amini@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

[why]
A display that supports DRR can never really be considered
"synchronized" with any other display because we can dynamically
enable DRR (i.e. without modeset).  this will cause their
relative CRTC positions to drift and lose sync.  this will disrupt
features such as MCLK switching that assume and depend on
their permanent alignment (that can only change with modeset)

[how]
check for ignore_msa in stream when considered synchronizability
this ignore_msa is basically actually implemented as "supports drr"
Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Use ERR_PTR to return back the error happened during amdgpu_ib_schedule.
Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Problem:
When run_job fails and HW fence returned is NULL we still signal
the s_fence to avoid hangs but the user has no way of knowing if
the actual HW job was ran and finished.

Fix:
Allow .run_job implementations to return ERR_PTR in the fence pointer
returned and then set this error for s_fence->finished fence so whoever
wait on this fence can inspect the signaled fence for an error.
Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>