If the GuC CTs are full and we need to stall the request submission
while waiting for space, we save the stalled request and where the stall
occurred; when the CTs have space again we pick up the request submission
from where we left off.
If a full GT reset occurs, the state of all contexts is cleared and all
non-guilty requests are unsubmitted, therefore we need to restart the
stalled request submission from scratch. To make sure that we do so,
clear the saved request after a reset.
Fixes note: the patch that introduced the bug is in 5.15, but no
officially supported platform had GuC submission enabled by default
in that kernel, so the backport to that particular version (and only
that one) can potentially be skipped.
Fixes: 925dc1cf58 ("drm/i915/guc: Implement GuC submission tasklet")
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: <stable@vger.kernel.org> # v5.15+
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220811210812.3239621-1-daniele.ceraolospurio@intel.com
(cherry picked from commit f922fbb0f2)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Pull drm updates from Dave Airlie:
"Highlights:
- New driver for logicvc - which is a display IP core.
- EDID parser rework to add new extensions
- fbcon scrolling improvements
- i915 has some more DG2 work but not enabled by default, but should
have enough features for userspace to work now.
Otherwise it's lots of work all over the place. Detailed summary:
New driver:
- logicvc
vfio:
- use aperture API
core:
- of: Add data-lane helpers and convert drivers
- connector: Remove deprecated ida_simple_get()
media:
- Add various RGB666 and RGB888 format constants
panel:
- Add HannStar HSD101PWW
- Add ETML0700Y5DHA
dma-buf:
- add sync-file API
- set dma mask for udmabuf devices
fbcon:
- Improve scrolling performance
- Sanitize input
fbdev:
- device unregistering fixes
- vesa: Support COMPILE_TEST
- Disable firmware-device registration when first native driver loads
aperture:
- fix segfault during hot-unplug
- export for use with other subsystems
client:
- use driver validated modes
dp:
- aux: make probing more reliable
- mst: Read extended DPCD capabilities during system resume
- Support waiting for HDP signal
- Port-validation fixes
edid:
- CEA data-block iterators
- struct drm_edid introduction
- implement HF-EEODB extension
gem:
- don't use fb format non-existing planes
probe-helper:
- use 640x480 as displayport fallback
scheduler:
- don't kill jobs in interrupt context
bridge:
- Add support for i.MX8qxp and i.MX8qm
- lots of fixes/cleanups
- Add TI-DLPC3433
- fy07024di26a30d: Optional GPIO reset
- ldb: Add reg and reg-name properties to bindings, Kconfig fixes
- lt9611: Fix display sensing;
- tc358767: DSI/DPI refactoring and DSI-to-eDP support, DSI lane handling
- tc358775: Fix clock settings
- ti-sn65dsi83: Allow GPIO to sleep
- adv7511: I2C fixes
- anx7625: Fix error handling; DPI fixes; Implement HDP timeout via callback
- fsl-ldb: Drop DE flip
- ti-sn65dsi86: Convert to atomic modesetting
amdgpu:
- use atomic fence helpers in DM
- fix VRAM address calculations
- export CRTC bpc via debugfs
- Initial devcoredump support
- Enable high priority gfx queue on asics which support it
- Adjust GART size on newer APUs for S/G display
- Soft reset for GFX 11 / SDMA 6
- Add gfxoff status query for vangogh
- Fix timestamps for cursor only commits
- Adjust GART size on newer APUs for S/G display
- fix buddy memory corruption
amdkfd:
- MMU notifier fixes
- P2P DMA support using dma-buf
- Add available memory IOCTL
- HMM profiler support
- Simplify GPUVM validation
- Unified memory for CWSR save/restore area
i915:
- General driver clean-up
- DG2 enabling (still under force probe)
- DG2 small BAR memory support
- HuC loading support
- DG2 workarounds
- DG2/ATS-M device IDs added
- Ponte Vecchio prep work and new blitter engines
- add Meteorlake support
- Fix sparse warnings
- DMC MMIO range checks
- Audio related fixes
- Runtime PM fixes
- PSR fixes
- Media freq factor and per-gt enhancements
- DSI fixes for ICL+
- Disable DMC flip queue handlers
- ADL_P voltage swing updates
- Use more the VBT for panel information
- Fix on Type-C ports with TBT mode
- Improve fastset and allow seamless M/N changes
- Accept more fixed modes with VRR/DMRRS panels
- Disable connector polling for a headless SKU
- ADL-S display PLL w/a
- Enable THP on Icelake and beyond
- Fix i915_gem_object_ggtt_pin_ww regression on old platforms
- Expose per tile media freq factor in sysfs
- Fix dma_resv fence handling in multi-batch execbuf
- Improve on suspend / resume time with VT-d enabled
- export CRTC bpc settings via debugfs
msm:
- gpu: a619 support
- gpu: Fix for unclocked GMU register access
- gpu: Devcore dump enhancements
- client utilization via fdinfo support
- fix fence rollover issue
- gem: Lockdep false-positive warning fix
- gem: Switch to pfn mappings
- WB support on sc7180
- dp: dropped custom bulk clock implementation
- fix link retraining on resolution change
- hdmi: dropped obsolete GPIO support
tegra:
- context isolation for host1x engines
- tegra234 soc support
mediatek:
- add vdosys0/1 for mt8195
- add MT8195 dp_intf driver
exynos:
- Fix resume function issue of exynos decon driver by calling
clk_disable_unprepare() properly if clk_prepare_enable() failed.
nouveau:
- set of misc fixes/cleanups
- display cleanups
gma500:
- Cleanup connector I2C handling
hyperv:
- Unify VRAM allocation of Gen1 and Gen2
meson:
- Support YUV422 output; Refcount fixes
mgag200:
- Support damage clipping
- Support gamma handling
- Protect concurrent HW access
- Fixes to connector
- Store model-specific limits in device-info structure
- fix PCI register init
panfrost:
- Valhall support
r128:
- Fix bit-shift overflow
rockchip:
- Locking fixes in error path
ssd130x:
- Fix built-in linkage
udl:
- Always advertize VGA connector
ast:
- Support multiple outputs
- fix black screen on resume
sun4i:
- HDMI PHY cleanups
vc4:
- Add support for BCM2711
vkms:
- Allocate output buffer with vmalloc()
mcde:
- Fix ref-count leak
mxsfb/lcdif:
- Support i.MX8MP LCD controller
stm/ltdc:
- Support dynamic Z order
- Support mirroring
ingenic:
- Fix display at maximum resolution"
* tag 'drm-next-2022-08-03' of git://anongit.freedesktop.org/drm/drm: (1480 commits)
drm/amd/display: Fix a compilation failure on PowerPC caused by FPU code
drm/amdgpu: enable support for psp 13.0.4 block
drm/amdgpu: add files for PSP 13.0.4
drm/amdgpu: add header files for MP 13.0.4
drm/amdgpu: correct RLC_RLCS_BOOTLOAD_STATUS offset and index
drm/amdgpu: send msg to IMU for the front-door loading
drm/amdkfd: use time_is_before_jiffies(a + b) to replace "jiffies - a > b"
drm/amdgpu: fix hive reference leak when reflecting psp topology info
drm/amd/pm: enable GFX ULV feature support for SMU13.0.0
drm/amd/pm: update driver if header for SMU 13.0.0
drm/amdgpu: move mes self test after drm sched re-started
drm/amdgpu: drop non-necessary call trace dump
drm/amdgpu: enable VCN cg and JPEG cg/pg
drm/amdgpu: vcn_4_0_2 video codec query
drm/amdgpu: add VCN_4_0_2 firmware support
drm/amdgpu: add VCN function in NBIO v7.7
drm/amdgpu: fix a vcn4 boot poll bug in emulation mode
drm/amd/amdgpu: add memory training support for PSP_V13
drm/amdkfd: remove an unnecessary amdgpu_bo_ref
drm/amd/pm: Add get_gfx_off_status interface for yellow carp
...
For execlists backend, current implementation of Wa_22011802037 is to
stop the CS before doing a reset of the engine. This WA was further
extended to wait for any pending MI FORCE WAKEUPs before issuing a
reset. Add the extended steps in the execlist path of reset.
In addition, extend the WA to gen11.
v2: (Tvrtko)
- Clarify comments, commit message, fix typos
- Use IS_GRAPHICS_VER for gen 11/12 checks
v3: (Daneile)
- Drop changes to intel_ring_submission since WA does not apply to it
- Log an error if MSG IDLE is not defined for an engine
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Fixes: f6aa0d713c ("drm/i915: Add Wa_22011802037 force cs halt")
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220621192105.2100585-1-umesh.nerlige.ramappa@intel.com
(cherry picked from commit 0667429ce6)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
This patch re-introduces support for GuC v69 in parallel to v70. As this
is a quick fix, v69 has been re-introduced as the single "fallback" guc
version in case v70 is not available on disk and only for platforms that
are out of force_probe and require the GuC by default. All v69 specific
code has been labeled as such for easy identification, and the same was
done for all v70 functions for which there is a separate v69 version,
to avoid accidentally calling the wrong version via the unlabeled name.
When the fallback mode kicks in, a drm_notice message is printed in
dmesg to inform the user of the required update. The existing
logging of the fetch function has also been updated so that we no
longer complain immediately if we can't find a fw and we only throw an
error if the fetch of both the base and fallback blobs fails.
The plan is to follow this up with a more complex rework to allow for
multiple different GuC versions to be supported at the same time.
v2: reduce the fallback to platform that require it, switch to
firmware_request_nowarn(), improve logs.
Fixes: 2584b3549f ("drm/i915/guc: Update to GuC version 70.1.1")
Link: https://lists.freedesktop.org/archives/intel-gfx/2022-July/301640.html
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Dave Airlie <airlied@gmail.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220718230732.1409641-1-daniele.ceraolospurio@intel.com
(cherry picked from commit 774ce1510e)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
For execlists backend, current implementation of Wa_22011802037 is to
stop the CS before doing a reset of the engine. This WA was further
extended to wait for any pending MI FORCE WAKEUPs before issuing a
reset. Add the extended steps in the execlist path of reset.
In addition, extend the WA to gen11.
v2: (Tvrtko)
- Clarify comments, commit message, fix typos
- Use IS_GRAPHICS_VER for gen 11/12 checks
v3: (Daneile)
- Drop changes to intel_ring_submission since WA does not apply to it
- Log an error if MSG IDLE is not defined for an engine
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Fixes: f6aa0d713c ("drm/i915: Add Wa_22011802037 force cs halt")
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220621192105.2100585-1-umesh.nerlige.ramappa@intel.com
Using two different types of workoads, it was observed that
guc_update_engine_gt_clks was being called too frequently and/or
causing a CPU-to-lmem bandwidth hit over PCIE. Details on
the workloads and numbers are in the notes below.
Background: At the moment, guc_update_engine_gt_clks can be invoked
via one of 3 ways. #1 and #2 are infrequent under normal operating
conditions:
1.When a predefined "ping_delay" timer expires so that GuC-
busyness can sample the GTPM clock counter to ensure it
doesn't miss a wrap-around of the 32-bits of the HW counter.
(The ping_delay is calculated based on 1/8th the time taken
for the counter go from 0x0 to 0xffffffff based on the
GT frequency. This comes to about once every 28 seconds at a
GT frequency of 19.2Mhz).
2.In preparation for a gt reset.
3.In response to __gt_park events (as the gt power management
puts the gt into a lower power state when there is no work
being done).
Root-cause: For both the workloads described farther below, it was
observed that when user space calls IOCTLs that unparks the
gt momentarily and repeats such calls many times in quick succession,
it triggers calling guc_update_engine_gt_clks as many times. However,
the primary purpose of guc_update_engine_gt_clks is to ensure we don't
miss the wraparound while the counter is ticking. Thus, the solution
is to ensure we skip that check if gt_park is calling this function
earlier than necessary.
Solution: Snapshot jiffies when we do actually update the busyness
stats. Then get the new jiffies every time intel_guc_busyness_park
is called and bail if we are being called too soon. Use half of the
ping_delay as a safe threshold.
NOTE1: Workload1: IGTs' gem_create was modified to create a file handle,
allocate memory with sizes that range from a min of 4K to the max supported
(in power of two step-sizes). Its maps, modifies and reads back the
memory. Allocations and modification is repeated until total memory
allocation reaches the max. Then the file handle is closed. With this
workload, guc_update_engine_gt_clks was called over 188 thousand times
in the span of 15 seconds while this test ran three times. With this patch,
the number of calls reduced to 14.
NOTE2: Workload2: 30 transcode sessions are created in quick succession.
While these sessions are created, pcm-iio tool was used to measure I/O
read operation bandwidth consumption sampled at 100 milisecond intervals
over the course of 20 seconds. The total bandwidth consumed over 20 seconds
without this patch was measured at average at 311KBps per sample. With this
patch, the number went down to about 175Kbps which is about a 43% savings.
Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220623023157.211650-2-alan.previn.teres.alexis@intel.com
We have long standing customer complaints that pressing Ctrl-C (or to the
effect of) causes engine resets with otherwise well behaving programs.
Not only is logging engine resets during normal operation not desirable
since it creates support incidents, but more fundamentally we should avoid
going the engine reset path when we can since any engine reset introduces
a chance of harming an innocent context.
Reason for this undesirable behaviour is that the driver currently does
not distinguish between banned contexts and non-persistent contexts which
have been closed.
To fix this we add the distinction between the two reasons for revoking
contexts, which then allows the strict timeout only be applied to banned,
while innocent contexts (well behaving) can preempt cleanly and exit
without triggering the engine reset path.
Note that the added context exiting category applies both to closed non-
persistent context, and any exiting context when hangcheck has been
disabled by the user.
At the same time we rename the backend operation from 'ban' to 'revoke'
which more accurately describes the actual semantics. (There is no ban at
the backend level since banning is a concept driven by the scheduling
frontend. Backends are simply able to revoke a running context so that
is the more appropriate name chosen.)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220527072452.2225610-1-tvrtko.ursulin@linux.intel.com
Pull drm updates from Dave Airlie:
"Intel have enabled DG2 on certain SKUs for laptops, AMD has started
some new GPU support, msm has user allocated VA controls
dma-buf:
- add dma_resv_replace_fences
- add dma_resv_get_singleton
- make dma_excl_fence private
core:
- EDID parser refactorings
- switch drivers to drm_mode_copy/duplicate
- DRM managed mutex initialization
display-helper:
- put HDMI, SCDC, HDCP, DSC and DP into new module
gem:
- rework fence handling
ttm:
- rework bulk move handling
- add common debugfs for resource managers
- convert to kvcalloc
format helpers:
- support monochrome formats
- RGB888, RGB565 to XRGB8888 conversions
fbdev:
- cfb/sys_imageblit fixes
- pagelist corruption fix
- create offb platform device
- deferred io improvements
sysfb:
- Kconfig rework
- support for VESA mode selection
bridge:
- conversions to devm_drm_of_get_bridge
- conversions to panel_bridge
- analogix_dp - autosuspend support
- it66121 - audio support
- tc358767 - DSI to DPI support
- icn6211 - PLL/I2C fixes, DT property
- adv7611 - enable DRM_BRIDGE_OP_HPD
- anx7625 - fill ELD if no monitor
- dw_hdmi - add audio support
- lontium LT9211 support, i.MXMP LDB
- it6505: Kconfig fix, DPCD set power fix
- adv7511 - CEC support for ADV7535
panel:
- ltk035c5444t, B133UAN01, NV3052C panel support
- DataImage FG040346DSSWBG04 support
- st7735r - DT bindings fix
- ssd130x - fixes
i915:
- DG2 laptop PCI-IDs ("motherboard down")
- Initial RPL-P PCI IDs
- compute engine ABI
- DG2 Tile4 support
- DG2 CCS clear color compression support
- DG2 render/media compression formats support
- ATS-M platform info
- RPL-S PCI IDs added
- Bump ADL-P DMC version to v2.16
- Support static DRRS
- Support multiple eDP/LVDS native mode refresh rates
- DP HDR support for HSW+
- Lots of display refactoring + fixes
- GuC hwconfig support and query
- sysfs support for multi-tile
- fdinfo per-client gpu utilisation
- add geometry subslices query
- fix prime mmap with LMEM
- fix vm open count and remove vma refcounts
- contiguous allocation fixes
- steered register write support
- small PCI BAR enablement
- GuC error capture support
- sunset igpu legacy mmap support for newer devices
- GuC version 70.1.1 support
amdgpu:
- Initial SoC21 support
- SMU 13.x enablement
- SMU 13.0.4 support
- ttm_eu cleanups
- USB-C, GPUVM updates
- TMZ fixes for RV
- RAS support for VCN
- PM sysfs code cleanup
- DC FP rework
- extend CG/PG flags to 64-bit
- SI dpm lockdep fix
- runtime PM fixes
amdkfd:
- RAS/SVM fixes
- TLB flush fixes
- CRIU GWS support
- ignore bogus MEC signals more efficiently
msm:
- Fourcc modifier for tiled but not compressed layouts
- Support for userspace allocated IOVA (GPU virtual address)
- DPU: DSC (Display Stream Compression) support
- DP: eDP support
- DP: conversion to use drm_bridge and drm_bridge_connector
- Merge DPU1 and MDP5 MDSS driver
- DPU: writeback support
nouveau:
- make some structures static
- make some variables static
- switch to drm_gem_plane_helper_prepare_fb
radeon:
- misc fixes/cleanups
mxsfb:
- rework crtc mode setting
- LCDIF CRC support
etnaviv:
- fencing improvements
- fix address space collisions
- cleanup MMU reference handling
gma500:
- GEM/GTT improvements
- connector handling fixes
komeda:
- switch to plane reset helper
mediatek:
- MIPI DSI improvements
omapdrm:
- GEM improvements
qxl:
- aarch64 support
vc4:
- add a CL submission tracepoint
- HDMI YUV support
- HDMI/clock improvements
- drop is_hdmi caching
virtio:
- remove restriction of non-zero blob types
vmwgfx:
- support for cursormob and cursorbypass 4
- fence improvements
tidss:
- reset DISPC on startup
solomon:
- SPI support
- DT improvements
sun4i:
- allwinner D1 support
- drop is_hdmi caching
imx:
- use swap() instead of open-coding
- use devm_platform_ioremap_resource
- remove redunant initializations
ast:
- Displayport support
rockchip:
- Refactor IOMMU initialisation
- make some structures static
- replace drm_detect_hdmi_monitor with drm_display_info.is_hdmi
- support swapped YUV formats,
- clock improvements
- rk3568 support
- VOP2 support
mediatek:
- MT8186 support
tegra:
- debugabillity improvements"
* tag 'drm-next-2022-05-25' of git://anongit.freedesktop.org/drm/drm: (1740 commits)
drm/i915/dsi: fix VBT send packet port selection for ICL+
drm/i915/uc: Fix undefined behavior due to shift overflowing the constant
drm/i915/reg: fix undefined behavior due to shift overflowing the constant
drm/i915/gt: Fix use of static in macro mismatch
drm/i915/audio: fix audio code enable/disable pipe logging
drm/i915: Fix CFI violation with show_dynamic_id()
drm/i915: Fix 'mixing different enum types' warnings in intel_display_power.c
drm/i915/gt: Fix build error without CONFIG_PM
drm/msm/dpu: handle pm_runtime_get_sync() errors in bind path
drm/msm/dpu: add DRM_MODE_ROTATE_180 back to supported rotations
drm/msm: don't free the IRQ if it was not requested
drm/msm/dpu: limit writeback modes according to max_linewidth
drm/amd: Don't reset dGPUs if the system is going to s2idle
drm/amdgpu: Unmap legacy queue when MES is enabled
drm: msm: fix possible memory leak in mdp5_crtc_cursor_set()
drm/msm: Fix fb plane offset calculation
drm/msm/a6xx: Fix refcount leak in a6xx_gpu_init
drm/msm/dsi: don't powerup at modeset time for parade-ps8640
drm/rockchip: Change register space names in vop2
dt-bindings: display: rockchip: make reg-names mandatory for VOP2
...
There are 2 ways an engine can get reset in i915 and the method of reset
affects how KMD labels a context as guilty/innocent.
(1) GuC initiated engine-reset: GuC resets a hung engine and notifies
KMD. The context that hung on the engine is marked guilty and all other
contexts are innocent. The innocent contexts are resubmitted.
(2) GT based reset: When an engine heartbeat fails to tick, KMD
initiates a gt/chip reset. All active contexts are marked as guilty and
discarded.
In order to correctly mark the contexts as guilty/innocent, pass a mask
of engines that were reset to __guc_reset_context.
Fixes: eb5e7da736 ("drm/i915/guc: Reset implementation for new GuC interface")
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220426003045.3929439-1-umesh.nerlige.ramappa@intel.com
(cherry picked from commit 303760aa91)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
There are 2 ways an engine can get reset in i915 and the method of reset
affects how KMD labels a context as guilty/innocent.
(1) GuC initiated engine-reset: GuC resets a hung engine and notifies
KMD. The context that hung on the engine is marked guilty and all other
contexts are innocent. The innocent contexts are resubmitted.
(2) GT based reset: When an engine heartbeat fails to tick, KMD
initiates a gt/chip reset. All active contexts are marked as guilty and
discarded.
In order to correctly mark the contexts as guilty/innocent, pass a mask
of engines that were reset to __guc_reset_context.
Fixes: eb5e7da736 ("drm/i915/guc: Reset implementation for new GuC interface")
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220426003045.3929439-1-umesh.nerlige.ramappa@intel.com
Userspace may leave predication enabled upon return from the batch
buffer, which has the consequent of preventing all operation from the
ring from being executed, including all the synchronisation, coherency
control, arbitration and user signaling. This is more than just a local
gpu hang in one client, as the user has the ability to prevent the
kernel from applying critical workarounds and can cause a full GT reset.
We could simply execute MI_SET_PREDICATE upon return from the user
batch, but this has the repercussion of modifying the user's context
state. Instead, we opt to execute a fixup batch which by mixing
predicated operations can determine the state of the
SET_PREDICATE_RESULT register and restore it prior to the next userspace
batch. This allows us to protect the kernel's ring without changing the
uABI.
Suggested-by: Zbigniew Kempczynski <zbigniew.kempczynski@intel.com>
Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Cc: Zbigniew Kempczynski <zbigniew.kempczynski@intel.com>
Cc: Thomas Hellstrom <thomas.hellstrom@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220425152317.4275-4-ramalingam.c@intel.com
Initiating a reset when the command streamer is not idle or in the
middle of executing an MI_FORCE_WAKE can result in a hang. Multiple
command streamers can be part of a single reset domain, so resetting one
would mean resetting all command streamers in that domain.
To workaround this, before initiating a reset, ensure that all command
streamers within that reset domain are either IDLE or are not executing
a MI_FORCE_WAKE.
Enable GuC PRE_PARSER WA bit so that GuC follows the WA sequence when
initiating engine-resets.
For gt-resets, ensure that i915 applies the WA sequence.
Opens to address in future patches:
- The part of the WA to wait for pending forcewakes is also applicable
to execlists backend.
- The WA also needs to be applied for gen11
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220415224025.3693037-3-umesh.nerlige.ramappa@intel.com
The latest GuC firmware drops the context descriptor pool in favour of
passing all creation data in the create H2G. It also greatly simplifies
the work queue and removes the process descriptor used for multi-LRC
submission. So, remove all mention of LRC and process descriptors and
update the registration code accordingly.
Unfortunately, the new API also removes the ability to set default
values for the scheduling policies at context registration time.
Instead, a follow up H2G must be sent. The individual scheduling
policy update H2G commands are also dropped in favour of a single KLV
based H2G. So, change the update wrappers accordingly and call this
during context registration..
Of course, this second H2G per registration might fail due to being
backed up. The registration code has a complicated state machine to
cope with the actual registration call failing. However, if that works
then there is no support for unwinding if a further call should fail.
Unwinding would require sending a H2G to de-register - but that can't
be done because the CTB is already backed up.
So instead, add a new flag to say whether the context has a pending
policy update. This is set if the policy H2G fails at registration
time. The submission code checks for this flag and retries the policy
update if set. If that call fails, the submission path early exists
with a retry error. This is something that is already supported for
other reasons.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220412225955.1802543-2-John.C.Harrison@Intel.com
Print the GuC captured error state register list (string names
and values) when gpu_coredump_state printout is invoked via
the i915 debugfs for flushing the gpu error-state that was
captured prior.
Since GuC could have reported multiple engine register dumps
in a single notification event, parse the captured data
(appearing as a stream of structures) to identify each dump as
a different 'engine-capture-group-output'.
Finally, for each 'engine-capture-group-output' that is found,
verify if the engine register dump corresponds to the
engine_coredump content that was previously populated by the
i915_gpu_coredump function. That function would have copied
the context's vma's including the bacth buffer during the
G2H-context-reset notification that occurred earlier. Perform
this verification check by comparing guc_id, lrca and engine-
instance obtained from the 'engine-capture-group-output' vs a
copy of that same info taken during i915_gpu_coredump. If
they match, then print those vma's as well (such as the batch
buffers).
NOTE: the output format was verified using the gem_exec_capture
IGT test.
Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220321164527.2500062-14-alan.previn.teres.alexis@intel.com
Add a flags parameter through all of the coredump creation
functions. Add a bitmask flag to indicate if the top
level gpu_coredump event is triggered in response to
a GuC context reset notification.
Using that flag, ensure all coredump functions that
read or print mmio-register values related to work submission
or command-streamer engines are skipped and replaced with
a calls guc-capture module equivalent functions to retrieve
or print the register dump.
While here, split out display related register reading
and printing into its own function that is called agnostic
to whether GuC had triggered the reset.
For now, introduce an empty printing function that can
filled in on a subsequent patch just to handle formatting.
Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220321164527.2500062-13-alan.previn.teres.alexis@intel.com
- Upon the G2H Notify-Err-Capture event, parse through the
GuC Log Buffer (error-capture-subregion) and generate one or
more capture-nodes. A single node represents a single "engine-
instance-capture-dump" and contains at least 3 register lists:
global, engine-class and engine-instance. An internal link
list is maintained to store one or more nodes.
- Because the link-list node generation happen before the call
to i915_gpu_codedump, duplicate global and engine-class register
lists for each engine-instance register dump if we find
dependent-engine resets in a engine-capture-group.
- When i915_gpu_coredump calls into capture_engine, (in a
subsequent patch) we detach the matching node (guc-id,
LRCA, etc) from the link list above and attach it to
i915_gpu_coredump's intel_engine_coredump structure when have
matching LRCA/guc-id/engine-instance.
Additional notes to be aware of:
- GuC generates the error capture dump into the GuC log buffer but
this buffer is one big log buffer with 3 independent subregions
within it. Each subregion is populated with different content
and used in different ways and timings but all regions operate
behave as independent ring buffers. Each guc-log subregion
(general-logs, crash-dump and error- capture) has it's own
guc_log_buffer_state that contain independent read and write
pointers.
Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220321164527.2500062-11-alan.previn.teres.alexis@intel.com
In the past we've always assumed that an RCS engine is present on every
platform. However now that we have compute engines there may be
platforms that have CCS engines but no RCS, or platforms that are
designed to have both, but have the RCS engine fused off.
Various engine-centric initialization that only needs to be done a
single time for the group of RCS+CCS engines can't rely on being setup
with the RCS now; instead we add a I915_ENGINE_FIRST_RENDER_COMPUTE flag
that will be assigned to a single engine in the group; whichever engine
has this flag will be responsible for some of the general setup
(RCU_MODE programming, initialization of certain workarounds, etc.).
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220303223435.2793124-1-matthew.d.roper@intel.com
The LRC descriptor pool is going away. Further, the function that was
populating it was also doing a bunch of logic about the context
registration sequence. So, split that code apart into separate state
setup and try to register functions. Note that some of those 'try to
register' code paths actually undo the state setup and leave it to be
redone again later (with potentially different values). This is
inefficient. The next patch will correct this.
Also, move a comment about ignoring return values to the place where
the return values are actually ignored.
v2: Move some more splitting from a later patch (and do it correctly).
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220302003357.4188363-5-John.C.Harrison@Intel.com
It is possible for reset notifications to arrive for a context that is
in the process of being banned. So don't flag these as an error, just
report it as informational (because it is still useful to know that
resets are happening even if they are being ignored).
v2: Better wording for the message (review feedback from Tvrtko).
v3: Fix rebase issue (review feedback from Daniele).
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220225015232.1939497-1-John.C.Harrison@Intel.com
A flag query helper was actually writing to the flags word rather than
just reading. Fix that. Also update the function's comment as it was
out of date.
NB: No need for a 'Fixes' tag. The test was only ever used inside a
BUG_ON during context registration. Rather than asserting that the
condition was true, it was making the condition true. So, in theory,
there was no consequence because we should never have hit a BUG_ON
anyway. Which means the write should always have been a no-op.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220217212942.629922-1-John.C.Harrison@Intel.com
UAPI Changes:
- Weak parallel submission support for execlists
Minimal implementation of the parallel submission support for
execlists backend that was previously only implemented for GuC.
Support one sibling non-virtual engine.
Core Changes:
- Two backmerges of drm/drm-next for header file renames/changes and
i915_regs reorganization
Driver Changes:
- Add new DG2 subplatform: DG2-G12 (Matt R)
- Add new DG2 workarounds (Matt R, Ram, Bruce)
- Handle pre-programmed WOPCM registers for DG2+ (Daniele)
- Update guc shim control programming on XeHP SDV+ (Daniele)
- Add RPL-S C0/D0 stepping information (Anusha)
- Improve GuC ADS initialization to work on ARM64 on dGFX (Lucas)
- Fix KMD and GuC race on accessing PMU busyness (Umesh)
- Use PM timestamp instead of RING TIMESTAMP for reference in PMU with GuC (Umesh)
- Report error on invalid reset notification from GuC (John)
- Avoid WARN splat by holding RPM wakelock during PXP unbind (Juston)
- Fixes to parallel submission implementation (Matt B.)
- Improve GuC loading status check/error reports (John)
- Tweak TTM LRU priority hint selection (Matt A.)
- Align the plane_vma to min_page_size of stolen mem (Ram)
- Introduce vma resources and implement async unbinding (Thomas)
- Use struct vma_resource instead of struct vma_snapshot (Thomas)
- Return some TTM accel move errors instead of trying memcpy move (Thomas)
- Fix a race between vma / object destruction and unbinding (Thomas)
- Remove short-term pins from execbuf (Maarten)
- Update to GuC version 69.0.3 (John, Michal Wa.)
- Improvements to GT reset paths in GuC backend (Matt B.)
- Use shrinker_release_pages instead of writeback in shmem object hooks (Matt A., Tvrtko)
- Use trylock instead of blocking lock when freeing GEM objects (Maarten)
- Allocate intel_engine_coredump_alloc with ALLOW_FAIL (Matt B.)
- Fixes to object unmapping and purging (Matt A)
- Check for wedged device in GuC backend (John)
- Avoid lockdep splat by locking dpt_obj around set_cache_level (Maarten)
- Allow dead vm to unbind vma's without lock (Maarten)
- s/engine->i915/i915/ for DG2 engine workarounds (Matt R)
- Use to_gt() helper for GGTT accesses (Michal Wi.)
- Selftest improvements (Matt B., Thomas, Ram)
- Coding style and compiler warning fixes (Matt B., Jasmine, Andi, Colin, Gustavo, Dan)
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Yg4i2aCZvvee5Eai@jlahtine-mobl.ger.corp.intel.com
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Fixed conflicts while applying, using the fixups/drm-intel-gt-next.patch
from drm-rerere's 1f2b1742abdd ("2022y-02m-23d-16h-07m-57s UTC: drm-tip
rerere cache update")]
Catch-up with 5.17-rc2 and trying to align with drm-intel-gt-next
for a possible topic branch for merging the split of i915_regs...
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
GuC updates shared memory and KMD reads it. Since this is not
synchronized, we run into a race where the value read is inconsistent.
Sometimes the inconsistency is in reading the upper MSB bytes of the
last_switch_in value. 2 types of cases are seen - upper 8 bits are zero
and upper 24 bits are zero. Since these are non-zero values, it is
not trivial to determine validity of these values. Instead we read the
values multiple times until they are consistent. In test runs, 3
attempts results in consistent values. The upper bound is set to 6
attempts and may need to be tuned as per any new occurences.
Since the duration that gt is parked can vary, the patch also updates
the gt timestamp on unpark before starting the worker.
v2:
- Initialize i
- Use READ_ONCE to access engine record
Fixes: 77cdd054dd ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu")
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220125020124.788679-2-umesh.nerlige.ramappa@intel.com
Update to the latest GuC release.
The latest GuC firmware introduces a number of interface changes:
GuC may return NO_RESPONSE_RETRY message for requests sent over CTB.
Add support for this reply and try resending the request again as a
new CTB message.
A KLV (key-length-value) mechanism is now used for passing
configuration data such as CTB management.
With the new KLV scheme, the old CTB management actions are no longer
used and are removed.
Register capture on hang is now supported by GuC. Full i915 support
for this will be added by a later patch. A minimum support of
providing capture memory and register lists is required though, so add
that in.
The device id of the current platform needs to be provided at init time.
The 'poll CS' w/a (Wa_22012773006) was blanket enabled by previous
versions of GuC. It must now be explicitly requested by the KMD. So,
add in the code to turn it on when relevant.
The GuC log entry format has changed. This requires adding a new field
to the log header structure to mark the wrap point at the end of the
buffer (as the buffer size is no longer a multiple of the log entry
size).
New CTB notification messages are now sent for some things that were
previously only sent via MMIO notifications.
Of these, the crash dump notification was not really being handled by
i915. It called the log flush code but that only flushed the regular
debug log and then only if relay logging was enabled. So just report
an error message instead.
The 'exception' notification was just being ignored completely. So add
an error message for that as well.
Note that in either the crash dump or the exception case, the GuC is
basically dead. The KMD will detect this via the heartbeat and trigger
both an error log (which will include the crash dump as part of the
GuC log) and a GT reset. So no other processing is really required.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220107000622.292081-3-John.C.Harrison@Intel.com
A weak implementation of parallel submission (multi-bb execbuf IOCTL) for
execlists. Doing as little as possible to support this interface for
execlists - basically just passing submit fences between each request
generated and virtual engines are not allowed. This is on par with what
is there for the existing (hopefully soon deprecated) bonding interface.
We perma-pin these execlists contexts to align with GuC implementation.
v2:
(John Harrison)
- Drop siblings array as num_siblings must be 1
v3:
(John Harrison)
- Drop single submission
v4:
(John Harrison)
- Actually drop single submission
- Use IS_ERR check on return value from intel_context_create
- Set last request to NULL on unpin
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211222223532.28698-1-matthew.brost@intel.com
If GuC encounters an error during engine reset, the i915 driver
promotes to full GT reset. This includes an info message about why the
reset is happening. However, that is not treated as a failure by any
of the CI systems because resets are an expected occurrance during
testing. This kind of failure is a major problem and should never
happen. So, complain more loudly and make sure CI notices.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211211065859.2248188-4-John.C.Harrison@Intel.com
