nv-tegra-mirror/linux-nvgpu
1
0
Fork 0
mirror of git://nv-tegra.nvidia.com/linux-nvgpu.git synced 2025-12-23 01:50:07 +03:00
Code Issues Packages Projects Releases Wiki Activity
Files
bf29bd942201e17f55ff0894833a84b8f6c0bf97
linux-nvgpu/drivers/gpu/nvgpu/gk20a
History
Deepak Nibade bf29bd9422 gpu: nvgpu: check engine ctx_status in wait_idle 
We have following bug where GPU Host reports non-idle
when it should report engine idle

- if a context is preempted off the GPU, and there is
  no other context to load, NV_PGRAPH_ENGINE_STATUS
  will not be idle until new context is loaded

- this could cause gr_gk20a_wait_idle() to fail since
  here we rely only on NV_PGRAPH_ENGINE_STATUS to
  decide if engine is busy or not

To fix this, first check if context is valid or not
from NV_PFIFO_ENGINE_STATUS_CTX_STATUS

If context is invalid, return immediately
Otherwise, continue as before

Also, add accessors for invalid ctx_status

Bug 1826768

Change-Id: Id627be3f02e79f4beac59a8b5195d08eabf651f2
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1237521
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
2016-10-19 10:49:04 -07:00
..
as_gk20a.c
gpu: nvgpu: Support multiple types of allocators
2016-07-19 11:21:46 -07:00
as_gk20a.h
gpu: nvgpu: User-space managed address space support
2015-11-18 09:45:07 -08:00
bitmap_allocator_priv.h
gpu: nvgpu: bitmap allocator optimization
2016-10-18 12:24:26 -07:00
buddy_allocator_priv.h
gpu: nvgpu: Implement carveout support in the buddy allocator
2016-09-20 14:56:39 -07:00
cde_gk20a.c
gpu: nvgpu: Suppress error msg from VBIOS overlay
2016-10-09 13:03:35 -07:00
cde_gk20a.h
gpu: nvgpu: Use device instead of platform_device
2016-04-08 09:42:41 -07:00
ce2_gk20a.c
gpu: nvgpu: unify nvgpu and pci probe
2016-09-08 09:43:51 -07:00
ce2_gk20a.h
gpu: nvgpu: Add nvgpu infra to allow kernel to create privileged CE channels
2016-07-20 03:09:28 -07:00
channel_gk20a.c
gpu: nvgpu: userd allocation from sysmem
2016-10-11 09:16:03 -07:00
channel_gk20a.h
gpu: nvgpu: fix use-after-free in case of error notifier
2016-10-10 10:13:56 -07:00
channel_sync_gk20a.c
gpu: nvgpu: Fix invalid test for signaled sync_fences
2016-09-15 21:58:44 -07:00
channel_sync_gk20a.h
gpu: nvgpu: implement sync refcounting
2016-04-19 08:16:13 -07:00
clk_gk20a.c
gpu: nvgpu: Use device instead of platform_device
2016-04-08 09:42:41 -07:00
clk_gk20a.h
gpu: nvgpu: add debugfs to dump clocks
2016-09-29 13:17:52 -07:00
css_gr_gk20a.c
gpu: nvgpu: vgpu: cyclestat snapshot support
2016-09-08 16:04:09 -07:00
css_gr_gk20a.h
gpu: nvgpu: vgpu: cyclestat snapshot support
2016-09-08 16:04:09 -07:00
ctrl_gk20a.c
gpu: nvgpu: add ioctl for querying memory state
2016-10-14 08:12:34 -07:00
ctrl_gk20a.h
gpu: nvgpu: rename gpu ioctls and structs to nvgpu
2015-03-18 12:11:33 -07:00
ctxsw_trace_gk20a.c
gpu: nvgpu: process granularity for FECS traces
2016-07-18 23:19:09 -07:00
ctxsw_trace_gk20a.h
gpu: nvgpu: Add checks for tracing enabled
2016-05-23 14:48:02 -07:00
dbg_gpu_gk20a.c
gpu: nvgpu: Skip calling undefined prod callbacks
2016-09-08 20:05:49 -07:00
dbg_gpu_gk20a.h
gpu: nvgpu: move dbg_session_ops to gops
2016-07-30 11:29:20 -07:00
debug_gk20a.c
gpu: nvgpu: unify nvgpu and pci probe
2016-09-08 09:43:51 -07:00
debug_gk20a.h
gpu: nvgpu: unify nvgpu and pci probe
2016-09-08 09:43:51 -07:00
fb_gk20a.c
nvgpu: gk20a: include proper header file
2015-07-15 21:54:16 -07:00
fb_gk20a.h
gpu: nvgpu: Reset FB also in gm20b
2015-07-03 19:00:22 -07:00
fecs_trace_gk20a.c
gpu: nvgpu: do not flush FECS record on engine reset
2016-08-29 16:14:40 -07:00
fecs_trace_gk20a.h
gpu: nvgpu: Add support for FECS ctxsw tracing
2016-03-23 07:48:47 -07:00
fence_gk20a.c
gpu: nvgpu: Optimize sync fence creation
2016-09-15 21:58:36 -07:00
fence_gk20a.h
gpu: nvgpu: Optimize sync fence creation
2016-09-15 21:58:36 -07:00
fifo_gk20a.c
gpu: nvgpu: userd allocation from sysmem
2016-10-11 09:16:03 -07:00
fifo_gk20a.h
gpu: nvgpu: userd allocation from sysmem
2016-10-11 09:16:03 -07:00
gk20a_allocator_bitmap.c
gpu: nvgpu: bitmap allocator optimization
2016-10-18 12:24:26 -07:00
gk20a_allocator_buddy.c
gpu: nvgpu: add space query in page and buddy allocs
2016-10-14 08:12:27 -07:00
gk20a_allocator_lockless.c
gpu: nvgpu: implement lockless allocator
2016-09-20 10:43:40 -07:00
gk20a_allocator_page.c
gpu: nvgpu: Avoid 64 bit division
2016-10-18 12:24:41 -07:00
gk20a_allocator.c
gpu: nvgpu: support free space query in allocators
2016-10-14 08:12:26 -07:00
gk20a_allocator.h
gpu: nvgpu: support free space query in allocators
2016-10-14 08:12:26 -07:00
gk20a_gating_reglist.c
gpu: nvgpu: gk20a: update slcg prod values
2015-09-21 16:41:55 -07:00
gk20a_gating_reglist.h
gpu: nvgpu: gm20b: Regenerate clock gating lists
2015-03-18 12:11:09 -07:00
gk20a_scale.c
gpu: nvgpu: unify nvgpu and pci probe
2016-09-08 09:43:51 -07:00
gk20a_scale.h
gpu: nvgpu: add QoS notifier for common clk framework
2016-06-27 09:14:04 -07:00
gk20a_sysfs.c
gpu: nvgpu: support suspend/resume with user disabled railgating
2016-10-18 11:08:01 -07:00
gk20a.c
gpu: nvgpu: implement PCIe Gen2 frequency swap
2016-10-18 12:24:25 -07:00
gk20a.h
gpu: nvgpu: implement PCIe Gen2 frequency swap
2016-10-18 12:24:25 -07:00
gr_ctx_gk20a_sim.c
gpu: nvgpu: program sw veid bundles
2016-10-14 08:11:19 -07:00
gr_ctx_gk20a.c
gpu: nvgpu: program sw veid bundles
2016-10-14 08:11:19 -07:00
gr_ctx_gk20a.h
gpu: nvgpu: Full chip support for ctxsw
2016-07-22 15:10:22 -07:00
gr_gk20a.c
gpu: nvgpu: check engine ctx_status in wait_idle
2016-10-19 10:49:04 -07:00
gr_gk20a.h
gpu: nvgpu: add func ptr for gpc exceptions
2016-10-17 14:46:44 -07:00
gr_pri_gk20a.h
gpu: nvgpu: Full chip support for ctxsw
2016-07-22 15:10:22 -07:00
hal_gk20a.c
gpu: nvgpu: Add dGPU clocks support
2016-09-29 13:17:46 -07:00
hal_gk20a.h
gpu: nvgpu: Add class numbers to characteristics
2015-03-18 12:12:06 -07:00
hal.c
gpu: nvgpu: Support third GPU version on T18x
2016-06-13 07:39:56 -07:00
hal.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_bus_gk20a.h
gpu: nvgpu: add PRAMIN support for mem accessors
2016-05-24 12:39:06 -07:00
hw_ccsr_gk20a.h
gpu: nvgpu: Use sysmem aperture for SoC memory
2016-04-15 08:50:34 -07:00
hw_ce2_gk20a.h
gpu: nvgpu: enable ce2 interrupts
2015-04-04 18:02:35 -07:00
hw_ctxsw_prog_gk20a.h
gpu: nvgpu: add accessors for global_esr values and sm_dbgr_control
2016-04-19 08:08:07 -07:00
hw_fb_gk20a.h
gpu: nvgpu: Program NISO sysmem flush addr
2016-04-25 08:16:10 -07:00
hw_fifo_gk20a.h
gpu: nvgpu: check engine ctx_status in wait_idle
2016-10-19 10:49:04 -07:00
hw_flush_gk20a.h
gpu: nvgpu: Do not use G_ELPG_FLUSH
2015-11-10 10:33:39 -08:00
hw_gmmu_gk20a.h
gpu: ngpu: add support for vidmem in page tables
2016-07-04 23:12:29 -07:00
hw_gr_gk20a.h
gpu: nvgpu: select target based on aperture
2016-09-15 12:24:31 -07:00
hw_ltc_gk20a.h
gpu: nvgpu: hwpm broadcast register support
2016-05-19 15:58:24 -07:00
hw_mc_gk20a.h
gpu: nvgpu: Ensure memory subsystem is enabled
2015-04-04 18:08:03 -07:00
hw_pbdma_gk20a.h
gpu: nvgpu: Use sysmem aperture for SoC memory
2016-04-15 08:50:34 -07:00
hw_perf_gk20a.h
gpu: nvgpu: cyclestats mode E snapshots support
2015-06-06 07:23:24 -07:00
hw_pram_gk20a.h
gpu: nvgpu: add PRAMIN support for mem accessors
2016-05-24 12:39:06 -07:00
hw_pri_ringmaster_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_pri_ringstation_fbp_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_pri_ringstation_gpc_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_pri_ringstation_sys_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_proj_gk20a.h
gpu: nvgpu: Full chip support for ctxsw
2016-07-22 15:10:22 -07:00
hw_pwr_gk20a.h
gpu: nvgpu: Use sysmem aperture for SoC memory
2016-04-15 08:50:34 -07:00
hw_ram_gk20a.h
gpu: nvgpu: Use sysmem aperture for SoC memory
2016-04-15 08:50:34 -07:00
hw_sim_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
hw_therm_gk20a.h
gpu: nvgpu: add support for therm gate ctrl
2016-01-27 09:53:31 -08:00
hw_timer_gk20a.h
gpu: nvgpu: Add support for FECS ctxsw tracing
2016-03-23 07:48:47 -07:00
hw_top_gk20a.h
gpu: nvgpu: Add multiple engine and runlist support
2016-06-07 12:31:34 -07:00
hw_trim_gk20a.h
gpu: nvgpu: Sync gk20a and gm20b headers
2015-03-18 12:11:52 -07:00
kind_gk20a.c
gpu: nvgpu: Use noinline_for_stack to avoid GCOV build break
2015-03-18 12:10:41 -07:00
kind_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
lockless_allocator_priv.h
gpu: nvgpu: implement lockless allocator
2016-09-20 10:43:40 -07:00
ltc_common.c
gpu: nvgpu: Add gpu_dbg_map_v message type
2016-08-30 10:04:23 -07:00
ltc_gk20a.c
gpu: nvgpu: add check for is_fmodel
2016-07-27 14:32:54 -07:00
ltc_gk20a.h
gpu: nvgpu: Add class numbers to characteristics
2015-03-18 12:12:06 -07:00
mc_gk20a.c
gpu: nvgpu: Add multiple engine and runlist support
2016-06-07 12:31:34 -07:00
mc_gk20a.h
gpu: nvgpu: Allow enabling/disabling MC interrupt
2015-03-18 12:12:31 -07:00
mm_gk20a.c
gpu: nvgpu: SLAB allocation for page allocator
2016-10-18 12:24:33 -07:00
mm_gk20a.h
gpu: nvgpu: allow skipping pramin barriers
2016-10-17 14:39:00 -07:00
page_allocator_priv.h
gpu: nvgpu: SLAB allocation for page allocator
2016-10-18 12:24:33 -07:00
platform_gk20a_generic.c
gpu: nvgpu: Use device instead of platform_device
2016-04-08 09:42:41 -07:00
platform_gk20a_tegra.c
gpu: nvgpu: gm20b (un)railgate common clock support
2016-10-17 02:09:28 -07:00
platform_gk20a.h
gpu: nvgpu: support suspend/resume with user disabled railgating
2016-10-18 11:08:01 -07:00
platform_vgpu_tegra.c
gpu: nvgpu: improve sync create/destroy logic
2016-09-20 10:43:35 -07:00
pmu_api.h
gpu: nvgpu: Add dGPU clocks support
2016-09-29 13:17:46 -07:00
pmu_common.h
gpu: nvgpu: Add dGPU clocks support
2016-09-29 13:17:46 -07:00
pmu_gk20a.c
gpu: nvgpu: Suppress error msg from VBIOS overlay
2016-10-09 13:03:35 -07:00
pmu_gk20a.h
gpu: nvgpu: Suppress error msg from VBIOS overlay
2016-10-09 13:03:35 -07:00
priv_ring_gk20a.c
gpu: nvgpu: add check for is_fmodel
2016-07-27 14:32:54 -07:00
priv_ring_gk20a.h
gpu: nvgpu: Add NVIDIA GPU Driver
2015-03-18 12:08:53 -07:00
regops_gk20a.c
gpu: nvgpu: move dbg_session_ops to gops
2016-07-30 11:29:20 -07:00
regops_gk20a.h
gpu: nvgpu: create new nvgpu ioctl header
2015-03-18 12:11:33 -07:00
sched_gk20a.c
gpu: nvgpu: add ref counting for GPU sched ctrl
2016-07-18 23:17:55 -07:00
sched_gk20a.h
gpu: nvgpu: add ref counting for GPU sched ctrl
2016-07-18 23:17:55 -07:00
semaphore_gk20a.c
gpu: nvgpu: Fix error handling of __semaphore_bitmap_alloc()
2016-08-31 14:04:12 -07:00
semaphore_gk20a.h
gpu: nvgpu: Add debugging to the semaphore code
2016-08-30 10:04:30 -07:00
sim_gk20a.h
gpu: nvgpu: Support >32bit addresses in simulation
2015-06-01 08:16:28 -07:00
sync_gk20a.c
gpu: nvgpu: Greatly simplify the semaphore detection
2016-09-15 21:58:37 -07:00
sync_gk20a.h
gpu: nvgpu: Greatly simplify the semaphore detection
2016-09-15 21:58:37 -07:00
therm_gk20a.c
gpu: nvgpu: gp10x: Add debugfs entry for temperature reading
2016-10-11 13:44:45 -07:00
therm_gk20a.h
gpu: nvgpu: update thermal programming
2015-10-27 15:42:08 -07:00
tsg_gk20a.c
gpu: nvgpu: add ref counting for GPU sched ctrl
2016-07-18 23:17:55 -07:00
tsg_gk20a.h
gpu: nvgpu: add ref counting for GPU sched ctrl
2016-07-18 23:17:55 -07:00