aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
AgeCommit message (Collapse)AuthorFilesLines
2017-11-01drm/amdgpu: Implement amdgpu SDMA functions for VIPhilip Cox2-11/+78
Signed-off-by: Philip Cox <[email protected]> Signed-off-by: shaoyun liu <[email protected]> Signed-off-by: Yong Zhao <[email protected]> Signed-off-by: Jay Cornwall <[email protected]> Signed-off-by: Felix Kuehling <[email protected]> Acked-by: Oded Gabbay <[email protected]> Signed-off-by: Oded Gabbay <[email protected]>
2017-11-01drm/amdgpu: Add support for resuming SDMA queues w/o HWSFelix Kuehling2-14/+23
Save wptr in hqd_sdma_destroy, restore it in hqd_sdma_load. Also read updated wptr from user mode when resuming an SDMA queue. Signed-off-by: Jay Cornwall <[email protected]> Signed-off-by: Yong Zhao <[email protected]> Signed-off-by: Felix Kuehling <[email protected]> Acked-by: Oded Gabbay <[email protected]> Signed-off-by: Oded Gabbay <[email protected]>
2017-11-01drm/amd: Update kgd_kfd interface for resuming SDMA queuesFelix Kuehling2-4/+8
Add wptr and mm parameters to hqd_sdma_load and pass these parameters from device_queue_manager through the mqd_manager. SDMA doesn't support polling while the engine believes it's idle. The driver must update the wptr. The new parameters will be used for looking up the updated value from the specified mm when SDMA queues are resumed after being disabled. Signed-off-by: Felix Kuehling <[email protected]> Reviewed-by: Oded Gabbay <[email protected]> Signed-off-by: Oded Gabbay <[email protected]>
2017-10-31drm/radeon: fix atombios on big endianRoman Kapl1-20/+18
The function for byteswapping the data send to/from atombios was buggy for num_bytes not divisible by four. The function must be aware of the fact that after byte-swapping the u32 units, valid bytes might end up after the num_bytes boundary. This patch was tested on kernel 3.12 and allowed us to sucesfully use DisplayPort on and Radeon SI card. Namely it fixed the link training and EDID readout. The function is patched both in radeon and amd drivers, since the functions and the fixes are identical. Signed-off-by: Roman Kapl <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2017-10-31drm/amdgpu: fix error handling in amdgpu_bo_do_createChristian König1-3/+3
The bo structure is freed up in case of an error, so we can't do any accounting if that happens. Signed-off-by: Christian König <[email protected]> Reviewed-by: Michel Dänzer <[email protected]> CC: [email protected] Signed-off-by: Alex Deucher <[email protected]>
2017-10-25drm/amd/amdgpu: Remove workaround for suspend/resume in uvd7Tom St Denis1-11/+5
The workaround is not required anymor and would result in hangs during suspend/resume cycles if the uvd block were busy. Signed-off-by: Tom St Denis <[email protected]> Acked-by: Leo Liu <[email protected]> Acked-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-25drm/amdgpu: don't flush the TLB before initializing GARTChristian König1-6/+7
No point in doing this. Signed-off-by: Christian König <[email protected]> Reviewed-by: Michel Dänzer <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-25drm/amdgpu: minor cleanup for amdgpu_ttm_bindChristian König1-1/+2
Filter the placement mask before using it. In theory it could be that we have other flags set here as well. Signed-off-by: Christian König <[email protected]> Reviewed-by: Michel Dänzer <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-25drm/amdgpu/psp: prevent page fault by checking write_frame address(v4)Evan Quan2-4/+24
- Prevent a possible buffer overflow when updating the ring buffer by bounds checking the command frame against the available space in the ring buffer. v2: update the ring_buffer_end address v3: update the commit log v4: squash in print fix (Michel) Signed-off-by: Evan Quan <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-25drm/amd/amdgpu: Remove workaround check for UVD6 on APUsTom St Denis1-11/+5
On APUs the uvd6 driver was skipping proper suspend/resume routines resulting in a broken state upon resume. Signed-off-by: Tom St Denis <[email protected]> Acked-by: Alex Deucher <[email protected]> Reviewed-by: Leo Liu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-25locking/atomics: COCCINELLE/treewide: Convert trivial ACCESS_ONCE() patterns ↵Mark Rutland2-4/+4
to READ_ONCE()/WRITE_ONCE() Please do not apply this to mainline directly, instead please re-run the coccinelle script shown below and apply its output. For several reasons, it is desirable to use {READ,WRITE}_ONCE() in preference to ACCESS_ONCE(), and new code is expected to use one of the former. So far, there's been no reason to change most existing uses of ACCESS_ONCE(), as these aren't harmful, and changing them results in churn. However, for some features, the read/write distinction is critical to correct operation. To distinguish these cases, separate read/write accessors must be used. This patch migrates (most) remaining ACCESS_ONCE() instances to {READ,WRITE}_ONCE(), using the following coccinelle script: ---- // Convert trivial ACCESS_ONCE() uses to equivalent READ_ONCE() and // WRITE_ONCE() // $ make coccicheck COCCI=/home/mark/once.cocci SPFLAGS="--include-headers" MODE=patch virtual patch @ depends on patch @ expression E1, E2; @@ - ACCESS_ONCE(E1) = E2 + WRITE_ONCE(E1, E2) @ depends on patch @ expression E; @@ - ACCESS_ONCE(E) + READ_ONCE(E) ---- Signed-off-by: Mark Rutland <[email protected]> Signed-off-by: Paul E. McKenney <[email protected]> Cc: Linus Torvalds <[email protected]> Cc: Peter Zijlstra <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Link: http://lkml.kernel.org/r/[email protected] Signed-off-by: Ingo Molnar <[email protected]>
2017-10-20drm/amdgpu:fix wb_clearMonk Liu1-2/+2
Properly shift the index when clearing so we clear the right bit Signed-off-by: Monk Liu <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-20drm/amdgpu:fix vf_error_putMonk Liu2-11/+17
1,it should not work on non-SR-IOV case 2,the NO_VBIOS error is incorrect, should handle it under detect_sriov_bios. 3,wrap the whole detect_sriov_bios with sriov check Signed-off-by: Monk Liu <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-20drm/amdgpu/sriov:now must reinit pspMonk Liu1-0/+1
otherwise after VF FLR the KIQ cannot work Signed-off-by: Monk Liu <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-20drm/amdgpu: merge bios post checking functionspding1-23/+18
Merge the post checking functions to avoid confusion and take virtualization into account in all cases. Signed-off-by: pding <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-20Merge branch 'drm-next-4.15' of git://people.freedesktop.org/~agd5f/linux ↵Dave Airlie34-258/+1291
into drm-next Last set of features for 4.15. Highlights: - Add a bo flag to allow buffers to opt out of implicit sync - Add ctx priority setting interface - Lots more powerplay cleanups - Start to plumb through vram lost infrastructure for gpu reset - ttm support for huge pages - misc cleanups and bug fixes * 'drm-next-4.15' of git://people.freedesktop.org/~agd5f/linux: (73 commits) drm/amd/powerplay: Place the constant on the right side of the test drm/amd/powerplay: Remove useless variable drm/amd/powerplay: Don't cast kzalloc() return value drm/amdgpu: allow GTT overcommit during bind drm/amdgpu: linear validate first then bind to GART drm/amd/pp: Fix overflow when setup decf/pix/disp dpm table. drm/amd/pp: thermal control not enabled on vega10. drm/amdgpu: busywait KIQ register accessing (v4) drm/amdgpu: report more amdgpu_fence_info drm/amdgpu:don't check soft_reset for sriov drm/amdgpu:fix duplicated setting job's vram_lost drm/amdgpu:reduce wb to 512 slot drm/amdgpu: fix regresstion on SR-IOV gpu reset failed drm/amd/powerplay: Tidy up cz_dpm_powerup_vce() drm/amd/powerplay: Tidy up cz_dpm_powerdown_vce() drm/amd/powerplay: Tidy up cz_dpm_update_vce_dpm() drm/amd/powerplay: Tidy up cz_dpm_update_uvd_dpm() drm/amd/powerplay: Tidy up cz_dpm_powerup_uvd() drm/amd/powerplay: Tidy up cz_dpm_powerdown_uvd() drm/amd/powerplay: Tidy up cz_start_dpm() ...
2017-10-19drm/amdgpu: allow GTT overcommit during bindChristian König1-3/+5
While binding BOs to GART we need to allow a bit overcommit in the GTT domain. Otherwise we can never use the full GART space when GART size=GTT size. Signed-off-by: Christian König <[email protected]> Reviewed-by: Chunming Zhou <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: linear validate first then bind to GARTChristian König1-9/+9
For VM emulation for old UVD/VCE we need to validate the BO with linear VRAM flag set first and then eventually bind it to GART. Validating with linear VRAM flag set can move the BO to GART making UVD/VCE read/write from an unbound GART BO. Signed-off-by: Christian König <[email protected]> Reviewed-by: Alex Deucher <[email protected]> CC: [email protected] Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: busywait KIQ register accessing (v4)pding6-25/+71
Register accessing is performed when IRQ is disabled. Never sleep in this function. Known issue: dead sleep in many use cases of index/data registers. v2: - wrap polling fence functions. - don't trigger IRQ for polling in case of wrongly fence signal. v3: - handle wrap round gracefully. - add comments for polling function v4: - don't return negative timeout confused with error code Signed-off-by: pding <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: report more amdgpu_fence_infopding1-0/+13
Only for GFX ring. This can help checking MCBP feature. The fence at the end of the frame will indicate the completion status. If the frame completed normally, the fence is written to the address given in the EVENT_WRITE_EOP packet. If preemption occurred in the previous IB the address is adjusted by 2 DWs. If work submitted in the frame was reset before completion, the fence address is adjusted by four DWs. In the case that preemption occurred, and before preemption completed a reset was initiated, the address will be adjusted with six DWs Signed-off-by: pding <[email protected]> Reviewed-by: Monk Liu <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu:don't check soft_reset for sriovMonk Liu1-0/+3
Signed-off-by: Monk Liu <[email protected]> Ack-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu:fix duplicated setting job's vram_lostMonk Liu2-2/+1
Signed-off-by: Monk Liu <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu:reduce wb to 512 slotMonk Liu1-1/+1
with current WB usage we only use 57 slots, so 512 is extreamly sufficient, and reduce to 512 can make WB fit into one page. Signed-off-by: Monk Liu <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: fix regresstion on SR-IOV gpu reset failedRex Zhu3-8/+14
fw ucode is corrupted after vf flr by PSP so ucode_init() is a must in psp_hw_init othewise KIQ/KCQ enabling will fail Revert "drm/amdgpu: refine code delete duplicated error handling" This reverts commit e57b87ff828f95efe992468e6d18c2c059b27aa9. Revert "drm/amdgpu: move amdgpu_ucode_init_bo to amdgpu_device.c" This reverts commit 815b8f8595148d06a64d2ce4282e8e80dfcb02f1. Reviewed-by: Monk Liu <[email protected]> Signed-off-by: Rex Zhu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: use ARRAY_SIZEJérémy Lefaure2-8/+10
Using the ARRAY_SIZE macro improves the readability of the code. Found with Coccinelle with the following semantic patch: @r depends on (org || report)@ type T; T[] E; position p; @@ ( (sizeof(E)@p /sizeof(*E)) | (sizeof(E)@p /sizeof(E[...])) | (sizeof(E)@p /sizeof(T)) ) Reviewed-by: Christian König <[email protected]> Reviewed-by: Thierry Reding <[email protected]> Signed-off-by: Jérémy Lefaure <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: reserve root PD while releasing itChristian König1-2/+11
Otherwise somebody could try to evict it at the same time and try to use half torn down structures. Signed-off-by: Christian König <[email protected]> Reviewed-and-Tested-by: Michel Dänzer <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: bump version for vram lost counter query (v2)Alex Deucher1-1/+2
v2: vram -> VRAM in comment Reviewed-by: Michel Dänzer <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: minor CS optimizationChristian König1-18/+19
We only need to loop over all IBs for old UVD/VCE command stream patching. Signed-off-by: Christian König <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: Add amdgpu_find_mm_node()Harish Kasiviswanathan1-22/+27
Replace some commonly repeated code with a function. v2: Use amdgpu_find_mm_node() in amdgpu_ttm_io_mem_pfn() Signed-off-by: Harish Kasiviswanathan <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: Refactor amdgpu_move_blitHarish Kasiviswanathan2-49/+132
Add more generic function amdgpu_copy_ttm_mem_to_mem() that supports arbitrary copy size, offsets and two BOs (source & dest.). This is useful for KFD Cross Memory Attach feature where data needs to be copied from BOs from different processes v2: Add struct amdgpu_copy_mem and changed amdgpu_copy_ttm_mem_to_mem() function parameters to use the struct v3: Minor function name change Signed-off-by: Harish Kasiviswanathan <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: Fix extra call to amdgpu_ctx_put.Andrey Grodzovsky1-4/+2
In amdgpu_cs_parser_init() in case of error handling amdgpu_ctx_put() is called without setting p->ctx to NULL after that, later amdgpu_cs_parser_fini() also calls amdgpu_ctx_put() again and mess up the reference count. Signed-off-by: Andrey Grodzovsky <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: add VRAM lost queryChristian König1-0/+3
Allows userspace to figure out if VRAM was lost. Signed-off-by: Christian König <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: set -ECANCELED when dropping jobsChristian König2-1/+7
And return from the wait functions the fence error code. Signed-off-by: Christian König <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: move the VRAM lost counter per contextChristian König4-24/+8
Instead of per device track the VRAM lost per context and return ECANCELED instead of ENODEV. Signed-off-by: Christian König <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: keep copy of VRAM lost counter in jobChristian König3-6/+10
Instead of reading the current counter from fpriv. Signed-off-by: Christian König <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: partial revert VRAM lost handling v2Christian König3-20/+10
Keep blocking the CS, but revert everything else. Mapping BOs and info IOCTL are harmless and can still happen even when VRAM content ist lost. Signed-off-by: Christian König <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: correct reference clock value on vega10Ken Wang1-4/+1
Old value from bringup was wrong. Cc: [email protected] Signed-off-by: Ken Wang <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: SR-IOV data exchange between PF&VFHorace Chen4-0/+262
SR-IOV need to exchange some data between PF&VF through shared VRAM PF will copy some necessary firmware and information to the shared VRAM. It also requires some information from VF. PF will send a key through mailbox2 to help guest calculate checksum so that it can verify whether the data is correct. So check the data on the specified offset of the shared VRAM, if the checksum is right, read values from it and write some VF information next to the data from PF. Signed-off-by: Horace Chen <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: Move old fence waiting before reservation lock is aquired v2Andrey Grodzovsky3-10/+34
Helps avoiding deadlock during GPU reset. Added mutex to amdgpu_ctx to preserve order of fences on a ring. v2: Put waiting logic in a function in a seperate function in amdgpu_ctx.c Signed-off-by: Andrey Grodzovsky <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: Refactor amdgpu_cs_ib_vm_chunk and amdgpu_cs_ib_fill.Andrey Grodzovsky1-51/+61
This enables old fence waiting before reservation lock is aquired which in turn is part of a bigger solution to deadlock happening when gpu reset with VRAM recovery accures during intensive rendering. Signed-off-by: Andrey Grodzovsky <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: make function uvd_v6_0_enc_get_destroy_msg staticColin Ian King1-2/+3
The function uvd_v6_0_enc_get_destroy_msg is local to the source and does not need to be in global scope, so make it static. Cleans up sparse warning: symbol 'uvd_v6_0_enc_get_destroy_msg' was not declared. Should it be static? Reviewed-by: Christian König <[email protected]> Signed-off-by: Colin Ian King <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: refine code delete duplicated error handlingRex Zhu1-2/+2
in function amdgpu_ucode_init_bo, when failed, it will set load_type to AMDGPU_FW_LOAD_DIRECT. Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Rex Zhu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: bump version for new AMDGPU_SCHED ioctlAlex Deucher1-1/+2
Signed-off-by: Alex Deucher <[email protected]>
2017-10-19drm/amdgpu: rename context priority levelsAndres Rodriguez1-4/+4
Don't leak implementation details about how each priority behaves to usermode. This allows greater flexibility in the future. Squash into c2636dc53abd8269a0930bccd564f2f195dba729 Signed-off-by: Andres Rodriguez <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-18drm/amd/amdgpu: Added asic_type as ACP DMA driver platform dataVijendar Mukunda1-0/+2
asic_type information is passed to ACP DMA Driver as platform data. Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Vijendar Mukunda <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-17Merge remote-tracking branch 'pfdo/drm-next' into drm-nextDave Airlie2-10/+10
Pull in drm-next for the object find API changes. Fix the one place the API crashes. Signed-off-by: Dave Airlie <[email protected]>
2017-10-12drm/amdgpu: fix placement flags in amdgpu_ttm_bindChristian König1-1/+1
Otherwise we lose the NO_EVICT flag and can try to evict pinned BOs. Signed-off-by: Christian König <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-12drm: Pass struct drm_file * to __drm_mode_object_find [v2]Keith Packard2-10/+10
This will allow __drm_mode_object_file to be extended to perform access control checks based on the file in use. v2: Also fix up vboxvideo driver in staging [airlied: merging early as this is an API change] Suggested-by: Daniel Vetter <[email protected]> Signed-off-by: Keith Packard <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
2017-10-09drm/amdgpu: add interface for editing a foreign process's priority v3Andres Rodriguez5-21/+147
The AMDGPU_SCHED_OP_PROCESS_PRIORITY_OVERRIDE ioctls are used to set the priority of a different process in the current system. When a request is dropped, the process's contexts will be restored to the priority specified at context creation time. A request can be dropped by setting the override priority to AMDGPU_CTX_PRIORITY_UNSET. An fd is used to identify the remote process. This is simpler than passing a pid number, which is vulnerable to re-use, etc. This functionality is limited to DRM_MASTER since abuse of this interface can have a negative impact on the system's performance. v2: removed unused output structure v3: change refcounted interface for a regular set operation Signed-off-by: Andres Rodriguez <[email protected]> Acked-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2017-10-09drm/amdgpu: add plumbing for ctx priority changes v2Andres Rodriguez2-1/+34
Introduce amdgpu_ctx_priority_override(). A mechanism to override a context's priority. An override can be terminated by setting the override to AMD_SCHED_PRIORITY_UNSET. v2: change refcounted interface for a direct set Signed-off-by: Andres Rodriguez <[email protected]> Acked-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.