Commit Graph

68 Commits (764ed8b91632276bc0d77461a62316b3bcb0cccf)

Author SHA1 Message Date
Pierre-Eric Pelloux-Prayer 118addfaf8 amdgpu: fix use-after-free
Closes: https://gitlab.freedesktop.org/mesa/drm/-/issues/96
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2024-01-04 11:00:37 +01:00
Marek Olšák 98e1db5011 amdgpu: add an environment variable that overrides the context priority
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
2023-05-11 12:53:29 -04:00
Samuel Pitoiset de84cdc563 amdgpu: implement new CTX OP to set/get stable pstates
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2022-01-20 18:32:46 +01:00
Huang Rui a42eb3dfde amdgpu: clean up the cs structure variable
This patch is to use generic variables as the input of amdgpu_cs_submit_raw2.
Because amdgpu_cs_submit_one won't handle IOCTL directly.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2020-02-12 19:55:22 +08:00
Huang Rui c483259248 amdgpu: remove the un-used chunk_array
This array won't be used.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2020-02-12 19:55:16 +08:00
Huang Rui cf9eb78a89 amdgpu: use amdgpu_cs_submit_raw2 in amdgpu_cs_submit
So far, amdgpu_cs_submit_raw2 is mainly used for upper layer (Mesa), however,
amdgpu_cs_submit is used for current all unit tests. Our intention is that the
unit tests can actually verify the API which is really used.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2020-02-12 19:55:09 +08:00
Huang Rui fca0849af3 amdgpu: use alloca for dependencies and sem_dependencies
Use alloca instead of malloc, then we don't need free them at the end of this
function.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2020-02-12 19:52:37 +08:00
Eric Engestrom 074947ee4b meson: always define whether headers exist
Combined with -Wundef (added in 75758d2ccf & enforced in ba17673eed),
this provides absolute safety against #ifdef typos.

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
2020-01-23 17:00:12 +00:00
Chunming Zhou 0a7ad7df14 libdrm: wrap new flexible syncobj query interface v2
v2: nit-picks fix

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Christian König <Christian.Koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
For the xf86drm.[ch] part : Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
2019-10-26 03:11:14 +00:00
Marek Olšák 00320d7d68 amdgpu: add amdgpu_cs_query_reset_state2 for AMDGPU_CTX_OP_QUERY_STATE2
This is a better GPU reset query.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
2019-10-15 15:20:38 -04:00
Michel Dänzer e246114c46 amdgpu: Move union declaration to top of amdgpu_cs_ctx_override_priority
Avoids compiler warning:

../../amdgpu/amdgpu_cs.c: In function 'amdgpu_cs_ctx_override_priority':
../../amdgpu/amdgpu_cs.c:155:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
  union drm_amdgpu_sched args;
  ^~~~~

Reviewed-by: Christian König <christian.koenig@amd.com>
2019-06-25 17:58:31 +02:00
Chunming Zhou 6a72661c33 wrap transfer interfaces
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
2019-05-16 12:07:18 +02:00
Chunming Zhou 6bb5cc174b expose timeline signal/export/import interfaces v2
v2: adapt to new one transfer ioctl

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
2019-05-16 12:07:18 +02:00
Chunming Zhou 46f930d962 wrap syncobj timeline query/wait APIs for amdgpu v3
v2: symbos are stored in lexical order.
v3: drop export/import and extra query indirection

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
2019-05-16 12:07:18 +02:00
John Stultz 51553dd0e0 libdrm: amdgpu: Initialize unions with memset rather than "= {0}"
Clang complains when initializing unions using "= {0}"
so instead use memset.

Cc: Sean Paul <seanpaul@chromium.org>
Cc: Alistair Strachan <astrachan@google.com>
Cc: Marissa Wall <marissaw@google.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
2019-04-25 10:57:26 +01:00
Bas Nieuwenhuizen 232dc3305d amdgpu: Add context priority override function.
This way we can override the priority of a single context using a
master fd.

Since we cannot usefully create an amdgpu device of a master fd
without the fd deduplication kicking in this takes a plain fd.

This can be used by e.g. radv to get high priority contexts using
a master fd from the primary node or a lease.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Andres Rodriguez <andresx7@gmail.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2019-04-18 10:39:32 +02:00
Marek Olšák f19afaa519 amdgpu: add a faster BO list API
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
2019-01-16 16:39:25 -05:00
Lucas De Marchi e15e3a65b8 amdgpu: annotate public functions
This was done with:
nm --dynamic --defined-only build/amdgpu/libdrm_amdgpu.so | \
	grep amdgpu_ | \
	cut -d' ' -f3 > /tmp/a.txt

while read sym; do
	read f func line _ <<<$(cscope -d -L -1 $sym)
	if [ ! -z "$f" ]; then
		line=$((line-1))
		sed -i "${line}s/^/drm_public /" $f
	fi
done < /tmp/a.txt

Then the alignment of function arguments were manually fixed all over.
The idea here will be to switch the default visibility to hidden so we
don't export symbols we shouldn't.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2018-09-19 22:46:45 -07:00
Eric Engestrom 0926f0af54 meson,configure: include config.h automatically
This will prevent any more missing `#include "config.h"` bug, at the
cost of having to recompile some files that didn't need to be when
changing build options.

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
2018-03-20 18:19:26 +00:00
Sabre Shao a5329cd990 drm/amdgpu: Remove IB count checking
Signed-off-by: Sabre Shao <Sabre.Shao@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2018-03-08 20:06:28 -05:00
Eric Engestrom 089d45456b remove unnecessary double-semicolon
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
2018-01-29 15:41:52 +00:00
Bas Nieuwenhuizen 5889f6ba1a amdgpu: Add syncobj reset & signal wrappers.
Signed-off-by: Dave Airlie <airlied@redhat.com>
2017-12-18 09:51:50 +10:00
David Mao 1749d56ea0 amdgpu: Adding amdgpu_cs_create_syncobj2 to create syncobj as signaled initially
Signed-off-by: David Mao <david.mao@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-11-29 22:05:25 -05:00
Andres Rodriguez 35bc82cee9 amdgpu: implement context priority for amdgpu_cs_ctx_create2 v3
Add a new context creation function that allows specifying the context
priority.

A high priority context has the potential of starving lower priority
contexts. The current kernel driver implementation allows only apps
that hold CAP_SYS_NICE or DRM_MASTER to acquire a priority above
AMDGPU_CTX_PRIORITY_NORMAL.

v2: corresponding changes for kernel patch v2
v3: Fixed 'make check' symbol error

Signed-off-by: Andres Rodriguez <andresx7@gmail.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-10-20 21:51:02 +02:00
Marek Olšák c74d461404 amdgpu: add amdgpu_cs_fence_to_handle
v2: update amdgpu-symbol-check

Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2017-10-12 20:35:57 +02:00
Marek Olšák 59aa57b363 amdgpu: add amdgpu_cs_syncobj_wait
v2: update amdgpu-symbol-check

Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2017-10-12 20:35:57 +02:00
Marek Olšák b6e24501a1 amdgpu: add sync_file import and export functions
v2: update amdgpu-symbol-check

Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2017-10-12 20:35:57 +02:00
Dave Airlie 22790a65d4 drm/amdgpu: add new low overhead command submission API. (v2)
This just sends chunks to the kernel API for a single command
stream.

This should provide a more future proof and extensible API
for command submission.

v2: use amdgpu_bo_list_handle, add two helper functions to
access bo and context internals.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2017-07-19 00:33:14 +01:00
Dave Airlie 69532d0188 drm/amdgpu: add syncobj create/destroy/import/export apis
These are just wrappers using the amdgpu device handle.

Acked-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2017-07-19 00:32:53 +01:00
Edward O'Callaghan 00aa2c1881 amdgpu: Use the canonical form in branch predicate
Suggested-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
2017-05-03 21:36:02 +10:00
Edward O'Callaghan 7cfcd5ef4b amdgpu/: concisely && consistently check null ptrs in canonical form
Be consistent and use the canonical form while sanity checking
null pointers, also combine a few branches for brevity.

v2: rebase on top of 'add amdgpu_cs_wait_fences' series.

Signed-off-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
2017-04-19 18:08:10 +10:00
Nicolai Hähnle d8d45a4938 amdgpu: add the interface of waiting multiple fences
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
[v2: allow returning the first signaled fence index]
Signed-off-by: monk.liu <Monk.Liu@amd.com>
[v3:
 - cleanup *status setting
 - fix amdgpu symbols check
v4: simplify return from amdgpu_cs_wait_fences (suggested
    by Edward O'Callaghan)]
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com> (v1)
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com> (v1)
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2017-04-18 21:05:00 +02:00
Ken Wang f884af9b57 amdgpu: fix for submition with no ibs
Avoid a crash if no IBs are specified.

Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-04 12:26:43 -05:00
Junwei Zhang 6b79c66b84 amdgpu: list each entry safely for sw semaphore when submit ib
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: David Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2016-01-20 02:14:46 +01:00
Marek Olšák 6afadeaf13 amdgpu: add semaphore support
the semaphore is a binary semaphore. the work flow is:
1. create sem
2. signal sem
3. wait sem, reset sem after signalled
4. destroy sem.

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2016-01-20 02:14:46 +01:00
Tom St Denis c05049b420 amdgpu: Make amdgpu_cs_calculate_timeout() return something sensible on error
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-20 19:21:48 -05:00
Jammy Zhou 9c15b4aacd amdgpu: fix overflow for timeout calculation
Set the timeout to AMDGPU_TIMEOUT_INFINITE when overflow happens

Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-20 19:21:48 -05:00
Alan Coopersmith 94ecdcb8b1 Include <alloca.h> when needed before calling alloca
Fixes "error: implicit declaration of function 'alloca'" failures
when building on Solaris

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2015-09-12 09:41:24 -07:00
Christian König f6f25d67a9 amdgpu: remove sequence mutex
It's not used any more.

Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
2015-08-25 10:26:58 -04:00
Emil Velikov bddf4df4a1 amdgpu: hide the final internal functions from global namespace
Thus the only symbols that we export are the ones officially provided by
the API.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:36 +01:00
Emil Velikov a30da8e9b3 amdgpu: cosmetic chances in license boilerplate
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:27 +01:00
Emil Velikov f4c2bfd63e amdgpu: add a bunch of missing config.h includes
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:24 +01:00
Michel Dänzer 2978f5c37e amdgpu: Remove unused local variables bo_size/offset from amdgpu_cs_submit
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-05 13:50:38 -04:00
Ken Wang 926c805686 amdgpu : move management of user fence from libdrm to UMD
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-05 13:47:52 -04:00
Michel Dänzer 95ecf91c7b amdgpu: Use drmIoctl in amdgpu_ioctl_wait_cs
This is safe now because the ioctl uses an absolute timeout.

This prevents amdgpu_cs_query_fence_status from returning early e.g.
when a signal is delivered, which in turn caused Mesa winsys code to
assume a BO was idle when it actually wasn't yet.

Reviewed-by: Christian König <christian.koenig@amd.com>
2015-08-05 13:47:52 -04:00
Christian König 5463d2e83a amdgpu: use common fence structure for dependencies as well.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-05 13:47:52 -04:00
Jammy Zhou f91b56dc8c amdgpu: improve the amdgpu_cs_query_fence_status interface
make amdgpu_cs_query_fence reusable to support multi-fence query

Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-08-05 13:47:52 -04:00
Marek Olšák 8097d08ee4 amdgpu: check the user fence only if the IP supports user fences
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-05 13:47:52 -04:00
Marek Olšák 67c994f057 drm/amdgpu: allow passing absolute timeouts to amdgpu_cs_query_fence_status
Useful when Mesa wants to wait for a lot of fences at the same time and
doesn't want to recalculate the relative timeout after every call.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-08-05 13:47:52 -04:00
Christian König 0f37bc9029 amdgpu: add CS dependencies v2
This allows the driver to specify on which previous CS to wait.

v2: fix spelling in comment

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2015-08-05 13:47:51 -04:00