amdgpu: Add wrappers for AMDGPU_VM IOCTL.
PATCH 1/2 amdgpu: Add wrappers for AMDGPU_VM IOCTL.
--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu.h | 20 +++++++++++++++++++ amdgpu/amdgpu_vm.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+) create mode 100644 amdgpu/amdgpu_vm.c
amdgpu/Makefile.sources +1 -0
@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \
amdgpu/amdgpu.h +20 -0
@@ -1489,6 +1489,26 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_alloc_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); + +/** +* Free reserved VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_free_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); + #ifdef __cplusplus }#endif
amdgpu/amdgpu_vm.c new file +52 -0
52 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
A message that is not in the archive (sent off-list or not yet received): 1509063355-5205-1-git-send-email-Andrey.Grodzovsky-5C7GfCeVMHo@public.gmane.org
PATCH 2/2 amdgpu: Add VMID reservation per GPU context test.
The test will Reserve a VMID, submit a command and unreserve the VMID.
--- tests/amdgpu/Makefile.am | 3 +- tests/amdgpu/amdgpu_test.c | 7 +++ tests/amdgpu/amdgpu_test.h | 15 +++++ tests/amdgpu/vm_tests.c | 151 +++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 175 insertions(+), 1 deletion(-) create mode 100644 tests/amdgpu/vm_tests.c
tests/amdgpu/Makefile.am +2 -1
@@ -31,4 +31,5 @@ amdgpu_test_SOURCES = \ uvd_enc_tests.c \ vcn_tests.c \ uve_ib.h \ - deadlock_tests.c + deadlock_tests.c \ + vm_tests.c
tests/amdgpu/amdgpu_test.c +7 -0
@@ -103,6 +103,13 @@ static CU_SuiteInfo suites[] = { .pCleanupFunc = suite_deadlock_tests_clean, .pTests = deadlock_tests, }, + { + .pName = "VM Tests", + .pInitFunc = suite_vm_tests_init, + .pCleanupFunc = suite_vm_tests_clean, + .pTests = vm_tests, + }, + CU_SUITE_INFO_NULL, };
tests/amdgpu/amdgpu_test.h +15 -0
@@ -150,6 +150,21 @@ int suite_deadlock_tests_clean();extern CU_TestInfo deadlock_tests[];/** +* Initialize vm test suite +*/ +int suite_vm_tests_init (); + +/** +* Deinitialize deadlock test suite +*/ +int suite_vm_tests_clean (); + +/** +* Tests in vm test suite +*/ +extern CU_TestInfo vm_tests[]; + +/** * Helper functions */ static inline amdgpu_bo_handle gpu_mem_alloc(
tests/amdgpu/vm_tests.c new file +151 -0
151 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Am 27.10.2017 um 02:15 schrieb Andrey Grodzovsky:
Quoted 79 lines
Change-Id I7eafb85c1ca96d6d255f0183bed0ce4129746fe0Signed-off-by Andrey Grodzovsky <[email protected]>--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu.h | 20 +++++++++++++++++++ amdgpu/amdgpu_vm.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+)create mode 100644
amdgpu/amdgpu_vm.camdgpu/Makefile.sources +1
-0@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \amdgpu/amdgpu.h +20
-0@@ -1489,6 +1489,26 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_alloc_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); + +/** +* Free reserved VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_free_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); + #ifdef __cplusplus }#endif amdgpu/amdgpu_vm.c new file +21
-0@@ -0,0 +1,52 @@ +/* +* Copyright 2017 Advanced Micro Devices, Inc. +* +* Permission is hereby granted, free of charge, to any person obtaining a +* copy of this software and associated documentation files (the "Software"), +* to deal in the Software without restriction, including without limitation +* the rights to use, copy, modify, merge, publish, distribute, sublicense, +* and/or sell copies of the Software, and to permit persons to whom the +* Software is furnished to do so, subject to the following conditions: +* +* The above copyright notice and this permission notice shall be included in +* all copies or substantial portions of the Software. +* +* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL +* THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR +* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, +* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR +* OTHER DEALINGS IN THE SOFTWARE. +*
amdgpu/amdgpu_vm.c line 31
Quoted 10 lines
amdgpu/amdgpu_vm.c new file +10
-0+*/ + + +#include "amdgpu.h" +#include "amdgpu_drm.h" +#include "xf86drm.h" +#include "amdgpu_internal.h" + +int amdgpu_vm_alloc_reserved_vmid (amdgpu_context_handle context, +uint32_t flags)
Could just be the mail client, but flags looks way to much indented to the right.
Additional to that I would keep the name of the kernel IOCTL for the function name.
E.g. amdgpu_vm_reserve_vmid() and amdgpu_vm_unreserve_vmid().
Apart from that both patches look good to me.
Christian.
Quoted 11 lines
amdgpu/amdgpu_vm.c new file +11
-0+{ +union drm_amdgpu_vm vm; + + vm.in.op =AMDGPU_VM_OP_RESERVE_VMID ; + vm.in.flags = flags; + +return drmCommandWriteRead (context->dev->fd,DRM_AMDGPU_VM , + &vm,sizeof (vm)); +} + +int amdgpu_vm_free_reserved_vmid(amdgpu_context_handle context,
amdgpu/amdgpu_vm.c line 52
Quoted 10 lines
amdgpu/amdgpu_vm.c new file +10
-0+ uint32_t flags) +{ +union drm_amdgpu_vm vm; + + vm.in.op =AMDGPU_VM_OP_UNRESERVE_VMID ; + vm.in.flags = flags; + +return drmCommandWriteRead (context->dev->fd,DRM_AMDGPU_VM , + &vm,sizeof (vm)); +}
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Quoted 39 lines · Andrey Grodzovsky
Change-Id I7eafb85c1ca96d6d255f0183bed0ce4129746fe0Signed-off-by Andrey Grodzovsky <[email protected]>--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu.h | 20 +++++++++++++++++++ amdgpu/amdgpu_vm.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+) create mode 100644 amdgpu/amdgpu_vm.camdgpu/Makefile.sources +1
-0@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \amdgpu/amdgpu.h +10
-0@@ -1489,6 +1489,26 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_alloc_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); +
amdgpu/amdgpu.h void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence, · line 1511
Quoted 10 lines
amdgpu/amdgpu.h +10
-0+/** +* Free reserved VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_free_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); +
Andrey, don't forget to run make check. You'll see it flag a lovely error ;-)
-Emil
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Quoted 44 lines · Emil Velikov
Quoted 42 lines · Andrey Grodzovsky
Change-Id I7eafb85c1ca96d6d255f0183bed0ce4129746fe0Signed-off-by Andrey Grodzovsky <[email protected]>--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu.h | 20 +++++++++++++++++++ amdgpu/amdgpu_vm.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+)create mode 100644
amdgpu/amdgpu_vm.camdgpu/Makefile.sources +1
-0@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \amdgpu/amdgpu.h +13
-0@@ -1489,6 +1489,26 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_alloc_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); + +/** +* Free reserved VMID +* \param context - \c [in] GPU Context
Quoted 8 lines
amdgpu/amdgpu.h void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence, · line 1511
Quoted 7 lines
amdgpu/amdgpu.h +7
-0+* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_free_reserved_vmid (amdgpu_context_handle context, +uint32_t flags); +Andrey, don't forget to run make check. You'll see it flag a lovely error ;-)
Thanks, didn't know about the check option, will fix.
Andrey
Quoted 2 lines
-Emil
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH v2 1/2 amdgpu: Add wrappers for AMDGPU_VM IOCTL.
In reply to Emil VelikovFrom: Andrey Grodzovsky <[email protected]>
v2:
Rename wrappers to match the IOCTL naming, fix
identation and fix make check error.
--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu-symbol-check | 2 ++ amdgpu/amdgpu.h | 18 +++++++++++++++++ amdgpu/amdgpu_vm.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 71 insertions(+) create mode 100644 amdgpu/amdgpu_vm.c
amdgpu/Makefile.sources +1 -0
@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \
amdgpu/amdgpu-symbol-check +2 -0
@@ -64,6 +64,8 @@ amdgpu_read_mm_registers amdgpu_va_range_alloc amdgpu_va_range_free amdgpu_va_range_query +amdgpu_vm_reserve_vmid +amdgpu_vm_unreserve_vmid EOF done)
amdgpu/amdgpu.h +18 -0
@@ -1489,6 +1489,24 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_reserve_vmid (amdgpu_context_handle context,uint32_t flags); + +/** +* Free reserved VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_unreserve_vmid (amdgpu_context_handle context,uint32_t flags); + #ifdef __cplusplus }#endif
amdgpu/amdgpu_vm.c new file +50 -0
50 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
A message that is not in the archive (sent off-list or not yet received): 1509116951-1250-1-git-send-email-andrey.grodzovsky-5C7GfCeVMHo@public.gmane.org
PATCH v2 2/2 amdgpu: Add VMID reservation per GPU context test.
From: Andrey Grodzovsky <[email protected]>
The test will Reserve a VMID, submit a command and unreserve the VMID.
v2:
Wrappers names were changed.
--- tests/amdgpu/Makefile.am | 3 +- tests/amdgpu/amdgpu_test.c | 7 +++ tests/amdgpu/amdgpu_test.h | 15 +++++ tests/amdgpu/vm_tests.c | 151 +++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 175 insertions(+), 1 deletion(-) create mode 100644 tests/amdgpu/vm_tests.c
tests/amdgpu/Makefile.am +2 -1
@@ -31,4 +31,5 @@ amdgpu_test_SOURCES = \ uvd_enc_tests.c \ vcn_tests.c \ uve_ib.h \ - deadlock_tests.c + deadlock_tests.c \ + vm_tests.c
tests/amdgpu/amdgpu_test.c +7 -0
@@ -103,6 +103,13 @@ static CU_SuiteInfo suites[] = { .pCleanupFunc = suite_deadlock_tests_clean, .pTests = deadlock_tests, }, + { + .pName = "VM Tests", + .pInitFunc = suite_vm_tests_init, + .pCleanupFunc = suite_vm_tests_clean, + .pTests = vm_tests, + }, + CU_SUITE_INFO_NULL, };
tests/amdgpu/amdgpu_test.h +15 -0
@@ -150,6 +150,21 @@ int suite_deadlock_tests_clean();extern CU_TestInfo deadlock_tests[];/** +* Initialize vm test suite +*/ +int suite_vm_tests_init (); + +/** +* Deinitialize deadlock test suite +*/ +int suite_vm_tests_clean (); + +/** +* Tests in vm test suite +*/ +extern CU_TestInfo vm_tests[]; + +/** * Helper functions */ static inline amdgpu_bo_handle gpu_mem_alloc(
tests/amdgpu/vm_tests.c new file +151 -0
151 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Am 27.10.2017 um 17:09 schrieb Andrey Grodzovsky:
Quoted 7 lines
From: Andrey Grodzovsky <[email protected]>
v2:
Rename wrappers to match the IOCTL naming, fix
identation and fix make check error.Signed-off-by Andrey Grodzovsky <[email protected]>
Quoted 108 lines
--- amdgpu/Makefile.sources | 1 + amdgpu/amdgpu-symbol-check | 2 ++ amdgpu/amdgpu.h | 18 +++++++++++++++++ amdgpu/amdgpu_vm.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 71 insertions(+)create mode 100644
amdgpu/amdgpu_vm.camdgpu/Makefile.sources +1
-0@@ -6,6 +6,7 @@ LIBDRM_AMDGPU_FILES := \ amdgpu_gpu_info.c \ amdgpu_internal.h \ amdgpu_vamgr.c \ + amdgpu_vm.c \ util_hash.c \ util_hash.h \ util_hash_table.c \amdgpu/amdgpu-symbol-check +2
-0@@ -64,6 +64,8 @@ amdgpu_read_mm_registers amdgpu_va_range_alloc amdgpu_va_range_free amdgpu_va_range_query +amdgpu_vm_reserve_vmid +amdgpu_vm_unreserve_vmid EOF done)amdgpu/amdgpu.h +18
-0@@ -1489,6 +1489,24 @@ void amdgpu_cs_chunk_fence_to_dep(struct amdgpu_cs_fence *fence,void amdgpu_cs_chunk_fence_info_to_data (struct amdgpu_cs_fence_info *fence_info,struct drm_amdgpu_cs_chunk_data *data); +/** +* Reserve VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_reserve_vmid (amdgpu_context_handle context,uint32_t flags); + +/** +* Free reserved VMID +* \param context - \c [in] GPU Context +* \param flags - \c [in] TBD +* +* \return 0 on success otherwise POSIX Error code +*/ +int amdgpu_vm_unreserve_vmid (amdgpu_context_handle context,uint32_t flags); + #ifdef __cplusplus }#endif amdgpu/amdgpu_vm.c new file +40
-0@@ -0,0 +1,50 @@ +/* +* Copyright 2017 Advanced Micro Devices, Inc. +* +* Permission is hereby granted, free of charge, to any person obtaining a +* copy of this software and associated documentation files (the "Software"), +* to deal in the Software without restriction, including without limitation +* the rights to use, copy, modify, merge, publish, distribute, sublicense, +* and/or sell copies of the Software, and to permit persons to whom the +* Software is furnished to do so, subject to the following conditions: +* +* The above copyright notice and this permission notice shall be included in +* all copies or substantial portions of the Software. +* +* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL +* THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR +* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, +* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR +* OTHER DEALINGS IN THE SOFTWARE. +* +*/ + + +#include "amdgpu.h" +#include "amdgpu_drm.h" +#include "xf86drm.h" +#include "amdgpu_internal.h" + +int amdgpu_vm_reserve_vmid (amdgpu_context_handle context,uint32_t flags) +{ +union drm_amdgpu_vm vm; + + vm.in.op =AMDGPU_VM_OP_RESERVE_VMID ; + vm.in.flags = flags; + +return drmCommandWriteRead (context->dev->fd,DRM_AMDGPU_VM , + &vm,sizeof (vm)); +} +
amdgpu/amdgpu_vm.c line 50
Quoted 10 lines
amdgpu/amdgpu_vm.c new file +10
-0+int amdgpu_vm_unreserve_vmid (amdgpu_context_handle context,uint32_t flags) +{ +union drm_amdgpu_vm vm; + + vm.in.op =AMDGPU_VM_OP_UNRESERVE_VMID ; + vm.in.flags = flags; + +return drmCommandWriteRead (context->dev->fd,DRM_AMDGPU_VM , + &vm,sizeof (vm)); +}
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel