Commit Graph

64521 Commits

Author SHA1 Message Date
taosang2 352932738a SWDEV-476125 - Fix synchronization tests
Fix issues of
    Unit_cache_coherency_cpu_gpu
    Unit_cache_coherency_gpu_gpu
Enable them for all devices.

Change-Id: I19ba2084fdddd7b173edddb4e9c1b16cf7a97314
2024-09-24 14:39:57 -04:00
Ajay 98a1b3b42a SWDEV-486816 - RenderOpDispatch usage in pal client
Change-Id: I11cae3e625b287b998c9500c547efdacf1034a2b


[ROCm/clr commit: 7a288ea8bf]
2024-09-24 14:28:16 -04:00
Ajay 7a288ea8bf SWDEV-486816 - RenderOpDispatch usage in pal client
Change-Id: I11cae3e625b287b998c9500c547efdacf1034a2b
2024-09-24 14:28:16 -04:00
Jonathan Kim 027af8dacd hsakmt: Update hsa capabilities with precise ALU ops
Update the HSA capabilities field with precise ALU ops bit support
for GPU debugging.

Change-Id: I796f2c2e0559577828aba510c401ed5187e10179
2024-09-24 13:45:04 -04:00
Jonathan Kim 4cd7f3b675 hsakmt: Update hsa capabilities with precise ALU ops
Update the HSA capabilities field with precise ALU ops bit support
for GPU debugging.

Change-Id: I796f2c2e0559577828aba510c401ed5187e10179


[ROCm/ROCR-Runtime commit: 027af8dacd]
2024-09-24 13:45:04 -04:00
Rahul Manocha 51620a1bb6 [SWDEV-467733] - Set Cache Config Negative Param test fix
Change-Id: I058ca23a05e63aaaa54ea22e2b509593764cd7bc


[ROCm/hip-tests commit: 4e99f68b03]
2024-09-24 12:26:35 -04:00
Rahul Manocha 4e99f68b03 [SWDEV-467733] - Set Cache Config Negative Param test fix
Change-Id: I058ca23a05e63aaaa54ea22e2b509593764cd7bc
2024-09-24 12:26:35 -04:00
Julia Jiang e96fc07482 SWDEV-485943 - Adding note for some semophore related HIP APIs NOT supported on Linux.
Change-Id: I6fb57d5014cdd6e1f8eb56ca5326b460a76033de


[ROCm/hip commit: d4f6a6672e]
2024-09-24 12:00:50 -04:00
Julia Jiang d4f6a6672e SWDEV-485943 - Adding note for some semophore related HIP APIs NOT supported on Linux.
Change-Id: I6fb57d5014cdd6e1f8eb56ca5326b460a76033de
2024-09-24 12:00:50 -04:00
Alex Xu 6cc8f7ec13 rocminfo fails when amdgpu is built into the kernel
When amdgpu is built into the kernel, /sys/module/amdgpu/initstate will not be created even when the driver is functional. However, test shows /sys/module/amdgpu will be present. Adding an additional check for /sys/module/amdgpu when /sys/module/amdgpu/instate is not present.

Change-Id: Ie5c67c7e1eff8ac1683b211aaec802d0d342aeeb


[ROCm/rocminfo commit: c6f7a17c85]
2024-09-24 10:59:22 -04:00
Alex Xu c6f7a17c85 rocminfo fails when amdgpu is built into the kernel
When amdgpu is built into the kernel, /sys/module/amdgpu/initstate will not be created even when the driver is functional. However, test shows /sys/module/amdgpu will be present. Adding an additional check for /sys/module/amdgpu when /sys/module/amdgpu/instate is not present.

Change-Id: Ie5c67c7e1eff8ac1683b211aaec802d0d342aeeb
2024-09-24 10:59:22 -04:00
Branislav Brzak b5ee5efbaa SWDEV-483315 - Add hipGraphNodeGetDependentNodes_v2
Change-Id: I1ac0fb50df248521dfaae283bf9017acce146279


[ROCm/hipother commit: f50835c1fd]
2024-09-24 12:23:16 +02:00
Branislav Brzak f50835c1fd SWDEV-483315 - Add hipGraphNodeGetDependentNodes_v2
Change-Id: I1ac0fb50df248521dfaae283bf9017acce146279
2024-09-24 12:23:16 +02:00
Marko Arandjelovic 0ad2a9912a SWDEV-486448 - Disable the coalesced group tests due to too long execution
Change-Id: I4e91034c812319cef15c540078064e70ab5b5da6


[ROCm/hip-tests commit: 1621b7ad85]
2024-09-24 05:13:00 -04:00
Marko Arandjelovic 1621b7ad85 SWDEV-486448 - Disable the coalesced group tests due to too long execution
Change-Id: I4e91034c812319cef15c540078064e70ab5b5da6
2024-09-24 05:13:00 -04:00
Jonathan Kim a926a070ee hsakmt: Update thunk doc comments for debug firmware support caps
Update commentary on HWS scheduler support bit for GPU debugging in
the HSA capabilities node properties field.

Change-Id: I59c519d74a528d5ecf5817ef94e75091314bd844
2024-09-23 16:34:31 -04:00
Jonathan Kim a5954a2442 hsakmt: Update thunk doc comments for debug firmware support caps
Update commentary on HWS scheduler support bit for GPU debugging in
the HSA capabilities node properties field.

Change-Id: I59c519d74a528d5ecf5817ef94e75091314bd844


[ROCm/ROCR-Runtime commit: a926a070ee]
2024-09-23 16:34:31 -04:00
Ranjith Ramakrishnan b1e16e76ea SWDEV-437189 - Provide option to enable/disable CPACK_SET_DESTDIR
The variable is set as cache, so that user can override.
But the hard coded setting is preventing override. Removed the same

Change-Id: Ica5f9074a6532f465705a8dd1a8f21c6643ef49e


[ROCm/hip-tests commit: 173abcfba0]
2024-09-23 15:46:24 -04:00
Ranjith Ramakrishnan 173abcfba0 SWDEV-437189 - Provide option to enable/disable CPACK_SET_DESTDIR
The variable is set as cache, so that user can override.
But the hard coded setting is preventing override. Removed the same

Change-Id: Ica5f9074a6532f465705a8dd1a8f21c6643ef49e
2024-09-23 15:46:24 -04:00
Shweta Khatri 681610937a hsakmt: queues.c: Fix Coverity reported warnings
Move variable declarations inline and add NULL checks to prevent errors

Change-Id: Ia5bf5e245bcc0f756a15bc799b55c5e2a8459f89
2024-09-23 15:07:28 -04:00
Shweta Khatri fb7b04e06e hsakmt: queues.c: Fix Coverity reported warnings
Move variable declarations inline and add NULL checks to prevent errors

Change-Id: Ia5bf5e245bcc0f756a15bc799b55c5e2a8459f89


[ROCm/ROCR-Runtime commit: 681610937a]
2024-09-23 15:07:28 -04:00
Harkirat Gill 8596afd618 Add cstdint header for gcc-15 compatibility
Common C++ headers (like <memory>) in GCC 15.0.0 (combined with libstdc++) don't transitively include uint64_t anymore.

Minimal reproducer: https://godbolt.org/z/dqGbnG8bY

Porting: https://github.com/ROCm/rocm_smi_lib/pull/198
Closes: https://github.com/ROCm/rocm_smi_lib/issues/191

Change-Id: I2786e968c107a78104c43c4c474b7f65eaf88c0a


[ROCm/rocm_smi_lib commit: c61ab4fa28]
2024-09-23 15:05:07 -04:00
Harkirat Gill c61ab4fa28 Add cstdint header for gcc-15 compatibility
Common C++ headers (like <memory>) in GCC 15.0.0 (combined with libstdc++) don't transitively include uint64_t anymore.

Minimal reproducer: https://godbolt.org/z/dqGbnG8bY

Porting: https://github.com/ROCm/rocm_smi_lib/pull/198
Closes: https://github.com/ROCm/rocm_smi_lib/issues/191

Change-Id: I2786e968c107a78104c43c4c474b7f65eaf88c0a
2024-09-23 15:05:07 -04:00
Longlong Yao 3b829d0e62 rocrtst: fix resource leak
Change-Id: Ib57dccad0b639539e1076daba31eef278f2cf638
Signed-off-by: Longlong Yao <Longlong.Yao@amd.com>
2024-09-23 15:04:43 -04:00
Longlong Yao 745d799f9b rocrtst: fix resource leak
Change-Id: Ib57dccad0b639539e1076daba31eef278f2cf638
Signed-off-by: Longlong Yao <Longlong.Yao@amd.com>


[ROCm/ROCR-Runtime commit: 3b829d0e62]
2024-09-23 15:04:43 -04:00
Marko Arandjelovic a52b3a0e52 SWDEV-465148 - Enable Unit_hipMemcpyAsync_Positive_Synchronization_Behavior
- For transfers from any host memory to any host memory,
   the function is fully synchronous with respect to the host.

Change-Id: Ie1cb3e136334af320d2596460faed58fcb714b83


[ROCm/hip-tests commit: e35cc19aaf]
2024-09-23 14:55:23 -04:00
Marko Arandjelovic e35cc19aaf SWDEV-465148 - Enable Unit_hipMemcpyAsync_Positive_Synchronization_Behavior
- For transfers from any host memory to any host memory,
   the function is fully synchronous with respect to the host.

Change-Id: Ie1cb3e136334af320d2596460faed58fcb714b83
2024-09-23 14:55:23 -04:00
Young Hui b530e0f619 docs: move .readthedocs.yaml to the root of repo
Change-Id: I6c5afb806c47c029359a2dee2a7e73c6d076cfb1
2024-09-23 15:49:19 +00:00
Young Hui ea8cf386f4 docs: move .readthedocs.yaml to the root of repo
Change-Id: I6c5afb806c47c029359a2dee2a7e73c6d076cfb1


[ROCm/ROCR-Runtime commit: b530e0f619]
2024-09-23 15:49:19 +00:00
Young Hui 75b674f0ad docs: path adjustments to allow documentation to build again
- adding doc files to .gitignore

Change-Id: Ia6b2358bb1f236298ad1d705c1bed0636026632d
2024-09-23 15:49:13 +00:00
Young Hui e2b14cc3ec docs: path adjustments to allow documentation to build again
- adding doc files to .gitignore

Change-Id: Ia6b2358bb1f236298ad1d705c1bed0636026632d


[ROCm/ROCR-Runtime commit: 75b674f0ad]
2024-09-23 15:49:13 +00:00
pghafari 1157776b30 SWDEV-479260,SWDEV-483599 - Check griddim Y,Z <= 65536
Gfx12 has 16 bits for grid dim Y/Z. Detect gfxIp and return error if dim y/z > 16 bits

Change-Id: I43dd14affc9e4073d0b1232e7523967f0180fa31


[ROCm/clr commit: 0a918c8f96]
2024-09-23 11:36:13 -04:00
pghafari 0a918c8f96 SWDEV-479260,SWDEV-483599 - Check griddim Y,Z <= 65536
Gfx12 has 16 bits for grid dim Y/Z. Detect gfxIp and return error if dim y/z > 16 bits

Change-Id: I43dd14affc9e4073d0b1232e7523967f0180fa31
2024-09-23 11:36:13 -04:00
Shweta Khatri 857200e28c hsakmt: events.c: Fix Coverity reported warnings
Fix data race by protecting events_page access with mutex in event create
Fix potential NULL dereference in hsaKmtWaitOnMultipleEvents_Ext
Fix unchecked return value in hsaKmtCreateEvent function

Change-Id: I434bef43666e5205a8b061259569c1d99a952752
2024-09-23 11:35:02 -04:00
Shweta Khatri 5dff20b660 hsakmt: events.c: Fix Coverity reported warnings
Fix data race by protecting events_page access with mutex in event create
Fix potential NULL dereference in hsaKmtWaitOnMultipleEvents_Ext
Fix unchecked return value in hsaKmtCreateEvent function

Change-Id: I434bef43666e5205a8b061259569c1d99a952752


[ROCm/ROCR-Runtime commit: 857200e28c]
2024-09-23 11:35:02 -04:00
Shweta Khatri 303c02690d hsakmt: hsakmttypes.h: Fix Coverity reported warnings
Eliminated declared but not referenced variables to fix warnings

Change-Id: I80032a699fb59ce4635c5001f669d009ba60e588
2024-09-23 11:34:44 -04:00
Shweta Khatri b4db09e269 hsakmt: hsakmttypes.h: Fix Coverity reported warnings
Eliminated declared but not referenced variables to fix warnings

Change-Id: I80032a699fb59ce4635c5001f669d009ba60e588


[ROCm/ROCR-Runtime commit: 303c02690d]
2024-09-23 11:34:44 -04:00
Shweta Khatri 659fa04d8c hsakmt: topology.c: Fix Coverity reported warnings
Refactor fscanf_str to use fgets for safer string handling, remove unused code

Change-Id: Ibf4b4b485f99bf2fabfe48e9609ca99111feaf1e
2024-09-23 11:34:28 -04:00
Shweta Khatri fa0978e19a hsakmt: topology.c: Fix Coverity reported warnings
Refactor fscanf_str to use fgets for safer string handling, remove unused code

Change-Id: Ibf4b4b485f99bf2fabfe48e9609ca99111feaf1e


[ROCm/ROCR-Runtime commit: 659fa04d8c]
2024-09-23 11:34:28 -04:00
Jatin Chaudhary 0a850d8725 SWDEV-466747 - call device sync once while unregistering
Basically embed hipDeviceSync in std::call_once.

Change-Id: I29ca926d61ed80e21acba5c388a8256d913487e4


[ROCm/clr commit: f8beeede22]
2024-09-23 08:00:10 -04:00
Jatin Chaudhary f8beeede22 SWDEV-466747 - call device sync once while unregistering
Basically embed hipDeviceSync in std::call_once.

Change-Id: I29ca926d61ed80e21acba5c388a8256d913487e4
2024-09-23 08:00:10 -04:00
German Andryeyev f8fc11c2d8 SWDEV-483586 - Unblock staging H2D transfers
Although unpinned copies require synchronizations
in HIP, runtime can avoid syncs for H2D copies with
a staging buffer

Change-Id: If2203c6bc0cbd89742823688dc8e89e9acd873b2


[ROCm/clr commit: 29cc678d8d]
2024-09-21 10:25:27 -04:00
German Andryeyev 29cc678d8d SWDEV-483586 - Unblock staging H2D transfers
Although unpinned copies require synchronizations
in HIP, runtime can avoid syncs for H2D copies with
a staging buffer

Change-Id: If2203c6bc0cbd89742823688dc8e89e9acd873b2
2024-09-21 10:25:27 -04:00
Yiannis Papadopoulos 48fdc17179 rocr/aie: Correct reporting of dev heap size
Storing the correct dev heap size in the memory region.

Change-Id: I14b053330c187da1d7d0213256625e50795b9902
2024-09-20 12:44:23 -04:00
Yiannis Papadopoulos 14b3b04250 rocr/aie: Correct reporting of dev heap size
Storing the correct dev heap size in the memory region.

Change-Id: I14b053330c187da1d7d0213256625e50795b9902


[ROCm/ROCR-Runtime commit: 48fdc17179]
2024-09-20 12:44:23 -04:00
Maneesh Gupta e87bf2096d SWDEV-485179 - Revert "SWDEV-459254 - Overwrite cacheline size to 256 for gfx12, as it is used for kernarg alignment."
This reverts commit 1f63650bf96e01e48f879aa58b80e2130dd4a567.

Reason for revert: <INSERT REASONING HERE>

Change-Id: I6d7ed87c09d9b77116548dce1f30ac4711c2c09d


[ROCm/clr commit: 2d1c6ee23e]
2024-09-20 11:33:34 -04:00
Maneesh Gupta 2d1c6ee23e SWDEV-485179 - Revert "SWDEV-459254 - Overwrite cacheline size to 256 for gfx12, as it is used for kernarg alignment."
This reverts commit 1f63650bf96e01e48f879aa58b80e2130dd4a567.

Reason for revert: <INSERT REASONING HERE>

Change-Id: I6d7ed87c09d9b77116548dce1f30ac4711c2c09d
2024-09-20 11:33:34 -04:00
Kent Russell daad183bf8 hsakmt: Undo HSAKMT prefix for PAGE_SHIFT
We had skipped doing it for PAGE_SIZE, but it should be left as the
regular PAGE_SHIFT name, especially for users who are using different
headers. We want PAGE_SHIFT and PAGE_SIZE to be consistent with one
another, so set them both explicitly to the same value if either
of them is undefined

Change-Id: I121d81c48409dd77351b59a192d824e2419a2410
Signed-off-by: Kent Russell <kent.russell@amd.com>
2024-09-20 11:04:34 -04:00
Kent Russell 4901f1a528 hsakmt: Undo HSAKMT prefix for PAGE_SHIFT
We had skipped doing it for PAGE_SIZE, but it should be left as the
regular PAGE_SHIFT name, especially for users who are using different
headers. We want PAGE_SHIFT and PAGE_SIZE to be consistent with one
another, so set them both explicitly to the same value if either
of them is undefined

Change-Id: I121d81c48409dd77351b59a192d824e2419a2410
Signed-off-by: Kent Russell <kent.russell@amd.com>


[ROCm/ROCR-Runtime commit: daad183bf8]
2024-09-20 11:04:34 -04:00
Anusha GodavarthySurya f278ac9ea0 SWDEV-485904 - Fix virtual,physical mem obj leaks
Change-Id: Ie0456b5dcfec206ae54a6aabfc2a15a620cac693


[ROCm/clr commit: 870842201d]
2024-09-19 23:04:20 -04:00