2
0
Gráfico de cometimentos

8920 Cometimentos

Autor(a) SHA1 Mensagem Data
David Yat Sin ec759c7995 Add rocrtst to Query agent memory available
Add new test for GPU agents memory available

Change-Id: Ib07e2003a21659b99732b535cd004081635d6aa1
Signed-off-by: David Yat Sin <david.yatsin@amd.com>
2022-08-11 09:36:58 -04:00
David Yat Sin 86e4cb1ddd Add max enum value to hsa_agent_info_t
Add max enum value to force size of enum and avoid clang compile
warnings.

Change-Id: I9cdf529517cc605a5039c3a924fd718ece16029d
2022-08-10 11:11:36 -04:00
Ranjith Ramakrishnan c3ea96dd71 SWDEV-350674 - Added backward compatibility for binary files and rdc.service
With file reorganization changes binaries are moved to /opt/rocm-ver/bin.
Similarly rdc.service moved to /opt/rocm-ver/libexec/rdc
Test suites still used old paths
Once test suites changes are made, backward compatibility for binaries and rdc.service can be removed
Corrcted binary path in rdc.service.in
Corrected GRPC runpath

Change-Id: I306924d81cedc19586305a79d51eea8af6e70e83
2022-08-09 17:45:22 -07:00
Ammar ELWazir 591db0b718 Changing NULL to nullptr (Tracer Tool)
Change-Id: I567bf7944599922e5d402e55142c2915ae24fb69
2022-08-08 20:45:34 -05:00
Jonathan R. Madsen 23fb3946c7 Handle --advanced printing for config generation (#137)
Handle --advanced printing for config
2022-08-08 15:28:34 -05:00
Jonathan R. Madsen 040da3fc6a Enable TRACE_THREAD_RW_LOCKS and TRACE_THREAD_SPIN_LOCKS by default (#136)
* Enable OMNITRACE_TRACE_THREAD_{RW,SPIN}_LOCKS by default

- updates timemory submodule with updated GOTCHA submodule
- fix to GOTCHA library which defaults to not wrapping dlopen and dlsym prevents deadlock

* Bump version to 1.4.0
2022-08-08 15:28:01 -05:00
Jonathan R. Madsen 34013bc539 OMNITRACE_TRACE_THREAD_SPIN_LOCKS config (#134)
- configuration setting to wrap pthread_spin_lock, pthread_spin_trylock, pthread_spin_unlock
2022-08-08 08:38:52 -05:00
Jonathan R. Madsen 95913c7135 Update python install + build-tree setup (#128)
- When one python version is used, install to proper lib/pythonX.Y/site-packages
- config files, etc. in build tree resemble the install-tree
2022-08-08 08:38:38 -05:00
Jonathan R. Madsen cbe8862ea4 Verbose messages based on ROCP_ONLOAD_TRACE env (#131)
Debugging based on ROCP_ONLOAD_TRACE env
2022-08-08 08:38:18 -05:00
Jonathan R. Madsen 49127db414 Fix some inconsistencies in debug messages w/in category_region (#135) 2022-08-08 08:38:04 -05:00
Jonathan R. Madsen bb400d5d61 Remove unused funcs + messages for excluding system lib (#133)
- instead of filtering system library opaquely, generate info messages
- remove unused are_file_include_exclude_lists_empty()
- remove unused module_constraint free func
- remove unused routine_counstraint free func
2022-08-08 08:37:51 -05:00
Jonathan R. Madsen 1343c6722a offset thread ids where possible (#130)
- configure known background threads to start indexing down from TIMEMORY_MAX_THREADS
- invoke sampling::shutdown() instead of just blocking signals where possible
2022-08-08 08:37:37 -05:00
Jonathan R. Madsen ac5ce00561 Update config generation, join fix, sampler (#129)
- update timemory submodule
  - update ring_buffer to remove all dynamic allocation
  - update sampler + sampler allocator to use ring-buffers
  - fix papi_array and papi_vector labels
- fix common::join with char delimiter
- fix config generation without --advanced
2022-08-08 07:03:24 -05:00
Kiumars Sabeti b53fd84ade SWDEV-302380: [ROCm QA][Mainline][Navi21] 6 tests are failing in rocprofiler-stg2
This is an attempt to support basic and derived counters for navi21.  This code will not work correctly unless we add navi counters to metrics.xml and gfx_metrics.xml

Change-Id: Ied06a81345a6fbb02fa0fde1889d94bbe64e9a03
2022-08-05 17:31:37 -04:00
Ranjith Ramakrishnan ba9633b74b Use function to find all libraries with absl prefix rather than hard coding the library names
Some absl libraries are missing in the list and results in error while loading rdci binary
Use cmake logic to find all the absl libraries required instead of hard coding them.
Similar implemenation was done  in rdc server side

Change-Id: Ia04bcd137f892c2577a4a458b92ca212d42aef80
2022-08-05 14:06:25 -07:00
Ranjith Ramakrishnan 52a3463147 File reorganization with backward compatibility
SWDEV-291455 -  Binary , header files and libraries installed in bin,include and lib folder under /opt/rocm-ver
Prebuilt ras library with updated search path
cmake config files in lib/cmake/rdc
grpc,sp3,hsaco and private libraries installed in lib/rdc
config  installed in share/rdc
authentication and python_binding installed in libexec/rdc
Backward compatibility added for header files and libraries

Depends-On: I3f3d192935923f71737b3fe55ded536654a73dd7
Change-Id: Ia1a6cadc59034b155631a1ee5fdbe692d2a8a71b
2022-08-04 23:42:42 -07:00
David Yat Sin 117495fe88 Fix image LUT for gfx11
For gfx11 the image type table has some different values compared to
previous asic families (e.g TYPE_SRGB). Creating a new LUT class to
use these new values.

Change-Id: Ifdfc6cd29bfd5f4ec2643c848fcb9986eb874f9e
2022-08-04 11:23:28 -04:00
Yifan Zhang daa01b8d57 Add gfx1103 support
This patch adds gfx1103 support

Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Change-Id: I7f1d580059fcd501bce2c8fea894637960c29bc1
2022-08-04 11:23:28 -04:00
David Yat Sin 574bea4a4c Use FAMILY_GFX1103 for gfx1103
Also adding elf entry

Change-Id: Id47ec379f2880961022b4607eb7f106b7e9d7048
2022-08-04 11:23:28 -04:00
David Yat Sin f971834d7a Update entries for gfx11
Update image table enums and format tables for gfx11.
Remove some entries that are not needed.

Change-Id: I060c1e285925a6d428ef1c5498f5dd89f5d79d97
2022-08-04 11:23:28 -04:00
David Yat Sin 319e71e79f Use FAMILY_GFX1100 for GFX11 devices
Change-Id: Ib182b647a91987040d655dbc05cbe5f867d4f61a
2022-08-04 11:23:28 -04:00
David Yat Sin a742b7e830 Update addrLib to support gfx11
This library was taken from public MESA library:
https://gitlab.freedesktop.org/mesa/mesa/-/tree/main/src/amd/addrlib

with top commit:
2866ae32da0348caf71ad2d11c353321df626ff4

Removing macros.h as it is no longer used by addrlib

Change-Id: I0fdabfe48b74c259b4d29d81beae89604bbc141a
2022-08-04 11:23:28 -04:00
David Yat Sin c2a60a4d5d Fix scratch memory alignment on GFX11
GFX11 requires scratch memory alignment of 256 Bytes instead of 1024.

Change-Id: I103de1c12f3a4877d7d36f13254301166c66e11f
2022-08-04 11:23:28 -04:00
David Yat Sin 90322899fe Update scratch register definitions for GFX11
Update scratch register definitions for GFX11 asics.

Change-Id: I6195e04b0a099fe84d1015c2f34ca3756a8175ef
2022-08-04 11:23:28 -04:00
Graham Sider 061aa04147 Make queue memory allocation non-paged
Non-paged allocation for queue memory necessary for binding wptr to
GART. Required to support usermode queue oversubscription with MES for
GFX11.

Adds AllocateNonPaged entry to MemoryRegion::AllocateEnum for clarity;
aliases AllocateIPC.

Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Change-Id: I1a97a1820da26cf2433d9c237b2e6d2b0b8628b4
2022-08-04 11:21:00 -04:00
Laurent Morichetti 5fd1c7e8e3 Fix vgpr count calculation for gfx90a and gfx940
Read accum_offset from compute_pgm_rsrc3 to report both the arch vgprs
and the accum vgprs

Change-Id: I99e746d54a6a1671e343da5658cc6ce970f79939
2022-08-03 14:02:36 -07:00
Jonathan R. Madsen afa3df8523 Advanced category for configuration options (#125)
Adds advanced category

- advanced category hides less relevant configuration options
- omnitrace-avail has new '--advanced' option which shows these flags
- increase verbosity level to print issue with reading ppid children
- OMNITRACE_ROCTRACER_HSA_ACTIVITY defaults to ON
- OMNITRACE_ROCTRACER_HSA_API defaults to ON
2022-08-03 12:13:00 -05:00
Graham Sider db1a13aa05 Clean up includes in queue.h
Formatting.

Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Change-Id: I141c8308d6b283b376035e21344629dc665289bb
2022-08-03 10:57:17 -04:00
David Yat Sin 907e05c1b3 Add new ImageManager for GFX11
Adding new ImageManager class for GFX11 GPUs

ImageManagerGfx11 functions copied from ImageManagerNv.
Register descriptions in resource_gfx11.h updated for gfx11.

Signed-off-by: David Yat Sin <david.yatsin@amd.com>
Change-Id: I48b39f6a633aef14aa829f7240a43fe0feb1c290
2022-08-03 10:57:09 -04:00
David Yat Sin cc3bd31591 Add gfx1102 support
Change-Id: I39cbda81a7a999aa2ecfad7a3e720000f7ca3408
Signed-off-by: David Yat Sin <David.YatSin@amd.com>
2022-08-03 10:56:54 -04:00
Graham Sider 446c5e9672 Add gfx1100 support
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Change-Id: Ic5d5559e43df5c73409ba900a42c6901aabae661
2022-08-03 10:56:49 -04:00
Jay Cornwall 710adcc252 Add gfx11 blit/trap shaders
David Yat Sin:
   Rebased to amd-staging branch
   Changed MSG_GET_DOORBELL to MSG_RTN_GET_DOORBELL

Change-Id: I6015e54c4d8897f4c796f58c7fbc298758c6d76d
2022-08-03 10:56:41 -04:00
Jonathan R. Madsen b79ce10fee RPATH to rocprofiler_LIBRARY_DIR for ROCm < v5.2 (#126)
* RPATH to rocprofiler_LIBRARY_DIR for ROCm < v5.2

- until v5.2 only librocprofiler64.so was symlinked in /opt/rocm. Thus linker using SOVERSION caused issues finding librocprofiler64.so.1

* Test ROCm w/ CMAKE_INSTALL_RPATH_USE_LINK_PATH=OFF

* INSTALL_RPATH_USE_LINK_PATH for omnitrace exe
2022-08-02 14:19:04 -05:00
Jonathan Kim 9d2fe1ac2a Fix GPU destruction when user disabled
GPUs excluded by RVD are not expected to have scratch, memory, trap
handling nor memory regions set up.  Now that these GPUs are added to
a new list, early return on agent destruction to prevent bad function
calls on destroy.

Also fix up broken memory releases between the gpu lists and ugly braces.

Change-Id: I52fc6e86ceba0a0383cedc63310eb409515eaf9f
2022-08-02 14:18:43 -04:00
Kent Russell 90ada94141 kfdtest: Add return statement for ReadSMIEventThread
This didn't return anything, so add a "return 0" at the end, since the
function expects to return an int value

Change-Id: I17c398e431b2ce4571e6ca4abe6d567f110ea2a7
2022-08-02 09:22:49 -04:00
Jonathan R. Madsen 97d17a8ef8 Fix RPATH handling (#122)
Fix rpath handling

- remove explicit add to CMAKE_INSTALL_RPATH
- remove overriding INSTALL_RPATH_USE_LINK_PATH for exes
2022-08-01 14:19:19 -05:00
Laurent Morichetti 602c67ee00 Add a HIP domain callback stress test
This test verifies that callback argument matches the callback function
as a race condition while setting and reading the pair could result in
mismatched arguments.

Change-Id: I2fe49d98d19bb780b6956ea6718762cfa0de93f8
2022-07-30 02:51:37 -04:00
jie1zhan 17fb40f1f6 Fix allocate memory failed in VRAM
: The kernel driver will do align VRAM allocations to 2MB, instead of 4KB.
Change-Id: Iea9d8c0f02999b9ea5fd931da82240a33f7bcc69
2022-07-30 01:18:50 -04:00
jie1zhan 8a0fe6a832 Free the executable memory , when it don't used
Fix the issue of rocrtst test - The runtime failed to allocate the necessary resources

Change-Id: Ie4ffeb939fb322db068f3132a7973a359c204176
2022-07-29 15:16:37 -04:00
skhatri 364715cbc6 Enabled allocation of pseudo fine grain memory where memory ordering is per point to point connection
Atomic memory operations on these memory buffers are not guaranteed
to be visible at system scope

Change-Id: I4cccde114632071a000384502a83bc191e77e85b
2022-07-29 15:15:56 -04:00
Konstantin Zhuravlyov d962fc39bb Add support for the following kernel symbol query:
- HSA_EXECUTABLE_SYMBOL_INFO_KERNEL_DYNAMIC_CALLSTACK

Change-Id: Idff5c1a2ce2a3e2d65bcc9cf1f66a68d37cd41ef
2022-07-29 15:15:24 -04:00
Konstantin Zhuravlyov 5a49b4d17f Bring AMDHSAKernelDescriptor.h in sync with llvm
Change-Id: Icd35100ad4d7eb8638786d306ecfbbb1c8842db1
2022-07-29 15:14:39 -04:00
Hao Zhou 4626bf8938 Merge amd-staging into amd-master 20220729
Signed-off-by: Hao Zhou <Hao.Zhou@amd.com>
Change-Id: Icb8fc012b23f5e991630ad5234339245bc68c273
2022-07-29 15:53:10 +08:00
Ammar ELWazir 2e08b7d8f8 Fixing HIP Callback Data Arguments
Change-Id: I0bd9686024eed8fa757cb9fd7042b0f9508c5af5
2022-07-28 19:29:04 -05:00
Ranjith Ramakrishnan c5159fa6d1 Remove the default setting of cmake install libdir from source code
Any default value if required should be controlled from outside.
For ROCM, build script is setting the value to "lib"

Change-Id: I12a2951307fe64e46a4e608476bfceb678bdc97d
2022-07-28 13:55:55 -04:00
Jonathan R. Madsen 641225f883 Fix uploading release assets (#118)
Fix uploading release assets
2022-07-27 15:10:19 -05:00
Jonathan R. Madsen 7e31d9f450 ROCm environment fixes + workflow updates (#117)
* Improve dlopen of ROCm libraries + rocprofiler test

- Use PROJECT_BINARY_DIR in tests
- Added rocprofiler test

* Revert OMNITRACE_FORCE_ROCPROFILER_INIT

* omnitrace-avail --all test

* Fix ROCP_METRICS for ROCm 5.2.0

* Fix ROCP_METRICS for ROCm 5.2.0

* Restrict containers workflow to AMDResearch/omnitrace

* Bump version to 1.3.1

* Update cpack workflow

- generate release draft
- upload installers as release assets

* Test rocprofiler w/o roctracer enabled

* Fix formatting

* verbose message
2022-07-27 06:36:52 -05:00
Laurent Morichetti 84ad727c38 Add agent enumeration
Intercept the first call to hsa_iterate_agents in order to number them.
The index assigned to agents will be used by a future commit.

Change-Id: I8db365f8fe913b6cde16a4dccb9bf09600846521
2022-07-27 02:16:42 -04:00
Laurent Morichetti 05d3cf3529 Cleanup roctracer_hsa.h
Remove declarations that are not meant to be part of the public API.

Change-Id: I47d9e83bf41bdb2f7ac25a1507200b51c616049b
2022-07-27 02:16:15 -04:00
Laurent Morichetti e282a82e29 Fix HSA intercept
Move the HSA intercept to the OnLoad function, so that it is available
as soon as the ROCR is loaded.

Layer the HSA API wrappers on top of the basic HSA activity intercept.

Change-Id: Ie636d59755543cda181e76ec29f0b55081136b63
2022-07-27 02:15:43 -04:00