If one API fail, tool will work properly for all other arguments
Change-Id: I7b996e6da87873efb9d3210398971ea22203ac44
Signed-off-by: Dalibor Stanisavljevic <Dalibor.Stanisavljevic@amd.com>
Bug Fixes for Set and Fan
Updated lib_amdsmi.so directory access
Backwards compatability fixes for gpuv-smi
Change-Id: I3b7977859c750c1c3d6f41eaa761c81d8b9e5184
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Updates:
* [rocm-smi] Added larger app width size, which helps
display missing device info
* [rocm-smi] Added better context when rsmi_ret_ok
does not return with RSMI_STATUS_SUCCESS
* [rocm-smi] Removed all references to an
undefined function (printLogNoDev())
* [rocm-smi] Fixed not detecting non-int
values when setting the voltage curve
* [rocm-smi] Added better context on missing
sysfs file when setting clock overdrive
values
* [rocm-smi] Fixed getMemInfo() calls not
referencing tuple values (making it easier
to read)
* [rocm-smi] Silenced concise info spitting
out errors for missing VRAM files, instead
display which metric is "unsupported" if
the files are missing
* [rocm-smi] Updated function descriptions for
rsmi_ret_ok & getMemInfo
* [rocm-smi] Updated getMemInfo to provide a
quiet call, to silence for concise info calls.
This provides a way to keep the output clean.
* [rocm-smi-lib] Added when using debug sysfs
files, to state, which enums are enabled
for debug
Change-Id: I0e9e0c97ccf71467ced0e1a1f71803327a8be2b7
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Updates:
* VoltRead - needed to properly send out RSMI_STATUS_NOT_SUPPORTED
when device does not have voltage hwmon files
* ComputePart. - test failure was likely caused due to EvtNotif
causing conflicts (unknown exactly why). Test passes when
moving it ahead of the event notifier. Both API calls may have
a system resource issue, TBD.
* rocm_smi_example - now indicates when an API call
returns RSMI_STATUS_NOT_SUPPORTED or
RSMI_STATUS_NOT_YET_IMPLEMENTED. Allows example to fully complete
on systems which may not provide support for all API calls.
Change-Id: I520b8584e078d412414e8e5797c664220a7e823a
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Added Rocm Set Commands
Wrapped all amdsmi_interface calls with error handling
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Change-Id: Ic6b3648ef01ded8ee1fb7f0f14f3ca7bc069c567
The rsmi_dev_subsystem_name_get() only matches subvendor id and
subdevice id for a vendor. The change will also match device id.
Change-Id: Ife3aedaf6fc7390ed7fa62edbde40c2340689b23
Python is no longer used for generating any CLI files
Change-Id: I70dff5d86df9a179f99b32630dada9e92391eaaa
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
Wrapper generation is still only compatible with python3.7 and above
Change-Id: I33cdc3925cd3fab80c9ce5f4540e1a981a5cc1f0
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
Generating .whl has proved to be difficult with AMD CI.
Switch to directly installing python projects instead.
Change-Id: I8a4c2fe22fc872865bba7da3a3386513efd5269a
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
- Added CMake packaging to the amdsmi_cli folder
- Updated Headings in the README.md to follow markdown standard
- Updated Compatibility to be based on the built package name in bin
- Added misc error handling and import corrections
- Updated py-interface amdsmi_exception imports to work by relative path
- Cleaned up py-interface cmake & generator code for finding libamd_smi.so
- Changed line endings in tools/generator.py file to unix
Change-Id: I91858ff3dd0cb57ed9b8cd61a0ada27b6af9c51c
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Fix was needed due to hwmon updates.
Several voltage sensors (ex. vddgfx/vddnb)
are unsupported or not applicable
to upcoming hardware. This was not the case
for previous hardware sensors, resulting in
the rocm-smi crash observed.
Change-Id: Ib8593e10811638def26fc7a1eda29309e328db09
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
- Add "BUILD_PACKAGE" option that is OFF by default (CI issues)
- Optionally package wrapper if python3.7 or above is present (hack for CI)
- Optionally build wrapper if clang is present (hack for CI)
- Improve dependency resolution for wrapper
- Use python venv
- Use pyproject.toml instead of setup.py
- Use CMake install path to search for libamd_smi.so
- Move python-specific CMakeLists into py-interface
- Search for libamd_smi.so more aggressively
Change-Id: Ie7dad676b0d4a5f58ad2b887db7fecf5b1297e3b
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
In case there is no device id to map to the corresponding market_name,
the rsmi_dev_brand_get function is used to retrieve the market_name
Signed-off-by: Dalibor Stanisavljevic <Dalibor.Stanisavljevic@amd.com>
Change-Id: I565089d34b2b7e5f714e0dd41062ac8d52095835
Using backward compatibility paths will provide an #error message. Compile time option added to enable/disable the #error message.
Disabling the same will provide a #warning message
Change-Id: Ib49633501aa6eb6d97158b1ecfc47de6f18fba85
If the amdsmi is fail to read ECC counter from sysfs file, it
still assumes the data is read into vector, which causes crash.
Add the check to make sure the data is read into the vector.
Change-Id: Ia24331ee6e75408120248b0ed7bc58581a2f23ad
The process ids of other container are still visible in the sysfs file,
filter it out to prevent crash.
Change-Id: I665912cd09c606804186aff8cba5c24f5e58ded7
Updates:
* Fixed rsmi_dev_compute_partition_get
& rsmi_dev_nps_mode_get to properly check
for invalid arguments
* Updated compute partition & NPS mode tests
- Now properly confirms the invalid
argument is seen
- Spacing for multiple devices is added
to better see distinction between
separate device's tests (for verbose output)
- Changed expect to assert calls, so errors
are observed faster for test failures
- Fixed multiple device testing where a
variable should have been unset, but
having multiple devices caused it to
set
- Updated multiple device testing to iterate
accross all devices (previously returned,
instead of continuing checking support
after RSMI_STATUS_NOT_SUPPORTED detected)
- Fixed a few spelling errors & verbose output
Change-Id: Ieba9e5b46763c6cd880fbf27fcdf58be8ecbc683
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Updates:
* Added RSMI_STATUS_SETTING_UNAVAILABLE for
rsmi_dev_compute_partition_set - gives users
better error output when attempting to set
compute partition to values not listed in
available_compute_partition SYSFS
* Updated python --setcomputepartition to
provide better output when receiving
RSMI_STATUS_SETTING_UNAVAILABLE
* Updated all test & example files to check for
RSMI_STATUS_SETTING_UNAVAILABLE when doing
rsmi_dev_compute_partition_set
Change-Id: Ida5d54880d9b9b6e4a0468cdb962fdc0c18d6257
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Dalibor Stanisavljevic
Maisam Arif
Marko Oblak
Hao Zhou
Bill(Shuzhou) Liu
Galantsev, Dmitrii
Charis Poag
AravindanC
Ranjith Ramakrishnan