scmi-tests merge requestshttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests2021-09-17T09:10:57Zhttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/24Add TC0 hardware under Linux platform2021-09-17T09:10:57Zmonalisa-jena-armAdd TC0 hardware under Linux platformIntroduced TC0 hardware as one of the supported Linux platforms.
Restructured the directories under linux directory, and moved
"transport_mailbox.c" to common directory, since it is common
to both SGM776 and TC0 hardware.
Signed-off-by:...Introduced TC0 hardware as one of the supported Linux platforms.
Restructured the directories under linux directory, and moved
"transport_mailbox.c" to common directory, since it is common
to both SGM776 and TC0 hardware.
Signed-off-by: Monalisa Jena <monalisa.jena@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/25perf/test_d007: Update logic for intermediate level2021-09-17T10:44:39ZNicola Mazzucatoperf/test_d007: Update logic for intermediate levelTest 007 iterates over the performance levels discovered to find min,
max and intermediate. The intermediate level is chosen to be the second
one found. In domains where there are only 2 levels the intermediate
level is in fact the max.
...Test 007 iterates over the performance levels discovered to find min,
max and intermediate. The intermediate level is chosen to be the second
one found. In domains where there are only 2 levels the intermediate
level is in fact the max.
In further tests, the intermediate level is chosen to replace the max,
while effectively the value used is the max, thus nullifying the test
case.
This patch adds an extra check to pick the minimum level as
intermediate when a performance domain exposes only two levels.
Signed-off-by: Nicola Mazzucato <nicola.mazzucato@arm.com>
Change-Id: If952eaf6a64168c00a7ba7ef75c5a3adf6aee0efchetan singh rathorechetan singh rathorehttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/26Since Mailbox driver is same for TC0 and SGM, so renamed guide_sgm_testing.md...2021-09-23T06:51:22Zmonalisa-jena-armSince Mailbox driver is same for TC0 and SGM, so renamed guide_sgm_testing.md into guide_linux_testing.mdSince Mailbox driver is same for TC0 and SGM, so renamed guide_sgm_testing.md into guide_linux_testing.md
Updated SGM reference also from corresponding files
Updated user_guide.md for the same
Signed-off-by: Monalisa Jena <monalisa.jen...Since Mailbox driver is same for TC0 and SGM, so renamed guide_sgm_testing.md into guide_linux_testing.md
Updated SGM reference also from corresponding files
Updated user_guide.md for the same
Signed-off-by: Monalisa Jena <monalisa.jena@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/27Replace HARDWARE with TARGET and TC0 with TC2021-10-19T07:43:49ZShruti AggarwalReplace HARDWARE with TARGET and TC0 with TCSince SGM/TC test are run on FVP, therefore replaced HARDWARE
with TARGET.
Replaced TC0 with TC because TC0 implies Total Compute 2020.
Signed-off-by: Shruti Aggarwal <shruti.aggarwal2@arm.com>Since SGM/TC test are run on FVP, therefore replaced HARDWARE
with TARGET.
Replaced TC0 with TC because TC0 implies Total Compute 2020.
Signed-off-by: Shruti Aggarwal <shruti.aggarwal2@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/28TC: Update implementation version from 2.7 to 2.82021-10-26T05:02:21ZShruti AggarwalTC: Update implementation version from 2.7 to 2.8Signed-off-by: Shruti Aggarwal <shruti.aggarwal2@arm.com>Signed-off-by: Shruti Aggarwal <shruti.aggarwal2@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/29Replace MATTERHORN with MAKALU2021-10-29T05:00:04Zmonalisa-jena-armReplace MATTERHORN with MAKALUReplace MATTERHORN with MAKALU
Since expected performance_domain_names changed for
TC1, therefore modified these values.
Signed-off-by: Monalisa Jena <monalisa.jena@arm.com>Replace MATTERHORN with MAKALU
Since expected performance_domain_names changed for
TC1, therefore modified these values.
Signed-off-by: Monalisa Jena <monalisa.jena@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/31Update README.md with SCMI V3 now.2022-01-12T11:04:53ZNitin SharmaUpdate README.md with SCMI V3 now.Added description that SCMI V3 is supported by this ACS and it supports SCMI V2 and V1 too.Added description that SCMI V3 is supported by this ACS and it supports SCMI V2 and V1 too.https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/30Added SCMI v3 Tests.2022-02-25T11:42:19ZNitin SharmaAdded SCMI v3 Tests.SCMI V3 ACS Upgrade.
Test case for Voltage, Sensor 2.0 and other minor changes.
Tested in Baremetal, TC0 and Mockers.
Test results at : https://confluence.arm.com/display/BSGSoftware/Stage-4+Upstreaming+to+GitLab+repo
Signed-off-by: Niti...SCMI V3 ACS Upgrade.
Test case for Voltage, Sensor 2.0 and other minor changes.
Tested in Baremetal, TC0 and Mockers.
Test results at : https://confluence.arm.com/display/BSGSoftware/Stage-4+Upstreaming+to+GitLab+repo
Signed-off-by: Nitin Sharma <nitin.sharma3@arm.com>chetan singh rathorechetan singh rathorehttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/32Add changes for tc0-2022.02.25 release2022-04-06T08:51:52Zmonalisa-jena-armAdd changes for tc0-2022.02.25 releaseAdd changes for tc0-2022.02.25 release
Added VOLTAGE_PROTOCOL as supported protocols.
Changed implementation version from 2.8 to 2.9
Replaced performance domain name MAKALU with MATTERHORN.
Replaced voltage domain names from USB to DUMM...Add changes for tc0-2022.02.25 release
Added VOLTAGE_PROTOCOL as supported protocols.
Changed implementation version from 2.8 to 2.9
Replaced performance domain name MAKALU with MATTERHORN.
Replaced voltage domain names from USB to DUMMY_VOLTD.
Signed-off-by: Monalisa Jena <monalisa.jena@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/33test_c007: correct processing of triplet rate arrays2022-07-04T11:41:20ZBrett Warrentest_c007: correct processing of triplet rate arraysThe test incorrectly accesses triplet rate arrays
by going out of bounds due to a incorrectly used
increment variable. This is replaced by fixed indexes
outside of the previous loop.
Signed-off-by: Brett Warren <brett.warren@arm.com>The test incorrectly accesses triplet rate arrays
by going out of bounds due to a incorrectly used
increment variable. This is replaced by fixed indexes
outside of the previous loop.
Signed-off-by: Brett Warren <brett.warren@arm.com>chetan singh rathoreNicola Mazzucatochetan singh rathorehttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/35Add Linux app helpers to map physical to virtual addresses2022-11-08T07:58:08ZCristian MarussiAdd Linux app helpers to map physical to virtual addressesCurrently while running the ACS tests as a Linux app the test suite segfaults during FastChannels testcases because it tries to access directly the physical memory addresses describing the FCs: this works fine in the baremetal setup but ...Currently while running the ACS tests as a Linux app the test suite segfaults during FastChannels testcases because it tries to access directly the physical memory addresses describing the FCs: this works fine in the baremetal setup but it cannot be done in Linux virtual memory world.
Add a couple of helpers to map physical to virtual addresses when running in Linux and patch accordingly the relevant testcases: the behavior is left instead unchanged in baremetal where physical addresses can be used directly.
A successful baremetal CI testrun with this series included is at: https://jenkins.oss.arm.com/job/scp-v2-cmake-tests-scmi/3939/
Tested manually under Linux setup on my local JUNO.
***NOTE THAT***: even though the change does not affect at runtime the baremetal setup (that is run in CI) it ***DOES BREAK the compilation of CI mobile platforms*** (sgm775 sgm776) due to some BL31 segment overflow. (see https://jenkins.oss.arm.com/job/scp-v2-cmake-tests-scmi/3905/)
In order to avoid this, I patched the supporting ***scp-framework*** linker definitions in the ***branch v2.10.0-mobile*** so as to increase the MAX BL31 size by 4kb with the following tentative patch:
https://gerrit.oss.arm.com/gitweb?p=scp/test-framework.git;a=commit;h=ba1b735d51fbc6cc5d1f775630b05fb8180afe04
Not sure which is the correct way to merge the ***above scp-framework change*** BUT, in the hypothesis is fine, it ***has go in BEFORE this series*** is merged to avoid breaking the internal CI compilation of sgm775/776.Cristian MarussiCristian Marussihttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/36Miscellaneous SCMI v3.0 fixes2023-01-12T12:01:41ZCristian MarussiMiscellaneous SCMI v3.0 fixesThis merge-request include a bunch of small fixes around the ACS suite and one small enhancement (last patch) to allow the optional relaxing of the protocol version checks so that a few testcases can be run even when the SCMI server unde...This merge-request include a bunch of small fixes around the ACS suite and one small enhancement (last patch) to allow the optional relaxing of the protocol version checks so that a few testcases can be run even when the SCMI server under test had advertised a newer protocol implementation than the one currently supported by the test-suite.
This could lead to false positives, so it is off by default and can be enabled (only for the linux app as of now) via a command-line option -r.
The series has been tested successfully on CI at:
https://jenkins.oss.arm.com/job/scp-v2-cmake-tests-scmi/4815/
Thanks,
CristianCristian MarussiCristian Marussihttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/39Add support for SCMI Raw Linux interface and new targets for Linux platform.2023-08-03T11:06:21ZCristian MarussiAdd support for SCMI Raw Linux interface and new targets for Linux platform.This series at first adds support for Linux SCMI Raw transport (available since Linux v6.3) as an alternative way for the ACS to inject/receive SCMI messages to/from the server. Compiling with TRANS=raw builds a Linux app using such new ...This series at first adds support for Linux SCMI Raw transport (available since Linux v6.3) as an alternative way for the ACS to inject/receive SCMI messages to/from the server. Compiling with TRANS=raw builds a Linux app using such new Linux debugfs-based interface, instead of the legacy one based on the mailbox test driver.
Then, a few of test-cases have been fixed:
- test_c010/m014 to cope with the new SCMI Raw interface for async commands (if Raw was chosen)
- test_c009/c010/c011/c015/c017 to gracefully handle SCMI_DENIED replies about clocks
- test_d026/d027 to properly unregister notifications also on the failure path
- test_c017 to continue on error
- test_d026/027 to properly keep on looking for the desired notification in case other unrelated notifs where emitted
Finally a couple of new TARGET are added for the Linux platform (app):
- juno: to support such board
- generic_scmi: to support testing a generic SCMI server without any fixed expectations in terms of implemented protocols, found agents and resources.
**IMPORTANT NOTE**
This series causes the internal CI to break on mobile platforms (sgm**) due to a lack of space in BL31 to keep the compiled ACS when running in baremetal mode. As a fix, this patch to our CI infra is pending review:
https://gerrit.oss.arm.com/c/scp/test-framework/+/255449
has to be reviewed and (possibly) merged BEFORE this series in order not to break our internal CI. (for such reason is marked as Draft)
A successful CI run with both the series and the above fix is at:
https://jenkins.oss.arm.com/job/scp-v2-cmake-tests-scmi/6215/Cristian MarussiCristian Marussihttps://gitlab.arm.com/tests/scmi-tests/-/merge_requests/40Remove executable permissions from non-script files2023-08-21T10:10:08ZBalaji GontumukkalaRemove executable permissions from non-script filesSigned-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>Signed-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/41Changing return value to 02023-12-18T09:40:44ZBalaji GontumukkalaChanging return value to 0- Getting errors from CI/CD pipeline bacause the executable
is returning error code 1, so changing to retun 0
Signed-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>- Getting errors from CI/CD pipeline bacause the executable
is returning error code 1, so changing to retun 0
Signed-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/42Update .gitlab-ci.yml file2023-12-18T10:43:55ZBalaji GontumukkalaUpdate .gitlab-ci.yml fileAdded pipeline configuration file to build the mocker and run on merge_requests or when the trigger is from the master source. This also stores mocker executable as an artifact.Added pipeline configuration file to build the mocker and run on merge_requests or when the trigger is from the master source. This also stores mocker executable as an artifact.https://gitlab.arm.com/tests/scmi-tests/-/merge_requests/43Changes for SCMI v3.2 BETA-0 release2024-02-02T09:29:22ZBalaji GontumukkalaChanges for SCMI v3.2 BETA-0 releaseSigned-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>Signed-off-by: Balaji Gontumukkala <balaji.gontumukkala@arm.com>