Skip to content

K
Kernel

openeuler / Kernel
接近 2 年 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 16 5月, 2023 7 次提交
    • O
      !634 Accelerator Linux Mainline Patch Round · a0a072f1
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @xiao_jiang_shui 
 
https://gitee.com/openeuler/kernel/issues/I6VW8E 
 
Link:https://gitee.com/openeuler/kernel/pulls/634 

Reviewed-by: Yang Shen <shenyang39@huawei.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      a0a072f1
    • O
      !763 DAMON features: Support for process-level hot and cold data recognition and adjustment. · a824b373
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @gesefudiao 
 
The higher version community DAMON features in rounds:

Removed the debugfs interface and enabled the sysfs interface.
Added the lru_sort module to achieve proactive adjustment of the system-level LRU list for physical address monitoring.
Support for online automation adjustment of DAMON strategies. 
 
Link:https://gitee.com/openeuler/kernel/pulls/763 

Reviewed-by: Kefeng Wang <wangkefeng.wang@huawei.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      a824b373
    • O
      !749 intel: backport intel-idle support for EMR · 8614faa4
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @jiayingbao 
 
intel_idle: add Emerald Rapids Xeon support
commit 74528edf upstream.
Emerald Rapids (EMR) is the next Intel Xeon processor after Sapphire
Rapids (SPR).

EMR C-states are the same as SPR C-states, and we expect that EMR
C-state characteristics (latency and target residency) will be the
same as in SPR. Therefore, add EMR support by using SPR C-states table.


no change from upstream patch.

driver can be initial successfully as SPR.

PR dependency: 
https://gitee.com/openeuler/kernel/pulls/469
 
 
Link:https://gitee.com/openeuler/kernel/pulls/749 

Reviewed-by: Aichun Shi <aichun.shi@intel.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      8614faa4
    • O
      !748 intel: backport intel-uncore-freq support for EMR · 1ee0232e
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @jiayingbao 
 
backport from upstream commit 9c252ecf
Make Intel uncore frequency driver support Emerald Rapids by adding its
CPU model to the match table.

Emerald Rapids uncore frequency control is the same as in Sapphire Rapids.

no change from upstream patch.

driver can be initial successfully as SPR.


PR dependency: 
https://gitee.com/openeuler/kernel/pulls/469 
 
Link:https://gitee.com/openeuler/kernel/pulls/748 

Reviewed-by: Jason Zeng <jason.zeng@intel.com> 
Reviewed-by: Aichun Shi <aichun.shi@intel.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      1ee0232e
    • O
      !742 SPR: EDAC driver enhance for driver decode and 2LM · e2888060
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @youquan_song 
 
[Description]​
SPR: EDAC driver enhance for driver decode and 2LM

Current i10nm_edac only supports firmware decoder (ACPI DSM methods) for Sapphire Rapids. So add the driver decoder(decoding DDR memory errors via extracting error information from the IMC MC error codes) for Sapphire Rapids for better decoding performance.
Enhance the SPR HBM EDAC driver to decode errors from the 1st level memory (the fast "near" memory as cache) of the 2-level memory system.

It will be valuable to avoid SMI triggered to call firware decoder, especially when CE (Correctable Error) triggered frequently on DDR memory.

6e8746cb EDAC/skx_common: Enable EDAC support for the "near" memory.
221aa03f EDAC/i10nm: Add driver decoder for Sapphire Rapids server.

[Testing]
download: https://git.kernel.org/pub/scm/linux/kernel/git/aegl/ras-tools.git
# echo 1 > /sys/module/i10nm_edac/parameters/decoding_via_mca
# modprobe einj
# cd /home/ras-tools/
# ./cmcistorm 1
0: vaddr = 0x117b490 paddr = 4884ef490
# dmesg and check the address was decoded.
 
 
Link:https://gitee.com/openeuler/kernel/pulls/742 

Reviewed-by: Jason Zeng <jason.zeng@intel.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      e2888060
    • O
      !622 Intel: Add PMU support for Intel Emerald Rapids · 9b40194a
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @yunyingsun 
 
Title: Add PMU support for Intel Emerald Rapids
 
Content:
This PR adds Performance Monitoring Unit(PMU) support for next Intel Xeon platform Emerald Rapids.

Totally 6 commits, including 4 EMR PMU enabling patches from v6.2 and 2 dependent patches from v5.14/v5.18:
(v6.2-rc6) 5a8a05f1 perf/x86/intel/cstate: Add Emerald Rapids
	(v5.18-rc4) 528c9f1d perf/x86/cstate: Add SAPPHIRERAPIDS_X CPU support
	(v5.14-rc1) 87bf399f perf/x86/cstate: Add ICELAKE_X and ICELAKE_D support
(v6.2-rc6) 6795e558 perf/x86/intel: Add Emerald Rapids
(v6.2-rc4) 5268a284 perf/x86/intel/uncore: Add Emerald Rapids
(v6.2-rc4) 69ced416 perf/x86/msr: Add Emerald Rapids

The four 6.2 patches above use a macro "INTEL_FAM6_EMERALDRAPIDS_X", which is introduced by:
(v6.1-rc1) 7beade0d x86/cpu: Add several Intel server CPU model numbers
This patch is already included in another PR:
https://gitee.com/openeuler/kernel/pulls/469

Note: this PR for PMU must be merged AFTER PR-469, otherwise there will be kernel compiling error complaining for missing definition of macro "INTEL_FAM6_EMERALDRAPIDS_X".

Intel-kernel issue:
https://gitee.com/openeuler/intel-kernel/issues/I6YO4Z
 
Test:
1. platform dependent core PMU event works with perf, like "L1-dcache-loads".
2. platform dependent uncore PMU event works with perf, like "uncore_imc_0/event=0x1/".
3. offcore event works with perf.
3. PEBS works with perf.
4. topdown works with perf.
With this PR(along with the patch from PR469) applied to kernel OLK-5.10, all tests above PASS on EMR.

Known issue:
N/A
 
Default config change:
N/A


 
 
Link:https://gitee.com/openeuler/kernel/pulls/622 

Reviewed-by: Jason Zeng <jason.zeng@intel.com> 
Reviewed-by: Aichun Shi <aichun.shi@intel.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      9b40194a
    • O
      !615 Intel: Add RAPL support for Emerald Rapids(EMR) · 5216ad71
      由 openeuler-ci-bot 提交于 
      Merge Pull Request from: @yunyingsun 
 
Running Average Power Limit (RAPL) is an interface for reporting accumulated energy consumption of various system-on-chip (SoC) power domains. The RAPL energy reporting feature has been available for many generations on Intel SoC products.

To enable perf collecting RAPL events on future Intel Xeon platform Emerald Rapids, below upstream commit is needed:
(v6.2-rc3)  57512b57 perf/x86/rapl: Add support for Intel Emerald Rapids

This patch uses a macro "INTEL_FAM6_EMERALDRAPIDS_X" which is introduced with upstream commit:
(v6.1-rc1) 7beade0d x86/cpu: Add several Intel server CPU model numbers
which has been backported to OLK-5.10 earlier with this PR(under review, not merged yet):
https://gitee.com/openeuler/kernel/pulls/469
So here setting PR-469 as a dependency of this RAPL backport.

Note: this PR must be merged AFTER PR-469, otherwise there will be kernel compiling error complaining for missing definition of macro "INTEL_FAM6_EMERALDRAPIDS_X".

Intel-Kernel Issue:
https://gitee.com/openeuler/intel-kernel/issues/I6YGL6

Test:
With the patches in this PR included, running command below will list energy-psys events on EMR:
$ perf list power | grep energy-psys
Without the patches, no energy-psys event is available.
It's verified that test is PASS on Intel EMR pre-production platform.

Known Issue:
N/A

Default config change:
N/A 
 
Link:https://gitee.com/openeuler/kernel/pulls/615 

Reviewed-by: Jason Zeng <jason.zeng@intel.com> 
Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
      5216ad71
  3. 15 5月, 2023 22 次提交
  5. 12 5月, 2023 10 次提交
  7. 11 5月, 2023 1 次提交