kernel: Enable ptp_kvm for 5.4 kernel on arm64. by jongwu · Pull Request #998 · kata-containers/packaging

jongwu · 2020-04-03T02:08:10Z

The enablement of ptp_kvm for arm is under review, see [1].
So we have to apply private patch to enable it in 5.4 kernel.
ptp_kvm can offer the capability of time sync in kata even there
is no network available and higher precision than time sync
service depend on network.

note:
If you want to use this feature on your arm machine, the host kernel
also need apply this patch. we recommend that your host kernel version
is the 5.4, then you can apply this patch smoothly.

[1] https://patchwork.kernel.org/cover/11372743/

Fixes: #997
Signed-off-by: Jianyong Wu jianyong.wu@arm.com

@devimc @grahamwhaley @Pennyzct

jodh-intel · 2020-04-03T10:49:50Z

/test

devimc · 2020-04-03T14:04:26Z

@jongwu couldn't not apply a patch

linux-5.4.15.tar.xz: OK
INFO: Applying patch /tmp/jenkins/workspace/kata-containers-packaging-ubuntu-18-04-snap-PR/go/src/github.com/kata-containers/packaging/kernel/patches//5.4.x/0001-arm-arm64-Provide-a-wrapper-for-SMCCC-1.1-calls.patch
patching file include/linux/arm-smccc.h
INFO: Applying patch /tmp/jenkins/workspace/kata-containers-packaging-ubuntu-18-04-snap-PR/go/src/github.com/kata-containers/packaging/kernel/patches//5.4.x/0003-ptp-arm64-Enable-ptp_kvm-for-arm64.patch
patching file drivers/clocksource/arm_arch_timer.c
patching file drivers/firmware/psci/psci.c
Hunk #1 FAILED at 71.
1 out of 1 hunk FAILED -- saving rejects to file drivers/firmware/psci/psci.c.rej

jongwu · 2020-04-08T01:09:24Z

@devimc it's strange as it works in my machine.

jongwu · 2020-04-08T08:02:41Z

@devimc I think that the order of patch applied cause that error, the patch should apply in order of 0001,0002,0003, if 0003 applied before 0002, error will occur.
So, should I squash patch serial to avoid that problem? @jodh-intel @devimc

jodh-intel · 2020-04-08T09:07:35Z

@jongwu, @devimc - I found the problem: #1005.

jongwu · 2020-04-08T09:15:41Z

thanks @jodh-intel

jodh-intel · 2020-04-09T08:09:49Z

Now that #1005 has landed, let's re-run the CI...

/retest

jongwu · 2020-04-09T09:09:57Z

thanks @jodh-intel , need retest as I just bump config version.

jodh-intel · 2020-04-09T09:46:57Z

/retest

jongwu · 2020-04-10T13:36:50Z

@jodh-intel can you help retest it? thanks

jongwu · 2020-04-13T11:33:20Z

hi @jodh-intel @devimc , help retest, thanks!

devimc · 2020-04-13T12:30:29Z

/test

devimc · 2020-04-13T18:14:09Z

/AzurePipelines run

azure-pipelines · 2020-04-13T18:14:22Z

Azure Pipelines successfully started running 2 pipeline(s).

devimc

thanks @jongwu

jongwu · 2020-04-15T09:15:45Z

hi @devimc @jodh-intel , help retest obs ci to see if that's a random error, thanks.

Pennyzct · 2020-04-15T09:26:11Z

/test

grahamwhaley · 2020-04-15T09:26:26Z

Here is the log. I'll try to kick a rebuild...

Project home:katacontainers:ci:x86_64:packaging-PR-998 has failed packages
 kata-containers-image
 | ksm-throttler
 | | linux-container
 | | | proxy
 | | | | qemu-vanilla
 | | | | | runtime
 | | | | | | shim
 % s s s s s s  CentOS_7 x86_64 (building)
 % s s s s s s  Debian_10 x86_64 (building)
 s s % s s s s  Debian_9 x86_64 (building)
 s s % s s s s  Fedora_30 x86_64 (building)
 % s s s s s s  SLE_15_SP1 x86_64 (building)
 s s % s s s s  openSUSE_Leap_15.1 x86_64 (building)
 % s f s s s s  openSUSE_Tumbleweed x86_64 (outdated)
 % s % s s s s  xUbuntu_16.04 x86_64 (building)
 f s % s s s s  xUbuntu_18.04 x86_64 (building)

 Legend:
  . succeeded           
    disabled            
  U unresolvable        
  F failed              
  B broken              
  b blocked             
  % building            
  f finished            
  s scheduled           
  L locked              
  x excluded            
  d dispatching

grahamwhaley · 2020-04-15T09:27:08Z

/test

jongwu · 2020-04-15T10:11:14Z

oh, building for linux-container all failed, is there any detail info to see why that happen? @grahamwhaley

grahamwhaley · 2020-04-15T10:26:03Z

I can't see anything on the Action/azure page - I think it will come down to some OBS stuff - we need to ask @jcvenegas for help I think.
It could be spurious, but let's check this failure first before we try a rebuild.

I'm also not sure why the ubuntu builds don't seem to be triggering....

devimc · 2020-04-15T13:34:54Z

@jongwu

https://build.opensuse.org/package/show/home:katacontainers:ci:x86_64:packaging-PR-998/linux-container

[  119s] + echo 'Patch #1 (0001-arm-arm64-Provide-a-wrapper-for-SMCCC-1.1-calls.patch):'
[  119s] Patch #1 (0001-arm-arm64-Provide-a-wrapper-for-SMCCC-1.1-calls.patch):
[  119s] + /usr/bin/cat /home/abuild/rpmbuild/SOURCES/0001-arm-arm64-Provide-a-wrapper-for-SMCCC-1.1-calls.patch
[  119s] + /usr/bin/patch -p1 --fuzz=0
[  119s] patching file include/linux/arm-smccc.h
[  119s] + echo 'Patch #2 (0003-ptp-arm64-Enable-ptp_kvm-for-arm64.patch):'
[  119s] Patch #2 (0003-ptp-arm64-Enable-ptp_kvm-for-arm64.patch):
[  119s] + /usr/bin/cat /home/abuild/rpmbuild/SOURCES/0003-ptp-arm64-Enable-ptp_kvm-for-arm64.patch
[  119s] + /usr/bin/patch -p1 --fuzz=0
[  119s] patching file drivers/clocksource/arm_arch_timer.c
[  119s] patching file drivers/firmware/psci/psci.c
[  119s] Hunk #1 FAILED at 71.
[  119s] 1 out of 1 hunk FAILED -- saving rejects to file drivers/firmware/psci/psci.c.rej
[  119s] patching file drivers/ptp/Kconfig
[  119s] patching file drivers/ptp/Makefile
[  119s] patching file drivers/ptp/ptp_kvm.h
[  119s] patching file drivers/ptp/ptp_kvm_arm64.c
[  119s] patching file drivers/ptp/ptp_kvm_common.c (renamed from drivers/ptp/ptp_kvm.c)
[  119s] patching file drivers/ptp/ptp_kvm_x86.c
[  119s] patching file include/linux/arm-smccc.h
[  119s] patching file include/linux/clocksource.h
[  119s] patching file include/linux/clocksource_ids.h
[  119s] patching file include/linux/timekeeping.h
[  119s] patching file include/uapi/linux/kvm.h
[  119s] patching file kernel/time/clocksource.c
[  119s] patching file kernel/time/timekeeping.c
[  119s] patching file virt/kvm/arm/arm.c
[  119s] patching file virt/kvm/arm/psci.c
[  119s] error: Bad exit status from /var/tmp/rpm-tmp.UvOSAI (%prep)

jongwu · 2020-04-16T01:55:39Z

@devimc @grahamwhaley this patch set should be applied in order. how to avoid that failure, should I squeezing these patch set?

jodh-intel · 2020-04-16T08:51:19Z

This PR is effectively blocked on #1016.

devimc · 2020-04-17T12:12:11Z

@jongwu this PR may need a rebase 😄

The enablement of ptp_kvm for arm is under review, see [1]. So we have to apply private patch to enable it in 5.4 kernel. ptp_kvm can offer the capability of time sync in kata even there is no network available and higher precision than time sync service depend on network. note: If you want to use this feature on your arm machine, the host kernel also need apply this patch. we recommend that your host kernel version is the 5.4, then you can apply this patch smoothly. [1] https://patchwork.kernel.org/cover/11372743/ Fixes: kata-containers#997 Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>

jongwu · 2020-04-17T13:50:31Z

thanks @devimc

devimc · 2020-04-17T14:05:44Z

/AzurePipelines run

azure-pipelines · 2020-04-17T14:06:00Z

Azure Pipelines successfully started running 2 pipeline(s).

jongwu · 2020-04-20T09:06:23Z

hi @devimc @jodh-intel, help trigger snap ci, thanks!

jodh-intel · 2020-04-20T09:40:49Z

/test-ubuntu

devimc · 2020-04-20T12:31:00Z

/test-snap

kvm-ptp is critical for mitigating time drift between host and guest. This implementation in kernel side is still one experimental feature on aarch64, and see kata-containers#998 for detailed instructions. Fixes: kata-containers#1004 Signed-off-by: Penny Zheng <penny.zheng@arm.com> Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>

jodh-intel approved these changes Apr 3, 2020

View reviewed changes

jodh-intel mentioned this pull request Apr 8, 2020

Kernel patches are not sorted correctly #1003

Closed

Pennyzct mentioned this pull request Apr 9, 2020

AArch64: Enable kernel fragment on aarch64 #1006

Merged

jongwu force-pushed the ptp branch 2 times, most recently from 0157524 to ea39591 Compare April 9, 2020 09:06

jongwu force-pushed the ptp branch from ea39591 to 27e64d9 Compare April 10, 2020 01:00

devimc approved these changes Apr 13, 2020

View reviewed changes

jongwu force-pushed the ptp branch from 27e64d9 to 3a0d53d Compare April 17, 2020 13:49

devimc merged commit fc8c3a0 into kata-containers:master Apr 20, 2020

Conversation

jongwu commented Apr 3, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jodh-intel commented Apr 3, 2020

Uh oh!

devimc commented Apr 3, 2020

Uh oh!

jongwu commented Apr 8, 2020

Uh oh!

jongwu commented Apr 8, 2020

Uh oh!

jodh-intel commented Apr 8, 2020

Uh oh!

jongwu commented Apr 8, 2020

Uh oh!

jodh-intel commented Apr 9, 2020

Uh oh!

jongwu commented Apr 9, 2020

Uh oh!

jodh-intel commented Apr 9, 2020

Uh oh!

jongwu commented Apr 10, 2020

Uh oh!

jongwu commented Apr 13, 2020

Uh oh!

devimc commented Apr 13, 2020

Uh oh!

devimc commented Apr 13, 2020

Uh oh!

azure-pipelines Bot commented Apr 13, 2020

Uh oh!

devimc left a comment

Choose a reason for hiding this comment

Uh oh!

jongwu commented Apr 15, 2020

Uh oh!

Pennyzct commented Apr 15, 2020

Uh oh!

grahamwhaley commented Apr 15, 2020

Uh oh!

grahamwhaley commented Apr 15, 2020

Uh oh!

jongwu commented Apr 15, 2020

Uh oh!

grahamwhaley commented Apr 15, 2020

Uh oh!

devimc commented Apr 15, 2020

Uh oh!

jongwu commented Apr 16, 2020

Uh oh!

jodh-intel commented Apr 16, 2020

Uh oh!

devimc commented Apr 17, 2020

Uh oh!

jongwu commented Apr 17, 2020

Uh oh!

devimc commented Apr 17, 2020

Uh oh!

azure-pipelines Bot commented Apr 17, 2020

Uh oh!

jongwu commented Apr 20, 2020

Uh oh!

jodh-intel commented Apr 20, 2020

Uh oh!

devimc commented Apr 20, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

jongwu commented Apr 3, 2020 •

edited

Loading