criu: Version 3.17 #1862

adrianreber · 2022-05-01T15:55:16Z

Amongst a huge number of fixes all over the place this release introduces:

mount-v2 engine
support for MAP_HUGETLB mappings
support for Linux Restartable Sequences
support for SOCK_SEQPACKET unix sockets
CRIU AMD GPU plugin
setsockopt(SO_BUF_LOCK) support for tcp sockets

@checkpoint-restore/maintainers this are the Makefile.versions changes for release 3.17. PTAL.

adrianreber · 2022-05-02T06:31:25Z

I see a couple of new failures with static/bridge:

2022-05-01T16:00:16.9754754Z ======================== Run zdtm/static/bridge in uns =========================
2022-05-01T16:00:16.9754999Z Start test
2022-05-01T16:00:16.9755188Z Test is SUID
2022-05-01T16:00:16.9755495Z ./bridge --pidfile=bridge.pid --outfile=bridge.out
2022-05-01T16:00:16.9755743Z Run criu dump
2022-05-01T16:00:16.9755985Z Wait for zdtm/static/bridge(87) to die for 0.100000
2022-05-01T16:00:16.9756353Z Run criu restore
2022-05-01T16:00:16.9756665Z =[log]=> dump/zdtm/static/bridge/87/1/restore.log
2022-05-01T16:00:16.9757064Z ------------------------ grep Error ------------------------
2022-05-01T16:00:16.9757444Z b'(00.006251)      1: Try to restore a link 10:2:zdtmbr0'
2022-05-01T16:00:16.9757775Z b'(00.006256)      1: Restoring link zdtmbr0 type 6'
2022-05-01T16:00:16.9758118Z b'(00.006262)      1: Restoring netdev zdtmbr0 idx 2'
2022-05-01T16:00:16.9758467Z b'(00.006269)      1: Restore ll addr (c6:../6) for device'
2022-05-01T16:00:16.9758914Z b'(00.006300)      1: Error (criu/libnetlink.c:54): -95 reported by netlink: Operation not supported'
2022-05-01T16:00:16.9759357Z b"(00.006304)      1: Error (criu/net.c:1816): Can't restore link: -95"
2022-05-01T16:00:16.9759806Z b"(00.006404)      1: Error (criu/util.c:1411): Can't wait or bad status: errno=0, status=65280"
2022-05-01T16:00:16.9760184Z b'(00.006841) uns: calling exit_usernsd (-1, 1)'
2022-05-01T16:00:16.9760514Z b'(00.006875) uns: daemon calls 0x476140 (117, -1, 1)'
2022-05-01T16:00:16.9760828Z b'(00.006886) uns: `- daemon exits w/ 0'
2022-05-01T16:00:16.9761131Z b'(00.007517) uns: daemon stopped'
2022-05-01T16:00:16.9761475Z b'(00.007524) Error (criu/cr-restore.c:2536): Restoring FAILED.'
2022-05-01T16:00:16.9761868Z ------------------------ ERROR OVER ------------------------
2022-05-01T16:00:16.9762200Z ################# Test zdtm/static/bridge FAIL at CRIU restore #################

The main difference between the working CI runs and the none working CI runs is that the CI kernel has been upgraded from Linux 5.13.0-1021-azure #24~20.04.1-Ubuntu SMP Tue Mar 29 15:34:22 UTC 2022 x86_64 to 5.13.0-1022-azure #26~20.04.1-Ubuntu SMP Thu Apr 7 19:42:45 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux.

I cannot reproduce on Fedora's 5.16.18. As the failure is in Arch and Fedora Rawhide maybe it is a combination of newer libnl3 and that specific kernel.

The Podman test seems to be failing because crun picks up the installed CRIU (/usr)and not the one from our CI run (/usr/local).

avagin · 2022-05-02T06:32:07Z

@adrianreber could you take a look at why presubmit checks failed?
@mihalicyn What does Mr. Jenkins think about this release?

GitHub Actions comes with pre-installed criu in /usr. configure scripts looking for CRIU will pickup the pre-installed version in /usr if we do not install CI criu also in /usr. Signed-off-by: Adrian Reber <areber@redhat.com>

Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

checkpoint-restore#1866 Suggested-by: Andrei Vagin <avagin@gmail.com> Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

adrianreber · 2022-05-05T06:05:44Z

Updated this PR with three patches to fix CI failures.

codecov-commenter · 2022-05-05T06:24:49Z

Codecov Report

Merging #1862 (c52498f) into master (d879220) will decrease coverage by 0.09%.
The diff coverage is 43.92%.

@@            Coverage Diff             @@
##           master    #1862      +/-   ##
==========================================
- Coverage   69.08%   68.98%   -0.10%     
==========================================
  Files         136      128       -8     
  Lines       33167    33346     +179     
==========================================
+ Hits        22914    23005      +91     
- Misses      10253    10341      +88

Impacted Files	Coverage Δ
compel/arch/x86/src/lib/thread_area.c	`86.36% <0.00%> (-4.55%)`	⬇️
compel/src/lib/infect-util.c	`33.33% <ø> (ø)`
criu/arch/x86/cpu.c	`74.82% <ø> (ø)`
criu/arch/x86/include/asm/types.h	`100.00% <ø> (ø)`
criu/arch/x86/sigaction_compat.c	`0.00% <0.00%> (ø)`
criu/eventpoll.c	`79.51% <ø> (+0.38%)`	⬆️
criu/include/autofs.h	`100.00% <ø> (ø)`
criu/include/files-reg.h	`100.00% <ø> (ø)`
criu/include/image.h	`100.00% <ø> (ø)`
criu/include/linux/mount.h	`100.00% <ø> (ø)`
... and 122 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2f0f128...c52498f. Read the comment docs.

mihalicyn · 2022-05-05T15:29:59Z

@adrianreber please take changes from these PRs:
#1871
#1869

Fixes: e2e02bc ("zdtm: Add MAP_HUGETLB memory mapping test") Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

Currently we check memfd_hugetlb by doing memfd_create("", MFD_HUGETLB). If we see EINVAL we report that it's not supported, but we can also get ENOENT error in such case in hugetlb_file_setup() while trying to find proper hugetlbfs mount. Reference: https://github.com/torvalds/linux/blob/06fb4ecfeac/fs/hugetlbfs/inode.c#L1465 Fixes: 4245e6b ("check: Add a check for using memfd with hugetlb") Reported-by: Mr. Jenkins (ppc64le) Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

This commit has to be reverted once we fix the issue. Issue: checkpoint-restore#1868 Reported-by: Mr. Jenkins Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

Amongst a huge number of fixes all over the place this release introduces: * mount-v2 engine * support for MAP_HUGETLB mappings * support for Linux Restartable Sequences * support for SOCK_SEQPACKET unix sockets * CRIU AMD GPU plugin * setsockopt(SO_BUF_LOCK) support for tcp sockets Signed-off-by: Adrian Reber <areber@redhat.com>

adrianreber · 2022-05-05T15:47:58Z

@adrianreber please take changes from these PRs: #1871 #1869

I added the three commits.

adrianreber · 2022-05-10T15:15:09Z

@avagin Please rebase criu-dev on top of master.

li-xiaocheng · 2022-09-21T09:23:10Z

hello! now criu support amd gpu， nvidia gpu when can it be supported？

adrianreber · 2022-09-21T09:42:48Z

hello! now criu support amd gpu， nvidia gpu when can it be supported？

As soon as somebody writes the patches to support it.

adrianreber mentioned this pull request May 2, 2022

test: install criu in /usr #1863

Merged

This was referenced May 3, 2022

Criu check fails on NI Linux RT #1864

Open

ci: print kernel modules list #1865

Closed

Race between zdtm/static/sit and zdtm/static/bridge #1866

Open

adrianreber and others added 3 commits May 5, 2022 06:04

test: install criu in /usr

814b847

GitHub Actions comes with pre-installed criu in /usr. configure scripts looking for CRIU will pickup the pre-installed version in /usr if we do not install CI criu also in /usr. Signed-off-by: Adrian Reber <areber@redhat.com>

ci: print kernel modules list

1139062

Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

ci: workaround race between sit module loading and bridge test

5d0ae8e

checkpoint-restore#1866 Suggested-by: Andrei Vagin <avagin@gmail.com> Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

adrianreber force-pushed the 2022-05-01-3-17 branch from 659ddfa to c52498f Compare May 5, 2022 06:05

mihalicyn and others added 4 commits May 5, 2022 15:46

zdtm: handle the case when hugetlb isn't supported

60b88ce

Fixes: e2e02bc ("zdtm: Add MAP_HUGETLB memory mapping test") Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

ci: skip new hugetlb maps09/maps10 tests for pre-dump

745b270

This commit has to be reverted once we fix the issue. Issue: checkpoint-restore#1868 Reported-by: Mr. Jenkins Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>

adrianreber force-pushed the 2022-05-01-3-17 branch from c52498f to fcd15df Compare May 5, 2022 15:47

avagin merged commit 4f8f295 into checkpoint-restore:master May 5, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

criu: Version 3.17 #1862

criu: Version 3.17 #1862

adrianreber commented May 1, 2022 •

edited

Loading

adrianreber commented May 2, 2022

avagin commented May 2, 2022

adrianreber commented May 5, 2022

codecov-commenter commented May 5, 2022

mihalicyn commented May 5, 2022

adrianreber commented May 5, 2022

adrianreber commented May 10, 2022

li-xiaocheng commented Sep 21, 2022

adrianreber commented Sep 21, 2022

criu: Version 3.17 #1862

criu: Version 3.17 #1862

Conversation

adrianreber commented May 1, 2022 • edited Loading

adrianreber commented May 2, 2022

avagin commented May 2, 2022

adrianreber commented May 5, 2022

codecov-commenter commented May 5, 2022

Codecov Report

mihalicyn commented May 5, 2022

adrianreber commented May 5, 2022

adrianreber commented May 10, 2022

li-xiaocheng commented Sep 21, 2022

adrianreber commented Sep 21, 2022

adrianreber commented May 1, 2022 •

edited

Loading