Verification:
- Install success onto a StarlingX system with two controller and
two compute nodes with std kernel.
- Install success onto a All-in-One Duplex lab with rt kernel.
- The network performance test is done with rt system. The result is
almost same with 5.10.99.
- The regression testing was run including kernel and applications
by our test team.
- The cyclictest benchmark was run by our test team, the result is
little better than 5.10.99. Test result is "samples:259200000
hrtimer_noload_min:3545 hrtimer_noload_avg: 4636
hrtimer_noload_max:8025 99.9999th perc.:6427".
- The cyclictest benchmark was also run on the starlingx lab, the
result is "samples:43199988 avg:1340 max:14633 99.9999th percentile:3030
overflows: 0". It is not big difference with 5.10.99.
Reviewed: https:/ /review. opendev. org/c/starlingx /kernel/ +/841891 /opendev. org/starlingx/ kernel/ commit/ 61723d0f6553160 1aedafbc7bf310c 5838df2db9
Committed: https:/
Submitter: "Zuul (22348)"
Branch: master
commit 61723d0f6553160 1aedafbc7bf310c 5838df2db9
Author: Jiping Ma <email address hidden>
Date: Wed Apr 27 03:52:57 2022 -0400
CentOS: kernel: Update 5.10 Kernel to 5.10.112
This commit updates kernel to 5.10.112 to fix many issues, including the /git.yoctoproje ct.org/ linux-yocto/ commit/ net/ipv4/ inet_hashtables .c?h=v5. 10/standard/ preempt- rt/base& id=bd6e7290bc76 6ef13e42a1e37f7 5e6e708b4e317
following two CVE issues which were of special concern.
- CVE-2022-0847: kernel: improper initialization of the "flags"
member of the new pipe_buffer
- CVE-2022-0435: kernel: remote stack overflow via kernel panic on
systems using TIPC may lead to DoS
Note we add more one commit
https:/
a minor fix-up on top of v5.10.112 for rt kernel.
Verification: noload_ min:3545 hrtimer_noload_avg: 4636 noload_ max:8025 99.9999th perc.:6427".
- Install success onto a StarlingX system with two controller and
two compute nodes with std kernel.
- Install success onto a All-in-One Duplex lab with rt kernel.
- The network performance test is done with rt system. The result is
almost same with 5.10.99.
- The regression testing was run including kernel and applications
by our test team.
- The cyclictest benchmark was run by our test team, the result is
little better than 5.10.99. Test result is "samples:259200000
hrtimer_
hrtimer_
- The cyclictest benchmark was also run on the starlingx lab, the
result is "samples:43199988 avg:1340 max:14633 99.9999th percentile:3030
overflows: 0". It is not big difference with 5.10.99.
Closes-Bug: 1969605 /review. opendev. org/c/starlingx /manifest/ +/841835
Depends-On: https:/
Signed-off-by: Jiping Ma <email address hidden> 46e8ca85c8f954a bcc7e5c9fec
Change-Id: I80d72c157e4db6