Ubuntu 24.04 LTS (noble) and Ubuntu 22.04 LTS (jammy)
#2021
Conversation
|
Jenkins, test this please |
1 similar comment
|
Jenkins, test this please |
There was a problem hiding this comment.
Reviewed 8 of 8 files at r1, 3 of 3 files at r2, 3 of 3 files at r3, 5 of 5 files at r4, 32 of 32 files at r5.
Reviewable status: 47 of 79 files reviewed, 5 unresolved discussions, not enough approvals from maintainers (2 more required), not enough approvals from different teams (1 more required, approved so far: ITL) (waiting on @woju)
-- commits line 5 at r1:
Removal of Ubuntu 20.04 and the removal of EPID support are two orthogonal things, so this commit message reads weird. But I won't block, just keeping my comment for history.
-- commits line 22 at r3:
Interestingly, we haven't used Fixes: tag in Gramine up until now.
-- commits line 27 at r4:
I would expect some explanation, why we do it and why we do it only now.
From what I recall, this is because of #1909 (comment). Could we add a comment, or is it too late?
Also, wouldn't this fix #1909? If yes, please add the PR description.
-- commits line 43 at r5:
I would prefer to not mention my name (who cares) and instead make it a streamlined text. We can fix it during final rebase.
-- commits line 48 at r6:
Blocking myself: stopped on rev6 ("Rework benchmark-http.sh" commit), must continue from here tomorrow.
There was a problem hiding this comment.
Reviewed 3 of 3 files at r6, 4 of 4 files at r7, 3 of 3 files at r8, 1 of 1 files at r9, 1 of 1 files at r10, 1 of 1 files at r11, 1 of 1 files at r12, 2 of 2 files at r13, 2 of 2 files at r14, 1 of 1 files at r15, 2 of 2 files at r16, 14 of 14 files at r17, all commit messages.
Reviewable status: all files reviewed, 11 unresolved discussions, not enough approvals from maintainers (2 more required), not enough approvals from different teams (1 more required, approved so far: ITL) (waiting on @woju)
-- commits line 78 at r9:
Do we know why we needed this? Could we add a quick explanation in commit message body? But not blocking.
-- commits line 100 at r12:
Would be good to add a quick comment why it got broken. IIRC, it was something about difference in Clang versions (we run ASan only on Gramine built with Clang), which refused to print libos_call.c (i.e. proper source file name) for some reason.
-- commits line 112 at r14:
I would remove I guess is fine, it. So the sentence will become simply which means that ...
-- commits line 119 at r15:
Is this just to ease CI debugging? Could you add a short explanation in the commit body message?
-- commits line 123 at r17:
Could you explain again why this is needed?
.ci/ubuntu22.04.dockerfile line 56 at r8 (raw file):
# libomp-dev: needed for libos/test/regression/openmp.c # libevent-dev: CI-Examples/memcached # libmemcached-tools: CI-Examples/memcache
memcached (typo)
.ci/ubuntu24.04.dockerfile line 57 at r8 (raw file):
# libomp-dev: needed for libos/test/regression/openmp.c # libevent-dev: CI-Examples/memcached # libmemcached-tools: CI-Examples/memcache
ditto
This coincides with IAS EOL, which happened on 1 October 2024. Our CI runners for Focal used to use EPID developer key, and those were disabled. As of this writing, IAS anwers 401 to our key. See also: https://www.intel.com/content/www/us/en/developer/articles/technical/software-security-guidance/resources/sgx-ias-using-epid-eol-timeline.html Moving those runners to DCAP environment would require some work for very little value. Focal is the third newest LTS at the time of this writing, so not supported anyway according to our policy. Therefore we just disable Ubuntu 20.04 LTS altogether. Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
This is a leftover from changing the default driver to upstream. Fixes: 12e5d9a ("[CI] Change default SGX driver to upstream") Signed-off-by: Wojtek Porczyk <[email protected]>
Busybox we were using had problem compiling on newer kernels for undeclared `TCA_CBQ_MAX`. Switch to system busybox, as the one provided by distribution is already compiled, and all we are exercising is `busybox echo "Hello, world!"`, so compatibility between busybox versions is not really a concern. Fixes: #1909 Signed-off-by: Wojtek Porczyk <[email protected]>
Dmitrii Kuvaiskii explained it this way:
3 threads is nothing, given that:
- There is one main app thread
- There is one IPC helper thread
- There is one Async helper thread
- When a pipe is created, another helper thread is created for a brief moment
[...] technically 4 is a tight bound. There can't be more than that
(at least in the current state of Gramine, until we add more helper
threads).
Signed-off-by: Wojtek Porczyk <[email protected]>
- Make wrk2 dependency optional. The wrk2 tool adds -R option to wrk tool, however, wrk2 is not packaged for Debian/Ubuntu, but wrk is. If wrk2 is not available, then we can just use vanilla wrk tool. - Convert bc arithmetic to python3 -c. This removes bc dependency in favour of python3, which is always available, because it's a dependency of both Gramine tooling and build system. (POSIX shell arithmetic substitution does not support decimal, so it's not suitable). Signed-off-by: Wojtek Porczyk <[email protected]>
And add bookworm (Debian 12) and noble (Ubuntu 24.04 LTS) to check-python-platlib. Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
Based on this suggestion by Dmitrii Kuvaiskii: #1780 (comment) Signed-off-by: Wojtek Porczyk <[email protected]>
- epoll01: 300 s -> skip - pipe: 30 s -> 60 s - clock_gettime04: disable test 1 Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
For some reason memcached-tool sometimes outputs more than 2 lines, which still means that there is data in the server, so it's working. Signed-off-by: Wojtek Porczyk <[email protected]>
Piping straight to OUTPUT hid the error from the log, which made diagnostics unwieldy in our CI. Signed-off-by: Wojtek Porczyk <[email protected]>
Fixes: 03a42b0 ("[CI] Add pipeline for noble (Ubuntu 24.04 LTS)") Fixes: 9d53dde ("[CI] Add pipeline for jammy (Ubuntu 22.04 LTS)") Signed-off-by: Wojtek Porczyk <[email protected]>
Signed-off-by: Wojtek Porczyk <[email protected]>
woju
force-pushed
the
woju/ubuntu-noble-3
branch
from
bcbb172
to
38e6579
Compare
There was a problem hiding this comment.
Reviewable status: all files reviewed, 11 unresolved discussions, not enough approvals from maintainers (2 more required), not enough approvals from different teams (1 more required, approved so far: ITL) (waiting on @dimakuv)
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Removal of Ubuntu 20.04 and the removal of EPID support are two orthogonal things, so this commit message reads weird. But I won't block, just keeping my comment for history.
I've rewritten the commit message.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Interestingly, we haven't used
Fixes:tag in Gramine up until now.
It's kernel's convention: https://kernel.org/doc/html/latest/process/submitting-patches.html#using-reported-by-tested-by-reviewed-by-suggested-by-and-fixes
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
I would expect some explanation, why we do it and why we do it only now.
From what I recall, this is because of #1909 (comment). Could we add a comment, or is it too late?
Also, wouldn't this fix #1909? If yes, please add the PR description.
Done. Added commit message.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
I would prefer to not mention my name (who cares) and instead make it a streamlined text. We can fix it during final rebase.
Well, that was your explanation, so normally I shouldn't appropriate your work, but if you don't want to be credited, I can remove the mention.
Pray people won't quote your full emails in commit messages: https://gitlab.com/libvirt/libvirt-python/-/commit/8c800b0adf4849aeb94cb712192dba82fa4720d3
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Do we know why we needed this? Could we add a quick explanation in commit message body? But not blocking.
I'm not sure, it just timed out several times.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Would be good to add a quick comment why it got broken. IIRC, it was something about difference in Clang versions (we run ASan only on Gramine built with Clang), which refused to print
libos_call.c(i.e. proper source file name) for some reason.
Added the comment in the file itself, plz see below.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
I would remove
I guess is fine, it. So the sentence will become simplywhich means that ...
Done.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Is this just to ease CI debugging? Could you add a short explanation in the commit body message?
Done.
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
Could you explain again why this is needed?
To make it so that full stack runs on this particular distro (both the VM and the docker).
.ci/ubuntu22.04.dockerfile line 56 at r8 (raw file):
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
memcached(typo)
Done.
.ci/ubuntu24.04.dockerfile line 57 at r8 (raw file):
Previously, dimakuv (Dmitrii Kuvaiskii) wrote…
ditto
Done.
There was a problem hiding this comment.
Reviewed 71 of 71 files at r18, 3 of 3 files at r19, 2 of 3 files at r20, 2 of 5 files at r21, 32 of 32 files at r22, 2 of 3 files at r23, 4 of 4 files at r24, 3 of 3 files at r25, 1 of 1 files at r27, 1 of 1 files at r28, 2 of 2 files at r30, 2 of 2 files at r31, 1 of 1 files at r32, 2 of 2 files at r33, 14 of 14 files at r34, all commit messages.
Reviewable status: all files reviewed, all discussions resolved, not enough approvals from maintainers (1 more required)
Previously, woju (Wojtek Porczyk) wrote…
I've rewritten the commit message.
Thanks, exactly what I expected
Previously, woju (Wojtek Porczyk) wrote…
It's kernel's convention: https://kernel.org/doc/html/latest/process/submitting-patches.html#using-reported-by-tested-by-reviewed-by-suggested-by-and-fixes
I know, just wondered that it would be good to introduce this convention to Gramine too.
Previously, woju (Wojtek Porczyk) wrote…
Well, that was your explanation, so normally I shouldn't appropriate your work, but if you don't want to be credited, I can remove the mention.
Pray people won't quote your full emails in commit messages: https://gitlab.com/libvirt/libvirt-python/-/commit/8c800b0adf4849aeb94cb712192dba82fa4720d3
Ok, whatever, I'll unblock here.
|
Jenkins, test Jenkins-Direct-22.04-Sanitizers please |
Description of the changes
This PR replaces #1904, which unfortunately got too many commits and too many revisions to be revievable-able.
How to test this PR?
CI
Special GitHub/CI considerations
After merging this PR, 20.04 pipelines need to be unmarked as required in GitHub, instead the new pipelines for 24.04 and 22.04 need to be set as required.
Fixes: #1893
Closes: #1904
This change is