TcpReceiveSendGetsCanceledByDispose: update test for change in Linux kernel. #93198
Conversation
ghost
added
the
community-contribution
|
Tagging subscribers to this area: @dotnet/ncl Issue DetailsFixes #91543. @antonfirsov @directhex ptal.
|
|
/azp run runtime-libraries-coreclr outerloop |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
oops, nevermind, wrong PR :D |
|
May be also useful here, as the ppc64le/s390x jobs this was observed on are very likely not part of the regular run. |
|
/azp list |
|
/azp run runtime-community |
|
Azure Pipelines successfully started running 1 pipeline(s). |
| // .NET uses connect(AF_UNSPEC) to abort on-going operations on Linux. | ||
| // Linux 6.4+ introduced a change (4faeee0cf8a5d88d63cdbc3bab124fb0e6aed08c) which disallows | ||
| // this operation while operations are on-going. |
There was a problem hiding this comment.
Assuming that distros will eventually update to 6.4+, does this mean that we should stop using connect(AF_UNSPEC)? Is there any alternative?
There was a problem hiding this comment.
I think for the time being we should keep connect(AF_UNSPEC) as there are many kernels still that provide this, and the resulting behavior matches closer to what Windows does.
There was a problem hiding this comment.
Eventually, though, this means our Linux implementation will not behave the way we want it to, right? Is there something better we can / should be doing instead?
There was a problem hiding this comment.
We're doing the best we can.
I've been investigating the issue with a kernel dev. He wrote this patch which fixes the regression.
We'll see if it gets included. 🤞
src/libraries/System.Net.Sockets/tests/FunctionalTests/SendReceive/SendReceive.cs
Show resolved
Hide resolved
|
So far it looks like this fixes the recent spike in @dotnet/dnceng were there any recent changes in the linux queues where the failures started to happen? (I suspect a kernel update.) |
Ubuntu kernel 5.15.0-81.90 included a backport of the breaking change in early august - but we were probably hit by a more recent kernel package which was uploaded to the security repo rather than the updates repo - 5.15.0-82.91 in early September, 5.15.0-83.92 in mid-September, or 5.15.0-86.96 in early October. |
|
The CI failures are unrelated, |
|
/backport to release/8.0 |
|
Started backporting to release/8.0: https://github.com/dotnet/runtime/actions/runs/6514772847 |
|
/backport to release/7.0 |
|
Started backporting to release/7.0: https://github.com/dotnet/runtime/actions/runs/6514775056 |
|
/backport to release/6.0 |
|
Started backporting to release/6.0: https://github.com/dotnet/runtime/actions/runs/6514776759 |
|
@antonfirsov backporting to release/6.0 failed, the patch most likely resulted in conflicts: $ git am --3way --ignore-whitespace --keep-non-patch changes.patch
Applying: TcpReceiveSendGetsCanceledByDispose: update test for change in Linux kernel.
Using index info to reconstruct a base tree...
M src/libraries/System.Net.Sockets/tests/FunctionalTests/SendReceive/SendReceive.cs
Falling back to patching base and 3-way merge...
Auto-merging src/libraries/System.Net.Sockets/tests/FunctionalTests/SendReceive/SendReceive.cs
CONFLICT (content): Merge conflict in src/libraries/System.Net.Sockets/tests/FunctionalTests/SendReceive/SendReceive.cs
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0001 TcpReceiveSendGetsCanceledByDispose: update test for change in Linux kernel.
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
Error: The process '/usr/bin/git' failed with exit code 128Please backport manually! |
|
@antonfirsov an error occurred while backporting to release/6.0, please check the run log for details! Error: git am failed, most likely due to a merge conflict. |
|
/backport to release/7.0-staging |
|
Started backporting to release/7.0-staging: https://github.com/dotnet/runtime/actions/runs/6514878161 |
ghost
locked as resolved and limited conversation to collaborators
Fixes #91543.
@antonfirsov @directhex ptal.