Improve cross-compilation support for arm and arm64 #6769
Conversation
| SET( HAVE_SHM_OPEN_THAT_WORKS_WELL_ENOUGH_WITH_MMAP_EXITCODE | ||
| 0 | ||
| CACHE STRING "Result from TRY_RUN" FORCE) | ||
|
|
There was a problem hiding this comment.
Was this intentional, looks like a mistake.
There was a problem hiding this comment.
Yes, to be like CoreCLR and CoreRT that have empty lines between the various SET commands.
There was a problem hiding this comment.
Yeah my bad sorry, saw a + and assumed it was a literal + forgetting it's used to mean addition.
|
LGTM modulo minor comments, won't be able to check it this weekend though. |
|
@manu-silicon, checking in on this. Is this PR still necessary? Are you planning to address the feedback that was left? |
|
Unless what is currently in CoreFX enables this, I would still think it is relevant. However, it has been a while I worked on the ARM side and haven't tested this recently. Maybe @leemgs can comment on this. |
|
@leemgs, @prajwal-aithal, could you comment on this and what you think the next steps are here? Thanks! |
|
@stephentoub @manu-silicon I agree that this PR brings consistency between cross build scripts in coreclr and corefx. The current PR has merge conflicts, and thus they should be fixed first. In the changes made currently, the changes in the file |
|
It would be great to have this PR come to a conclusion. Should we close it and track it with an issue until someone has time to finish it up or will someone be able to get to it in the next couple days? |
|
I've updated the PR to match master. Not sure what the Linux ARM Emulator failed in Release mode though. |
/CC: @sjsinju, @prajwal-aithal |
|
@dotnet-bot test Linux ARM Emulator Release |
|
@manu-silicon , looks like there are still merge conflicts with master. Did you need some help understanding the CI failure? Who can help with that? |
|
@dotnet-bot test Innerloop Linux ARM Emulator Release Cross Build please |
|
I'll check the conflicts and if anyone has an idea about the failure |
|
I believe the "device is busy" error should have been fixed. You may need to close this PR and re-open it (or open a new one) and rebase to pick up the revised netci.groovy changes that cause the legs to be produced. |
|
@manu-silicon, can you close this and submit a new rebased PR with the changes to pick up the Jenkins fix as Stephen suggests? Thanks for continuing to push this through! |
|
There are now too many changes and I need to check how things are done before I can re-apply my changes. I'm closing this now and when I'll have more time I'll re-open if there is a need to. |
8 participants
Merge parts for PR #6194 which were not included in PR #6585 so that cross compilation scripts are similar to the ones in CoreCLR and CoreRT.