Bug 260850 - emulators/qemu-user-static: build failure
Summary: emulators/qemu-user-static: build failure
Status: Closed FIXED
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Some People
Assignee: Warner Losh
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2021-12-31 17:10 UTC by Colin Percival
Modified: 2022-02-21 05:16 UTC (History)
6 users (show)

See Also:
bugzilla: maintainer-feedback? (emulation)


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Colin Percival freebsd_committer freebsd_triage 2021-12-31 17:10:10 UTC
The release engineering team uses qemu-user-static for building arm64 VM images, and it has been failing to build recently:

3 warnings generated.
3 warnings generated.
1 warning generated.
3 warnings generated.
gmake[2]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4/arm-bsd-user'
gmake[1]: *** [Makefile:483: subdir-arm-bsd-user] Error 2
gmake[2]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4/aarch64-bsd-user'
gmake[1]: *** [Makefile:483: subdir-aarch64-bsd-user] Error 2
gmake[2]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4/mips64el-bsd-user'
gmake[1]: *** [Makefile:483: subdir-mips64el-bsd-user] Error 2
gmake[2]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4/sparc-bsd-user'
gmake[1]: *** [Makefile:483: subdir-sparc-bsd-user] Error 2
gmake[2]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4/ppc-bsd-user'
gmake[1]: *** [Makefile:483: subdir-ppc-bsd-user] Error 2
gmake[1]: Leaving directory '/usr/ports/emulators/qemu-user-static/work/qemu-bsd-user-6ffee39fe2e4'
===> Compilation failed unexpectedly.
Try to set MAKE_JOBS_UNSAFE=yes and rebuild before reporting the failure to
the maintainer.
*** Error code 1

Stop.
make: stopped in /usr/ports/emulators/qemu-user-static
*** Error code 1
Comment 1 Colin Percival freebsd_committer freebsd_triage 2021-12-31 17:11:52 UTC
Since we (at present) only use this for arm64, I was wondering if it would make our lives easier to have a stripped-down and rarely-updated "qemu-user-releng" port.
Comment 2 Kyle Evans freebsd_committer freebsd_triage 2021-12-31 17:28:54 UTC
I sent a request out-of-band for more of the log, since this excerpt didn't capture the failure.

releng should already be using emulators/qemu-user-static, which doesn't see updates but once in a blue moon (last real update was in May) -- I don't think a separate port will do us much good, typical breakage is from base updates.
Comment 3 Kyle Evans freebsd_committer freebsd_triage 2021-12-31 17:33:22 UTC
11:32 <@kevans91_> looks like some header change probably broke it
11:32 <@kevans91_> target_os_user.h is now pulling in sys/user.h and getting smacked by macro expansion of kve_vn_fsid
Comment 4 Warner Losh freebsd_committer freebsd_triage 2021-12-31 19:35:16 UTC
I'll take a look at this when I'm back on Tuesday (or maybe a little before)...
Comment 5 Warner Losh freebsd_committer freebsd_triage 2021-12-31 19:43:38 UTC
(In reply to Warner Losh from comment #4)
We've had this error several times before...
Comment 6 Kyle Evans freebsd_committer freebsd_triage 2021-12-31 20:01:58 UTC
I'll see if I can get to it this weekend.

Given that re@ relies on qemu-user-static, we should really get a CI job in Jenkins going for base changes like this that break it. I don't know if re@ has other ports they'd like to add to that, but CC lwhsu@ to start the conversation.

If re@ doesn't have anything else to add, then a job that builds qemu-user-static and notifies imp/kevans (please not the committers involved) on job failures would be fine.
Comment 7 Warner Losh freebsd_committer freebsd_triage 2022-02-21 00:19:34 UTC
I think this is fixed
Comment 8 Kyle Evans freebsd_committer freebsd_triage 2022-02-21 05:16:11 UTC
(In reply to Warner Losh from comment #7)

Yup, we've gotten qemu back out of the way; the latest failures were kld funkiness. Closing...