Bug 196984

Summary: [arm] sysutils/cbsd attempts to use SSE 4.1 instructions on all architectures
Product: Ports & Packages Reporter: Sean Bruno <sbruno>
Component: Individual Port(s)Assignee: freebsd-ports-bugs (Nobody) <ports-bugs>
Status: Closed FIXED    
Severity: Affects Many People CC: amdmi3, olevole
Priority: --- Flags: bugzilla: maintainer-feedback? (olevole)
Version: Latest   
Hardware: arm   
OS: Any   
URL: http://chips.ysv.freebsd.org/data/11armv6-11armv6/2015-01-19_15h28m19s/logs/errors/cbsd-10.1.3.log
Attachments:
Description Flags
add ONLY_FOR_ARCHS=i386 amd64 for cbsd port none

Description Sean Bruno freebsd_committer freebsd_triage 2015-01-22 01:43:28 UTC 
I suspect a configure option needs to be passed to cbsd in order to not use SSE 4.1 instructions.

If there is no option for this, then the port should be marked as ONLY_FOR_ARCHS=i386 amd64
Comment 1 Bugzilla Automation freebsd_committer freebsd_triage 2015-01-22 01:43:28 UTC 
Maintainer CC'd
Comment 2 Oleg Ginzburg 2015-01-22 10:47:04 UTC 
Created attachment 152006 [details]
add ONLY_FOR_ARCHS=i386 amd64 for cbsd port
Comment 3 Oleg Ginzburg 2015-01-22 10:48:40 UTC 
(In reply to Sean Bruno from comment #0)

Yes, I agree with the addition ONLY_FOR_ARCHS=i386 amd64 until I find the possibility to test the port to other architectures.

Can you add? Or I'll do it at the next update version.
Comment 4 Sean Bruno freebsd_committer freebsd_triage 2015-01-22 14:19:55 UTC 
Ugh ... I typo'd the pr number when I committed the update.

https://svnweb.freebsd.org/ports?view=revision&revision=377654

I added a REASON variable so folks know why this is being added.  Good enough?
Comment 5 Oleg Ginzburg 2015-01-22 14:28:05 UTC 
(In reply to Sean Bruno from comment #4)

All fine, Thanks!
Comment 6 Dmitry Marakasov freebsd_committer freebsd_triage 2015-01-23 12:28:57 UTC 
What is the justification for this PR? cbsd builds fine on armv6 here.
Comment 7 Sean Bruno freebsd_committer freebsd_triage 2015-01-23 16:29:16 UTC 
(In reply to Dmitry Marakasov from comment #6)

Oh?  I seem to get a failure looking for SSE4.1 instructions (compiling via QEMU).

http://chips.ysv.freebsd.org/data/11armv6-11armv6/2015-01-19_15h28m19s/logs/errors/cbsd-10.1.3.log
Comment 8 Dmitry Marakasov freebsd_committer freebsd_triage 2015-01-23 18:01:33 UTC 
Hm, interesting. Here (under qemu as well) it compiles fine despite the error with -j8, however fails with DISABLE_MAKE_JOBS. Maybe there's some problem with make which doesn't catch child process failures on arm. Anyway you are right and the bug is valid.