Bug 238802 - devel/sdl20: enable altivec on powerpc64
Summary: devel/sdl20: enable altivec on powerpc64
Status: New
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Only Me
Assignee: Dmitry Marakasov
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-06-25 14:20 UTC by Piotr Kubaj
Modified: 2019-09-01 18:47 UTC (History)
3 users (show)

See Also:
bugzilla: maintainer-feedback? (amdmi3)


Attachments
patch (647 bytes, patch)
2019-06-25 14:20 UTC, Piotr Kubaj
no flags Details | Diff
v2 (1.20 KB, patch)
2019-09-01 18:06 UTC, Piotr Kubaj
pkubaj: maintainer-approval? (amdmi3)
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Piotr Kubaj freebsd_committer 2019-06-25 14:20:26 UTC
Created attachment 205329 [details]
patch

AFAIK there's no reason to have altivec disabled on powerpc64.
Comment 1 Justin Hibbits freebsd_committer 2019-07-05 19:41:54 UTC
It should be conditional, and default enabled.  Some powerpc64 implementations do not have altivec.
Comment 2 Piotr Kubaj freebsd_committer 2019-07-09 10:21:03 UTC
(In reply to Justin Hibbits from comment #1)
The proper solution would be to patch /usr/share/mk/bsd.cpu.mk and then check with:
.if defined(MACHINE_CPU) && ${MACHINE_CPU:Maltivec}
Comment 3 Piotr Kubaj freebsd_committer 2019-09-01 18:06:14 UTC
Created attachment 207072 [details]
v2

Is that patch ok, jhibbits?

Altivec is on by default and you can disable it.
Comment 4 Justin Hibbits freebsd_committer 2019-09-01 18:16:07 UTC
Looks okay to me, but why not name it SIMD instead of ALTIVEC?  That would match other ports.  Not a big deal either way.
Comment 5 Piotr Kubaj freebsd_committer 2019-09-01 18:30:33 UTC
(In reply to Justin Hibbits from comment #4)
SIMD is used when it enables also e.g. SSE on amd64.
Comment 6 Justin Hibbits freebsd_committer 2019-09-01 18:47:47 UTC
Yes, and the software does allow for SSE to be disabled.  So making it SIMD would naturally allow the x86 maintainer(s) to apply the same type of change for SSE/2/3/etc.

But that's something for the maintainer to weigh in on as well.  The change itself is fine.