[Bug 729276] audioresample: Misdetected and/or misused intrinsics headers
GStreamer (bugzilla.gnome.org)
bugzilla at gnome.org
Fri May 9 01:03:36 PDT 2014
https://bugzilla.gnome.org/show_bug.cgi?id=729276
GStreamer | gst-plugins-base | 1.x
LRN <lrn1986> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |lrn1986 at gmail.com
--- Comment #3 from LRN <lrn1986 at gmail.com> 2014-05-09 08:03:29 UTC ---
What i've been told initially is that you should isolate SSE-, MMX- and any
other extended-instruction-using code into separate source files, compile them
with appropriate compiler options, and make sure that *any* functions from
these files are only called if runtime CPU detection permits.
I suspect that for some cases (such as code from speex) this is not very
convenient.
I raised this issue again today, here's what i've got:
ktietz: I can't do much about gcc's intrinsic header
ktietz: the compromize I suggest is to use instead intrin.h header and
omitting direct use of gcc's intrin-headers
LRN: isn't intrin.h MS-specific?
ktietz: well, it is a standard header present in our runtime for exactly
such issues
ktietz: if user insist to use directly the intrin-headers provided by gcc,
then they have to deal with the bugs there themself
LRN: well, GStreamer code is hardly W32-only, so i somehow doubt that they
would like to rely on W32-only header for such things...
ktietz: well, the issues reported here aren't W-specific AFAICS
ktietz: so at least they need to guard the use of those headers
--
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.
More information about the gstreamer-bugs
mailing list