[MPlayer-users] VESA stopped working with Radeon.

[Ruben Garcia]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



TJ wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> 
>> I and others on this list have had the same problem - VESA mode not
>> working with newer Linux distros.  (Redhat 8 in my case.)  I believe the
>> culprit is newer versions of glibc.  Several months back, someone traced
>> the cause of the problem to some of the code used in mplayer using an
>> older function call that's no longer supported in newer versions of
>> glibc.  I don't know enought about programming to be able to provide
>> much more meaningful detail.
> 
> 
> I've now tried to study this a bit more. It seems that the reason is not 
> the use of an older function call (vm86old) instead of a new one (vm86). 
> I tried changing mplayer-sources (or lrmi) to use the new call (via 
> assembler), but nothing really changed.
> 
>> Any chance of getting VESA working again?  I'm sure there are plenty of
>> mplayer users out there using VESA for tv-out, and the problem will only
>> grow as more users switch to newer distros.
> 
> 
> I've now become confident that the problem isn't in mplayer, so it's 
> unfortunately of little use to complain here :( The problem isn't even 
> in the lrmi (a newer version of which can be used without any change in 
> behaviour), but in the glibc itself. See the following post:
> 
> http://sources.redhat.com/ml/bug-glibc/2002-12/msg00024.html
> 
> It seems that the problem is in pthreads. I asked the author of the post 
> about the problem, and he said the problem should've vanished with glibc 
> 2.3.x-versions. It seems that's not correct or the problem we're seeing 
> is a bit elsewhere... Glibc/pthread-stuff is too much for me as I don't 
> know anything about them, but someone could ask about this on glibc 
> mailing lists - provided he'd know what to actually ask :)
> 
What about using "normal, unoptimized pthreads" when compiling glibc?
anyway, MPlayer is supposed to be single threaded, so the optimized 
pthreads wouldn't be too useful.

This is what the report says, so if we manage to watch TV-out with
unoptimized pthreads,
we can tell them with an example that the problem is still there.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org

iD8DBQE++H3Okzndl97WsggRArL6AKDMKiL3Dk/9CCd7bGXjp48KzQoNvACdFRgV
Dh7Tk8uOJ1f8Uc7PVzMCaZc=
=n0YO
-----END PGP SIGNATURE-----