Subject: Re: [linux-audio-user] MusE Race Condition
From: Robert Jonsson (rj_AT_spamatica.se)
Date: Sat Aug 07 2004 - 00:13:49 EEST
fredagen den 6 augusti 2004 18.14 skrev Christian Henz:
> On Fri, Aug 06, 2004 at 08:46:59AM -0700, davidrclark_AT_earthlink.net wrote:
> > Thanks to those who responded, and in particular to Christian Henz.
> > I had not seen this before (from Paul Davies):
> >
> > "The implementation of DRI by certain video interface drivers means
> > that we end up trying to lock the video memory as well, and this tends
> > to fail for various reasons."
>
> It works with Linux 2.4.x though, so maybe one should bother dri-devel
> again and also let linux-kernel know.
Hi,
Currently I'm running pretty much the configuration that David suggested
without these problems. Though I'm not sure about which driver for the Radeon
card is the culprit? I'm using ATI's drivers.
I am doing one other thing though that MIGHT be what makes it work. I added
this line to the section "Device" in XF86Config:
VideoRam 32768
I don't do games so it's not a problem for me (not that 32mb is that little)
I added this because I had issues with memory locking failed (under kernel
2.4). It seems that QT indeed maps video memory into it's memory space, and
not only that, it appears to add it several times. I have a Radeon 9200 with
128mb memory, which should mean that memory would increase with 128mb but I
remember launching MusE ,which normally takes about 50mb of memory in debug
build, taking over 300mb of memory! My (non verified) conclusion was that it
has added the video memory several times.
I did however not experience any problems as serious as what is being
described and I don't recall ever having seen it, apps failed to launch
though.
Do try this trick and let me know if this makes any difference. If it does
not, then I assume there are more pieces to this puzzle that still are not
identified.
Regards,
Robert
-- http://spamatica.se/music/
This archive was generated by hypermail 2b28 : Sat Aug 07 2004 - 00:17:42 EEST