On Die, 2002-09-03 at 21:20, Mike A. Harris wrote: > On Wed, 28 Aug 2002, Egbert Eich wrote: > > >Date: Wed, 28 Aug 2002 12:33:02 +0200 (MEST) > >From: Egbert Eich <[EMAIL PROTECTED]> > >To: [EMAIL PROTECTED] > >Reply-To: [EMAIL PROTECTED] > >Content-Type: text/plain; charset=us-ascii > >Subject: Re: Re: [Dri-devel] Radeon switch to VT and back X freeze > > POSSIBLE FIX > > > >Carlos O'Donell writes: > > > dri-devel, > > > > > > Just coming out of the woodwork to post a few ideas... > > > > > > It would be interesting to see if bus mastering gets disabled > > > on any other PCI devices? > > > > > > Anyone have a PCI logic analyzer? :) > > > > > > Would it be the norm to expect that a VT switch (going through > > > though the BIOS) would cause a BM disable in order to cancel > > > pending transactions (similar to soft reset)? > > > >A VT switch restores the PCI state to the situation when the > >Xserver was started. Therefore if BM was enabled when X was > >started (ie by the BIOS at boot time) it will remain enabled. > >Otherwise it will be disabled and will remain disabled until the > >driver explicitely reenables it. > >At the time this decision was made it sounded like the most reasonable > >approach to leave as little as possible behind when terminating/VT > >switching X. > > I'm pretty sure some of the X drivers still force busmastering on > when it is needed and found off during startup to get around BIOS > bugs in some laptop computers that do not enable BM but require > it. > > In the Radeon VT switch case however this is a different issue. > What is happening, is: > > Start computer, do not start X: Busmastering is enabled > Start up X: Busmastering still enabled > VTswitch: Now in console, Busmaster is disabled > VTswitch back to X: Busmastering disabled still - BOOM. > > Why busmastering is getting disabled during VTswitch is probably > something that should be determined. Why it only happens for > some people, and the busmaster patch fixes it is probably just a > hack around the real problem.
I don't think we can rely on it staying enabled, the question is if the drivers have to enable it or if common code could do it. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast _______________________________________________ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert