On Mon, 2003-02-03 at 22:59, Keith Whitwell wrote: > Michel Dänzer wrote: > > On Mon, 2003-02-03 at 22:31, Chris Ison wrote: > > > >>Yeah, it works, > > > > > > Great! > > > > > >>just a little weary of it cause remembering without the read it won't work, > >>but it still appears to work with the read at the end. > > > > > > Well, I think Ben has given the best explanation of what's going on so > > far. Keith, are you happy with this? > > Relatively, though a little uncertain why we hadn't seen problems with this > before.
Well, I think we have. I've seen lots of reports of random crashes after hours to days with AGP cards, with no activity when the crash occurs. Two examples from my own experience are the AGP Radeon 7500 I used to have in an Athlon box at work, which would usually die after a couple days, and an AGP Radeon 7200 in a Cube here which used to crash after an hour maximum, but now it's been running for several hours with Chris' patch (going to test the one based on Ben's suggestions afterwards). Also, we used to at least have a radeon_flush_write_combining() call before the ring write pointer update when it was still done in ADVANCE_RING(). That probably got lost when COMMIT_RING() was split out? -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
