Linus Torvalds wrote: > > On Tue, 28 May 2002, Keith Whitwell wrote: > >>It's committed on the tcl branch now (it's also a minor optimization). >> > > Kevin, a quick look seems to imply that this _will_ cause a lock-up if > some ioctl were to try to do an operation that fills more than one full > ring-buffer. Which seems to be possible (even if it is probably > unlikely). In that case, "radeon_wait_ring()" will always fail, because > you haven't even told the card yet about the fact that you've updated the > ring. > > Should you not add a COMMIT_RING() in the overflow case to before the call > to radeon_wait_ring() in BEGIN_RING()? Or did I miss something?
Yes, I thought about this but it fell out of my brain... At the moment it's not a practical problem as the client doesn't send large buffers (the ring is 1mb), but I'll commit a fix. Keith _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
