On Tue, 2008-05-27 at 17:48 +0200, Thomas Hellström wrote: > I guess as long as the X server is allowed to touch the ring, you need > to take the lock from > user-space for that operation.
It would be nice to fix everything all at once, but not very practical. So, compatibility with the existing X environment is required. Obviously, the kernel will need to make sure the DRM lock is held for all operations which might touch the ring. Gem will need checks for that, and then we'll need to figure out where mesa is failing to lock. I suspect it's in some other buffer mapping operation. > Isn't the lock taken properly in intelSpanRenderStart? Yeah, that one looks right. I'm traveling this week and didn't bring another laptop so I'm hoping Eric will have a chance to work this out. I'm hopeful that this will be easy to fix -- I'm hitting this often when re-using batch buffers, something we did in the old TTM branch but hadn't done in GEM until this weekend. That, of course, eliminates the overhead of memory allocation from GEM operation, saving some of kernel CPU time and improving performance a bit. -- [EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part
------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
-- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
