Andi Kleen wrote:
The DRM code in Linux 2.5.73 has a new radeon getparam that is clearly
not 64bit clean. It tries to return a pointer to an lock in an int,
which clearly cannot work. The AMD64 compiler prints a warning for this.

RADEON_PARAM_SAREA_HANDLE does not seem to be used in the code, so I just disabled it.

Another issue I noticed is that the padding of the hw_lock is wrong
on P4. It only pads to 64bytes, but the L2 cache size there is 128 bytes. Use the Linux ____cacheline_aligned macro for that.

I don't think that will work. The problem is that those locks go in datastructures that are shared between user and kernel. The user-mode part gets built and distributed with XFree86. The kernel-mode part comes with the kernel (duh). They have to be in sync, for all sub-architectures. If our user-mode part is built for i586, but the kernel is built for P4, won't that cause problems? In either case, installing a new kernel with that macro with a pre-installed version of XFree86 (say, 4.1.0) will cause a binary compatability problem.


This is also a potential problem on some PowerPC flavors that have 128 byte cache-lines.



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to