From: F. <j_r...@ya...> - 2002-06-22 14:45:48
|
On Thu, Jun 20, 2002 at 08:05:00PM +0100, José Fonseca wrote: >On Thu, Jun 20, 2002 at 02:05:45PM -0400, Leif Delgass wrote: >>On Thu, 20 Jun 2002, Leif Delgass wrote: >> >> ... >> >>OK, I was missing the obvious problem. ;) The end-of-list flag is set for >>almost all the descriptors on the ring. More likely this is a bug in the >>clear_bit. Is it a problem that "mask" is used as a variable both in the >>RING* macros and the inline clear_bit? >> >Stupid me..! I dont't think that clear_bit preserves endianess. That 31 >has to be something different on PowerPC. Ok. I've fixed this in CVS. I don't think that the use of a variable "mask" inside the inline function represents any problem, but I'll eventually rename the local variables in the MACROS to eliminate this kind of doubts. José Fonseca |