Re: [PATCH][2.6.7-rc3-mm1] perfctr cpumask cleanup

["Randy.Dunlap" <rddunlap@osdl.org>]

On Thu, 10 Jun 2004 09:03:50 -0700 William Lee Irwin III wrote:

| Mikael wrote:
| >> Doesn't work because cpus_andnot() requires all three parameters
| >> to be lvalues. ... CPU_MASK_NONE ...
| 
| On Thu, Jun 10, 2004 at 09:01:57AM -0700, Paul Jackson wrote:
| > I think your other fix (also done by Bill Irwin), make the above possible:
| >  #define CPU_MASK_NONE							\
| > -{ {									\
| > +((cpumask_t) { {							\
| 
| I've posted this at least twice, I think once in isolation for some
| driver (MSI?) and once as part of the Alpha fixes.
| 
| Please get some cross-compilers together so we don't have every non-x86
| arch exploding at once. Alpha vs. cpu_possible_map was horrendous.

or submit such patches to OSDL PLM.  PLM will apply the patch
(to whatever base you say, although you may have to supply the
"base" also; linus and -mm bases are already there)
and then try to cross-compile it on ia32, ia64, ppc32, ppc64, alpha,
sparc32, sparc64, and x86_64.

E.g., here's the results of 2.6.7-rc3-bk3:
https://www.osdl.org/plm-cgi/plm?module=patch_info&patch_id=3010

--
~Randy