Re: [RFC Patch] use MTRR for write combining if PAT is not available

[Thomas Schlichter <thomas.schlichter@web.de>]

Suresh Siddha wrote:
> On Thu, 2009-10-22 at 05:14 -0700, H. Peter Anvin wrote:
> > On 10/22/2009 09:08 PM, Thomas Schlichter wrote:
> > > And in that case (shared "struct file", one single release() call in the end) this
> > > implementation should be completely safe...
> > 
> > struct file is shared between forked processes.
> 
> That is correct. But I am referring to the ref-count getting incremented
> in Thomas's patch only in the pci_mmap_page_range() which will be called
> only during first mmap.
> 
> We need to keep track of the counts of later forks too.

When forked processes do mmap() PCI additional memory,
pci_mmap_page_range() will be called again and the corresponding (shared)
mtrr_usage_count wil be incremented. So we do keep track of later forks...

Yes, the MTRR reference count has nothing to do with the processes using this
memory, but if you want this, arch/x86/kernel/cpu/mtrr/if.c must be changed, too.

Regards,
  Thomas