On Monday 24 September 2007 10:23:29 Jan Beulich wrote:
quoted text
> >@@ -162,7 +198,7 @@ __change_page_attr(unsigned long address
> > 	/* on x86-64 the direct mapping set at boot is not using 4k pages */
> >  	BUG_ON(PageReserved(kpte_page));
> > 
> >-	save_page(kpte_page);
> >+	save_page(kpte_page, 0);
> > 	if (page_private(kpte_page) == 0)
> > 		revert_page(address, ref_prot);
> > 	return 0;
> 
> What is the point of continuing to launder kpte_page here? Page table pages
> never get their caching attributes changed, nor would their direct mapping
> ever change. (Same for i386, obviously.)

We can only free the page table after all the TLBs have been flushed;
otherwise other CPUs can walk already overwritten data as page tables (which can 
then cause various problems). This is similar to the lazy TLB flush logic in the 
standard VM.

We don't need to cache flush it though, that is why 0 is passed here.

Admittedly the function argument is a little bogus because the caller
could just do it; didn't change that yet.

-Andi

-
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/