Re: Linux 2.4.1-ac15

[Andrea Arcangeli <andrea@suse.de>]

On Fri, Feb 23, 2001 at 01:09:02PM -0800, Linus Torvalds wrote:
> 
> 
> On Fri, 23 Feb 2001, Andrea Arcangeli wrote:
> > 
> > I think that can't happen. Infact I think the whole section:
> > 
> > 		pmd = pmd_offset(pgd, address);
> > 		pmd_k = pmd_offset(pgd_k, address);
> > 
> > 		if (pmd_present(*pmd) || !pmd_present(*pmd_k))
> > 			goto bad_area_nosemaphore;
> > 		set_pmd(pmd, *pmd_k);
> > 		return;
> > 
> > is superflows.
> 
> No. Think about the differences in PAE and non-PAE mode. 
> 
> > The middle pagetable is shared and the pmd_t entry is set by vmalloc itself (it
> > just points to the as well shared pte), it's only the pgd that is setup lazily
> 
> In non-PAE mode, the pgd entry doesn't exist. pgd_present() returns a
> unconditional 1. Its' the pmd functions that kick in then.

Woops, I see, I was thinking only PAE mode indeed 8), sorry. Thanks for the
correction.

All the rest still applies (and the patch still looks fine for PAE mode). I
think I only need to rediff my patch after resurrecting the pmd thing inside
the cli critical section in a #ifndef CONFIG_X86_PAE region.

Andrea