Re: [patch] thp, memcg: split hugepage for memcg oom on cow

[David Rientjes <rientjes@google.com>]

On Tue, 10 Apr 2012, KAMEZAWA Hiroyuki wrote:

> > @@ -3502,13 +3503,24 @@ int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
> >  							  pmd, flags);
> >  	} else {
> >  		pmd_t orig_pmd = *pmd;
> > +		int ret;
> > +
> >  		barrier();
> >  		if (pmd_trans_huge(orig_pmd)) {
> >  			if (flags & FAULT_FLAG_WRITE &&
> >  			    !pmd_write(orig_pmd) &&
> > -			    !pmd_trans_splitting(orig_pmd))
> > -				return do_huge_pmd_wp_page(mm, vma, address,
> > -							   pmd, orig_pmd);
> > +			    !pmd_trans_splitting(orig_pmd)) {
> > +				ret = do_huge_pmd_wp_page(mm, vma, address, pmd,
> > +							  orig_pmd);
> > +				/*
> > +				 * If COW results in an oom memcg, the huge pmd
> > +				 * will already have been split, so retry the
> > +				 * fault on the pte for a smaller charge.
> > +				 */
> 
> 
> IIUC, do_huge_pmd_wp_page_fallback() can return VM_FAULT_OOM. So, this check
> is not related only to memcg.
> 

You're right, and if we do that then we infinitely loop trying to handle 
the pagefault instead of returning.  I'll post a v2 of the patch that 
fixes this, thanks for catching it.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>