Re: [PATCH 3/3] ia64: Call migration code on correctable errors v2

[Christoph Lameter]

On Thu, 1 May 2008, Russ Anderson wrote:

> +int
> +freeOneBadPage(unsigned long paddr)
> +{
> +	struct page *page, *page2, *target;
> +
> +	/*
> +	 * Verify page address
> +	 */
> +	target = phys_to_page(paddr);
> +	list_for_each_entry_safe(page, page2, &badpagelist, lru) {
> +		if (page != target)
> +			continue;
> +
> +		ClearPageMemError(page);        /* Mark the page as good */
> +		move_to_lru(page);
> +		list_del(&page->lru);

This is the wrong order. move_to_lru() uses the lru field. So do the 
list_del first then the move. Note that there is a function 
putback_lru_pages() that is used to put a list of pages back to the lru if 
migration has failed and we give up on them. Could you use 
putback_lru_pages() and then avoid the exporting of move_to_lru()?


> +
> +int
> +freeAllBadPages(void)
> +{
> +	struct page *page, *page2;
> +
> +	list_for_each_entry_safe(page, page2, &badpagelist, lru) {
> +		ClearPageMemError(page);        /* Mark the page as good */
> +		totalbad_pages--;
> +	}
> +	putback_lru_pages(&badpagelist);
> +	return 0;
> +}

Ahh.. Here you use it.

> Index: linus/include/asm-ia64/page.h
> ===================================================================
> --- linus.orig/include/asm-ia64/page.h	2008-05-01 19:36:40.000000000 -0500
> +++ linus/include/asm-ia64/page.h	2008-05-01 19:36:49.000000000 -0500
> @@ -122,6 +122,7 @@ extern unsigned long max_low_pfn;
>  #endif
>  
>  #define page_to_phys(page)	(page_to_pfn(page) << PAGE_SHIFT)
> +#define	phys_to_page(kaddr)	(pfn_to_page(kaddr >> PAGE_SHIFT))

#define phys_to_page virt_to_page ?

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