Re: [PATCH 7/8] badpage: ratelimit print_bad_pte and bad_page

From: Andrew Morton
Date: Tue Dec 02 2008 - 19:58:00 EST


On Mon, 1 Dec 2008 00:46:53 +0000 (GMT)
Hugh Dickins <hugh@xxxxxxxxxxx> wrote:

> print_bad_pte() and bad_page() might each need ratelimiting - especially
> for their dump_stacks, almost never of interest, yet not quite dispensible.
> Correlating corruption across neighbouring entries can be very helpful,
> so allow a burst of 60 reports before keeping quiet for the remainder
> of that minute (or allow a steady drip of one report per second).
>
> Signed-off-by: Hugh Dickins <hugh@xxxxxxxxxxx>
> ---
>
> mm/memory.c | 23 +++++++++++++++++++++++
> mm/page_alloc.c | 26 +++++++++++++++++++++++++-
> 2 files changed, 48 insertions(+), 1 deletion(-)
>
> --- badpage6/mm/memory.c 2008-11-28 20:40:48.000000000 +0000
> +++ badpage7/mm/memory.c 2008-11-28 20:40:50.000000000 +0000
> @@ -383,6 +383,29 @@ static void print_bad_pte(struct vm_area
> pmd_t *pmd = pmd_offset(pud, addr);
> struct address_space *mapping;
> pgoff_t index;
> + static unsigned long resume;
> + static unsigned long nr_shown;
> + static unsigned long nr_unshown;
> +
> + /*
> + * Allow a burst of 60 reports, then keep quiet for that minute;
> + * or allow a steady drip of one report per second.
> + */
> + if (nr_shown == 60) {
> + if (time_before(jiffies, resume)) {
> + nr_unshown++;
> + return;
> + }
> + if (nr_unshown) {
> + printk(KERN_EMERG
> + "Bad page map: %lu messages suppressed\n",
> + nr_unshown);
> + nr_unshown = 0;
> + }
> + nr_shown = 0;
> + }
> + if (nr_shown++ == 0)
> + resume = jiffies + 60 * HZ;
>
> mapping = vma->vm_file ? vma->vm_file->f_mapping : NULL;
> index = linear_page_index(vma, addr);
> --- badpage6/mm/page_alloc.c 2008-11-28 20:40:42.000000000 +0000
> +++ badpage7/mm/page_alloc.c 2008-11-28 20:40:50.000000000 +0000
> @@ -223,6 +223,30 @@ static inline int bad_range(struct zone
>
> static void bad_page(struct page *page)
> {
> + static unsigned long resume;
> + static unsigned long nr_shown;
> + static unsigned long nr_unshown;
> +
> + /*
> + * Allow a burst of 60 reports, then keep quiet for that minute;
> + * or allow a steady drip of one report per second.
> + */
> + if (nr_shown == 60) {
> + if (time_before(jiffies, resume)) {
> + nr_unshown++;
> + goto out;
> + }
> + if (nr_unshown) {
> + printk(KERN_EMERG
> + "Bad page state: %lu messages suppressed\n",
> + nr_unshown);
> + nr_unshown = 0;
> + }
> + nr_shown = 0;
> + }
> + if (nr_shown++ == 0)
> + resume = jiffies + 60 * HZ;
> +

gee, that's pretty elaborate. There's no way of using the
possibly-enhanced ratelimit.h?

--
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/