This is a patch for counting the number of pages for bounce buffer.
It's shown in /proc/vmstat.
Currently, the number of bounce pages are not counted anywhere.
So, if there are many bounce pages, it seems that there are
leaked pages. And it's difficult for a user to imagine the usage of
bounce pages. So, it's meaningful to show # of bouce pages.
Signed-off-by: KAMEZAWA Hiroyuki <[email protected]>
---
linux-2.6.12-rc2-mm3-kamezawa/include/linux/page-flags.h | 1 +
linux-2.6.12-rc2-mm3-kamezawa/mm/highmem.c | 2 ++
linux-2.6.12-rc2-mm3-kamezawa/mm/page_alloc.c | 1 +
3 files changed, 4 insertions(+)
diff -puN mm/highmem.c~count_bounce mm/highmem.c
--- linux-2.6.12-rc2-mm3/mm/highmem.c~count_bounce 2005-04-25 12:04:28.000000000 +0900
+++ linux-2.6.12-rc2-mm3-kamezawa/mm/highmem.c 2005-04-27 10:25:51.000000000 +0900
@@ -325,6 +325,7 @@ static void bounce_end_io(struct bio *bi
continue;
mempool_free(bvec->bv_page, pool);
+ dec_page_state(nr_bounce);
}
bio_endio(bio_orig, bio_orig->bi_size, err);
@@ -405,6 +406,7 @@ static void __blk_queue_bounce(request_q
to->bv_page = mempool_alloc(pool, q->bounce_gfp);
to->bv_len = from->bv_len;
to->bv_offset = from->bv_offset;
+ inc_page_state(nr_bounce);
if (rw == WRITE) {
char *vto, *vfrom;
diff -puN mm/page_alloc.c~count_bounce mm/page_alloc.c
--- linux-2.6.12-rc2-mm3/mm/page_alloc.c~count_bounce 2005-04-27 10:15:39.000000000 +0900
+++ linux-2.6.12-rc2-mm3-kamezawa/mm/page_alloc.c 2005-04-27 10:17:18.000000000 +0900
@@ -1902,6 +1902,7 @@ static char *vmstat_text[] = {
"nr_page_table_pages",
"nr_mapped",
"nr_slab",
+ "nr_bounce",
"pgpgin",
"pgpgout",
diff -puN include/linux/page-flags.h~count_bounce include/linux/page-flags.h
--- linux-2.6.12-rc2-mm3/include/linux/page-flags.h~count_bounce 2005-04-27 10:23:15.000000000 +0900
+++ linux-2.6.12-rc2-mm3-kamezawa/include/linux/page-flags.h 2005-04-27 10:24:11.000000000 +0900
@@ -89,6 +89,7 @@ struct page_state {
unsigned long nr_page_table_pages;/* Pages used for pagetables */
unsigned long nr_mapped; /* mapped into pagetables */
unsigned long nr_slab; /* In slab */
+ unsigned long nr_bounce; /* pages for bounce buffers */
#define GET_PAGE_STATE_LAST nr_slab
/*
_
In article <[email protected]> (at Wed, 27 Apr 2005 15:42:13 +0900), KAMEZAWA Hiroyuki <[email protected]> says:
> diff -puN include/linux/page-flags.h~count_bounce include/linux/page-flags.h
> --- linux-2.6.12-rc2-mm3/include/linux/page-flags.h~count_bounce 2005-04-27 10:23:15.000000000 +0900
> +++ linux-2.6.12-rc2-mm3-kamezawa/include/linux/page-flags.h 2005-04-27 10:24:11.000000000 +0900
> @@ -89,6 +89,7 @@ struct page_state {
> unsigned long nr_page_table_pages;/* Pages used for pagetables */
> unsigned long nr_mapped; /* mapped into pagetables */
> unsigned long nr_slab; /* In slab */
> + unsigned long nr_bounce; /* pages for bounce buffers */
> #define GET_PAGE_STATE_LAST nr_slab
>
> /*
I think you need to change GET_PAGE_STATE_LAST as well.
--yoshfuji
KAMEZAWA Hiroyuki <[email protected]> wrote:
>
> diff -puN include/linux/page-flags.h~count_bounce include/linux/page-flags.h
> --- linux-2.6.12-rc2-mm3/include/linux/page-flags.h~count_bounce 2005-04-27 10:23:15.000000000 +0900
> +++ linux-2.6.12-rc2-mm3-kamezawa/include/linux/page-flags.h 2005-04-27 10:24:11.000000000 +0900
> @@ -89,6 +89,7 @@ struct page_state {
> unsigned long nr_page_table_pages;/* Pages used for pagetables */
> unsigned long nr_mapped; /* mapped into pagetables */
> unsigned long nr_slab; /* In slab */
> + unsigned long nr_bounce; /* pages for bounce buffers */
> #define GET_PAGE_STATE_LAST nr_slab
That's not really right.
There are two functions: get_page_state() and get_full_page_state().
get_page_state() only gets those fields up to and including
GET_PAGE_STATE_LAST. The way you have the code laid out there implies that
get_page_state() also calculates the total nr_bounce, only it doesn't.
I'll fix it up.
Andrew Morton wrote:
>
> That's not really right.
>
> There are two functions: get_page_state() and get_full_page_state().
> get_page_state() only gets those fields up to and including
> GET_PAGE_STATE_LAST. The way you have the code laid out there implies that
> get_page_state() also calculates the total nr_bounce, only it doesn't.
>
> I'll fix it up.
>
Oh, I misunderstood that.
Thank you.
-- Kame