2013-07-24 18:17:36

by Adrian Brzezinski

Subject: [PATCH 1/1] fs/proc: cached = pagecache - shmem

Cached value taken from /proc/meminfo counts also shmem pages.
This can be very confusing when we encounter OOM problems.

Maybe the better solution would be, to not increment NR_FILE_PAGES,
when we have NR_SHMEM page? Should I rewrote and resend this patch?

Signed-off-by: Adrian Brzezinski <[email protected]>
---
fs/proc/meminfo.c | 7 +++++++
1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 5aa847a..d00aef5 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -40,7 +40,12 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
allowed = ((totalram_pages - hugetlb_total_pages())
* sysctl_overcommit_ratio / 100) + total_swap_pages;

+ /*
+ * NR_SHMEM pages can't be treated as a cache,
+ * but they are counted in NR_FILE_PAGES
+ */
cached = global_page_state(NR_FILE_PAGES) -
+ global_page_state(NR_SHMEM) -
total_swapcache_pages() - i.bufferram;
if (cached < 0)
cached = 0;
@@ -59,6 +64,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
"Buffers: %8lu kB\n"
"Cached: %8lu kB\n"
"SwapCached: %8lu kB\n"
+ "PageCache: %8lu kB\n"
"Active: %8lu kB\n"
"Inactive: %8lu kB\n"
"Active(anon): %8lu kB\n"
@@ -111,6 +117,7 @@ static int meminfo_proc_show(struct seq_file *m,
void *v)
K(i.bufferram),
K(cached),
K(total_swapcache_pages()),
+ K(global_page_state(NR_FILE_PAGES)),
K(pages[LRU_ACTIVE_ANON] + pages[LRU_ACTIVE_FILE]),
K(pages[LRU_INACTIVE_ANON] + pages[LRU_INACTIVE_FILE]),
K(pages[LRU_ACTIVE_ANON]),
--
1.7.1

2013-08-08 11:12:34

by tingwei liu

Subject: Re: [PATCH 1/1] fs/proc: cached = pagecache - shmem

On Thu, Jul 25, 2013 at 2:12 AM, Adrian Brzezinski
<[email protected]> wrote:
>
> Cached value taken from /proc/meminfo counts also shmem pages.
> This can be very confusing when we encounter OOM problems.
>
> Maybe the better solution would be, to not increment NR_FILE_PAGES,
> when we have NR_SHMEM page? Should I rewrote and resend this patch?
>
I suggest to consider push this patch to mainline.
> Signed-off-by: Adrian Brzezinski <[email protected]>
> ---
> fs/proc/meminfo.c | 7 +++++++
> 1 files changed, 7 insertions(+), 0 deletions(-)
>
> diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
> index 5aa847a..d00aef5 100644
> --- a/fs/proc/meminfo.c
> +++ b/fs/proc/meminfo.c
> @@ -40,7 +40,12 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
> allowed = ((totalram_pages - hugetlb_total_pages())
> * sysctl_overcommit_ratio / 100) + total_swap_pages;
>
> + /*
> + * NR_SHMEM pages can't be treated as a cache,
> + * but they are counted in NR_FILE_PAGES
> + */
> cached = global_page_state(NR_FILE_PAGES) -
> + global_page_state(NR_SHMEM) -
> total_swapcache_pages() - i.bufferram;
> if (cached < 0)
> cached = 0;
> @@ -59,6 +64,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
> "Buffers: %8lu kB\n"
> "Cached: %8lu kB\n"
> "SwapCached: %8lu kB\n"
> + "PageCache: %8lu kB\n"
> "Active: %8lu kB\n"
> "Inactive: %8lu kB\n"
> "Active(anon): %8lu kB\n"
> @@ -111,6 +117,7 @@ static int meminfo_proc_show(struct seq_file *m,
> void *v)
> K(i.bufferram),
> K(cached),
> K(total_swapcache_pages()),
> + K(global_page_state(NR_FILE_PAGES)),
> K(pages[LRU_ACTIVE_ANON] + pages[LRU_ACTIVE_FILE]),
> K(pages[LRU_INACTIVE_ANON] + pages[LRU_INACTIVE_FILE]),
> K(pages[LRU_ACTIVE_ANON]),
> --
> 1.7.1
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/