lkml.org 
[lkml]   [2015]   [Dec]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v4 2/2] zram: try vmalloc() after kmalloc()
On (12/01/15 21:36), Sergey Senozhatsky wrote:
> When we're using LZ4 multi compression streams for zram swap, we found out
> page allocation failure message in system running test. That was not only
> once, but a few(2 - 5 times per test). Also, some failure cases were
> continually occurring to try allocation order 3.
>
> In order to make parallel compression private data, we should call
> kzalloc() with order 2/3 in runtime(lzo/lz4). But if there is no order
> 2/3 size memory to allocate in that time, page allocation fails. This
> patch makes to use vmalloc() as fallback of kmalloc(), this prevents page
> alloc failure warning.
>
> After using this, we never found warning message in running test, also It
> could reduce process startup latency about 60-120ms in each case.
>

Hello Kyeongdon,

just to make sure, the patch works fine for you and we can move
forward and Cc -stable. correct?

-ss

> For reference a call trace :
>
> Binder_1: page allocation failure: order:3, mode:0x10c0d0
> CPU: 0 PID: 424 Comm: Binder_1 Tainted: GW 3.10.49-perf-g991d02b-dirty #20
> Call trace:
> [<ffffffc0002069c8>] dump_backtrace+0x0/0x270
> [<ffffffc000206c48>] show_stack+0x10/0x1c
> [<ffffffc000cb51c8>] dump_stack+0x1c/0x28
> [<ffffffc0002bbfc8>] warn_alloc_failed+0xfc/0x11c
> [<ffffffc0002bf518>] __alloc_pages_nodemask+0x724/0x7f0
> [<ffffffc0002bf5f8>] __get_free_pages+0x14/0x5c
> [<ffffffc0002ed6a4>] kmalloc_order_trace+0x38/0xd8
> [<ffffffc0005d9738>] zcomp_lz4_create+0x2c/0x38
> [<ffffffc0005d78f0>] zcomp_strm_alloc+0x34/0x78
> [<ffffffc0005d7a58>] zcomp_strm_multi_find+0x124/0x1ec
> [<ffffffc0005d7c14>] zcomp_strm_find+0xc/0x18
> [<ffffffc0005d8fa0>] zram_bvec_rw+0x2fc/0x780
> [<ffffffc0005d9680>] zram_make_request+0x25c/0x2d4
> [<ffffffc00040f8ac>] generic_make_request+0x80/0xbc
> [<ffffffc00040f98c>] submit_bio+0xa4/0x15c
> [<ffffffc0002e8bb0>] __swap_writepage+0x218/0x230
> [<ffffffc0002e8c04>] swap_writepage+0x3c/0x4c
> [<ffffffc0002c7384>] shrink_page_list+0x51c/0x8d0
> [<ffffffc0002c7e88>] shrink_inactive_list+0x3f8/0x60c
> [<ffffffc0002c86c8>] shrink_lruvec+0x33c/0x4cc
> [<ffffffc0002c8894>] shrink_zone+0x3c/0x100
> [<ffffffc0002c8c10>] try_to_free_pages+0x2b8/0x54c
> [<ffffffc0002bf308>] __alloc_pages_nodemask+0x514/0x7f0
> [<ffffffc0002bf5f8>] __get_free_pages+0x14/0x5c
> [<ffffffc0003446cc>] proc_info_read+0x50/0xe4
> [<ffffffc0002f5204>] vfs_read+0xa0/0x12c
> [<ffffffc0002f59c8>] SyS_read+0x44/0x74
> DMA: 3397*4kB (MC) 26*8kB (RC) 0*16kB 0*32kB 0*64kB 0*128kB 0*256kB
> 0*512kB 0*1024kB 0*2048kB 0*4096kB = 13796kB
>
> [minchan@kernel.org: change vmalloc gfp and adding comment about gfp]
> [sergey.senozhatsky@gmail.com: tweak comments and styles]
> Signed-off-by: Kyeongdon Kim <kyeongdon.kim@lge.com>
> Signed-off-by: Minchan Kim <minchan@kernel.org>
> Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> ---
> drivers/block/zram/zcomp_lz4.c | 23 +++++++++++++++++++++--
> drivers/block/zram/zcomp_lzo.c | 23 +++++++++++++++++++++--
> 2 files changed, 42 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/block/zram/zcomp_lz4.c b/drivers/block/zram/zcomp_lz4.c
> index ee44b51..dd60831 100644
> --- a/drivers/block/zram/zcomp_lz4.c
> +++ b/drivers/block/zram/zcomp_lz4.c
> @@ -10,17 +10,36 @@
> #include <linux/kernel.h>
> #include <linux/slab.h>
> #include <linux/lz4.h>
> +#include <linux/vmalloc.h>
> +#include <linux/mm.h>
>
> #include "zcomp_lz4.h"
>
> static void *zcomp_lz4_create(void)
> {
> - return kzalloc(LZ4_MEM_COMPRESS, GFP_NOIO);
> + void *ret;
> +
> + /*
> + * This function can be called in swapout/fs write path
> + * so we can't use GFP_FS|IO. And it assumes we already
> + * have at least one stream in zram initialization so we
> + * don't do best effort to allocate more stream in here.
> + * A default stream will work well without further multiple
> + * streams. That's why we use NORETRY | NOWARN.
> + */
> + ret = kzalloc(LZ4_MEM_COMPRESS, GFP_NOIO | __GFP_NORETRY |
> + __GFP_NOWARN);
> + if (!ret)
> + ret = __vmalloc(LZ4_MEM_COMPRESS,
> + GFP_NOIO | __GFP_NORETRY | __GFP_NOWARN |
> + __GFP_ZERO | __GFP_HIGHMEM,
> + PAGE_KERNEL);
> + return ret;
> }
>
> static void zcomp_lz4_destroy(void *private)
> {
> - kfree(private);
> + kvfree(private);
> }
>
> static int zcomp_lz4_compress(const unsigned char *src, unsigned char *dst,
> diff --git a/drivers/block/zram/zcomp_lzo.c b/drivers/block/zram/zcomp_lzo.c
> index 683ce04..edc5499 100644
> --- a/drivers/block/zram/zcomp_lzo.c
> +++ b/drivers/block/zram/zcomp_lzo.c
> @@ -10,17 +10,36 @@
> #include <linux/kernel.h>
> #include <linux/slab.h>
> #include <linux/lzo.h>
> +#include <linux/vmalloc.h>
> +#include <linux/mm.h>
>
> #include "zcomp_lzo.h"
>
> static void *lzo_create(void)
> {
> - return kzalloc(LZO1X_MEM_COMPRESS, GFP_NOIO);
> + void *ret;
> +
> + /*
> + * This function can be called in swapout/fs write path
> + * so we can't use GFP_FS|IO. And it assumes we already
> + * have at least one stream in zram initialization so we
> + * don't do best effort to allocate more stream in here.
> + * A default stream will work well without further multiple
> + * streams. That's why we use NORETRY | NOWARN.
> + */
> + ret = kzalloc(LZO1X_MEM_COMPRESS, GFP_NOIO | __GFP_NORETRY |
> + __GFP_NOWARN);
> + if (!ret)
> + ret = __vmalloc(LZO1X_MEM_COMPRESS,
> + GFP_NOIO | __GFP_NORETRY | __GFP_NOWARN |
> + __GFP_ZERO | __GFP_HIGHMEM,
> + PAGE_KERNEL);
> + return ret;
> }
>
> static void lzo_destroy(void *private)
> {
> - kfree(private);
> + kvfree(private);
> }
>
> static int lzo_compress(const unsigned char *src, unsigned char *dst,
> --
> 2.6.2
>


\
 
 \ /
  Last update: 2015-12-18 02:01    [W:0.115 / U:0.064 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site