Re: decode GFP flags in oom killer output.

["Michal Nazarewicz" <mina86@mina86.com>]

On Thu, 08 Mar 2012 00:39:39 +0100, Dave Jones <davej@redhat.com> wrote:

> Decoding these flags by hand in oom reports is tedious,
> and error-prone.
>
> Signed-off-by: Dave Jones <davej@redhat.com>
>
> diff -durpN '--exclude-from=/home/davej/.exclude' -u src/git-trees/kernel/linux/include/linux/gfp.h linux-dj/include/linux/gfp.h
> --- linux/include/linux/gfp.h	2012-01-11 16:54:21.736395499 -0500
> +++ linux-dj/include/linux/gfp.h	2012-03-06 13:17:37.294692113 -0500
> @@ -10,6 +10,7 @@
>  struct vm_area_struct;
> /* Plain integer GFP bitmasks. Do not use this directly. */
> +/* Update mm/oom_kill.c gfp_flag_texts when adding to/changing this list */
>  #define ___GFP_DMA		0x01u
>  #define ___GFP_HIGHMEM		0x02u
>  #define ___GFP_DMA32		0x04u
> diff -durpN '--exclude-from=/home/davej/.exclude' -u src/git-trees/kernel/linux/mm/oom_kill.c linux-dj/mm/oom_kill.c
> --- linux/mm/oom_kill.c	2012-01-17 17:54:14.541881964 -0500
> +++ linux-dj/mm/oom_kill.c	2012-03-06 13:17:44.071680535 -0500
> @@ -416,13 +416,40 @@ static void dump_tasks(const struct mem_
>  	}
>  }
>+static unsigned char *gfp_flag_texts[32] = {
> +	"DMA", "HIGHMEM", "DMA32", "MOVABLE",
> +	"WAIT", "HIGH", "IO", "FS",
> +	"COLD", "NOWARN", "REPEAT", "NOFAIL",
> +	"NORETRY", NULL, "COMP", "ZERO",
> +	"NOMEMALLOC", "HARDWALL", "THISNODE", "RECLAIMABLE",
> +	NULL, "NOTRACK", "NO_KSWAPD", "OTHER_NODE",
> +};
> +
> +static void decode_gfp_mask(gfp_t gfp_mask, char *out_string)
> +{
> +	unsigned int i;
> +
> +	for (i = 0; i < 32; i++) {
> +		if (gfp_mask & (1 << i)) {
> +			if (gfp_flag_texts[i])
> +				out_string += sprintf(out_string, "%s ", gfp_flag_texts[i]);
> +			else
> +				out_string += sprintf(out_string, "reserved! ");
> +		}
> +	}
> +	out_string = "\0";

Uh?  Did you mean “*out_string = 0;” which is redundant anyway?

Also, this leaves a trailing space at the end of the string.

> +}
> +
>  static void dump_header(struct task_struct *p, gfp_t gfp_mask, int order,
>  			struct mem_cgroup *memcg, const nodemask_t *nodemask)
>  {
> +	char gfp_string[80];

For ~0, the string will be 256 characters followed by a NUL byte byte at the end.
This combination may make no sense, but the point is that you need to take length
of the buffer into account, probably by using snprintf() and a counter.

>  	task_lock(current);
> -	pr_warning("%s invoked oom-killer: gfp_mask=0x%x, order=%d, "
> +	decode_gfp_mask(gfp_mask, gfp_string);
> +	pr_warning("%s invoked oom-killer: gfp_mask=0x%x [%s], order=%d, "
>  		"oom_adj=%d, oom_score_adj=%d\n",
> -		current->comm, gfp_mask, order, current->signal->oom_adj,
> +		current->comm, gfp_mask, gfp_string,
> +		order, current->signal->oom_adj,
>  		current->signal->oom_score_adj);
>  	cpuset_print_task_mems_allowed(current);
>  	task_unlock(current);
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
>


-- 
Best regards,                                         _     _
.o. | Liege of Serenely Enlightened Majesty of      o' \,=./ `o
..o | Computer Science,  Michał “mina86” Nazarewicz    (o o)
ooo +----<email/xmpp: mpn@google.com>--------------ooO--(_)--Ooo--

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>