linux-kernel - Re: [patch -mm] mm, page_alloc: warn_alloc nodemask is NULL when cpusets are disabled

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <279f10c2-3eaa-c641-094f-3070db67d84f@suse.cz>
Date:   Thu, 19 Jan 2017 08:29:45 +0100
From:   Vlastimil Babka <vbabka@...e.cz>
To:     David Rientjes <rientjes@...gle.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...nel.org>
Cc:     Johannes Weiner <hannes@...xchg.org>, Mel Gorman <mgorman@...e.de>,
        linux-mm@...ck.org, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [patch -mm] mm, page_alloc: warn_alloc nodemask is NULL when
 cpusets are disabled

On 01/18/2017 10:51 PM, David Rientjes wrote:
> The patch "mm, page_alloc: warn_alloc print nodemask" implicitly sets the 
> allocation nodemask to cpuset_current_mems_allowed when there is no 
> effective mempolicy.  cpuset_current_mems_allowed is only effective when 
> cpusets are enabled, which is also printed by warn_alloc(), so setting 
> the nodemask to cpuset_current_mems_allowed is redundant and prevents 
> debugging issues where ac->nodemask is not set properly in the page 
> allocator.
> 
> This provides better debugging output since 
> cpuset_print_current_mems_allowed() is already provided.
> 
> Signed-off-by: David Rientjes <rientjes@...gle.com>

Yes, with my current cpuset vs mempolicy debugging experience, this is
more useful (except how both nodemask and mems_allowed can change under
us, so what we print here is not necessarily the same that what
get_page_from_freelist() has seen, but that's another thing...).

But I would suggest you change the oom killer's dump_header() the same
way than warn_alloc().

Thanks,
Vlastimil

> ---
>  mm/page_alloc.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -3037,7 +3037,6 @@ void warn_alloc(gfp_t gfp_mask, nodemask_t *nodemask, const char *fmt, ...)
>  	va_list args;
>  	static DEFINE_RATELIMIT_STATE(nopage_rs, DEFAULT_RATELIMIT_INTERVAL,
>  				      DEFAULT_RATELIMIT_BURST);
> -	nodemask_t *nm = (nodemask) ? nodemask : &cpuset_current_mems_allowed;
>  
>  	if ((gfp_mask & __GFP_NOWARN) || !__ratelimit(&nopage_rs) ||
>  	    debug_guardpage_minorder() > 0)
> @@ -3051,11 +3050,16 @@ void warn_alloc(gfp_t gfp_mask, nodemask_t *nodemask, const char *fmt, ...)
>  	pr_cont("%pV", &vaf);
>  	va_end(args);
>  
> -	pr_cont(", mode:%#x(%pGg), nodemask=%*pbl\n", gfp_mask, &gfp_mask, nodemask_pr_args(nm));
> +	pr_cont(", mode:%#x(%pGg), nodemask=", gfp_mask, &gfp_mask);
> +	if (nodemask)
> +		pr_cont("%*pbl\n", nodemask_pr_args(nodemask));
> +	else
> +		pr_cont("(null)\n");
> +
>  	cpuset_print_current_mems_allowed();
>  
>  	dump_stack();
> -	warn_alloc_show_mem(gfp_mask, nm);
> +	warn_alloc_show_mem(gfp_mask, nodemask);
>  }
>  
>  static inline struct page *
>