Re: Misleading OOM messages

[Randy Dunlap]

Christoph Lameter wrote:
> Subject: Remove misleading kernel log entries about "Out of Memory" conditions
> 
> What we traditionally call an "out of memory" failure is mostly not really
> related to having enough physical memory. "out of memory" occurs when the
> memory reclaim attempts fail to provide enough memory for an allocation.
> 
> Typically there is a misconfiguration or kernel bug that is at the root
> of an out of memory issue. The message suggests that the machine does
> not have enough memory which is not true.
> 
> People have done strange things as a result of these messages. Some
> put more physical memory into their machines others limit the memory
> use of their applications with ulimit. Having a clear message avoids
> these reactions.
> 
> So change the messages to describe what actually went wrong.
> 
> Signed-off-by: Christoph Lameter <cl@xxxxxxxxxxxxxxxxxxxx>
> 
> ---
>  mm/oom_kill.c |   14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> Index: linux-2.6/mm/oom_kill.c
> ===================================================================
> --- linux-2.6.orig/mm/oom_kill.c	2009-05-12 12:37:52.000000000 -0500
> +++ linux-2.6/mm/oom_kill.c	2009-05-12 12:44:36.000000000 -0500
> @@ -387,7 +387,7 @@ static int oom_kill_process(struct task_
>  	struct task_struct *c;
> 
>  	if (printk_ratelimit()) {
> -		printk(KERN_WARNING "%s invoked oom-killer: "
> +		printk(KERN_WARNING "%s invoked process-killer: "
>  			"gfp_mask=0x%x, order=%d, oomkilladj=%d\n",
>  			current->comm, gfp_mask, order, current->oomkilladj);
>  		task_lock(current);
> @@ -519,7 +519,7 @@ static void __out_of_memory(gfp_t gfp_ma
> 
>  	if (sysctl_oom_kill_allocating_task)
>  		if (!oom_kill_process(current, gfp_mask, order, 0, NULL,
> -				"Out of memory (oom_kill_allocating_task)"))
> +				"Failure to reclaim enough memory (oom_kill_allocating_task)"))
>  			return;
>  retry:
>  	/*
> @@ -534,11 +534,11 @@ retry:
>  	/* Found nothing?!?! Either we hang forever, or we panic. */
>  	if (!p) {
>  		read_unlock(&tasklist_lock);
> -		panic("Out of memory and no killable processes...\n");
> +		panic("Failure to reclaim enough memory and no killable processes...\n");
>  	}
> 
>  	if (oom_kill_process(p, gfp_mask, order, points, NULL,
> -			     "Out of memory"))
> +			     "Memory reclaim failure"))
>  		goto retry;
>  }
> 
> @@ -563,7 +563,7 @@ void pagefault_out_of_memory(void)
>  		goto rest_and_return;
> 
>  	if (sysctl_panic_on_oom)
> -		panic("out of memory from page fault. panic_on_oom is selected.\n");
> +		panic("failure to reclaim enough memory. panic_on_oom is selected.\n");
> 
>  	read_lock(&tasklist_lock);
>  	__out_of_memory(0, 0); /* unknown gfp_mask and order */
> @@ -600,7 +600,7 @@ void out_of_memory(struct zonelist *zone
>  		return;
> 
>  	if (sysctl_panic_on_oom == 2)
> -		panic("out of memory. Compulsory panic_on_oom is selected.\n");
> +		panic("failure to reclaim enough memory. Compulsory panic_on_oom is selected.\n");
> 
>  	/*
>  	 * Check if there were limitations on the allocation (only relevant for
> @@ -617,7 +617,7 @@ void out_of_memory(struct zonelist *zone
> 
>  	case CONSTRAINT_NONE:
>  		if (sysctl_panic_on_oom)
> -			panic("out of memory. panic_on_oom is selected\n");
> +			panic("failure to enough reclaim memory. panic_on_oom is selected\n");

huh?

But I think that normal users won't know what reclaiming memory is anyway,
so the patch doesn't help IMO.

>  		/* Fall-through */
>  	case CONSTRAINT_CPUSET:
>  		__out_of_memory(gfp_mask, order);


-- 
~Randy
LPC 2009, Sept. 23-25, Portland, Oregon
http://linuxplumbersconf.org/2009/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/