[Devel] Re: [PATCH] remove BUG() in possible but rare condition

[Michal Hocko]

On Wed 11-04-12 15:10:24, Glauber Costa wrote:
> While stressing the kernel with with failing allocations today,
> I hit the following chain of events:
> 
> alloc_page_buffers():
> 
> 	bh = alloc_buffer_head(GFP_NOFS);
> 	if (!bh)
> 		goto no_grow; <= path taken
> 
> grow_dev_page():
>         bh = alloc_page_buffers(page, size, 0);
>         if (!bh)
>                 goto failed;  <= taken, consequence of the above
> 
> and then the failed path BUG()s the kernel.
> 
> The failure is inserted a litte bit artificially, but even then,
> I see no reason why it should be deemed impossible in a real box.
> 
> Even though this is not a condition that we expect to see
> around every time, failed allocations are expected to be handled,
> and BUG() sounds just too much. As a matter of fact, grow_dev_page()
> can return NULL just fine in other circumstances, so I propose we just
> remove it, then.

I am not familiar with the code much but a trivial call chain walk up to
write_dev_supers (in btrfs) shows that we do not check for the return value
from __getblk so we would nullptr and there might be more. 
I guess these need some treat before the BUG might be removed, right?

> 
> Signed-off-by: Glauber Costa <glommer at parallels.com>
> CC: Linus Torvalds <torvalds at linux-foundation.org>
> CC: Andrew Morton <akpm at linux-foundation.org>
> ---
>  fs/buffer.c |    1 -
>  1 files changed, 0 insertions(+), 1 deletions(-)
> 
> diff --git a/fs/buffer.c b/fs/buffer.c
> index 36d6665..351e18e 100644
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -985,7 +985,6 @@ grow_dev_page(struct block_device *bdev, sector_t block,
>  	return page;
>  
>  failed:
> -	BUG();
>  	unlock_page(page);
>  	page_cache_release(page);
>  	return NULL;
> -- 
> 1.7.7.6
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo at 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 at kvack.org"> email at kvack.org </a>

-- 
Michal Hocko
SUSE Labs
SUSE LINUX s.r.o.
Lihovarska 1060/12
190 00 Praha 9    
Czech Republic