lkml.org 
[lkml]   [2012]   [Aug]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH] block: replace __getblk_slow misfix by grow_dev_page fix
Date
Hugh Dickins <hughd@google.com> writes:

> [PATCH] block: replace __getblk_slow misfix by grow_dev_page fix
>
> Commit 91f68c89d8f3 ("block: fix infinite loop in __getblk_slow")
> is not good: a successful call to grow_buffers() cannot guarantee
> that the page won't be reclaimed before the immediate next call to
> __find_get_block(), which is why there was always a loop there.
[snip]
> Revert 91f68c89d8f3, restoring __getblk_slow() to how it was (plus
> a checkpatch nitfix). Simplify the interface between grow_buffers()
> and grow_dev_page(), and avoid the infinite loop beyond end of device
> by instead checking init_page_buffers()'s end_block there (I presume
> that's more efficient than a repeated call to blkdev_max_block()),
> returning -ENXIO to __getblk_slow() in that case.
>
> And remove akpm's ten-year-old "__getblk() cannot fail ... weird"
> comment, but that is worrying: are all users of __getblk() really
> now prepared for a NULL bh beyond end of device, or will some oops??

Hi, Hugh,

Thanks for digging into this. I had wondered whether that loop was
required for other purposes, and you've verified that it was. I mostly
like your fix. Returning end_block from init_page_buffers is a little
strange, but I understand not wanting to redo the call to
blkdev_max_block.

I went ahead to re-tested with the reproducer that I had, and your patch
works fine. Thanks again for tracking this down, and sorry I wasn't
more diligent to begin with.

Reviewed-and-Tested-by: Jeff Moyer <jmoyer@redhat.com>


\
 
 \ /
  Last update: 2012-08-23 23:41    [W:1.431 / U:2.088 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site