Re: [PATCH v2] nilfs2: remove double bio_put() in nilfs_end_bio_write() for BIO_EOPNOTSUPP error

[Vyacheslav Dubeyko <slava@dubeyko.com>]

Hi Ryusuke,

On Sat, 2013-07-27 at 12:02 +0900, Ryusuke Konishi wrote:

[snip]
> 
> This decrement looks wrong.
> 
> Otherwise, your change of nilfs_segbuf_submit_bio() is just a
> equivalent transformation and doesn't fix problem, that is, a mismatch
> of the number of calls between complete() and wait_for_completion().
> 
> In your patch, nilfs_end_bio_write() function calls the complete()
> routine as before even if it received an EOPNOTSUPP error.
> 
> In that case, segbuf->sb_nbio must be incremented to call
> wait_for_completion() the right number of times.
> 
> Note that wait_for_completion() is called based on the count of
> segbuf->sb_nbio even if nilfs_segbuf_submit_bio() returns an error.
> This is performed through the following path:
> 
>             nilfs_segbuf_submit_bio
>           nilfs_segbuf_submit_bh
>         nilfs_segbuf_write
>       nilfs_write_logs
>     nilfs_segctor_write
>   nilfs_segctor_do_construct
>     nilfs_segctor_abort_construction
>       nilfs_wait_on_logs
>         nilfs_segbuf_wait
>           wait_for_completion
>             (repeated for the number of times of segbuf->sb_nbio)
> 
> 
> If you think this is a separate problem, then it should be fixed in
> another patch and nilfs_segbuf_submit_bio() should not be touched in
> this patch.
> 

Sorry for delay with working on this patch. Now I am investigating
the issue (Kernel Bug: unable to handle kernel paging request) that it
was reported by Jérôme Poulin <jeromepoulin@gmail.com>. I suspect that
the reported issue is related to this patch. So, I feel a necessity to
investigate the issue more deeply before continuation to work on this
patch.

With the best regards,
Vyacheslav Dubeyko.


--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html