public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Neil Brown <neilb@suse.de>
To: Tejun Heo <htejun@gmail.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH 025 of 35] Treat rq->hard_nr_sectors as setting an overriding limit in the size of the request
Date: Thu, 2 Aug 2007 13:31:16 +1000	[thread overview]
Message-ID: <18097.20484.698570.267166@notabene.brown> (raw)
In-Reply-To: message from Tejun Heo on Thursday August 2

On Thursday August 2, htejun@gmail.com wrote:
> 
> This is pretty confusing.  In all other places, bi_size -> #sector
> conversion is done by rounding down but only in blk_rq_bio_prep() it's
> being rounded up.
> 
> Is my following reasoning correct?
> 
> It was okay till now because unaligned requests don't get merged and
> also haven't done partial completions (end_that_request_first with
> partial count)?  So till now, hard_nr_sectors and nr_sectors didn't
> really matter for unaligned requests but now it matters because it's
> considered while iterating over bvecs in rq.

Yes, that reasoning matches mine.

> 
> If so, I think the correct thing to do would be changing bio_sectors()
> to round up first or let block layer measure transfer in bytes not in
> sectors.  I don't think everyone would agree with the latter tho.  I
> (tentatively) think it would be better to represent length in bytes
> tho.  A lot of requests which aren't aligned to 512 bytes pass through
> the block layer and the mismatch can result in subtle bugs.

I suspect that having a byte count in 'struct request' would make
sense too.  However I would rather avoid making that change myself - I
think it would require reading and understanding a lot more code....

I cannot see anything that would go wrong with rounding up bio_sectors
unconditionally, so I think I will take that approach for this patch
series.

Thanks.

NeilBrown

  reply	other threads:[~2007-08-02  3:31 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-07-31  2:15 [PATCH 000 of 35] Refactor block layer to improve support for stacked devices NeilBrown
2007-07-31  2:15 ` [PATCH 001 of 35] Replace bio_data with blk_rq_data NeilBrown
2007-07-31  2:15 ` [PATCH 002 of 35] Replace bio_cur_sectors with blk_rq_cur_sectors NeilBrown
2007-07-31  2:16 ` [PATCH 003 of 35] Introduce rq_for_each_segment replacing rq_for_each_bio NeilBrown
2007-07-31  2:16 ` [PATCH 004 of 35] Merge blk_recount_segments into blk_recalc_rq_segments NeilBrown
2007-07-31  2:16 ` [PATCH 005 of 35] Stop updating bi_idx, bv_len, bv_offset when a request completes NeilBrown
2007-08-01 14:54   ` Tejun Heo
2007-07-31  2:16 ` [PATCH 006 of 35] Only call bi_end_io once for any bio NeilBrown
2007-07-31  2:16 ` [PATCH 007 of 35] Drop 'size' argument from bio_endio and bi_end_io NeilBrown
2007-08-01 15:17   ` Tejun Heo
2007-07-31  2:16 ` [PATCH 008 of 35] Introduce bi_iocnt to count requests sharing the one bio NeilBrown
2007-08-01 15:49   ` Tejun Heo
2007-07-31  2:16 ` [PATCH 009 of 35] Remove overloading of bi_hw_segments in raid5 NeilBrown
2007-07-31  2:16 ` [PATCH 010 of 35] New function blk_req_append_bio NeilBrown
2007-08-01 15:54   ` Christoph Hellwig
2007-07-31  2:16 ` [PATCH 011 of 35] Stop exporting blk_rq_bio_prep NeilBrown
2007-07-31  2:16 ` [PATCH 012 of 35] Share code between init_request_from_bio and blk_rq_bio_prep NeilBrown
2007-07-31  2:16 ` [PATCH 013 of 35] Don't update bi_hw_*_size if we aren't going to merge NeilBrown
2007-08-01 15:57   ` Tejun Heo
2007-08-02  3:37     ` Neil Brown
2007-07-31  2:17 ` [PATCH 014 of 35] Change blk_phys/hw_contig_segment to take requests, not bios NeilBrown
2007-07-31  2:17 ` [PATCH 015 of 35] Move hw_front_size and hw_back_size from bio to request NeilBrown
2007-07-31  2:17 ` [PATCH 016 of 35] Centralise setting for REQ_NOMERGE NeilBrown
2007-07-31  2:17 ` [PATCH 017 of 35] Fix various abuse of bio fields in umem.c NeilBrown
2007-07-31  2:17 ` [PATCH 018 of 35] Remove bi_idx NeilBrown
2007-07-31  2:17 ` [PATCH 019 of 35] Convert bio_for_each_segment to fill in a fresh bio_vec NeilBrown
2007-08-01 16:21   ` Tejun Heo
2007-07-31  2:17 ` [PATCH 020 of 35] Add bi_offset and allow a bio to reference only part of a bi_io_vec NeilBrown
2007-07-31  2:17 ` [PATCH 021 of 35] Teach umem.c about bi_offset and to limit to bi_size NeilBrown
2007-07-31  2:17 ` [PATCH 022 of 35] Teach dm-crypt to honour bi_offset and bi_size NeilBrown
2007-07-31  2:17 ` [PATCH 023 of 35] Teach pktcdvd.c " NeilBrown
2007-07-31  2:17 ` [PATCH 024 of 35] Allow request bio list not to end with NULL NeilBrown
2007-07-31  2:17 ` [PATCH 025 of 35] Treat rq->hard_nr_sectors as setting an overriding limit in the size of the request NeilBrown
2007-08-01 17:44   ` Tejun Heo
2007-08-02  3:31     ` Neil Brown [this message]
2007-08-02  5:03       ` Tejun Heo
2007-07-31  2:18 ` [PATCH 026 of 35] Split any large bios that arrive at __make_request NeilBrown
2007-08-01 17:56   ` Tejun Heo
2007-08-02  0:49     ` Neil Brown
2007-08-02  2:59       ` Tejun Heo
2007-08-02  3:16         ` Neil Brown
2007-07-31  2:18 ` [PATCH 028 of 35] Split arbitrarily large requests to md/raid0 and md/linear NeilBrown
2007-07-31  2:18 ` [PATCH 029 of 35] Teach md/raid10 to split arbitrarily large bios NeilBrown
2007-07-31  2:18 ` [PATCH 030 of 35] Teach raid5 to split incoming bios NeilBrown
2007-07-31  2:18 ` [PATCH 031 of 35] Use bio_multi_split to fully split bios for pktcdvd NeilBrown
2007-07-31  2:18 ` [PATCH 032 of 35] Remove blk_queue_merge_bvec and bio_split and related code NeilBrown
2007-07-31  2:18 ` [PATCH 033 of 35] Simplify stacking of IO restrictions NeilBrown
2007-07-31  2:18 ` [PATCH 034 of 35] Simplify bio_add_page and raid1/raid10 resync which use it NeilBrown
2007-07-31  2:18 ` [PATCH 035 of 35] Simplify bio splitting in dm NeilBrown
2007-07-31 15:28 ` [PATCH 000 of 35] Refactor block layer to improve support for stacked devices Avi Kivity
2007-08-01 14:37   ` Tejun Heo
2007-08-01 15:52     ` John Stoffel
2007-08-01 15:59       ` Tejun Heo
2007-08-02  3:43       ` Neil Brown

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=18097.20484.698570.267166@notabene.brown \
    --to=neilb@suse.de \
    --cc=htejun@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox