LKML Archive on
help / color / mirror / Atom feed
From: Jeff Garzik <>
To: IDE/ATA development list <>
Cc: Tejun Heo <>, Jens Axboe <>,
	Linus Torvalds <>,
	Linux Kernel <>
Subject: Re: request to revert libata-convert-to-block-tagging patches
Date: Mon, 10 Nov 2008 17:55:16 -0500	[thread overview]
Message-ID: <> (raw)
In-Reply-To: <>

Tejun Heo wrote:
> Hello, all.
> I went through libata-convert-to-block-tagging today and found several
> issues.
> 1. libata internal data structure for command context (qc) allocation is
> bound to tag allocation, which means that block layer tagging should be
> enabled for all controllers which have can_queue > 1.
> 2. blk-tag offsets allocation for non-sync requests.  I'm not confident
> this is safe.  Till now, if there was only single command in flight for
> the port, it was guaranteed that the qc gets tag zero whether the device
> is NCQ capable or not.  qc allocation is tied tightly with hardware
> command slot allocation and I don't think it's wise to change this
> assumption.
> #1 is easy to fix but #2 requires either adding a spinlock or two atomic
> variables to struct blk_queue_tag to keep the current behavior while
> guaranteeing that tags are used in order.  Also, there's delay between
> libata marks a request complete and the request actually gets completed
> and the tag is freed.  If another request gets issued inbetween, the tag
> number can't be guaranteed.  This can be worked around by re-mapping tag
> number in libata depending on command type but, well then, it's worse
> than the original implementation.
> So, please revert the following commits.
> 43a49cbdf31e812c0d8f553d433b09b421f5d52c
> e013e13bf605b9e6b702adffbe2853cfc60e7806
> 2fca5ccf97d2c28bcfce44f5b07d85e74e3cd18e

A bit late, since they're already in, but, ACK.  (I'm on East Coast 
Vampire time, apparently)

Now that this is resolved, please allow me a bit of grumbling.  I always 
thought the original course -- 2.6.29 -- was best for these patches.  I 
had even queued them for 2.6.29, when they found their way into 2.6.28 
anyway.  Without /any/ testing by libata maintainers or linux-next. 
Without even being tested on non-NCQ setups, apparently.

The process broke down completely with this patchset :(

I still want to see this stuff in 2.6.29 though; it is the right way to 
go: following the theme of using block rather than SCSI bits in libata 
generic code [when those bits are, themselves, generic rather than 


  parent reply	other threads:[~2008-11-10 22:55 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-11-10  5:19 Tejun Heo
2008-11-10  5:30 ` Tejun Heo
2008-11-10  5:48   ` [PATCH] libata: revert convert-to-block-tagging patches Tejun Heo
2008-11-10 12:05 ` request to revert libata-convert-to-block-tagging patches Jens Axboe
2008-11-10 12:09   ` Jens Axboe
2008-11-10 16:00     ` Linus Torvalds
2008-11-10 17:10       ` Jens Axboe
2008-11-10 13:24   ` Tejun Heo
2008-11-10 16:03     ` Linus Torvalds
2008-11-10 22:55 ` Jeff Garzik [this message]
2008-11-12  1:11   ` Linus Torvalds

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:

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

  git send-email \ \ \ \ \ \ \ \
    --subject='Re: request to revert libata-convert-to-block-tagging patches' \

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).