From: Jens Axboe <axboe@suse.de>
To: Alan Chandler <alan@chandlerfamily.org.uk>
Cc: linux-kernel@vger.kernel.org
Subject: Re: ide-cd problem
Date: Mon, 22 Nov 2004 14:02:02 +0100 [thread overview]
Message-ID: <20041122130202.GO10463@suse.de> (raw)
In-Reply-To: <E1CWDhN-00040Y-E6@home.chandlerfamily.org.uk>
On Mon, Nov 22 2004, Alan Chandler wrote:
> Jens Axboe writes:
>
> >On Mon, Nov 22 2004, Alan Chandler wrote:
> >>Jens Axboe writes:
>
> >>>400ns is the correctl value. Your writing is a little unclear to me -
> >>>did it work or not, with that change alone?
> >>>
> >>
> >>To be clear ...
> >>
> >>
> >>I have modified ide-cd.c with
> >>
> >>1) ndelay(400) at the head of cdrom_newpc_intr()
> >>
> >>2) Alan Cox's patch in the place he originally identified for it to go
> >>
> >>3) Some printk's in cdrom_newpc_intr() after the point where it reads the
> >>status and IREASON and length registers and just for the purposes of
> >>diagnostics.
> >>
> >>With only those changes it now works.
> >
> >You are not answering my question :-)
> >
> >Here's is Alans patch as I posted some mails ago. Does it work with that
> >alone?? I'm curious of it is enough. It should not be necessary to incur
> >extra delay in the interrupt handler, if it is invoked from a real irq.
>
> Sorry, I misunderstood what you meant. I presume you think that the
> interrupt may be triggered immediately the command packet has been sent but
> before 400ns delay had occurred.
>
> NO - with Alan's patch alone, this did not work.
>
> The delay seesm to be needed in the path between the interrupt occuring and
> the IDE_STATUS_REG being read.
>
> I had seen an note on a web site that said that there was two delays
> required in the ATA/ATAPI spec - the 400ns which Alan's patch deals with
> and a shorter delay (one PIO cycle) between busy being cleared and DRQ
> reaching the correct state where the technique had been to read the
> ALTSTATUS register. That was why I had tried that approach but found it
> not to work.
> (I have subsequently downloaded a copy of the full spec and haven't been
> able to find this - but then its just short of 500 pages of dense text:-)).
>
> Thinking about it now, I tried the ALTSTATUS delay before applying Alan's
> patch, so maybe its the some of the two delays that maybe necessary. If
> you think its appropriate I will try that again this evening.
I think the more correct patch is the following. It seems I was wrong in
assuming that the ide_intr() path already waited 400ns for us, I think
this should work for you. Can you test it?
===== drivers/ide/ide-iops.c 1.31 vs edited =====
--- 1.31/drivers/ide/ide-iops.c 2004-11-01 18:06:50 +01:00
+++ edited/drivers/ide/ide-iops.c 2004-11-22 13:59:27 +01:00
@@ -476,10 +476,8 @@
if (drive->waiting_for_dma)
return hwif->ide_dma_test_irq(drive);
-#if 0
/* need to guarantee 400ns since last command was issued */
- udelay(1);
-#endif
+ ndelay(400);
#ifdef CONFIG_IDEPCI_SHARE_IRQ
/*
--
Jens Axboe
next prev parent reply other threads:[~2004-11-22 13:03 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-11-20 18:42 ide-cd problem Alan Chandler
2004-11-20 19:47 ` Jens Axboe
2004-11-21 0:53 ` Alan Chandler
2004-11-21 8:56 ` Jens Axboe
2004-11-21 10:25 ` Alan Chandler
2004-11-21 16:13 ` Alan Chandler
2004-11-22 7:52 ` Alan Chandler
2004-11-22 8:01 ` Jens Axboe
2004-11-22 10:30 ` Alan Chandler
2004-11-22 10:51 ` Jens Axboe
2004-11-22 11:29 ` Alan Chandler
2004-11-22 11:31 ` Jens Axboe
2004-11-22 12:53 ` Alan Chandler
2004-11-22 13:02 ` Jens Axboe [this message]
2004-11-22 19:19 ` Alan Chandler
2004-11-22 23:48 ` Alan Chandler
2004-11-23 7:13 ` Alan Chandler
2004-11-23 14:51 ` Jens Axboe
2004-11-23 21:49 ` Alan Chandler
2004-11-26 23:39 ` Alan Chandler
2004-11-29 17:29 ` Bill Davidsen
2004-11-30 8:59 ` Alan Chandler
2004-12-10 21:32 ` ide-cd problem revisited - more brainpower needed Alan Chandler
2004-12-10 23:14 ` Alan Cox
2004-12-12 0:17 ` Alan Chandler
2004-12-12 11:39 ` Alan Cox
2004-12-12 13:34 ` Alan Chandler
2004-12-14 0:20 ` Alan Chandler
2004-12-16 15:56 ` Bill Davidsen
2004-12-17 23:59 ` Alan Chandler
2004-11-24 23:19 ` ide-cd problem Alan Cox
2004-11-25 15:29 ` Jens Axboe
2004-11-25 16:25 ` Alan Cox
2004-11-25 18:12 ` Jens Axboe
2004-11-25 18:45 ` Alan Chandler
2004-11-23 18:34 ` Jeff Garzik
2004-11-23 19:13 ` Jens Axboe
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=20041122130202.GO10463@suse.de \
--to=axboe@suse.de \
--cc=alan@chandlerfamily.org.uk \
--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;
as well as URLs for NNTP newsgroup(s).