From: Jeff Garzik <jgarzik@pobox.com>
To: Jens Axboe <axboe@suse.de>
Cc: linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: Playing with SATA NCQ
Date: Fri, 27 May 2005 03:29:32 -0400 [thread overview]
Message-ID: <4296CC5C.5000807@pobox.com> (raw)
In-Reply-To: <20050527072046.GN1435@suse.de>
[-- Attachment #1: Type: text/plain, Size: 353 bytes --]
Jens Axboe wrote:
> I double checked this. If you agree to move the setting of QCFLAG_ACTIVE
> _after_ successful ap->ops->qc_issue(qc) and clear it _after_
> __ata_qc_complete(qc) then I can just use that bit and kill
> ATA_QCFLAG_ACCOUNT.
>
> What do you think?
Fine with me.
Keep in mind that the attached patch was applied recently...
Jeff
[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 1933 bytes --]
author Albert Lee <albertcc@tw.ibm.com> Fri, 29 Apr 2005 17:34:59 +0800
committer Jeff Garzik <jgarzik@pobox.com> Mon, 16 May 2005 02:46:59 -0400
[PATCH] libata: Prevent the interrupt handler from completing a command twice
Problem:
During the libata CD-ROM stress test, sometimes the "BUG: timeout
without command" error is seen.
Root cause:
Unexpected interrupt occurs after the ata_qc_complete() is called,
but before the SCSI error handler. The interrupt handler is invoked
before the SCSI error handler, and it clears the command by calling
ata_qc_complete() again. Later when the SCSI error handler is run,
the ata_queued_cmd is already gone, causing the "BUG: timeout without
command" error.
Changes:
- Use the ATA_QCFLAG_ACTIVE flag to prevent the interrupt handler
from completing the command twice, before the scsi_error_handler.
Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
diff --git a/drivers/scsi/libata-core.c b/drivers/scsi/libata-core.c
--- a/drivers/scsi/libata-core.c
+++ b/drivers/scsi/libata-core.c
@@ -2539,7 +2539,7 @@ static void atapi_request_sense(struct a
ata_sg_init_one(qc, cmd->sense_buffer, sizeof(cmd->sense_buffer));
qc->dma_dir = DMA_FROM_DEVICE;
- memset(&qc->cdb, 0, sizeof(ap->cdb_len));
+ memset(&qc->cdb, 0, ap->cdb_len);
qc->cdb[0] = REQUEST_SENSE;
qc->cdb[4] = SCSI_SENSE_BUFFERSIZE;
@@ -2811,6 +2811,7 @@ void ata_qc_complete(struct ata_queued_c
/* call completion callback */
rc = qc->complete_fn(qc, drv_stat);
+ qc->flags &= ~ATA_QCFLAG_ACTIVE;
/* if callback indicates not to complete command (non-zero),
* return immediately
@@ -3229,7 +3230,8 @@ irqreturn_t ata_interrupt (int irq, void
struct ata_queued_cmd *qc;
qc = ata_qc_from_tag(ap, ap->active_tag);
- if (qc && (!(qc->tf.ctl & ATA_NIEN)))
+ if (qc && (!(qc->tf.ctl & ATA_NIEN)) &&
+ (qc->flags & ATA_QCFLAG_ACTIVE))
handled |= ata_host_intr(ap, qc);
}
}
next prev parent reply other threads:[~2005-05-27 7:29 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-05-26 14:00 Playing with SATA NCQ Jens Axboe
2005-05-26 16:25 ` Jeff Garzik
2005-05-26 17:07 ` Jens Axboe
2005-05-26 17:11 ` Jens Axboe
2005-05-26 17:15 ` Jeff Garzik
2005-05-26 17:33 ` Jens Axboe
2005-05-26 19:49 ` Jeff Garzik
2005-05-26 20:30 ` Jens Axboe
2005-05-27 7:20 ` Jens Axboe
2005-05-27 7:29 ` Jeff Garzik [this message]
2005-05-27 7:33 ` Jens Axboe
2005-05-27 7:51 ` Jeff Garzik
2005-05-27 8:00 ` Jens Axboe
2005-05-27 8:23 ` Jeff Garzik
2005-05-26 21:50 ` Mark Lord
2005-05-27 6:28 ` Jens Axboe
2005-05-27 6:58 ` Jeff Garzik
2005-05-27 7:15 ` Jens Axboe
2005-05-27 4:41 ` Jeff Garzik
2005-05-27 6:39 ` Jens Axboe
2005-05-27 21:40 ` Michael Thonke
2005-05-27 22:16 ` Jeff Garzik
2005-05-27 22:30 ` Michael Thonke
2005-05-28 12:12 ` Jens Axboe
2005-05-29 13:01 ` Michael Thonke
2005-05-29 14:09 ` Mark Lord
2005-05-29 14:24 ` Tyler
2005-05-29 15:22 ` Eric D. Mudama
2005-05-29 19:04 ` Jens Axboe
2005-05-29 19:05 ` Jeff Garzik
2005-05-29 19:21 ` Jens Axboe
2005-05-29 19:03 ` Jens Axboe
2005-05-29 20:12 ` Michael Thonke
2005-05-29 20:17 ` Jeff Garzik
2005-05-30 6:05 ` Jens Axboe
2005-05-30 6:07 ` Jens Axboe
2005-05-29 18:10 ` Michael Thonke
2005-05-29 19:06 ` Jens Axboe
2005-05-29 16:03 ` Erik Slagter
2005-05-29 16:34 ` Jeff Garzik
2005-05-29 16:50 ` Erik Slagter
2005-05-29 16:59 ` Jeff Garzik
2005-05-29 17:23 ` Erik Slagter
2005-05-29 17:29 ` Jeff Garzik
2005-05-29 17:45 ` Erik Slagter
2005-05-29 18:01 ` Jeff Garzik
2005-05-29 18:10 ` Erik Slagter
2005-05-29 18:14 ` Jeff Garzik
2005-05-29 18:27 ` Michael Thonke
2005-05-29 18:31 ` Jeff Garzik
2005-05-29 16:57 ` Michael Thonke
2005-05-29 17:26 ` Erik Slagter
2005-05-30 0:06 ` Mark Lord
2005-05-30 7:29 ` Erik Slagter
2005-05-30 18:09 ` Mark Lord
2005-05-30 18:22 ` Erik Slagter
2005-05-30 18:25 ` Mark Lord
2005-05-30 18:34 ` Michael Thonke
2005-05-30 18:51 ` Jeff Garzik
2005-05-30 18:48 ` Jeff Garzik
2005-05-30 18:50 ` Jeff Garzik
2005-05-30 20:03 ` Erik Slagter
2005-05-30 20:19 ` Jeff Garzik
2005-05-31 7:44 ` Erik Slagter
2005-05-30 23:14 ` Mark Lord
2005-05-31 7:48 ` Erik Slagter
2005-05-31 8:05 ` Patrick McFarland
2005-05-29 21:49 ` Michael Thonke
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=4296CC5C.5000807@pobox.com \
--to=jgarzik@pobox.com \
--cc=axboe@suse.de \
--cc=linux-ide@vger.kernel.org \
--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).