From: "Bene Martin" <martin.bene@icomedias.com>
To: "Jeff Garzik" <jgarzik@pobox.com>
Cc: <linux-ide@vger.kernel.org>,
"Linux Kernel" <linux-kernel@vger.kernel.org>
Subject: Re: [SATA] libata-dev queue updated
Date: Tue, 7 Jun 2005 09:00:48 +0200 [thread overview]
Message-ID: <FA095C015271B64E99B197937712FD02510D0B@freedom.grz.icomedias.com> (raw)
> The two reasons why passthru is not upstream are:
>
> * I need to make sure that it is 100% up-to-date for the latest ATA
> passthru spec, which clarified some SCSI CDB and descriptor codes.
>
> * There was at least one report of ATA passthru use causing a device
> under load to start timing out, which could possibly indicate a driver
> bug. I need to do a serious analysis and final review of the patch, to
> make sure that it is plugged into the ATA host state machine at the
> correct places.
Just as a data point:
I've had a couple of problems on a system using sata+passthrough patch;
not latest code though. I am reading smart info at fairly high frequency
(smartd + a smartctl call for each device every 5 minutes to graph drive
temperatures).
Versions used: 2.6.11.7 + 2.6.11-bk6-libata-dev1.patch.bz2
I got several ata errors, mostly in high - throughput situations (rsync
over local network or rebuilding raid5), example from the logs:
Jun 5 04:13:50 backup kernel: ata3: command 0x25 timeout, stat 0x50
host_stat 0x1
Jun 5 06:30:51 backup kernel: ata3: command 0x35 timeout, stat 0x50
host_stat 0x1
Jun 5 10:46:50 backup kernel: ata3: command 0x35 timeout, stat 0x50
host_stat 0x1
Jun 5 10:47:20 backup kernel: ata3: command 0x35 timeout, stat 0x50
host_stat 0x1
Jun 5 10:53:45 backup kernel: ATA: abnormal status 0xD0 on port
0xE0802287
Jun 5 10:53:45 backup last message repeated 2 times
Jun 5 10:54:15 backup kernel: ata3: command 0x35 timeout, stat 0x50
host_stat 0x1
Jun 5 11:43:11 backup kernel: ATA: abnormal status 0xD0 on port
0xE0802287
Jun 5 11:43:11 backup last message repeated 2 times
Jun 5 11:43:41 backup kernel: ata3: command 0x35 timeout, stat 0x50
host_stat 0x1
Jun 5 14:23:45 backup kernel: ata3: command 0x35 timeout, stat 0x51
host_stat 0x1
Jun 5 14:23:45 backup kernel: ata3: status=0x51 { DriveReady
SeekComplete Error }
Jun 5 14:23:45 backup kernel: ata3: error=0x04 { DriveStatusError }
Jun 5 14:23:45 backup kernel: SCSI error : <3 0 0 0> return code =
0x8000002
Jun 5 14:23:45 backup kernel: sdc: Current: sense key: Aborted Command
Jun 5 14:23:45 backup kernel: Additional sense: No additional sense
information
Jun 5 14:23:45 backup kernel: end_request: I/O error, dev sdc, sector
135917718
Jun 5 14:23:45 backup kernel: raid5: Disk failure on sdc6, disabling
device. Operation continuing on 2 devices
What makes me doubt the theory that smart/passthrough might be to blame
is that I've only seen these errors on one of the 3 disks (always
sdc/ata3). On the other hand, smart info for the drive doesn't show any
errors or suspicious readings.
Boot log for the sata devices:
May 18 08:51:53 backup kernel: libata version 1.10 loaded.
May 18 08:51:53 backup kernel: sata_sil version 0.8
May 18 08:51:53 backup kernel: ACPI: PCI interrupt 0000:02:00.0[A] ->
GSI 21 (level, low) -> IRQ 21
May 18 08:51:53 backup kernel: ata1: SATA max UDMA/100 cmd 0xE0802080
ctl 0xE080208A bmdma 0xE0802000 irq 21
May 18 08:51:53 backup kernel: ata2: SATA max UDMA/100 cmd 0xE08020C0
ctl 0xE08020CA bmdma 0xE0802008 irq 21
May 18 08:51:53 backup kernel: ata3: SATA max UDMA/100 cmd 0xE0802280
ctl 0xE080228A bmdma 0xE0802200 irq 21
May 18 08:51:53 backup kernel: ata4: SATA max UDMA/100 cmd 0xE08022C0
ctl 0xE08022CA bmdma 0xE0802208 irq 21
May 18 08:51:53 backup kernel: ata1: dev 0 cfg 49:2f00 82:346b 83:7d01
84:4023 85:3469 86:3c01 87:4023 88:407f
May 18 08:51:53 backup kernel: ata1: dev 0 ATA-7, max UDMA/133,
781422768 sectors: LBA48
May 18 08:51:53 backup kernel: ata1: dev 0 configured for UDMA/100
May 18 08:51:53 backup kernel: scsi1 : sata_sil
May 18 08:51:53 backup kernel: ata2: dev 0 cfg 49:2f00 82:346b 83:7d01
84:4023 85:3469 86:3c01 87:4023 88:407f
May 18 08:51:53 backup kernel: ata2: dev 0 ATA-7, max UDMA/133,
781422768 sectors: LBA48
May 18 08:51:53 backup kernel: ata2: dev 0 configured for UDMA/100
May 18 08:51:53 backup kernel: scsi2 : sata_sil
May 18 08:51:53 backup kernel: ata3: dev 0 cfg 49:2f00 82:346b 83:7d01
84:4023 85:3469 86:3c01 87:4023 88:407f
May 18 08:51:53 backup kernel: ata3: dev 0 ATA-7, max UDMA/133,
781422768 sectors: LBA48
May 18 08:51:53 backup kernel: ata3: dev 0 configured for UDMA/100
May 18 08:51:53 backup kernel: scsi3 : sata_sil
May 18 08:51:53 backup kernel: ata4: no device found (phy stat 00000000)
May 18 08:51:53 backup kernel: scsi4 : sata_sil
May 18 08:51:53 backup kernel: Vendor: ATA Model: ST3400832AS
Rev: 3.01
May 18 08:51:53 backup kernel: Type: Direct-Access
ANSI SCSI revision: 05
May 18 08:51:53 backup kernel: Vendor: ATA Model: ST3400832AS
Rev: 3.01
May 18 08:51:53 backup kernel: Type: Direct-Access
ANSI SCSI revision: 05
May 18 08:51:53 backup kernel: Vendor: ATA Model: ST3400832AS
Rev: 3.01
May 18 08:51:53 backup kernel: Type: Direct-Access
ANSI SCSI revision: 05
SATA controller used:
0000:02:00.0 RAID bus controller: Silicon Image, Inc. SiI 3114
[SATALink/SATARaid] Serial ATA Controller (rev 02)
Bye, Martin
next reply other threads:[~2005-06-07 7:01 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-06-07 7:00 Bene Martin [this message]
-- strict thread matches above, loose matches on Subject: below --
2005-06-04 6:08 [SATA] libata-dev queue updated Jeff Garzik
2005-06-04 15:17 ` Greg Stark
2005-06-04 17:48 ` Jeff Garzik
2005-06-04 18:50 ` Adrian Bunk
2005-06-04 19:05 ` Jeff Garzik
2005-06-04 19:15 ` Greg Stark
2005-06-04 19:19 ` Jeff Garzik
2005-06-05 2:25 ` Greg Stark
2005-06-05 2:30 ` Jeff Garzik
2005-06-05 5:30 ` Greg Stark
2005-06-22 16:15 ` Mark Lord
2005-06-22 20:23 ` Jeff Garzik
2005-06-23 3:39 ` Mark Lord
2005-06-05 2:40 ` Drew Winstel
2005-06-05 2:58 ` Jim Crilly
2005-06-06 16:01 ` Mark Lord
2005-06-07 1:57 ` Greg Stark
2005-06-07 11:20 ` Brad Campbell
2005-06-08 8:33 ` Mark Lord
2005-06-07 13:15 ` Brad Campbell
2005-06-08 8:37 ` Mark Lord
2005-06-07 13:51 ` Greg Stark
2005-06-08 10:30 ` Mark Lord
2005-06-08 8:51 ` Mark Lord
2005-06-07 13:09 ` Domenico Andreoli
2005-06-07 17:53 ` Jeff Garzik
2005-03-10 5:40 Jeff Garzik
2005-03-29 9:12 ` kern.petr
2005-03-08 11:20 Jeff Garzik
[not found] <3Ds62-5AS-3@gated-at.bofh.it>
2005-03-02 20:34 ` Joerg Sommrey
2005-03-02 22:43 ` Jeff Garzik
2005-03-03 19:32 ` Joerg Sommrey
2005-03-04 4:09 ` Jeff Garzik
2005-03-04 6:37 ` Joerg Sommrey
2005-03-04 7:10 ` Jeff Garzik
2005-03-04 17:49 ` Joerg Sommrey
2005-03-04 18:07 ` Jeff Garzik
2005-03-04 20:33 ` Joerg Sommrey
2005-03-04 20:43 ` Jeff Garzik
2005-03-04 22:06 ` Joerg Sommrey
2005-03-04 22:32 ` Joerg Sommrey
2005-03-02 2:40 Jeff Garzik
2005-02-19 8:40 Jeff Garzik
2005-02-19 16:58 ` Andre Tomt
2005-02-19 17:09 ` Jeff Garzik
2005-02-13 20:33 Jeff Garzik
2004-10-27 6:52 Jeff Garzik
2004-10-27 6:53 ` Jeff Garzik
2004-10-27 22:53 ` Rob van Nieuwkerk
2004-10-27 23:03 ` Jeff Garzik
2004-10-15 8:04 Jeff Garzik
2004-10-15 17:15 ` Jeff Garzik
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=FA095C015271B64E99B197937712FD02510D0B@freedom.grz.icomedias.com \
--to=martin.bene@icomedias.com \
--cc=jgarzik@pobox.com \
--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).