Re: [Bug #14622] Second IDE device not found
From: Jeff Garzik
Date: Fri Nov 20 2009 - 06:31:54 EST
On 11/20/2009 04:29 AM, Zeno Davatz wrote:
On Fri, Nov 20, 2009 at 9:41 AM, Jeff Garzik<jeff@xxxxxxxxxx> wrote:
On 11/20/2009 02:59 AM, Zeno Davatz wrote:
On Thu, Nov 19, 2009 at 9:05 PM, David Miller<davem@xxxxxxxxxxxxx> wrote:
From: "Rafael J. Wysocki"<rjw@xxxxxxx>
Date: Mon, 16 Nov 2009 23:37:47 +0100 (CET)
This message has been generated automatically as a part of a report
of recent regressions.
The following bug entry is on the current list of known regressions
from 2.6.31. Please verify if it still should be listed and let me know
(either way).
Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=14622
Subject : Second IDE device not found
Submitter : Zeno Davatz<zdavatz@xxxxxxxxx>
Date : 2009-11-11 17:31 (6 days old)
References : http://marc.info/?l=linux-kernel&m=125796105822353&w=4
We're going to need more information to diagnose this.
And linux-ide should have been at least CC:'d from the very
beginning.
From what I can discern the problem is introduced somewhere between
2.6.27 and 2.6.31, you have a Serverworks CSB5 and primarily the issue
is that attaching or detaching your CD-ROM driver influences whether
both of your disks are properly detected. Correct?
You seem to have played around with using the IDE layer vs. the
ATA layer. Can you see any difference in behavior if you try
using just the IDE layer vs. just the ATA layer with the 2.6.31
kernel?
Please see:
http://www.flickr.com/photos/zrr/4118682747/
and
http://www.flickr.com/photos/zrr/4119453092/
Unfortunately, both of these photos only show that MD (block major 9) device
could not be found for root.
For either ATA or IDE, we would need to see full dmesg somehow -- perhaps
capturing serial console output? (Documentation/serial-console.txt)
Ok, I appended
console=ttyS1,9600 console=tty0
to the kernel command line at lilo after choosing the kernel image and
I get a ton of output. But how to I save the output to a file?
I use minicom and a null modem serial cable. One of minicom's commands
will capture everything sent to the serial port, to a file. Other
options are available. Googling for "linux serial console" found
several useful starting-point links.
The basic idea is to sent console output to a serial port, and then have
some method of reading and capturing the serial port's data.
Another alternative is netconsole (google for "linux netconsole"), which
permits output over the network.
I makes no difference if I choose ATA or only IDE.
I am also attaching you my Kernel .config
I could also test with 2.6.32-rc8 (torvalds-git)
I suggest you send me two .config that I can both test with 2.6.31.6
I bet libata fails because CONFIG_BLK_DEV_SD is not enabled. Probably want
to enable CONFIG_BLK_DEV_SR too, for libata CD-ROM support. I would be
interested to see your failing ATA config, with IDE disabled.
I enabled CONFIG_BLK_DEV_SR. Do you want above test with or without
CONFIG_BLK_DEV_SVWKS (OSB4/CSB5) enabled?
Well, this goes back to David's basic request: IDE-only or ATA-only.
You really, really, really should not to enable both at the same time.
If you are choosing ATA-only (libata), then you should disable
CONFIG_IDE and everything associated with CONFIG_IDE, including
CONFIG_BLK_DEV_SVWKS.
libata will want something like
CONFIG_ATA
CONFIG_ATA_VERBOSE_ERROR <-- optional, but helpful
CONFIG_PATA_SERVERWORKS
CONFIG_SCSI
CONFIG_SCSI_LOGGING <-- ditto
CONFIG_SCSI_CONSTANTS <-- ditto
CONFIG_BLK_DEV_SD
CONFIG_BLK_DEV_SR <-- required only for CD-ROM support
Regards,
Jeff
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/