Re: libATA error message doc

Previous thread: [GIT PULL] Blackfin fixing for 2.6.23-rc9 by Bryan Wu on Wednesday, October 3, 2007 - 12:50 pm. (1 message)

Next thread: [git patch] small ocfs2 fix by Mark Fasheh on Wednesday, October 3, 2007 - 2:20 pm. (1 message)
To: <linux-kernel@...>
Date: Wednesday, October 3, 2007 - 1:39 pm

Does a doc exist that provides a detailed description of the different
errors that could be produced by libata and its associated drivers?
The libATA developer's guide provides a very good general description
of the types of messages, but I'm looking for more detail.

For example I got this error message last week:

Sep 27 17:29:26 server kernel: ata2.00: exception Emask 0x0 SAct 0x0
SErr 0x0 action 0x1
Sep 27 17:29:26 server kernel: ata2.00: (irq_stat 0x40000001)
Sep 27 17:29:26 server kernel: ata2.00: tag 0 cmd 0xb0 Emask 0x1 stat
0x51 err 0x4 (device error)

I'd like to know what the following items means?

(irq_stat 0x40000001)
tag 0
cmd 0xb0
emask 0x1
stat 0x51
err 0x4

Can anybody help?

Thanks,
-Mike
-

To: Mike Robak <mrobak@...>
Cc: <linux-kernel@...>
Date: Wednesday, October 3, 2007 - 2:24 pm

> (irq_stat 0x40000001)

Internal libata error category (used to decide what to do to
Error register of drive

You want the ATA specification document from T13 (www.t13.org)

Alan
-

To: Alan Cox <alan@...>
Cc: Mike Robak <mrobak@...>, <linux-kernel@...>
Date: Friday, October 5, 2007 - 3:13 pm

The ISDN driver documentation always listed the numbers in the kernel documentation
even though people in theory could have gotten it from the specification too.

I always found that quite convenient. Putting at least a summary in Documentation
would probably user-friendly. It tends to be hard to find stuff in specifications,
especially if you're not very familiar with the standard.

-Andi
-

To: Andi Kleen <andi@...>
Cc: Mike Robak <mrobak@...>, <linux-kernel@...>
Date: Friday, October 5, 2007 - 4:51 pm

There are over 100 commands, and the status bits can only be read in
combination and according to some interpretation rules - it would be
quite a long document. The standards are also all publically available in
this case, albeit not neccessarily the most interesting read

AFAIK there is a patch queued to print some of the data in a more user
friendly format.
-

To: Mike Robak <mrobak@...>
Cc: <linux-kernel@...>
Date: Wednesday, October 3, 2007 - 1:54 pm

It's pretty much all defined in the ATA or controller specifications...

Well, except for emask, which is a libata-only invention.

Jeff

-

Previous thread: [GIT PULL] Blackfin fixing for 2.6.23-rc9 by Bryan Wu on Wednesday, October 3, 2007 - 12:50 pm. (1 message)

Next thread: [git patch] small ocfs2 fix by Mark Fasheh on Wednesday, October 3, 2007 - 2:20 pm. (1 message)