2009-08-14 03:41:50

by Sujith

[permalink] [raw]
Subject: Revert commit in ath9k

John,

Can you revert the commit - "ath9k: Change DEBUG level for certain interrupts" ?
This interrupt seems to trigger fairly often and floods the log.
The driver does seem to work with this message suppressed.
(which was the original behaviour).

The actual cause for this bug is still unknown, but until it is
found and fixed, the earlier behaviour can be retained.

Sorry for the trouble. :)

Sujith


2009-08-14 12:30:10

by John W. Linville

[permalink] [raw]
Subject: Re: Revert commit in ath9k

On Fri, Aug 14, 2009 at 09:12:08AM +0530, Sujith wrote:

> Can you revert the commit - "ath9k: Change DEBUG level for certain interrupts" ?
> This interrupt seems to trigger fairly often and floods the log.
> The driver does seem to work with this message suppressed.
> (which was the original behaviour).
>
> The actual cause for this bug is still unknown, but until it is
> found and fixed, the earlier behaviour can be retained.
>
> Sorry for the trouble. :)

OK

--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.

2009-08-14 06:09:13

by Pavel Roskin

[permalink] [raw]
Subject: Re: Revert commit in ath9k

Quoting Sujith <[email protected]>:

> John,
>
> Can you revert the commit - "ath9k: Change DEBUG level for certain
> interrupts" ?
> This interrupt seems to trigger fairly often and floods the log.

I think it would be better to revert only the part for the actually
triggered AR_INTR_SYNC_LOCAL_TIMEOUT. This way we'll learn of other
unexpected interrupts if they happen for some users.

--
Regards,
Pavel Roskin

2009-08-14 06:17:23

by Sujith

[permalink] [raw]
Subject: Re: Revert commit in ath9k

Pavel Roskin wrote:
> Quoting Sujith <[email protected]>:
>
> > John,
> >
> > Can you revert the commit - "ath9k: Change DEBUG level for certain
> > interrupts" ?
> > This interrupt seems to trigger fairly often and floods the log.
>
> I think it would be better to revert only the part for the actually
> triggered AR_INTR_SYNC_LOCAL_TIMEOUT. This way we'll learn of other
> unexpected interrupts if they happen for some users.

Well, the other interrupts are handled appropriately (marked as FATAL).
And the ISR does a chip reset for all interrupts marked as FATAL.
SYNC_LOCAL_TIMEOUT was the one I was really interested in, since I saw it
occurring in my machine - and was curious about its occurrence elsewhere.
Apparently, it is widespread. :)

Sujith