Messages in this thread | | | Date | Fri, 09 Jan 2009 07:03:14 -0500 | From | Jeff Garzik <> | Subject | Re: [PATCH] libata: use WARN_ON_ONCE on hot paths |
| |
Zdenek Kabelac wrote: > 2009/1/9 Tejun Heo <tj@kernel.org>: >> Convert WARN_ON() on command issue/completion paths to WARN_ON_ONCE() >> so that libata doesn't spam the machine even when one of those >> conditions triggers repeatedly. >> > > > Hi > > I've not yet applied this patch - but now I'm getting these WARNS on my T61: > (commit: 2150edc6c5cf00f7adb54538b9ea2a3e9cedca3f) > > Is this patch solving the problem or just hiding warnings ?
Hiding warnings. As Tejun noted in another email, this condition should not be happening in the first place.
If I had to guess, I would say that Arjan's patches assume it is OK to treat two ports on a single IDE controller as completely independent, when that is not really the reality of the hardware.
Jeff
| |