Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 06 Aug 2003 23:03:01 -0600
From:      Scott Long <scottl@freebsd.org>
To:        Nate Lawson <nate@root.org>
Cc:        Scott <scottl@pooker.samsco.home>
Subject:   Re: Ongoing U320 AIC7902 Seagate ST318453LW issues, SCB timed out
Message-ID:  <3F31DD85.6090801@freebsd.org>
In-Reply-To: <20030806213643.F74720@root.org>
References:  <FE045D4D9F7AED4CBFF1B3B813C853370274216B@mail.sandvine.com> <1695190000.1060183917@aslan.scsiguy.com> <20030806093347.N73637@root.org> <20030806123207.F4081@pooker.samsco.home> <20030806213643.F74720@root.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Nate Lawson wrote:
> On Wed, 6 Aug 2003, Scott wrote:
> 
>>On Wed, 6 Aug 2003, Nate Lawson wrote:
>>
>>>On Wed, 6 Aug 2003, Justin T. Gibbs wrote:
>>>
>>>>>I offer this patch to make a loader tunable out of the maximum
>>>>>tags supported.
>>>>
>>>>It should be a cam tunable.
>>>
>>>Terse.  I think he means it should be "kern.cam.ahd%d.max_tags"
>>
>>I'll go out on a limb and suggest 'kern.cam.da%d.max_tags" or something
>>similar, as this will be useful for more than just the Adaptec chip.
> 
> 
> Max tags is a property of cam (i.e. cam_sim_alloc()) and thus not a scsi
> (da(4)) issue.
> 
> -Nate


Not neccesarily.  As Justin pointed out, having global and per-device
settings are both useful.  One could have a bus full of well-behaving
Fujitsu drives, but have one ill-behaving Seagate.  Why penalize the
whole system?  Cam is designed to recognise both controller and device
tag limits.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3F31DD85.6090801>