Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 May 2007 22:28:32 +0200
From:      Hans Petter Selasky <>
To:        "M. Warner Losh" <>
Subject:   Re: UMASS pbm at startup
Message-ID:  <>
In-Reply-To: <>
References:  <000001c7a148$94c5d9a0$0202a8c0@artemis> <> <>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
On Monday 28 May 2007 20:01, M. Warner Losh wrote:
> In message: <>
>             Hans Petter Selasky <> writes:
> : On Monday 28 May 2007 18:52, Aymeric MUNTZ (Cerbere3) wrote:
> : > Hi,
> : >
> : > I have a pbm with my usb disk-key under freebsd.
> : > When i start the system (FreeBSD 5.2) with the key already plugged, it
> : > is not recognised. If I unplug and replug it, it is recognised.
> : > With an other motherboard, I do not have this problem.
> : >
> : > Is there a way to rescan for usb devices? I already tried "camcontrol
> : > rescan all" with no success.
> :
> : Have you tried the SVN version of the new USB stack with FreeBSD 6.2
> : installed?
> Let's just change one thing at a time...  Try going to 6.2 first.
> changing both the usb stack and the freebsd version at one time might
> not be wise.  We also need to track down problems in the current USB
> stack, since we are never going to put hps' stack into RELENG_6.
> Since it hasn't yet been committed to current and there's a freeze
> coming up shortly, it is doubtful that it will be in RELENG_7 either.
> We need to not view hps' stack as this magical thing that solves all
> problems because we have at least RELENG_6 to support for a while, and
> may have to support the current stack in RELENG_7 too.
> To be honest, I really wish Hans would be more pro-active in merging
> changes to the old stack so that the delta between it and his would be
> more manageable.  I've said this plenty of times privately, but no
> fixes to the old stack have been forthcoming, not even new vendor IDs
> for existing drivers...

My schedule is full. I'm tracking all changes going into HEAD at least. There 
are some deltas that could have been backported, but I do not specifically 
keep track of all the small things I fix. I would need a secretary for 
that :-) Any volunteers ;-)

With regard to the topic of this thread I know about some specific issues that 
leaves the result that attached devices are not detected. This is related to 
missed PCD interrupts. You have to fake a PCD interrupt every time you enable 
the PCD again, to catch any missed PCD interrupts. The problem is the same 
with OHCI, but there this interrupt is called something else. The PCD 
interrupt is connected to the root interrupt endpoint on the root HUB, which 
is a virtual device. I think that this is the problem.


Want to link to this message? Use this URL: <>