On the other hand, if that line of code triggered this,
then power cycling the disk couldn't work.  The ohci module
would need reinitialization.  How did "-58" happen then?

- Dave

p.s. The normal sequence of events there:  OHCI sees something
    on the order of a bus error, then gives a Unrecoverable
    Error (UE) interrupt, and "disabled" is set.  Recovery
    is manual:  rmmod usb-ohci, modprobe usb-ohci works.


----- Original Message -----
From: "Johannes Erdfelt" <[EMAIL PROTECTED]>
To: "Matthew Dharm" <[EMAIL PROTECTED]>
Cc: "Matti Aarnio" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Tuesday, 25 April, 2000 4:38 PM
Subject: Re: [linux-usb] HP USB CD-Writer Plus -- nearly works (?)


> On Tue, Apr 25, 2000, Matthew Dharm <[EMAIL PROTECTED]> wrote:
> > On Wed, 26 Apr 2000, Matti Aarnio wrote:
> > > Apr 25 23:13:50 mea kernel: usb.c: registered new driver usb-storage
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: Searching unusual device list for 
>(0x3f0, 0x107,
0x200)...
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: -- found matching device: USB/SCSI 
>bridge
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: USB Mass Storage device detected
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: Endpoints: In 2 Out 1 Int 3 (Period 
>32)
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: Result from usb_set_interface is -58
> > > Apr 25 23:13:50 mea kernel: usb-storage.c: -- Unknown error.  Rejecting device
> > > Apr 25 23:13:50 mea kernel: USB Mass Storage support registered.
> > >
> > > Weird..  That *used* to work.  Power cycling the device.
> > >
> > >         { 0x03f0, 0x0107, 0x0200,
> > >           "HP USB CD-Writer Plus", US_SC_SCSI, US_PR_CB, 0},
> >
> > I've seen this before.  I think it might be an HCD error -- if I
> > physically detach the device and reattach it and it works.  I'm not sure
> > why this is necessary.  Does someone with more of an HCD clue have any
> > ideas?
> >
> > Have you tried this on an UHCI system?
>
> This was on an alpha system, right? If so, -58 is -ESHUTDOWN. According
> to my sources, the only way to get an -ESHUTDOWN is:
>
>         /* when controller's hung, permit only hub cleanup attempts
>          * such as powering down ports */
>         if (ohci->disabled)
>                 return -ESHUTDOWN;
>
> That comment makes me wonder
>
> JE
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to