[martin%duskware.de@localhost: Re: [pullup-5 #1123] (...)] / umass quirk for PR 42225

[Ignatios Souvatzis <is%netbsd.org@localhost>]

Hi,

so, maybe I was too hasty accepting the proposed workaround for those
broken devices into the tree.

But how should that be handled? Just accept *any* device that claims
zero bytes were transferred, but transferred everything? How would
we know the difference to a device that really didn't transfer
anything this time?

Regards,
        -is

----- Forwarded message from Martin Husemann <martin%duskware.de@localhost> 
-----

Date: Mon, 2 Nov 2009 12:30:06 +0100
From: Martin Husemann <martin%duskware.de@localhost>
To: Ignatios Souvatzis <is%netbsd.org@localhost>
Cc: pullup-5%netbsd.org@localhost
Subject: Re: [pullup-5 #1123] [is%netbsd.org@localhost: CVS commit: 
src/sys/dev/usb] umass quirk for PR 42225

On Mon, Nov 02, 2009 at 12:28:16PM +0100, Ignatios Souvatzis wrote:
> Hello,
> 
> please pull this up; this lets us read/write data from/to (at least
> one) broken usb-ata-bridge. Implemented as a Quirk, so more broken
> devices can be added when necessary (I understand there are several
> handled in a similar way in Linux).

This should not be a quirk, please stall untill this has been properly
discussed.

Check the usb quirks history...

Martin

----- End forwarded message -----