There should likely be limitations on the
number of retries ... there was that recent
UHCI trace were seemingly the controller
got into a halt/reset loop. Surely there's
some happy medium between zero and unlimited
retries ... ;-)
----- Original Message -----
From: "Alan Cox" <[EMAIL PROTECTED]>
To: "David Brownell" <[EMAIL PROTECTED]>
Cc: "Johannes Erdfelt" <[EMAIL PROTECTED]>; "Mark Douglas Corner"
<[EMAIL PROTECTED]>;
<[EMAIL PROTECTED]>
Sent: Friday, 16 June, 2000 1:21 PM
Subject: Re: [linux-usb] control_msg question
> > Should all HCDs stop dead in such cases? That's safer than
> > continuing into an all but certain error, I'd think.
>
> I think I'd prefer to get a kernel message, all existing requests zapped,
> the bus reset and re-enumerated as if the root hub was unplugged and then
> back into service.
>
> I dont see what guarantees the pci master about the hcd got was software
> at least not once the code works nicely )
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
