On Thu, Apr 20, 2000, Michael W Zappe <[EMAIL PROTECTED]> wrote:
> Huh. I saw stuff like that happen before on mine, but after removing the
> SPD disable flag, that behaviour stopped.
Remove the SPD disable flag? Is there some sort of new driver available
past 0.2?
The driver I have specifically set USB_DISABLE_SPD. That's the only reason
I can think of for the kaweth_control function which otherwise duplicates
usb_control_msg.
> Which chipset are you using?
> I've been running on the PIIX4 (371AB rev 2). I've also noted that
> people with the VIA 586B chips (I think its the VC 3 or something chipset,
> if VIAs site wasn't #*$&#* down i'd be able to tell you... ;-) It also
> happens to be the chipset on the iOpener, so my friend should have tons
> of fun tonight... ;-)
Both with my 82371AB PIIX4 and VIA VT82C586B do the same things with both
drivers.
> So when and why was the spin_lock added to the urb structure? It has to
> have been VERY recently!
A couple of versions ago. I added it to ensure locking for the URB
structure. The hcpriv field was the one I was interested in.
Here's the offending code in the kaweth.c driver:
static inline void kaweth_resubmit_rx_urb(struct kaweth_device *kaweth)
{
int result;
memset(kaweth->rx_urb, 0, sizeof(*kaweth->rx_urb));
spin_lock_init(&urb->lock);
There's no urb variable, so the code doesn't compile. Dunno why it's
there at all.
You also shouldn't clear the urb with memset. It's very evil.
JE
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
