On Tue, Jan 28, 2003 at 08:19:25PM +0100, Oliver Neukum wrote: > Am Dienstag, 28. Januar 2003 13:54 schrieb Oliver Kurth: > > Hello! > > > > I wrote a new driver for Atmel at76c503 based USB wlan adapters. See > > the link below in my .signature.
> > Maybe someone wants to look over it. It is still in development,
> > and supports only ad-hoc mode at the moment. There is one 'exotic'
> > thing about it: it uses another module which I wrote to load the
> > firmware, it is included in the package. It works as follows:
>
> block = kmalloc(EXT_FW_BLOCK_SIZE, GFP_KERNEL|GFP_DMA);
> if(block == NULL) return -ENOMEM;
>
> Wrong. Use GFP_KERNEL only.
Alright. I just wanted to be sure it's DMA.
> if(urb->status != 0){
> if((urb->status != -ENOENT) &&
> (urb->status != -ECONNRESET)) {
> dbg(__FUNCTION__ " - nonzero write bulk status received: %d",
> urb->status);
> }
> return;
>
> Any error will lock up the device? Not nice.
Oh, thanks for that. Fixed.
>
> struct ieee802_11_hdr *i802_11_hdr =
> (struct ieee802_11_hdr *)&(tx_buffer->packet);
>
> down(&dev->sem);
Fixed as well, I now use spin_lock_irq(). Strange thing, it never
crashed here, so I thought this is not in an interrupt.
> down() with spinlock held. Absolute taboo. There are severaly bugs
> of this kind.
Thank you for looking... if anyone finds more, please tell me.
Greetings,
Oliver
--
An MTA for home users, notebooks and PDAs: http://masqmail.cx/masqmail/
A driver for Atmel at76c503 based WLAN Adapters: http://masqmail.cx/at76c503/
msg11129/pgp00000.pgp
Description: PGP signature
