Re: [PATCH 1/1] Char: istallion, init+locking fixes

[Jiri Slaby <jirislaby@gmail.com>]

Ingo Korb napsal(a):
> Move brdp->iosize assignment in stli_initecp up a few lines to stop the
> driver from requesting an I/O region of length 0.
> 
> Remove spin_lock_irqsave/spin_unlock_irqrestore from __stli_sendcmd as
> all users of that function take the lock already.
> 
> Signed-off-by: Ingo Korb <ml@akana.de>
> ---
> One thing I dislike about this driver: It polls its cards every jiffy to
> look for new data. The cards are able to use interrupts (some of the
> drivers for other OSes use them), but as far as I know there is no open
> documentation. Unfortunately the DOS driver (probably easiest to
> analyze) also uses polling. =(
> 
> 
> diff --git a/drivers/char/istallion.c b/drivers/char/istallion.c
> index 7b279d1..0bc1c37 100644
> --- a/drivers/char/istallion.c
> +++ b/drivers/char/istallion.c
> @@ -2168,8 +2168,6 @@ static void __stli_sendcmd(struct stlibrd *brdp, struct stliport *portp, unsigne
>  	unsigned char __iomem *bits;
>  	unsigned long flags;

Remove the flags and you will see the bug, you introduced ;).

> -	spin_lock_irqsave(&brd_lock, flags);
> -
>  	if (test_bit(ST_CMDING, &portp->state)) {
>  		printk(KERN_ERR "STALLION: command already busy, cmd=%x!\n",
>  				(int) cmd);
> @@ -2194,7 +2192,6 @@ static void __stli_sendcmd(struct stlibrd *brdp, struct stliport *portp, unsigne
>  	writeb(readb(bits) | portp->portbit, bits);
>  	set_bit(ST_CMDING, &portp->state);
>  	EBRDDISABLE(brdp);
> -	spin_unlock_irqrestore(&brd_lock, flags);
>  }


regards,
-- 
http://www.fi.muni.cz/~xslaby/            Jiri Slaby
faculty of informatics, masaryk university, brno, cz
e-mail: jirislaby gmail com, gpg pubkey fingerprint:
B674 9967 0407 CE62 ACC8  22A0 32CC 55C3 39D4 7A7E