Re: [PATCH] mISDN GIT: hfcpci cleanup failure on IRQ weirdness

[Karsten Keil <isdn@linux-pingi.de>]

Hi Andreas,

On Dienstag, 2. Juni 2009 15:06:49 Andreas Mohr wrote:

Thanks for discovering this issue.

...
> ChangeLog:
> Make sure to properly bail out (call free_irq())
> after hfcpci interrupt detection failure.
> Improve comment.
>
> Signed-off-by: Andreas Mohr <andi@lisas.de>
>
> --- ./drivers/isdn/hardware/mISDN/hfcpci.c.orig	2009-06-02
> 13:54:10.000000000 +0200 +++
> ./drivers/isdn/hardware/mISDN/hfcpci.c	2009-06-02 14:29:20.000000000 +0200
> @@ -1172,7 +1172,7 @@ hfcpci_int(int intno, void *dev_id)
>  			printk(KERN_DEBUG
>  			    "HFC-PCI: stat(%02x) s1(%02x)\n", stat, val);
>  	} else {
> -		/* shared */
> +		/* (shared) IRQ triggered by other device */
>  		spin_unlock(&hc->lock);
>  		return IRQ_NONE;
>  	}

Hmm, I think shared as comment in the IRQ function should be
enough -  but  I do not complain to a longer comment.

> @@ -1807,8 +1807,7 @@ init_card(struct hfc_pci *hc)
>  			    "HFC PCI: IRQ(%d) getting no interrupts "
>  			    "during init %d\n", hc->irq, 4 - cnt);
>  			if (cnt == 1) {
> -				spin_unlock_irqrestore(&hc->lock, flags);
> -				return -EIO;
> +				goto fail;
>  			} else {
>  				reset_hfcpci(hc);
>  				cnt--;

easier:
			if (cnt == 1)
				break;
			else {


I will send this patch together with some other ISDN fixes again soon.

Karsten