Re: [DIFF] 2.6.10 add MSI to tg3

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Grant,

The latest Broadcom errata for the 57XX controllers indicates that MSI
isn't supported on most of their shipping controllers.  There is a problem
where interrupts aren't masked when MSI is used.

Dave

> Dave,
> Commit Log:
> 	Patch to enable tg3 to use MSI instead of line based interrupts.
> 	Tested on ia64 with BCM5701 (PCI mode) and it works fine there.
>
> Signed-off-by: Grant Grundler (grundler@xxxxxxxxxxxxxxxx)
>
>
> I thought at first I should add a runtime driver parameter like mthca
> (Infiniband) driver does (e.g. msi=1). But I would prefer if tg3 left (a
> new) TG3_FLG2_USE_MSI flag unset for bcm5xxx chips with MSI bugs.
>
> And if the platform claims MSI support but doesn't work right,
> then pci_enable_msi() will cause problems with any device whose
> driver attempts to enable MSI. A kernel "command line" parameter
> could turn off MSI and/or MSIX support for all devices on that platform.
> (I know this is going to be a problem on parisc...I'll propose something
> more concrete when I get there).
>
> BTW, adding MSI support will enable tg3 to drop one MMIO read
> in tg3_interrupt. Once this patch is accepted/rejected I'll work
> on a tg3_msi_interrupt() routine and investigate status_tag
> implementation.
>
> thanks,
> grant
>
>
> ===== drivers/net/tg3.c 1.222 vs edited =====
> --- 1.222/drivers/net/tg3.c	2004-11-15 15:53:08 -08:00
> +++ edited/drivers/net/tg3.c	2005-01-15 11:43:51 -08:00
> @@ -5679,10 +5679,19 @@
>  	if (err)
>  		return err;
>
> -	err = request_irq(dev->irq, tg3_interrupt,
> -			  SA_SHIRQ, dev->name, dev);
> +	if (pci_enable_msi(tp->pdev))
> +		printk(KERN_WARNING PFX "MSI NOT enabled for %s\n",
> +				tp->pdev->slot_name);
> +	else {
> +		tw32(MSGINT_MODE, MSGINT_MODE_ENABLE);
> +		printk(KERN_INFO PFX "MSI enabled on %s\n",
> +				tp->pdev->slot_name);
> +	}
> +
> +	err = request_irq(dev->irq, tg3_interrupt, SA_SHIRQ, dev->name, dev);
>
>  	if (err) {
> +		pci_disable_msi(tp->pdev);
>  		tg3_free_consistent(tp);
>  		return err;
>  	}
> @@ -5713,6 +5722,7 @@
>
>  	if (err) {
>  		free_irq(dev->irq, dev);
> +		pci_disable_msi(tp->pdev);
>  		tg3_free_consistent(tp);
>  		return err;
>  	}
> @@ -5988,6 +5998,7 @@
>  	spin_unlock_irq(&tp->lock);
>
>  	free_irq(dev->irq, dev);
> +	pci_disable_msi(tp->pdev);
>
>  	memcpy(&tp->net_stats_prev, tg3_get_stats(tp->dev),
>  	       sizeof(tp->net_stats_prev));
> -
> : send the line "unsubscribe linux-net" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Netdev]     [Ethernet Bridging]     [Linux 802.1Q VLAN]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Git]     [Bugtraq]     [Yosemite News and Information]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux PCI]     [Linux Admin]     [Samba]

  Powered by Linux