On Fri, 2008-10-31 at 14:48 +1100, Vincent Rizza wrote:
> Michael raised the issue of dev->stop() calling pci_disable_msi()
> on ifconfig down. The rtnl_lock eliminates that problem.
I did, but that's only one example of something that can call
pci_disable_msi(). It needs to be robust against anything that might
call it, eg. a hotplug disk controller or something.
So I think using list_for_each_entry_safe() is the right option, using
rtnl_lock() definitely isn't.
cheers
> @@ -45,6 +47,27 @@ pci_config_attr(subsystem_device, "0x%04x\n");
> pci_config_attr(class, "0x%06x\n");
> pci_config_attr(irq, "%u\n");
>
> +#ifdef CONFIG_PCI_MSI
> +static ssize_t msix_show(struct device *dev,
> + struct device_attribute *attr, char *buf)
> +{
> + struct pci_dev *pdev = to_pci_dev(dev);
> + struct msi_desc *entry;
> + u32 total = 0;
> +
> + if (pdev->msix_enabled) {
> + /* netdev interface lock in case someone does an ifdown */
> + rtnl_lock();
> + list_for_each_entry(entry, &pdev->msi_list, list) {
> + total += sprintf(buf + total, "%u\n", entry->irq);
> + }
> + rtnl_unlock();
> + }
> +
> + return total;
> +}
--
Michael Ellerman
OzLabs, IBM Australia Development Lab
wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)
We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
Attachment:
signature.asc
Description: This is a digitally signed message part
