Re: AMD 8131 and MSI quirk
From: Kristen Accardi
Date: Tue Feb 14 2006 - 16:59:49 EST
On Tue, 2006-02-14 at 23:21 +0200, Michael S. Tsirkin wrote:
> Quoting Kristen Accardi <kristen.c.accardi@xxxxxxxxx>:
> > > It turns out AMD 8131 quirk only affects MSI for devices behind the 8131
> > > bridge. Handle this by adding a flags field in pci_bus, inherited from
> > > parent to child.
> > It seems like we have a way to turn of msi already (the no_msi bit in
> > the pci_dev structure). Does it make sense to just have the child bus
> > pci_dev structure inherit the no_msi bit from the parent's pci_dev
> > structure when doing an allocation, or does that unnecessarily remove
> > the msi capability for devices that may not need it?
> > Kristen
> This bit is already used to mean that msi is disabled for a specific device,
> which appears to be a PCI Express to PCI bridge (PCI_DEVICE_ID_INTEL_PXH). So
> it seems that disabling MSI for child devices as well might break things (i.e.
> disable msi unnecessarily).
> Working for Intel, I guess you would know about the PCI_DEVICE_ID_INTEL_PXH
> best: what do you say?
> In my opinion, it is cleaner to separate the two concepts: suppress msi
> for child devices versus suppress it for the specific device.
I was thinking something along these lines might work for you. I think
it does the same thing as the other patch, without needing to add extra
flags to pci_bus. I guess the assumption I made was that if msi is
turned off for a bridge, then all devices under the bridge may not use
drivers/pci/msi.c | 2 +-
drivers/pci/probe.c | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
@@ -701,7 +701,7 @@ int pci_enable_msi(struct pci_dev* dev)
if (!pci_msi_enable || !dev)
- if (dev->no_msi)
+ if (dev->no_msi || dev->bus->self->no_msi)
temp = dev->irq;
@@ -344,6 +344,7 @@ pci_alloc_child_bus(struct pci_bus *pare
child->self = bridge;
+ child->self->no_msi = parent->self->no_msi;
child->parent = parent;
child->ops = parent->ops;
child->sysdata = parent->sysdata;
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/