Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-16 Thread Jason Gunthorpe
On Thu, Nov 12, 2020 at 08:15:02PM +0100, Thomas Gleixner wrote: > The recent changes to store the MSI irqdomain pointer in struct device > missed that Intel DMAR does not register virtual function devices. Due to > that a VF device gets the plain PCI-MSI domain assigned and then issues > compat M

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-16 Thread Lu Baolu
On 2020/11/16 17:47, Geert Uytterhoeven wrote: Hi Thomas, On Thu, Nov 12, 2020 at 8:16 PM Thomas Gleixner wrote: The recent changes to store the MSI irqdomain pointer in struct device missed that Intel DMAR does not register virtual function devices. Due to that a VF device gets the plain PCI

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-16 Thread Thomas Gleixner
Geert, On Mon, Nov 16 2020 at 10:47, Geert Uytterhoeven wrote: > On Thu, Nov 12, 2020 at 8:16 PM Thomas Gleixner wrote: >> The recent changes to store the MSI irqdomain pointer in struct device >> missed that Intel DMAR does not register virtual function devices. Due to >> that a VF device gets

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-16 Thread Geert Uytterhoeven
Hi Thomas, On Thu, Nov 12, 2020 at 8:16 PM Thomas Gleixner wrote: > The recent changes to store the MSI irqdomain pointer in struct device > missed that Intel DMAR does not register virtual function devices. Due to > that a VF device gets the plain PCI-MSI domain assigned and then issues > compa

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-13 Thread Thomas Gleixner
On Fri, Nov 13 2020 at 09:19, Marc Zyngier wrote: > On 2020-11-12 21:34, Thomas Gleixner wrote: >> That would allow to add a irq_find_matching_fwspec() based lookup to >> pci_msi_get_device_domain(). > > Just so that I understand the issue: is the core of the problem that > there is no 1:1 mapping

[tip: x86/urgent] iommu/vt-d: Cure VF irqdomain hickup

2020-11-13 Thread tip-bot2 for Thomas Gleixner
Committer: Thomas Gleixner CommitterDate: Fri, 13 Nov 2020 12:00:40 +01:00 iommu/vt-d: Cure VF irqdomain hickup The recent changes to store the MSI irqdomain pointer in struct device missed that Intel DMAR does not register virtual function devices. Due to that a VF device gets the plain PCI-MSI

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-13 Thread Marc Zyngier
On 2020-11-12 21:34, Thomas Gleixner wrote: On Thu, Nov 12 2020 at 20:15, Thomas Gleixner wrote: The recent changes to store the MSI irqdomain pointer in struct device missed that Intel DMAR does not register virtual function devices. Due to that a VF device gets the plain PCI-MSI domain assig

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-12 Thread Lu Baolu
Hi Thomas, On 2020/11/13 3:15, Thomas Gleixner wrote: The recent changes to store the MSI irqdomain pointer in struct device missed that Intel DMAR does not register virtual function devices. Due to that a VF device gets the plain PCI-MSI domain assigned and then issues compat MSI messages whic

Re: iommu/vt-d: Cure VF irqdomain hickup

2020-11-12 Thread Thomas Gleixner
On Thu, Nov 12 2020 at 20:15, Thomas Gleixner wrote: > The recent changes to store the MSI irqdomain pointer in struct device > missed that Intel DMAR does not register virtual function devices. Due to > that a VF device gets the plain PCI-MSI domain assigned and then issues > compat MSI messages

iommu/vt-d: Cure VF irqdomain hickup

2020-11-12 Thread Thomas Gleixner
The recent changes to store the MSI irqdomain pointer in struct device missed that Intel DMAR does not register virtual function devices. Due to that a VF device gets the plain PCI-MSI domain assigned and then issues compat MSI messages which get caught by the interrupt remapping unit. Cure that