Re: [PATCH 2/3] vmd: disable MSI remapping bypass under Xen

[Jonathan Derrick <jonathan.derrick@linux.dev>]

Hi Bjorn,

On 1/10/25 3:25 PM, Bjorn Helgaas wrote:
> Match historical subject line style for prefix and capitalization:
> 
>    PCI: vmd: Set devices to D0 before enabling PM L1 Substates
>    PCI: vmd: Add DID 8086:B06F and 8086:B60B for Intel client SKUs
>    PCI: vmd: Fix indentation issue in vmd_shutdown()
> 
> On Fri, Jan 10, 2025 at 03:01:49PM +0100, Roger Pau Monne wrote:
>> MSI remapping bypass (directly configuring MSI entries for devices on the VMD
>> bus) won't work under Xen, as Xen is not aware of devices in such bus, and
>> hence cannot configure the entries using the pIRQ interface in the PV case, and
>> in the PVH case traps won't be setup for MSI entries for such devices.
>>
>> Until Xen is aware of devices in the VMD bus prevent the
>> VMD_FEAT_CAN_BYPASS_MSI_REMAP capability from being used when running as any
>> kind of Xen guest.
> 
> Wrap to fit in 75 columns.
> 
> Can you include a hint about *why* Xen is not aware of devices below
> VMD?  That will help to know whether it's a permanent unfixable
> situation or something that could be done eventually.
> 
I wasn't aware of the Xen issue with VMD but if I had to guess it's 
probably due to the special handling of the downstream device into the 
dmar table.

[1] 4fda230ecddc2
[2] 9b4a824b889e1


>> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
>> ---
>>   drivers/pci/controller/vmd.c | 9 +++++++++
>>   1 file changed, 9 insertions(+)
>>
>> diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
>> index 264a180403a0..d9b7510ace29 100644
>> --- a/drivers/pci/controller/vmd.c
>> +++ b/drivers/pci/controller/vmd.c
>> @@ -965,6 +965,15 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
>>   	struct vmd_dev *vmd;
>>   	int err;
>>   
>> +	if (xen_domain())
>> +		/*
>> +		 * Xen doesn't have knowledge about devices in the VMD bus.
> 
> Also here.
> 
>> +		 * Bypass of MSI remapping won't work in that case as direct
>> +		 * write to the MSI entries won't result in functional
>> +		 * interrupts.
>> +		 */
>> +		features &= ~VMD_FEAT_CAN_BYPASS_MSI_REMAP;
>> +
>>   	if (resource_size(&dev->resource[VMD_CFGBAR]) < (1 << 20))
>>   		return -ENOMEM;
>>   
>> -- 
>> 2.46.0
>>