On 11/14/23 04:13, Jan Beulich wrote:
> On 13.11.2023 23:21, Stewart Hildebrand wrote:
>> --- a/xen/arch/x86/include/asm/domain.h
>> +++ b/xen/arch/x86/include/asm/domain.h
>> @@ -503,6 +503,8 @@ struct arch_domain
>> #define has_vpit(d) (!!((d)->arch.emulation_flags & X86_EMU_PIT))
>> #define has_pirq(d) (!!((d)->arch.emulation_flags &
>> X86_EMU_USE_PIRQ))
>>
>> +#define arch_needs_vpci(d) ({ (void)(d); false; })
>
> See my comments on the v5 thread on both this and ...
So, the goal here is to return true for a PVH dom0, and false otherwise (for
now). Since dom0 can't feasibly be full HVM, and is_hvm_domain(d) returns true
for PVH, how about the following?
/* TODO: re-visit when vPCI is enabled for PVH domUs. */
#define arch_needs_vpci(d) ({ \
const struct domain *_d = (d); \
is_hardware_domain(_d) && is_hvm_domain(_d); })
Link to v5 thread for reference [1]
[1] https://lists.xenproject.org/archives/html/xen-devel/2023-11/msg00968.html
>
>> --- a/xen/drivers/passthrough/pci.c
>> +++ b/xen/drivers/passthrough/pci.c
>> @@ -1542,6 +1542,18 @@ void iommu_dev_iotlb_flush_timeout(struct domain *d,
>> struct pci_dev *pdev)
>> pcidevs_unlock();
>> }
>>
>> +static bool needs_vpci(const struct domain *d)
>> +{
>> + if ( is_hardware_domain(d) )
>> + return false;
>
> ... this.
I'll move this check to the Arm arch_needs_vpci() in
xen/arch/arm/include/asm/domain.h
> (It is generally a good idea to wait a little with sending new
> versions, when you can't be sure yet whether the earlier discussion has
> settled.)
(Sorry, I'll be better about this going forward.)
>
> Jan
