Hello Jan,

> On 17 Nov 2020, at 11:03 am, Jan Beulich <jbeul...@suse.com> wrote:
> 
> On 16.11.2020 13:25, Rahul Singh wrote:
>> passthrough/pci.c file is common for all architecture, but there is x86
>> specific code in this file.
> 
> In how far is ...
> 
>> @@ -1370,13 +1301,6 @@ static int __init setup_dump_pcidevs(void)
>> }
>> __initcall(setup_dump_pcidevs);
>> 
>> -int iommu_update_ire_from_msi(
>> -    struct msi_desc *msi_desc, struct msi_msg *msg)
>> -{
>> -    return iommu_intremap
>> -           ? iommu_call(&iommu_ops, update_ire_from_msi, msi_desc, msg) : 0;
>> -}
> 
> ... this code x86-specific? The hook being called lives in a
> #ifdef CONFIG_HAS_PCI section, and MSI is a general PCI sub-
> feature. IOW if this is another workaround, it should be
> called so (if there's really no other way to address whatever
> issue there is), which in turn likely means it wants to be in
> a separate patch.
> 

As of now there is no implementation for ARM to remap the interrupt and 
interrupt remapping is enabled for x86 only so I thought I will move this code 
to x86 file as this function is called from the x86 specific code only.

I will remove this code form this patch and will fix this once once will have 
proper MSI implementation for ARM.

> Jan


Reply via email to