Hi Jan,
> On 30 Sep 2021, at 4:00 pm, Jan Beulich <jbeul...@suse.com> wrote:
>
> On 28.09.2021 20:18, Rahul Singh wrote:
>> @@ -62,8 +63,21 @@ static int __init acpi_pci_init(void)
>> }
>> #endif
>>
>> +/*
>> + * By default pci passthrough is disabled
>> + */
>
> Nit: This is a single line comment.
Ack.
>
>> +bool __read_mostly pci_passthrough_enabled = false;
>
> Nit: Unnecessary initializer.
Ack.
>
>> --- a/xen/common/physdev.c
>> +++ b/xen/common/physdev.c
>> @@ -19,6 +19,9 @@ ret_t do_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void)
>> arg)
>> struct pci_dev_info pdev_info;
>> nodeid_t node;
>>
>> + if ( !is_pci_passthrough_enabled() )
>> + return -ENOSYS;
>
> Here and ...
>
>> @@ -54,6 +57,9 @@ ret_t do_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void)
>> arg)
>> case PHYSDEVOP_pci_device_remove: {
>> struct physdev_pci_device dev;
>>
>> + if ( !is_pci_passthrough_enabled() )
>> + return -ENOSYS;
>
> ... here -EOPNOTSUPP (or any other suitable value, which -ENOSYS is not
> from all I can tell) please.
I checked all the possible errors and came to the conclusion that
-EOPNOTSUPP is the right choice here.
>
>> --- a/xen/include/asm-x86/pci.h
>> +++ b/xen/include/asm-x86/pci.h
>> @@ -32,4 +32,12 @@ bool_t pci_ro_mmcfg_decode(unsigned long mfn, unsigned
>> int *seg,
>> extern int pci_mmcfg_config_num;
>> extern struct acpi_mcfg_allocation *pci_mmcfg_config;
>>
>> +/*
>> + * Unlike ARM, PCI passthrough is always enabled for x86.
>> + */
>
> Nit: This again is a single line comment.
Ack.
Regards,
Rahul
>
> Jan
>
