Re: [RFC PATCH V1 09/12] libxl: Handle virtio-mmio irq in more correct way

[Oleksandr <olekstysh@gmail.com>]

On 05.08.20 02:22, Stefano Stabellini wrote:

Hi Stefano

> On Mon, 3 Aug 2020, Oleksandr Tyshchenko wrote:
>> From: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
>>
>> This patch makes possible to use device passthrough again.
>>
>> Signed-off-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
>> ---
>>   tools/libxl/libxl_arm.c | 33 +++++++++++++++++++++++----------
>>   1 file changed, 23 insertions(+), 10 deletions(-)
>>
>> diff --git a/tools/libxl/libxl_arm.c b/tools/libxl/libxl_arm.c
>> index 620b499..4f748e3 100644
>> --- a/tools/libxl/libxl_arm.c
>> +++ b/tools/libxl/libxl_arm.c
>> @@ -9,6 +9,10 @@
>>   #include <assert.h>
>>   #include <xen/device_tree_defs.h>
>>   
>> +#define GUEST_VIRTIO_MMIO_BASE  xen_mk_ullong(0x02000000)
>> +#define GUEST_VIRTIO_MMIO_SIZE  xen_mk_ullong(0x200)
>> +#define GUEST_VIRTIO_MMIO_SPI   33
> They should be in xen/include/public/arch-arm.h

ok


>
> Is one interrupt enough if there are multiple virtio devices? Is it one
> interrupt for all virtio devices, or one for each device?

   One interrupt for each virtio device. I experimented with current 
series and assigned 4 disk partitions to the guest. This resulted in 4 
separate device-tree nodes, and each node had individual SPI and MMIO range.


>
> Of course this patch should be folded in the patch to add virtio support
> to libxl.

ok


-- 
Regards,

Oleksandr Tyshchenko