On Wed, Apr 12, 2017 at 02:57:23AM -0600, Jan Beulich wrote:
>>>> On 12.04.17 at 02:04, <chao....@intel.com> wrote:
>> --- a/xen/drivers/passthrough/vtd/intremap.c
>> +++ b/xen/drivers/passthrough/vtd/intremap.c
>> @@ -200,8 +200,9 @@ static void update_irte(struct iommu *iommu, struct
>> iremap_entry *entry,
>> else
>> {
>> /*
>> - * If the caller requests an atomic update but we can't meet it,
>> - * a bug will be raised.
>> + * VT-d hardware doesn't update IRTEs behind us, nor the software.
>
>Hmm, so far I was under the impression that in posted mode the
>IRTE could be updated by hardware. Is that not the case? As to
No. I have confirmed this with VT-d architect that VT-d hardware
doesn't update IRTEs. For posted format, VT-d hardware will atomically
update Posted Interrupt Descriptor, an address recorded in posted
format IRTE.
>software not updating, with there not being any synchronization
>clearly visible around here, I'm afraid this also needs expanding
>on (in the commit message at least, not necessarily in the
>comment).
Will do.
>
>> + * If a caller want an atomic update from the views of VT-d
>
>wants
>
>Also what do you mean by "from the views of VT-d"?
OK. will fix this too. Do you mean it is (and should be) atomic from software's
view
, so these words are redundant.
Thanks
Chao
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
