[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] x86/vpt: fix a bug in pt_update_irq()



>>> On 13.10.17 at 10:14, <chao.gao@xxxxxxxxx> wrote:
> On Fri, Oct 13, 2017 at 02:25:38AM -0600, Jan Beulich wrote:
>>>>> On 09.10.17 at 23:32, <chao.gao@xxxxxxxxx> wrote:
>>> +{
>>> +    if ( unlikely(vec < 16) )
>>> +        return false;
>>> +
>>> +    if ( hvm_funcs.sync_pir_to_irr )
>>> +        hvm_funcs.sync_pir_to_irr(vlapic_vcpu(vlapic));
>>
>>Question is whether this is really necessary, of whether instead
>>you could just return the state of the respective PIR bit here. I'd
>>prefer that over giving the function a name no longer suggesting
>>it leaves all state alone.
> 
> It is a good suggestion. But I incline to check the PIR bit and if the
> bit is set, return true and otherwise return the state of the vIRR bit
> in case PIR bits are already synced to vIRR.

Oh, right, of course you should return the effective "or" of the
two bits.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.