Re: [Xen-devel] [PATCH V3 13/29] x86/vvtd: Set Interrupt Remapping Table Pointer through GCMD

[Chao Gao <chao.gao@xxxxxxxxx>]

On Thu, Oct 19, 2017 at 12:56:45PM +0100, Roger Pau Monné wrote:
>On Thu, Sep 21, 2017 at 11:01:54PM -0400, Lan Tianyu wrote:
>> From: Chao Gao <chao.gao@xxxxxxxxx>
>> 
>> Software sets this field to set/update the interrupt remapping table pointer
>> used by hardware. The interrupt remapping table pointer is specified through
>> the Interrupt Remapping Table Address (IRTA_REG) register.
>> 
>> This patch emulates this operation and adds some new fields in VVTD to track
>> info (e.g. the table's gfn and max supported entries) of interrupt remapping
>> table.
>> 
>> Signed-off-by: Chao Gao <chao.gao@xxxxxxxxx>
>> Signed-off-by: Lan Tianyu <tianyu.lan@xxxxxxxxx>
>> 
>> ---
>> @@ -148,6 +205,18 @@ static int vvtd_write(struct vcpu *v, unsigned long 
>> addr,
>>              break;
>>          }
>>      }
>> +    else /* len == 8 */
>> +    {
>> +        switch ( offset )
>> +        {
>> +        case DMAR_IRTA_REG:
>> +            vvtd_set_reg_quad(vvtd, DMAR_IRTA_REG, val);
>
>I have kind of a generic comment regarding the handlers in general,
>which I will just make here. Don't you need some kind of locking to
>prevent concurrent read/write accesses to the registers?

I think guest should be responsible to avoid concurrency.
Xen only needs to not be fooled (crashed) by a malicious guest.

>
>Also the 'if' to handle different sized accesses to the same registers
>seems quite cumbersome. I would think there's a better way to handle
>this with a single switch statement.

Will use only one switch statement and maybe add if-else for the
cases which can be accessed with different size.

Thanks
chao

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