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

Re: [Xen-devel] [v3][PATCH 1/1] xen:vtd: missing RMRR mapping while share EPT

>>> On 25.07.14 at 10:19, <yang.z.zhang@xxxxxxxxx> wrote:
> Jan Beulich wrote on 2014-07-25:
>>>>> On 25.07.14 at 02:56, <yang.z.zhang@xxxxxxxxx> wrote:
>>> Jan Beulich wrote on 2014-07-24:
>>>>>>> On 24.07.14 at 13:35, <tim@xxxxxxx> wrote:
>>>>> (BTW, I think this patch can go in as-is, and the address-space
>>>>> clash be fixed up separately.)
>>>> Not sure - I agree it shouldn't make things worse except when there
>>>> is an RMRR but the device doesn't actually access it (i.e. would be
>>>> fine without this patch). But I'm not sure committing half
>>>> solutions is really appropriate for not really pressing issues.
>>> I think we are talking two issues. The address-apace clash has
>>> nothing to do with the issue this patch tries to fix. We will work
>>> out another patch to fix it as Tim suggested.
>> No, we aren't: This patch _introduces_ the address space clash.
> In non-sharept case, do we reserve the RMRR in e820 table? If not, the guest 
> may use it as DMA buffer and it actually the same problem we are talking 
> here. 

True, indeed. Yet I still consider the patch is fixing a problem at the
price of making that situation worse.


Xen-devel mailing list



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