[Xen-devel] Re: Changeset 20209 causes an issue in xen_in_range()

[Keir Fraser <keir.fraser@xxxxxxxxxxxxx>]

Well, there maybe a bit of unneeded code hanging around, but I don't think
it's that bad or particularly misleading.

 -- Keir


On 14/01/2010 13:13, "Zhang, Xiantao" <xiantao.zhang@xxxxxxxxx> wrote:

> Okay, seems a cleanup patch for that is needed, otherwise the related logic
> maybe misleading...
> Xiantao
> Keir Fraser wrote:
>> At some point we will introduce hotplug notifiers and have per-cpu
>> areas set up only for present CPUs. The possible_map is basically
>> deprecated/obsolete.
>> 
>>  -- keir
>> 
>> On 14/01/2010 12:19, "Zhang, Xiantao" <xiantao.zhang@xxxxxxxxx> wrote:
>> 
>>> I am curious that all bits of cpu_possible_map are all set to 1, so
>>> the related checks in percpu_free_unused_areas are unnecessary ?
>>> Xiantao
>>> 
>>> 
>>> Keir Fraser wrote:
>>>> On 14/01/2010 11:21, "Jan Beulich" <JBeulich@xxxxxxxxxx> wrote:
>>>> 
>>>>>>> However, in iommu_set_dom0_mapping(), xen_in_range() is still
>>>>>>> True for the freed pages above, so devices in Dom0 can meet with
>>>>>>> DMA fault.
>>>>>> 
>>>>>> Should be fixed by c/s 20803.
>>>>> 
>>>>> Do you really think so? Masking "start" with PERCPU_SIZE-1 doesn't
>>>>> make sense, as it may be arbitrarily before __percpu_start and
>>>>> not aligned. Below my take on it.
>>>> 
>>>> How about c/s 20806?
>>>> 
>>>> 
>>>> _______________________________________________
>>>> Xen-devel mailing list
>>>> Xen-devel@xxxxxxxxxxxxxxxxxxx
>>>> http://lists.xensource.com/xen-devel
> 



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel