Re: [Xen-devel] [PATCH v3 23/38] arm/p2m: Cosmetic fixes -- __p2m_get_mem_access

[Julien Grall <julien.grall@xxxxxxx>]

On 13/09/16 14:42, Sergej Proskurin wrote:
> Hi Julien,
>
>
> On 09/13/2016 03:30 PM, Julien Grall wrote:
> > On 13/09/16 14:27, Sergej Proskurin wrote:
> > > Hi Julien,
> > Hello Sergej,
> >
> > > On 09/12/2016 10:53 AM, Julien Grall wrote:
> > > > Hello Sergej,
> > > >
> > > > On 16/08/2016 23:16, Sergej Proskurin wrote:
> > > > > This commit extends the function prototypes of the functions:
> > > > > * __p2m_get_mem_access
> > > > > * p2m_mem_access_check_and_get_page
> > > > >
> > > > > We extend the function prototype of "__p2m_get_mem_access" to hold an
> > > > > argument of type "struct p2m_domain*", as we need to distinguish
> > > > > between
> > > > > the host's p2m and different altp2m views. While doing so, we
> > > > > needed to
> > > > > extend the function's prototype of "p2m_mem_access_check_and_get_page"
> > > > > to hold an argument of type "struct vcpu*".
> > > > Please details in the commit message why it is necessary to pass a
> > > > "struct vcpu *" to p2m_mem_access_check_and_get_page.
> > > Actually, it is already sufficient to provide a "struct domain *".
> > > Thanks for the hint. I will extend the commit message in the next patch
> > > accordingly.
> > No it is not, the stage-1 page table (VTTBR0, VTTBR1...) are per-vCPU
> > and if the current vCPU is not the correct one, you will have to
> > context switch few registers before.
> >
> > It is not handled today, but we should avoid to have a broken interface.
> I see. Ok, since it is not really used today, I will state that we will
> need the "struct vcpu*" to access vcpu-related registers in the future.
> Would that be ok?
I would recommend you to have a think on the problem mentioned in patch 
#24 first. You may need the vCPU to fix the bug.
Regards,

--
Julien Grall

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