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

Re: [Xen-devel] [PATCH 2/2] domain: use PGC_extra domheap page for shared_info

On 06.03.2020 14:13, Paul Durrant wrote:
>> -----Original Message-----
>> From: Xen-devel <xen-devel-bounces@xxxxxxxxxxxxxxxxxxxx> On Behalf Of Jan 
>> Beulich
>> Sent: 06 March 2020 13:10
>> To: David Woodhouse <dwmw2@xxxxxxxxxxxxx>; Durrant, Paul 
>> <pdurrant@xxxxxxxxxxxx>
>> Cc: sstabellini@xxxxxxxxxx; julien@xxxxxxx; wl@xxxxxxx; 
>> konrad.wilk@xxxxxxxxxx;
>> andrew.cooper3@xxxxxxxxxx; ian.jackson@xxxxxxxxxxxxx; 
>> george.dunlap@xxxxxxxxxx; xen-
>> devel@xxxxxxxxxxxxxxxxxxxx; Volodymyr_Babchuk@xxxxxxxx
>> Subject: Re: [Xen-devel] [PATCH 2/2] domain: use PGC_extra domheap page for 
>> shared_info
>> On 06.03.2020 13:57, David Woodhouse wrote:
>>> On Fri, 2020-03-06 at 13:36 +0100, Jan Beulich wrote:
>>>> And of course this means you're intending to (at least
>>>> partially) resurrect the distinction between domheap and xenheap,
>>>> which isn't said anywhere in Paul's series, I don't think.
>>> Right. Secret hiding makes the distinction (xenheap is mapped, domheap
>>> is not) mostly go away. We are talking about restoring *a* distinction
>>> between one type of page (Xen ephemeral pages which don't need to be
>>> preserved over live update) and another (must be preserved), but
>>> whether that should still be called "xenheap" vs. "domheap", despite
>>> the massive parallels, isn't entirely clear.
>>>>  If this
>>>> is a sufficiently correct understanding of mine, then on one hand
>>>> I start seeing the point of the conversion Paul wants to make, but
>>>> otoh this then feels a little like making the 2nd step before the
>>>> 1st.
>>> What would you suggest is the first step?
>> Establishing of what the new separation rule and mechanism is going
>> to be (no matter how the two resulting pieces are going to be
>> named).
> Would you be ok with a comment to that effect?

Not sure. It would certainly help if the cover letter at least
mentioned other related aspects like this one.

> My aim is to make the separation abundantly obvious by getting rid
> of shared xenheap pages (for non-system domains at least) but I
> can't do that before converting shared_info and grant shared/status
> frames to domheap.

Following David's various replies - instead of going this route of
replacing the sharing of xenheap pages by different logic, the
same ought to be achievable by making the backing allocations come
from the correct pool?


Xen-devel mailing list



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