Re: [PATCH 3/6] add_to_physmap: Move the code for XENMEM_add_to_physmap

[Keir Fraser <keir.xen@gmail.com>]

On 20/11/2011 13:25, "Olaf Hering" <olaf@aepfle.de> wrote:

> On Sat, Nov 19, Keir Fraser wrote:
> 
>> On 19/11/2011 21:58, "Olaf Hering" <olaf@aepfle.de> wrote:
>> 
>>> On Wed, Nov 16, Jean Guyader wrote:
>>> 
>>>> Move the code for the XENMEM_add_to_physmap case into it's own
>>>> function (xenmem_add_to_physmap).
>>> 
>>> This changeset 24163:7a9a1261a6b0 seems to cause the current testsuite
>>> failures.
>>> (XEN) Assertion '!in_atomic()' failed at softirq.c:61
>>> 
>>> preempt_count is like fffffc52 or fffffc00 in my testing.
>> 
>> Thanks, hopefully fixed by c/s 24167.
> 
> Yes, the ASSERT does not trigger anymore.
> 
> The remaining issue is this:
> 
> Nov 20 06:21:11.744519 (XEN) hvm.c:2312:d1 guest attempted write to read-only
> memory page. gfn=0xc0, mfn=0x201979

Is that new behaviour? It may be unrelated to whatever HVM test failure
we're seeing, or else be a mere symptom of a guest gone haywire for other
reasons (we write-protect that memory range because it is supposed to be
ROM).

 -- Keir

> See
> http://www.chiark.greenend.org.uk/~xensrcts/logs/9893/test-amd64-i386-rhel6hvm
> -amd/serial-potato-beetle.log
> 
> Olaf