Re: [Qemu-devel] [PATCH v4] s390: Implement dump-guest-memory support for target s390x

[Alexander Graf]

On 10.07.2013, at 14:15, Andreas Färber wrote:

> Am 10.07.2013 12:37, schrieb Alexander Graf:
>> On 10.07.2013, at 12:27, Christian Borntraeger wrote:
>>> On 05/07/13 14:02, Jens Freimann wrote:
>>>> With this patch dump-guest-memory on s390 produces an ELF formatted,
>>>> crash-readable dump.
>>>> In order to implement this, the arch-specific part of dump-guest-memory
>>>> was added:
>>>> target-s390x/arch_dump.c contains the whole set of function for writing
>>>> Elf note sections of all types for s390x.
>>>> 
>>>> Signed-off-by: Ekaterina Tumanova <address@hidden>
>>>> Signed-off-by: Jens Freimann <address@hidden>
>>> Acked-by: Christian Borntraeger <address@hidden>
>>> 
>>> I can confirm that this patch creates proper dumps readable by crash with 
>>> valid
>>> memory and valid register content. Would be good to see this upstream.
>>> 
>>> Meanwhile this patch no longer cleanly applies due to
>>> commit 878096eeb278a8ac1ccd6667af73e026f29b4cf5
>>>   cpu: Turn cpu_dump_{state,statistics}() into CPUState hooks
>>> 
>>> The fixup is trivial.
>>> Alex, please apply.
>> 
>> I'd like to get an Ack from Andreas too :). But then I'll happily apply it.
> 
> I am mostly happy and had offered Alex to reply with a fixup patch to
> squash on top, but am still buried in downstream work and didn't get to
> preparing that yet: v4 started #including cpu-qom.h, which it shouldn't
> (only cpu.h does in the tree) - if Alex wants to drop that line himself,
> please add my Reviewed-by on top, Alex. I did not check whether the
> s390x implementation is functionally correct, just the changed API and
> any style issues.

The patch needs one more round to apply on top of current master anyway, so I'd 
prefer if Ekaterina or Christian could quickly do a respin with your comment 
addressed.


Alex