[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v3 0/3] Bitmap based CPU enumeration
From: |
Paolo Bonzini |
Subject: |
Re: [Qemu-devel] [PATCH v3 0/3] Bitmap based CPU enumeration |
Date: |
Thu, 04 Jun 2015 10:51:35 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0 |
On 04/06/2015 10:39, Peter Crosthwaite wrote:
>> > I wouldn't mind separating the "CPU" parts of exec.c and moving them
>> > under Andreas and Eduardo's mantainership. In fact, Peter, in your
>> > patch to move stuff from cpu-exec.c to cpus.c, perhaps you can use
>> > qom/cpu.c instead? Then qom/cpu.c can also be the place where we can
>> > move the CPU parts of exec.c.
>> >
> So that relocated code uses conditional compile based on
> CONFIG_SOFTMMU. Is that def accessible from common-obj-y code which
> qom/cpu.c is?
>
> My choice of cpus.c was based on the fact that it was obj-y.
Hmm, right---qom/cpu.c is indeed common-obj-y, so it has to be a new
file. cpu-exec.c is taken, so I guess I'll move the memory parts of
exec.c out to exec-memory.c and leave exec.c for the obj-y part of CPU
object handling. We have:
- cpu-exec.c: TCG only, arch-obj-y
- cpus.c: thread management, obj-y
- exec.c: CPU object management, obj-y
- qom/cpu.c: CPU object management, common-obj-y
And you can move the stuff from cpu-exec.c to exec.c in your patches.
> I assume this is all follow up work out of scope of Bharata's code. Do
> you have a queue I can rebase my conflicting ENV_GET_CPU work on?
No, I don't, because I'm not going to be the one who merge these
patches. Sorry.
Paolo