Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[Programmingkid]

On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:

> On 13/06/2016 22:35, Andrew Jones wrote:
>> On Mon, Jun 13, 2016 at 07:04:01PM +0200, Paolo Bonzini wrote:
>>> On 10/06/2016 19:40, Andrew Jones wrote:
>>>> +    if (sockets == -1 || cores == -1 || threads == -1 ||
>>>> +        maxcpus == -1 || cpus == -1) {
>>>> +        error_report("cpu topology: "
>>>> +                     "all machine properties must be specified");
>>>> +        exit(1);
>>>> +    }
>>>> +
>>> 
>>> I think it's sane to accept some defaults.  It must not be the DWIM
>>> thing that -smp does (which is targeted to Windows's dislike of
>>> multi-socket machine on consumer hardware).  It must be something that
>>> makes sense, and my proposal is:
>>> 
>>> - threads: 1
>>> - cores: 1
>>> - sockets:
>>>  - maxcpus / (cores * threads) if maxcpus given
>>>  - cpus / (cores * threads) if cpus given
>>>  - else 1
>>> - maxcpus: cores * threads * sockets
>>> - cpus: maxcpus
>> 
>> I think some machines may prefer
>> 
>> - threads: 1
>> - sockets: 1
>> - cores:
>>  - maxcpus / (sockets * threads) if maxcpus given
>>  - cpus / (sockets * threads) if cpus given
>>  - else 1
> 
> smp_cores is only used by pseries and x86 machines.  I expect machines
> that must be single-socket to disregard smp_sockets altogether.

Could smp support be added to the beigeg3 and mac99 targets?

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[Programmingkid]

On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:

> On 13/06/2016 22:35, Andrew Jones wrote:
>> On Mon, Jun 13, 2016 at 07:04:01PM +0200, Paolo Bonzini wrote:
>>> On 10/06/2016 19:40, Andrew Jones wrote:
>>>> +    if (sockets == -1 || cores == -1 || threads == -1 ||
>>>> +        maxcpus == -1 || cpus == -1) {
>>>> +        error_report("cpu topology: "
>>>> +                     "all machine properties must be specified");
>>>> +        exit(1);
>>>> +    }
>>>> +
>>> 
>>> I think it's sane to accept some defaults.  It must not be the DWIM
>>> thing that -smp does (which is targeted to Windows's dislike of
>>> multi-socket machine on consumer hardware).  It must be something that
>>> makes sense, and my proposal is:
>>> 
>>> - threads: 1
>>> - cores: 1
>>> - sockets:
>>>  - maxcpus / (cores * threads) if maxcpus given
>>>  - cpus / (cores * threads) if cpus given
>>>  - else 1
>>> - maxcpus: cores * threads * sockets
>>> - cpus: maxcpus
>> 
>> I think some machines may prefer
>> 
>> - threads: 1
>> - sockets: 1
>> - cores:
>>  - maxcpus / (sockets * threads) if maxcpus given
>>  - cpus / (sockets * threads) if cpus given
>>  - else 1
> 
> smp_cores is only used by pseries and x86 machines.  I expect machines
> that must be single-socket to disregard smp_sockets altogether.

Could smp support be added to the beigeg3 and mac99 targets?

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[BALATON Zoltan]

On Tue, 14 Jun 2016, Programmingkid wrote:
> On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:
>> smp_cores is only used by pseries and x86 machines.  I expect machines
>> that must be single-socket to disregard smp_sockets altogether.
>
> Could smp support be added to the beigeg3 and mac99 targets?

I think the machines these are emulating had no SMP so it would not make 
much sense. Maybe you'd need to create a new target emulating a Mac model 
with multiple CPUs instead or update mac99 to a newer model that had SMP?
But what would this bring besides more complexity and more possible bugs?

Regards,
BALATON Zoltan

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[Programmingkid]

On Jun 14, 2016, at 5:13 PM, BALATON Zoltan wrote:

> On Tue, 14 Jun 2016, Programmingkid wrote:
>> On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:
>>> smp_cores is only used by pseries and x86 machines.  I expect machines
>>> that must be single-socket to disregard smp_sockets altogether.
>> 
>> Could smp support be added to the beigeg3 and mac99 targets?
> 
> I think the machines these are emulating had no SMP so it would not make much sense. Maybe you'd need to create a new target emulating a Mac model with multiple CPUs instead or update mac99 to a newer model that had SMP?

http://www.macworld.com/article/1002601/multiprocessor.html
You will find this article enlightening.

> But what would this bring besides more complexity and more possible bugs?

- An emulator that can take better advantage of the host system's multiprocessor offering.
- An emulator that can run more programs and still feel responsive.
- A simulator for testing how a program behaves in a single and multiprocessor environment.

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[Paolo Bonzini]

On 14/06/2016 23:32, Programmingkid wrote:
> 
> On Jun 14, 2016, at 5:13 PM, BALATON Zoltan wrote:
> 
>> On Tue, 14 Jun 2016, Programmingkid wrote:
>>> On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:
>>>> smp_cores is only used by pseries and x86 machines.  I expect machines
>>>> that must be single-socket to disregard smp_sockets altogether.
>>>
>>> Could smp support be added to the beigeg3 and mac99 targets?
>>
>> I think the machines these are emulating had no SMP so it would not make much sense. Maybe you'd need to create a new target emulating a Mac model with multiple CPUs instead or update mac99 to a newer model that had SMP?
> 
> http://www.macworld.com/article/1002601/multiprocessor.html

Do you also have a link to the specs or hardware programming manuals?

Paolo

Re: [Qemu-devel] [Qemu-ppc] [PATCH RFC 06/16] vl: move smp parsing to machine pre_init

[Thomas Huth]

On 14.06.2016 23:32, Programmingkid wrote:
> 
> On Jun 14, 2016, at 5:13 PM, BALATON Zoltan wrote:
> 
>> On Tue, 14 Jun 2016, Programmingkid wrote:
>>> On Jun 14, 2016, at 7:39 AM, qemu-ppc-request@nongnu.org wrote:
>>>> smp_cores is only used by pseries and x86 machines.  I expect machines
>>>> that must be single-socket to disregard smp_sockets altogether.
>>>
>>> Could smp support be added to the beigeg3 and mac99 targets?
>>
>> I think the machines these are emulating had no SMP so it would not make much sense. Maybe you'd need to create a new target emulating a Mac model with multiple CPUs instead or update mac99 to a newer model that had SMP?
> 
> http://www.macworld.com/article/1002601/multiprocessor.html
> You will find this article enlightening.
> 
>> But what would this bring besides more complexity and more possible bugs?
> 
> - An emulator that can take better advantage of the host system's multiprocessor offering.

Please note that as long as the MTTCG patches are not included yet, this
point is currently not valid for QEMU yet (unless you use KVM as
accelerator).

 Thomas