Anthony Liguori wrote:
> Jan Kiszka wrote:
>> [ Also available via git://git.kiszka.org/qemu.git queue/gdb ]
>>
>> In order to set the VCPU for the next single-step command, you need gdb
>> 6.8 or better - and this patch. It enhances the existing support for
>> representing VCPUs as threads to the gdb frontend by introducing the
>> vCont remote gdb command. This is used by gdb to switch the debugging
>> focus for single-stepping multi-threaded targets.
>>
>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>>   
> 
> I think the consensus from the last posting of this was that modeling
> threads was pretty broken and that we should model as processes.  Did I
> miss something there?

There was this concern brought up by Paul Brook. I looked into this
again, but the picture didn't change for me:

 a) Modeling cpus as processes buys us nothing compared to threads given
    the fact that we cannot provide a stable memory mapping to the gdb
    frontend anyway. (*)

 b) The model is already part of mainline qemu. This patch is just
    about adding even more usefulness to it.

Jan

(*) Process abstraction is, if used at all, guest business. At best we
could try to invent (likely OS-specific) heuristics to identify
identical mappings and call them processes. I don't see a reasonable
benefit compared to the expected effort and unreliability.