Re: [PATCH] gdb/s390: Fill guess_tracepoint_registers hook.

["Marcin Kościelnicki" <koriakin@0x04.net>]

On 18/02/16 17:48, Ulrich Weigand wrote:
> Marcin Kościelnicki wrote:
>
>> This version fills PSWM too.  This causes two failures in tests when
>> combined with the later patches:
>> 		
>> - gdb.trace/mi-trace-frame-collected.exp: will have to be changed to expect
>>    $pswa and $cc to be available.
>> - gdb.trace/unavailable.exp: likewise.
>>
>> An unfortunate side effect is that $cc is now considered available (which
>> unavailable.exp rightly complains about).  There isn't a way to mark
>> individual bitfields as unavailable, is there?
>
> That's bit unfortunate.  Right now there's no way to mark *parts* of
> a register a unavailable (this is possible for the value subsystem,
> but not for the plain regcache at this point).
>
> It probably shouldn't matter much to show a defined $cc in those cases.
>
> Patch is OK.
>
> Bye,
> Ulrich
>

Thanks, pushed.  I will update patch #5 shortly to fix the test issues. 
  FWIW patches #3 and #4 are currently ready for review, while #6-#8 
have some pending issues I'll have to fix (Antoine's comment about auxv 
+ making it work with the high GPRs patch + doc comments everywhere).