This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] remove set_tfile_traceframe and cur_traceframe_number


On Wednesday, June 27, 2012 09:47:55 PM Yao Qi wrote:
> I can't see the necessity to use function set_tfile_traceframe and
> variable cur_traceframe_number.  IIUC, both set_tfile_traceframe
> and cur_traceframe_number are equivalent to remote.c:set_remote_traceframe
> and remote.c:remote_traceframe_number.  set_remote_traceframe
> and remote_traceframe_number are used to switch between traceframe
> and live inferior in a lazy mode.  However, this requirement doesn't
> exists in tfile, because GDB only reads from trace file.  This is
> the reason I propose to remove them.  Regression tested on native
> and gdbserver on x86_64-linux.  OK to apply?

The code this patch tries to remove was added by Pedro in this patch,

  [unavailable values part 1, 05/17] move traceframe memory reading fallback 
to read-only sections to GDB side
  http://sourceware.org/ml/gdb-patches/2011-02/msg00136.html

in order to make GDB to read read-only memory from the live target.  Pedro's 
patch did something similar to both remote target and tfile target.  It makes 
sense to remote target, because there is a live inferior that GDB can access.  
However, it is not necessary for tfile target, because there is no live 
inferior at all.  IMO, it is correct to remove them in this patch.

-- 
Yao (éå)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]