On 08/08/2011 09:31 PM, Pedro Alves wrote:
> On Monday 08 August 2011 03:46:29, Yao Qi wrote:
>> @@ -1182,6 +1197,7 @@ static const struct qxfer qxfer_packets[] =
>>      { "statictrace", handle_qxfer_statictrace },
>>      { "threads", handle_qxfer_threads },
>>      { "traceframe-info", handle_qxfer_traceframe_info },
>> +    { "fdpic", handle_qxfer_fdpic},
> 
> Please keep the list alpha sorted.
> 

OK.  Fixed.

>> +  memcpy (myaddr, ((char *)data) + offset, copy_length);
> 
>   memcpy (myaddr, (char *) data + offset, copy_length);
> 
>> @@ -5112,6 +5172,11 @@ static struct target_ops linux_target_ops = {
>>    NULL,
>>  #endif
>>    linux_core_of_thread,
>> +#if defined PT_GETDSBT
>> +  linux_read_loadmap,
>> +#else
>> +  NULL,
>> +#endif
>>    linux_process_qsupported,
>>    linux_supports_tracepoints,
>>    linux_read_pc,
> 
> Looks like you may be breaking all !tic6x ports.  Either
> put the new method at the end, or update all 
> struct target_ops instances to adjust for the new field.
> 

There are five instances of target_ops in gdbserver, they are in
linux-low.c, lynx-low.c, nto-low.c, spu-low.c, and win32-low.c.  Adding
new field linux_read_loadmap has only affect on target_ops instance in
win32-low.c.  The rest instances are too short to reach field
linux_read_loadmap and fields after it.  Updated target_ops instance in
win32-low.c.

> Otherwise, with Mark's comments addressed, this looks
> fine to me.
> 
I take Mark's choice #2 on the macro of PTRACE_GETDSBT.  PTRACE_GETDSBT
is replaced with PT_GETDSBT in new patch.  I leave PTRACE_GETDSBT_EXEC
and PTRACE_GETDSBT_INTERP there, because they are macros and no PT_*
counterparts defined in headers.

> Is there any patch in the series missing review?
> 

The last bit is about test case patch,

  [RFA 7/8] New port: TI C6x: test case fixes
  http://sourceware.org/ml/gdb-patches/2011-07/msg00682.html

> Please take care to apply the patches in a sequence
> where none breaks the build, to avoid breaking bisects.
> 

Thanks for reminding this.  I'll take care of it.

-- 
Yao (齐尧)