Pedro Alves wrote:
> On Monday 27 July 2009 20:44:51, Aleksandar Ristovski wrote:
>> Pedro Alves wrote:
>>> (( Note: It isn't correct to defer to the target beneath for
>>> TARGET_OBJECT_MEMORY (and similars) objects, but, I see that nto-procfs.c
>>> is still implementing deprecated_xfer_memory.  ))
>> Did you want me to make changes here?
> 
> Nope, consider it an FYI.
> 
>>>> +  /* For 32-bit architecture, size of auxv_t is 8 bytes.  */
>>>> +  const unsigned int sizeof_auxv_t = 8;
>>> NTO doesn't support any 64-bit architecture?
>> No, not yet.
> 
> Okay.
> 
>>>> +  enum bfd_endian byte_order;
>>>> +
>>>> +  /* Skip over argc, argv and envp... (see comment in ldd.c)  */
>>>> +  if (target_read_memory (initial_stack + data_ofs, targ32, sizeof (targ32))
>>>> +      != 0)
>>>> +    return 0;
>>> Can you paste here that ldd.c comment too?  You had pasted it in the
>>> equivalent gdbserver bit.
>> Done.
> 
> Thanks!
> 
> On Monday 27 July 2009 20:44:51, Aleksandar Ristovski wrote:
>> +static LONGEST
>> +procfs_xfer_partial (struct target_ops *ops, enum target_object object,
>> +                    const char *annex, gdb_byte *readbuf,
>> +                    const gdb_byte *writebuf, ULONGEST offset, LONGEST len)
>> +{
>> +  if (object == TARGET_OBJECT_AUXV
>> +      && readbuf)
>> +    {
>> +      int err;
>> +      CORE_ADDR initial_stack;
>> +      debug_process_t procinfo;
>> +
>> +      if (offset > 0)
>> +       return 0;
> 
> This offset > 0 check isn't really correct.  Would it be hard to
> make this work with partial transfers?
> 

Implemented.

-- 
Aleksandar Ristovski
QNX Software Systems


ChangeLog:


* nto-procfs.c (procfs_xfer_partial): New function.
(init_procfs_ops): Register to_xfer_partial.
* nto-tdep.c (nto_read_auxv_from_initial_stack): Define.
* nto-tdep.h (nto_read_auxv_from_initial_stack): Declare.