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: [PATCH 2/2] gdbserver: xtensa: support THREADPTR register


On Mon, Feb 13, 2017 at 10:26 AM, Luis Machado
<lgustavo@codesourcery.com> wrote:
> On 01/18/2017 05:56 PM, Max Filippov wrote:
>>
>> Provide aceess to the THREADPTR register to remote gdb.
>>
>> 2017-01-18  Max Filippov  <jcmvbkbc@gmail.com>
>> gdb/gdbserver/
>
>
>
> Similarly, drop the gdb/gdbserver/
>
>>         * linux-xtensa-low.c (regnum::R_THREADPTR): New enum member.
>>         (xtensa_fill_gregset): Call collect_register_by_name for
>>         threadptr register.
>>         (xtensa_store_gregset): Call supply_register_by_name for
>>         threadptr register.
>> ---
>>  gdb/gdbserver/linux-xtensa-low.c | 9 +++++++++
>>  1 file changed, 9 insertions(+)
>>
>> diff --git a/gdb/gdbserver/linux-xtensa-low.c
>> b/gdb/gdbserver/linux-xtensa-low.c
>> index 73fbfe2..eca0289 100644
>> --- a/gdb/gdbserver/linux-xtensa-low.c
>> +++ b/gdb/gdbserver/linux-xtensa-low.c
>> @@ -36,6 +36,7 @@ enum regnum {
>>         R_LBEG, R_LEND, R_LCOUNT,
>>         R_SAR,
>>         R_WS, R_WB,
>> +       R_THREADPTR,
>>         R_A0 = 64
>>  };
>
>
> Does xtensa have xml register description support?

AFAIK no.

> It would be nice to have
> those described instead of hardcoding the entries like this.
>>
>> @@ -86,6 +87,10 @@ xtensa_fill_gregset (struct regcache *regcache, void
>> *buf)
>>        collect_register (regcache, i, ptr);
>>        ptr += register_size (tdesc, i);
>>      }
>> +
>> +#if XCHAL_HAVE_THREADPTR
>> +  collect_register_by_name (regcache, "threadptr",
>> (char*)&rset[R_THREADPTR]);
>
>
> Missing spaces in the cast. "(char *) &rset[R_THREADPTR]"

Will add.

> How is XCHAL_HAVE_THREADPTR set?

It is defined in the core configuration overlay, include/xtensa-config.h
This file is generated for each Xtensa core and is copied over to the
source tree before building gdb for that specific core.

> Is it a reliable way of confirming the presence of this register?

This is the primary source of configuration-specific information.

> Can a runtime check be used instead?

AFAIK this information is not provided anywhere else.

>> +#endif
>>  }
>>
>>  static void
>> @@ -133,6 +138,10 @@ xtensa_store_gregset (struct regcache *regcache,
>> const void *buf)
>>    supply_register_by_name (regcache, "ps", (char*)&rset[R_PS]);
>>    supply_register_by_name (regcache, "windowbase", (char*)&rset[R_WB]);
>>    supply_register_by_name (regcache, "windowstart", (char*)&rset[R_WS]);
>> +
>> +#if XCHAL_HAVE_THREADPTR
>> +  supply_register_by_name (regcache, "threadptr",
>> (char*)&rset[R_THREADPTR]);
>
>
> Same as above regarding spaces and the cast construct and the
> XCHAL_HAVE_THREADPTR check.

Will fix.

-- 
Thanks.
-- Max


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