Re: GDB6.2.1 : Queries about mips simulator configuration

Re: GDB6.2.1 : Queries about mips simulator configuration

[cgd]

I've been really busy.  I've not had time to look at mips16 patch
yet...

At Tue, 14 Sep 2004 16:30:41 +0530, Monika Chaddha wrote:
> Queries: 
> 
> 1. 'sim_gen': According to my understanding this is name of simulator
> generator. For the targets in which 32-bit instructions are supported it
> is assigned as 'IGEN' and the targets in which 16-bit instructions or 32
> and 16-bit both are supported, it is assigned as 'M16'.

Mostly.  There's another option, "MULTI" for use by simulators which
support multiple CPU types at once.

What these do is control the way 'igen' is invoked to generate the
simulator.

> Are 'IGEN' and 'M16' two different simulator generators or they are
> sharing something? 
> Is M16 using 'IGEN' indirectly?
> 
> I have gone through 'makefile.in' file also but could not gather much
> information.

AFAICT, they both generate simulators in approximately the same way,
but different parameters are used for each.

In Makefile.in, compare tmp-igen with tmp-m16.

Most of this stuff predates my maintainership -- i've never touched
MIPS16, really, for instance.


> 2. sim_m16_machine="-M mips16,mipsIII"
> 
> I had reported this bug that gdb6.0 was not building for mips16 after
> that you have done this modification " (sim_m16_machine): Include
> mipsIII" to fix this problem. I have also attached the mails for
> reference.

Right.  MIPS16 doesn't exist in a vacuum.  Any machine that implements
MIPS16 must also implement a MIPS I...MIPS64 ISA as well, which in
fact is the underlying ISA used by the machine.

The MIPS16 support needs some basic information and functions related
to the underlying ISA.



cgd

RE: GDB6.2.1 : Queries about mips simulator configuration

[Monika Chaddha]

Thanks a lot for the reply.
My queries regarding configuration part of Simulator are clear now. :-)

Thanks & regards
Monika Chaddha


>-----Original Message-----
>From: gdb-owner@sources.redhat.com
[mailto:gdb-owner@sources.redhat.com] On
>Behalf Of cgd@broadcom.com
>Sent: Wednesday, September 15, 2004 11:21 AM
>To: Monika Chaddha
>Cc: gdb@sources.redhat.com; acmet@vsnl.net
>Subject: Re: GDB6.2.1 : Queries about mips simulator configuration
>
>I've been really busy.  I've not had time to look at mips16 patch
>yet...
>
>At Tue, 14 Sep 2004 16:30:41 +0530, Monika Chaddha wrote:
>> Queries:
>>
>> 1. 'sim_gen': According to my understanding this is name of simulator
>> generator. For the targets in which 32-bit instructions are supported
it
>> is assigned as 'IGEN' and the targets in which 16-bit instructions or
32
>> and 16-bit both are supported, it is assigned as 'M16'.
>
>Mostly.  There's another option, "MULTI" for use by simulators which
>support multiple CPU types at once.
>
>What these do is control the way 'igen' is invoked to generate the
>simulator.
>
>> Are 'IGEN' and 'M16' two different simulator generators or they are
>> sharing something?
>> Is M16 using 'IGEN' indirectly?
>>
>> I have gone through 'makefile.in' file also but could not gather much
>> information.
>
>AFAICT, they both generate simulators in approximately the same way,
>but different parameters are used for each.
>
>In Makefile.in, compare tmp-igen with tmp-m16.
>
>Most of this stuff predates my maintainership -- i've never touched
>MIPS16, really, for instance.
>
>
>> 2. sim_m16_machine="-M mips16,mipsIII"
>>
>> I had reported this bug that gdb6.0 was not building for mips16 after
>> that you have done this modification " (sim_m16_machine): Include
>> mipsIII" to fix this problem. I have also attached the mails for
>> reference.
>
>Right.  MIPS16 doesn't exist in a vacuum.  Any machine that implements
>MIPS16 must also implement a MIPS I...MIPS64 ISA as well, which in
>fact is the underlying ISA used by the machine.
>
>The MIPS16 support needs some basic information and functions related
>to the underlying ISA.
>
>
>
>cgd