recent changes to frame pointer handling on trunk?

recent changes to frame pointer handling on trunk?

[Jafa]

Hi guys,



I am working from the gdb trunk and the ip2k port of gdb.



I have been running a trunk snapshot from about 4 weeks ago and when asked
for the frame-pointer I go and figure it out (the fp isn't normally stored
by the c-compiler).



I just updated to the trunk and now the frames are all wrong.



What I am seeing is that my code figures the FP out correctly (for example
SP+10) but instead of looking for the first parameter at FP+0, it looks for
the first parameter at FP+10.



Does this make any sense to anyone?



Thanks



Nick

Re: recent changes to frame pointer handling on trunk?

[Andrew Cagney]

> Hi guys,
> 
> 
> 
> I am working from the gdb trunk and the ip2k port of gdb.

(unless I'm missing something) GDB does not include a port to the ip2k.
Suggest contacting who ever supplied you with this GDB.

Andrew


> I have been running a trunk snapshot from about 4 weeks ago and when asked
> for the frame-pointer I go and figure it out (the fp isn't normally stored
> by the c-compiler).
> 
> 
> 
> I just updated to the trunk and now the frames are all wrong.
> 
> 
> 
> What I am seeing is that my code figures the FP out correctly (for example
> SP+10) but instead of looking for the first parameter at FP+0, it looks for
> the first parameter at FP+10.
> 
> 
> 
> Does this make any sense to anyone?
> 
> 
> 
> Thanks
> 
> 
> 
> Nick
> 
> 
> 

Re: recent changes to frame pointer handling on trunk?

[Daniel Jacobowitz]

On Tue, Dec 03, 2002 at 10:36:30PM -0500, Andrew Cagney wrote:
> >Hi guys,
> >
> >
> >
> >I am working from the gdb trunk and the ip2k port of gdb.
> 
> (unless I'm missing something) GDB does not include a port to the ip2k.
> Suggest contacting who ever supplied you with this GDB.

<translation>

He's the one writing the ip2k port, I believe...

> >I have been running a trunk snapshot from about 4 weeks ago and when asked
> >for the frame-pointer I go and figure it out (the fp isn't normally stored
> >by the c-compiler).
> >
> >
> >
> >I just updated to the trunk and now the frames are all wrong.
> >
> >
> >
> >What I am seeing is that my code figures the FP out correctly (for example
> >SP+10) but instead of looking for the first parameter at FP+0, it looks for
> >the first parameter at FP+10.
> >
> >
> >
> >Does this make any sense to anyone?
> >
> >
> >
> >Thanks
> >
> >
> >
> >Nick
> >
> >
> >
> 
> 
> 

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: recent changes to frame pointer handling on trunk?

[Roland Egger]

Hi,

On Tue, Dec 03, 2002 at 07:08:51PM -0800, Jafa wrote:
> Hi guys,
> 
> 
> 
> I am working from the gdb trunk and the ip2k port of gdb.
> 
> 
> 
> I have been running a trunk snapshot from about 4 weeks ago and when asked
> for the frame-pointer I go and figure it out (the fp isn't normally stored
> by the c-compiler).
> 
> 
> 
> I just updated to the trunk and now the frames are all wrong.
> 
> 
> 
> What I am seeing is that my code figures the FP out correctly (for example
> SP+10) but instead of looking for the first parameter at FP+0, it looks for
> the first parameter at FP+10.
I don't know much about ip2k but for sparc and pa-risc fp is the old sp.
If fp = sp+10 then fp+10 is the old fp. It looks like the +10 is done twice
perhaps at your chages and somewhere else but unfortunately I'm not a 
gdb developer ...
Do you know where (in which cpp file) the backtrace functionality is?
(I've programmed it for sparc solaris and intel linux and try to do it
 for pa-risc hp-ux, but on hp-ux the most things seems to be a little bit
 more complicated ...)

Bye 
  Roland