[ECOS] Re: cannot able to debug

[Yoshinori Sato <ysato at users dot sourceforge dot jp>]

At Tue, 24 Feb 2004 15:29:05 +0530,
Ganeshan wrote:
> 
> Hi
> 
> Yoshinori Sato wrote:
> 
> > At Mon, 23 Feb 2004 11:53:45 +0530,
> > Ganeshan wrote:
> > > Hi
> > > I am using h8/300H - 3064, gcc-3.2.1,  insight-5.3
> > > I could not able to debug the a small application and applied the patch,
> > > http://ysato.dip.jp/~ysato/h8300hal.diff.gz
> > >
> > > 1. If break is given at a specified location, in this case 0x204000, the
> > > gdb is not sending the proper packet
> > >     (instread of sending/reading $m,204000, it is sending/reading as
> > > $m,4000).  The gdb command
> > >     'set remoteaddresssize 24' is used to send a proper packet.  Even
> > > though doing this, could not able to set
> > >     the breakpoint (disassembled the code and saw the instruction.).
> >
> > Please execute "set machine h8300h" in the first place.
> 
> after executing above command, the state of the program is same i.e., it is
> not stopping at the breakpoint.
> Since I have limited memory, I have removed the CYGPKG_KERNEL while build the
> eCos application (hello.c),
> is it OK?
> I have removed library 'libsupc++.a' from h8300_h8300h.ld, since it is not
> able to generated in gcc-3.2.1.

I understood it.
It is the problem that happens so that HAL overwrites with exception vector.
I made a patch.
Turn CYGSEM_HAL_H8300_SAVE_STUB_VECTOR into 1, and please do build.
 
> >
> >
> > > 2. If set the breakpoint ("trapa") manually, then also the program is
> > > not stopping at this location.
> >
> > Because manage the address which set break point inside,
> > it is ignored even if bury it in it directly.
> 
> I could able to understand.  Please explain me.

This was specification of former stub.
Unconditional stop in current version.
But when it is not regular breakpoint, PC slips off.

-- 
Yoshinori Sato
<ysato@users.sourceforge.jp>

Attachment: hal-vector.diff
Description: Binary data

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss