Re: [linuxsh-dev] Compile fix - CONFIG_GDB_STUB_VBR

[Greg B. <gn...@al...>]

NIIBE Yutaka wrote:
> 
> Jesper Skov writes:
>  > I don't know if this is right, but building without GDB
>  > support, but with CONFIG_SH_STANDARD_BIOS causes a failure
>  > since the STUB_VBR option is not defined.
> 
> This is bug of config.in.  I've fixed.  Thanks.
> --

  Arrrgh, you beat me to the punch ;-)

  For the edification of Jesper and anyone else who might care,
here's the explanation of what just happened (from the message
I'll now not need to send)...

  The code in entry.S which is activated by those #if's (specifically
the code at debug_kernel: which delegates certain traps to the BIOS'
or stub's trap handler) is needed in *both* these cases:

x  CONFIG_DEBUG_KERNEL_WITH_GDB_STUB=y
   so that traps which go to the stub (e.g. breakpoints) are delegated
   to the stub.

x  CONFIG_SH_EARLY_PRINTK=y
   so that traps which go the BIOS (i.e. BIOS calls) are delegated
   to the BIOS.

  The actual problem here is that the definition of CONFIG_GDB_STUB_VBR
in
config.in is wrongly dependant on CONFIG_DEBUG_KERNEL_WITH_GDB_STUB=y
[instead of both the variables, or neither].

  Anyway I have no idea why CONFIG_GDB_STUB_VBR exists at all.  I was
under the impression that the value 0xa0000000 is essentially fixed by
the CPU design, as a VBR needs to exist at physical 0x00000000 to handle
resets and powerup.  But I'm sure there's a good reason.

  BTW, nice job Jesper ;-)

Greg.
-- 
Fight spam http://www.caube.org.au/
If it's a choice between being a paranoid, hyper-suspicious global
village
idiot, or a gullible, mega-trusting sheep, I don't look good in mint
sauce. 
 - jd, slashdot, 11Feb2000.