This is the mail archive of the gdb-patches@sources.redhat.com 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/RFA] Fix busted logic in find_saved_register()


> On Sat, Apr 20, 2002 at 08:00:06PM -0400, Andrew Cagney wrote:
> 
>  > > Directly.  The alpha_get_saved_register() is nearly identical to the
>  > > mips_get_saved_register().
>  > > If you've implemented INIT_SAVED_REGS() do you need a custom 
>  > get_saved_register()?
> 
> Ok, this is where my understanding of the call_dummy stuff starts to get
> muddled :-)
> 
> The Alpha target doesn't use the generic call_dummy frame stuff, but
> rather its own.  I suspect it could be converted to use the generic
> stuff, but an explanation of how it's supposed to work would really help :-)

Traditional dummy frames were implemented by creating a fake frame on 
the stack that contained all the registers.

Generic dummy frames instead, don't bother.  They create a fake, local 
to gdb, frame that contains the saved registers, and then calls the 
inferior function directly.  On the stack they appear like a special 
frameless function.  On paper, getting them to work is fairly easy.  See 
xstormy16-elf, cris-elf, or any new/obscure target.  Suggest grepping 
for the word ``dummy'' and copying the corresponding settings.

> Would it be terribly evil to commit the alpha multi-arch stuff that I
> have now (which would meet your condition for committing my alpha-netbsd
> configuration :-), and then take a stab at converting it to generic dummy
> frames?

Getting it multi-arched has a higher priority.

Andrew


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