RE: [PATCH v2 9/9] gdb, mi: Add -shadow-stack-list-frames command

["Schimpe, Christina" <christina.schimpe@intel.com>]

> -----Original Message-----
> From: Tom Tromey <tom@tromey.com>
> Sent: Donnerstag, 19. Februar 2026 19:26
> To: Schimpe, Christina <christina.schimpe@intel.com>
> Cc: gdb-patches@sourceware.org; thiago.bauermann@linaro.org
> Subject: Re: [PATCH v2 9/9] gdb, mi: Add -shadow-stack-list-frames command
> 
> >>>>> Christina Schimpe <christina.schimpe@intel.com> writes:
> 
> > +void
> > +mi_cmd_shadow_stack_list_frames (const char *command,
> > +				 const char *const *argv,
> > +				 int argc)
> 
> > +  /* Check if START_SSP points to a shadow stack memory range and use
> > +     the returned range to determine when to stop unwinding.
> > +     Note that a shadow stack memory range can change, due to shadow
> stack
> > +     switches for instance on x86 for an inter-privilege far call or when
> > +     calling an interrupt/exception handler at a higher privilege level.
> > +     Shadow stack for userspace is supported for amd64 linux starting with
> > +     Linux kernel v6.6.  However, shadow stack switches are not supported
> > +     due to missing kernel space support.  We therefore implement this
> > +     command without support for shadow stack switches for now.  */
> > +  std::pair<CORE_ADDR, CORE_ADDR> range;
> > +  if (!gdbarch_address_in_shadow_stack_memory_range (gdbarch,
> *start_ssp,
> > +						     &range))
> > +    {
> > +      /* If START_SSP points off the shadow stack memory range, we cannot
> > +	 print the shadow stack backtrace.  This is possible, for
> > +	 instance, on x86 if NEW_SSP points to the end of RANGE which
> > +	 means that the shadow stack is empty.  */
> > +      return;
> > +    }
> 
> It seems like more of this could be shared with the earlier patch.
> Replicating a lot of the logic doesn't seem good.
> The gdb_printfs in the other code can be handled by changing them to use
> ui_out::text or ui_out::message instead.
> 
> > +  if (!curr.has_value ())
> > +    error (_("-shadow-stack-list-frames: Not enough frames on the shadow "
> > +	     "stack."));
> 
> For things like this maybe the API has to be changed a little but that seems
> fine.

I cannot completely reuse the backtrace_shadow_command logic (which contains some
gdb_printfs) since the interface for -shadow-stack-list-frames is different from bt -shadow:

-shadow-stack-list-frames: Usage: [FRAME_LOW FRAME_HIGH]

With bt -shadow, we call get_trailing_outermost_shadow_stack_frame_info, for instance.
This is handled differently for -shadow-stack-list-frames, as we don’t have a COUNT option
here. This is in line with the normal bt command, in which we call trailing_outermost_frame
instead of get_trailing_outermost_shadow_stack_frame_info.

But for the startup code I fully agree with your feedback, there is a lot of code that is
replicated. I basically can share the code until 
~~~
  /* Let's position curr on the shadow stack frame at which to start the
    display.  This could be the innermost frame if the whole shadow stack
    needs displaying, or if frame_low is 0.  */
  int frame_num = 0;
  for (; curr.has_value () && frame_num < frame_low; frame_num++)
  [...]
~~~

Thanks a lot for pointing that out.

Christina
Intel Deutschland GmbH
Registered Address: Dornacher Strasse 1, 85622 Feldkirchen, Germany
Tel: +49 89 991 430, www.intel.de
Managing Directors: Harry Demas, Jeffrey Schneiderman, Yin Chong Sorrell
Chairperson of the Supervisory Board: Nicole Lau
Registered Seat: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928