Re: [PATCH 6/9] gdb: Implement 'bt shadow' to print the shadow stack backtrace.

[Simon Marchi <simark@simark.ca>]

On 9/25/25 9:19 AM, Eli Zaretskii wrote:
>> From: "Schimpe, Christina" <christina.schimpe@intel.com>
>> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>, "Thiago Jung
>>  Bauermann" <thiago.bauermann@linaro.org>
>> Date: Thu, 25 Sep 2025 11:06:32 +0000
>>
>>>> (gdb) help bt shadow
>>>> Print backtrace of all shadow stack frames, or innermost COUNT frames.
>>>> Usage: backtrace shadow [OPTION]... [COUNT | -COUNT]
>>>
>>> Thanks, but I wonder if this UI is the best we can come up with.
>>> The "backtrace" command doesn't currently have non-option arguments
>>> except COUNT. 
>>
>> Hm, not sure if I fully understand what you mean by non-option arguments.
>> "bt shadow" does not introduce more non-option arguments, except if you say
>> that "shadow" itself is interpreted as "non-option" argument.
> 
> That's what I'm saying, yes.
> 
>>> We have just got rid of qualifier arguments like "full"
>>> and "hide".  Instead of introducing qualifier arguments anew, why not go the
>>> way of thread-related commands and add a new command "shadow", so the
>>> user could say "shadow backtrace" and maybe in the future also other sub-
>>> commands?
>>
>> This has been discussed in a thread years ago:
>> https://sourceware.org/pipermail/gdb/2023-December/051024.html
>> The direction was to make it part of the ordinary bt command. 
> 
> OK, if that's what the others want...

My initial reaction was: why not use a "-shadow" option to the backtrace
command, instead of introducing "backtrace shadow", which is effectively
a separate command?  I then re-read the thread above, and I saw that
"-shadow" was actually part of the options considered, and even seemed
like the preferred option of many.  I didn't see "backtrace shadow"
proposed there, so I'm confused.

The thing I don't like about "backtrace shadow" is: I don't like when a
command ("backtrace") is both a prefix and a regular command.  It can
paint us in a corner at some point.  Let's say we want to introduce the
syntax:

  (gdb) backtrace SYMBOL

at some point, then the sub-command "shadow" would conflict if someone
wanted to call the command "backtrace" with the symbol "shadow".

From what I read in the thread above, the downside of the "backtrace
-shadow" option is that some other options wouldn't make sense for the
shadow backtrace.  I don't think it's a problem, those other options can
just be ignored when printing the shadow stack.

Simon