RE: Proposed Addition to GDB packets

["Alkallas, Mouhanad" <Mouhanad.Alkallas@amd.com>]

[-- Attachment #1: Type: text/plain, Size: 1592 bytes --]

[AMD Official Use Only - AMD Internal Distribution Only]

Hi Aaron,

You are right, I didn’t know that you can use this from a serial connection.
This makes my previous proposal unnecessary.

Thanks for the feedback.

Regards,
Mouhanad Alkallas

From: Aaron Griffith <aargri@gmail.com>
Sent: Thursday, August 21, 2025 4:08 PM
To: Alkallas, Mouhanad <Mouhanad.Alkallas@amd.com>
Cc: gdb-patches@sourceware.org
Subject: Re: Proposed Addition to GDB packets

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.

On Thu, Aug 21, 2025 at 5:05 PM Alkallas, Mouhanad <Mouhanad.Alkallas@amd.com<mailto:Mouhanad.Alkallas@amd.com>> wrote:
> The proposal is to add a packet “J” for customization in privately
> defined KGDB IO Module.
>
> Packet format would “J <custom data>”, instead of gdbstub.c handling
> the packet, it would forward the call to the attached KGDB IO Module
> to service the packet.

I'm not familiar with KGDB specifically, but a common solution used by other stubs is to use the "qRcmd" packet for custom commands. I know the Black Magic debug probe uses this to allow the user to scan the JTAG chain it's connected to and a few other things.

The documentation for KGDB [1] indicates it already uses this packet for some commands. Is it possible to hook in to that existing mechanism? Or is that unsuitable for the problem you're trying to solve?

 [1]: https://www.kernel.org/doc/html/v4.14/dev-tools/kgdb.html#running-kdb-commands-from-gdb

 -Aaron

[-- Attachment #2: Type: text/html, Size: 5216 bytes --]