[PATCH] gdb/MicroBlaze: Add support for native linux gdb

[Tom Tromey]

>>>>> Gopi Kumar Bulusu <gopi@sankhya.com> writes:

> Attached is a patch to add native linux support for gdb including
> cache target support.

Thanks for the patch.

I read through it and have some nits.  There's nothing very serious
except this:

> The original sources for the patch come from Xilinx/AMD Yocto git
> repository (2025.2)

Who wrote them and are they covered by the copyright assignment?

This is the most important consideration, the patch can't land without
this being clear.

> From a3328b4cdfc58783b5f290f22255b36e26a6fb14 Mon Sep 17 00:00:00 2001
> From: Gopi Kumar Bulusu <gopi@sankhya.com>
> Date: Thu, 7 May 2026 17:47:40 +0530
> Subject: [PATCH] gdb/MicroBlaze: Add support for native linux gdb

It's best to just git send-email rather than using an attachment.

>     bfd/ChangeLog:

>     * elf32-microblaze.c (microblaze_elf_grok_prstatus): New function.
>       (microblaze_elf_grok_psinfo): Likewise.

>     gdb/ChangeLog:

>     * Makefile.in (HFILES_NO_SRCDIR): Add microblaze-linux-tdep.h
>       ALLDEPFILES: Add microblaze-linux-nat.c

gdb doesn't use ChangeLogs any more and this text should all be removed.

> +#define MICROBLAZE_TARGET_HAS_GETREGS 0
> +#define MICROBLAZE_TARGET_HAS_SETREGS 0

This seems weird:

> +/* Wrapper function around ptrace.  */
> +
> +static void
> +fetch_target_gp_regs (int tid, elf_gregset_t *gregs)
> +{
> +  elf_greg_t *gregp = *gregs;
> +
> +#if MICROBLAZE_TARGET_HAS_GETREGS
> +  if (ptrace (PTRACE_GETREGS, tid, 0, (long) gregp) < 0)
> +    {
> +      perror_with_name (_("Couldn't get registers"));
> +      return;
> +    }
> +#error "this configuration did not work during testing"

Normally we don't put new dead code into gdb.

I think this and the defines could just be replaced with a comment
explaining that PTRACE_GETREGS doesn't work on this platform.

> +++ b/gdb/microblaze-linux-tdep.h
> @@ -0,0 +1,24 @@
> +/* Target-dependent code for GNU/Linux on MicroBlaze.
> +
> +   Copyright (C) 2021-2026 Free Software Foundation, Inc.
> +
> +   This file is part of GDB.
> +
> +   This program is free software; you can redistribute it and/or modify
> +   it under the terms of the GNU General Public License as published by
> +   the Free Software Foundation; either version 3 of the License, or
> +   (at your option) any later version.
> +
> +   This program is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +   GNU General Public License for more details.
> +
> +   You should have received a copy of the GNU General Public License
> +   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
> +#ifndef MICROBLAZE_LINUX_TDEP_H
> +#define MICROBLAZE_LINUX_TDEP_H
> + /* Target descriptions.  */
> + extern struct target_desc *tdesc_microblaze_linux;

Blank line between the #define and the comment.

Also this define should have a slightly different name, see
check-include-guards.py.

>  	  non_stack_instruction_found = 0;
> +          cache->register_offsets[rd] = -imm;
>  	  continue;

Indentation looks wrong, probably not using tabs.

> @@ -399,8 +400,7 @@ microblaze_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
>      {
>        sal = find_sal_for_pc (func_start, 0);
 
> -      if (sal.end < func_end
> -	  && start_pc <= sal.end)
> +      if (sal.line != 0 && sal.end <= func_end  && start_pc <= sal.end)
>  	start_pc = sal.end;

No objection from me but this seems somewhat unrelated.

 
>    /* Call for side effects.  */
> -  get_frame_func (next_frame);
> +  cache->pc = get_frame_func (next_frame);

Please remove that comment as it is now wrong.

> +  if (regnum == MICROBLAZE_SP_REGNUM)
> +        regnum = 1;

Wrong indentation.

> +  if (regnum == -1) {
> +    int i;
> +
> +    for (i = 0; i < MICROBLAZE_REDR_REGNUM; i++) {
> +      regcache->raw_supply (i, regs + i * MICROBLAZE_REGISTER_SIZE);
> +    }

Wrong brace placement in this hunk.

> +  cb (".reg", tdep->sizeof_gregset, tdep->sizeof_gregset, tdep->gregset, NULL,
> +      cb_data);

Normally in new code we're using 'nullptr' now.  Best would be to go
through the whole patch and check for this.  You don't have to fix
pre-existing code that is using NULL though.

>    gdbarch *gdbarch
>      = gdbarch_alloc (&info, gdbarch_tdep_up (new microblaze_gdbarch_tdep));

> +  microblaze_gdbarch_tdep *tdep
> +    = gdbarch_tdep<microblaze_gdbarch_tdep> (gdbarch);
> +
> +  tdep->gregset = NULL;
> +  tdep->sizeof_gregset = 0;
> +  tdep->fpregset = NULL;
> +  tdep->sizeof_fpregset = 0;

Here it would be better to just add inline initializers to the new
fields in microblaze_gdbarch_tdep.  Then they'll automatically be
initialized by the 'new'.

Tom