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

[Michael Eager]

Please post an updated patch for review.

On 5/31/26 11:38 PM, Gopi Kumar Bulusu wrote:
> On Fri, May 29, 2026 at 8:26 PM Tom Tromey <tom@tromey.com> wrote:
>>
>>>>>>> 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.
> 
> This work is being performed under a contract with Xilinx/AMD as
> confirmed in a different response to
> Simon and therefore covered by the copyright assignment.
> 
> I will make the changes as described below and retest.
> 
> Approved for commit ?
> 
> dhanyavaadaaha
> gopi
> 
>>
>>>  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
>