Re: [RFC PATCH 00/34] The rest of the x86_64-gnu port

[Sergey Bugaev]

On Tue, Mar 21, 2023 at 1:28 PM Sergey Bugaev <bugaevc@gmail.com> wrote:
>
> On Tue, Mar 21, 2023 at 12:48 PM Luca <luca@orpolo.org> wrote:
> > Great! if you still have issues with the long mode, you could try to
> > hack gnumach to wait at some point (e.g. user_bootstrap() or
> > setup_main()), attach with gdb and then continue. Either with some delay
> > or polling a flag that you flip from gdb once attached. In this way you
> > should be able to set a breakpoint in userspace.
>
> OK, but how do I actually get the task to start / run? :D
> I can't debug it if it's not run...

OK, I was able to attach GDB prior to bootscript parsing thanks to
your suggestion (volatile flag that I flip from gdb).

It seems to parse the bootscript correctly. It starts the initial
thread, which loads the ELF and then hangs right here (on printf???):

(gdb) s
83 if (ph->p_flags & PF_R) type |= EXEC_SECTYPE_READ;
(gdb) bt
#0  exec_load (read=read@entry=0xffffffff8104a5a0 <boot_read>,
read_exec=read_exec@entry=0xffffffff8104a650 <read_exec>,
handle=0xffffffffbf11bf20,
    out_info=out_info@entry=0xffffffffbf112f78) at ../kern/elf-load.c:83
#1  0xffffffff8104ad5e in user_bootstrap () at ../kern/bootstrap.c:810
#2  0x0000000000000000 in ?? ()
(gdb) finish
Run till exit from #0  exec_load (read=read@entry=0xffffffff8104a5a0
<boot_read>, read_exec=read_exec@entry=0xffffffff8104a650 <read_exec>,
handle=0xffffffffbf11bf20,
    out_info=out_info@entry=0xffffffffbf112f78) at ../kern/elf-load.c:83
user_bootstrap () at ../kern/bootstrap.c:811
811  if (err)
Value returned is $61 = 0
(gdb) s
815  printf ("task loaded:");
(gdb) n

Does that ring any bells?

Sergey