[PATCH RESEND] scripts/gdb: add lx_current support for riscv

Song Shuai suagrfillet at gmail.com
Wed Feb 22 02:13:04 PST 2023


Conor Dooley <conor.dooley at microchip.com> 于2023年2月22日周三 09:53写道:
>
> Hey Song
>
> On Wed, Feb 22, 2023 at 05:37:30PM +0800, Song Shuai wrote:
> > RISC-V uses the tp register to save the current task_struct address
> > as its current() defines. So lx_current() of riscv just returns the
> > dereference of the address cast via task_ptr_type.
>
> FYI, there's already been some patches posted for this:
> https://lore.kernel.org/all/mhng-bdcd75a5-e6d6-4090-8292-a32eb46e4464@palmer-ri-x1c9a/
>
Deepak's patch is much better than mine, which also takes the
userspace and trap handler into count.
So no need to care about this patch.
> That version already has review and ack tags from Palmer and others.
> It seems to me that Andrew is the usual applier of patches for this
> script, so I'm gonna add him to CC here so he is aware - but as it is
> the middle of the merge window the best course of action might just be
> reposting that patch, CCing Andrew, after v6.3-rc1.
>
> Cheers,
> Conor.
>
> >
> > Signed-off-by: Song Shuai <suagrfillet at gmail.com>
> > ---
> >  scripts/gdb/linux/cpus.py | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/scripts/gdb/linux/cpus.py b/scripts/gdb/linux/cpus.py
> > index 15fc4626d236..6825a9834394 100644
> > --- a/scripts/gdb/linux/cpus.py
> > +++ b/scripts/gdb/linux/cpus.py
> > @@ -173,6 +173,9 @@ def get_current_task(cpu):
> >           else:
> >               raise gdb.GdbError("Sorry, obtaining the current task is not allowed "
> >                                  "while running in userspace(EL0)")
> > +    elif utils.is_target_arch("riscv"):
> > +         current_task_addr = gdb.parse_and_eval("$tp")
> > +         return current_task_addr.cast(task_ptr_type).dereference()
> >      else:
> >          raise gdb.GdbError("Sorry, obtaining the current task is not yet "
> >                             "supported with this arch")
> > --
> > 2.20.1
> >
> >



-- 
Thanks,
Song



More information about the linux-riscv mailing list