Re: [PATCH v3 2/3] MIPS: Remove noreturn attribute for die()

From: Maciej W. Rozycki
Date: Thu Aug 17 2023 - 22:43:06 EST


On Mon, 14 Aug 2023, Tiezhu Yang wrote:

> If notify_die() returns NOTIFY_STOP, honor the return value from the
> handler chain invocation in die() as, through a debugger, the fault
> may have been fixed. It makes sense even if ignoring the event will
> make the system unstable, by allowing access through a debugger it
> has been compromised already anyway. So we can remove the noreturn
> attribute for die() to make our port consistent with x86, arm64,
> riscv and csky.

I find it weird that you say that it is specifically the removal of the
`noreturn' attribute that makes our port consistent with the other ones
(and make it the change heading too). I don't think you need to mention
the removal of `noreturn' even as you can see it in the code itself and
it's a natural consequence of the change proper. How about:

"
MIPS: Do not kill the task in die() if notify_die() returns NOTIFY_STOP

If notify_die() returns NOTIFY_STOP, honor the return value from the
handler chain invocation in die() and return without killing the task
as, through a debugger, the fault may have been fixed. It makes sense
even if ignoring the event will make the system unstable: by allowing
access through a debugger it has been compromised already anyway. It
makes our port consistent with x86, arm64, riscv and csky.
"

then (notice the use of a colon rather than a comma changing the meaning
of the sentence above)?

> Commit 20c0d2d44029 ("[PATCH] i386: pass proper trap numbers to die
> chain handlers") may be the earliest of similar changes.
>
> Link: https://lore.kernel.org/all/alpine.DEB.2.21.2308132148500.8596@xxxxxxxxxxxxxxxxx/

I think you meant:

Link: https://lore.kernel.org/r/43DDF02E.76F0.0078.0@xxxxxxxxxx/

didn't you?

> diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
> index 7a34674..4f5140f 100644
> --- a/arch/mips/kernel/traps.c
> +++ b/arch/mips/kernel/traps.c
> @@ -391,16 +391,15 @@ void show_registers(struct pt_regs *regs)
>
> static DEFINE_RAW_SPINLOCK(die_lock);
>
> -void __noreturn die(const char *str, struct pt_regs *regs)
> +void die(const char *str, struct pt_regs *regs)
> {
> static int die_counter;
> - int sig = SIGSEGV;
> + int ret;
>
> oops_enter();
>
> - if (notify_die(DIE_OOPS, str, regs, 0, current->thread.trap_nr,
> - SIGSEGV) == NOTIFY_STOP)
> - sig = 0;
> + ret = notify_die(DIE_OOPS, str, regs, 0,
> + current->thread.trap_nr, SIGSEGV);
>
> console_verbose();
> raw_spin_lock_irq(&die_lock);
> @@ -422,7 +421,8 @@ void __noreturn die(const char *str, struct pt_regs *regs)
> if (regs && kexec_should_crash(current))
> crash_kexec(regs);
>
> - make_task_dead(sig);
> + if (ret != NOTIFY_STOP)
> + make_task_dead(SIGSEGV);

It doesn't appear to me we should panic or execute the crash kernel if
the oops is to be suppressed. Can we just do what the x86 port does, that
is return if !sig after the call to `oops_exit'?

Also I note that the individual ports aren't exactly consistent here with
respect to each other, so maybe that's something you might want to post a
combined follow-up clean-up patch series for too?

Maciej