MIME-Version: 1.0
References: <20230126161559.1467374-1-guoren@kernel.org> <0abbbdd4-6b85-9659-03ee-97c56a5b77c1@huawei.com>
 <CAJF2gTS0s4X_uwLaEeSqKAyRmxCR2vxRuHhz7-SP2w4bBqzr+Q@mail.gmail.com> <87r0vc9h4g.fsf@all.your.base.are.belong.to.us>
In-Reply-To: <87r0vc9h4g.fsf@all.your.base.are.belong.to.us>
From:   Guo Ren <guoren@kernel.org>
Date:   Tue, 31 Jan 2023 09:01:31 +0800
Message-ID: <CAJF2gTQZWO8Ubi60nHn-J-8czqBnsp1Kdyoim3uLNkjbYd-H8Q@mail.gmail.com>
Subject: Re: [PATCH] riscv: kprobe: Optimize kprobe with accurate atomicity
To:     =?UTF-8?B?QmrDtnJuIFTDtnBlbA==?= <bjorn@kernel.org>
Cc:     "liaochang (A)" <liaochang1@huawei.com>, palmer@dabbelt.com,
        paul.walmsley@sifive.com, mhiramat@kernel.org,
        conor.dooley@microchip.com, penberg@kernel.org,
        mark.rutland@arm.com, linux-riscv@lists.infradead.org,
        linux-kernel@vger.kernel.org, Guo Ren <guoren@linux.alibaba.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Jan 30, 2023 at 11:28 PM Bj=C3=B6rn T=C3=B6pel <bjorn@kernel.org> w=
rote:
>
> Guo Ren <guoren@kernel.org> writes:
>
> >> In the serie of RISCV OPTPROBES [1], it patches a long-jump instructio=
ns pair
> >> AUIPC/JALR in kernel text, so in order to ensure other CPUs does not e=
xecute
> >> in the instructions that will be modified, it is still need to stop ot=
her CPUs
> >> via patch_text API, or you have any better solution to achieve the pur=
pose?
> >  - The stop_machine is an expensive way all architectures should
> > avoid, and you could keep that in your OPTPROBES implementation files
> > with static functions.
> >  - The stop_machine couldn't work with PREEMPTION, so your
> > implementation needs to work with !PREEMPTION.
>
> ...and stop_machine() with !PREEMPTION is broken as well, when you're
> replacing multiple instructions (see Mark's post at [1]). The
> stop_machine() dance might work when you're replacing *one* instruction,
> not multiple as in the RISC-V case. I'll expand on this in a comment in
> the OPTPROBES v6 series.
>
> >> >  static void __kprobes arch_prepare_simulate(struct kprobe *p)
> >> > @@ -114,16 +120,23 @@ void *alloc_insn_page(void)
> >> >  /* install breakpoint in text */
> >> >  void __kprobes arch_arm_kprobe(struct kprobe *p)
> >> >  {
> >> > -     if ((p->opcode & __INSN_LENGTH_MASK) =3D=3D __INSN_LENGTH_32)
> >> > -             patch_text(p->addr, __BUG_INSN_32);
> >> > -     else
> >> > -             patch_text(p->addr, __BUG_INSN_16);
> >> > +#ifdef CONFIG_RISCV_ISA_C
> >> > +     u32 opcode =3D __BUG_INSN_16;
> >> > +#else
> >> > +     u32 opcode =3D __BUG_INSN_32;
> >> > +#endif
> >> > +     patch_text_nosync(p->addr, &opcode, GET_INSN_LENGTH(opcode));
> >>
> >> Sounds good, but it will leave some RVI instruction truncated in kerne=
l text,
> >> i doubt kernel behavior depends on the rest of the truncated instructi=
on, well,
> >> it needs more strict testing to prove my concern :)
> > I do this on purpose, and it doesn't cause any problems. Don't worry;
> > IFU hw must enforce the fetch sequence, and there is no way to execute
> > broken instructions even in the speculative execution path.
>
> This is stretching reality a bit much. ARMv8, e.g., has a chapter in the
> Arm ARM [2] Appendix B "Concurrent modification and execution of
> instructions" (CMODX). *Some* instructions can be replaced concurrently,
> and others cannot without caution. Assuming that that all RISC-V
> implementations can, is a stretch. RISC-V hasn't even specified the
> behavior of CMODX (which is problematic).
Here we only use one sw/sh instruction to store a 32bit/16bit aligned eleme=
nt:

INSN_0 <- ebreak (16bit/32bit aligned)
INSN_1
INSN_2

The ebreak would cause an exception which implies a huge fence here.
No machine could give a speculative execution for the ebreak path.

>
> If anything it would be more likely that the existing
> "stop_machine()-to-replace-with-ebreak" works (again, replacing one
> instruction does not have the !PREEMPTION issues). Then again, no spec,
> so mostly guessing from my side. :-(
>
> Oh, but the existing "ebreak replace" might be broken like [3].
>
>
> Bj=C3=B6rn
>
>
> [1] https://lore.kernel.org/linux-riscv/Y7%2F6AtX5X0+5qF6Y@FVFF77S0Q05N/
> [2] https://developer.arm.com/documentation/ddi0487/latest
> [3] https://lore.kernel.org/linux-riscv/20230126170607.1489141-2-guoren@k=
ernel.org/


--=20
Best Regards
 Guo Ren