Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 2 Apr 2021 21:32:06 +0800
From:   Jisheng Zhang <jszhang3@mail.ustc.edu.cn>
To:     "liaochang (A)" <liaochang1@huawei.com>
Cc:     Masami Hiramatsu <mhiramat@kernel.org>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Guo Ren <guoren@linux.alibaba.com>,
        <linux-riscv@lists.infradead.org>, <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] riscv: keep interrupts disabled for BREAKPOINT
 exception
Message-ID: <20210402213206.7db93785@xhacker>
In-Reply-To: <096752a7-21f9-8339-2e76-8c65859fa1d7@huawei.com>
References: <20210330021624.2b776386@xhacker>
        <20210330183316.942215efe8e6e8455ad14113@kernel.org>
        <20210331222244.45a5807c@xhacker>
        <096752a7-21f9-8339-2e76-8c65859fa1d7@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, 1 Apr 2021 16:49:47 +0800
"liaochang (A)" <liaochang1@huawei.com> wrote:

> Hi Jisheng,

Hi,

>=20
> =E5=9C=A8 2021/3/31 22:22, Jisheng Zhang =E5=86=99=E9=81=93:
> > On Tue, 30 Mar 2021 18:33:16 +0900
> > Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >  =20
> >> Hi Jisheng, =20
> >=20
> > Hi Masami,
> >  =20
> >>
> >> On Tue, 30 Mar 2021 02:16:24 +0800
> >> Jisheng Zhang <jszhang3@mail.ustc.edu.cn> wrote:
> >> =20
> >>> From: Jisheng Zhang <jszhang@kernel.org>
> >>>
> >>> Current riscv's kprobe handlers are run with both preemption and
> >>> interrupt enabled, this violates kprobe requirements. Fix this issue
> >>> by keeping interrupts disabled for BREAKPOINT exception.   =20
> >>
> >> Not only while the breakpoint exception but also until the end of
> >> the single step (maybe you are using __BUG_INSN_32 ??) need to be
> >> disable interrupts. Can this do that?
> >> =20
> >=20
> > interrupt is disabled during "single step" by kprobes_save_local_irqfla=
g()
> > and kprobes_restore_local_irqflag(). The code flow looks like:=20
> >=20
> > do_trap_break()   // for bp
> >   kprobe_breakpoint_handler()
> >     setup_singlestep()
> >       kprobes_restore_local_irqflag()
> >=20
> > do_trap_break()  // for ss
> >   kprobe_single_step_handler()
> >     kprobes_restore_local_irqflag() =20
>=20
> Recently, kernel hit BUG_ON() on QEMU after I install a probe at "sys_rea=
d" via kprobe,

TIPS: Each line should not exceed 80 chars

> accoriding to my debugging and analysis it looks like caused by the "irq =
disable" operation for single-stepping.
>=20
> I present a detailed description about this problem in the mail with titl=
e "[PATCH] riscv/kprobe: fix kernel panic when invoking sys_read traced by =
kprobe".
> Looking forward to some feedback,Thanks.
>=20

I will comment that patch.

thanks