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:   Thu, 22 Jul 2021 21:37:24 +0800
From:   Jisheng Zhang <jszhang3@mail.ustc.edu.cn>
To:     Atish Patra <atishp@atishpatra.org>
Cc:     Palmer Dabbelt <palmer@dabbelt.com>,
        Andreas Schwab <schwab@linux-m68k.org>, tongtiangen@huawei.com,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        linux-riscv <linux-riscv@lists.infradead.org>,
        "linux-kernel@vger.kernel.org List" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH -next v2] riscv: add VMAP_STACK overflow detection
Message-ID: <20210722213724.1f12a0e7@xhacker>
In-Reply-To: <CAOnJCU+Ss0cO1mqr=GDVnpxV075uR+KipSnr7dN93099dAH+vQ@mail.gmail.com>
References: <87bl6yrcmd.fsf@igel.home>
        <mhng-e14c3232-cc4d-4146-8c93-c60ec81ed272@palmerdabbelt-glaptop>
        <CAOnJCU+Ss0cO1mqr=GDVnpxV075uR+KipSnr7dN93099dAH+vQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Thu, 22 Jul 2021 01:35:23 -0700
Atish Patra <atishp@atishpatra.org> wrote:

> On Wed, Jul 21, 2021 at 11:12 PM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> >
> > On Mon, 19 Jul 2021 00:23:06 PDT (-0700), schwab@linux-m68k.org wrote:  
> > > On Jul 19 2021, tongtiangen wrote:
> > >  
> > >> On 2021/7/17 14:55, Andreas Schwab wrote:  
> > >>> Please use
> > >>> https://download.opensuse.org/repositories/home:/Andreas_Schwab:/riscv:/jeos/images/openSUSE-Tumbleweed-RISC-V-JeOS-efi.riscv64.raw.xz
> > >>> and run it in qemu with u-boot as kernel.
> > >>>
> > >>> Andreas.
> > >>>  
> > >>
> > >> Hi andreas:
> > >> I used today's latest mainline code and .config provided by you, and I
> > >> can't reproduce this panic.  
> > >
> > > Did you test it like I said above?
> > >
> > > Andreas.  
> >
> > I'm getting this on and off, with just
> >
> > CONFIG_VMAP_STACK=y
> >
> > on top of defconfig, when running on QEMU.  It's not showing up right
> > now: I'd thought it was an issue with that initrd patch, but it went
> > away when I re-ran the tests so I'm guessing it's something
> > non-deterministic.  I'll try to take a look if it comes back.
> >  
> 
> I got it very frequently on beagleV with the following branch & config.
> https://github.com/esmil/linux/commits/beaglev
> 
> beaglev_defconfig
> 
> Disabling CONFIG_VMAP_STACK avoids the crash.

Hi all,

I think we need to pin the stack before calling get_wchan(), could you please
try below patch?

Thanks

diff --git a/arch/riscv/kernel/stacktrace.c b/arch/riscv/kernel/stacktrace.c
index ff467b98c3e3..ac7593607fa6 100644
--- a/arch/riscv/kernel/stacktrace.c
+++ b/arch/riscv/kernel/stacktrace.c
@@ -132,8 +132,12 @@ unsigned long get_wchan(struct task_struct *task)
 {
 	unsigned long pc = 0;
 
-	if (likely(task && task != current && !task_is_running(task)))
+	if (likely(task && task != current && !task_is_running(task))) {
+		if (!try_get_task_stack(task))
+			return 0;
 		walk_stackframe(task, NULL, save_wchan, &pc);
+		put_task_stack(task);
+	}
 	return pc;
 }
 

> 
> > _______________________________________________
> > linux-riscv mailing list
> > linux-riscv@lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/linux-riscv  
> 
> 
>