[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190218104359.GB10743@fuggles.cambridge.arm.com>
Date: Mon, 18 Feb 2019 10:43:59 +0000
From: Will Deacon <will.deacon@....com>
To: Qian Cai <cai@....pw>
Cc: rostedt@...dmis.org, mingo@...hat.com, catalin.marinas@....com,
andreyknvl@...gle.com, aryabinin@...tuozzo.com,
linux-arm-kernel@...ts.infradead.org, kasan-dev@...glegroups.com,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] trace: skip hwasan
On Sat, Feb 16, 2019 at 11:34:34PM -0500, Qian Cai wrote:
> Enabling function tracer with CONFIG_KASAN_SW_TAGS=y (hwasan) tracer
> causes the whole system frozen on ThunderX2 systems with 256 CPUs,
> because there is a burst of too much pointer access, and then KASAN will
> dereference each byte of the shadow address for the tag checking which
> will kill all the CPUs.
>
> Signed-off-by: Qian Cai <cai@....pw>
> ---
> kernel/trace/Makefile | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
> index c2b2148bb1d2..fdd547a68385 100644
> --- a/kernel/trace/Makefile
> +++ b/kernel/trace/Makefile
> @@ -28,6 +28,11 @@ ifdef CONFIG_GCOV_PROFILE_FTRACE
> GCOV_PROFILE := y
> endif
>
> +# Too much pointer access will kill hwasan.
> +ifdef CONFIG_KASAN_SW_TAGS
> +KASAN_SANITIZE := n
> +endif
I don't maintain this file, but I think that my comments on your related
patch are relevant here as well:
https://lkml.org/lkml/2019/2/18/223
Will
Powered by blists - more mailing lists