Re: [PATCH 1/3] ubsan: Add trap instrumentation option
From: Kees Cook
Date: Thu Nov 21 2019 - 12:20:59 EST
On Thu, Nov 21, 2019 at 03:52:52PM +0300, Andrey Ryabinin wrote:
> On 11/20/19 4:06 AM, Kees Cook wrote:
> > +config UBSAN_TRAP
> > + bool "On Sanitizer warnings, stop the offending kernel thread"
> That description seems inaccurate and confusing. It's not about kernel threads.
> UBSAN may trigger in any context - kernel thread/user process/interrupts...
> Probably most of the kernel code runs in the context of user process, so "stop the offending kernel thread"
> doesn't sound right.
> > + depends on UBSAN
> > + depends on $(cc-option, -fsanitize-undefined-trap-on-error)
> > + help
> > + Building kernels with Sanitizer features enabled tends to grow
> > + the kernel size by over 5%, due to adding all the debugging
> > + text on failure paths. To avoid this, Sanitizer instrumentation
> > + can just issue a trap. This reduces the kernel size overhead but
> > + turns all warnings into full thread-killing exceptions.
> I think we should mention that enabling this option also has a potential to
> turn some otherwise harmless bugs into more severe problems like lockups, kernel panic etc..
> So the people who enable this would better understand what they signing up for.
Good point about other contexts. I will attempt to clarify and send a
BTW, which tree should ubsan changes go through? The files are actually
not mentioned by anything in MAINTAINERS. Should the KASAN entry gain
paths to cover ubsan too? Something like:
diff --git a/MAINTAINERS b/MAINTAINERS
index 9dffd64d5e99..585434c013c4 100644
@@ -8824,7 +8824,7 @@ S: Maintained
+KERNEL SANITIZERS (KASAN, UBSAN)
M: Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx>
R: Alexander Potapenko <glider@xxxxxxxxxx>
R: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
@@ -8834,9 +8834,13 @@ F: arch/*/include/asm/kasan.h
M: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>