lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 14 Jul 2022 14:24:56 -0700
From:   Nathan Chancellor <nathan@...nel.org>
To:     Nick Desaulniers <ndesaulniers@...gle.com>
Cc:     Kees Cook <keescook@...omium.org>,
        Sudip Mukherjee <sudipm.mukherjee@...il.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Tom Rix <trix@...hat.com>, Marco Elver <elver@...gle.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Josh Poimboeuf <jpoimboe@...nel.org>,
        "Peter Zijlstra (Intel)" <peterz@...radead.org>,
        linux-kernel@...r.kernel.org, llvm@...ts.linux.dev
Subject: Re: [PATCH] ubsan: disable UBSAN_DIV_ZERO for clang

On Thu, Jul 14, 2022 at 01:56:43PM -0700, Nick Desaulniers wrote:
> Building with UBSAN_DIV_ZERO with clang produces numerous fallthrough
> warnings from objtool.
> 
> In the case of uncheck division, UBSAN_DIV_ZERO may introduce new
> control flow to check for division by zero. Because the result of the
> division is undefined, LLVM may optimize the control flow such that
> after the call to __ubsan_handle_divrem_overflow doesn't matter. If
> panic_on_warn was set, __ubsan_handle_divrem_overflow would panic. The
> problem is is that panic_on_warn is run time configurable. If it's
> disabled, then we cannot guarantee that we will be able to recover
> safely.  Disable this config for clang until we can come up with a
> solution in LLVM.
> 
> Link: https://github.com/ClangBuiltLinux/linux/issues/1657
> Link: https://github.com/llvm/llvm-project/issues/56289
> Link: https://lore.kernel.org/lkml/CAHk-=wj1qhf7y3VNACEexyp5EbkNpdcu_542k-xZpzmYLOjiCg@mail.gmail.com/
> Reported-by: Sudip Mukherjee <sudipm.mukherjee@...il.com>
> Suggested-by: Linus Torvalds <torvalds@...ux-foundation.org>
> Signed-off-by: Nick Desaulniers <ndesaulniers@...gle.com>

Acked-by: Nathan Chancellor <nathan@...nel.org>

> ---
> Linus,
> I still think we should add explicit checks to gaurd against divide by
> zero.
> 
>  lib/Kconfig.ubsan | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/lib/Kconfig.ubsan b/lib/Kconfig.ubsan
> index a9f7eb047768..fd15230a703b 100644
> --- a/lib/Kconfig.ubsan
> +++ b/lib/Kconfig.ubsan
> @@ -84,6 +84,9 @@ config UBSAN_SHIFT
>  config UBSAN_DIV_ZERO
>  	bool "Perform checking for integer divide-by-zero"
>  	depends on $(cc-option,-fsanitize=integer-divide-by-zero)
> +	# https://github.com/ClangBuiltLinux/linux/issues/1657
> +	# https://github.com/llvm/llvm-project/issues/56289
> +	depends on !CC_IS_CLANG
>  	help
>  	  This option enables -fsanitize=integer-divide-by-zero which checks
>  	  for integer division by zero. This is effectively redundant with the
> -- 
> 2.37.0.170.g444d1eabd0-goog
> 
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ