linux-kernel - Re: [PATCH] lib/crypto: curve25519-hacl64: Fix older clang KASAN workaround for GCC

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251103185500.GA3953783@google.com>
Date: Mon, 3 Nov 2025 18:55:00 +0000
From: Eric Biggers <ebiggers@...nel.org>
To: Nathan Chancellor <nathan@...nel.org>
Cc: "Jason A. Donenfeld" <Jason@...c4.com>,
	Ard Biesheuvel <ardb@...nel.org>, linux-crypto@...r.kernel.org,
	linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH] lib/crypto: curve25519-hacl64: Fix older clang KASAN
 workaround for GCC

On Mon, Nov 03, 2025 at 11:48:46AM -0700, Nathan Chancellor wrote:
> On Mon, Nov 03, 2025 at 09:00:36AM -0800, Eric Biggers wrote:
> > On Sun, Nov 02, 2025 at 09:35:03PM -0500, Nathan Chancellor wrote:
> > > Commit 2f13daee2a72 ("lib/crypto/curve25519-hacl64: Disable KASAN with
> > > clang-17 and older") inadvertently disabled KASAN in curve25519-hacl64.o
> > > for GCC unconditionally because clang-min-version will always evaluate
> > > to nothing for GCC. Add a check for CONFIG_CC_IS_GCC to avoid the
> > > workaround, which is only needed for clang-17 and older.
> > > 
> > > Additionally, invert the 'ifeq (...,)' into 'ifneq (...,y)', as it is a
> > > little easier to read and understand the intention ("if not GCC or at
> > > least clang-18, disable KASAN").
> > > 
> > > Cc: stable@...r.kernel.org
> > > Fixes: 2f13daee2a72 ("lib/crypto/curve25519-hacl64: Disable KASAN with clang-17 and older")
> > > Signed-off-by: Nathan Chancellor <nathan@...nel.org>
> > > ---
> > >  lib/crypto/Makefile | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/lib/crypto/Makefile b/lib/crypto/Makefile
> > > index bded351aeace..372b7a12b371 100644
> > > --- a/lib/crypto/Makefile
> > > +++ b/lib/crypto/Makefile
> > > @@ -90,7 +90,7 @@ else
> > >  libcurve25519-$(CONFIG_CRYPTO_LIB_CURVE25519_GENERIC) += curve25519-fiat32.o
> > >  endif
> > >  # clang versions prior to 18 may blow out the stack with KASAN
> > > -ifeq ($(call clang-min-version, 180000),)
> > > +ifneq ($(CONFIG_CC_IS_GCC)$(call clang-min-version, 180000),y)
> > >  KASAN_SANITIZE_curve25519-hacl64.o := n
> > >  endif
> > 
> > Thanks for catching this!
> > 
> > Using CONFIG_CC_IS_GCC == "" to check for clang seems a bit odd when
> > there's already a CONFIG_CC_IS_CLANG available.
> > 
> > How about we do it like this?
> > 
> >     ifeq ($(CONFIG_CC_IS_CLANG)_$(call clang-min-version, 180000),y_)
> 
> Yeah, I am not really sure why I was being so cryptic with the original
> way it was written :) I think it made a little more sense when it was
> 'ifeq'. Technically we could do without the _ but would you prefer that
> I keep it? I can send a v2 with whatever you prefer and an updated
> commit message.

While it shouldn't be possible to have a y just from the second part, I
think I'd prefer still having the separator so that it's clear which
part the y is coming from.

- Eric