Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221118194421.160414-1-ebiggers@kernel.org> <20221118194421.160414-9-ebiggers@kernel.org>
 <Y3fmskgfAb/xxzpS@sol.localdomain> <CABCJKudPXbDx2MSURDxGanTLhBkJjpMx=G=2RPDi6+96LGxcvw@mail.gmail.com>
In-Reply-To: <CABCJKudPXbDx2MSURDxGanTLhBkJjpMx=G=2RPDi6+96LGxcvw@mail.gmail.com>
From:   Sami Tolvanen <samitolvanen@google.com>
Date:   Fri, 18 Nov 2022 12:53:01 -0800
Message-ID: <CABCJKueoEkn7rWnDs7hb0nm84kKyyQuj5EVS_MtFNcfdT0D-EA@mail.gmail.com>
Subject: Re: [PATCH v2 08/12] crypto: x86/sm4 - fix crash with CFI enabled
To:     Eric Biggers <ebiggers@kernel.org>
Cc:     x86@kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-crypto@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Nov 18, 2022 at 12:27 PM Sami Tolvanen <samitolvanen@google.com> wrote:
>
> On Fri, Nov 18, 2022 at 12:10 PM Eric Biggers <ebiggers@kernel.org> wrote:
> > Sami, is it expected that a CFI check isn't being generated for the indirect
> > call to 'func' in sm4_avx_cbc_decrypt()?  I'm using LLVM commit 4a7be42d922af0.
>
> If the compiler emits an indirect call, it should also emit a CFI
> check. What's the assembly code it generates here?

With CONFIG_RETPOLINE, the check is emitted as expected, but I can
reproduce this issue without retpolines. It looks like the cfi-type
attribute is dropped from the machine instruction in one of the X86
specific passes. I'll take a look.

Sami