Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 11 Aug 2023 19:57:22 -0700
In-Reply-To: <8edc91f9-ce20-9528-a496-5b6e650bb63f@redhat.com>
Mime-Version: 1.0
References: <20230811155255.250835-1-seanjc@google.com> <8edc91f9-ce20-9528-a496-5b6e650bb63f@redhat.com>
Message-ID: <ZNb1EpG1GDzRspPB@google.com>
Subject: Re: [PATCH] x86/retpoline: Don't clobber RFLAGS during srso_safe_ret()
From:   Sean Christopherson <seanjc@google.com>
To:     "Mika =?utf-8?B?UGVudHRpbMOk?=" <mpenttil@redhat.com>
Cc:     Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org,
        linux-kernel@vger.kernel.org, Srikanth Aithal <sraithal@amd.com>,
        kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Aug 11, 2023, Mika Penttil=C3=A4 wrote:
> > @@ -252,11 +252,10 @@ SYM_START(srso_untrain_ret, SYM_L_GLOBAL, SYM_A_N=
ONE)
> >   	.byte 0x48, 0xb8
> >   SYM_INNER_LABEL(srso_safe_ret, SYM_L_GLOBAL)
> > -	add $8, %_ASM_SP
> > +	lea 8(%_ASM_SP), %_ASM_SP
> >   	ret
> >   	int3
> >   	int3
> > -	int3
> >   	lfence
> >   	call srso_safe_ret
> >   	int3
> >=20
> > base-commit: 25aa0bebba72b318e71fe205bfd1236550cc9534
>=20
> Don't we have the same kind of problems with __x86_return_skl ?

Yep, forcing that path via "retbleed=3Dforce retbleed=3Dstuff spectre_v2=3D=
retpoline,generic"
yields the same failures.  I have no idea how to go about cleanly fixing th=
at.
The logic effectively requires modifying flags, the only thing I can think =
of is
to save/restore flags across the thunk, which seems beyond gross.

Given that no one has complained about this, I think I'd vote to simply dis=
able
KVM if call depth tracking is being used.