Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Wed, 30 Mar 2022 17:54:14 +0300
From:   Jarkko Sakkinen <jarkko@kernel.org>
To:     Reinette Chatre <reinette.chatre@intel.com>
Cc:     Shuah Khan <skhan@linuxfoundation.org>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Shuah Khan <shuah@kernel.org>,
        "open list:INTEL SGX" <linux-sgx@vger.kernel.org>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 1/2] selftests/sgx: Use rip relative addressing for
 encl_stack
Message-ID: <YkRvFkEO8Ic1VdQ8@iki.fi>
References: <20220322074313.7444-1-jarkko@kernel.org>
 <7b7732ec-c7ff-cf92-510f-64c83ed985cd@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <7b7732ec-c7ff-cf92-510f-64c83ed985cd@intel.com>
Precedence: bulk

On Mon, Mar 28, 2022 at 02:49:04PM -0700, Reinette Chatre wrote:
> Hi Jarkko,
> 
> On 3/22/2022 12:43 AM, Jarkko Sakkinen wrote:
> > Simplify the test_encl_bootstrap.S flow by using rip-relative addressing.
> > Compiler does the right thing here, and this removes dependency on where
> > TCS entries need to be located in the binary, i.e. allows the binary layout
> > changed freely in the future.
> > 
> > Cc: Reinette Chatre <reinette.chatre@intel.com>
> > Cc: Dave Hansen <dave.hansen@linux.intel.com>
> > Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
> > ---
> >  tools/testing/selftests/sgx/test_encl_bootstrap.S | 6 +-----
> >  1 file changed, 1 insertion(+), 5 deletions(-)
> > 
> > diff --git a/tools/testing/selftests/sgx/test_encl_bootstrap.S b/tools/testing/selftests/sgx/test_encl_bootstrap.S
> > index 82fb0dfcbd23..1c1b5c6c4ffe 100644
> > --- a/tools/testing/selftests/sgx/test_encl_bootstrap.S
> > +++ b/tools/testing/selftests/sgx/test_encl_bootstrap.S
> > @@ -40,11 +40,7 @@
> >  	.text
> >  
> >  encl_entry:
> > -	# RBX contains the base address for TCS, which is the first address
> > -	# inside the enclave for TCS #1 and one page into the enclave for
> > -	# TCS #2. By adding the value of encl_stack to it, we get
> > -	# the absolute address for the stack.
> > -	lea	(encl_stack)(%rbx), %rax
> > +	lea	(encl_stack)(%rip), %rax
> >  	xchg	%rsp, %rax
> >  	push	%rax
> >  
> 
> The goal of the above snippet is to set RSP to ensure that each thread has its own stack.
> 
> Since EENTER computes RIP as EnclaveBase + TCS.OENTRY, by using offset from RIP this
> would result in all TCS with OENTRY of encl_entry to use the same stack, no?
> 
> Could you please consider the following as an alternative:
> https://lore.kernel.org/lkml/65c137c875bd4da675eaba35316ff43d7cfd52f8.1644274683.git.reinette.chatre@intel.com/
> 
> The idea in that patch is that a new TCS would always need to be accompanied by a
> dedicated stack so, at least for testing purposes, the TCS and stack can be dynamically
> allocated together with the TCS page following its stack.  This seems much simpler
> to me and also makes the following patch unnecessary.

There's no better alternative than use rip. Compiler will fix it up.

So, no, I won't consider that. This a dead obvious change.

BR, Jarkko