Message-ID: <5ecc383c-621b-57d9-7f6d-d63496fca3b3@zytor.com>
Date:   Mon, 23 Jan 2023 16:01:11 -0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.0
Subject: Re: the x86 sysret_rip test fails on the Intel FRED architecture
Content-Language: en-US
To:     Andrew Cooper <Andrew.Cooper3@citrix.com>,
        Ammar Faizi <ammarfaizi2@gnuweeb.org>
Cc:     Dave Hansen <dave.hansen@intel.com>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        "Li, Xin3" <xin3.li@intel.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Peter Zijlstra <peterz@infradead.org>,
        "x86@kernel.org" <x86@kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
References: <SA1PR11MB6734FA9139B9C9F6CC2ED123A8C59@SA1PR11MB6734.namprd11.prod.outlook.com>
 <5d4ad3e3-034f-c7da-d141-9c001c2343af@intel.com>
 <18B5DB6D-AEBD-4A67-A7B3-CE64940819B7@zytor.com>
 <SA1PR11MB673498933098295BFC7C2900A8CB9@SA1PR11MB6734.namprd11.prod.outlook.com>
 <b6e36a5c-6f5e-eda6-54ad-a0c20eb00402@intel.com>
 <25b96960-a07e-a952-5c23-786b55054126@zytor.com>
 <fb1cab9f-a373-38e6-92e6-456332010653@gnuweeb.org>
 <6cd0db14-c9e2-3598-fd10-4b473d78c373@citrix.com>
From:   "H. Peter Anvin" <hpa@zytor.com>
In-Reply-To: <6cd0db14-c9e2-3598-fd10-4b473d78c373@citrix.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 1/23/23 15:53, Andrew Cooper wrote:
>>
>> A simple workaround is: just compile it with "-mno-red-zone" flag.
> 
> You can't compile userspace like that, unless you recompile and
> statically link everything including libc.
> 
> The proper way to fix this is to add a "+r" constraint on the stack
> pointer, at which point the compiler will arrange for nothing in the
> redzone to be live around the asm block.
> 

So you are of course right that that is The Right Way[TM] to fix this.

However, to be picky: since the redzone is not in use across function 
calls, it is perfectly legitimate to inhibit the redzone for one 
function only.

	-hpa