Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [RFC v2 PATCH 7/7] x86/entry: use int for syscall number; handle
 all invalid syscall nrs
To:     "H. Peter Anvin" <hpa@zytor.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Andy Lutomirski <luto@kernel.org>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
References: <871racf928.ffs@nanos.tec.linutronix.de>
 <60495dd3-ea68-4db3-47ad-b7b45796bf76@zytor.com>
 <87o8dfer7k.ffs@nanos.tec.linutronix.de>
 <b0ca952a-fd82-7833-ac82-ea8ecad48b53@zytor.com>
 <87mtsz619u.ffs@nanos.tec.linutronix.de>
 <b6f47bd3-3e4c-7200-befe-5e14fecf3da7@zytor.com>
From:   Andy Lutomirski <andy@linux.luto.us>
Message-ID: <57bf2ff9-11ff-0907-975e-15177df72144@linux.luto.us>
Date:   Thu, 13 May 2021 20:18:37 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.8.1
MIME-Version: 1.0
In-Reply-To: <b6f47bd3-3e4c-7200-befe-5e14fecf3da7@zytor.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 5/13/21 5:38 PM, H. Peter Anvin wrote:
> On 5/12/21 3:22 PM, Thomas Gleixner wrote:
>>>
>>> As far as this being a user ABI change, this is actually a revert to the
>>> original x86-64 ABI; see my message to Ingo.
>>
>> I'm not against that change, but it has to be well justified and the
>> reasoning wants to be in the changelog. You know the drill :)
>>
> 
> FYI:
> 
> So in the process of breaking up and better document this patch, I have
> looked at the syscall_numbering_64 (and have rewritten it to be more
> complete.)
> 
> I found that running it under strace fails, as strace (possibly ptrace,
> possibly the strace binary) causes %rax = 2^32 to be clobbered to zero
> already...
> 
> More motivation, I guess.
> 

Indeed.

I would love to go back in time and switch to long, but there are plenty
of things that use int now.  I suppose we could try to make it long for
real, but seccomp has u32 baked into its ABI.