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;
MIME-Version: 1.0
References: <20210921130127.24131-1-rpalethorpe@suse.com> <CAK8P3a29ycNqOC_pD-UUtK37jK=Rz=nik=022Q1XtXr6-o6tuA@mail.gmail.com>
 <87o88mkor1.fsf@suse.de> <87lf3qkk72.fsf@suse.de> <87ilytkngp.fsf@suse.de>
 <CAK8P3a2S=a0aw8GY8fZxaU5fz7ZkdehtHgStkn2=u9gO28GVEw@mail.gmail.com> <87fstvlifu.fsf@suse.de>
In-Reply-To: <87fstvlifu.fsf@suse.de>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Thu, 23 Sep 2021 12:25:07 +0200
Message-ID: <CAK8P3a3H3ZikSfqoeh0=PQQuFw6PTg2KgP+p=4G26ayuzgwQeQ@mail.gmail.com>
Subject: Re: ia32 signed long treated as x64 unsigned int by __ia32_sys*
To:     rpalethorpe@suse.de
Cc:     Arnd Bergmann <arnd@arndb.de>,
        Linux FS-devel Mailing List <linux-fsdevel@vger.kernel.org>,
        Linux API <linux-api@vger.kernel.org>,
        linux-aio <linux-aio@kvack.org>,
        y2038 Mailman List <y2038@lists.linaro.org>,
        Andy Lutomirski <luto@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        "the arch/x86 maintainers" <x86@kernel.org>,
        "H. Peter Anvin" <hpa@zytor.com>,
        Deepa Dinamani <deepa.kernel@gmail.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        LTP List <ltp@lists.linux.it>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Sep 23, 2021 at 12:01 PM Richard Palethorpe <rpalethorpe@suse.de> wrote:
> Arnd Bergmann <arnd@arndb.de> writes:
> > On Wed, Sep 22, 2021 at 10:46 AM Richard Palethorpe <rpalethorpe@suse.de> wrote:
> >> Richard Palethorpe <rpalethorpe@suse.de> writes:
> >
> > I also noticed that only x86 and s390 even have separate entry
> > points for normal syscalls when called in compat mode, while
> > the others all just zero the upper halves of the registers in the
> > low-level entry code and then call the native entry point.
>
> It looks to me like aarch64 also has something similar? At any rate, I
> can try to fix it for x86 and investigate what else might be effected.

arm64 also has a custom asm/syscall_wrapper.h, but it only does
this for accessing pt_regs (as x86 does), not for doing any
argument conversion. x86 does the 32-to-64 widening in the
wrapper, arm64 relies on the pt_regs already having the upper
halves zeroed.

        Arnd