Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20180516081910.10067-1-ynorov@caviumnetworks.com>
 <20180724173957.GA22106@yury-thinkpad> <20181010141017.GA2881@asgard.redhat.com>
 <7aac1a08-8948-1a04-cbd3-fbc6a53f9ff0@arm.com> <20181013020731.GD21972@asgard.redhat.com>
In-Reply-To: <20181013020731.GD21972@asgard.redhat.com>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Sun, 14 Oct 2018 21:53:22 +0200
Message-ID: <CAK8P3a1qUVoE0CYkCOU2N+-WkqMad15fFsakbwkpiiPPRFz-eA@mail.gmail.com>
Subject: Re: [PATCH v9 00/24] ILP32 for ARM64
To:     Eugene Syromiatnikov <esyr@redhat.com>
Cc:     Szabolcs Nagy <szabolcs.nagy@arm.com>,
        Yury Norov <ynorov@caviumnetworks.com>, nd@arm.com,
        Catalin Marinas <catalin.marinas@arm.com>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "open list:DOCUMENTATION" <linux-doc@vger.kernel.org>,
        linux-arch <linux-arch@vger.kernel.org>,
        Linux API <linux-api@vger.kernel.org>,
        Adam Borowski <kilobyte@angband.pl>,
        Alexander Graf <agraf@suse.de>,
        Alexey Klimov <klimov.linux@gmail.com>,
        Andreas Schwab <schwab@suse.de>,
        Andrew Pinski <pinskia@gmail.com>,
        Bamvor Zhangjian <bamv2005@gmail.com>,
        Chris Metcalf <cmetcalf@mellanox.com>,
        Christoph Muellner <christoph.muellner@theobroma-systems.com>,
        Dave Martin <Dave.Martin@arm.com>,
        David Miller <davem@davemloft.net>,
        Florian Weimer <fweimer@redhat.com>,
        Geert Uytterhoeven <geert@linux-m68k.org>,
        Heiko Carstens <heiko.carstens@de.ibm.com>,
        James Hogan <james.hogan@imgtec.com>,
        James Morse <james.morse@arm.com>,
        Joseph Myers <joseph@codesourcery.com>,
        Lin Yongting <linyongting@huawei.com>,
        "Manuel A. Fernandez Montecelo" <manuel.montezelo@gmail.com>,
        Mark Brown <broonie@kernel.org>,
        Martin Schwidefsky <schwidefsky@de.ibm.com>,
        Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>,
        Nathan Lynch <Nathan_Lynch@mentor.com>,
        Philipp Tomsich <philipp.tomsich@theobroma-systems.com>,
        Prasun Kapoor <Prasun.Kapoor@caviumnetworks.com>,
        Ramana Radhakrishnan <ramana.gcc@googlemail.com>,
        Steve Ellcey <sellcey@caviumnetworks.com>,
        Pavel Machek <pavel@ucw.cz>,
        Palmer Dabbelt <palmer@sifive.com>,
        Wookey <wookey@wookware.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sat, Oct 13, 2018 at 4:07 AM Eugene Syromiatnikov <esyr@redhat.com> wrote:
>
> On Wed, Oct 10, 2018 at 03:39:07PM +0100, Szabolcs Nagy wrote:
> > On 10/10/18 15:10, Eugene Syromiatnikov wrote:
> > >  * What's the reasoning behind capping syscall arguments to 32 bit? x32
> > >    and MIPS N32 do not have such a restriction (and do not need special
> > >    wrappers for syscalls that pass 64-bit values as a result, except
> > >    when they do,  as it is the case for preadv2 on x32); moreover, that
> > >    would lead to insurmountable difficulties for AArch64 ILP32 tracers
> > >    that try to trace LP64 tracees, as it would be impossible to pass
> > >    64-bit addresses to process_vm_{read,write} or ptrace PEEK/POKE.
> >
> > but that's necessarily the case for all ilp32 abis:
> > the userspace syscall function receives 32bit
> > arguments so even if the kernel abi takes 64bit
> > args you cannot use that from c code. (the libc
> > does not even know which args should be sign or
> > zero extended.)
>
> glibc's syscall() prototype has kernel_ulong_t as its arguments (more
> specifically, to __syscall_ulong_t, which is 64-bit wide on x32; it
> should also have kernel_long_t as its return type instead of long,
> but that's another story), so it works perfectly fine in case of x32.
>
> > process_vm_readv/writev is limited by the ilp32
> > iovec struct, not by the syscall arguments.
>
> Right, on x32/N32 this issue is worked around by the usage of the respective
> x86_64/N64 call, and it looks like another thing that is impossible
> with AArch64 ilp32.
>
> > ptrace is specified to take void* addr argument,
> > and void* is 32bit on all ilp32 targets.
> > so again on the c language level there is no
> > way around the 32bit limitation.
>
> Which is an issue.

I have no idea why you think this is a problem specific to aarch64-ilp32:
If we want to be able to debug 64-bit tasks from a 32-bit task on any
architecture that has compat mode, we should solve it once and
extend the ptrace interface to allow it on *all* of them. We certainly
don't need /more/ special cases for the x32 hack, there should really
be fewer of them.

      Arnd