Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Zhangjin Wu <falcon@tinylab.org>
To:     w@1wt.eu
Cc:     arnd@arndb.de, david.laight@aculab.com, falcon@tinylab.org,
        linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
        linux-riscv@lists.infradead.org, thomas@t-8ch.de
Subject: Re: [PATCH v5 09/14] tools/nolibc: add missing my_syscall6() for mips
Date:   Mon,  3 Jul 2023 18:13:50 +0800
Message-Id: <20230703101350.493374-1-falcon@tinylab.org>
In-Reply-To: <20230702185549.GH16233@1wt.eu>
References: <20230702185549.GH16233@1wt.eu>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Feedback-ID: bizesmtp:tinylab.org:qybglogicsvrgz:qybglogicsvrgz5a-1
Precedence: bulk

Hi, Willy

> Hi Zhangjin,
> 
> On Wed, Jun 28, 2023 at 09:37:29PM +0800, Zhangjin Wu wrote:
> > It is able to pass the 6th argument like the 5th argument via the stack
> > for mips, let's add a new my_syscall6() now, see [1] for details:
> > 
> >   The mips/o32 system call convention passes arguments 5 through 8 on
> >   the user stack.
> > 
> > Both mmap() and pselect6() require my_syscall6().
> 
> Very interesting, I didn't manage to make it work previously. Did you
> test it to confirm that it works ? I guess so but since you didn't
> mention, I preferred to ask.
>

Yes, I'm curious too ;-) 

we did add the mmap test cases and run it for mips, as Thomas
suggested, we pass a none-zero pa_offset to mmap() to make sure the 6th
argument is used.

I just re-tested it for mips and printed something like this:

    44 mmap_bad = <0xffffffff> EINVAL                                [OK]
    45 munmap_bad = -1 EINVAL                                        [OK]
    46 mmap_munmap_good
    pa_offset: 8192
    length: 1
    file_size: 12287
     = 0                                          [OK]
    47 open_tty = 3                                                  [OK]
    48 open_blah = -1 ENOENT                                         [OK]

And I also checked the mips support of musl, it evan provide a
__syscall7, so, it should be ok ;-)

The only difference is, musl provide a different clobber list for
'__mips_isa_rev >= 6', I didn't look into the details yet:
https://elixir.bootlin.com/musl/latest/source/arch/mips/syscall_arch.h

Best regards,
Zhangjin

> Thanks!
> Willy