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;
References: <20190530190800.7633-1-luke.r.nels@gmail.com> <CAPhsuW4kMBSjpATqHrEhTmuqje=XZNGOrMyNur8f6K0RNQP=yw@mail.gmail.com> <CAB-e3NSidgz8gLRTL796A0DyRVePPjVDpSC6=gSA4hH8q6VqvQ@mail.gmail.com> <CAPhsuW7rOzyJTac7d9PPHeWW39Hu5pV6Mk0xJr8jyr0HH=-W2A@mail.gmail.com>
User-agent: mu4e 0.9.18; emacs 25.2.2
From:   Jiong Wang <jiong.wang@netronome.com>
To:     Luke Nelson <luke.r.nels@gmail.com>,
        Song Liu <liu.song.a23@gmail.com>
Cc:     Xi Wang <xi.wang@gmail.com>,
        =?utf-8?B?QmrDtnJuIFTDtnBlbA==?= <bjorn.topel@gmail.com>,
        Palmer Dabbelt <palmer@sifive.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
        Networking <netdev@vger.kernel.org>,
        linux-riscv@lists.infradead.org, bpf <bpf@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/2] bpf, riscv: fix bugs in JIT for 32-bit ALU operations
In-reply-to: <CAPhsuW7rOzyJTac7d9PPHeWW39Hu5pV6Mk0xJr8jyr0HH=-W2A@mail.gmail.com>
Date:   Fri, 31 May 2019 08:22:45 +0100
Message-ID: <87d0jzgkai.fsf@netronome.com>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


Song Liu writes:

> On Thu, May 30, 2019 at 3:34 PM Luke Nelson <luke.r.nels@gmail.com> wrote:
>>
>> On Thu, May 30, 2019 at 1:53 PM Song Liu <liu.song.a23@gmail.com> wrote:
>> >
>> > This is a little messy. How about we introduce some helper function
>> > like:
>> >
>> > /* please find a better name... */
>> > emit_32_or_64(bool is64, const u32 insn_32, const u32 inst_64, struct
>> > rv_jit_context *ctx)
>> > {
>> >        if (is64)
>> >             emit(insn_64, ctx);
>> >        else {
>> >             emit(insn_32, ctx);
>> >            rd = xxxx;
>> >            emit_zext_32(rd, ctx);
>> >        }
>> > }
>>
>> This same check is used throughout the file, maybe clean it up in a
>> separate patch?

We also need to enable the recent 32-bit opt (on bpf-next) on these missing
insns, like what has been done at:

  https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/commit/?id=66d0d5a854a6625974e7de4b874e7934988b0ef8

Perhaps the best way is to wait this patch merged back to bpf-next, then we
do two patches, the first one to enable the opt, the second one then do the
re-factor. I guess this could avoid some code conflict.

Regards,
Jiong

>
> Yes, let's do follow up patch.
>
> Thanks,
> Song