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;
MIME-Version: 1.0
References: <20230607091814.46080-1-puranjay12@gmail.com> <20230607091814.46080-4-puranjay12@gmail.com>
 <CAPhsuW7soOQasGw5fHB2qTeJnqR4ZrGBodyO87k=vg=TYqCsWA@mail.gmail.com> <CANk7y0ihF2RvwGXK0p0gOaGrMoPxfb6fcfn3-J40zn8LS8vROQ@mail.gmail.com>
In-Reply-To: <CANk7y0ihF2RvwGXK0p0gOaGrMoPxfb6fcfn3-J40zn8LS8vROQ@mail.gmail.com>
From:   Song Liu <song@kernel.org>
Date:   Thu, 8 Jun 2023 14:13:00 -0700
Message-ID: <CAPhsuW4KBYvE8oPz+4tF0_7zFoLhO-86Dh1LfZTiW1t59ty24g@mail.gmail.com>
Subject: Re: [PATCH bpf-next v2 3/3] bpf, arm64: use bpf_jit_binary_pack_alloc
To:     Puranjay Mohan <puranjay12@gmail.com>
Cc:     ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org,
        martin.lau@linux.dev, catalin.marinas@arm.com,
        mark.rutland@arm.com, bpf@vger.kernel.org, kpsingh@kernel.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Jun 8, 2023 at 9:51=E2=80=AFAM Puranjay Mohan <puranjay12@gmail.com=
> wrote:
>
> Hi Song,
>
> On Thu, Jun 8, 2023 at 6:28=E2=80=AFPM Song Liu <song@kernel.org> wrote:
> >
> > On Wed, Jun 7, 2023 at 2:18=E2=80=AFAM Puranjay Mohan <puranjay12@gmail=
.com> wrote:
> > >
> > [...]
> > > +
> > >  static inline int epilogue_offset(const struct jit_ctx *ctx)
> > >  {
> > >         int to =3D ctx->epilogue_offset;
> > > @@ -701,7 +716,8 @@ static int add_exception_handler(const struct bpf=
_insn *insn,
> > >                                  struct jit_ctx *ctx,
> > >                                  int dst_reg)
> > >  {
> > > -       off_t offset;
> > > +       off_t ins_offset;
> > > +       off_t fixup_offset;
> >
> > Please add some comments for these two offsets.
>
> Here I am using two variables because I need to change from the RO
> buffer for calculating offsets
> to the RW buffer for writing.
>
> Earlier, a single variable could work because it was being reused for
> calculating the second offset
> after writing the first one. Here, I can't re-calculate using the same
> variable because I have to change
> to the RW buffer, and using the same variable would need changing back
> to the RO buffer.
>
> So, I am calculating both offsets first, changing to RW buffer and
> writing both offsets.
>
> But I will add comments explaining what these offsets are being used for.

Thanks for the explanation! These comments will be useful for future work.

Song