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: <20220901171252.2148348-1-song@kernel.org> <Y3expGRt4cPoZgHL@alley>
 <CAPhsuW4qYpX7wzHn5J5Hn9cnOFSZwwQPCjTM_HPTt_zbBS03ww@mail.gmail.com> <Y5Me5dTGv+GznvtO@alley>
In-Reply-To: <Y5Me5dTGv+GznvtO@alley>
From:   Song Liu <song@kernel.org>
Date:   Tue, 13 Dec 2022 11:31:29 -0800
Message-ID: <CAPhsuW5eVyb10VM6WsHS_Y4uwDK3pS_9uuL-cKdOdMo+1p-peA@mail.gmail.com>
Subject: Re: powerpc-part: was: Re: [PATCH v6] livepatch: Clear relocation
 targets on a module removal
To:     Petr Mladek <pmladek@suse.com>
Cc:     live-patching@vger.kernel.org, linux-kernel@vger.kernel.org,
        jpoimboe@kernel.org, jikos@kernel.org, mbenes@suse.cz,
        x86@kernel.org, joe.lawrence@redhat.com,
        linuxppc-dev@lists.ozlabs.org,
        Josh Poimboeuf <jpoimboe@redhat.com>,
        Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>,
        Michael Ellerman <mpe@ellerman.id.au>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Dec 9, 2022 at 3:41 AM Petr Mladek <pmladek@suse.com> wrote:
>
> Hi,
>
> this reply is only about the powerpc-specific part.
>
> Also adding Kamalesh and Michael into Cc who worked on the related
> commit a443bf6e8a7674b86221f49 ("powerpc/modules: Add REL24 relocation
> support of livepatch symbols").
>
>
> On Mon 2022-11-28 17:57:06, Song Liu wrote:
> > On Fri, Nov 18, 2022 at 8:24 AM Petr Mladek <pmladek@suse.com> wrote:
> > >
> > > > --- a/arch/powerpc/kernel/module_64.c
> > > > +++ b/arch/powerpc/kernel/module_64.c
>
> I put back the name of the modified file so that it is easier
> to know what changes we are talking about.
>
> [...]
> > > > +#ifdef CONFIG_LIVEPATCH
> > > > +void clear_relocate_add(Elf64_Shdr *sechdrs,
> > > > +                    const char *strtab,
> > > > +                    unsigned int symindex,
> > > > +                    unsigned int relsec,
> > > > +                    struct module *me)
> > > > +{
> > > > +     unsigned int i;
> > > > +     Elf64_Rela *rela = (void *)sechdrs[relsec].sh_addr;
> > > > +     Elf64_Sym *sym;
> > > > +     unsigned long *location;
> > > > +     const char *symname;
> > > > +     u32 *instruction;
> > > > +
> > > > +     pr_debug("Clearing ADD relocate section %u to %u\n", relsec,
> > > > +              sechdrs[relsec].sh_info);
> > > > +
> > > > +     for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rela); i++) {
> > > > +             location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
> > > > +                     + rela[i].r_offset;
> > > > +             sym = (Elf64_Sym *)sechdrs[symindex].sh_addr
> > > > +                     + ELF64_R_SYM(rela[i].r_info);
> > > > +             symname = me->core_kallsyms.strtab
> > > > +                     + sym->st_name;
>
> The above calculation is quite complex. It seems to be copied from
> apply_relocate_add(). If I maintained this code I would want to avoid
> the duplication. definitely.
>

Back to this one...

If we go with option 2 that clear_relocate_add() only does things
needed to make the next apply_relocate_add() succeed, we are
not likely to have one write_relocate_add(), which is shared by
apply_relocate_add() and clear_relocate_add(). To avoid
duplication, shall we have two helpers to calculate location and
sym? Or maybe one more to calculate symname? I personally
don't like such one liner helper with multiple arguments, such as

static unsigned long *get_location(Elf64_Shdr *sechdrs,
           unsigned int relsec, unsigned int idx)
{
    Elf64_Rela *rela = (void *)sechdrs[relsec].sh_addr;

    return (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
         + rela[idx].r_offset;
}

Then use it as
     location = get_location(sechdrs, relsec, i);

We also need get_sym(), which is similar to get_location.

We can probably pass in different arguments. But I don't find
any options that I like. I think duplicate some code is better in
this case. However, if you do think these helpers are better,
or have other suggestions, I won't insist further.

Thanks,
Song