Re: [PATCH v3 8/9] riscv: introduce interfaces to patch kernel code

[Zong Li <zong.li@sifive.com>]

On Thu, Apr 2, 2020 at 9:17 AM Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> Hi,
>
> On Wed, 1 Apr 2020 15:42:30 +0800
> Zong Li <zong.li@sifive.com> wrote:
>
> > > > +
> > > > +static int __kprobes riscv_insn_write(void *addr, const void *insn, size_t len)
> > >
> > > Why would you add "riscv_" prefix for those functions? It seems a bit odd.
> >
> > There is no particular reason, I just was used to adding a prefix for
> > arch-related stuff. I have no preference here, it's OK to me to remove
> > the prefix of these functions, do you think we need to remove them?
>
> Yeah, it will be better, unless it can mixed up with arch-independent
> functions.

OK. I'll remove it and use NOKPROBE_SYMBOL() instead of __kprobes annotation.

>
> > > > +{
> > > > +     void *waddr = addr;
> > > > +     bool across_pages = (((uintptr_t) addr & ~PAGE_MASK) + len) > PAGE_SIZE;
> > > > +     unsigned long flags = 0;
> > > > +     int ret;
> > > > +
> > > > +     raw_spin_lock_irqsave(&patch_lock, flags);
> > >
> > > This looks a bit odd since stop_machine() is protected by its own mutex,
> > > and also the irq is already disabled here.
> >
> > We need it because we don't always enter the riscv_patch_text_nosync()
> > through stop_machine mechanism. If we call the
> > riscv_patch_text_nosync() directly, we need a lock to protect the
> > page.
>
> Oh, OK, but it leads another question. Is that safe to patch the
> text without sync? Would you use it for UP system?
> I think it is better to clarify "in what case user can call _nosync()"
> and add a comment on it.

The ftrace is one of the cases, as documentation of ftrace said, when
dynamic ftrace is initialized, it calls kstop_machine to make the
machine act like a uniprocessor so that it can freely modify code
without worrying about other processors executing that same code. So
the ftrace called the _nosync interface here directly.

>
> Thank you,
>
> >
> > >
> > > Thank you,
> > >
> > > > +
> > > > +     if (across_pages)
> > > > +             patch_map(addr + len, FIX_TEXT_POKE1);
> > > > +
> > > > +     waddr = patch_map(addr, FIX_TEXT_POKE0);
> > > > +
> > > > +     ret = probe_kernel_write(waddr, insn, len);
> > > > +
> > > > +     patch_unmap(FIX_TEXT_POKE0);
> > > > +
> > > > +     if (across_pages)
> > > > +             patch_unmap(FIX_TEXT_POKE1);
> > > > +
> > > > +     raw_spin_unlock_irqrestore(&patch_lock, flags);
> > > > +
> > > > +     return ret;
> > > > +}
> > > > +#else
> > > > +static int __kprobes riscv_insn_write(void *addr, const void *insn, size_t len)
> > > > +{
> > > > +     return probe_kernel_write(addr, insn, len);
> > > > +}
> > > > +#endif /* CONFIG_MMU */
> > > > +
> > > > +int __kprobes riscv_patch_text_nosync(void *addr, const void *insns, size_t len)
> > > > +{
> > > > +     u32 *tp = addr;
> > > > +     int ret;
> > > > +
> > > > +     ret = riscv_insn_write(tp, insns, len);
> > > > +
> > > > +     if (!ret)
> > > > +             flush_icache_range((uintptr_t) tp, (uintptr_t) tp + len);
> > > > +
> > > > +     return ret;
> > > > +}
> > > > +
> > > > +static int __kprobes riscv_patch_text_cb(void *data)
> > > > +{
> > > > +     struct riscv_insn_patch *patch = data;
> > > > +     int ret = 0;
> > > > +
> > > > +     if (atomic_inc_return(&patch->cpu_count) == 1) {
> > > > +             ret =
> > > > +                 riscv_patch_text_nosync(patch->addr, &patch->insn,
> > > > +                                         GET_INSN_LENGTH(patch->insn));
> > > > +             atomic_inc(&patch->cpu_count);
> > > > +     } else {
> > > > +             while (atomic_read(&patch->cpu_count) <= num_online_cpus())
> > > > +                     cpu_relax();
> > > > +             smp_mb();
> > > > +     }
> > > > +
> > > > +     return ret;
> > > > +}
> > > > +
> > > > +int __kprobes riscv_patch_text(void *addr, u32 insn)
> > > > +{
> > > > +     struct riscv_insn_patch patch = {
> > > > +             .addr = addr,
> > > > +             .insn = insn,
> > > > +             .cpu_count = ATOMIC_INIT(0),
> > > > +     };
> > > > +
> > > > +     return stop_machine_cpuslocked(riscv_patch_text_cb,
> > > > +                                    &patch, cpu_online_mask);
> > > > +}
> > > > --
> > > > 2.25.1
> > > >
> > >
> > >
> > > --
> > > Masami Hiramatsu <mhiramat@kernel.org>
>
>
> --
> Masami Hiramatsu <mhiramat@kernel.org>