Re: [PATCH bpf-next 04/13] uprobes: Add arch_uprobe_verify_opcode function

[Jiri Olsa <olsajiri@gmail.com>]

On Fri, Dec 13, 2024 at 01:11:30PM -0800, Andrii Nakryiko wrote:
> On Fri, Dec 13, 2024 at 5:21 AM Jiri Olsa <olsajiri@gmail.com> wrote:
> >
> > On Thu, Dec 12, 2024 at 04:48:05PM -0800, Andrii Nakryiko wrote:
> > > On Wed, Dec 11, 2024 at 5:34 AM Jiri Olsa <jolsa@kernel.org> wrote:
> > > >
> > > > Adding arch_uprobe_verify_opcode function, so we can overload
> > > > verification for each architecture in following changes.
> > > >
> > > > Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> > > > ---
> > > >  include/linux/uprobes.h |  5 +++++
> > > >  kernel/events/uprobes.c | 19 ++++++++++++++++---
> > > >  2 files changed, 21 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/include/linux/uprobes.h b/include/linux/uprobes.h
> > > > index cc723bc48c1d..8843b7f99ed0 100644
> > > > --- a/include/linux/uprobes.h
> > > > +++ b/include/linux/uprobes.h
> > > > @@ -215,6 +215,11 @@ extern void uprobe_handle_trampoline(struct pt_regs *regs);
> > > >  extern void *arch_uretprobe_trampoline(unsigned long *psize);
> > > >  extern unsigned long uprobe_get_trampoline_vaddr(void);
> > > >  extern void uprobe_copy_from_page(struct page *page, unsigned long vaddr, void *dst, int len);
> > > > +extern int uprobe_verify_opcode(struct page *page, unsigned long vaddr, uprobe_opcode_t *new_opcode);
> > > > +extern int arch_uprobe_verify_opcode(struct arch_uprobe *auprobe, struct page *page,
> > > > +                                    unsigned long vaddr, uprobe_opcode_t *new_opcode,
> > > > +                                    int nbytes);
> > > > +extern bool arch_uprobe_is_register(uprobe_opcode_t *insn, int nbytes);
> > > >  #else /* !CONFIG_UPROBES */
> > > >  struct uprobes_state {
> > > >  };
> > > > diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
> > > > index 7c2ecf11a573..8068f91de9e3 100644
> > > > --- a/kernel/events/uprobes.c
> > > > +++ b/kernel/events/uprobes.c
> > > > @@ -263,7 +263,13 @@ static void uprobe_copy_to_page(struct page *page, unsigned long vaddr, const vo
> > > >         kunmap_atomic(kaddr);
> > > >  }
> > > >
> > > > -static int verify_opcode(struct page *page, unsigned long vaddr, uprobe_opcode_t *new_opcode)
> > > > +__weak bool arch_uprobe_is_register(uprobe_opcode_t *insn, int nbytes)
> > > > +{
> > > > +       return is_swbp_insn(insn);
> > >
> > > a bit weird that we ignore nbytes here... should we have nbytes ==
> > > UPROBE_SWBP_INSN_SIZE check somewhere here or inside is_swbp_insn()?
> >
> > the original is_swbp_insn function does not need that and we need
> > nbytes in the overloaded arch_uprobe_is_register to distinguish
> > between 1 byte and 5 byte update..
> >
> 
> and that's my point, if some architecture forgot to override it for
> nop5 (or similar stuff), this default implementation should reject
> instruction that's not an original nop, no?

ok, makes sense, will add that

thanks,
jirka