Re: [PATCH v4 7/7] powerpc/64s: Blacklist rtas entry/exit from kprobes

[Nicholas Piggin <npiggin@gmail.com>]

On Thu, 29 Jun 2017 17:24:14 +0530
"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> wrote:

> On 2017/06/29 09:01PM, Nicholas Piggin wrote:
> > On Thu, 29 Jun 2017 16:11:10 +0530
> > "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> wrote:
> >   
> > > We can't take traps with relocation off, so blacklist enter_rtas() and
> > > rtas_return_loc(). However, instead of blacklisting all of enter_rtas(),
> > > introduce a new symbol __enter_rtas from where on we can't take a trap
> > > and blacklist that.
> > > 
> > > Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
> > > ---
> > >  arch/powerpc/kernel/entry_64.S | 3 +++
> > >  1 file changed, 3 insertions(+)
> > > 
> > > diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
> > > index 0c27084800b6..16f4c4a1a294 100644
> > > --- a/arch/powerpc/kernel/entry_64.S
> > > +++ b/arch/powerpc/kernel/entry_64.S
> > > @@ -1082,6 +1082,7 @@ _GLOBAL(enter_rtas)
> > >  	sync				/* disable interrupts so SRR0/1 */
> > >  	mtmsrd	r0			/* don't get trashed */
> > >  
> > > +__enter_rtas:  
> > 
> > Hmm, am I missing something, or is there a reason to put these labels
> > after the mtmsr? Even if kprobes does the right thing, I think it's
> > easier to read the code if you cover the mtmsr as well.  
> 
> I thought you asked for this, per your previous review comment:
> https://www.mail-archive.com/linuxppc-dev@lists.ozlabs.org/msg119667.html
> 
> Or, did I get that wrong?

No you're right, I'm contradicting myself. Let me start again.

I think we'd like to put the label before the mtmsrd if possible. So
in that case, should we adjust the system call code instead (then you
wouldn't have to add a comment for it).

And then you could move this label back above the mtmsrd. Sorry for
the confusion.

Thanks,
Nick