Re: [PATCH] rcu: shift by 1UL rather than 1 to fix sign extension error

[Boqun Feng <boqun.feng@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 3496 bytes --]

On Wed, Dec 14, 2016 at 08:47:55AM +0800, Boqun Feng wrote:
> On Tue, Dec 13, 2016 at 10:36:47AM -0800, Paul E. McKenney wrote:
> > On Wed, Dec 14, 2016 at 02:09:27AM +0800, Boqun Feng wrote:
> > > 2016年12月14日 上午1:17，"Mark Rutland" <mark.rutland@arm.com>写道：
> > > >
> > > > Hi,
> > > >
> > > > On Tue, Dec 13, 2016 at 10:56:46AM +0000, Colin King wrote:
> > > > > From: Colin Ian King <colin.king@canonical.com>
> > > > >
> > > > > mask and bit are unsigned longs, so if bit is 31 we end up sign
> > > > > extending the 1 and mask ends up as 0xffffffff80000000. Fix this
> > > > > by explicitly adding integer suffix UL ensure 1 is a unsigned long
> > > > > rather than an signed int.
> > > > >
> > > > > Issue found with static analysis with CoverityScan, CID 1388564
> > > > >
> > > > > Fixes: 8965c3ce4718754db ("rcu: Use
> > > leaf_node_for_each_mask_possible_cpu() in force_qs_rnp()")
> > > > > Signed-off-by: Colin Ian King <colin.king@canonical.com>
> > > > > ---
> > > > >  kernel/rcu/tree.c | 2 +-
> > > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > >
> > > > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > > > > index 10162ac..6ecedd8 100644
> > > > > --- a/kernel/rcu/tree.c
> > > > > +++ b/kernel/rcu/tree.c
> > > > > @@ -3051,7 +3051,7 @@ static void force_qs_rnp(struct rcu_state *rsp,
> > > > >
> > > > >               leaf_node_for_each_mask_possible_cpu(rnp, rnp->qsmask,
> > > bit, cpu)
> > > > >                       if (f(per_cpu_ptr(rsp->rda, cpu), isidle, maxj))
> > > > > -                             mask |= 1 << bit;
> > > > > +                             mask |= 1UL << bit;
> > > >
> > > > So as to match the rest of the code altered in commit bc75e99983df1efd
> > > > ("rcu: Correctly handle sparse possible cpus"), and regardless of
> > > > naming, I think it'd be nicer to use leaf_node_cpu_bit(), e.g.
> > > >
> > > >         leaf_node_for_each_mask_possible_cpu(rnp, rnp->qsmask, bit, cpu)
> > > >                 if (f(per_cpu_ptr(rsp->rda, cpu), isidle, maxj))
> > > >                         mask |= leaf_node_cpu_bit(rnp, cpu);
> > > >
> > > > IMO, it would be nice to hide the iterator bit somehow, to match
> > > > for_each_leaf_node_possible_cpu(), which this largely looks similar to
> > > > otherwise.
> > > 
> > > Good point. ;-)
> > > 
> > > We can
> > > 
> > > #define for_each_leaf_node_cpu(rnp, mask, cpu) \
> > >         for((cpu) = (rnp)->grplo + find _first_bit(mask, MASK_BITS(mask)); \
> > >             (cpu) >= (rnp)->grplo && (cpu) <= (rnp)->grphi; \
> > >             (cpu) = (rnp)->grplo + find _next_bit(mask, ...,
> > > leaf_node_cpu_bit(rnp, cpu) + 1))) \
> > >                 if (!cpu_possible(cpu)) \
> > >                         continue; \
> > >                 else
> > 
> > What is the purpose of the cpu_possible() check?
> > 
> 
> To filter out CPUs in range [grplo, grphi] but not in cpu_possible_mask.
> 

Hmm.. if rcu_cpu_starting(cpu) is never called with "impossible" cpu,
IOW, ->qsmask and ->expmask never mask "impossible" cpus, then this is
just an over-care check.

I think I probably will remove this check eventually, let me audit the
code a little more for safety ;-)

Regards,
Boqun

> Regards,
> Boqun
> 
> > 							Thanx, Paul
> > 
> > > Typing from my cellphone, plz ignore the bad formatting ;-)
> > > 
> > > Regards,
> > > Boqun
> > > 
> > > > Thanks,
> > > > Mark.
> > 



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]