[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f4e8d869-401b-43f0-a326-52522730fb17@paulmck-laptop>
Date: Tue, 5 Sep 2023 06:26:11 -0700
From: "Paul E. McKenney" <paulmck@...nel.org>
To: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
Cc: David Laight <David.Laight@...lab.com>,
Denis Arefev <arefev@...mel.ru>,
Lai Jiangshan <jiangshanlai@...il.com>,
Josh Triplett <josh@...htriplett.org>,
Steven Rostedt <rostedt@...dmis.org>,
"rcu@...r.kernel.org" <rcu@...r.kernel.org>,
"lvc-project@...uxtesting.org" <lvc-project@...uxtesting.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"trufanov@...mel.ru" <trufanov@...mel.ru>,
"vfh@...mel.ru" <vfh@...mel.ru>
Subject: Re: [PATCH v2] The value may overflow
On Tue, Sep 05, 2023 at 08:26:51AM -0400, Mathieu Desnoyers wrote:
> On 9/5/23 05:31, David Laight wrote:
> > From: Mathieu Desnoyers
> > > Sent: 04 September 2023 11:24
> > >
> > > On 9/4/23 05:42, Denis Arefev wrote:
> > > > The value of an arithmetic expression 1 << (cpu - sdp->mynode->grplo)
> > > > is subject to overflow due to a failure to cast operands to a larger
> > > > data type before performing arithmetic
> > >
> > > The patch title should identify more precisely its context, e.g.:
> > >
> > > "srcu: Fix srcu_struct node grpmask overflow on 64-bit systems"
> > >
> > > Also, as I stated in my reply to the previous version, the patch commit
> > > message should describe the impact of the bug it fixes.
> >
> > And is the analysis complete?
> > Is 1UL right for 32bit archs??
> > Is 64 bits even enough??
>
> I understand from include/linux/rcu_node_tree.h and kernel/rcu/Kconfig
> RCU_FANOUT and RCU_FANOUT_LEAF ranges that a 32-bit integer is sufficient to
> hold the mask on 32-bit architectures, and a 64-bit integer is enough on
> 64-bit architectures given the current implementation.
>
> At least this appears to be the intent. I did not do a thorough analysis of
> the various parameter limits.
Mathieu has it right.
32-bit kernels are unaffected by this bug.
RCU_FANOUT_LEAF defaults to 16, which means that a 64-bit kernel would
need more than 32 leaf rcu_node structures for the parent rcu_node
structure to need more than 32 bit to track its children. This means
that more than 32*16=512 CPUs are required for this bug to affect 64-bit
systems. And there really are systems this big, so I am surprised that
this has not shown up long ago. But it would not be the first time that
objective reality surprised me. ;-)
Thanx, Paul
> Thanks,
>
> Mathieu
>
> >
> > David
> >
> > >
> > > Thanks,
> > >
> > > Mathieu
> > >
> > >
> > > >
> > > > Found by Linux Verification Center (linuxtesting.org) with SVACE.
> > > >
> > > > Signed-off-by: Denis Arefev <arefev@...mel.ru>
> > > > ---
> > > > v2: Added fixes to the srcu_schedule_cbs_snp function as suggested by
> > > > Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
> > > > kernel/rcu/srcutree.c | 4 ++--
> > > > 1 file changed, 2 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
> > > > index 20d7a238d675..6c18e6005ae1 100644
> > > > --- a/kernel/rcu/srcutree.c
> > > > +++ b/kernel/rcu/srcutree.c
> > > > @@ -223,7 +223,7 @@ static bool init_srcu_struct_nodes(struct srcu_struct *ssp, gfp_t gfp_flags)
> > > > snp->grplo = cpu;
> > > > snp->grphi = cpu;
> > > > }
> > > > - sdp->grpmask = 1 << (cpu - sdp->mynode->grplo);
> > > > + sdp->grpmask = 1UL << (cpu - sdp->mynode->grplo);
> > > > }
> > > > smp_store_release(&ssp->srcu_sup->srcu_size_state, SRCU_SIZE_WAIT_BARRIER);
> > > > return true;
> > > > @@ -833,7 +833,7 @@ static void srcu_schedule_cbs_snp(struct srcu_struct *ssp, struct srcu_node *snp
> > > > int cpu;
> > > >
> > > > for (cpu = snp->grplo; cpu <= snp->grphi; cpu++) {
> > > > - if (!(mask & (1 << (cpu - snp->grplo))))
> > > > + if (!(mask & (1UL << (cpu - snp->grplo))))
> > > > continue;
> > > > srcu_schedule_cbs_sdp(per_cpu_ptr(ssp->sda, cpu), delay);
> > > > }
> > >
> > > --
> > > Mathieu Desnoyers
> > > EfficiOS Inc.
> > > https://www.efficios.com
> >
> > -
> > Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> > Registration No: 1397386 (Wales)
>
> --
> Mathieu Desnoyers
> EfficiOS Inc.
> https://www.efficios.com
>
Powered by blists - more mailing lists