Re: [PATCH net-next v8 3/4] ptr_ring: move free-space check into separate helper

[Eric Dumazet <edumazet@google.com>]

On Thu, Mar 12, 2026 at 2:48 PM Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Thu, Mar 12, 2026 at 02:17:16PM +0100, Eric Dumazet wrote:
> > On Thu, Mar 12, 2026 at 2:06 PM Simon Schippers
> > <simon.schippers@tu-dortmund.de> wrote:
> > >
> > > This patch moves the check for available free space for a new entry into
> > > a separate function. As a result, __ptr_ring_produce() remains logically
> > > unchanged, while the new helper allows callers to determine in advance
> > > whether subsequent __ptr_ring_produce() calls will succeed. This
> > > information can, for example, be used to temporarily stop producing until
> > > __ptr_ring_peek() indicates that space is available again.
> > >
> > > Co-developed-by: Tim Gebauer <tim.gebauer@tu-dortmund.de>
> > > Signed-off-by: Tim Gebauer <tim.gebauer@tu-dortmund.de>
> > > Signed-off-by: Simon Schippers <simon.schippers@tu-dortmund.de>
> > > ---
> > >  include/linux/ptr_ring.h | 14 ++++++++++++--
> > >  1 file changed, 12 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h
> > > index 534531807d95..a5a3fa4916d3 100644
> > > --- a/include/linux/ptr_ring.h
> > > +++ b/include/linux/ptr_ring.h
> > > @@ -96,6 +96,14 @@ static inline bool ptr_ring_full_bh(struct ptr_ring *r)
> > >         return ret;
> > >  }
> > >
> > > +static inline int __ptr_ring_produce_peek(struct ptr_ring *r)
> > > +{
> > > +       if (unlikely(!r->size) || r->queue[r->producer])
> >
> > I think this should be
> >
> >        if (unlikely(!r->size) || READ_ONCE(r->queue[r->producer]))
> >
> > And of course:
> >
> > @@ -194,7 +194,7 @@ static inline void *__ptr_ring_peek(struct ptr_ring *r)
> >  static inline bool __ptr_ring_empty(struct ptr_ring *r)
> >  {
> >         if (likely(r->size))
> > -               return !r->queue[READ_ONCE(r->consumer_head)];
> > +               return !READ_ONCE(r->queue[READ_ONCE(r->consumer_head)]);
> >         return true;
> >  }
>
>
> I don't understand why it's necessary. consumer_head etc are
> all lock protected.
>
> queue itself is not but we are only checking it for NULL -
> it is fine if compiler reads it in many chunks and not all
> at once.
>
> > @@ -256,7 +256,7 @@ static inline void __ptr_ring_zero_tail(struct
> > ptr_ring *r, int consumer_head)
> >          * besides the first one until we write out all entries.
> >          */
> >         while (likely(head > r->consumer_tail))
> > -               r->queue[--head] = NULL;
> > +               WRITE_ONCE(r->queue[--head], NULL);
> >
> >         r->consumer_tail = consumer_head;
> >  }
> >
> >
> > Presumably we should fix this in net tree first.
>
>
> Maybe this one yes but I am not sure at all - KCSAN is happy.
>

Hmmm.. what about this trace ?

BUG: KCSAN: data-race in pfifo_fast_dequeue / pfifo_fast_enqueue

write to 0xffff88811d5ccc00 of 8 bytes by interrupt on cpu 0:
__ptr_ring_zero_tail include/linux/ptr_ring.h:259 [inline]
__ptr_ring_discard_one include/linux/ptr_ring.h:291 [inline]
__ptr_ring_consume include/linux/ptr_ring.h:311 [inline]
__skb_array_consume include/linux/skb_array.h:98 [inline]
pfifo_fast_dequeue+0x770/0x8f0 net/sched/sch_generic.c:770
dequeue_skb net/sched/sch_generic.c:297 [inline]
qdisc_restart net/sched/sch_generic.c:402 [inline]
__qdisc_run+0x189/0xc80 net/sched/sch_generic.c:420
qdisc_run include/net/pkt_sched.h:120 [inline]
net_tx_action+0x379/0x590 net/core/dev.c:5793
handle_softirqs+0xb9/0x280 kernel/softirq.c:622
do_softirq+0x45/0x60 kernel/softirq.c:523
__local_bh_enable_ip+0x70/0x80 kernel/softirq.c:450
local_bh_enable include/linux/bottom_half.h:33 [inline]
bpf_test_run+0x2db/0x620 net/bpf/test_run.c:426
bpf_prog_test_run_skb+0x9a4/0xef0 net/bpf/test_run.c:1159
bpf_prog_test_run+0x204/0x340 kernel/bpf/syscall.c:4721
__sys_bpf+0x52e/0x7e0 kernel/bpf/syscall.c:6246
__do_sys_bpf kernel/bpf/syscall.c:6341 [inline]
__se_sys_bpf kernel/bpf/syscall.c:6339 [inline]
__x64_sys_bpf+0x41/0x50 kernel/bpf/syscall.c:6339
x64_sys_call+0x10cb/0x3020 arch/x86/include/generated/asm/syscalls_64.h:322
do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
do_syscall_64+0x12c/0x370 arch/x86/entry/syscall_64.c:94
entry_SYSCALL_64_after_hwframe+0x77/0x7f

read to 0xffff88811d5ccc00 of 8 bytes by task 22632 on cpu 1:
__ptr_ring_produce include/linux/ptr_ring.h:106 [inline]
ptr_ring_produce include/linux/ptr_ring.h:129 [inline]
skb_array_produce include/linux/skb_array.h:44 [inline]
pfifo_fast_enqueue+0xd5/0x2c0 net/sched/sch_generic.c:741
dev_qdisc_enqueue net/core/dev.c:4144 [inline]
__dev_xmit_skb net/core/dev.c:4188 [inline]
__dev_queue_xmit+0x6a4/0x1f20 net/core/dev.c:4795
dev_queue_xmit include/linux/netdevice.h:3384 [inline]
__bpf_tx_skb net/core/filter.c:2153 [inline]
__bpf_redirect_common net/core/filter.c:2197 [inline]
__bpf_redirect+0x862/0x990 net/core/filter.c:2204
____bpf_clone_redirect net/core/filter.c:2487 [inline]
bpf_clone_redirect+0x20c/0x290 net/core/filter.c:2450
bpf_prog_53f18857bc887b09+0x22/0x2a
bpf_dispatcher_nop_func include/linux/bpf.h:1402 [inline]
__bpf_prog_run include/linux/filter.h:723 [inline]
bpf_prog_run include/linux/filter.h:730 [inline]
bpf_test_run+0x29d/0x620 net/bpf/test_run.c:423
bpf_prog_test_run_skb+0x9a4/0xef0 net/bpf/test_run.c:1159
bpf_prog_test_run+0x204/0x340 kernel/bpf/syscall.c:4721
__sys_bpf+0x52e/0x7e0 kernel/bpf/syscall.c:6246
__do_sys_bpf kernel/bpf/syscall.c:6341 [inline]
__se_sys_bpf kernel/bpf/syscall.c:6339 [inline]
__x64_sys_bpf+0x41/0x50 kernel/bpf/syscall.c:6339
x64_sys_call+0x10cb/0x3020 arch/x86/include/generated/asm/syscalls_64.h:322
do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
do_syscall_64+0x12c/0x370 arch/x86/entry/syscall_64.c:94
entry_SYSCALL_64_after_hwframe+0x77/0x7f

value changed: 0xffff888104a93a00 -> 0x0000000000000000

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 UID: 0 PID: 22632 Comm: syz.0.4135 Tainted: G W syzkaller #0
PREEMPT(full)
Tainted: [W]=WARN
Hardware name: Google Google Compute Engine/Google Compute Engine,
BIOS Google 01/24/2026