I missed the obvious case where netif_ix() is invoked from hard-IRQ
context.
Disabling bottom halves is only needed in process context. This ensures
that the code remains on the current CPU and that the soft-interrupts
are processed at local_bh_enable() time.
In hard- and soft-interrupt context this is already the case and the
soft-interrupts will be processed once the context is left (at irq-exit
time).
Disable bottom halves if neither hard-interrupts nor soft-interrupts are
disabled. Update the kernel-doc, mention that interrupts must be enabled
if invoked from process context.
Fixes: baebdf48c3600 ("net: dev: Makes sure netif_rx() can be invoked in any context.")
Reported-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
---
Marek, does this work for you?
net/core/dev.c | 11 ++++++++---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index 909fb38159108..87729491460fc 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4860,7 +4860,9 @@ EXPORT_SYMBOL(__netif_rx);
* congestion control or by the protocol layers.
* The network buffer is passed via the backlog NAPI device. Modern NIC
* driver should use NAPI and GRO.
- * This function can used from any context.
+ * This function can used from interrupt and from process context. The
+ * caller from process context must not disable interrupts before invoking
+ * this function.
*
* return values:
* NET_RX_SUCCESS (no congestion)
@@ -4870,12 +4872,15 @@ EXPORT_SYMBOL(__netif_rx);
int netif_rx(struct sk_buff *skb)
{
int ret;
+ bool need_bh_off = !(hardirq_count() | softirq_count());
- local_bh_disable();
+ if (need_bh_off)
+ local_bh_disable();
trace_netif_rx_entry(skb);
ret = netif_rx_internal(skb);
trace_netif_rx_exit(ret);
- local_bh_enable();
+ if (need_bh_off)
+ local_bh_enable();
return ret;
}
EXPORT_SYMBOL(netif_rx);
--
2.34.1
