5.15-stable review patch. If anyone has any objections, please let me know.
------------------
From: Heiner Kallweit <hkallweit1@xxxxxxxxx>
commit eabb8a9be1e4a12f3bf37ceb7411083e3775672d upstream.
This reverts commit 7274c4147afbf46f45b8501edbdad6da8cd013b9.
Ken reported that RTL8125b can lock up if gro_flush_timeout has the
default value of 20000 and napi_defer_hard_irqs is set to 0.
In this scenario device interrupts aren't disabled, what seems to
trigger some silicon bug under heavy load. I was able to reproduce this
behavior on RTL8168h. Fix this by reverting 7274c4147afb.
Fixes: 7274c4147afb ("r8169: don't try to disable interrupts if NAPI is scheduled already")
Cc: stable@xxxxxxxxxxxxxxx
Reported-by: Ken Milmore <ken.milmore@xxxxxxxxx>
Signed-off-by: Heiner Kallweit <hkallweit1@xxxxxxxxx>
Reviewed-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Link: https://lore.kernel.org/r/9b5b6f4c-4f54-4b90-b0b3-8d8023c2e780@xxxxxxxxx
Signed-off-by: Paolo Abeni <pabeni@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/net/ethernet/realtek/r8169_main.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
--- a/drivers/net/ethernet/realtek/r8169_main.c
+++ b/drivers/net/ethernet/realtek/r8169_main.c
@@ -4617,10 +4617,8 @@ static irqreturn_t rtl8169_interrupt(int
rtl_schedule_task(tp, RTL_FLAG_TASK_RESET_PENDING);
}
- if (napi_schedule_prep(&tp->napi)) {
- rtl_irq_disable(tp);
- __napi_schedule(&tp->napi);
- }
+ rtl_irq_disable(tp);
+ napi_schedule(&tp->napi);
out:
rtl_ack_events(tp, status);
