Am 09.12.2019 14:31, schrieb Mao Wenan: > There is softlockup when using TPACKET_V3: > ... > NMI watchdog: BUG: soft lockup - CPU#2 stuck for 60010ms! > (__irq_svc) from [<c0558a0c>] (_raw_spin_unlock_irqrestore+0x44/0x54) > (_raw_spin_unlock_irqrestore) from [<c027b7e8>] (mod_timer+0x210/0x25c) > (mod_timer) from [<c0549c30>] > (prb_retire_rx_blk_timer_expired+0x68/0x11c) > (prb_retire_rx_blk_timer_expired) from [<c027a7ac>] > (call_timer_fn+0x90/0x17c) > (call_timer_fn) from [<c027ab6c>] (run_timer_softirq+0x2d4/0x2fc) > (run_timer_softirq) from [<c021eaf4>] (__do_softirq+0x218/0x318) > (__do_softirq) from [<c021eea0>] (irq_exit+0x88/0xac) > (irq_exit) from [<c0240130>] (msa_irq_exit+0x11c/0x1d4) > (msa_irq_exit) from [<c0209cf0>] (handle_IPI+0x650/0x7f4) > (handle_IPI) from [<c02015bc>] (gic_handle_irq+0x108/0x118) > (gic_handle_irq) from [<c0558ee4>] (__irq_usr+0x44/0x5c) > ... > > If __ethtool_get_link_ksettings() is failed in > prb_calc_retire_blk_tmo(), msec and tmo will be zero, so tov_in_jiffies > is zero and the timer expire for retire_blk_timer is turn to > mod_timer(&pkc->retire_blk_timer, jiffies + 0), > which will trigger cpu usage of softirq is 100%. > > Fixes: f6fb8f100b80 ("af-packet: TPACKET_V3 flexible buffer implementation.") > Tested-by: Xiao Jiangfeng <xiaojiangfeng@xxxxxxxxxx> > Signed-off-by: Mao Wenan <maowenan@xxxxxxxxxx> > --- > net/packet/af_packet.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c > index 53c1d41fb1c9..118cd66b7516 100644 > --- a/net/packet/af_packet.c > +++ b/net/packet/af_packet.c > @@ -544,7 +544,8 @@ static int prb_calc_retire_blk_tmo(struct packet_sock *po, > msec = 1; > div = ecmd.base.speed / 1000; > } > - } > + } else > + return DEFAULT_PRB_RETIRE_TOV; > > mbits = (blk_size_in_bytes * 8) / (1024 * 1024); > With some litrle refactoring you can save one indent and make it more readable. err = __ethtool_get_settings(dev, &ecmd); if (err) return DEFAULT_PRB_RETIRE_TOV; speed = ethtool_cmd_speed(&ecmd); if (speed < SPEED_1000 || speed == SPEED_UNKNOWN) return DEFAULT_PRB_RETIRE_TOV; msec = 1; // never changes - needed at all ?? div = speed / 1000; jm2c re, wh