Re: [PATCH] net: udp: increase UDP_MIB_RCVBUFERRORS when ENOBUFS

2020-10-26 Thread Menglong Dong
Hello~ On Mon, Oct 26, 2020 at 5:52 PM Paolo Abeni wrote: > > Hello, > > On Mon, 2020-10-26 at 17:39 +0800, Menglong Dong wrote: > > The error returned from __udp_enqueue_schedule_skb is ENOMEM or ENOBUFS. > > For now, only ENOMEM is counted into UDP_MIB_RCVBUFERRORS in > > __udp_queue_rcv_skb. U

Re: [PATCH] net: udp: increase UDP_MIB_RCVBUFERRORS when ENOBUFS

2020-10-26 Thread Paolo Abeni
Hello, On Mon, 2020-10-26 at 17:39 +0800, Menglong Dong wrote: > The error returned from __udp_enqueue_schedule_skb is ENOMEM or ENOBUFS. > For now, only ENOMEM is counted into UDP_MIB_RCVBUFERRORS in > __udp_queue_rcv_skb. UDP_MIB_RCVBUFERRORS should count all of the > failed skb because of memor

[PATCH] net: udp: increase UDP_MIB_RCVBUFERRORS when ENOBUFS

2020-10-26 Thread Menglong Dong
The error returned from __udp_enqueue_schedule_skb is ENOMEM or ENOBUFS. For now, only ENOMEM is counted into UDP_MIB_RCVBUFERRORS in __udp_queue_rcv_skb. UDP_MIB_RCVBUFERRORS should count all of the failed skb because of memory errors during udp receiving, not just those because of the limit of so