[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20201124113655.77b44aad@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date: Tue, 24 Nov 2020 11:36:55 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: davem@...emloft.net, netdev@...r.kernel.org, kernel-team@...com
Subject: Re: [PATCH net-next 1/3] net: remove napi_hash_del() from
driver-facing API
On Tue, 24 Nov 2020 20:11:42 +0100 Eric Dumazet wrote:
> >> gro_cells_init() is setting NAPI_STATE_NO_BUSY_POLL, and this was enough
> >> to not have one synchronize_net() call per netif_napi_del()
> >>
> >> I will test something like :
> >> I am not yet convinced the synchronize_net() is needed, since these
> >> NAPI structs are not involved in busy polling.
> >
> > IDK how this squares against netpoll, though?
> >
>
> Can we actually attach netpoll to a virtual device using gro_cells ?
Not 100% sure but I grabbed a machine running 5.2 and it appears to
work:
# ip tunnel add gre1 mode gre local 192.168.123.1 remote 192.168.123.2 ttl 255
# ip link set gre1 up
# ip addr add 10.12.34.56/30 dev gre1
# cd /sys/kernel/config/netconsole/
# cat enabled
0
# echo gre1 > dev_name
# echo 10.12.34.57 > remote_ip
# echo cb:a9:87:65:43:21 > remote_mac
# echo 1 > enabled
# dmesg | tail
[16859.016632] gre: GRE over IPv4 demultiplexor driver
[16859.019831] ip_gre: GRE over IPv4 tunneling driver
[16945.776625] netpoll: netconsole: local port 6665
[16945.776627] netpoll: netconsole: local IPv4 address 0.0.0.0
[16945.776628] netpoll: netconsole: interface 'gre1'
[16945.776629] netpoll: netconsole: remote port 6666
[16945.776630] netpoll: netconsole: remote IPv4 address 10.12.34.57
[16945.776631] netpoll: netconsole: remote ethernet address cb:a9:87:65:43:21
[16945.776633] netpoll: netconsole: local IP 10.12.34.56
[16945.776635] netconsole: network logging started
# ip li show dev gre1
6: gre1@...E: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1476 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/gre 192.168.123.1 peer 192.168.123.2
Not sure anything actually comes out on the other end of the pipe :)
Powered by blists - more mailing lists