RE: 3.14 regression: huge latency in read/select on tun

2014-04-21 Thread Lukas Tribus
>> Hmm.. probably already solved by commit >> 1e785f48d29a09b6cf96db7b49b6320dada332e1 >> ("net: Start with correct mac_len in skb_network_protocol") > > Indeed that fixes it. I have cherry picked that onto 3.14 and it works. > That means 1e785f48 should be applied to stable 3.13.y and 3.14.y too.

Re: 3.14 regression: huge latency in read/select on tun

2014-04-21 Thread Ortwin Glück
On 20.04.2014 18:31, Eric Dumazet wrote:> On Sat, 2014-04-19 at 22:13 +0200, Ortwin Glück wrote: >> On 03.04.2014 15:50, Eric Dumazet wrote: >> >> I finally found time to bisect this: >> >> commit 53d6471cef17262d3ad1c7ce8982a234244f68ec >> Author: Vlad Yasevich >> Date: Thu Mar 27 17:26:18 20

Re: 3.14 regression: huge latency in read/select on tun

2014-04-20 Thread Eric Dumazet
On Sat, 2014-04-19 at 22:13 +0200, Ortwin Glück wrote: > On 03.04.2014 15:50, Eric Dumazet wrote: > > On Thu, 2014-04-03 at 06:19 -0700, Eric Dumazet wrote: > > > >> It seems TSO support is broken. > > I finally found time to bisect this: > > 53d6471cef17262d3ad1c7ce8982a234244f68ec is the first

Re: 3.14 regression: huge latency in read/select on tun

2014-04-19 Thread Ortwin Glück
On 03.04.2014 15:50, Eric Dumazet wrote: > On Thu, 2014-04-03 at 06:19 -0700, Eric Dumazet wrote: > >> It seems TSO support is broken. I finally found time to bisect this: 53d6471cef17262d3ad1c7ce8982a234244f68ec is the first bad commit commit 53d6471cef17262d3ad1c7ce8982a234244f68ec Author: Vla

Re: 3.14 regression: huge latency in read/select on tun

2014-04-04 Thread Ortwin Glück
On 03.04.2014 15:50, Eric Dumazet wrote: > If bug is in tun.c you have this list of changes you could bisect > from : > > # git log --oneline v3.13..v3.14 drivers/net/tun.c > 6671b2240c54 tun: remove bogus hardware vlan acceleration flags from > vlan_features > 99932d4fc03a netdevice: add queu

Re: 3.14 regression: huge latency in read/select on tun

2014-04-03 Thread Eric Dumazet
On Thu, 2014-04-03 at 06:19 -0700, Eric Dumazet wrote: > It seems TSO support is broken. If bug is in tun.c you have this list of changes you could bisect from : # git log --oneline v3.13..v3.14 drivers/net/tun.c 6671b2240c54 tun: remove bogus hardware vlan acceleration flags from vlan_feature

Re: 3.14 regression: huge latency in read/select on tun

2014-04-03 Thread Eric Dumazet
On Thu, 2014-04-03 at 09:52 +0200, Ortwin Glück wrote: > On 02.04.2014 18:42, Eric Dumazet wrote: > > You could post a tcpdump maybe ? > > This is a dump from on the tun interface (traffic inside the tunnel). > It's an scp upload of a 2MB file. I aborted after a while when not much > progress was

Re: 3.14 regression: huge latency in read/select on tun

2014-04-02 Thread Eric Dumazet
On Wed, 2014-04-02 at 18:17 +0200, Ortwin Glück wrote: > Hi, > > Since 3.14 the openconnect VPN tunnel becomes unusable for me because > packets appear on the tun device at a horribly low rate. 3.12 and 3.13 > do not exhibt the problem. > > Here is an strace of openconnect trying to read from its