Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.networking > #276 > unrolled thread

Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised

Started byMark Hobley <markhobley@yahoo.donottypethisbit.co>
First post2011-05-21 14:58 +0000
Last post2011-05-22 00:44 -0400
Articles 9 on this page of 29 — 8 participants

Back to article view | Back to comp.os.linux.networking

Contents

  Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-21 14:58 +0000
    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised "David W. Hodgins" <dwhodgins@nomail.afraid.org> - 2011-05-21 15:16 -0400
      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-21 22:02 +0000
        Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2011-05-22 19:59 +0300
          Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 17:09 +0000
    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-21 20:20 +0000
      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Richard Kettlewell <rjk@greenend.org.uk> - 2011-05-21 22:47 +0100
      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-21 22:28 +0000
        Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 01:15 +0000
          Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-22 06:01 +0000
            Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 07:28 +0000
              Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Andy Furniss <spam@andyfurniss.entadsl.com> - 2011-05-22 11:40 +0100
                Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 12:25 +0000
                  Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Andy Furniss <spam@andyfurniss.entadsl.com> - 2011-05-22 20:40 +0100
                    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Andy Furniss <spam@andyfurniss.entadsl.com> - 2011-05-22 20:56 +0100
                      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 22:57 +0000
                    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 23:12 +0000
              Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-22 11:31 +0000
                Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 11:50 +0000
                  Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-22 12:52 +0000
                    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 13:27 +0000
                      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-05-22 19:19 +0000
                        Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-22 22:12 +0000
                          Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-25 15:55 +0000
                            ip route - Wrong Maximum Segment Size being advertised Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-25 17:27 +0000
                              Re: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions Mark Hobley <markhobley@yahoo.donottypethisbit.co> - 2011-05-29 00:34 +0000
                                Re: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions Andy Furniss <spam@andyfurniss.entadsl.com> - 2011-05-29 11:11 +0100
                    Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Martijn Lievaart <m@rtij.nl.invlalid> - 2011-05-23 08:13 +0200
      Re: Linux 2.6.39-rc7 - Wrong Maximum Segment Size being advertised Barry Margolin <barmar@alum.mit.edu> - 2011-05-22 00:44 -0400

Page 2 of 2 — ← Prev page 1 [2]

#291

FromMark Hobley <markhobley@yahoo.donottypethisbit.co>
Date2011-05-22 13:27 +0000
Message-ID<irb30e$fnn$1@dont-email.me>
In reply to#290
On Sun, 22 May 2011 12:52:59 +0000, Jorgen Grahn wrote:
> It seems to me it's more likely that you're doing something wrong --
> people would immediately notice if these parts of the Linux kernel
> broke! Is it the right route? Are there others?  What does route -ne
> say?

Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
0.0.0.0         10.0.0.1        0.0.0.0         UG        0 0          0 eth0
10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0

That is interesting. Shouldn't the MSS be 1392 in the top line?

> I also can't help wondering if you have caused all these problems
> somehow. If the ISP breaks things like this, it's commercial suicide --
> unless they somehow manage to keep Windows unaffected ...

Yeah. They don't support Linux at all.

-- 
Mark Hobley
Linux User: #370818  http://markhobley.yi.org/

[toc] | [prev] | [next] | [standalone]

#294

FromJorgen Grahn <grahn+nntp@snipabacken.se>
Date2011-05-22 19:19 +0000
Message-ID<slrnitioed.gtj.grahn+nntp@frailea.sa.invalid>
In reply to#291
On Sun, 2011-05-22, Mark Hobley wrote:
> On Sun, 22 May 2011 12:52:59 +0000, Jorgen Grahn wrote:
>> It seems to me it's more likely that you're doing something wrong --
>> people would immediately notice if these parts of the Linux kernel
>> broke! Is it the right route? Are there others?  What does route -ne
>> say?
>
> Kernel IP routing table
> Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
> 0.0.0.0         10.0.0.1        0.0.0.0         UG        0 0          0 eth0
> 10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0
>
> That is interesting. Shouldn't the MSS be 1392 in the top line?

I don't know about 1392, but if you have forced it, that column should
say so.  0 means "use the interface's MTU", and the interface is pure
Ethernet.

/Jorgen

-- 
  // Jorgen Grahn <grahn@  Oo  o.   .  .
\X/     snipabacken.se>   O  o   .

[toc] | [prev] | [next] | [standalone]

#297

FromMark Hobley <markhobley@yahoo.donottypethisbit.co>
Date2011-05-22 22:12 +0000
Message-ID<irc1od$ivj$1@dont-email.me>
In reply to#294
On Sun, 22 May 2011 19:19:42 +0000, Jorgen Grahn wrote:

> I don't know about 1392, but if you have forced it, that column should
> say so.  0 means "use the interface's MTU", and the interface is pure
> Ethernet.

Right. The ip command is not making the change then. I just tried to fix
this with route:

route del 0.0.0.0
SIOCDELRT: No such process
route add 0.0.0.0 gw 10.0.0.1 netmask 0.0.0.0 mss 1392

route -ne

Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
0.0.0.0         10.0.0.1        255.255.255.255 UGH    1392 0          0 eth0
0.0.0.0         10.0.0.1        0.0.0.0         UG        0 0          0 eth0
10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0

Hmmm. Not quite what I had in mind. How do I force the revised value into
the right place in the table?

Mark.

-- 
Mark Hobley
Linux User: #370818  http://markhobley.yi.org/

[toc] | [prev] | [next] | [standalone]

#312

FromMark Hobley <markhobley@yahoo.donottypethisbit.co>
Date2011-05-25 15:55 +0000
Message-ID<irj8pe$ke0$1@dont-email.me>
In reply to#297
On Sun, 22 May 2011 22:12:31 +0000, Mark Hobley wrote:

> route del 0.0.0.0
> SIOCDELRT: No such process
> route add 0.0.0.0 gw 10.0.0.1 netmask 0.0.0.0 mss 1392
> 
> route -ne
> 
> Kernel IP routing table
> Destination     Gateway         Genmask         Flags   MSS Window  irtt
> Iface 0.0.0.0         10.0.0.1        255.255.255.255 UGH    1392 0     
>     0 eth0 0.0.0.0         10.0.0.1        0.0.0.0         UG        0 0
>          0 eth0 10.0.0.0        0.0.0.0         255.255.255.0   U       
>  0 0          0 eth0
> 
> Hmmm. Not quite what I had in mind. How do I force the revised value
> into the right place in the table?

Could I change this value through the /proc interface? Or is there a system
call that I could use to change the MSS for the first entry in the table?

Mark.

-- 
Mark Hobley
Linux User: #370818  http://markhobley.yi.org/

[toc] | [prev] | [next] | [standalone]

#314 — ip route - Wrong Maximum Segment Size being advertised

FromMark Hobley <markhobley@yahoo.donottypethisbit.co>
Date2011-05-25 17:27 +0000
Subjectip route - Wrong Maximum Segment Size being advertised
Message-ID<irje56$1to$1@dont-email.me>
In reply to#312
On Wed, 25 May 2011 15:55:28 +0000, Mark Hobley wrote:

> Could I change this value through the /proc interface? Or is there a
> system call that I could use to change the MSS for the first entry in
> the table?

Actually, it looks as though the ip command manipulates the table. The value
it produces does not look right to me though:

ip route change default via 10.0.0.1 mtu lock 1412 advmss 1392

route -ne
Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
0.0.0.0         10.0.0.1        0.0.0.0         UG     1432 0          0 eth0
10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0

Huh! The value of MSS is now 1432 (which is bigger than the maximum
transmission unit). I ordered 1392.

Mark.

-- 
Mark Hobley
Linux User: #370818  http://markhobley.yi.org/

[toc] | [prev] | [next] | [standalone]

#321 — Re: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions

FromMark Hobley <markhobley@yahoo.donottypethisbit.co>
Date2011-05-29 00:34 +0000
SubjectRe: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions
Message-ID<irs4a7$eha$1@dont-email.me>
In reply to#314
On Wed, 25 May 2011 17:27:03 +0000, Mark Hobley wrote:

> Huh! The value of MSS is now 1432 (which is bigger than the maximum
> transmission unit). I ordered 1392.

Even with an MSS value in the table, the maximum segment size in the
datagrams exceeds the maximum transmission unit that can be carried:

ip route change default via 10.0.0.1 mtu lock 1412 advmss 1372
route -ne

Kernel IP routing table
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
0.0.0.0         10.0.0.1        0.0.0.0         UG     1412 0          0 eth0
10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0

01:13:58.554039 IP markhobley.yi.org.34525 > rosettacode.org.www: Flags [S], seq 2344941968, win 14600, options [mss 1460,sackOK,TS val 65981322 ecr 0,nop,wscale 2], length 0
01:13:59.163369 IP rosettacode.org.www > markhobley.yi.org.34525: Flags [S.], seq 1837671167, ack 2344941969, win 5792, options [mss 1380,sackOK,TS val 486273651 ecr 65981322,nop,wscale 7], length 0

(Note that the MSS is too big in both directions.)

Also, should the entry headed "MSS" in the routing table read "MTU"?

Mark.

-- 
Mark Hobley
Linux User: #370818  http://markhobley.yi.org/

[toc] | [prev] | [next] | [standalone]

#322 — Re: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions

FromAndy Furniss <spam@andyfurniss.entadsl.com>
Date2011-05-29 11:11 +0100
SubjectRe: Linux 2.6.39.1 - Wrong Maximum Segment Size in both directions
Message-ID<irt64t$sv4$1@localhost.localdomain>
In reply to#321
Mark Hobley wrote:
> On Wed, 25 May 2011 17:27:03 +0000, Mark Hobley wrote:
>
>> Huh! The value of MSS is now 1432 (which is bigger than the maximum
>> transmission unit). I ordered 1392.
>
> Even with an MSS value in the table, the maximum segment size in the
> datagrams exceeds the maximum transmission unit that can be carried:
>
> ip route change default via 10.0.0.1 mtu lock 1412 advmss 1372

try advmss lock 1372

> route -ne

ip ro ls. As you say, route (depreciated by ip?) is showing + 40 for 
whatever reason. tcpdump will show what is really happening.

>
> Kernel IP routing table
> Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
> 0.0.0.0         10.0.0.1        0.0.0.0         UG     1412 0          0 eth0
> 10.0.0.0        0.0.0.0         255.255.255.0   U         0 0          0 eth0
>
> 01:13:58.554039 IP markhobley.yi.org.34525>  rosettacode.org.www: Flags [S], seq 2344941968, win 14600, options [mss 1460,sackOK,TS val 65981322 ecr 0,nop,wscale 2], length 0
> 01:13:59.163369 IP rosettacode.org.www>  markhobley.yi.org.34525: Flags [S.], seq 1837671167, ack 2344941969, win 5792, options [mss 1380,sackOK,TS val 486273651 ecr 65981322,nop,wscale 7], length 0
>
> (Note that the MSS is too big in both directions.)

Incoming mss is set by the other end - if it's smaller than your mtu 
tcp/ip stack will respect that. If it's larger the packets sent should 
still be limited by your mtu setting.

Andy.

[toc] | [prev] | [next] | [standalone]

#301

FromMartijn Lievaart <m@rtij.nl.invlalid>
Date2011-05-23 08:13 +0200
Message-ID<3iapa8-tpf.ln1@news.rtij.nl>
In reply to#290
On Sun, 22 May 2011 12:52:59 +0000, Jorgen Grahn wrote:

> I also can't help wondering if you have caused all these problems
> somehow.  If the ISP breaks things like this, it's commercial suicide --
> unless they somehow manage to keep Windows unaffected ...

The non server versions of Windows use a smaller MTU (I recall 1300, but 
that may be wrong) to avoid pmtu blackholes, so yes, this is not 
impossible.

M4

[toc] | [prev] | [next] | [standalone]

#283

FromBarry Margolin <barmar@alum.mit.edu>
Date2011-05-22 00:44 -0400
Message-ID<barmar-5AE40C.00440122052011@news.eternal-september.org>
In reply to#278
In article <ir96q3$565$1@dont-email.me>,
 Mark Hobley <markhobley@yahoo.donottypethisbit.co> wrote:

> On Sat, 21 May 2011 14:58:32 +0000, Mark Hobley wrote:
> 
> > 18: Flags [S.], seq 3385954279, ack 3902988975, win 5792, options [mss
> > 1380,sackOK,TS val 140346418 ecr 47079743,nop,wscale 7], length 0
> 
> I'm not sure what this first line is, but win 5792 is bigger than any
> datagram that my ISP can carry.

Large window sizes are normal.  It's basically how many data bytes can 
be sent before waiting for acknowledgement.

The "wscale" option scales this up, so the actual window size is 
5792*2^7, i.e. 724KB.

http://en.wikipedia.org/wiki/TCP_window_scale_option

-- 
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.os.linux.networking

csiph-web