public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Almost 80% of UDP packets dropped
@ 2006-01-07  7:03 kernel coder
  2006-01-07 15:04 ` Denis Vlasenko
  0 siblings, 1 reply; 7+ messages in thread
From: kernel coder @ 2006-01-07  7:03 UTC (permalink / raw)
  To: linux-kernel

hi,
    I was trying to measure the UDP reception speed on my borad which
has MIPS 4kc processor with 133 MHZ speed.I was transfering 10mb file
from intel pentium 4 machine to MIPS board,but the recieved file was
only 900kB.

When i further investigated the problem ,i came to know that the user
application was not getting enough opportunities to get data from
socket queue which caused almost 80% of packets to be dropped as
socket queue had no free space.

When i increased the socket recieve buffer size,it resulted in
increase in no. of packets recieved .When i slow slowed down the
transmitter , it also caused more packets to be recieved.

But the above mentioned mechanism only decreased no. of lost
packets.But there was no way that i could increase UDP reception speed
because the user application was not getting enough opportunities to
read packets in burst of UDP packets.

I noticed that user application started recieveing packets after
Kernel had recieved all the UDP packets.

Please tell me how can i make sure that user application or udp client
running MIPS 4kc processor gets enough opportunities to dequeue
packets from socket buffer so that lost of packets could be reduced to
minimal and also the size of UDP recieved file in a specific interval
of time could be increased.

lhrkernelcoder

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07  7:03 Almost 80% of UDP packets dropped kernel coder
@ 2006-01-07 15:04 ` Denis Vlasenko
  2006-01-07 15:23   ` Tomasz Torcz
  2006-01-07 22:35   ` Bernd Eckenfels
  0 siblings, 2 replies; 7+ messages in thread
From: Denis Vlasenko @ 2006-01-07 15:04 UTC (permalink / raw)
  To: kernel coder; +Cc: linux-kernel

On Saturday 07 January 2006 09:03, kernel coder wrote:
> hi,
>     I was trying to measure the UDP reception speed on my borad which
> has MIPS 4kc processor with 133 MHZ speed.I was transfering 10mb file
> from intel pentium 4 machine to MIPS board,but the recieved file was
> only 900kB.

UDP is connectionless. There is no way for sender to know that it must
stop sending UDP packets because receiver cannot keep up. If sender
and your network is producing and delivering UDP packets faster
than receiver can consume them, packets will be lost.

Use TCP instead.
--
vda

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07 15:04 ` Denis Vlasenko
@ 2006-01-07 15:23   ` Tomasz Torcz
  2006-01-07 21:19     ` Jan Engelhardt
  2006-01-07 22:35   ` Bernd Eckenfels
  1 sibling, 1 reply; 7+ messages in thread
From: Tomasz Torcz @ 2006-01-07 15:23 UTC (permalink / raw)
  To: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 818 bytes --]

On Sat, Jan 07, 2006 at 05:04:52PM +0200, Denis Vlasenko wrote:
> On Saturday 07 January 2006 09:03, kernel coder wrote:
> > hi,
> >     I was trying to measure the UDP reception speed on my borad which
> > has MIPS 4kc processor with 133 MHZ speed.I was transfering 10mb file
> > from intel pentium 4 machine to MIPS board,but the recieved file was
> > only 900kB.
> 
> UDP is connectionless. There is no way for sender to know that it must
> stop sending UDP packets because receiver cannot keep up. If sender
> and your network is producing and delivering UDP packets faster
> than receiver can consume them, packets will be lost.
> 
> Use TCP instead.

 Or DCCP.

-- 
Tomasz Torcz                 "God, root, what's the difference?"
zdzichu@irc.-nie.spam-.pl         "God is more forgiving."


[-- Attachment #2: Type: application/pgp-signature, Size: 229 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07 15:23   ` Tomasz Torcz
@ 2006-01-07 21:19     ` Jan Engelhardt
  2006-01-07 22:44       ` Tomasz Torcz
  0 siblings, 1 reply; 7+ messages in thread
From: Jan Engelhardt @ 2006-01-07 21:19 UTC (permalink / raw)
  To: Tomasz Torcz; +Cc: linux-kernel


>> Use TCP instead.
>
> Or DCCP.

Don't you mean SCTP?



Jan Engelhardt
-- 

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07 15:04 ` Denis Vlasenko
  2006-01-07 15:23   ` Tomasz Torcz
@ 2006-01-07 22:35   ` Bernd Eckenfels
  1 sibling, 0 replies; 7+ messages in thread
From: Bernd Eckenfels @ 2006-01-07 22:35 UTC (permalink / raw)
  To: linux-kernel

Denis Vlasenko <vda@ilport.com.ua> wrote:
> UDP is connectionless. There is no way for sender to know that it must
> stop sending UDP packets because receiver cannot keep up. If sender
> and your network is producing and delivering UDP packets faster
> than receiver can consume them, packets will be lost.

Yes, but the problem of the OP sounds like the interrupt handling on mips
pretty badly affects the scheduler. I guess softwareinterrupts and network
polling would help here.

Gruss
Bernd

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07 21:19     ` Jan Engelhardt
@ 2006-01-07 22:44       ` Tomasz Torcz
  2006-01-09  5:52         ` kernel coder
  0 siblings, 1 reply; 7+ messages in thread
From: Tomasz Torcz @ 2006-01-07 22:44 UTC (permalink / raw)
  To: Jan Engelhardt; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 459 bytes --]

On Sat, Jan 07, 2006 at 10:19:56PM +0100, Jan Engelhardt wrote:
> 
> >> Use TCP instead.
> >
> > Or DCCP.
> 
> Don't you mean SCTP?

  No, DCCP, UDP-like protocol with TCP-like congestion control. In
mainline kernel since 2.6.14. DCCPv6 in the works.
http://www.wlug.org.nz/DCCP   http://lwn.net/Articles/149756/

-- 
Tomasz Torcz                 "God, root, what's the difference?"
zdzichu@irc.-nie.spam-.pl         "God is more forgiving."


[-- Attachment #2: Type: application/pgp-signature, Size: 229 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Almost 80% of UDP packets dropped
  2006-01-07 22:44       ` Tomasz Torcz
@ 2006-01-09  5:52         ` kernel coder
  0 siblings, 0 replies; 7+ messages in thread
From: kernel coder @ 2006-01-09  5:52 UTC (permalink / raw)
  To: Jan Engelhardt, linux-kernel

Thanks for all your help.Ethernet driver is already using NAPI.

The main problem was that the sender was running on P4 which has 3 gHz
speed and the reciever had speed 133MHz.So the sender was transmitting
packets far quicker than the reciever could handle.

Then i used the sme MIPS board as transmitter and another one as
receiver so now i recieved almost 80% of transmitted file.

shahzad

On 1/7/06, Tomasz Torcz <zdzichu@irc.pl> wrote:
> On Sat, Jan 07, 2006 at 10:19:56PM +0100, Jan Engelhardt wrote:
> >
> > >> Use TCP instead.
> > >
> > > Or DCCP.
> >
> > Don't you mean SCTP?
>
>  No, DCCP, UDP-like protocol with TCP-like congestion control. In
> mainline kernel since 2.6.14. DCCPv6 in the works.
> http://www.wlug.org.nz/DCCP   http://lwn.net/Articles/149756/
>
> --
> Tomasz Torcz                 "God, root, what's the difference?"
> zdzichu@irc.-nie.spam-.pl         "God is more forgiving."
>
>
>
>

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2006-01-09  5:52 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-01-07  7:03 Almost 80% of UDP packets dropped kernel coder
2006-01-07 15:04 ` Denis Vlasenko
2006-01-07 15:23   ` Tomasz Torcz
2006-01-07 21:19     ` Jan Engelhardt
2006-01-07 22:44       ` Tomasz Torcz
2006-01-09  5:52         ` kernel coder
2006-01-07 22:35   ` Bernd Eckenfels

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox