Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Joel Cunningham 

It's worth comparing the behavior of the netconn API.  Even though the netconn 
API is only available with NO_SYS = 0, it is still a user of the raw API like 
the applications mentioned in this email thread.



lwip_netconn_do_writemore() calls tcp_write() until all segments have been put 
into the TCP send buffer, then calls tcp_output().



Joel


On Jul 13, 2016, at 10:36 AM, Pîrvu Mihai  wrote:


Again, it's probably not be the wrong approach, it's just what I've been told 
on this thread:


https://lists.gnu.org/archive/html/lwip-users/2016-06/msg00055.html



Since it worked for me at that moment, I just let it like that, and reading the 
responses here confirms that it's the correct approach if you don't want to 
wait for the delay.


Mihai


On Wed, Jul 13, 2016 at 5:59 PM, Jakub Schmidtke  wrote:


I think you might have missed reading the documentation: tcp_write enqueue 
data, tcp_outout tries to send it. You always have to call both.




I actually found the recommended call flow on wiki (Here: 
http://lwip.wikia.com/wiki/Raw/TCP ) after Mihai mentioned using tcp_output().

And yes, I haven't fully read the docs/wiki yet, since I am trying to fix 
existing code written by someone else - who must have missed that...


However, Mihai mentions that calling tcp_output() might be the wrong 
approach... Is it really?


Thanks!

___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users



___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Pîrvu Mihai 
Again, it's probably not be the wrong approach, it's just what I've been
told on this thread:

https://lists.gnu.org/archive/html/lwip-users/2016-06/msg00055.html

Since it worked for me at that moment, I just let it like that, and reading
the responses here confirms that it's the correct approach if you don't
want to wait for the delay.

Mihai

On Wed, Jul 13, 2016 at 5:59 PM, Jakub Schmidtke  wrote:

> I think you might have missed reading the documentation: tcp_write enqueue
>> data, tcp_outout tries to send it. You always have to call both.
>>
>> I actually found the recommended call flow on wiki (Here:
> http://lwip.wikia.com/wiki/Raw/TCP ) after Mihai mentioned using
> tcp_output().
> And yes, I haven't fully read the docs/wiki yet, since I am trying to fix
> existing code written by someone else - who must have missed that...
>
> However, Mihai mentions that calling tcp_output() might be the wrong
> approach... Is it really?
>
> Thanks!
>
> ___
> lwip-users mailing list
> lwip-users@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/lwip-users
>
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Jakub Schmidtke 
>
> I think you might have missed reading the documentation: tcp_write enqueue
> data, tcp_outout tries to send it. You always have to call both.
>
> I actually found the recommended call flow on wiki (Here:
http://lwip.wikia.com/wiki/Raw/TCP ) after Mihai mentioned using
tcp_output().
And yes, I haven't fully read the docs/wiki yet, since I am trying to fix
existing code written by someone else - who must have missed that...

However, Mihai mentions that calling tcp_output() might be the wrong
approach... Is it really?

Thanks!
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Sergio R. Caprile 
Confusion might arise from the examples in contrib, neither tcpecho_raw 
nor smtp do call tcp_output(). Not everyone seems to look at httpd... 
nor read the wiki, nor the docs...


___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread goldsimon 
I think you might have missed reading the documentation: tcp_write enqueue 
data, tcp_outout tries to send it. You always have to call both.


Simon


Gesendet mit AquaMail für Android
http://www.aqua-mail.com


Am 13. Juli 2016 2:35:59 nachm. schrieb Jakub Schmidtke :



Maybe your sys_now() or sys_jiffies() functions are too course?
They should have a resolution of approx. 1 msec.



I am not even talking about observed delays.
I am looking at the code, and based on that, those packets will be delayed
by up to twice the value of TCP_TMR_INTERVAL in milliseconds.
TCP_TMR_INTERVAL is set, by default, to 250.

The TCP timer is scheduled to run every TCP_TMR_INTERVAL milliseconds.
Actually every TCP_TMR_INTERVAL values returned by sys_now(),
which is supposed to return milliseconds. That timer calls tcp_tmr(), which
will always call tcp_fasttmr() and call tcp_slowtmr() every other time.
Even the comments in tcp_tmr() say that tcp_slowtmr() will be called every
500ms.

And I am looking at tcp_write(). It is a pretty big function so I may be
missing something,
but it doesn't look like it's actually sending any of the packets being
written,
it just appends them to 'unsent' queue.

The actual writing happens inside tcp_slowtmr() which, again, is called
every 500ms.
So if something is written using tcp_write() 10ms after tcp_slowtmr()
triggered,
it will be actually sent out 490ms later. Even assuming perfect resolution
of sys_now
and being able to call things at exactly the right time. It's still up to a
half a second later!
Which is the problem I am experiencing.

Now I could modify TCP_TMR_INTERVAL to much lower value,
and call sys_check_timeouts() much more often to cause tcp_slowtmr() to be
called much more frequently.

Or modify sys_now() to return values that don't represent milliseconds but
something smaller
(and still call sys_check_timeouts() more often), but this changes the
whole time scale and I have no idea
what other effects this might have.

Both of these solutions feel ugly and could affect the performance.

I even tried calling tcp_slowtmr() manually, and it seems to send out those
TCP packets
sooner, but it doesn't feel like the right solution.

I haven't looked into manually calling tcp_output after each tcp_write as
suggested in an earlier response yet,
but the same response also says that it's a wrong thing to do...

That is why I am asking if I am missing something...

But depending on your hardware it could not be so easy to accomplish. BTW,

what is your hardware?



I am running it as a part of  a system that generates IP packets in
software.
Right now I am running it on a powerful i5 machine, so performance is not
an issue.



I've myself built with NO_SYS=1 on djgpp/MSDOS with no long delays
as you describe. Hence I think it's a problem with your 'clock-tick'
function.



What is your TCP_TMR_INTERVAL?
What do you return in sys_now()?


Thanks!



--
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Jakub Schmidtke 
>
> Maybe your sys_now() or sys_jiffies() functions are too course?
> They should have a resolution of approx. 1 msec.


I am not even talking about observed delays.
I am looking at the code, and based on that, those packets will be delayed
by up to twice the value of TCP_TMR_INTERVAL in milliseconds.
TCP_TMR_INTERVAL is set, by default, to 250.

The TCP timer is scheduled to run every TCP_TMR_INTERVAL milliseconds.
Actually every TCP_TMR_INTERVAL values returned by sys_now(),
which is supposed to return milliseconds. That timer calls tcp_tmr(), which
will always call tcp_fasttmr() and call tcp_slowtmr() every other time.
Even the comments in tcp_tmr() say that tcp_slowtmr() will be called every
500ms.

And I am looking at tcp_write(). It is a pretty big function so I may be
missing something,
but it doesn't look like it's actually sending any of the packets being
written,
it just appends them to 'unsent' queue.

The actual writing happens inside tcp_slowtmr() which, again, is called
every 500ms.
So if something is written using tcp_write() 10ms after tcp_slowtmr()
triggered,
it will be actually sent out 490ms later. Even assuming perfect resolution
of sys_now
and being able to call things at exactly the right time. It's still up to a
half a second later!
Which is the problem I am experiencing.

Now I could modify TCP_TMR_INTERVAL to much lower value,
and call sys_check_timeouts() much more often to cause tcp_slowtmr() to be
called much more frequently.

Or modify sys_now() to return values that don't represent milliseconds but
something smaller
(and still call sys_check_timeouts() more often), but this changes the
whole time scale and I have no idea
what other effects this might have.

Both of these solutions feel ugly and could affect the performance.

I even tried calling tcp_slowtmr() manually, and it seems to send out those
TCP packets
sooner, but it doesn't feel like the right solution.

I haven't looked into manually calling tcp_output after each tcp_write as
suggested in an earlier response yet,
but the same response also says that it's a wrong thing to do...

That is why I am asking if I am missing something...

But depending on your hardware it could not be so easy to accomplish. BTW,
> what is your hardware?
>

I am running it as a part of  a system that generates IP packets in
software.
Right now I am running it on a powerful i5 machine, so performance is not
an issue.


> I've myself built with NO_SYS=1 on djgpp/MSDOS with no long delays
> as you describe. Hence I think it's a problem with your 'clock-tick'
> function.
>

What is your TCP_TMR_INTERVAL?
What do you return in sys_now()?


Thanks!
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Gisle Vanem 
Jakub Schmidtke wrote:

> I am using lwip with NO_SYS=1, and I have noticed
> that outgoing TCP packets are very delayed.
> 
> It looks like between writing a packet (using tcp_write),
> and the time it actually written to the interface it takes usually about 
> 200ms.

Maybe your sys_now() or sys_jiffies() functions are too course?
They should have a resolution of approx. 1 msec. But depending on your
hardware it could not be so easy to accomplish. BTW, what is your hardware?

I've myself built with NO_SYS=1 on djgpp/MSDOS with no long delays
as you describe. Hence I think it's a problem with your 'clock-tick'
function.

-- 
--gv

___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-13 Thread Pîrvu Mihai 
I've worked around this disabling Nagle Algorithm for each pcb and also
calling tcp_output after every write. But somebody told me it was the wrong
approach doing so, but it works for me so I don't know what to say.

Mihai

On Wed, Jul 13, 2016 at 3:20 AM, Jakub Schmidtke  wrote:

> Hi,
>
> I am using lwip with NO_SYS=1, and I have noticed
> that outgoing TCP packets are very delayed.
>
> It looks like between writing a packet (using tcp_write),
> and the time it actually written to the interface it takes usually about
> 200ms.
>
> I examined the code, and it looks like lwip only writes regular packets
> every 500ms.
>
> Basically tcp_write only adds segments to a list, and they are only
> actually sent inside tcp_slowtmr.
> Which, by default, is only called once every two calls to tcp_tmr().
> Which gets called every 250ms, so the data being written with tcp_write
> will be sent out up to 500ms later.
>
> This is a really, really long time!
>
> Am I missing something?
>
>
> Thanks!
>
>
> ___
> lwip-users mailing list
> lwip-users@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/lwip-users
>
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

[lwip-users] lwIP delays outgoing TCP packets by up to 500ms

2016-07-12 Thread Jakub Schmidtke 
Hi,

I am using lwip with NO_SYS=1, and I have noticed
that outgoing TCP packets are very delayed.

It looks like between writing a packet (using tcp_write),
and the time it actually written to the interface it takes usually about
200ms.

I examined the code, and it looks like lwip only writes regular packets
every 500ms.

Basically tcp_write only adds segments to a list, and they are only
actually sent inside tcp_slowtmr.
Which, by default, is only called once every two calls to tcp_tmr().
Which gets called every 250ms, so the data being written with tcp_write
will be sent out up to 500ms later.

This is a really, really long time!

Am I missing something?


Thanks!
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users