TCP/IPv4/uIP stack appears to be broken with recent commit


Paul Sokolovsky
 

Hello,

To make sure that my comment to
https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd like to
dup it with this email. The patch above has broke echo_server
TCP/IPv4 for me. When run on frdm_k64f over Ethernet, the application
answers pings, but being telneted to port 4242, it doesn't echo TCP
data. Note that the issue of "The trouble is that that linkage between
the outgoing buffer and the connection hasn't been established at the
point data_is_sent_and_acked() is called" has been fixed some time ago
by https://gerrit.zephyrproject.org/r/#/c/4888/ .

Reverting just that commit fixes issue for me.

--
Best Regards,
Paul

Linaro.org | Open source software for ARM SoCs
Follow Linaro: http://www.facebook.com/pages/Linaro
http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog


Andy Ross
 

Paul Sokolovsky wrote (on Tuesday, October 11, 2016 10:22AM):
To make sure that my comment to
https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd like to
dup it with this email. The patch above has broke echo_server
FWIW, I don't know enough about the uIP stack to add anything
technical, but:

An even better way to make sure that report isn't lost is to submit it
to the bug tracker. :)

https://jira.zephyrproject.org/

It should take the same LF authentication that gerrit does.

Andy


Jukka Rissanen
 

Hi Paul,

we can revert https://gerrit.zephyrproject.org/r/#/c/4282/6 if needed.
The patch fixed an issue Rohit had thou. Best would be to find a
solution that fixes both problems.

Cheers,
Jukka

On Tue, 2016-10-11 at 20:22 +0300, Paul Sokolovsky wrote:
Hello,

To make sure that my comment to
https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd like to
dup it with this email. The patch above has broke echo_server
TCP/IPv4 for me. When run on frdm_k64f over Ethernet, the application
answers pings, but being telneted to port 4242, it doesn't echo TCP
data. Note that the issue of "The trouble is that that linkage
between
the outgoing buffer and the connection hasn't been established at the
point data_is_sent_and_acked() is called" has been fixed some time
ago
by https://gerrit.zephyrproject.org/r/#/c/4888/ .

Reverting just that commit fixes issue for me.


Rohit Grover
 

Paul,

Can your issue be reproduced by the vanilla echo_server sample?

rohit

-----Original Message-----
> From: Jukka Rissanen [mailto:jukka.rissanen(a)linux.intel.com]
> Sent: 12 October 2016 13:40
> To: Paul Sokolovsky; devel(a)lists.zephyrproject.org; Rohit Grover
> Subject: Re: TCP/IPv4/uIP stack appears to be broken with recent commit
>
> Hi Paul,
>
> we can revert https://gerrit.zephyrproject.org/r/#/c/4282/6 if needed.
> The patch fixed an issue Rohit had thou. Best would be to find a solution
> that fixes both problems.
>
> Cheers,
> Jukka
>
>
> On Tue, 2016-10-11 at 20:22 +0300, Paul Sokolovsky wrote:
> > Hello,
> >
> > To make sure that my comment to
> > https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd like to
> > dup it with this email. The patch above has broke echo_server
> > TCP/IPv4 for me. When run on frdm_k64f over Ethernet, the application
> > answers pings, but being telneted to port 4242, it doesn't echo TCP
> > data. Note that the issue of "The trouble is that that linkage between
> > the outgoing buffer and the connection hasn't been established at the
> > point data_is_sent_and_acked() is called" has been fixed some time ago
> > by https://gerrit.zephyrproject.org/r/#/c/4888/ .
> >
> > Reverting just that commit fixes issue for me.
> >

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.


Paul Sokolovsky
 

Hello Rohit,

On Wed, 12 Oct 2016 12:57:26 +0000
Rohit Grover <Rohit.Grover(a)arm.com> wrote:

Paul,

Can your issue be reproduced by the vanilla echo_server sample?
I'm not sure what you mean by "vanilla". Default configuration of
echo_server doesn't have IPv4 or TCP enabled, and one needs to enable
those to see the issue. There're no changes to echo_server beyond
that. Otherwise, the issue is reproducible with QEMU SLIP driver (setup
described in
https://wiki.zephyrproject.org/index.php?title=Networking-with-Qemu) (my
original test was with FDRM-K64F Ethernet driver). After "git revert
627feb92d4b01556891e232db6b9c9181899965b", QEMU SLIP starts to work,
just as FDRM-K64F Ethernet before.


rohit

> -----Original Message-----
> From: Jukka Rissanen [mailto:jukka.rissanen(a)linux.intel.com]
> Sent: 12 October 2016 13:40
> To: Paul Sokolovsky; devel(a)lists.zephyrproject.org; Rohit Grover
> Subject: Re: TCP/IPv4/uIP stack appears to be broken with recent
> commit
>
> Hi Paul,
>
> we can revert https://gerrit.zephyrproject.org/r/#/c/4282/6 if
> needed. The patch fixed an issue Rohit had thou. Best would be to
> find a solution that fixes both problems.
>
> Cheers,
> Jukka
>
>
> On Tue, 2016-10-11 at 20:22 +0300, Paul Sokolovsky wrote:
> > Hello,
> >
> > To make sure that my comment to
> > https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd
> > like to dup it with this email. The patch above has broke
> > echo_server TCP/IPv4 for me. When run on frdm_k64f over
> > Ethernet, the application answers pings, but being telneted to
> > port 4242, it doesn't echo TCP data. Note that the issue of
> > "The trouble is that that linkage between the outgoing buffer
> > and the connection hasn't been established at the point
> > data_is_sent_and_acked() is called" has been fixed some time
> > ago by https://gerrit.zephyrproject.org/r/#/c/4888/ .
> >
> > Reverting just that commit fixes issue for me.
> >

IMPORTANT NOTICE: The contents of this email and any attachments are
confidential and may also be privileged. If you are not the intended
recipient, please notify the sender immediately and do not disclose
the contents to any other person, use it for any purpose, or store or
copy the information in any medium. Thank you.


--
Best Regards,
Paul

Linaro.org | Open source software for ARM SoCs
Follow Linaro: http://www.facebook.com/pages/Linaro
http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog


Paul Sokolovsky
 

Hello Jukka,

On Wed, 12 Oct 2016 15:39:44 +0300
Jukka Rissanen <jukka.rissanen(a)linux.intel.com> wrote:

Hi Paul,

we can revert https://gerrit.zephyrproject.org/r/#/c/4282/6 if needed.
The patch fixed an issue Rohit had though. Best would be to find a
solution that fixes both problems.
I'm aware of issues Rohit experienced, as I literally followed on his
footsteps to re-discover them. What weren't clear were root causes of
different issues and the best way to fix them, so there were
alternative and partial solutions proposed. In particular, he produced
https://gerrit.zephyrproject.org/r/#/c/4282, while I, later, - an
alternative solution https://gerrit.zephyrproject.org/r/#/c/4888/ ,
which happened to be merged first. I don't think I heard from Rohit
whether my patch fixed his issue (even though I tried to keep him in
loop both as a reviewer a via email updates), and I guess now it would
be really helpful to hear if
https://gerrit.zephyrproject.org/r/#/c/4282 makes any difference for
him, with #4888 merged couple of weeks before it.

#4282 clearly leads to regressions though (on its own, in interaction
with #4888, or some other patch which was merged since #4282 was
initially submitted).


Cheers,
Jukka


On Tue, 2016-10-11 at 20:22 +0300, Paul Sokolovsky wrote:
Hello,

To make sure that my comment to
https://gerrit.zephyrproject.org/r/#/c/4282/6 isn't lost, I'd like
to dup it with this email. The patch above has broke echo_server
TCP/IPv4 for me. When run on frdm_k64f over Ethernet, the
application answers pings, but being telneted to port 4242, it
doesn't echo TCP data. Note that the issue of "The trouble is that
that linkage between
the outgoing buffer and the connection hasn't been established at
the point data_is_sent_and_acked() is called" has been fixed some
time ago
by https://gerrit.zephyrproject.org/r/#/c/4888/ .

Reverting just that commit fixes issue for me.


--
Best Regards,
Paul

Linaro.org | Open source software for ARM SoCs
Follow Linaro: http://www.facebook.com/pages/Linaro
http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog