Redboot networking problem

[Ilko Iliev-2]

Hi,

I have a problem when I perform two http loads one after other (in
Redboot) - the second load put the RedBoot in an endless loop.

If there is a small delay between the both loads, then sometimes it works.
If the the first load is a http and the second load a tftp, then the
problem doesn't exist.

The endless loop is in tcp.c, __tcp_handler():
    for (prev = NULL, s = tcp_list; s; prev = s, s = s->next) {
        if (s->our_port == ntohs(tcp->dest_port)) {
        if (s->his_port == 0)

It loops forever because *s == s->next* and *s->our_port* and
*ntohs(tcp->dest_port)* differs with one.

Can somebody try this situation?

--
With best regards,
Ilko Iliev
Ronetix Development Tools GmbH
CPU Modules, JTAG/BDM Emulators and Flash Programmers
Waidhausenstrasse 13/5, 1140 Vienna, Austria
E-Mail: iliev@...; Web: www.ronetix.at


--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[John Dallaway-2]

Hi Ilko

RONETIX - Ilko Iliev wrote:

> I have a problem when I perform two http loads one after other (in
> Redboot) - the second load put the RedBoot in an endless loop.
>
> If there is a small delay between the both loads, then sometimes it works.
> If the the first load is a http and the second load a tftp, then the
> problem doesn't exist.
>
> The endless loop is in tcp.c, __tcp_handler():
>    for (prev = NULL, s = tcp_list; s; prev = s, s = s->next) {
>        if (s->our_port == ntohs(tcp->dest_port)) {
>        if (s->his_port == 0)
>
> It loops forever because *s == s->next* and *s->our_port* and
> *ntohs(tcp->dest_port)* differs with one.
>
> Can somebody try this situation?

Thank you for the report. This does sound like a bug. Are you using
recent sources (such as eCos 3.0)? If not, could you verify that this
issue exists with recent sources please? What type of file are you
loading (ELF, SREC, raw) and how long is it? Which processor and
ethernet part are you using?

The best way to raise formal bug reports is via our Bugzilla system:

  http://bugs.ecos.sourceware.org

This helps to keep all information in one place and to ensure prompt
investigation.

John Dallaway

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[Ilko Iliev-2]

Hello John,

John Dallaway wrote:

> Hi Ilko
>
> RONETIX - Ilko Iliev wrote:
>
>  
>> I have a problem when I perform two http loads one after other (in
>> Redboot) - the second load put the RedBoot in an endless loop.
>>
>> If there is a small delay between the both loads, then sometimes it works.
>> If the the first load is a http and the second load a tftp, then the
>> problem doesn't exist.
>>
>> The endless loop is in tcp.c, __tcp_handler():
>>    for (prev = NULL, s = tcp_list; s; prev = s, s = s->next) {
>>        if (s->our_port == ntohs(tcp->dest_port)) {
>>        if (s->his_port == 0)
>>
>> It loops forever because *s == s->next* and *s->our_port* and
>> *ntohs(tcp->dest_port)* differs with one.
>>
>> Can somebody try this situation?
>>    
>
> Thank you for the report. This does sound like a bug. Are you using
> recent sources (such as eCos 3.0)? If not, could you verify that this
> issue exists with recent sources please?

I tested with the latest eCos/Redboot.

> What type of file are you
> loading (ELF, SREC, raw) and how long is it?
Raw format, the file length doesn't matter, I tested with small (200
bytes) and large files (1.5MB).

> Which processor and
> ethernet part are you using?
>  
ARM7, S3C4530A (KS32C5000).

> The best way to raise formal bug reports is via our Bugzilla system:
>
>   http://bugs.ecos.sourceware.org
>
> This helps to keep all information in one place and to ensure prompt
> investigation.
>  
I will report the problem.

--
With best regards,
Ilko Iliev
Ronetix Development Tools GmbH
CPU Modules, JTAG/BDM Emulators and Flash Programmers
Waidhausenstrasse 13/5, 1140 Vienna, Austria
E-Mail: iliev@...; Web: www.ronetix.at


--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[rd.asha]

Hi,

I am finding same issue on iMX27 processor and am using LTIB Release 3. I found that it is not dependent on size of file being downloaded.
This issue is observed when http load is invoked twice in source code flow. If http load is invoked from command prompt twice/thrice, no such issue is seen.

Please help in fixing this issue.

Regards,
Asha R

Ilko Iliev-2 wrote:

Hi,

I have a problem when I perform two http loads one after other (in
Redboot) - the second load put the RedBoot in an endless loop.

If there is a small delay between the both loads, then sometimes it works.
If the the first load is a http and the second load a tftp, then the
problem doesn't exist.

The endless loop is in tcp.c, __tcp_handler():
    for (prev = NULL, s = tcp_list; s; prev = s, s = s->next) {
        if (s->our_port == ntohs(tcp->dest_port)) {
        if (s->his_port == 0)

It loops forever because *s == s->next* and *s->our_port* and
*ntohs(tcp->dest_port)* differs with one.

Can somebody try this situation?

--
With best regards,
Ilko Iliev
Ronetix Development Tools GmbH
CPU Modules, JTAG/BDM Emulators and Flash Programmers
Waidhausenstrasse 13/5, 1140 Vienna, Austria
E-Mail: [e-mail]; Web: www.ronetix.at


--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[Andrew Lunn-2]

On Tue, Apr 28, 2009 at 03:11:05AM -0700, rd.asha wrote:
>
> Hi,
>
> I am finding same issue on iMX27 processor and am using LTIB Release 3. I
> found that it is not dependent on size of file being downloaded.
> This issue is observed when http load is invoked twice in source code flow.
> If http load is invoked from command prompt twice/thrice, no such issue is
> seen.

It should be fixed in the eCos trunk.

   Andrew

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[rd.asha]

Hi Andrew,
Correct me if am wrong. Is this already fixed in ecos 3.0?
I verified the same by taking tcp,http releated files to my code base.
The same behaviour is observerd.

Regards,
Asha

Andrew Lunn-2 wrote:

On Tue, Apr 28, 2009 at 03:11:05AM -0700, rd.asha wrote:
>
> Hi,
>
> I am finding same issue on iMX27 processor and am using LTIB Release 3. I
> found that it is not dependent on size of file being downloaded.
> This issue is observed when http load is invoked twice in source code flow.
> If http load is invoked from command prompt twice/thrice, no such issue is
> seen.

It should be fixed in the eCos trunk.

   Andrew

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[Andrew Lunn-2]

On Wed, Apr 29, 2009 at 11:10:17PM -0700, rd.asha wrote:
>
> Hi Andrew,
> Correct me if am wrong. Is this already fixed in ecos 3.0?
> I verified the same by taking tcp,http releated files to my code base.
> The same behaviour is observerd.

Please don't top post.

>
> Regards,
> Asha
>
>
> Andrew Lunn-2 wrote:
> >
> > On Tue, Apr 28, 2009 at 03:11:05AM -0700, rd.asha wrote:
> >>
> >> Hi,
> >>
> >> I am finding same issue on iMX27 processor and am using LTIB Release 3. I
> >> found that it is not dependent on size of file being downloaded.
> >> This issue is observed when http load is invoked twice in source code
> >> flow.
> >> If http load is invoked from command prompt twice/thrice, no such issue
> >> is
> >> seen.
> >
> > It should be fixed in the eCos trunk.

                                   ^^^^^^

        Andrew

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss