RTEMS 4.9.5 vs 4.10.2 Network Stack

[Joel Sherrill]

The new TCP/IP stack port isn't ready for you to actually test on your target hardware but it should compile fine. If that is enough to check if these structure packing and alignment issues exist in the new code base, please check.

The port is far enough along to get help completing it. Please

--joel

SAeeD <salpha.2004 at gmail.com> wrote:

>As I tested the new patch with real board, it is OK with TCP client/server establishment and is OK with UDP server establishment.
>But after a call to "sendto" function (defined in cpukit/libnetworking/rtems/rtems_syscall.c) in UDP client in order to send some data to server, that familiar data abort problem with "INSN_LDR" happens again. Excluding this one, seems that all other network functionalities are fine.
>Thanks Bacon for your new patch. :)
>
>On Sun, Sep 2, 2012 at 4:57 PM, Pierre Ficheux <pierre.ficheux at openwide.fr<mailto:pierre.ficheux at openwide.fr>> wrote:
>Le 02/09/12 13:06, Bacon Xu a écrit :
>> Hi Pierre & Saeed,  I have reviewed my patch of RTEMS 4.10.2. It's
>> indeed that I add ONE (__attribute__((packed))) in patch. It's not
>> enough for arm platform. It must add all of places were mentioned in
>> previous mail with __attribute__((packed)).
>>
>> I found those problems in RTEMS 4.9 for porting ARM at9260. When I add
>> those __attribute__((packed)), the version work properly. So in RTEMS
>> 4.10, I add them too. Because of some specific problems,  the patch
>> haven't included those modifications.
>>
>> Attachment is the newest patch that I have test it several times. I
>> haven't real mini2440 board, so I can't ensure some important things,
>> such as s3c2440's pll, and other drivers. It just work smoothly in
>> QEMU.
>>
>
>I will test it on real board tomorrow.
>
>Thanks for it :)
>
>regards
>
>--
>Pierre FICHEUX -/- CTO OW/OWI, France -\- pierre.ficheux at openwide.fr<mailto:pierre.ficheux at openwide.fr>
>                                         http://ingenierie.openwide.fr
>                                         http://www.ficheux.org
>I would love to change the world, but they won't give me the source code
>
>