[cgl_discussion] TIPC on 2.5.62

Jon Maloy jon.maloy at ericsson.com
Wed Feb 19 16:21:09 PST 2003


It is quite ok that you change the directory structure to fit 
within the Linux, but I am not happy with the fact that
it blurs the distinction between core code/includes and
other code. I think Mika's core directory was a good idea, 
and I would even like to se a corresponding  inlude/core
for the three 'core' inlude files which are now under 
TipcAdaptation_SWI/incl. (We could even turn it around
and put the core code at the top level and the adaptation
code in subdirectoties, but at first glance I prefer Mika's 
structure)

Since most people read documentation only as last resort 
we can not rely on that to make them understand or respect 
this distinction. I think the only way is to clearly show
it by putting the files into different directories.

For me it is very important to keep the core code portable, since
we already have this code running on three other OS:es, and there
will probably be more soon.

Regards /Jon


Mark Haverkamp wrote:


On Tue, 2003-02-18 at 15:18, Mika Kukkonen wrote:

  

I did some low down dirty hacking on tipc-0.92 (still available at

http://tipc.sourceforge.net/tipc-0.92.tar.gz
<http://tipc.sourceforge.net/tipc-0.92.tar.gz> ), on a single purpose to

make tipc compile against 2.5 kernels.



I was succesfull (the compressed diff and diffstat are in 

http://www.osdl.org/archive/mika/ <http://www.osdl.org/archive/mika/> ),
but:

  a) it is not even a module currently

  b) I just made it to compile; I am sure for example that the

     socket_api.c is badly broken.

  c) I have not even booted this kernel, as like I said, my purpose

     was just to get it to compile.

So you have been warned; try to boot this on your own risk.



I'll be mucking around it some more to remove more of C++-isms out of it

and make other changes so that it could have some chance of even getting

kernel people look into it. 



And by the way, I renamed most of non-core files, and run Lindent on all

files, so it does not make much sense to try to back port this to

tipc-0.93. 



--MiKu

    





I moved the code in the core directory up to the tipc directory and

changed the Makefile to get it to build as a module.

The module loaded, but with a debug/Call trace:



Feb 19 10:29:29 markh2 kernel: -----> TIPC: Compiled at Feb 19 2003
10:11:30 starting...

Feb 19 10:29:29 markh2 kernel: -----> TIPC: Own Address = <1.1.1>

Feb 19 10:29:29 markh2 kernel: -----> TIPC: Bearer type <ethernet>
registered

Feb 19 10:29:29 markh2 kernel: -----> TIPC/UDP: Using system id 0 

Feb 19 10:29:29 markh2 kernel: -----> TIPC: Bearer type <udp> registered

Feb 19 10:29:29 markh2 kernel: Debug: sleeping function called from
illegal context at net/ipv4/af_inet.c:516

Feb 19 10:29:29 markh2 kernel: Call Trace:

Feb 19 10:29:29 markh2 kernel:  [<c03cc050>] inet_bind+0xf0/0x310

Feb 19 10:29:29 markh2 kernel:  [<d89ba80e>] bearer_instances+0x2e/0x460
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893a927>] enableBearer+0x107/0x3d0
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89ba0e4>] bearer+0x4/0x640 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8940475>]
debug_consolePrint+0x45/0x80 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89bb125>] stripped_name+0x5/0x40
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8940186>]
bearer_extractType+0x96/0xd0 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89ba80e>] bearer_instances+0x2e/0x460
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893f70b>]
tipc_enableBearerInstance+0x15b/0x2b0 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89ba7e0>] bearer_instances+0x0/0x460
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89bad34>] bearer_types+0xf4/0x4a0
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89bad34>] bearer_types+0xf4/0x4a0
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89ba6c0>] bearer+0x5e0/0x640 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89ba7e0>] bearer_instances+0x0/0x460
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89bacd4>] bearer_types+0x94/0x4a0
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d895b37f>] +0xb5/0xdd6 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893ad35>]
udpAdaptationStart+0x135/0x150 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d895b36f>] +0xa5/0xdd6 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8961f80>] +0x0/0x20 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893a820>] enableBearer+0x0/0x3d0
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893abf0>] disableBearer+0x0/0x10
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893a3f0>] dataReq+0x0/0x60 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893a800>] prepareLink+0x0/0x20 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d893a640>] sendConfigMsg+0x0/0x110
[tipc]

Feb 19 10:29:29 markh2 kernel:  [<d895b37f>] +0xb5/0xdd6 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d894a987>] mem_alloc+0x17/0x70 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89488c0>]
manager_error_event+0x0/0x10 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89488d0>]
manager_connected_error_event+0x0/0x10 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8949ca0>]
manager_named_message_event+0x0/0x890 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8949c10>]
manager_connection_message_event+0x0/0x90 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d89d00eb>] +0xeb/0x151 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<d8961420>] +0x0/0x4e0 [tipc]

Feb 19 10:29:29 markh2 kernel:  [<c0134e33>] sys_init_module+0x153/0x200

Feb 19 10:29:29 markh2 kernel:  [<c01098c7>] syscall_call+0x7/0xb

Feb 19 10:29:29 markh2 kernel: 

Feb 19 10:29:29 markh2 kernel: -----> TIPC/UDP: Unicast listener socket
on port 55555

Feb 19 10:29:29 markh2 kernel: -----> TIPC: Bearer instance <udp:udp>
enabled

Feb 19 10:29:29 markh2 kernel: -----> TIPC: Socket adaptation started. 





  


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.linux-foundation.org/pipermail/cgl_discussion/attachments/20030219/200bb3e7/attachment-0001.htm


More information about the cgl_discussion mailing list