
|
[openbeosnetteam]
||
[Date Prev]
[04-2004 Date Index]
[Date Next]
||
[Thread Prev]
[04-2004 Thread Index]
[Thread Next]
[openbeosnetteam] Re: route show, problem found
- From: "Brennan Cleveland" <warriorspot@xxxxxxxxxxxxx>
- To: openbeosnetteam@xxxxxxxxxxxxx
- Date: Mon, 12 Apr 2004 09:11:21 -0500
Phillipe,
Ok. So you're guessing the problem is simply manifested in the
loops, but the loops themselves should be ok. I will hunt around and
see if I can find the root of the issue... The weird thing is that
the output I get from "route show" using my exit condition is correct.
Brennan
>> > Here is the problem with the 'route show' command:
>> >
>> > Endless loops are encountered in the file route.c in the function
>
>> > char
>> > *netname(sa). They occur in the two while() loops (approx. lines
>> > 524 and 534). The value of 'mask' gets shifted to 0, and the
>exit
>> > conditions of these loops are never met. This is the case on my
>> > 192.168.0.0 network, I can't vouch for behavior of any other
>> > subnets. A
>> > 'fix' that worked temporarily is to change the exit conditons of
>> > both
>> > loops to be:
>> > while(mask > 0). This produced correct output for me.
>> >
>> > Can some who is more familiar with this code verify a correct fix
>
>> > for
>> > this? I don't fully understand this section of the code at this
>> > point.
>>
>> I think it tries to look if there is a subnetmask (192.168.1.x
>> 192.168.2.x, etc.). We should take a look at the FreeBSD tool.
>Probably
>they have a working function (but only this function, better we do
>not
>replace the socket communication code).
>
>Well, it's a very similar code, no suprised as our route command is a
>
>port of BSD one:
>http://www.freebsd.org/cgi/cvsweb.cgi/src/sbin/route/route.c?rev=1.71
>&
>content-type=text/x-cvsweb-markup
>
>In the first loop, we don't cast shifted mask value into a (long)
>like
>BSD but an (int). I don't how it could screw us, but, anyway...
>I guess the issue is not really in these loops but in the values of
>in
>and mask variable.
>
>> > Finally, can we get rid of (comment out by default) the printf
>> > statements in compat.c? They relly mess up the formatting of the
>
>> > output
>> > for route show and other things.. It might be a good idea if we
>> > start
>> > getting some users playing with our stack that they don't by
>> > default see
>> > this type of debugging output.
>>
>> Yes, that should be commented out.
>
>Agreed.
>
>- Philippe
>
>--
>Fortune Cookie Says:
>
>"This is a job for BOB VIOLENCE and SCUM, the INCREDIBLY STUPID
>MUTANT
>DOG."
> -- Bob Violence
>
|

|