[Click] Alignment issue

Eddie Kohler ekohler at gmail.com
Mon Jan 9 13:19:27 EST 2012


Hi Markku, Cliff,

On reflection, I think the right answer here is to degrade the 
Classifier error to a warning. I did this, and added more context to 
describe the different ways to fix the error.

Eddie


On 11/18/11 11:01 AM, Cliff Frey wrote:
> if you modify click-align.cc to specify that the FromOcteon element
> produces packets that are (4,2) aligned, then you won't get the extra
> Align() element.
>
> If you really want to do what you are asking for, you could likely add
> -DHAVE_INDIFFERENT_ALIGNMENT to CXXFLAGS/CFLAGS, but I'm not positive if
> that would work.  I'd definitely recommend just keeping click-align.cc up
> to date with your elements.
>
> Cliff
>
> On Fri, Nov 18, 2011 at 12:45 AM, Markku Savela<Markku.Savela at vtt.fi>wrote:
>
>> When configuring for mips 64, it does not set the
>>
>>   HAVE_INDIFFERENT_ALIGNMENT
>>
>> in include/click/config.h.
>>
>> This is fine, but it does trouble me that click doesn't
>> even run the configuration without it being passed through
>> the click-align tool.
>>
>> I would like to have an option to run the unaligned
>> configuration and take my risks on "bus error". Maybe
>> I missed a command line switch to disable this check?
>>
>> In my test case, non-aligned click configuration works
>> just fine (the packets just happen to be aligned), but
>> running conf through click align, adds alingning and
>> causes unnecessary extra copy of the packet.
>>
>> (see attached confs).
>>
>> _______________________________________________
>> click mailing list
>> click at amsterdam.lcs.mit.edu
>> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
>>
>>
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click


More information about the click mailing list