Re: [pim] Last Call Expired: <draft-ietf-pim-assert-packing-08.txt>

[Toerless Eckert <[e-mail]>]

Thanks, Alvaro

Will definitely change the Reserved to 24 bits.

Would it help to avoid a lot of duplicate feedback on this issue if
i was doing another quick rev for this right now, or
does this just confuse the IESG review process ?

Btw:

I had actually overlooked stuff in rfc7761 and was under the misguided
impression that there was only 16 bit alignment in some fields.

So i just went back to investigate a bit more if we should still consider
"native" alignment up to 32 bits (aka: 32 bit address like IPv4 address
starts at 32 bit boundary) to be important enough, and i could not
find sufficient evidence to not be concerned about it anymore.
Interestingly enough it's less the CPU/instruction set capabilities to
do misaligned access - that has improved a lot since e.g.: 2010. But
it seems to actually be ongoing C language level expectations to
have "native" alignments.

Cheers
    Toerless

On Thu, Mar 09, 2023 at 03:25:14AM -0800, Alvaro Retana wrote:
> On March 8, 2023 at 9:03:06 PM, Toerless Eckert wrote:
> 
> Toerless:
> 
> Hi!
> 
> 
> > Hope this is ready for you to go to IESG!
> 
> I have issued by ballot and it should be placed in next week's IESG
> Telechat.  This should give us enough time to address any comments by
> IETF 116.
> 
> 
> ...
> > - I did remove "Count" from message headers and added a "Zero" field
> > to resolve the discuss with Stigs.
> 
> The only issue I have with this is that the packet format now looks as
> if the first Assert Record has a length of multiple of 32 + 16 bits --
> but they are are all 32-bit aligned.  The fix seems to be to make the
> Reserved field 24 bits long.
> 
> 
> Thanks!
> 
> Alvaro.

-- 
---
[e-mail]