Re: [PATCH bluetooth-next] mac802154: rename seq to sequence_number

[Alexander Aring <alex.aring@gmail.com>]

Hi,

On Thu, Jun 25, 2015 at 02:53:18PM +0530, Varka Bhadram wrote:
> Hello,
> 
> On 06/25/2015 02:10 PM, Phoebe Buckheister wrote:
> 
> >On Thu, June 25, 2015 10:23 am, Varka Bhadram wrote:
> >>Hi Phoebe Buckheister,
> >>
> >>On 06/25/2015 01:13 PM, Phoebe Buckheister wrote:
> >>>On Thu, June 25, 2015 9:29 am, Stefan Schmidt wrote:
> >>>>Hello.
> >>>>
> >>>>On 25/06/15 08:31, Varka Bhadram wrote:
> >>>>>This patch rename ieee802154_hdr member seq to sequence_number.
> >>>>Any good reason for this? I think seq is quite clear in this context
> >>>>and
> >>>>making it sequence_number has no real benefit.
> >>>>
> >>>>If others disagree I'm fine to let that one in though. No hard feelings
> >>>>about it.
> >>>Don't see the in renaming this either.
> >>I didn't get your point. ?
> >The "point" went missing. I don't see why renaming this symbol is
> >necessary or useful. Like Stefan said, it is pretty clear from context,
> >and sequence_number is kind of long too.
> >
> As part of the rework on frame parsing, Alex used this naming convention [1].
> 
> If you think that it's too long, can we use *seq_num* instead of *seq* ?
> 

Yea, shame on me. The rework dev branch was just fun for me, I was
looking for mechanism like (nl802154, cfg802154, (frame parsing?)) which
wireless/mac80211 do.

I think I deleted for that the complete actual frame parsing stuff. Then
I grab the 802.15.4 standard and simple use the naming convention from
there.

Nevertheless in the C programmers nature is to use short names. I am
fine with "seq", maybe add comments that seq means sequence_number.

I need also to remember that doing things in a dirty dev branch and
bringing things mainline are complete different parts. Bringing things
mainline will be a longer process that all people are fine with the new
solution. Maybe we should first start a new mailing thread to talking
about the new strategies and listing the pros and cons of each solution.


Another point is: when you try to bring this frame parsing stuff
mainline (which is similar like mac80211 stuff), I don't ack patches
until we have some working llsec state in nl802154 and wpan-tools.

Because the llsec module is a core part of the frame parsing stuff.
The actual state for doing the frame parsing stuff in the control block
of skb (skb->cb (mac_cb)) will occur many side effects when we changing
something in there, which I can't ack at the moment without a working
llsec layer.


The second thing is: don't make per day 1-2 patches. If you plan to do
something big, then send patch series (please 20 patches at maximum per
patch series).

- Alex