netdev development stats for 6.7

Hi!

General stats
-------------

The cycle started on August 29th and ended on Oct 31st, 1 day shorter than previous one, due to the timing of our PR.

We have seen total of 14243 messages on the list (226 / day) which is 17% lower than the (very busy) 6.6 cycle. The number of commits directly applied by netdev maintainers dropped by 13% to 19 commits a day, close to our long term average.

There was a fluctuation in the number of participants. While the number of people "reviewing" (replying in threads) remained constant (around
410) the number of people exclusively starting threads ("authors"?) decreased by 40 (344 -> 306).

Fraction of changes with Review/Ack tags has dropped again to 58% (counting all tags) and 50% (counting tags from different company than the author).

Rankings
--------

Top reviewers (thr): Top reviewers (msg):

   1 (   ) [24] Simon Horman            1 ( +1) [43] Jakub Kicinski      
   2 (   ) [23] Jakub Kicinski          2 ( -1) [33] Simon Horman        
   3 (   ) [12] Andrew Lunn             3 (   ) [33] Andrew Lunn         
   4 ( +1) [10] Paolo Abeni             4 ( +5) [15] Eric Dumazet        
   5 ( +1) [ 8] Eric Dumazet            5 ( +1) [14] David Ahern         
   6 ( +1) [ 7] David Ahern             6 ( +2) [13] Paolo Abeni         
   7 (+21) [ 7] Kees Cook               7 (+21) [13] Florian Fainelli    
   8 (+12) [ 6] Jiri Pirko              8 ( +5) [12] Jiri Pirko          
   9 (+14) [ 6] Florian Fainelli        9 (+23) [12] Jacob Keller        
  10 (+19) [ 6] Jacob Keller           10 ( +8) [10] Vladimir Oltean     
  11 ( -2) [ 4] Russell King           11 (+24) [ 9] Kees Cook           
  12 ( -2) [ 4] Willem de Bruijn       12 ( +5) [ 8] Rob Herring         
  13 ( +2) [ 4] Vladimir Oltean        13 ( -8) [ 8] Russell King        
  14 ( -1) [ 4] Rob Herring            14 ( +1) [ 8] Jason Wang          
  15 (***) [ 3] Wojciech Drewek        15 ( -4) [ 7] Willem de Bruijn    

No surprises in the top 6. Kees takes #7, helping to review string op safety and __conted_by patches. Wojciech enters the ranking at #15. Thanks for all the review work, folks!

Top authors (thr): Top authors (msg):

   1 (***) [7] Justin Stitt             1 ( +3) [18] Eric Dumazet        
   2 ( +1) [6] Eric Dumazet             2 (***) [16] David Howells       
   3 ( -1) [6] Jakub Kicinski           3 ( +6) [16] Dmitry Safonov      
   4 (+10) [3] Jiri Pirko               4 ( -2) [14] Saeed Mahameed      
   5 ( -1) [2] Tony Nguyen              5 ( +9) [14] Herve Codina        
   6 (***) [2] Oleksij Rempel           6 (   ) [13] Jiri Pirko          
   7 (***) [2] Kees Cook                7 ( -4) [12] Jakub Kicinski      
   8 (***) [2] Ivan Vecera              8 (+26) [11] Aurelien Aptel      
   9 (***) [2] Dan Carpenter            9 ( -8) [11] Tony Nguyen         

10 (+15) [2] MD Danish Anwar 10 (***) [10] Uwe Kleine-König

Justin has posted the most individual patches, replacing the use of unsafe string APIs throughout the drivers. Jiri jumps into top 5 with his devlink and YNL work. David H posted a few series for iov and network file systems (somewhat netdev-adjacent). Dmitry contributed the TCP Auth Option support. Herve worked on a HDLC framer for QMC.

Top reviewers (thr): Top reviewers (msg):

   1 ( +2) [42] RedHat                  1 ( +2) [71] RedHat              
   2 (   ) [27] Meta                    2 ( -1) [52] Meta                
   3 ( +2) [23] Intel                   3 ( +2) [46] Intel               
   4 ( +2) [15] Google                  4 ( +2) [33] Andrew Lunn         
   5 ( -1) [12] nVidia                  5 ( +2) [29] Google              
   6 ( +1) [12] Andrew Lunn             6 ( -2) [23] nVidia              
   7 ( +3) [ 7] Enfabrica               7 ( +4) [14] Broadcom            

The biggest change in the company statistics is the disappearance of Corigine. Simon is now employed at Red Hat, giving Red Hat the #1 spot with quite some margin.

With Corigine dropping out, Enfabrica (David Ahern) and Broadcom (Florian Fainelli) ascend to the top #7.

Top authors (thr): Top authors (msg):

   1 ( +5) [22] Google                  1 (   ) [76] Intel               
   2 (   ) [19] Intel                   2 (   ) [59] nVidia              
   3 (   ) [17] RedHat                  3 (   ) [55] RedHat              
   4 (   ) [10] Meta                    4 ( +2) [50] Google              
   5 (   ) [ 9] nVidia                  5 (   ) [38] Meta                
   6 ( -5) [ 6] Huawei                  6 (+10) [26] Bootlin             
   7 ( +6) [ 5] Linaro                  7 ( -3) [23] Huawei       

Top scores (positive): Top scores (negative):

   1 ( +3) [341] RedHat                 1 (+13) [97] Bootlin             
   2 (   ) [219] Meta                   2 (***) [72] nVidia              
   3 (   ) [183] Andrew Lunn            3 ( -2) [66] Huawei              
   4 ( +2) [ 95] Enfabrica              4 ( +4) [59] Arista              
   5 ( +4) [ 59] Broadcom               5 (+10) [48] Alibaba             
   6 (+17) [ 52] Isovalent              6 (***) [41] Pengutronix         

7 ( +3) [ 47] ARM
8 ( +5) [ 46] Oracle
9 ( -1) [ 44] Linux Foundation
10 ( -5) [ 32] Linaro

A few things worth noting in the "community score" metrics.

Intel moved from the "negative" to the "positive" side (at #13, so not high enough to make the "top"). Shout out to Jake, Wojciech and Przemek for their review work! This move may have been helped slightly by the lower volume of Intel patches and external contributions to Intel drivers. So please do not rest on your laurels :)

nVidia makes the opposite switch and ironically takes negative spot #2, the exact spot previously occupied by Intel. Jiri's efforts are not enough to counter balance the flow of patches there :(

Arista is likely a blip as Dmitry had to repost his work a few times.

Bootlin and Pengutronix return to the same (negative) positions they held in 6.5 cycles. It may be the time to carve out more review time for folks working at those companies.

Code: https://github.com/kuba-moo/ml-stat

Hi Jakub,

I just noticed this stats report from Simon. Thanks for your work and sharing. I want to know if there is a way to bind my personal email with my company so my review could increase my company's score :) I think there are some other company developers who use their own email also curious.

Thanks and Best Regards
Hangbin

Jonathan Corbet <[email protected]> maintains a list of email addresses to organisations mapping. Let him know who you work for and the next cycle you should count to your company.

Andrew

Thanks for asking, the company association is my least favorite part of this work, and you gave me an excuse to rant about it :)

Yes, unfortunately I do not have access to that list. The LWN list was compiled with some assurances of the list not being shared. Jon / Greg understandably did not want to send me the list and break that promise.

So telling Jon won't help me.

Now, CNCF has a similar setup: https://github.com/cncf/gitdm
and they do share their database. So I use that, plus my local hacky mapping. Unfortunately the CNCF DB is not very up to date for kernel folks.

Hangbin, according to CNCF you're at Red Hat, which seems sane, and that's how I count you :)

Now the rant.

Unfortunately I can't handle creating a company/developer DB myself, because of GDPR etc. I work for a company which takes personal information very, very seriously.

I brought creating a public DB up at Linux Foundation TAB meetings, but after some poking there's no movement.

To add insult to injury, if you watch past the end of the recent (excellent) talk from Jon - https://lwn.net/Articles/949647/ you will see Jim Zemlin pop up on the stage after Jon finishes, and you will hear him tout the great "analytics tools" that Linux Foundation has been working. I think he's talking about this:
https://insights.lfx.linuxfoundation.org/projects/korg/dashboard

It would be great if Linux Foundation helped the community with the developer/company DB, which is ACTUALLY USEFUL BEFORE WASTING TIME ON SOME WEB STUFF THAT DOESN'T WORK FOR THE KERNEL.

This makes me so angry.

Hi Jakub,

Thanks for this info. Glad to know my email and company are mapping correctly.

I personally agree Linux Foundation should maintain a developer/company DB. The developer could submit their information on a voluntary basis, instead of letting some tools search the website and collect data.

Thanks
Hangbin