netdev development stats for 6.3

Hi!

Here are the stats for the last release cycle: based on 15,383 emails netdev has seen since Dec 15th, over 68 days (226 msgs / day).

Roughly 55% of the patches merged had a review tag (4% of which were exclusively from the same email domain as the author). This means an increase of 7% compared to the previous cycle (from 48%).

When comparing the stats note the fluctuation in total number of messages (6.2 cycle had roughly 3k more messages).

Top 10 reviewers (thr): Top 10 reviewers (msg):

   1. [276] Jakub Kicinski              1. [453] Jakub Kicinski 
   2. [118] Simon Horman                2. [234] Andrew Lunn
   3. [105] Leon Romanovsky             3. [232] Simon Horman
   4. [104] Andrew Lunn                 4. [184] Leon Romanovsky
   5. [ 88] Paolo Abeni                 5. [162] Jiri Pirko
   6. [ 62] Eric Dumazet                6. [125] Krzysztof Kozlowski
   7. [ 61] Jiri Pirko                  7. [121] Paolo Abeni
   8. [ 51] Krzysztof Kozlowski         8. [105] Eric Dumazet
   9. [ 43] Kalle Valo                  9. [102] Vladimir Oltean

10. [ 42] Alexander Lobakin 10. [ 91] Michael S. Tsirkin

Simon (and Olek) were noticeably more active in this release cycle. Russell has dropped out (he mentioned battling COVID).

Top 10 authors (thr): Top 10 authors (msg):

   1. [ 44] Jakub Kicinski              1. [283] Vladimir Oltean
   2. [ 41] Vladimir Oltean             2. [211] Saeed Mahameed
   3. [ 35] Tony Nguyen                 3. [184] Oleksij Rempel
   4. [ 33] Lorenzo Bianconi            4. [172] Jakub Kicinski
   5. [ 27] Stephen Rothwell            5. [145] Tony Nguyen
   6. [ 27] Eric Dumazet                6. [127] Lorenzo Bianconi
   7. [ 24] Arnd Bergmann               7. [126] Daniel Golle
   8. [ 21] Horatiu Vultur              8. [114] Jiri Pirko
   9. [ 18] Saeed Mahameed              9. [104] Aurelien Aptel

10. [ 18] Jiri Pirko 10. [100] Paul Blakey

Vladimir claims the top author spot continuing work on TSN (taprio, mqprio, MAC Merge uAPI and driver support).

Stephen Rothwell takes spot number 5, meaning we must have had 27 merge conflicts, lets try to have fewer conflicts going forward...

As for "scores" which are supposed to point out who doesn't help with reviewing (10 * reviews - 3 * authorship):

Top 10 scores (positive): Top 10 scores (negative):

   1. [3446] Jakub Kicinski             1. [140] Lorenzo Bianconi
   2. [1590] Simon Horman               2. [ 91] Paul Blakey
   3. [1482] Andrew Lunn                3. [ 87] Oleksij Rempel
   4. [1336] Leon Romanovsky            4. [ 84] Stephen Rothwell
   5. [1069] Paolo Abeni                5. [ 83] Daniel Golle
   6. [ 821] Jiri Pirko                 6. [ 75] Alejandro Lucero
   7. [ 751] Krzysztof Kozlowski        7. [ 71] Yoshihiro Shimoda
   8. [ 720] Eric Dumazet               8. [ 66] Aurelien Aptel
   9. [ 559] Russell King               9. [ 60] Jason Xing

10. [ 552] Alexander Lobakin 10. [ 58] Piergiorgio Beruto

Lorenzo, Paul and Oleksij author much more than they review :) Although in all fairness most folks on the list are there due to reving the same series rather than generating a lot of code.

Interestingly the flow of semi-automated trivial patches seem to have slowed down significantly (authors of those previously occupied most of the top 5 spots).

As for the corporate stats:

Top 7 reviewers (thr): Top 7 reviewers (msg):

   1. [340] Meta                        1. [623] Meta
   2. [210] Intel                       2. [463] Intel
   3. [191] nVidia                      3. [439] RedHat
   4. [178] RedHat                      4. [437] nVidia
   5. [125] Google                      5. [238] Corigine
   6. [120] Corigine                    6. [234] Andrew Lunn
   7. [104] Andrew Lunn                 7. [225] Google

Intel takes the #2 spot from nVidia. Corigine makes an appearance thanks to Simon's work.

Top 7 authors (thr): Top 7 authors (msg):

   1. [142] RedHat                      1. [923] nVidia
   2. [137] nVidia                      2. [445] RedHat
   3. [128] Intel                       3. [406] Intel 
   4. [ 77] Google                      4. [324] NXP
   5. [ 60] Meta                        5. [271] Microchip
   6. [ 58] NXP                         6. [268] Pengutronix
   7. [ 54] Microchip                   7. [213] Meta

Top 12 scores (positive): Top scores (negative):

   1. [4358] Meta                       1. [127] Marvell
   2. [2437] Intel                      2. [120] Alibaba
   3. [2008] RedHat                     ...
   4. [1910] nVidia                     5. [ 71] Bootlin
   5. [1622] Corigine                   ...
   6. [1482] Andrew Lunn                8. [ 52] Renesas
   7. [1371] Google                     9. [ 51] ZTE
   8. [ 916] Linaro                     ... one-time contributors follow

9. [ 741] Oracle
10. [ 504] Kalle Valo
11. [ 498] Rob Herring
12. [ 489] Broadcom

Positive scores mostly match the review ranking (with RedHat jumping to #3, given nVidia's large patch volume). As mentioned bots have mostly disappeared in this release cycle, Marvell and Alibaba are the top "net negative" contributors in terms of reviews. Microchip has dropped off the negative list.

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