netdev development stats for 6.18
Subsystems (2)
Jakub Kicinski
maintainerIntro
-----
As is tradition here are the development statistics based on mailing list traffic on netdev@vger.
These stats are somewhat like LWN stats: https://lwn.net/Articles/1004998/
but more focused on mailing list participation. And by participation we mean reviewing code more than producing patches.
In particular "review score" tries to capture the balance between reviewing other people's code vs posting patches. It's roughly number of patches reviewed minus number of patches posted. Those who post more than they review will have a negative score.
Previous 3 reports:
- for 6.15: https://lore.kernel.org/[email protected]
- for 6.16: https://lore.kernel.org/[email protected]
- for 6.17: https://lore.kernel.org/[email protected]
General stats
-------------
The 6.18 release was very similar in size to 6.17 for the overall kernel, only -0.3% change in the number of patches queued in linux-next when the merge window opened.
For netdev 6.18 was +4% larger in terms of commits merged by core maintainers than 6.17 (which in turn was 3% larger than 6.16). We have seen -3% fewer messages on the list (previous delta was also negative: -1%).
Average number of revisions of a patchset decreased by -2%.
Testing
-------
Proportion of test commits went up by 3% (from 8% to 11% of all commits touching the selftests directory). This is mostly due to efforts of a handful of people, the number of unique contributors to selftest unfortunately went down from 44 to 41.
Contributions to selftests:
1 [ 41] Jakub Kicinski
2 [ 20] Kuniyuki Iwashima
3 [ 18] Petr Machata
4 [ 16] Matthieu Baerts
5 [ 13] Ido Schimmel
6 [ 7] Hangbin Liu 7 [ 4] Geliang Tang 8 [ 3] Stanislav Fomichev 9 [ 3] Eric Dumazet 10 [ 3] Dimitri Daskalakis
Kuniyuki added a solid chunk of packetdrill tests for TCP Fast Open. Petr cleaned up bash helpers and their use of defer constructs. Metthieu worked on MPTCP tests, and Ido on routing tests.
I should mention that the stats only cover work applied by core networking maintainers, work done by trees which send us PRs (Netfilter for instance) does not show up here.
Developer rankings
------------------
Top reviewers (cs): Top reviewers (msg):
1 ( ) [33] Jakub Kicinski 1 ( ) [70] Jakub Kicinski 2 ( ) [22] Simon Horman 2 ( +1) [46] Andrew Lunn 3 ( ) [17] Andrew Lunn 3 ( -1) [44] Simon Horman 4 ( ) [ 7] Paolo Abeni 4 ( ) [17] Paolo Abeni 5 ( ) [ 6] Eric Dumazet 5 ( ) [16] Eric Dumazet 6 ( +7) [ 6] Russell King 6 ( ) [15] Willem de Bruijn 7 ( +4) [ 5] Vadim Fedorenko 7 ( +7) [14] Russell King 8 ( -2) [ 4] Kuniyuki Iwashima 8 ( ) [12] Kuniyuki Iwashima 9 ( +5) [ 4] Aleksandr Loktionov 9 ( +2) [12] Vadim Fedorenko 10 (+15) [ 4] Paul Menzel 10 ( -3) [12] Krzysztof Kozlowski 11 ( -3) [ 3] Willem de Bruijn 11 (+14) [11] Vladimir Oltean 12 ( ) [ 3] Jacob Keller 12 (+23) [ 9] Michael S. Tsirkin 13 ( -4) [ 3] Rob Herring 13 ( +7) [ 8] Aleksandr Loktionov 14 ( -4) [ 3] Krzysztof Kozlowski 14 ( -1) [ 8] Mina Almasry 15 ( +1) [ 3] Vladimir Oltean 15 ( +2) [ 7] Rob Herring
Russell King returns to top 10 maintainers after a short period of lower activity. Reviewing various PHY / phylink and embedded networking patches, when he's not busy untangling stmmac :)
Vadim continues to review PTP but increasingly also other driver patches.
Aleksandr focuses on driver reviews, similarly to Paul Menzel who specifically reviews Intel driver patches.
The list doesn't cover the less frequent reviews, but all review work is appreciated! Thank you all!
Top authors (cs): Top authors (msg):
1 ( ) [6] Jakub Kicinski 1 ( ) [25] Jakub Kicinski 2 ( ) [5] Eric Dumazet 2 (+49) [24] Russell King 3 ( +8) [4] Russell King 3 ( ) [22] Kuniyuki Iwashima 4 (***) [3] Daniel Golle 4 ( +9) [18] Daniel Zahka 5 ( ) [3] Alok Tiwari 5 ( -1) [16] Tony Nguyen 6 ( ) [3] Tariq Toukan 6 ( +5) [15] Eric Dumazet 7 ( +3) [2] Heiner Kallweit 7 ( +7) [15] Marc Kleine-Budde 8 ( -5) [2] Kuniyuki Iwashima 8 (+11) [14] Tariq Toukan 9 (+13) [2] Hangbin Liu 9 (+11) [13] Fan Gong
10 (***) [2] Matthieu Baerts 10 (+20) [12] Laura Nao
Daniel Zahka worked on adding PSP support. Fan Gong hinic3 driver. Alok sends a lot of fixes and improvements to random code. One could speculate what technology lays behind those patches but the patches are generally solid :)
Laura Nao resends a 27 patch series for MediaTek clock drivers, which is of moderate relevance to netdev, but stats are stats.
The remaining names should be familiar.
Top scores (positive): Top scores (negative):
1 ( +1) [436] Jakub Kicinski 1 ( +7) [68] Daniel Zahka 2 ( -1) [338] Simon Horman 2 (+11) [52] Fan Gong 3 ( ) [304] Andrew Lunn 3 (+12) [51] Marc Kleine-Budde 4 ( ) [104] Paolo Abeni 4 ( +1) [50] Tony Nguyen 5 ( ) [ 96] Willem de Bruijn 5 (+17) [50] Laura Nao 6 ( +1) [ 67] Vadim Fedorenko 6 ( -5) [45] Chia-Yu Chang 7 ( -1) [ 62] Krzysztof Kozlowski 7 (+18) [42] Tariq Toukan 8 ( ) [ 55] Rob Herring 8 (***) [42] David Hildenbrand 9 ( ) [ 49] Eric Dumazet 9 (***) [42] Christian Brauner
10 (+19) [ 48] Paul Menzel 10 (***) [40] Amery Hung
Company rankings
----------------
Top reviewers (cs): Top reviewers (msg):
1 ( +1) [39] Meta 1 ( +1) [98] Meta 2 ( -1) [33] RedHat 2 ( -1) [86] RedHat 3 ( ) [17] Intel 3 ( ) [47] Google 4 ( +1) [17] Andrew Lunn 4 ( +1) [46] Andrew Lunn 5 ( -1) [13] Google 5 ( -1) [37] Intel 6 ( +2) [ 8] Oracle 6 ( +2) [25] Oracle 7 ( ) [ 6] nVidia 7 ( -1) [17] nVidia
Top authors (cs): Top authors (msg):
1 ( ) [12] Meta 1 ( ) [86] Meta 2 ( +1) [10] RedHat 2 ( +1) [57] RedHat 3 ( -1) [10] Google 3 ( -1) [54] Google 4 ( +1) [ 8] Intel 4 ( +2) [50] Intel 5 ( -1) [ 7] nVidia 5 ( -1) [43] nVidia 6 ( +1) [ 7] Oracle 6 (+13) [29] Oracle 7 (***) [ 3] Daniel Golle 7 ( +1) [23] Pengutronix
Top scores (positive): Top scores (negative):
1 ( +1) [340] Meta 1 ( +8) [70] Collabora 2 ( -1) [337] RedHat 2 ( +8) [68] Qualcomm 3 ( ) [304] Andrew Lunn 3 ( +2) [66] Pengutronix 4 ( ) [ 80] Intel 4 ( -1) [60] nVidia 5 ( ) [ 76] Linaro 5 ( -4) [54] Huawei 6 ( ) [ 73] Google 6 (***) [51] Broadcom 7 ( ) [ 65] ARM 7 ( -5) [45] Nokia
The aforementioned cross-poster Laura Nao works at Collabora. Broadcom ranking gets pretty negative due to multiple revisions the patches for their new driver are going.
Meta (very narrowly) interrupts Red Hat's period of domination across the rankings.
Signature · 2 lines
Code: https://github.com/kuba-moo/ml-stat Raw output: https://netdev.bots.linux.dev/static/nipa/stats-6.18/stdout