Re: [GIT PULL] Networking for v7.0-rc3
From: Jakub Kicinski
Date: Thu Mar 05 2026 - 12:24:23 EST
On Thu, 5 Mar 2026 09:10:04 -0800 Jakub Kicinski wrote:
> A few PRs here narrowly missed the previous rc, but I'm not sure
> if that explains why this PR is almost 2x the usual size.
> I had a feeling that AI tools are helping people investigate
> and send fixes for issues which would previously be waved off
> as one-off / not worth the time. But the data doesn't necessarily
> bear that out.
FWIW when I say data this is supposedly breakdown of how many people
who contributed fixes to networking have only been known to git for
3, 6 or 12 months:
Commits Date Avg Age Med Age <3mo <3mo% <6mo <6mo% <12mo <12m%
------------------------------------------------------------------------------------------
100 2024-09-04 9.9yr 10.5yr 10 10.0% 13 13.0% 15 15.0%
100 2024-10-01 8.5yr 8.1yr 5 5.0% 5 5.0% 10 10.0%
100 2024-10-13 9.2yr 9.2yr 11 11.0% 11 11.0% 18 18.0%
100 2024-10-24 7.5yr 5.8yr 5 5.0% 6 6.0% 15 15.0%
100 2024-11-06 7.1yr 5.7yr 18 18.0% 18 18.0% 21 21.0%
100 2024-11-22 8.4yr 7.3yr 8 8.0% 10 10.0% 10 10.0%
100 2024-12-08 9.4yr 10.5yr 4 4.0% 9 9.0% 9 9.0%
100 2024-12-20 7.1yr 5.1yr 9 9.0% 12 12.0% 28 28.0%
100 2025-01-10 8.5yr 7.7yr 8 8.0% 8 8.0% 12 12.0%
100 2025-01-26 10.9yr 12.9yr 5 5.0% 7 7.0% 9 9.0%
100 2025-02-07 9.7yr 9.6yr 7 7.0% 9 9.0% 10 10.0%
100 2025-02-24 8.1yr 7.4yr 7 7.0% 12 12.0% 13 13.0%
100 2025-03-07 8.2yr 6.9yr 6 6.0% 7 7.0% 10 10.0%
100 2025-03-24 10.4yr 13.1yr 7 7.0% 8 8.0% 8 8.0%
100 2025-04-07 8.1yr 6.9yr 10 10.0% 13 13.0% 15 15.0%
100 2025-04-18 7.3yr 6.1yr 14 14.0% 19 19.0% 22 22.0%
100 2025-04-30 9.3yr 10.1yr 6 6.0% 9 9.0% 12 12.0%
100 2025-05-14 7.7yr 6.5yr 8 8.0% 10 10.0% 14 14.0%
100 2025-06-02 9.3yr 9.8yr 5 5.0% 6 6.0% 9 9.0%
100 2025-06-11 9.8yr 6.8yr 8 8.0% 8 8.0% 10 10.0%
100 2025-06-24 9.0yr 6.9yr 9 9.0% 10 10.0% 13 13.0%
100 2025-06-12 8.8yr 7.1yr 11 11.0% 13 13.0% 15 15.0%
100 2025-07-17 10.3yr 11.4yr 7 7.0% 8 8.0% 19 19.0%
100 2025-08-06 8.3yr 7.8yr 5 5.0% 6 6.0% 8 8.0%
100 2025-08-19 7.6yr 6.4yr 10 10.0% 12 12.0% 20 20.0%
100 2025-08-18 10.2yr 10.1yr 2 2.0% 3 3.0% 6 6.0%
100 2025-09-08 9.7yr 9.9yr 7 7.0% 9 9.0% 13 13.0%
100 2025-08-13 8.1yr 7.4yr 17 17.0% 19 19.0% 22 22.0%
100 2025-10-09 9.1yr 9.3yr 12 12.0% 16 16.0% 17 17.0%
100 2025-11-05 7.5yr 5.9yr 12 12.0% 16 16.0% 19 19.0%
100 2025-11-18 8.7yr 10.2yr 12 12.0% 18 18.0% 19 19.0%
100 2025-12-09 9.2yr 8.4yr 12 12.0% 13 13.0% 18 18.0%
100 2025-12-21 8.5yr 7.6yr 20 20.0% 21 21.0% 26 26.0%
100 2026-01-09 9.8yr 8.3yr 8 8.0% 9 9.0% 13 13.0%
100 2026-01-12 6.9yr 5.7yr 14 14.0% 17 17.0% 24 24.0%
95 2026-02-27 10.0yr 10.4yr 11 11.6% 11 11.6% 11 11.6%
We have a sequence of high percentages since ~Nov but not in a crazy
way.
The numbers looking at just "core parts" (net/core net/ipv4
net/ipv6 net/sched) would support the theory a little more
but again not very strongly and the numbers are small in absolute
terms:
Commits Date Avg Age Med Age <3mo <3mo% <6mo <6mo% <12mo <12m%
------------------------------------------------------------------------------------------
24 2024-09-11 13.7yr 14.7yr 0 0.0% 0 0.0% 0 0.0%
24 2024-10-15 9.4yr 9.1yr 1 4.2% 1 4.2% 1 4.2%
24 2024-11-28 9.4yr 9.2yr 4 16.7% 4 16.7% 4 16.7%
24 2024-12-19 11.5yr 12.9yr 1 4.2% 1 4.2% 2 8.3%
24 2025-01-27 16.0yr 19.6yr 0 0.0% 0 0.0% 0 0.0%
24 2025-02-07 10.7yr 13.0yr 1 4.2% 1 4.2% 2 8.3%
24 2025-02-25 7.8yr 5.3yr 1 4.2% 1 4.2% 1 4.2%
24 2025-03-28 10.8yr 13.1yr 1 4.2% 1 4.2% 1 4.2%
24 2025-04-03 10.3yr 13.1yr 0 0.0% 0 0.0% 0 0.0%
24 2025-04-26 9.6yr 11.2yr 2 8.3% 2 8.3% 3 12.5%
24 2025-06-04 11.8yr 13.3yr 1 4.2% 1 4.2% 3 12.5%
24 2025-07-09 10.7yr 12.4yr 3 12.5% 3 12.5% 5 20.8%
24 2025-08-08 9.4yr 11.4yr 2 8.3% 3 12.5% 3 12.5%
24 2025-09-10 11.1yr 13.6yr 2 8.3% 4 16.7% 5 20.8%
24 2025-10-28 10.9yr 13.8yr 3 12.5% 5 20.8% 5 20.8%
24 2026-01-05 14.5yr 20.5yr 0 0.0% 0 0.0% 2 8.3%
20 2026-02-23 9.1yr 7.2yr 4 20.0% 4 20.0% 4 20.0%