diff options
| author | Aleksandr Nogikh <nogikh@google.com> | 2024-11-01 13:05:28 +0100 |
|---|---|---|
| committer | Aleksandr Nogikh <nogikh@google.com> | 2024-11-01 13:24:37 +0000 |
| commit | 28194a719952cbe880a9f0f53f2dda6c86019b07 (patch) | |
| tree | 41922a619e964dddf8f3f11f4b2fcc51df88fd37 /dashboard/config/linux/bits | |
| parent | 96eb609f6eb9558e2fab491a2c63a11bb0556471 (diff) | |
dashboard/config: set RCU_EXP_CPU_STALL_TIMEOUT to 0
The zero values makes the kernel use the RCU_CPU_STALL_TIMEOUT value,
which is already set to a big enough value.
Context: see the recent LKML discussion at
https://lore.kernel.org/all/6723aa4d.050a0220.35b515.0161.GAE@google.com/T/
Diffstat (limited to 'dashboard/config/linux/bits')
| -rw-r--r-- | dashboard/config/linux/bits/timeouts_emu.yml | 6 | ||||
| -rw-r--r-- | dashboard/config/linux/bits/timeouts_native.yml | 7 |
2 files changed, 4 insertions, 9 deletions
diff --git a/dashboard/config/linux/bits/timeouts_emu.yml b/dashboard/config/linux/bits/timeouts_emu.yml index c8f4b01d0..390459752 100644 --- a/dashboard/config/linux/bits/timeouts_emu.yml +++ b/dashboard/config/linux/bits/timeouts_emu.yml @@ -8,7 +8,5 @@ config: # See the comment in x86_64.yml re these numbers. - RCU_CPU_STALL_TIMEOUT: 300 - DEFAULT_HUNG_TASK_TIMEOUT: [420, -nonoise, -kmsan, -kcsan, -kmemleak] - # 21000 is the max value for the config, so we don't multiply it. - # TODO: `rcu: Allow up to five minutes expedited RCU CPU stall-warning timeouts` lets us set more than 21k, - # but we first need to change the config generation tool to support multiple possible parameter values. - - RCU_EXP_CPU_STALL_TIMEOUT: [21000, v5.19] + # A zero RCU_EXP_CPU_STALL_TIMEOUT value makes it equal to RCU_CPU_STALL_TIMEOUT. + - RCU_EXP_CPU_STALL_TIMEOUT: [0, v5.19] diff --git a/dashboard/config/linux/bits/timeouts_native.yml b/dashboard/config/linux/bits/timeouts_native.yml index 08c88641b..56e43b8db 100644 --- a/dashboard/config/linux/bits/timeouts_native.yml +++ b/dashboard/config/linux/bits/timeouts_native.yml @@ -30,8 +30,5 @@ config: - RCU_CPU_STALL_TIMEOUT: 100 # The constraints repeat constraints on inclusion of debug.yml which enables hung task detection. - DEFAULT_HUNG_TASK_TIMEOUT: [140, -nonoise, -kmsan, -kcsan, -kmemleak] - # By default it is set to a too small value, which causes tons of - # `rcu_preempt detected expedited stalls on CPUs/tasks`. - # TODO: `rcu: Allow up to five minutes expedited RCU CPU stall-warning timeouts` lets us set more than 21k, - # but we first need to change the config generation tool to support multiple possible parameter values. - - RCU_EXP_CPU_STALL_TIMEOUT: [21000, v5.19] + # A zero RCU_EXP_CPU_STALL_TIMEOUT value makes it equal to RCU_CPU_STALL_TIMEOUT. + - RCU_EXP_CPU_STALL_TIMEOUT: [0, v5.19] |