zygo: config: expedited RCU stall timeout back to the default (was 20 ms)

The v6.18.41 oldconfig left CONFIG_RCU_EXP_CPU_STALL_TIMEOUT at 20 ms, which
with lockdep on turns ordinary work into "stalls": at the end of a degraded
octopus pass the unmount's workqueue teardown spent 27 jiffies in lockdep's
zap_class() (lockdep_unregister_key from pwq_release_workfn) and the RCU
expedited-stall NMI backtrace fired, failing the kernel-health row of a pass
that had nothing wrong with it (dhive lane618 P2, 2026-09-13).  0 is the
Kconfig default, derives the expedited timeout from the normal one, and is
what the misc-next lane already carries.

Assisted-by: Claude:claude-fable-5-1

diff --git a/.config b/.config
index d4aac19930cbde854a7aca8f69d5ec11e6bdbef0..ac9e797e544b63183071ecbe911a2d1dbc3e5421 100644 (file)
--- a/.config
+++ b/.config
@@ -13560,7 +13560,7 @@ CONFIG_RCU_TORTURE_TEST=m
 # CONFIG_RCU_TORTURE_TEST_LOG_GP is not set
 CONFIG_RCU_REF_SCALE_TEST=m
 CONFIG_RCU_CPU_STALL_TIMEOUT=60
-CONFIG_RCU_EXP_CPU_STALL_TIMEOUT=20
+CONFIG_RCU_EXP_CPU_STALL_TIMEOUT=0
 CONFIG_RCU_CPU_STALL_CPUTIME=y
 CONFIG_RCU_TRACE=y
 CONFIG_RCU_EQS_DEBUG=y