Upgrade to NCS 3.0.2 causes fatal error

[Moh]

Hi there,

I have recently upgraded my software to latest ncs 3.2.0. I have noticed that my initialisation does complete but it ends up with fatal error that seems to be related to bt_enable. Please help me sort out this issue as I have tried various options. Here is the error and prj file 

[00:00:03.970,489] <err> os: ***** MPU FAULT *****
[00:00:03.978,637] <err> os: Data Access Violation
[00:00:03.986,968] <err> os: MMFAR Address: 0x20012300
[00:00:03.995,727] <err> os: r0/a1: 0x200122f4 r1/a2: 0x00078981 r2/a3: 0x00000000
[00:00:04.007,263] <err> os: r3/a4: 0x200122ff r12/ip: 0x00078a68 r14/lr: 0x00058f8b
[00:00:04.018,829] <err> os: xpsr: 0x81000000
[00:00:04.026,763] <err> os: Faulting instruction address (r15/pc): 0x000715f6
[00:00:04.037,445] <err> os: >>> ZEPHYR FATAL ERROR 2: Stack overflow on CPU 0
[00:00:04.048,126] <err> os: Current thread: 0x20004c88 (BT LW WQ)
[00:00:04.057,769] <err> fatal_error: Resetting system

7384.prj.conf

Thanks

[Susheel Nuguru]

Seems like BT workqueue size and increased in your use case.

Can you attempt to increase these below in your prj.conf to something more than you already have (check your .config in your build folder for the current values). For example check below.

CONFIG_BT_L2CAP=y
CONFIG_BT_LONG_WQ_STACK_SIZE=2048
CONFIG_SYSTEM_WORKQUEUE_STACK_SIZE=2048

[Moh]

Thanks for your response. I tried but no luck. As per map file, this is where it causes

.bss.bt_long_wq

0x0000000020004c88 0xe8 zephyr/subsys/bluetooth/host/libsubsys__bluetooth__host.a(long_wq.c.obj)

[Susheel Nuguru]

Can you reproduce it on nRF52840 DK? If yes, help me reproduce it and I can try to tweek around more configs to see if it is neighboring stacks that are spilling into the BT LW WQ stack.