Petitboot crash on Talos 2 with Broadcom SAS controller

Christian Müller cmueller at trufflepig-forensics.com
Mon Jul 6 21:04:42 AEST 2020


Hi guys,
I've bought a Talos 2 and currently trying to get it working.
There's a Broadcom SAS 9305-24i controller on the board and petitboot
comes up normally.
My Problem: When kexec is being executed (I can select the Alpine image
I want to boot from the petitboot TUI) it crashes with the following
stack trace:

SIGTERM received, booting...
cpu 0x0: Vector: 380 (Data Access Out of Range) at [c000001fe6c47620]
    pc: c0000000001d05f0: __free_pages+0x10/0x50
    lr: c000000000123c24: dma_direct_free_pages+0x54/0x90
    sp: c000001fe6c478b0
   msr: 900000000280b033
   dar: c04240000000f8b4
  current = 0xc000001fe6b88400
  paca    = 0xc000001fff7ff480   irqmask: 0x03   irq_happened: 0x01
    pid   = 1021, comm = kexec
Linux version 5.5.0-openpower1 (root at raptor-build-public-staging-01) (gcc version 6.5.0 (Buildroot 2019.05.3-06769-g7bdd570165)) #2 SMP Thu Feb 20 02:19:47 UTC 2020
enter ? for help
[c000001fe6c478b0] c000000000123c24 dma_direct_free_pages+0x54/0x90 (unreliable)
[c000001fe6c478d0] c000000000038728 dma_iommu_free_coherent+0x98/0xc0
[c000001fe6c47920] c000000000123020 dma_free_attrs+0x100/0x110
[c000001fe6c47970] c0000000001d9bf4 dma_pool_destroy+0x174/0x200
[c000001fe6c47a10] c0080000011617e8 _base_release_memory_pools+0x1e0/0x498 [mpt3sas]
[c000001fe6c47aa0] c00800000116b428 mpt3sas_base_detach+0x40/0x160 [mpt3sas]
[c000001fe6c47b10] c00800000117bb5c scsih_shutdown+0xc4/0x110 [mpt3sas]
[c000001fe6c47b70] c0000000003cda10 pci_device_shutdown+0x50/0xc0
[c000001fe6c47ba0] c00000000064a908 device_shutdown+0x1f8/0x330
[c000001fe6c47c40] c0000000000cfe2c kernel_restart_prepare+0x4c/0x60
[c000001fe6c47c60] c0000000001505c0 kernel_kexec+0xa0/0xe0
[c000001fe6c47cd0] c0000000000d03f4 __do_sys_reboot+0x234/0x2c0
[c000001fe6c47e20] c00000000000b50c system_call+0x5c/0x68
--- Exception: c01 (System Call) at 00007fffa70f10a4
SP (7ffff4af0240) is in userspace
0:mon> 

As far as I can tell it's before the entry point of the Alpine kernel is beeing called, so this might be an Petitboot issue or (more likely) a firmware issue with the card. I have updated the firmware to the current version and the problem persists.
Can you give me some hints how I can proceed/debug this issue?

Thanks for your time and have a nice day!

Chris


More information about the Petitboot mailing list