8.1-RC2: mvs/ZFS Fatal trap 19: non-maskable interrupt trap while in kernel mode

[Florian Smeets]

Hi,

a few days ago i installed a server with the latest RC. It has an mvs 
based controller, attached to it are 6 disks in a RAIDZ, the server has 
8GB of RAM and is running amd64. The CPU is a single core Xeon with HT 
turned on.

This morning it hanged for the 2nd time, this time i managed to capture 
the panic massage:

Fatal trap 19: non-maskable interrupt trap while in kernel mode
Fatal trap 19: non-maskable interrupt trap while in kernel mode
cpuid = 1; cpuid = 0; apic id = 01
apic id = 00
instruction pointer     = 0x20:0xffffffff80e75c65
instruction pointer     = 0x20:0xffffffff80870f94
stack pointer           = 0x28:0xffffff8000022fe0
stack pointer           = 0x28:0xffffffff80ca9740
frame pointer           = 0x28:0xffffff80000abb40
frame pointer           = 0x28:0xffffff80ee2108b0
code segment            = base 0x0, limit 0xfffff, type 0x1b
code segment            = base 0x0, limit 0xfffff, type 0x1b
                         = DPL 0, pres 1, long 1, def32 0, gran 1
                         = DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags        = processor eflags      = interrupt enabled, 
IOPL = 0
IOPL = 0
current process         = current process               = 0 (spa_zio)
12 (irq48: mvs0)
trap number             = 19
trap number             = 19
panic: non-maskable interrupt trap
cpuid = 0

Here are some details about the controller.

mvs0 at pci0:6:1:0:	class=0x010000 card=0x11ab11ab chip=0x608111ab rev=0x09 
hdr=0x00
     vendor     = 'Marvell Semiconductor (Was: Galileo Technology Ltd)'
     device     = 'MV88SX6081 8-port SATA II PCI-X Controller'
     class      = mass storage
     subclass   = SCSI

After the panic the server just hangs, so i was not able to get a dump 
or get a backtrace. I'm just building a custom kernel with debugging 
turned on.

Anything i could do in the meantime?

Thanks,
Florian