pcictl pci0 list panics on ROCKPro64

[Masanobu SAITOH <msaitoh%execsw.org@localhost>]

 Hi.

 I've not used my ROCKPro64 for more than two years.
I've updated whole of the system (boot loader, dtb,
kernel and userland) today and noticed a problem.

If an PCIe card is inserted the PCIe slot, "pcictl pci0 list"
panics with high probability.

-----------------------
[ 310.1429507] panic: kernel diagnostic assertion "l->l_nopreempt == 0" failed: file "/usr/src/sys/sys/userret.h", line 63
[ 310.1429507] cpu4: Begin traceback...
[ 310.1429507] trace fp ffffc0016774fce0
[ 310.1429507] fp ffffc0016774fd10 vpanic() at ffffc000005a9f88 netbsd:vpanic+0x178
[ 310.1429507] fp ffffc0016774fd70 kern_assert() at ffffc00000843408 netbsd:kern_assert+0x58
[ 310.1429507] fp ffffc0016774fe00 mi_userret() at ffffc000000b3c8c netbsd:mi_userret+0xe8
[ 310.1529513] fp ffffc0016774fe20 syscall() at ffffc000000b3928 netbsd:syscall+0x108
[ 310.1529513] tf ffffc0016774fed0 el0_trap() at ffffc000000b67f0 netbsd:el1_trap_exit+0x68
[ 310.1529513] ---- SVC Instruction Execution: trapframe 0xffffc0016774fed0 (304 bytes) ----
[ 310.1529513]     pc=0000f5b8a2088900,   spsr=0000000080000000
[ 310.1529513]    esr=0000000056000036,    far=0000000000000000
[ 310.1529513]     x0=0000000000000000,     x1=0000000000000000
[ 310.1529513]     x2=0000ffffffed8b78,     x3=0000000000000005
[ 310.1529513]     x4=0000000000000000,     x5=0000000000000000
[ 310.1529513]     x6=00000000000000a8,     x7=7f7f7f7f7f7f7f7f
[ 310.1529513]     x8=0000ffffffed8200,     x9=7369766572202c6b
[ 310.1529513]    x10=0000ffffffed89f0,    x11=0101010101010101
[ 310.1529513]    x12=0000000003082048,    x13=0000ffffffed8a30
[ 310.1529513]    x14=0000ffffffed8a30,    x15=0000ffffffed8a30
[ 310.1529513]    x16=0000f5b8a23e0008,    x17=0000f5b8a20888fc
[ 310.1529513]    x18=000000000000000a,    x19=0000ffffffed8c08
[ 310.1529513]    x20=0000000000000007,    x21=0000ffffffed8c08
[ 310.1629505]    x22=000000000000ffff,    x23=00000000ffffffff
[ 310.1629505]    x24=0000000003081684,    x25=0000000000000005
[ 310.1629505]    x26=0000ffffffed8c0c,    x27=0000000000000000
[ 310.1629505]    x28=00000000000000a8, fp=x29=0000ffffffed8b50
[ 310.1629505] lr=x30=0000f5b8a237cef4,     sp=0000ffffffed8b50
[ 310.1629505] ------------------------------------------------
[ 310.1629505] cpu4: End traceback...
Stopped in pid 1418.1418 (pcictl) at    netbsd:cpu_Debugger+0xc:        ldp
x29, x30, [sp],#16
db{4}>
--------------------

I read rk3399_pcie.c. It seems it has no such problem.
I tried the above command several times and noticed that it occurs only on cpu4 or 5.

Not panic:
	schedctl -A 0,1,2,3 pcictl pci0 list

Panic:
	schedctl -A 4,5 pcictl pci0 list

Is there a bug in gic or somewhere else?

Patch is welcomed.


-- 
-----------------------------------------------
                SAITOH Masanobu (msaitoh%execsw.org@localhost
                                 msaitoh%netbsd.org@localhost)