Re: -current kernel hang with a 4.0 binary

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Fri, Oct 24, 2008 at 08:54:22PM +0200, Manuel Bouyer wrote:
> Hi,
> I've got a -current NetBSD/sparc (sun4U system) system hang, reliably,
> when building pkgsrc/python24 on a 4.0 userland. Here's what ddb says:
> [BREAK]
> Stopped in pid 7114.1 (conftest) at     netbsd:cpu_Debugger+0x4:        nop
> db> tr
> sparc_interrupt(b945a30, cc7d700, cc7d700, 0, ded7fc0, ded1f80) at 
> netbsd:sparc_interrupt+0x1f0
> sys_sa_enable(cc7d700, d4bbdd0, d4bbe10, 1, 0, 1c14000) at 
> netbsd:sys_sa_enable+0xd8
> syscall_plain(d4bbed0, cc52d20, 400f4f34, d4bbdd0, 400, 400f4f38) at 
> netbsd:syscall_plain+0x318
> ?(ffe00034, ffffb744, c, 3, ffffb7f7, ffffb7f8) at 0x1008c74
> db> 
> 
> the stack trace is always the same.
> [...]

more play with ddb shows that it's looping on sparc_interrupt+0x1f0:
[BREAK]
Stopped in pid 13330.1 (conftest) at    netbsd:cpu_Debugger+0x4:        nop
db> tr
sparc_interrupt(b9459f0, cc7ea80, cc7ea80, 0, dee5fc0, dedff80) at netbsd:sparc_
interrupt+0x1f0
sys_sa_enable(cc7ea80, d7ffdd0, d7ffe10, 1, 0, 1c14000) at netbsd:sys_sa_enable+
0xd8
syscall_plain(d7ffed0, cc523c0, 400f4f34, d7ffdd0, 400, 400f4f38) at netbsd:sysc
all_plain+0x318
?(ffe00034, ffffb744, c, 3, ffffb7f7, ffffb7f8) at 0x1008c74
db> break sys_sa_enable+0xd8
db> c
[BREAK]
Stopped in pid 13330.1 (conftest) at    netbsd:cpu_Debugger+0x4:        nop
db> break sparc_interrupt+0x1f0
db> c
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7
db> s
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7
db> 
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7
db> 
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7
db> 
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7
db> 
Breakpoint in pid 13330.1 (conftest) at netbsd:sparc_interrupt+0x1f0:   jmpl    
        [%o4 + %g0], %o7

db> sh reg
tstate      0x44
pc          0
npc         0
ipl         0xa
y           0xa000
g0          0
g1          0
g2          0
g3          0
g4          0
g5          0
g6          0
g7          0xffffffff
o0          0
o1          0
o2          0
o3          0
o4          0
o5          0
o6          0
o7          0
l0          0x82000e04
l1          0x2807540
l2          0xe0018078
l3          0xa
l4          0xb9459f0
l5          0xcc7ea80
l6          0xdee5fc0
l7          0xd7ffd08
i0          0
i1          0
i2          0
i3          0
i4          0
i5          0
i6          0
i7          0
f0          0
f2          0
f4          0
f6          0
f8          0
f10         0
f12         0
f14         0
[f* are all 0]
fsr         0
gsr         0
db> print sparc_interrupt+0x1f0
 1008f74

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--