Re: kern/39173: random processes get stuck in "D" state

[Brett Lymn <blymn%internode.on.net@localhost>]

On Sun, Jul 20, 2008 at 02:17:08PM +0100, Mindaugas Rasiukevicius wrote:
> 
> - Do you have a back-trace of stuck process?
> 

I realised you meant a ddb backtrace of the process.  Here is one from a
wedged firefox3 process - the waitchannel label was uvn_putpages which
agrees with the backtrace.

trace: pid 870 lid 1 at 0xd554e92c
sleepq_block(0,0,c0a1e29f,c0a96a38,c042a4c0,c0b4907c,75,c12f758c,d577da10,c12f45
80) at netbsd:sleepq_block+0x102
mtsleep(c12f758c,204,c0a1e29f,0,d577da10,13,d554e9cc,c0496e68,cf90b4cc,d554eac4)
 at netbsd:mtsleep+0x11d
genfs_do_putpages(d577da10,0,0,0,0,13,0,1,0,0) at netbsd:genfs_do_putpages+0x9c6
genfs_putpages(d554eb6c,d55052a0,0,0,1,d577dab0,c080cd40,d577da10,0,0) at netbsd
:genfs_putpages+0x3d
VOP_PUTPAGES(d577da10,0,0,0,0,13,d554ebbc,d13cf000,d0ec9be0,0) at netbsd:VOP_PUT
PAGES+0x7f
ffs_full_fsync(d577da10,1,1,c047ced4,0,0,c04d3ff5,0,d55052a4,fffffff4) at netbsd
:ffs_full_fsync+0x284
ffs_fsync(d554ec88,d584b964,c080cac0,d577da10,2,c080c780,d577da10,d408df00,1,0) 
at netbsd:ffs_fsync+0x62
VOP_FSYNC(d577da10,d408df00,1,0,0,0,0,1000,d576ec40,0) at netbsd:VOP_FSYNC+0x7e
sys_fsync(d55052a0,d554ed00,d554ed28,17,0,bfbfdd58,bbbe04e6,0,28,bfbfdc98) at ne
tbsd:sys_fsync+0x82
syscall(d554ed48,b3,bb8100ab,b940001f,bfbf001f,0,bfa00000,bfbfdd58,bbbe71a0,0) a
t netbsd:syscall+0xa0

-- 
Brett Lymn