> So what does truss say?
System with timeout:
0.3767 open("/dev/lo0", O_RDWR) = 3
0.3768 putmsg(3, 0x08047878, 0x00000000, RS_HIPRI) = 0
0.3769 pollsys(0x08043818, 1, 0x080437E8, 0x00000000) = 1
0.3769 getmsg(3, 0x08043820, 0x0804582C, 0x08047884) = 0
pollsys(0x08043818, 1, 0x080437E8, 0x00000000) (sleeping...)
5.3824 pollsys(0x08043818, 1, 0x080437E8, 0x00000000) = 0
5.3826 close(3) = 0
5.3827 open("/dev/lo", O_RDWR) Err#2 ENOENT
5.3828 open("/devices/pseudo/clone at 0:lo", O_RDWR) = 3
5.3828 putmsg(3, 0x08047878, 0x00000000, RS_HIPRI) = 0
5.3828 pollsys(0x08043818, 1, 0x080437E8, 0x00000000) = 1
5.3829 getmsg(3, 0x08043820, 0x0804582C, 0x08047884) = 0
pollsys(0x08043818, 1, 0x080437E8, 0x00000000) (sleeping...)
10.3924 pollsys(0x08043818, 1, 0x080437E8, 0x00000000) = 0
10.3925 close(3) = 0
10.3926 fstat64(2, 0x08046DD0) = 0
10.3927 write(2, " d l a d m", 5) = 5
10.3928 write(2, " : w a r n i n g : ", 11) = 11
10.3928 write(2, " c a n n o t o p e n ", 12) = 12
10.3929 write(2, " l o 0", 3) = 3
10.3929 write(2, " : ", 2) = 2
10.3930 write(2, " D L P I o p e r a t i".., 24) = 24
10.3930 ioctl(1, TCGETA, 0x08047C1C) = 0
10.3930 fstat64(1, 0x08047B80) = 0
10.3941 write(1, "\n", 1) = 1
10.3942 _exit(1)
System without timeout:
0.1434 open("/dev/lo0", O_RDWR) = 3
0.1437 open("/dev/lo0", O_RDWR) = 4
0.1438 brk(0x0806DFE8) = 0
0.1440 brk(0x0807DFE8) = 0
0.1446 ioctl(4, I_STR, 0x08047500) Err#22 EINVAL
0.1447 close(4) = 0
0.1450 open("/devices/pseudo/dld at 0:ctl", O_RDWR) = 4
0.1452 ioctl(4, I_STR, 0x080476D0) Err#2 ENOENT
0.1453 close(4) = 0
0.1455 ioctl(1, TCGETA, 0x08046A6C) = 0
0.1457 fstat64(1, 0x080469D0) = 0
LINK PROPERTY VALUE DEFAULT POSSIBLE
0.1458 write(1, " L I N K ".., 81) = 81
lo0 zone -- -- --
0.1460 write(1, " l o 0 ".., 81) = 81
0.1461 close(3) = 0
0.1463 _exit(0)
Seems in line with Jim's explanation. Though neither system has the
/dev/lo link. Is "/devices/pseudo/clone at 0:lo" hard-coded somewhere?
-Artem
