[OpenAFS-devel] pr_Initialize falis when called within a simple application

[John Hayes]

One striking difference between the two programs can be seen in the
following output from strace. Remember that some debugger code was put in
tyhe libraries and that will show up as well -


strace output from klog:
getitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={2147483, 646000}}) = 0
write(1, "in rx1\n", 7)                 = 7
write(1, "got mem!\n", 9)               = 9
socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 4
ioctl(4, 0x8912, 0xbfffc188)            = 0
ioctl(4, 0x8915, 0xbfffc190)            = 0
ioctl(4, 0x8915, 0xbfffc1b0)            = 0
ioctl(4, 0x8913, 0xbfffc1b0)            = 0
ioctl(4, 0x8921, 0xbfffc1b0)            = 0
ioctl(4, 0x891b, 0xbfffc1b0)            = 0
close(4)                                = 0
brk(0)                                  = 0x80b5000
brk(0x80dc000)                          = 0x80dc000


strace output from my test program:
etitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={2147483, 646000}}) = 0
open("/home/jhh/.AFSSERVER", O_RDONLY)  = -1 ENOENT (No such file or
directory)
open("/.AFSSERVER", O_RDONLY)           = -1 ENOENT (No such file or
directory)
afs_syscall(0x14, 0, 0x400c5608, 0xbfffc770, 0) = 0
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 2), ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x4001a000
write(1, "in rx1\n", 7)                 = 7
write(1, "got mem!\n", 9)               = 9
socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 4
ioctl(4, 0x8912, 0xbfffc4f8)            = 0
ioctl(4, 0x8915, 0xbfffc500)            = 0
ioctl(4, 0x8915, 0xbfffc520)            = 0
ioctl(4, 0x8913, 0xbfffc520)            = 0
ioctl(4, 0x8921, 0xbfffc520)            = 0
ioctl(4, 0x891b, 0xbfffc520)            = 0
close(4)                                = 0
brk(0)                                  = 0x8086000
brk(0x80ad000)                          = 0x80ad000


Why is my test program looking for this .AFSSERVER file?



>> Unlikely...  Byteswapping of 42166 yields 46756 (7004 swaps to 23579).
>> So I don't think it's pure byteswapping.  It's not bit-swapped,
>> either..  So I have no clue where this port number came from.  Perhaps
>> an uninitialized variable?
>
> Well, I think one thing to ask, again, is what stracing and grepping for
> calls to open returned different in each case; but that's probably not it
> either.