Results of doing a LTP test run with the p4 patches
Gary Corcoran
gcorcoran at rcn.com
Sun Oct 8 10:28:10 PDT 2006
Alexander Leidinger wrote:
> Hi,
>
> a test run for 2.4.2 reveals:
>
> -clone05 FAIL 1
> +clone05 PASS 0
>
> -fcntl17 PASS 0
> +fcntl17 FAIL 1
>
> -setpriority04 FAIL 2
> +setpriority04 PASS 0
>
> -wait402 FAIL 2
> +wait402 PASS 0
>
> -waitpid04 FAIL 1
> +waitpid04 PASS 0
>
> -Total Failures: 164
> +Total Failures: 162
>
>
> The fcntl17 test output is:
> ---snip---
> fcntl17 0 INFO : Enter preparation phase
> fcntl17 0 INFO : child 3 starting
> fcntl17 0 INFO : child 3 pid 45870 locked
> fcntl17 0 INFO : child 3 resuming
> fcntl17 0 INFO : child 3 lockw err 35
> fcntl17 0 INFO : child 3 exiting
> fcntl17 0 INFO : Enter preparation phase
> fcntl17 0 INFO : child 1 starting
> fcntl17 0 INFO : child 1 pid 45868 locked
> fcntl17 0 INFO : child 1 resuming
> fcntl17 0 INFO : child 1 unlocked
> fcntl17 0 INFO : child 1 exiting
> fcntl17 0 INFO : Enter preparation phase
> fcntl17 0 INFO : child 2 starting
> fcntl17 0 INFO : child 2 pid 45869 locked
> fcntl17 0 INFO : child 2 resuming
> fcntl17 0 INFO : child 2 lockw locked
> fcntl17 0 INFO : child 2 exiting
> fcntl17 0 INFO : Enter preparation phase
> fcntl17 0 INFO : Exit preparation phase
> fcntl17 0 INFO : Enter block 1
> fcntl17 1 FAIL : locking pid is wrong, should be -19668 is 45868
> fcntl17 0 INFO : do_test failed child 1
> fcntl17 2 FAIL : locking pid is wrong, should be -19667 is 45869
> fcntl17 0 INFO : do_test failed child 2
> fcntl17 3 FAIL : locking pid is wrong, should be -19666 is 45870
> fcntl17 0 INFO : do_test failed child 3
> fcntl17 4 FAIL : locking pid is wrong, should be -19667 is 45869
> fcntl17 5 FAIL : locking pid is wrong, should be -19666 is 45870
> fcntl17 0 INFO : Block 1 FAILED
> fcntl17 0 INFO : Exit block 1
> ---snip---
>
> I'm a little bit puzzled. The PIDs correspond to the PIDs of the
> childs, but the expectation looks completely wrong.
Those look like somebody is using a signed 16-bit int when it should be
an unsigned int. -19668 (as 16 bits unsigned) is the same as 45868 ...
Gary
More information about the freebsd-emulation
mailing list