Drop table overrides in x86_syscall_resolve_name() to pseudo syscall … #41

Closed
wants to merge 1 commit into
from

Conversation

Projects
None yet
1 participant

xnox commented Jun 10, 2016

…numbers

A bunch of syscalls got converted from pseudo to real syscall
numbers. And the table now corectly uses pseudo syscall numbers where
needed (e.g. accept, send) thus there is no need to repeat these
overrides in the x86_syscall_resolve_name() which has now got out of
sync with the table.

Fixes #40

Signed-off-by: Dimitri John Ledkov xnox@ubuntu.com

Drop table overrides in x86_syscall_resolve_name() to pseudo syscall …
…numbers

A bunch of syscalls got converted from pseudo to real syscall
numbers. And the table now corectly uses pseudo syscall numbers where
needed (e.g. accept, send) thus there is no need to repeat these
overrides in the x86_syscall_resolve_name() which has now got out of
sync with the table.

Fixes #40

Signed-off-by: Dimitri John Ledkov <xnox@ubuntu.com>

xnox commented Jun 10, 2016

After this, my tests as pointed out in #40 pass, however, the upstream test-suite now fails with:

Test 30-sim-socket_syscalls%%012-00001 result: FAILURE bpf_sim resulted in ALLOW

xnox commented Jun 10, 2016

Horum.

30-sim-socket_syscalls  +x86    accept4         0               1               2       N       N       N       KILL

Not sure why it should be a KILL and/or how&why my code change results in this test suite failure, when e.g. "accept" ALLOW/KILL tests remain operating correctly.

xnox commented Jun 10, 2016

Actually, i"m not sure if this makes sense anymore.

accept4 is 364 on x86 from now on. Thus these two tests are identical, yet expect different results

30-sim-socket_syscalls  +x86    364             0               1               2       N       N       N       ALLOW
30-sim-socket_syscalls  +x86    accept4         0               1               2       N       N       N       KILL

which is not possible....

xnox commented Jun 10, 2016

actually this is intentional.

@xnox xnox closed this Jun 10, 2016

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment