Join GitHub today
GitHub is home to over 20 million developers working together to host and review code, manage projects, and build software together.
tests: check for negative syscalls in runBpf() and skip those tests #3850
Conversation
mvo5
added this to the 2.28 milestone
Sep 5, 2017
codecov-io
commented
Sep 5, 2017
Codecov Report
@@ Coverage Diff @@
## master #3850 +/- ##
==========================================
+ Coverage 75.71% 75.71% +<.01%
==========================================
Files 413 413
Lines 35555 35555
==========================================
+ Hits 26921 26922 +1
+ Misses 6730 6729 -1
Partials 1904 1904
Continue to review full report at Codecov.
|
mvo5
requested a review
from
jdstrand
Sep 5, 2017
mvo5
merged commit 29180f1
into
snapcore:master
Sep 5, 2017
2 of 8 checks passed
| +// can be skipped when "socketcall()" is used instead of "socket()". | ||
| +// | ||
| +// Some architectures (i386, s390x) use the "socketcall" syscall instead | ||
| +// of "socket". This is the case on Ubuntu 14.04, 17.04, 17.10 |
jdstrand
Sep 5, 2017
Contributor
17.04 and 17.10 shouldn't be using socketcall on i386 and s390x. The kernel and glibc should only be using socket and failures here would indicate a regression on those releases. Three pieces are needed, the kernel, glibc and libseccomp 2.3 for this to work. On 16.04, all this was in place-- why did this regress?
References:
- https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=809556 (i386 fixes)
- https://bugs.launchpad.net/ubuntu/+source/libseccomp/+bug/1554098 (s390x fixes)
- https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/commit/?h=x86/asm&id=9dea5dc921b5f4045a18c63eb92e84dc274d17eb (kernel commit to 4.3)
- https://bugs.launchpad.net/ubuntu/+source/libseccomp/+bug/1576066
mvo5 commentedSep 5, 2017
Some syscalls are resolved to negative numbers by libseccomp. An
example is the "socket" syscall which turns into -101 on i386 and
s390x on 14.04, 17.04 and 17.10.
We cannot run the unittests against a negative number so we need to
skip the unit test on these platforms for now.
This causes a FTBFS on {zesty,artful}/{i386,s390x}
apd/blob/master/CONTRIBUTING.md)?