[strace/strace] make check fails with -j16 and higher (#74)

Dmitry V. Levin notifications at github.com
Mon Dec 9 19:08:44 UTC 2019

On Mon, Dec 09, 2019 at 04:42:52AM -0800, Quentin Monnet wrote:
> There is a delay indeed when the locked memory is being reclaimed after a BPF program or map is removed. If I understand correctly it could come from the `call_rcu()` in [`__bpf_put_prog_noref()`](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/kernel/bpf/syscall.c?h=v5.4#n1339).
> For the record, in bpftool we call `setrlimit()` to work around it and [raise the limit](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8302b9bd31d29f29dd24dd6b1e1e5682c302c11c) prior to running a succession of tests to check what BPF objects are supported by the system.

Thanks Quentin.

We could use the same approach in the strace test suite, but unlike
iproute2 and bpftool, strace test suite does not usually have any
privileges to raise its RLIMIT_MEMLOCK above the hard limit
which is set by default to 64K.

You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20191209/66f492db/attachment.html>

More information about the Strace-devel mailing list