Hello Willem, On Mon, Jul 07, 2025 at 10:56:32PM -0400, Willem de Bruijn wrote: > > +def test_netpoll(cfg: NetDrvEpEnv) -> None: <snip> > > + bpftrace_stop() > > One risk with stateful tests is that the state is not reset if the > test exists (or crashes) before reaching the cleanup logic. There > are ways around it. Jakub added defer for this purpose, for one. Agree. Jakub suggested about "defer" a while ago, but I didn't realize we had a "defer" helper in our test framework. For a second I though Jakub was talking calling the stop later, my bad. Now that you raised it, I found that we have "defer" as a helper, and we definitely should use it. How about something as simple as: diff --git a/tools/testing/selftests/drivers/net/netpoll_basic.py b/tools/testing/selftests/drivers/net/netpoll_basic.py index 398ac959151b3..6017b71f154b2 100755 --- a/tools/testing/selftests/drivers/net/netpoll_basic.py +++ b/tools/testing/selftests/drivers/net/netpoll_basic.py @@ -27,6 +27,7 @@ from typing import Optional from lib.py import ( bpftrace, + defer, ip, ethtool, GenerateTraffic, @@ -251,6 +252,7 @@ def do_netpoll_flush_monitored(cfg: NetDrvEpEnv, ifname: str, target_name: str) # Start bpftrace in parallel, so, it is watching # netpoll_poll_dev() while we are sending netconsole messages bpftrace_start() + defer(bpftrace_stop) do_netpoll_flush(cfg, ifname, target_name) @@ -338,7 +340,6 @@ def test_netpoll(cfg: NetDrvEpEnv) -> None: # Revert RX/TX queues ethtool_set_rx_tx_queue(ifname, original_queues[0], original_queues[1]) netcons_delete_target(target_name) Thanks for th heads-up and review, --breno
