On Fri, Jan 23, 2026 at 8:16 AM Michal Luczaj <[email protected]> wrote: > > Attempt to trigger warnings/crashes by racing connect() against sockmap > updates and signals. > > Follow-up to the discussion regarding af_vsock connect(): > https://lore.kernel.org/netdev/[email protected]/ > > Suggested-by: John Fastabend <[email protected]> > Signed-off-by: Michal Luczaj <[email protected]> > --- > .../bpf/prog_tests/sockmap_interrupted_connect.c | 200 > +++++++++++++++++++++ > 1 file changed, 200 insertions(+)
Not sure why, but it seems to be causing reliable failures in what should have been an unrelated test. See bpf CI. I suspect there is a cleanup issue with this new test. Pls investigate. pw-bot: cr
