On Fri, 29 Jan 2021 at 11:00, Peter Maydell <peter.mayd...@linaro.org> wrote: > > From: Mihai Carabas <mihai.cara...@oracle.com> > > Add a test case for pvpanic-pci device. The scenario is the same as pvpanic > ISA device, but is using the PCI bus. > > Signed-off-by: Mihai Carabas <mihai.cara...@oracle.com> > Acked-by: Thomas Huth <th...@redhat.com> > Reviewed-by: Peter Maydell <peter.mayd...@linaro.org> > Signed-off-by: Mihai Carabas <mihai.cara...@oracle.com> > Signed-off-by: Peter Maydell <peter.mayd...@linaro.org>
This turns out to trigger the oss-fuzz build's leak-sanitizer: https://gitlab.com/qemu-project/qemu/-/jobs/995177788 > +static void test_panic_nopause(void) > +{ > + uint8_t val; > + QDict *response, *data; > + QTestState *qts; > + QPCIBus *pcibus; > + QPCIDevice *dev; > + QPCIBar bar; > + > + qts = qtest_init("-device pvpanic-pci,addr=04.0 -action panic=none"); > + pcibus = qpci_new_pc(qts, NULL); > + dev = qpci_device_find(pcibus, QPCI_DEVFN(0x4, 0x0)); > + qpci_device_enable(dev); > + bar = qpci_iomap(dev, 0, NULL); > + > + qpci_memread(dev, bar, 0, &val, sizeof(val)); > + g_assert_cmpuint(val, ==, 3); > + > + val = 1; > + qpci_memwrite(dev, bar, 0, &val, sizeof(val)); > + > + response = qtest_qmp_eventwait_ref(qts, "GUEST_PANICKED"); > + g_assert(qdict_haskey(response, "data")); > + data = qdict_get_qdict(response, "data"); > + g_assert(qdict_haskey(data, "action")); > + g_assert_cmpstr(qdict_get_str(data, "action"), ==, "run"); > + qobject_unref(response); > + I'm going to squash in the trivial fix, which is to add g_free(dev); qpci_free_pc(pcibus); here and similarly in the test_panic() function. > + qtest_quit(qts); > + } thanks -- PMM
