lib: Add a GPU error detector

If we listen to the uevents from the kernel, we can detect when the GPU
hangs. This requires us to fork a helper process to do so and send a
signal back to the parent.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

3 files changed, 93 insertions, 2 deletions

diff --git a/lib/Makefile.am b/lib/Makefile.am
index a8a1eb6d..d2f2e16b 100644
--- a/lib/Makefile.am
+++ b/lib/Makefile.am
@@ -15,12 +15,20 @@ if HAVE_VC4
 endif
 
 AM_CPPFLAGS = -I$(top_srcdir)
-AM_CFLAGS = $(DRM_CFLAGS) $(CWARNFLAGS) $(LIBUNWIND_CFLAGS) $(DEBUG_CFLAGS) \
+AM_CFLAGS = $(CWARNFLAGS) $(DRM_CFLAGS) $(PCIACCESS_CFLAGS) $(LIBUNWIND_CFLAGS) $(DEBUG_CFLAGS) \
 	    -DIGT_SRCDIR=\""$(abs_top_srcdir)/tests"\" \
 	    -DIGT_DATADIR=\""$(pkgdatadir)"\" \
 	    -DIGT_LOG_DOMAIN=\""$(subst _,-,$*)"\" \
 	    -pthread
 
-LDADD = $(CAIRO_LIBS) $(LIBUNWIND_LIBS) $(TIMER_LIBS) -lm
 AM_CFLAGS += $(CAIRO_CFLAGS)
 
+libintel_tools_la_LIBADD = \
+	$(DRM_LIBS) \
+	$(PCIACCESS_LIBS) \
+	$(CAIRO_LIBS) \
+	$(LIBUDEV_LIBS) \
+	$(LIBUNWIND_LIBS) \
+	$(TIMER_LIBS) \
+	-lm
+
diff --git a/lib/igt_aux.c b/lib/igt_aux.c
index b4c301e5..bfeaa168 100644
--- a/lib/igt_aux.c
+++ b/lib/igt_aux.c
@@ -42,6 +42,7 @@
 #include <stdlib.h>
 #include <time.h>
 #include <unistd.h>
+#include <sys/poll.h>
 #include <sys/wait.h>
 #include <sys/time.h>
 #include <sys/types.h>
@@ -359,6 +360,85 @@ void igt_stop_signal_helper(void)
 	sig_stat = 0;
 }
 
+#if HAVE_UDEV
+#include <libudev.h>
+
+static struct igt_helper_process hang_detector;
+static void __attribute__((noreturn))
+hang_detector_process(pid_t pid, dev_t rdev)
+{
+	struct udev_monitor *mon =
+		udev_monitor_new_from_netlink(udev_new(), "kernel");
+	struct pollfd pfd;
+
+	udev_monitor_filter_add_match_subsystem_devtype(mon, "drm", NULL);
+	udev_monitor_enable_receiving(mon);
+
+	pfd.fd = udev_monitor_get_fd(mon);
+	pfd.events = POLLIN;
+
+	while (poll(&pfd, 1, -1) > 0) {
+		struct udev_device *dev = udev_monitor_receive_device(mon);
+		dev_t devnum;
+
+		if (dev == NULL)
+			break;
+
+		devnum = udev_device_get_devnum(dev);
+		if (memcmp(&rdev, &devnum, sizeof(dev_t)) == 0) {
+			const char *str;
+
+			str = udev_device_get_property_value(dev, "ERROR");
+			if (str && atoi(str) == 1)
+				kill(pid, SIGRTMAX);
+		}
+
+		udev_device_unref(dev);
+		if (kill(pid, 0)) /* Parent has died, so must we. */
+			break;
+	}
+
+	exit(0);
+}
+
+static void sig_abort(int sig)
+{
+	igt_assert(!"GPU hung");
+}
+
+void igt_fork_hang_detector(int fd)
+{
+	struct stat st;
+
+	if (igt_only_list_subtests())
+		return;
+
+	igt_assert(fstat(fd, &st) == 0);
+
+	signal(SIGRTMAX, sig_abort);
+	igt_fork_helper(&hang_detector)
+		hang_detector_process(getppid(), st.st_rdev);
+}
+
+void igt_stop_hang_detector(void)
+{
+	if (igt_only_list_subtests())
+		return;
+
+	igt_stop_helper(&hang_detector);
+}
+#else
+void igt_fork_hang_detector(int fd)
+{
+	if (igt_only_list_subtests())
+		return;
+}
+
+void igt_stop_hang_detector(void)
+{
+}
+#endif
+
 /**
  * igt_check_boolean_env_var:
  * @env_var: environment variable name
diff --git a/lib/igt_aux.h b/lib/igt_aux.h
index 101fad12..cdaed297 100644
--- a/lib/igt_aux.h
+++ b/lib/igt_aux.h
@@ -40,6 +40,9 @@ extern int num_trash_bos;
 void igt_fork_signal_helper(void);
 void igt_stop_signal_helper(void);
 
+void igt_fork_hang_detector(int fd);
+void igt_stop_hang_detector(void);
+
 struct igt_sigiter {
 	unsigned pass;
 };