Hibernation test

Hibernation test

[David Weinehall]

[-- Attachment #1: Type: text/plain, Size: 1380 bytes --]

intel-gpu-tools currently has a bunch of tests for suspend,
but currently none (that I could find) for hibernate.

Attached is a rudimentary patch to add said test.  It does so
by repurposing the drv_suspend driver to handle both suspend
and hibernate, since the difference is miniscule.

I decided to split the suspend/autoresume functions in
igt_aux.c though, to be able to leave the igt_system_uspend_autoresume()
function unchanged (the other option would be to
introduce a boolean function argument and have that
decide what parameters to pass to rtcwake).

The timeout passed to rtcwake probably needs tuning (it might
even need to be dynamically adjusted, since the time hibernation takes
varies wildly depending on the amount of non-cache memory in use).


Kind regards, David Weinehall


PS: Go easy on me, mkay -- first patch to intel-gfx :P (meh, who am I 
kidding,
I should be flame tolerant enough, bring it on).
---------------------------------------------------------------------
Intel Finland Oy
Registered Address: PL 281, 00181 Helsinki 
Business Identity Code: 0357606 - 4 
Domiciled in Helsinki 

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: igt-hib.patch --]
[-- Type: text/x-patch; name="igt-hib.patch", Size: 6544 bytes --]

diff --git a/lib/igt_aux.c b/lib/igt_aux.c
index aefa0863e9e9..c39ccd86db63 100644
--- a/lib/igt_aux.c
+++ b/lib/igt_aux.c
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2007, 2011, 2013, 2014 Intel Corporation
+ * Copyright © 2007, 2011, 2013, 2014, 2015 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -411,7 +411,7 @@ void igt_cleanup_aperture_trashers(void)
  * igt_system_suspend_autoresume:
  *
  * Execute a system suspend-to-mem cycle and automatically wake up again using
- * the firmwares resume timer.
+ * the firmware's resume timer.
  *
  * This is very handy for implementing any kind of suspend/resume test.
  */
@@ -430,6 +430,34 @@ void igt_system_suspend_autoresume(void)
 }
 
 /**
+ * igt_system_hibernate_autoresume:
+ *
+ * Execute a system suspend-to-disk cycle and automatically wake up again using
+ * the firmware's resume timer.
+ *
+ * This is very handy for implementing any kind of hibernate/resume test.
+ */
+void igt_system_hibernate_autoresume(void)
+{
+	int ret;
+
+	/* FIXME: I'm guessing simulation behaves the same way as with
+	 * suspend/resume, but it might be prudent to make sure
+	 */
+	/* FIXME: Simulation doesn't like suspend/resume, and not even a lighter
+	 * approach using /sys/power/pm_test to just test our driver's callbacks
+	 * seems to fare better. We need to investigate what's going on. */
+	igt_skip_on_simulation();
+
+	/* The timeout might need to be adjusted if hibernation takes too long
+	 * or if we have to wait excessively long before resume
+	 */
+	ret = system("rtcwake -s 90 -m disk");
+	igt_assert_f(ret == 0, "This failure means that something is wrong with the rtcwake tool "
+		     "or how your distro is set up. This is not a i915.ko or i-g-t bug.");
+}
+
+/**
  * igt_drop_root:
  *
  * Drop root privileges and make sure it actually worked. Useful for tests
diff --git a/lib/igt_aux.h b/lib/igt_aux.h
index 798a5b45fcb9..636ab3dc06fc 100644
--- a/lib/igt_aux.h
+++ b/lib/igt_aux.h
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2014 Intel Corporation
+ * Copyright © 2014, 2015 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -58,8 +58,9 @@ void igt_init_aperture_trashers(drm_intel_bufmgr *bufmgr);
 void igt_trash_aperture(void);
 void igt_cleanup_aperture_trashers(void);
 
-/* suspend and auto-resume system */
+/* suspend/hibernate and auto-resume system */
 void igt_system_suspend_autoresume(void);
+void igt_system_hibernate_autoresume(void);
 
 /* dropping priviledges */
 void igt_drop_root(void);
diff --git a/tests/drv_suspend.c b/tests/drv_suspend.c
index 80f6a2254b75..8165cef9b333 100644
--- a/tests/drv_suspend.c
+++ b/tests/drv_suspend.c
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2013 Intel Corporation
+ * Copyright © 2013, 2015 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,6 +22,7 @@
  *
  * Authors:
  *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ *    David Weinehall <david.weinehall@intel.com>
  *
  */
 
@@ -45,7 +46,7 @@
 #define OBJECT_SIZE (16*1024*1024)
 
 static void
-test_fence_restore(int fd, bool tiled2untiled)
+test_fence_restore(int fd, bool tiled2untiled, bool hibernate)
 {
 	uint32_t handle1, handle2, handle_tiled;
 	uint32_t *ptr1, *ptr2, *ptr_tiled;
@@ -80,7 +81,10 @@ test_fence_restore(int fd, bool tiled2untiled)
 	else
 		gem_set_tiling(fd, handle_tiled, I915_TILING_X, 2048);
 
-	igt_system_suspend_autoresume();
+	if (hibernate)
+		igt_system_hibernate_autoresume();
+	else
+		igt_system_suspend_autoresume();
 
 	igt_info("checking the first canary object\n");
 	for (i = 0; i < OBJECT_SIZE/sizeof(uint32_t); i++)
@@ -100,7 +104,7 @@ test_fence_restore(int fd, bool tiled2untiled)
 }
 
 static void
-test_debugfs_reader(void)
+test_debugfs_reader(bool hibernate)
 {
 	struct igt_helper_process reader = {};
 	reader.use_SIGKILL = true;
@@ -117,7 +121,10 @@ test_debugfs_reader(void)
 
 	sleep(1);
 
-	igt_system_suspend_autoresume();
+	if (hibernate)
+		igt_system_hibernate_autoresume();
+	else
+		igt_system_suspend_autoresume();
 
 	sleep(1);
 
@@ -125,7 +132,7 @@ test_debugfs_reader(void)
 }
 
 static void
-test_sysfs_reader(void)
+test_sysfs_reader(bool hibernate)
 {
 	struct igt_helper_process reader = {};
 	reader.use_SIGKILL = true;
@@ -142,7 +149,10 @@ test_sysfs_reader(void)
 
 	sleep(1);
 
-	igt_system_suspend_autoresume();
+	if (hibernate)
+		igt_system_hibernate_autoresume();
+	else
+		igt_system_suspend_autoresume();
 
 	sleep(1);
 
@@ -150,13 +160,18 @@ test_sysfs_reader(void)
 }
 
 static void
-test_forcewake(void)
+test_forcewake(bool hibernate)
 {
 	int fw_fd;
 
 	fw_fd = igt_open_forcewake_handle();
 	igt_assert(fw_fd >= 0);
-	igt_system_suspend_autoresume();
+
+	if (hibernate)
+		igt_system_hibernate_autoresume();
+	else
+		igt_system_suspend_autoresume();
+
 	close (fw_fd);
 }
 
@@ -169,20 +184,35 @@ igt_main
 	igt_fixture
 		fd = drm_open_any();
 
-	igt_subtest("fence-restore-tiled2untiled")
-		test_fence_restore(fd, true);
+	igt_subtest("fence-restore-tiled2untiled-suspend")
+		test_fence_restore(fd, true, false);
+
+	igt_subtest("fence-restore-untiled-suspend")
+		test_fence_restore(fd, false, false);
+
+	igt_subtest("debugfs-reader-suspend")
+		test_debugfs_reader(false);
+
+	igt_subtest("sysfs-reader-suspend")
+		test_sysfs_reader(false);
+
+	igt_subtest("forcewake-suspend")
+		test_forcewake(false);
+
+	igt_subtest("fence-restore-tiled2untiled-hibernate")
+		test_fence_restore(fd, true, true);
 
-	igt_subtest("fence-restore-untiled")
-		test_fence_restore(fd, false);
+	igt_subtest("fence-restore-untiled-hibernate")
+		test_fence_restore(fd, false, true);
 
-	igt_subtest("debugfs-reader")
-		test_debugfs_reader();
+	igt_subtest("debugfs-reader-hibernate")
+		test_debugfs_reader(true);
 
-	igt_subtest("sysfs-reader")
-		test_sysfs_reader();
+	igt_subtest("sysfs-reader-hibernate")
+		test_sysfs_reader(true);
 
-	igt_subtest("forcewake")
-		test_forcewake();
+	igt_subtest("forcewake-hibernate")
+		test_forcewake(true);
 
 	igt_fixture
 		close(fd);

[-- Attachment #3: Type: text/plain, Size: 159 bytes --]

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Hibernation test

[Imre Deak]

On ke, 2015-02-11 at 16:46 +0200, David Weinehall wrote:
> intel-gpu-tools currently has a bunch of tests for suspend,
> but currently none (that I could find) for hibernate.
> 
> Attached is a rudimentary patch to add said test.  It does so
> by repurposing the drv_suspend driver to handle both suspend
> and hibernate, since the difference is miniscule.
> 
> I decided to split the suspend/autoresume functions in
> igt_aux.c though, to be able to leave the igt_system_uspend_autoresume()
> function unchanged (the other option would be to
> introduce a boolean function argument and have that
> decide what parameters to pass to rtcwake).
> 
> The timeout passed to rtcwake probably needs tuning (it might
> even need to be dynamically adjusted, since the time hibernation takes
> varies wildly depending on the amount of non-cache memory in use).

I think in general we should try to keep the existing subtest names, but
I couldn't find any concrete problem changing them in this case. The
patch looks ok to me (please inline the patch next time):
Reviewed-by: Imre Deak <imre.deak@intel.com>


_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Hibernation test

[Imre Deak]

On ke, 2015-02-18 at 18:43 +0200, Imre Deak wrote:
> On ke, 2015-02-11 at 16:46 +0200, David Weinehall wrote:
> > intel-gpu-tools currently has a bunch of tests for suspend,
> > but currently none (that I could find) for hibernate.
> > 
> > Attached is a rudimentary patch to add said test.  It does so
> > by repurposing the drv_suspend driver to handle both suspend
> > and hibernate, since the difference is miniscule.
> > 
> > I decided to split the suspend/autoresume functions in
> > igt_aux.c though, to be able to leave the igt_system_uspend_autoresume()
> > function unchanged (the other option would be to
> > introduce a boolean function argument and have that
> > decide what parameters to pass to rtcwake).
> > 
> > The timeout passed to rtcwake probably needs tuning (it might
> > even need to be dynamically adjusted, since the time hibernation takes
> > varies wildly depending on the amount of non-cache memory in use).
> 
> I think in general we should try to keep the existing subtest names, but
> I couldn't find any concrete problem changing them in this case. The
> patch looks ok to me (please inline the patch next time):
> Reviewed-by: Imre Deak <imre.deak@intel.com>

Thanks for the patch, I pushed it to igt.

--Imre

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx