[RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Dmitry Torokhov]

Hi,

I am asking everyone to give this patch a try. It is reported to speed
up suspend process, however there is a risk that there are boxes out
there that do require full reset and I would prefer to know that sooner
than later.

Thanks!

-- 
Dmitry


Input: speed up suspend/shutdown for PS/2 mice and keyboards

Instead of doing full-blown reset while suspending or shutting down
the box use lighter form of reset that should take less time.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
---

 drivers/input/keyboard/atkbd.c     |    5 +++--
 drivers/input/mouse/psmouse-base.c |    5 ++++-
 2 files changed, 7 insertions(+), 3 deletions(-)


diff --git a/drivers/input/keyboard/atkbd.c b/drivers/input/keyboard/atkbd.c
index a357357..1f5e2ce 100644
--- a/drivers/input/keyboard/atkbd.c
+++ b/drivers/input/keyboard/atkbd.c
@@ -134,7 +134,8 @@ static const unsigned short atkbd_unxlate_table[128] = {
 #define ATKBD_CMD_GETID		0x02f2
 #define ATKBD_CMD_SETREP	0x10f3
 #define ATKBD_CMD_ENABLE	0x00f4
-#define ATKBD_CMD_RESET_DIS	0x00f5
+#define ATKBD_CMD_RESET_DIS	0x00f5	/* Reset to defaults and disable */
+#define ATKBD_CMD_RESET_DEF	0x00f6	/* Reset to defaults */
 #define ATKBD_CMD_SETALL_MBR	0x00fa
 #define ATKBD_CMD_RESET_BAT	0x02ff
 #define ATKBD_CMD_RESEND	0x00fe
@@ -836,7 +837,7 @@ static void atkbd_cleanup(struct serio *serio)
 	struct atkbd *atkbd = serio_get_drvdata(serio);
 
 	atkbd_disable(atkbd);
-	ps2_command(&atkbd->ps2dev, NULL, ATKBD_CMD_RESET_BAT);
+	ps2_command(&atkbd->ps2dev, NULL, ATKBD_CMD_RESET_DEF);
 }
 
 
diff --git a/drivers/input/mouse/psmouse-base.c b/drivers/input/mouse/psmouse-base.c
index fd0bc09..401ac6b 100644
--- a/drivers/input/mouse/psmouse-base.c
+++ b/drivers/input/mouse/psmouse-base.c
@@ -1137,7 +1137,10 @@ static void psmouse_cleanup(struct serio *serio)
 	if (psmouse->cleanup)
 		psmouse->cleanup(psmouse);
 
-	psmouse_reset(psmouse);
+/*
+ * Reset the mouse to defaults (bare PS/2 protocol).
+ */
+	ps2_command(&psmouse->ps2dev, NULL, PSMOUSE_CMD_RESET_DIS);
 
 /*
  * Some boxes, such as HP nx7400, get terribly confused if mouse

Re: [linux-pm] [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Alan Stern]

On Sun, 20 Dec 2009, Dmitry Torokhov wrote:

> Hi,
> 
> I am asking everyone to give this patch a try. It is reported to speed
> up suspend process, however there is a risk that there are boxes out
> there that do require full reset and I would prefer to know that sooner
> than later.

It works okay on my machine (Intel ICH4).

Alan Stern

Re: [linux-pm] [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Rafael J. Wysocki]

On Monday 21 December 2009, Alan Stern wrote:
> On Sun, 20 Dec 2009, Dmitry Torokhov wrote:
> 
> > Hi,
> > 
> > I am asking everyone to give this patch a try. It is reported to speed
> > up suspend process, however there is a risk that there are boxes out
> > there that do require full reset and I would prefer to know that sooner
> > than later.
> 
> It works okay on my machine (Intel ICH4).

Likewise on my Toshiba Portege R500 (ICH7).

[Of course, it also works on my other test boxes, the Wind has ICH7 like the
Toshiba, and the nx6325 has ATI SB400).

Rafael

Re: [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Alexander Clouter]

Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote:
> 
> I am asking everyone to give this patch a try. It is reported to speed
> up suspend process, however there is a risk that there are boxes out
> there that do require full reset and I would prefer to know that sooner
> than later.
> 
Seems to work on my ICH8 lapdog (a Fujitsu T2010), can't say it actually 
speeds the suspend/resume cycle up though.

Cheers

-- 
Alexander Clouter
.sigmonster says: Bus error -- driver executed.

Re: [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Rafael J. Wysocki]

On Saturday 26 December 2009, Alexander Clouter wrote:
> Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote:
> > 
> > I am asking everyone to give this patch a try. It is reported to speed
> > up suspend process, however there is a risk that there are boxes out
> > there that do require full reset and I would prefer to know that sooner
> > than later.
> > 
> Seems to work on my ICH8 lapdog (a Fujitsu T2010), can't say it actually 
> speeds the suspend/resume cycle up though.

With 2.6.33-rc2, if you do

dmesg | grep "of devices"

that will give you total device suspend/resume times.  By comparing the times
with and without the patch you should be able to tell the difference (if any).

Rafael

Re: [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Henrique de Moraes Holschuh]

Works on my ThinkPad T43 (KBC emulated by the embedded controller firmware),
although I don't think it made much difference in speed.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Dmitry Torokhov]

On Sat, Dec 26, 2009 at 10:26:58PM -0200, Henrique de Moraes Holschuh wrote:
> Works on my ThinkPad T43 (KBC emulated by the embedded controller firmware),
> although I don't think it made much difference in speed.
> 

Thanks. The speedup should not be that noticeable (even worst hardware
usually manages to do full reset in 0.5 sec), still 0.5 sec here and 0.5
sec there may add up.

-- 
Dmitry

Re: [RFT] Input: speed up suspend/shutdown for PS/2 mice and keyboards

[Andreas Mohr]

Hi,

On an ATI SB600 (RS690), no problems either on mouse/kbd after a S2R cycle,
on current git with this patch applied (verified).
And this with a rather problematic "new-fangled, designer (tm)"
Logitech Wireless Keyboard M/N Y-RR54 , P/N 867512-0102 (plus mouse!),
plugged into PS/2 connectors!

Thanks for your work!

Andreas Mohr