converting appletouch to usb autosuspend again...
Dear all,
I wonder whether anyone else tried to get appletouch usb autosuspend to work. I've seen Oliver's patch but as it kept oopsing on me and I couldn't parse the usb_mark_last_busy logic I came up with the attached patch against (current git; at least didn't oops for me over the last several days and suspend/resume cycles).
However things I don't understand,
a) is autosuspend on the appletouch driver used at all (how can I find out)
b) do I need dev->intf->needs_remote_wakeup = 1; ?
c) is line 33 in the patch safe to do ?
And while we are at it I am still seeing these X falls back to hid mouse mode, only switching to console and back resolves this (looks like this only worked in previous kernels as resume from s2ram took a lot longer due to the IDE driver doing a couple of resets) - is there anything one could do about it?
Soeren
drivers/input/mouse/appletouch.c +51 -8
@@ -34,6 +34,7 @@#include <linux/slab.h> #include <linux/module.h> #include <linux/usb/input.h> +#include <linux/mutex.h> /* Type of touchpad */ enum atp_touchpad_type {
@@ -140,10 +141,12 @@ MODULE_DEVICE_TABLE(usb, atp_table);struct atp {char phys[64 ];struct usb_device *udev;/* usb device */ +struct usb_interface *intf;/* our interface */ struct urb *urb;/* usb request block */ signed char *data;/* transferred data */ struct input_dev *input;/* input dev */ enum atp_touchpad_type type;/* type of touchpad */ +struct mutex pm_mutex;/* serialize access to open/suspend */ bool open;bool valid;/* are the samples valid? */ bool size_detect_done;
@@ -259,6 +262,11 @@ static void atp_reinit(struct work_struct *work)if (retval)err ("atp_reinit: usb_submit_urb failed with error %d" , retval); + +/* device is idle and available for autosuspend */ +mutex_lock (&dev->pm_mutex); +usb_autopm_disable (dev->intf); +mutex_unlock (&dev->pm_mutex); } static int atp_calculate_abs(int *xy_sensors, int nb_sensors, int fact,
@@ -547,21 +555,37 @@ static void atp_complete(struct urb *urb)static int atp_open (struct input_dev *input) {struct atp *dev =input_get_drvdata (input); +int error =usb_autopm_get_interface (dev->intf); +if (error) +return error; -if (usb_submit_urb (dev->urb,GFP_ATOMIC )) -return -EIO ; +mutex_lock (&dev->pm_mutex); + error=usb_submit_urb (dev->urb,GFP_ATOMIC ); - dev->open =1 ; -return 0 ; +if (!error) + dev->open =1 ; + +mutex_unlock (&dev->pm_mutex); + +if (error) +usb_autopm_put_interface (dev->intf); + +return error; }static void atp_close (struct input_dev *input) {struct atp *dev =input_get_drvdata (input); +mutex_lock (&dev->pm_mutex); +usb_kill_urb (dev->urb);cancel_work_sync (&dev->work); dev->open =0 ; + +mutex_unlock (&dev->pm_mutex); + +usb_autopm_put_interface (dev->intf); }static int atp_handle_geyser (struct atp *dev)
@@ -615,9 +639,12 @@ static int atp_probe(struct usb_interface *iface, } dev->udev = udev; + dev->intf = iface; dev->input = input_dev; dev->type = id->driver_info; dev->overflow_warned = false; +mutex_init (&dev->pm_mutex); +if (dev->type ==ATP_FOUNTAIN || dev->type ==ATP_GEYSER1 ) dev->datalen =81 ;else
@@ -750,27 +777,41 @@ static int atp_suspend(struct usb_interface *iface, pm_message_t message) {struct atp *dev =usb_get_intfdata (iface); +mutex_lock (&dev->pm_mutex); +usb_kill_urb (dev->urb); dev->valid = false; +mutex_unlock (&dev->pm_mutex); +return 0 ; }static int atp_resume (struct usb_interface *iface) {struct atp *dev =usb_get_intfdata (iface); +int error =0 ; -if (dev->open &&usb_submit_urb (dev->urb,GFP_ATOMIC )) -return -EIO ; +mutex_lock (&dev->pm_mutex); -return 0 ; +if (dev->open) + error =usb_submit_urb (dev->urb,GFP_ATOMIC ); + +mutex_unlock (&dev->pm_mutex); + +return error; }static int atp_reset_resume (struct usb_interface *iface) {struct atp *dev =usb_get_intfdata (iface); +int error =0 ; -return atp_recover (dev); +mutex_lock (&dev->pm_mutex); + error =atp_recover (dev); +mutex_unlock (&dev->pm_mutex); + +return error; }static struct usb_driver atp_driver = {
@@ -781,6 +822,7 @@ static struct usb_driver atp_driver = { .resume = atp_resume, .reset_resume = atp_reset_resume, .id_table = atp_table, + .supports_autosuspend = 1, };static int __init atp_init (void )
@@ -795,3 +837,4 @@ static void __exit atp_exit(void)module_init (atp_init);module_exit (atp_exit); +
Am Samstag 23 August 2008 23:00:18 schrieb Soeren Sonnenburg:
Quoted 4 lines
Dear all,
I wonder whether anyone else tried to get appletouch usb autosuspend to work. I've seen Oliver's patch but as it kept oopsing on me and I
You might post the oops. I haven't seen it.
Quoted 8 lines
couldn't parse the
usb_mark_last_busylogic I came up with the attached patch against (current git; at least didn't oops for me over the last several days and suspend/resume cycles).However things I don't understand,
a) is autosuspend on the appletouch driver used at all (how can I find out)
Compile your kernel with CONFIG_USB_DEBUG and autosuspend will be logged. You need to activate autosuspend via sysfs.
Quoted 1 line
b) do I need
dev->intf->needs_remote_wakeup= 1; ?
Yes.
Quoted 1 line
c) is line 33 in the patch safe to do ?
Why do you want to add it?
Quoted 5 lines
And while we are at it I am still seeing these X falls back to hid mouse mode, only switching to console and back resolves this (looks like this only worked in previous kernels as resume from s2ram took a lot longer due to the IDE driver doing a couple of resets) - is there anything one could do about it?
Post a log.
Regards
Oliver
Quoted 7 lines · Oliver Neukum
Am Samstag 23 August 2008 23:00:18 schrieb Soeren Sonnenburg:
Quoted 4 lines
Dear all,
I wonder whether anyone else tried to get appletouch usb autosuspend to work. I've seen Oliver's patch but as it kept oopsing on me and I
You might post the oops. I haven't seen it.
Do you have an updated version of your patch that applies to git-current? I will give your patch another try then.
Quoted 11 lines
Quoted 8 lines
couldn't parse the
usb_mark_last_busylogic I came up with the attached patch against (current git; at least didn't oops for me over the last several days and suspend/resume cycles).However things I don't understand,
a) is autosuspend on the appletouch driver used at all (how can I find out)
Compile your kernel with
CONFIG_USB_DEBUGand autosuspend will be logged. You need to activate autosuspend via sysfs.
OK, indeed it never autosuspends... I guess this is this due to the keyboard and the mice being on the same usb port and as the kbd part does not support autosuspend it never suspends?!
Quoted 7 lines
Quoted 1 line
b) do I need
dev->intf->needs_remote_wakeup= 1; ?Yes.
Quoted 1 line
c) is line 33 in the patch safe to do ?
Why do you want to add it?
Well reaching this line the touchpad is idling (hasn't been touched+no button was pressed) - I somehow wanted to force it to suspend in this case.
Quoted 7 lines
Quoted 5 lines
And while we are at it I am still seeing these X falls back to hid mouse mode, only switching to console and back resolves this (looks like this only worked in previous kernels as resume from s2ram took a lot longer due to the IDE driver doing a couple of resets) - is there anything one could do about it?
Post a log.
My fault. I was still rmmod'ing appletouch before suspending. It works perfectly + reliably the way it is.
Thanks,
Soeren