fujitsu-laptop: Use RFKILL support bitmask from firmware
Subsystems (3)
PATCH 1/1 fujitsu-laptop: Use RFKILL support bitmask from firmware
Up until now, we polled the rfkill status for every incoming FUJ02E3 ACPI event. It turns out that the firmware has a bitmask which indicates what rfkill-related state it can report.
The rfkill_supported bitmask is now used to avoid polling for rfkill at all in the notification handler if there is no support. Also, it is used in the platform device callbacks. As before we register all callbacks and report "unknown" if the firmware does not give us status updates for that particular bit.
This was fed through checkpatch.pl and tested on the S6420 platform. Jonathan, your platform is of particular importance as you are the only tester on a platform with no S000 function in the DSDT. Please confirm you still get a unknown status in the platform files. With debugging on full, you should see the 0x1000 0x4 0x0 0x0 call disappear from your trace entirely.
drivers/platform/x86/fujitsu-laptop.c renamed from drivers/platform/x86/fujitsu-laptop.c.orig +18 -7
@@ -166,6 +166,7 @@struct platform_device *pf_device;struct kfifo *fifo;spinlock_t fifo_lock; +int rfkill_supported;int rfkill_state;int logolamp_registered;int kblamps_registered;
@@ -526,7 +527,7 @@ show_lid_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x100)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x100) return sprintf(buf, "open\n");
@@ -538,7 +539,7 @@ show_dock_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x200)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x200) return sprintf(buf, "docked\n");
@@ -550,7 +551,7 @@ show_radios_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x20)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x20) return sprintf(buf, "on\n");
@@ -928,8 +929,17 @@ ;/* No action, result is discarded */ vdbg_printk (FUJLAPTOP_DBG_INFO ,"Discarded %i ringbuffer entries\n" , i); - fujitsu_hotkey->rfkill_state = -call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 ); + fujitsu_hotkey->rfkill_supported = +call_fext_func (FUNC_RFKILL ,0x0 ,0x0 ,0x0 ); + +/* Make sure our bitmask of supported functions is cleared if the +RFKILL function block is not implemented, like on the S7020. */ +if (fujitsu_hotkey->rfkill_supported ==UNSUPPORTED_CMD ) + fujitsu_hotkey->rfkill_supported =0 ; + +if (fujitsu_hotkey->rfkill_supported) + fujitsu_hotkey->rfkill_state = +call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 );/* Suspect this is a keymap of the application panel, print it */ printk(KERN_INFO "fujitsu-laptop: BTNI: [0x%x]\n",
@@ -1005,8 +1015,9 @@ input = fujitsu_hotkey->input; - fujitsu_hotkey->rfkill_state = -call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 ); +if (fujitsu_hotkey->rfkill_supported) + fujitsu_hotkey->rfkill_state = +call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 );switch (event) {case ACPI_FUJITSU_NOTIFY_CODE1 :
The /sys/devices/platform/fujitsu-laptop/{dock,radios,lid} files still work on my P8010. I applied this patch to 2.6.29-rc2.
Hi Tony
Quoted 8 lines
Up until now, we polled the rfkill status for every incoming FUJ02E3 ACPI event. It turns out that the firmware has a bitmask which indicates what rfkill-related state it can report.
Therfkill_supportedbitmask is now used to avoid polling for rfkill at all in the notification handler if there is no support. Also, it is used in the platform device callbacks. As before we register all callbacks and report "unknown" if the firmware does not give us status updates for that particular bit.
Quoted 5 lines
This was fed through checkpatch.pl and tested on the S6420 platform. Jonathan, your platform is of particular importance as you are the only tester on a platform with no S000 function in the DSDT. Please confirm you still get a unknown status in the platform files. With debugging on full, you should see the
0x10000x40x00x0call disappear from your trace entirely.
I'll check this as soon as I can. I'm at Linux.conf.au this week though so I'm not sure I'll be able to. We'll see.
What kernel is the patch against?
Regards
jonathan
Quoted 1 line · Jonathan Woithe
What kernel is the patch against?
2.6.29-rc2 (as it assumes the move from misc to x86/platform has been made). Fairly sure it would apply to -rc1 as well, should that be convenient.
Quoted 2 lines
Regards
jonathan
Regards,
Tony V.
Re: [PATCH RFT] fujitsu-laptop: Use RFKILL support bitmask from
Hi Tony
Quoted 6 lines · Jonathan Woithe
Quoted 1 line
What kernel is the patch against?
2.6.29-rc2 (as it assumes the move from misc to x86/platform has been made). Fairly sure it would apply to -rc1 as well, should that be convenient.
Ah, they resolved the new location issue then? A lot happens in a week. :-) Thanks for the heads-up.
Regards
jonathan
Hi Tony
Quoted 5 lines
Jonathan, your platform is of particular importance as you are the only tester on a platform with no S000 function in the DSDT. Please confirm you still get a unknown status in the platform files. With debugging on full, you should see the
0x10000x40x00x0call disappear from your trace entirely.
I can confirm that this patch does not break the S7020 - I still get "unknown" in the status platform files, as expected.
Patch reproduced below for convenience.
Regards
jonathan
Up until now, we polled the rfkill status for every incoming FUJ02E3 ACPI event. It turns out that the firmware has a bitmask which indicates what rfkill-related state it can report.
The rfkill_supported bitmask is now used to avoid polling for rfkill at all in the notification handler if there is no support. Also, it is used in the platform device callbacks. As before we register all callbacks and report "unknown" if the firmware does not give us status updates for that particular bit.
This was fed through checkpatch.pl and tested on the S6420 platform. Jonathan, your platform is of particular importance as you are the only tester on a platform with no S000 function in the DSDT. Please confirm you still get a unknown status in the platform files. With debugging on full, you should see the 0x1000 0x4 0x0 0x0 call disappear from your trace entirely.
drivers/platform/x86/fujitsu-laptop.c renamed from drivers/platform/x86/fujitsu-laptop.c.orig +18 -7
@@ -166,6 +166,7 @@struct platform_device *pf_device;struct kfifo *fifo;spinlock_t fifo_lock; +int rfkill_supported;int rfkill_state;int logolamp_registered;int kblamps_registered;
@@ -526,7 +527,7 @@ show_lid_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x100)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x100) return sprintf(buf, "open\n");
@@ -538,7 +539,7 @@ show_dock_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x200)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x200) return sprintf(buf, "docked\n");
@@ -550,7 +551,7 @@ show_radios_state(struct device *dev, struct device_attribute *attr, char *buf) { - if (fujitsu_hotkey->rfkill_state == UNSUPPORTED_CMD) + if (!(fujitsu_hotkey->rfkill_supported & 0x20)) return sprintf(buf, "unknown\n"); if (fujitsu_hotkey->rfkill_state & 0x20) return sprintf(buf, "on\n");
@@ -928,8 +929,17 @@ ;/* No action, result is discarded */ vdbg_printk (FUJLAPTOP_DBG_INFO ,"Discarded %i ringbuffer entries\n" , i); - fujitsu_hotkey->rfkill_state = -call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 ); + fujitsu_hotkey->rfkill_supported = +call_fext_func (FUNC_RFKILL ,0x0 ,0x0 ,0x0 ); + +/* Make sure our bitmask of supported functions is cleared if the +RFKILL function block is not implemented, like on the S7020. */ +if (fujitsu_hotkey->rfkill_supported ==UNSUPPORTED_CMD ) + fujitsu_hotkey->rfkill_supported =0 ; + +if (fujitsu_hotkey->rfkill_supported) + fujitsu_hotkey->rfkill_state = +call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 );/* Suspect this is a keymap of the application panel, print it */ printk(KERN_INFO "fujitsu-laptop: BTNI: [0x%x]\n",
@@ -1005,8 +1015,9 @@ input = fujitsu_hotkey->input; - fujitsu_hotkey->rfkill_state = -call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 ); +if (fujitsu_hotkey->rfkill_supported) + fujitsu_hotkey->rfkill_state = +call_fext_func (FUNC_RFKILL ,0x4 ,0x0 ,0x0 );switch (event) {case ACPI_FUJITSU_NOTIFY_CODE1 :