Skip to contentSkip to conversation

RTC class uses subsys_init

PATCH In review David Brownell 4 participants 6 messages Sep 29, 2006 – Oct 1, 2006 Real time clock (RTC)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 RTC class uses subsys_init

[RESEND]

This makes RTC core components use "subsys_init" instead of "module_init", as appropriate for subsystem infrastructure. This is mostly useful for statically linking drivers in other parts of the tree that may provide an RTC interface as a secondary functionality (e.g. part of a multifunction chip); they won't need to worry so much about drivers/Makefile link order.

drivers/rtc/class.c +2 -2
Index: linux/drivers/rtc/class.c
===================================================================
@@ -142,9 +142,9 @@
 	class_destroy(rtc_class);
 }
 
-module_init(rtc_init);
+subsys_initcall(rtc_init);
 module_exit(rtc_exit);
 
-MODULE_AUTHOR("Alessandro Zummo <[email protected]>");
+MODULE_AUTHOR("Alessandro Zummo <[email protected]>");
 MODULE_DESCRIPTION("RTC class support");
 MODULE_LICENSE("GPL");
drivers/rtc/rtc-dev.c +1 -1
Index: linux/drivers/rtc/rtc-dev.c
===================================================================
@@ -496,7 +496,7 @@
 	unregister_chrdev_region(rtc_devt, RTC_DEV_MAX);
 }
 
-module_init(rtc_dev_init);
+subsys_initcall(rtc_dev_init);
 module_exit(rtc_dev_exit);
 
 MODULE_AUTHOR("Alessandro Zummo <[email protected]>");
drivers/rtc/rtc-proc.c +1 -1
Index: linux/drivers/rtc/rtc-proc.c
===================================================================
@@ -156,7 +156,7 @@
 	class_interface_unregister(&rtc_proc_interface);
 }
 
-module_init(rtc_proc_init);
+subsys_initcall(rtc_proc_init);
 module_exit(rtc_proc_exit);
 
 MODULE_AUTHOR("Alessandro Zummo <[email protected]>");
drivers/rtc/rtc-sysfs.c +1 -1
Index: linux/drivers/rtc/rtc-sysfs.c
===================================================================
@@ -116,7 +116,7 @@
 	class_interface_unregister(&rtc_sysfs_interface);
 }
 
-module_init(rtc_sysfs_init);
+subsys_init(rtc_sysfs_init);
 module_exit(rtc_sysfs_exit);
 
 MODULE_AUTHOR("Alessandro Zummo <[email protected]>");
#1 lore raw
Quoted 10 lines · David Brownell

[RESEND]

This makes RTC core components use "subsys_init" instead of "module_init", as appropriate for subsystem infrastructure. This is mostly useful for statically linking drivers in other parts of the tree that may provide an RTC interface as a secondary functionality (e.g. part of a multifunction chip); they won't need to worry so much about drivers/Makefile link order.

ACK

#2 lore raw
In reply to David Brownell

drivers/rtc/rtc-sysfs.c line 120

Quoted 8 lines · David Brownell

+++ linux/drivers/rtc/rtc-sysfs.c 2006-07-30 16:15:50.000000000 -0700

drivers/rtc/rtc-sysfs.c +1 -1
@@ -116,7 +116,7 @@
 	class_interface_unregister(&rtc_sysfs_interface);
 }
 
-module_init(rtc_sysfs_init);
+subsys_init(rtc_sysfs_init);
 module_exit(rtc_sysfs_exit);

subsys_init is not defined, but the change is in Linus tree now.

#3 lore raw
Quoted 8 lines · On Sun, 1 Oct 2006 11:07:17 +0200 Olaf Hering <[email protected]> wrote:
Quoted 5 lines · David Brownell

+++ linux/drivers/rtc/rtc-sysfs.c 2006-07-30 16:15:50.000000000 -0700

drivers/rtc/rtc-sysfs.c +0 -0
@@ -116,7 +116,7 @@
 	class_interface_unregister(&rtc_sysfs_interface);
 }
 
Quoted 5 lines

drivers/rtc/rtc-sysfs.c line 120

Quoted 3 lines
drivers/rtc/rtc-sysfs.c +1 -1
-module_init(rtc_sysfs_init);
+subsys_init(rtc_sysfs_init);
 module_exit(rtc_sysfs_exit);

subsys_init is not defined, but the change is in Linus tree now.

doh. But it still compiled.

drivers/rtc/rtc-sysfs.c:119: warning: data definition has no type or storage class drivers/rtc/rtc-sysfs.c:119: warning: type defaults to 'int' in declaration of 'subsys_init' drivers/rtc/rtc-sysfs.c:119: warning: parameter names (without types) in function declaration drivers/rtc/rtc-sysfs.c:110: warning: 'rtc_sysfs_init' defined but not used

I'll fix it up.

(Wonders how it passed runtime testing..)

#4 lore raw
Quoted 3 lines · Andrew Morton

I'll fix it up.

(Wonders how it passed runtime testing..)

Thanks ... this is a "wrong patch version" got forwarded" issue, sorry.

- Dave

#5 lore raw
In reply to Andrew Morton

Hallo, Andrew

Quoted 2 lines · Andrew Morton

(Wonders how it passed runtime testing..)

Somebody is not going to use and test sysfs.

[ start->control panel->device manager->.... ]
[ /sys/devices/system/...]

But sorry, i will try to not loosing it next time.

--