Skip to contentSkip to conversation

Export symbols so CONFIG_INPUT works as a module

PATCH In review Martin Michlmayr 9 participants 14 messages Jan 24, 2006 – Feb 3, 2006

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Export symbols so CONFIG_INPUT works as a module

Currently, modular input support fails to load with the following error:

qube:# modprobe input
input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

In the short run, this can be solved by exporting these two symbols. There have been discussions about fixing this in a different manner, see http://www.ussg.iu.edu/hypermail/linux/kernel/0505.2/1068.html
Since this was in the days of 2.6.12-rc4 and modular input support is still broken, I suggest these symbols to be exported for now.

---

 drivers/char/random.c |    2 ++
 lib/kobject.c         |    1 +
 2 files changed, 3 insertions(+)
drivers/char/random.c +2 -0
@@ -647,6 +647,8 @@
 			     (type << 4) ^ code ^ (code >> 4) ^ value);
 }
 
+EXPORT_SYMBOL(add_input_randomness);
+
 void add_interrupt_randomness(int irq)
 {
 	if (irq >= NR_IRQS || irq_timer_state[irq] == 0)
lib/kobject.c +1 -0
@@ -527,6 +527,7 @@
 EXPORT_SYMBOL(kobject_init);
 EXPORT_SYMBOL(kobject_register);
 EXPORT_SYMBOL(kobject_unregister);
+EXPORT_SYMBOL(kobject_get_path);
 EXPORT_SYMBOL(kobject_get);
 EXPORT_SYMBOL(kobject_put);
 EXPORT_SYMBOL(kobject_add);
Signature · 2 lines
Martin Michlmayr
http://www.cyrius.com/
#1 lore raw
Quoted 11 lines · Martin Michlmayr

Currently, modular input support fails to load with the following error:

qube:# modprobe input
input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

In the short run, this can be solved by exporting these two symbols. There have been discussions about fixing this in a different manner, see http://www.ussg.iu.edu/hypermail/linux/kernel/0505.2/1068.html
Since this was in the days of 2.6.12-rc4 and modular input support is still broken, I suggest these symbols to be exported for now.

better make these _GPL exports to make sure people understand these are internal things...

#2 lore raw
In reply to Martin Michlmayr
Quoted 11 lines · Martin Michlmayr

Currently, modular input support fails to load with the following error:

qube:# modprobe input
input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

In the short run, this can be solved by exporting these two symbols. There have been discussions about fixing this in a different manner, see http://www.ussg.iu.edu/hypermail/linux/kernel/0505.2/1068.html
Since this was in the days of 2.6.12-rc4 and modular input support is still broken, I suggest these symbols to be exported for now.

Is there actually any practical reason why you would want to make the input layer modular ?

Dave

#3 lore raw
Quoted 7 lines · Dave Jones
Quoted 6 lines · Martin Michlmayr

Currently, modular input support fails to load with the following error:

qube:# modprobe input
input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

Quoted 8 lines
Quoted 5 lines

In the short run, this can be solved by exporting these two symbols. There have been discussions about fixing this in a different manner, see http://www.ussg.iu.edu/hypermail/linux/kernel/0505.2/1068.html
Since this was in the days of 2.6.12-rc4 and modular input support is still broken, I suggest these symbols to be exported for now.

Is there actually any practical reason why you would want to make the input layer modular ?

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

maintainer
#4 lore raw
Quoted 14 lines · Al Viro
Quoted 13 lines · Dave Jones
Quoted 11 lines · Martin Michlmayr

Currently, modular input support fails to load with the following error:

qube:# modprobe input
input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

In the short run, this can be solved by exporting these two symbols. There have been discussions about fixing this in a different manner, see http://www.ussg.iu.edu/hypermail/linux/kernel/0505.2/1068.html
Since this was in the days of 2.6.12-rc4 and modular input support is still broken, I suggest these symbols to be exported for now.

Quoted 5 lines
Quoted 2 lines

Is there actually any practical reason why you would want to make the input layer modular ?

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

--
Dmitry

#5 lore raw
Quoted 6 lines · Dmitry Torokhov
Quoted 2 lines

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

So do you want an updated patch using _GPL to export the symbols or to change CONFIG_INPUT to boolean?

Signature · 2 lines
Martin Michlmayr
http://www.cyrius.com/
#6 lore raw
Quoted 10 lines · Martin Michlmayr
Quoted 6 lines · Dmitry Torokhov
Quoted 2 lines

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

So do you want an updated patch using _GPL to export the symbols or to change CONFIG_INPUT to boolean?

I guess having input core as a module does not make much sense, so we should change CONFIG_INPUT to be boolean _and_ clean up the core code removing module unloading support.

Signature · 1 line
Dmitry
#7 lore raw
Quoted 9 lines · Dmitry Torokhov
Quoted 8 lines · Martin Michlmayr
Quoted 6 lines · Dmitry Torokhov
Quoted 2 lines

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

Quoted 6 lines
Quoted 2 lines

So do you want an updated patch using _GPL to export the symbols or to change CONFIG_INPUT to boolean?

I guess having input core as a module does not make much sense, so we should change CONFIG_INPUT to be boolean _and_ clean up the core code removing module unloading support.

Well, USB or SCSI cores are also modules, so I think there is some point in having that functionality.

What were the required symbols?

Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#8 lore raw
Quoted 4 lines · Vojtech Pavlik

Well, USB or SCSI cores are also modules, so I think there is some point in having that functionality.

What were the required symbols?

input: Unknown symbol kobject_get_path
input: Unknown symbol add_input_randomness

Signature · 2 lines
Martin Michlmayr
http://www.cyrius.com/
#9 lore raw
In reply to Vojtech Pavlik
Quoted 13 lines · Vojtech Pavlik
Quoted 12 lines · Dmitry Torokhov
Quoted 10 lines · Martin Michlmayr
Quoted 6 lines · Dmitry Torokhov
Quoted 2 lines

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

So do you want an updated patch using _GPL to export the symbols or to change CONFIG_INPUT to boolean?

Quoted 7 lines
Quoted 3 lines

I guess having input core as a module does not make much sense, so we should change CONFIG_INPUT to be boolean _and_ clean up the core code removing module unloading support.

Well, USB or SCSI cores are also modules, so I think there is some point in having that functionality.
...

The difference is that USB and SCSI are not that essential, and therefore not always enabled if CONFIG_EMBEDDED=n. It's therefore e.g. not uncommon that distributions offer modular USB and SCSI cores.

Are there really people building kernels for that much space limited environments that they set CONFIG_EMBEDDED=y, and at the same time want CONFIG_INPUT=m?

I'd have expected people using CONFIG_EMBEDDED=y to usually also set CONFIG_MODULES=n for getting a smaller kernel.

Quoted 1 line

Vojtech Pavlik

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#10 lore raw
In reply to Dmitry Torokhov
Quoted 4 lines

I guess having input core as a module does not make much sense, so we should change CONFIG_INPUT to be boolean _and_ clean up the core code removing module unloading support.

Embedded devices with keyboard hotplugging?

Jan Engelhardt

#11 lore raw

PATCH 1/1 make INPUT a bool

In reply to Dmitry Torokhov
Quoted 9 lines · Dmitry Torokhov
Quoted 8 lines · Martin Michlmayr
Quoted 6 lines · Dmitry Torokhov
Quoted 2 lines

More interesting question: is pis^H^H^Hsysfs interaction in there safe for modular code?

The core should be safe, at least I was trying to make it this way, so if you see something wrong - shout. Locking is another question though...

Quoted 6 lines
Quoted 2 lines

So do you want an updated patch using _GPL to export the symbols or to change CONFIG_INPUT to boolean?

I guess having input core as a module does not make much sense, so we should change CONFIG_INPUT to be boolean _and_ clean up the core code removing module unloading support.

Is the patch below what you were thinking of?

Quoted 1 line

Dmitry

cu
Adrian

<-- snip -->

Make INPUT a bool.

INPUT!=y is only possible if EMBEDDED=y, and in such cases it doesn't make that much sense to make it modular.

---

 drivers/input/Kconfig |    2 +-
 drivers/input/input.c |    8 --------
 2 files changed, 1 insertion(+), 9 deletions(-)
drivers/input/Kconfig renamed from drivers/input/Kconfig.old +1 -1
@@ -5,7 +5,7 @@
 menu "Input device support"
 
 config INPUT
-	tristate "Generic input layer (needed for keyboard, mouse, ...)" if EMBEDDED
+	bool "Generic input layer (needed for keyboard, mouse, ...)" if EMBEDDED
 	default y
 	---help---
 	  Say Y here if you have any input device (mouse, keyboard, tablet,
drivers/input/input.c renamed from drivers/input/input.c.old +0 -8
@@ -984,12 +984,4 @@
 	return err;
 }
 
-static void __exit input_exit(void)
-{
-	input_proc_exit();
-	unregister_chrdev(INPUT_MAJOR, "input");
-	class_unregister(&input_class);
-}
-
 subsys_initcall(input_init);
-module_exit(input_exit);
#12 lore raw
In reply to Dave Jones
Quoted 2 lines · Dave Jones

Is there actually any practical reason why you would want to make the input layer modular ?

Not really, but if it doesn't work as a module then Kbuild shouldn't allow you to configure it like that.

Signature · 2 lines
Martin Michlmayr
http://www.cyrius.com/
#13 lore raw
Quoted 6 lines · Martin Michlmayr
Quoted 2 lines · Dave Jones

Is there actually any practical reason why you would want to make the input layer modular ?

Not really, but if it doesn't work as a module then Kbuild shouldn't allow you to configure it like that.

That's the point I'm trying to make. If there's no good reason, it may as well be boolean.

Dave