sgiioc4: Disable module unload

PATCH In review Jeremy Higdon 3 participants 5 messages Nov 21, 2006 – Nov 22, 2006 libata

lore

Share & tools

lore

Jeremy Higdon

author

2006-11-21 03:15 UTC #0 lore raw

PATCH 1/1 sgiioc4: Disable module unload

This patch removes a module_exit function that it sgiioc4 should not have had.

Signed-off-by Jeremy Higdon <[email protected]>

drivers/ide/pci/sgiioc4.c +0 -7

@@ -768,14 +768,7 @@
 	return ioc4_register_submodule(&ioc4_ide_submodule);
 }
 
-static void __devexit
-ioc4_ide_exit(void)
-{
-	ioc4_unregister_submodule(&ioc4_ide_submodule);
-}
-
 late_initcall(ioc4_ide_init); /* Call only after IDE init is done */
-module_exit(ioc4_ide_exit);
 
 MODULE_AUTHOR("Aniket Malatpure/Jeremy Higdon");
 MODULE_DESCRIPTION("IDE PCI driver module for SGI IOC4 Base-IO Card");

Andrew Morton

2006-11-22 01:28 UTC #1 lore raw

Quoted 12 lines · On Mon, 20 Nov 2006 19:15:52 -0800 Jeremy Higdon <[email protected]> wrote:

This patch removes a module_exit function that it sgiioc4 should not have had.
Signed-off-by Jeremy Higdon <[email protected]>
drivers/ide/pci/sgiioc4.c +0 -1
@@ -768,14 +768,7 @@
 	return ioc4_register_submodule(&ioc4_ide_submodule);
 }
 
-static void __devexit

drivers/ide/pci/sgiioc4.c line 774

Quoted 10 lines

drivers/ide/pci/sgiioc4.c +0 -6

-ioc4_ide_exit(void)
-{
-	ioc4_unregister_submodule(&ioc4_ide_submodule);
-}
-
 late_initcall(ioc4_ide_init); /* Call only after IDE init is done */
-module_exit(ioc4_ide_exit);
 
 MODULE_AUTHOR("Aniket Malatpure/Jeremy Higdon");
 MODULE_DESCRIPTION("IDE PCI driver module for SGI IOC4 Base-IO Card");

Why? We'd prefer that the module be unloadable, wouldn't we?

Jeremy Higdon

author

2006-11-22 01:59 UTC #2 lore raw

Quoted 23 lines · Andrew Morton

Quoted 20 lines · On Mon, 20 Nov 2006 19:15:52 -0800 Jeremy Higdon <[email protected]> wrote:
This patch removes a module_exit function that it sgiioc4 should not have had.
Signed-off-by Jeremy Higdon <[email protected]>
drivers/ide/pci/sgiioc4.c +0 -7
@@ -768,14 +768,7 @@
 	return ioc4_register_submodule(&ioc4_ide_submodule);
 }
 
-static void __devexit
-ioc4_ide_exit(void)
-{
-	ioc4_unregister_submodule(&ioc4_ide_submodule);
-}
-
 late_initcall(ioc4_ide_init); /* Call only after IDE init is done */
-module_exit(ioc4_ide_exit);
 

Quoted 4 lines

drivers/ide/pci/sgiioc4.c line 774
Quoted 2 lines
drivers/ide/pci/sgiioc4.c +0 -0
 MODULE_AUTHOR("Aniket Malatpure/Jeremy Higdon");
 MODULE_DESCRIPTION("IDE PCI driver module for SGI IOC4 Base-IO Card");
Why? We'd prefer that the module be unloadable, wouldn't we?

It seems that the IDE layer doesn't support that. sgiioc4 was the only driver in drivers/ide/pci that had an exit function. After an unload, the devices would stay around and the next attempt to reference would crash...

Thanks,

jeremy

Andrew Morton

2006-11-22 02:13 UTC #3 lore raw

Quoted 29 lines · On Tue, 21 Nov 2006 17:59:52 -0800 Jeremy Higdon <[email protected]> wrote:

Quoted 23 lines · Andrew Morton
Quoted 20 lines · On Mon, 20 Nov 2006 19:15:52 -0800 Jeremy Higdon <[email protected]> wrote:
This patch removes a module_exit function that it sgiioc4 should not have had.
Signed-off-by Jeremy Higdon <[email protected]>
drivers/ide/pci/sgiioc4.c +0 -7
@@ -768,14 +768,7 @@
 	return ioc4_register_submodule(&ioc4_ide_submodule);
 }
 
-static void __devexit
-ioc4_ide_exit(void)
-{
-	ioc4_unregister_submodule(&ioc4_ide_submodule);
-}
-
 late_initcall(ioc4_ide_init); /* Call only after IDE init is done */
-module_exit(ioc4_ide_exit);
 
Quoted 4 lines
drivers/ide/pci/sgiioc4.c line 774
Quoted 2 lines
drivers/ide/pci/sgiioc4.c +0 -0
 MODULE_AUTHOR("Aniket Malatpure/Jeremy Higdon");
 MODULE_DESCRIPTION("IDE PCI driver module for SGI IOC4 Base-IO Card");
Why? We'd prefer that the module be unloadable, wouldn't we?

Quoted 4 lines

It seems that the IDE layer doesn't support that. sgiioc4 was the only driver in drivers/ide/pci that had an exit function. After an unload, the devices would stay around and the next attempt to reference would crash...

oh, OK, thanks. I updated the changelog.

Bartlomiej Zolnierkiewicz

2006-11-22 20:29 UTC #4 lore raw

In reply to Jeremy Higdon

Quoted 24 lines · Jeremy Higdon

Quoted 23 lines · Andrew Morton
Quoted 20 lines · On Mon, 20 Nov 2006 19:15:52 -0800 Jeremy Higdon <[email protected]> wrote:
This patch removes a module_exit function that it sgiioc4 should not have had.
Signed-off-by Jeremy Higdon <[email protected]>
drivers/ide/pci/sgiioc4.c 2006-11-20 18:45:08.506724280 -0800 renamed from drivers/ide/pci/sgiioc4.c 2006-11-15 20:03:40.000000000 -0800 +0 -7
@@ -768,14 +768,7 @@
    return ioc4_register_submodule(&ioc4_ide_submodule);
 }
 
-static void __devexit
-ioc4_ide_exit(void)
-{
-   ioc4_unregister_submodule(&ioc4_ide_submodule);
-}
-
 late_initcall(ioc4_ide_init); /* Call only after IDE init is done */
-module_exit(ioc4_ide_exit);
 

Quoted 4 lines

Quoted 4 lines
drivers/ide/pci/sgiioc4.c 2006-11-20 18:45:08.506724280 -0800 line 774
Quoted 2 lines
drivers/ide/pci/sgiioc4.c 2006-11-20 18:45:08.506724280 -0800 renamed from drivers/ide/pci/sgiioc4.c 2006-11-15 20:03:40.000000000 -0800 +0 -0
 MODULE_AUTHOR("Aniket Malatpure/Jeremy Higdon");
 MODULE_DESCRIPTION("IDE PCI driver module for SGI IOC4 Base-IO Card");
Why? We'd prefer that the module be unloadable, wouldn't we?

not supported in the current IDE layer model
(lack of dynamic allocation for ide_hwifs[] and some needed locking)

Quoted 4 lines

It seems that the IDE layer doesn't support that. sgiioc4 was the only driver in drivers/ide/pci that had an exit function. After an unload, the devices would stay around and the next attempt to reference would crash...

ack

Andrew, please apply.

Bart