switch mtd to new driver model & cleanups
PATCH 1/1 switch mtd to new driver model & cleanups
Switch mtd to new power management.
---
commit c608c935000f919426316a2f76084e31cab0171b tree e5bbc83aac75f5916cf79f7e531530c9655e38c6 parent 81a2889a21eb2bfdf8b242f2cc2cc1d6ad424ea2 author <pavel@amd.(none)> Fri, 30 Sep 2005 00:19:32 +0200 committer <pavel@amd.(none)> Fri, 30 Sep 2005 00:19:32 +0200
drivers/mtd/mtdcore.c | 68 ++++++++++++++++++++++++++++++++----------------- 1 files changed, 44 insertions(+), 24 deletions(-)
drivers/mtd/mtdcore.c +44 -24
@@ -19,6 +19,7 @@#include <linux/ioctl.h> #include <linux/init.h> #include <linux/mtd/compatmac.h> +#include <linux/sysdev.h> #ifdef CONFIG_PROC_FS #include <linux/proc_fs.h> #endif
@@ -80,6 +81,7 @@ int add_mtd_device(struct mtd_info *mtd)return 1 ; } +/** * del_mtd_device - unregister an MTD device * @mtd: pointer to MTD device info structure
@@ -303,30 +305,59 @@ EXPORT_SYMBOL(default_mtd_readv);#include <linux/pm.h> -static struct pm_dev *mtd_pm_dev =NULL ; +static int mtd_resume (struct sys_device *dev) +{ +int ret =0 , i; -static int mtd_pm_callback (struct pm_dev *dev,pm_request_t rqst,void *data) +if (down_trylock (&mtd_table_mutex)) +return -EAGAIN ; + +for (i =MAX_MTD_DEVICES -1 ; i >=0 ; i--) { +if (mtd_table[i] && mtd_table[i]->resume) + mtd_table[i]->resume (mtd_table[i]); + } + +up (&mtd_table_mutex); +return ret; +} + +static int mtd_suspend (struct sys_device *dev,pm_message_t state) {int ret =0 , i;if (down_trylock (&mtd_table_mutex))return -EAGAIN ; -if (rqst ==PM_SUSPEND ) { -for (i =0 ; ret ==0 && i <MAX_MTD_DEVICES ; i++) { -if (mtd_table[i] && mtd_table[i]->suspend) - ret = mtd_table[i]->suspend (mtd_table[i]); - } - }else i =MAX_MTD_DEVICES -1 ; -if (rqst ==PM_RESUME || ret) { -for ( ; i >=0 ; i--) { -if (mtd_table[i] && mtd_table[i]->resume) - mtd_table[i]->resume (mtd_table[i]); - } +for (i =0 ; ret ==0 && i <MAX_MTD_DEVICES ; i++) { +if (mtd_table[i] && mtd_table[i]->suspend) + ret = mtd_table[i]->suspend (mtd_table[i]); } +up (&mtd_table_mutex);return ret; } + +static struct sysdev_class mtd_sysdev_class = { +set_kset_name ("mtd" ), + .suspend = mtd_suspend, + .resume = mtd_resume, +}; + +static struct sys_device device_mtd = { + .id =0 , + .cls = &mtd_sysdev_class, +}; + +static int __init mtd_init_sysfs (void ) +{ +int error =sysdev_class_register (&mtd_sysdev_class); +if (!error) + error =sysdev_register (&device_mtd); +return error; +} + +device_initcall (mtd_init_sysfs); +#endif /*====================================================================*/
@@ -388,22 +419,11 @@ static int __init init_mtd(void)if ((proc_mtd =create_proc_entry ("mtd" ,0 ,NULL ))) proc_mtd->read_proc = mtd_read_proc;#endif - -#ifdef CONFIG_PM - mtd_pm_dev =pm_register (PM_UNKNOWN_DEV ,0 , mtd_pm_callback); -#endif return 0 ; } static void __exit cleanup_mtd(void) { -#ifdef CONFIG_PM -if (mtd_pm_dev) { -pm_unregister (mtd_pm_dev); - mtd_pm_dev =NULL ; - } -#endif -#ifdef CONFIG_PROC_FS if (proc_mtd)remove_proc_entry ("mtd" ,NULL );
!-------------------------------------------------------------flip-
Kill useless ifdefs from mtd.
---
commit b6ee6a389e7679a4f81e61a4c1ce84232df1a6f5 tree b71d9dbd8d035ef167134b88e67ed1d662b43146 parent c608c935000f919426316a2f76084e31cab0171b author <pavel@amd.(none)> Fri, 30 Sep 2005 00:20:52 +0200 committer <pavel@amd.(none)> Fri, 30 Sep 2005 00:20:52 +0200
drivers/mtd/mtdcore.c | 6 ------ 1 files changed, 0 insertions(+), 6 deletions(-)
drivers/mtd/mtdcore.c +0 -6
@@ -20,9 +20,7 @@#include <linux/init.h> #include <linux/mtd/compatmac.h> #include <linux/sysdev.h> -#ifdef CONFIG_PROC_FS #include <linux/proc_fs.h> -#endif #include <linux/mtd/mtd.h>
@@ -415,19 +413,15 @@ done:static int __init init_mtd (void ) { -#ifdef CONFIG_PROC_FS if ((proc_mtd =create_proc_entry ("mtd" ,0 ,NULL ))) proc_mtd->read_proc = mtd_read_proc; -#endif return 0 ; } static void __exit cleanup_mtd(void) { -#ifdef CONFIG_PROC_FS if (proc_mtd)remove_proc_entry ("mtd" ,NULL ); -#endif }module_init (init_mtd);
!-------------------------------------------------------------flip-
Signature · 1 line
if you have sharp zaurus hardware you don't need... you know my address
Quoted 1 line · Pavel Machek
Switch mtd to new power management.
You don't need to do it this way - and this will probably mess up the SA1100 suspend code. MTD map drivers should probably be converted to the driver model rather than the core so that the PM ordering can be correctly handled.
(Consider the case where there's flash on a daughter board accessed over some kind of bridge.)
Signature · 3 lines
Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: 2.6 Serial core
Hi Pavel,
it looks like your patch is not against the latest linux-mtd CVS sources since there's no such things as mtd_pm_dev in the current one. Please correct me if I'm mistaken.
Vitaly
Quoted 189 lines · Pavel Machek
Switch mtd to new power management.
Signed-off-by Pavel Machek <[email protected]>---
commitc608c935000f919426316a2f76084e31cab0171btreee5bbc83aac75f5916cf79f7e531530c9655e38c6parent81a2889a21eb2bfdf8b242f2cc2cc1d6ad424ea2author <pavel@amd.(none)> Fri, 30 Sep 2005 00:19:32 +0200 committer <pavel@amd.(none)> Fri, 30 Sep 2005 00:19:32 +0200
drivers/mtd/mtdcore.c| 68 ++++++++++++++++++++++++++++++++-----------------1 files changed, 44 insertions(+), 24 deletions(-)drivers/mtd/mtdcore.c +44
-24@@ -19,6 +19,7 @@#include <linux/ioctl.h> #include <linux/init.h> #include <linux/mtd/compatmac.h> +#include <linux/sysdev.h> #ifdef CONFIG_PROC_FS #include <linux/proc_fs.h> #endif @@ -80,6 +81,7 @@ int add_mtd_device(struct mtd_info *mtd)return 1 ; } +/** * del_mtd_device - unregister an MTD device * @mtd: pointer to MTD device info structure @@ -303,30 +305,59 @@ EXPORT_SYMBOL(default_mtd_readv);#include <linux/pm.h> -static struct pm_dev *mtd_pm_dev =NULL ; +static int mtd_resume (struct sys_device *dev) +{ +int ret =0 , i; -static int mtd_pm_callback (struct pm_dev *dev,pm_request_t rqst,void *data) +if (down_trylock (&mtd_table_mutex)) +return -EAGAIN ; + +for (i =MAX_MTD_DEVICES -1 ; i >=0 ; i--) { +if (mtd_table[i] && mtd_table[i]->resume) + mtd_table[i]->resume (mtd_table[i]); + } + +up (&mtd_table_mutex); +return ret; +} + +static int mtd_suspend (struct sys_device *dev,pm_message_t state) {int ret =0 , i;if (down_trylock (&mtd_table_mutex))return -EAGAIN ; -if (rqst ==PM_SUSPEND ) { -for (i =0 ; ret ==0 && i <MAX_MTD_DEVICES ; i++) { -if (mtd_table[i] && mtd_table[i]->suspend) - ret = mtd_table[i]->suspend (mtd_table[i]); - } - }else i =MAX_MTD_DEVICES -1 ; -if (rqst ==PM_RESUME || ret) { -for ( ; i >=0 ; i--) { -if (mtd_table[i] && mtd_table[i]->resume) - mtd_table[i]->resume (mtd_table[i]); - } +for (i =0 ; ret ==0 && i <MAX_MTD_DEVICES ; i++) { +if (mtd_table[i] && mtd_table[i]->suspend) + ret = mtd_table[i]->suspend (mtd_table[i]); } +up (&mtd_table_mutex);return ret; } + +static struct sysdev_class mtd_sysdev_class = { +set_kset_name ("mtd" ), + .suspend = mtd_suspend, + .resume = mtd_resume, +}; + +static struct sys_device device_mtd = { + .id =0 , + .cls = &mtd_sysdev_class, +}; + +static int __init mtd_init_sysfs (void ) +{ +int error =sysdev_class_register (&mtd_sysdev_class); +if (!error) + error =sysdev_register (&device_mtd); +return error; +} + +device_initcall (mtd_init_sysfs); +#endif /*====================================================================*/ @@ -388,22 +419,11 @@ static int __init init_mtd(void)if ((proc_mtd =create_proc_entry ("mtd" ,0 ,NULL ))) proc_mtd->read_proc = mtd_read_proc;#endif - -#ifdef CONFIG_PM - mtd_pm_dev =pm_register (PM_UNKNOWN_DEV ,0 , mtd_pm_callback); -#endif return 0 ; } static void __exit cleanup_mtd(void) { -#ifdef CONFIG_PM -if (mtd_pm_dev) { -pm_unregister (mtd_pm_dev); - mtd_pm_dev =NULL ; - } -#endif -#ifdef CONFIG_PROC_FS if (proc_mtd)remove_proc_entry ("mtd" ,NULL );!-------------------------------------------------------------flip-
Kill useless ifdefs from mtd.
---
commitb6ee6a389e7679a4f81e61a4c1ce84232df1a6f5treeb71d9dbd8d035ef167134b88e67ed1d662b43146parentc608c935000f919426316a2f76084e31cab0171bauthor <pavel@amd.(none)> Fri, 30 Sep 2005 00:20:52 +0200 committer <pavel@amd.(none)> Fri, 30 Sep 2005 00:20:52 +0200
drivers/mtd/mtdcore.c| 6 ------1 files changed, 0 insertions(+), 6 deletions(-)drivers/mtd/mtdcore.c +0
-6@@ -20,9 +20,7 @@#include <linux/init.h> #include <linux/mtd/compatmac.h> #include <linux/sysdev.h> -#ifdef CONFIG_PROC_FS #include <linux/proc_fs.h> -#endif #include <linux/mtd/mtd.h> @@ -415,19 +413,15 @@ done:static int __init init_mtd (void ) { -#ifdef CONFIG_PROC_FS if ((proc_mtd =create_proc_entry ("mtd" ,0 ,NULL ))) proc_mtd->read_proc = mtd_read_proc; -#endif return 0 ; } static void __exit cleanup_mtd(void) { -#ifdef CONFIG_PROC_FS if (proc_mtd)remove_proc_entry ("mtd" ,NULL ); -#endif }module_init (init_mtd);!-------------------------------------------------------------flip-
Quoted 5 lines · Vitaly Wool
Hi Pavel,
it looks like your patch is not against the latest linux-mtd CVS sources since there's no such things as
mtd_pm_devin the current one. Please correct me if I'm mistaken.
I'm assuming Pavel is just fixing up what's in the mainline kernel. This happens quite often, since MTD doesn't sync all too much for various reasons.
When a sync is done, stuff like this usually gets picked up in some form for the MTD CVS tree. See the commit messages that say things like "Merge with upstream".
josh
As far as I've heard, there's gonna be a megre quite soon... :)
Anyway, I guess it'd be better if Pavel ntroduced the new patch against the latest linux-mtd code.
Best regards,
Vitaly
Quoted 22 lines · Josh Boyer
Quoted 5 lines · Vitaly Wool
Hi Pavel,
it looks like your patch is not against the latest linux-mtd CVS sources since there's no such things as
mtd_pm_devin the current one. Please correct me if I'm mistaken.I'm assuming Pavel is just fixing up what's in the mainline kernel. This happens quite often, since MTD doesn't sync all too much for various reasons.
When a sync is done, stuff like this usually gets picked up in some form for the MTD CVS tree. See the commit messages that say things like "Merge with upstream".
josh
Quoted 3 lines · Vitaly Wool
As far as I've heard, there's gonna be a megre quite soon... :) Anyway, I guess it'd be better if Pavel ntroduced the new patch against the latest linux-mtd code.
I don't think it's required. The SA1100 map driver handles the PM already, as I'm sure the PXA map driver will also.
All it takes is for the other map drivers to be converted and hey presto, the MTD subsystem has power management support done the right way.
Signature · 3 lines
Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: 2.6 Serial core