Driver Core and sysfs stuff for 2.6.16
Subsystems (20)
PATCH 1/1 Driver Core and sysfs stuff for 2.6.16
Here are some driver core and sysfs patches for 2.6.16. They contain the following changes:
- EXPORT_SYMBOL_GPL_FUTURE() added to the kernel
- mark a few subsystems with this new export
- fix a few sysfs bugs (fixes a few bugs with USB device removals).
- kref optimization
- semaphore to mutex conversions
- module sysfs files reference counting fixes
- make sysfs a bit more verbous about errors/stupid usages.
- mark where people are using the driver model wrong more obvious.
- add kobject_add_dir() function
- add binary blob file type helper to debugfs
- remove a few sysfs functions from being exported as no one except the kobject code should be using it (it is impossible to use these, so no one has.)
- firmware error path fixes
All of these patches have been in the -mm tree for a number of weeks, if not months.
Please pull from:
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-2.6.git/
or if master.kernel.org hasn't synced up yet:
master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6.git/
Patches will be sent as a follow-on to this message to lkml for people to see.
thanks,
greg k-h
Documentation/feature-removal-schedule.txt | 18 ++ arch/arm/common/locomo.c | 2 arch/arm/common/sa1111.c | 2 arch/m68knommu/kernel/vmlinux.lds.S | 10 + arch/v850/kernel/vmlinux.lds.S | 8 + block/genhd.c | 31 ++-- drivers/base/cpu.c | 2 drivers/base/firmware_class.c | 6 drivers/base/map.c | 21 +-- drivers/base/platform.c | 4 drivers/char/s3c2410-rtc.c | 4 drivers/char/watchdog/mpcore_wdt.c | 4 drivers/i2c/busses/i2c-iop3xx.c | 9 + drivers/i2c/busses/i2c-mpc.c | 5 drivers/i2c/busses/i2c-mv64xxx.c | 4 drivers/ide/mips/au1xxx-ide.c | 5 drivers/mmc/pxamci.c | 2 drivers/net/arm/am79c961a.c | 4 drivers/net/fs_enet/mac-fcc.c | 2 drivers/net/fs_enet/mac-fec.c | 2 drivers/net/fs_enet/mac-scc.c | 2 drivers/net/gianfar.c | 4 drivers/net/smc91x.c | 4 drivers/pcmcia/omap_cf.c | 2 drivers/serial/s3c2410.c | 2 drivers/usb/core/driver.c | 6 drivers/usb/host/ohci-omap.c | 9 + drivers/video/epson1355fb.c | 1 drivers/video/sa1100fb.c | 2 drivers/video/vfb.c | 1 fs/char_dev.c | 17 +- fs/debugfs/file.c | 46 ++++++ fs/sysfs/dir.c | 37 ++++- fs/sysfs/file.c | 9 - fs/sysfs/inode.c | 9 - fs/sysfs/symlink.c | 6 fs/sysfs/sysfs.h | 1 include/asm-generic/vmlinux.lds.h | 14 ++ include/linux/cpu.h | 2 include/linux/debugfs.h | 15 ++ include/linux/device.h | 2 include/linux/kobj_map.h | 4 include/linux/kobject.h | 4 include/linux/module.h | 10 + kernel/ksysfs.c | 3 kernel/module.c | 199 +++++++++++++++++------------ kernel/params.c | 10 - kernel/rcupdate.c | 6 lib/kobject.c | 60 +++++++- lib/kobject_uevent.c | 2 lib/kref.c | 7 - scripts/genksyms/keywords.c_shipped | 91 ++++++------- scripts/genksyms/keywords.gperf | 1 53 files changed, 514 insertions(+), 219 deletions(-)
---------------
Adrian Bunk:
Kobject: kobject.h: fix a typo
Andrew Morton:
get_cpu_sysdev() signedness fix
David Vrabel:
driver core: platform_get_irq*(): return -ENXIO on error handle errors returned by platform_get_irq*()
Eric Dumazet:
kref: avoid an atomic operation in kref_put()
Eric Sesterhenn:
sysfs: kzalloc conversion
Greg Kroah-Hartman:
sysfs: sysfs_remove_dir() needs to invalidate the dentry Mark empty release functions as broken
add EXPORT_SYMBOL_GPL_FUTURE()
add EXPORT_SYMBOL_GPL_FUTURE() to RCU subsystem
add EXPORT_SYMBOL_GPL_FUTURE() to USB subsystem
fix module sysfs files reference counting
Kobject: provide better warning messages when people do stupid things
sysfs: don't export dir symbols
sysfs: fix a kobject leak in sysfs_add_link on the error path
Jeff Moyer:
firmware: fix BUG: in fw_realloc_buffer
Jes Sorensen:
kobj_map semaphore to mutex conversion
Jun'ichi Nomura:
kobject: fix build error if CONFIG_SYSFS=n
kobject_add_dir
Maneesh Soni:
sysfs: fix problem with duplicate sysfs directories and files
Michael Ellerman:
debugfs: Add debugfs_create_blob() helper for exporting binary data
Sam Ravnborg:
Clean up module.c symbol searching logic
Tilman Schmidt:
Driver core: add macros notice(), dev_notice()
PATCH 1/23 sysfs: sysfs_remove_dir() needs to invalidate the dentry
When calling sysfs_remove_dir() don't allow any further sysfs functions to work for this kobject anymore. This fixes a nasty USB cdc-acm oops on disconnect.
Many thanks to Bob Copeland and Paul Fulghum for taking the time to track this down.
---
fs/sysfs/dir.c | 1 + fs/sysfs/inode.c | 6 +++++- 2 files changed, 6 insertions(+), 1 deletions(-)
641e6f30a095f3752ed84fd9d279382f5d3ef4c1
fs/sysfs/dir.c +1 -0
@@ -302,6 +302,7 @@ void sysfs_remove_dir(struct kobject * k* Drop reference from dget() on entrance. */ dput (dentry); + kobj->dentry =NULL ; }int sysfs_rename_dir (struct kobject * kobj,const char *new_name)
fs/sysfs/inode.c +5 -1
@@ -227,12 +227,16 @@ void sysfs_drop_dentry(struct sysfs_direvoid sysfs_hash_and_remove (struct dentry * dir,const char * name) {struct sysfs_dirent * sd; -struct sysfs_dirent * parent_sd = dir->d_fsdata; +struct sysfs_dirent * parent_sd; + +if (!dir) +return ;if (dir->d_inode ==NULL )/* no inode means this hasn't been made visible yet */ return ; + parent_sd = dir->d_fsdata;mutex_lock (&dir->d_inode->i_mutex);list_for_each_entry (sd, &parent_sd->s_children, s_sibling) {if (!sd->s_element)
Signature · 1 line
1.2.4
PATCH 2/23 kobject: fix build error if CONFIG_SYSFS=n
Moving uevent_seqnum and uevent_helper to kobject_uevent.c because they are used even if CONFIG_SYSFS=n while kernel/ksysfs.c is built only if CONFIG_SYSFS=y,
---
kernel/ksysfs.c | 3 --- lib/kobject_uevent.c | 2 ++ 2 files changed, 2 insertions(+), 3 deletions(-)
51107301b629640f9ab76fe23bf385e187b9ac29
kernel/ksysfs.c +0 -3
@@ -15,9 +15,6 @@#include <linux/module.h> #include <linux/init.h> -u64 uevent_seqnum; -char uevent_helper[UEVENT_HELPER_PATH_LEN ] ="/sbin/hotplug" ; -#define KERNEL_ATTR_RO (_name) \ static struct subsys_attribute _name##_attr = __ATTR_RO(_name)
lib/kobject_uevent.c +2 -0
@@ -26,6 +26,8 @@#define NUM_ENVP 32/* number of env pointers */ #if defined(CONFIG_HOTPLUG ) && defined(CONFIG_NET ) +u64 uevent_seqnum; +char uevent_helper[UEVENT_HELPER_PATH_LEN ] ="/sbin/hotplug" ;static DEFINE_SPINLOCK (sequence_lock);static struct sock *uevent_sock;
Signature · 1 line
1.2.4
PATCH 3/23 Mark empty release functions as broken
Come on people, this is just wrong...
---
drivers/video/epson1355fb.c | 1 + drivers/video/vfb.c | 1 + 2 files changed, 2 insertions(+), 0 deletions(-)
972de6c8bfd8b36618563be454df1e95a36dc379
drivers/video/epson1355fb.c +1 -0
@@ -607,6 +607,7 @@ static void clearfb16(struct fb_info *instatic void epson1355fb_platform_release (struct device *device) { +dev_err (device,"This driver is broken, please bug the authors so they will fix it.\n" ); }static int epson1355fb_remove (struct platform_device *dev)
drivers/video/vfb.c +1 -0
@@ -401,6 +401,7 @@ static int __init vfb_setup(char *optionstatic void vfb_platform_release (struct device *device) {// This is called when the reference count goes to zero. +dev_err (device,"This driver is broken, please bug the authors so they will fix it.\n" ); }static int __init vfb_probe (struct platform_device *dev)
Signature · 1 line
1.2.4
PATCH 4/23 driver core: platform_get_irq*(): return -ENXIO on error
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return -ENXIO instead.
---
drivers/base/platform.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
305b3228f9ff4d59f49e6d34a7034d44ee8ce2f0
drivers/base/platform.c +2 -2
@@ -61,7 +61,7 @@ int platform_get_irq(struct platform_dev {struct resource *r =platform_get_resource (dev,IORESOURCE_IRQ , num); -return r ? r->start :0 ; +return r ? r->start : -ENXIO ; }EXPORT_SYMBOL_GPL (platform_get_irq);
@@ -98,7 +98,7 @@ int platform_get_irq_byname(struct platf {struct resource *r =platform_get_resource_byname (dev,IORESOURCE_IRQ , name); -return r ? r->start :0 ; +return r ? r->start : -ENXIO ; }EXPORT_SYMBOL_GPL (platform_get_irq_byname);
Signature · 1 line
1.2.4
PATCH 5/23 handle errors returned by platform_get_irq*()
platform_get_irq*() now returns on -ENXIO when the resource cannot be found. Ensure all users of platform_get_irq*() handle this error appropriately.
---
arch/arm/common/locomo.c | 2 ++ arch/arm/common/sa1111.c | 2 ++ drivers/char/s3c2410-rtc.c | 4 ++-- drivers/char/watchdog/mpcore_wdt.c | 4 ++++ drivers/i2c/busses/i2c-iop3xx.c | 9 +++++++-- drivers/i2c/busses/i2c-mpc.c | 5 +++++ drivers/i2c/busses/i2c-mv64xxx.c | 4 ++++ drivers/ide/mips/au1xxx-ide.c | 5 +++++ drivers/mmc/pxamci.c | 2 +- drivers/net/arm/am79c961a.c | 4 +++- drivers/net/fs_enet/mac-fcc.c | 2 ++ drivers/net/fs_enet/mac-fec.c | 2 ++ drivers/net/fs_enet/mac-scc.c | 2 ++ drivers/net/gianfar.c | 4 ++++ drivers/net/smc91x.c | 4 ++++ drivers/pcmcia/omap_cf.c | 2 +- drivers/serial/s3c2410.c | 2 ++ drivers/usb/host/ohci-omap.c | 9 +++++++-- drivers/video/sa1100fb.c | 2 +- 19 files changed, 60 insertions(+), 10 deletions(-)
489447380a2921ec0e9154f773c44ab3167ede4b
arch/arm/common/locomo.c +2 -0
@@ -788,6 +788,8 @@ static int locomo_probe(struct platform_if (!mem)return -EINVAL ; irq =platform_get_irq (dev,0 ); +if (irq <0 ) +return -ENXIO ;return __locomo_probe (&dev->dev, mem, irq); }
arch/arm/common/sa1111.c +2 -0
@@ -943,6 +943,8 @@ static int sa1111_probe(struct platform_if (!mem)return -EINVAL ; irq =platform_get_irq (pdev,0 ); +if (irq <0 ) +return -ENXIO ;return __sa1111_probe (&pdev->dev, mem, irq); }
drivers/char/s3c2410-rtc.c +2 -2
@@ -448,13 +448,13 @@ static int s3c2410_rtc_probe(struct plat/* find the IRQs */ s3c2410_rtc_tickno =platform_get_irq (pdev,1 ); -if (s3c2410_rtc_tickno <=0 ) { +if (s3c2410_rtc_tickno <0 ) {dev_err (&pdev->dev,"no irq for rtc tick\n" );return -ENOENT ; } s3c2410_rtc_alarmno =platform_get_irq (pdev,0 ); -if (s3c2410_rtc_alarmno <=0 ) { +if (s3c2410_rtc_alarmno <0 ) {dev_err (&pdev->dev,"no irq for alarm\n" );return -ENOENT ; }
drivers/char/watchdog/mpcore_wdt.c +4 -0
@@ -338,6 +338,10 @@ static int __devinit mpcore_wdt_probe(st wdt->dev = &dev->dev; wdt->irq =platform_get_irq (dev,0 ); +if (wdt->irq <0 ) { + ret = -ENXIO ; + goto err_free; + } wdt->base =ioremap (res->start, res->end - res->start +1 );if (!wdt->base) { ret = -ENOMEM ;
drivers/i2c/busses/i2c-iop3xx.c +7 -2
@@ -434,7 +434,7 @@ static int iop3xx_i2c_probe(struct platform_device *pdev) {struct resource *res; -int ret; +int ret, irq;struct i2c_adapter *new_adapter;struct i2c_algo_iop3xx_data *adapter_data;
@@ -470,7 +470,12 @@ iop3xx_i2c_probe(struct platform_device goto release_region; } - ret =request_irq (platform_get_irq (pdev,0 ), iop3xx_i2c_irq_handler,0 , + irq =platform_get_irq (pdev,0 ); +if (irq <0 ) { + ret = -ENXIO ; + goto unmap; + } + ret =request_irq (irq, iop3xx_i2c_irq_handler,0 , pdev->name, adapter_data);if (ret) {
drivers/i2c/busses/i2c-mpc.c +5 -0
@@ -302,6 +302,10 @@ static int fsl_i2c_probe(struct platform } i2c->irq =platform_get_irq (pdev,0 ); +if (i2c->irq <0 ) { + result = -ENXIO ; + goto fail_get_irq; + } i2c->flags = pdata->device_flags;init_waitqueue_head (&i2c->queue);
@@ -340,6 +344,7 @@ static int fsl_i2c_probe(struct platform fail_irq:iounmap (i2c->base); fail_map: + fail_get_irq:kfree (i2c);return result; };
drivers/i2c/busses/i2c-mv64xxx.c +4 -0
@@ -516,6 +516,10 @@ mv64xxx_i2c_probe(struct platform_device drv_data->freq_m = pdata->freq_m; drv_data->freq_n = pdata->freq_n; drv_data->irq =platform_get_irq (pd,0 ); +if (drv_data->irq <0 ) { + rc = -ENXIO ; + goto exit_unmap_regs; + } drv_data->adapter.id =I2C_HW_MV64XXX ; drv_data->adapter.algo = &mv64xxx_i2c_algo; drv_data->adapter.owner =THIS_MODULE ;
drivers/ide/mips/au1xxx-ide.c +5 -0
@@ -674,6 +674,11 @@ static int au_ide_probe(struct device *d ret = -ENODEV ; goto out; } +if (ahwif->irq <0 ) { +pr_debug ("%s %d: no IRQ\n" ,DRV_NAME , pdev->id); + ret = -ENODEV ; + goto out; + }if (!request_mem_region (res->start, res->end-res->start, pdev->name)) {pr_debug ("%s: request_mem_region failed\n" ,DRV_NAME );
drivers/mmc/pxamci.c +1 -1
@@ -438,7 +438,7 @@ static int pxamci_probe(struct platform_ r =platform_get_resource (pdev,IORESOURCE_MEM ,0 ); irq =platform_get_irq (pdev,0 ); -if (!r || irq ==NO_IRQ ) +if (!r || irq <0 )return -ENXIO ; r =request_mem_region (r->start,SZ_4K ,DRIVER_NAME );
drivers/net/arm/am79c961a.c +3 -1
@@ -696,7 +696,9 @@ static int __init am79c961_probe(struct dev->base_addr = res->start; dev->irq =platform_get_irq (pdev,0 ); - ret = -ENODEV ; + ret = -ENODEV ; +if (dev->irq <0 ) + goto nodev;if (!request_region (dev->base_addr,0x18 , dev->name)) goto nodev;
drivers/net/fs_enet/mac-fcc.c +2 -0
@@ -118,6 +118,8 @@ static int do_pd_setup(struct fs_enet_pr/* Fill out IRQ field */ fep->interrupt =platform_get_irq (pdev,0 ); +if (fep->interrupt <0 ) +return -EINVAL ;/* Attach the memory for the FCC Parameter RAM */ r =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"fcc_pram" );
drivers/net/fs_enet/mac-fec.c +2 -0
@@ -144,6 +144,8 @@ static int do_pd_setup(struct fs_enet_pr/* Fill out IRQ field */ fep->interrupt =platform_get_irq_byname (pdev,"interrupt" ); +if (fep->interrupt <0 ) +return -EINVAL ; r =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"regs" ); fep->fec.fecp =(void *)r->start;
drivers/net/fs_enet/mac-scc.c +2 -0
@@ -118,6 +118,8 @@ static int do_pd_setup(struct fs_enet_pr/* Fill out IRQ field */ fep->interrupt =platform_get_irq_byname (pdev,"interrupt" ); +if (fep->interrupt <0 ) +return -EINVAL ; r =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"regs" ); fep->scc.sccp = (void *)r->start;
drivers/net/gianfar.c +4 -0
@@ -193,8 +193,12 @@ static int gfar_probe(struct platform_de priv->interruptTransmit =platform_get_irq_byname (pdev,"tx" ); priv->interruptReceive =platform_get_irq_byname (pdev,"rx" ); priv->interruptError =platform_get_irq_byname (pdev,"error" ); +if (priv->interruptTransmit <0 || priv->interruptReceive <0 || priv->interruptError <0 ) + goto regs_fail; }else { priv->interruptTransmit =platform_get_irq (pdev,0 ); +if (priv->interruptTransmit <0 ) + goto regs_fail; }/* get a pointer to the register memory */
drivers/net/smc91x.c +4 -0
@@ -2221,6 +2221,10 @@ static int smc_drv_probe(struct platform ndev->dma = (unsigned char )-1 ; ndev->irq =platform_get_irq (pdev,0 ); +if (ndev->irq <0 ) { + ret = -ENODEV ; + goto out_free_netdev; + } ret =smc_request_attrib (pdev);if (ret)
drivers/pcmcia/omap_cf.c +1 -1
@@ -218,7 +218,7 @@ static int __init omap_cf_probe(struct d/* either CFLASH.IREQ (INT_1610_CF) or some GPIO */ irq =platform_get_irq (pdev,0 ); -if (!irq) +if (irq <0 )return -EINVAL ; cf =kcalloc (1 ,sizeof *cf,GFP_KERNEL );
drivers/serial/s3c2410.c +2 -0
@@ -1066,6 +1066,8 @@ static int s3c24xx_serial_init_port(stru port->mapbase = res->start; port->membase =S3C24XX_VA_UART + (res->start -S3C24XX_PA_UART ); port->irq =platform_get_irq (platdev,0 ); +if (port->irq <0 ) + port->irq =0 ; ourport->clk =clk_get (&platdev->dev,"uart" );
drivers/usb/host/ohci-omap.c +7 -2
@@ -286,7 +286,7 @@ void usb_hcd_omap_remove (struct usb_hcdint usb_hcd_omap_probe (const struct hc_driver *driver,struct platform_device *pdev) { -int retval; +int retval, irq;struct usb_hcd *hcd =0 ;struct ohci_hcd *ohci;
@@ -329,7 +329,12 @@ int usb_hcd_omap_probe (const struct hc_if (retval <0 ) goto err2; - retval =usb_add_hcd (hcd,platform_get_irq (pdev,0 ),SA_INTERRUPT ); + irq =platform_get_irq (pdev,0 ); +if (irq <0 ) { + retval = -ENXIO ; + goto err2; + } + retval =usb_add_hcd (hcd, irq,SA_INTERRUPT );if (retval ==0 )return retval;
drivers/video/sa1100fb.c +1 -1
@@ -1457,7 +1457,7 @@ static int __init sa1100fb_probe(structint ret, irq; irq =platform_get_irq (pdev,0 ); -if (irq <=0 ) +if (irq <0 )return -EINVAL ;if (!request_mem_region (0xb0100000 ,0x10000 ,"LCD" ))
Signature · 1 line
1.2.4
PATCH 6/23 kref: avoid an atomic operation in kref_put()
Avoid an atomic operation in kref_put() when the last reference is dropped. On most platforms, atomic_read() is a plan read of the counter and involves no atomic at all.
---
lib/kref.c | 7 ++++++- 1 files changed, 6 insertions(+), 1 deletions(-)
8b5536bbee53620f8d5f367987e5727ba36d886d
lib/kref.c +6 -1
@@ -52,7 +52,12 @@ int kref_put(struct kref *kref, void (*rWARN_ON (release ==NULL );WARN_ON (release == (void (*)(struct kref *))kfree); -if (atomic_dec_and_test (&kref->refcount)) { +/* +* if current count is one, we are the last user and can release object +* right now, avoiding an atomic operation on 'refcount' +*/ +if ((atomic_read (&kref->refcount) ==1 ) || + (atomic_dec_and_test (&kref->refcount))) {release (kref);return 1 ; }
Signature · 1 line
1.2.4
PATCH 7/23 kobj_map semaphore to mutex conversion
Convert the kobj_map code to use a mutex instead of a semaphore. It converts the single two users as well, genhd.c and char_dev.c.
---
block/genhd.c | 31 ++++++++++++++++--------------- drivers/base/map.c | 21 +++++++++++---------- fs/char_dev.c | 17 +++++++++-------- include/linux/kobj_map.h | 4 ++-- 4 files changed, 38 insertions(+), 35 deletions(-)
58383af629efb07e5a0694e445eda0c65b16e1de
block/genhd.c +16 -15
@@ -15,12 +15,13 @@#include <linux/kmod.h> #include <linux/kobj_map.h> #include <linux/buffer_head.h> +#include <linux/mutex.h> #define MAX_PROBE_HASH 255/* random */ static struct subsystem block_subsys; -static DECLARE_MUTEX (block_subsys_sem); +static DEFINE_MUTEX (block_subsys_lock);/* * Can be deleted altogether. Later.
@@ -46,7 +47,7 @@ struct blkdev_info {/* * iterate over a list of blkdev_info structures. allows * the major_names array to be iterated over from outside this file -* must be called with the block_subsys_sem held +* must be called with the block_subsys_lock held */ void *get_next_blkdev (void *dev) {
@@ -85,20 +86,20 @@ out:void *acquire_blkdev_list (void ) { -down (&block_subsys_sem); +mutex_lock (&block_subsys_lock);return get_next_blkdev (NULL ); }void release_blkdev_list (void *dev) { -up (&block_subsys_sem); +mutex_unlock (&block_subsys_lock);kfree (dev); }/* * Count the number of records in the blkdev_list. -* must be called with the block_subsys_sem held +* must be called with the block_subsys_lock held */ int count_blkdev_list (void ) {
@@ -118,7 +119,7 @@ int count_blkdev_list(void)/* * extract the major and name values from a blkdev_info struct * passed in as a void to *dev. Must be called with -* block_subsys_sem held +* block_subsys_lock held */ int get_blkdev_info (void *dev,int *major,char **name) {
@@ -138,7 +139,7 @@ int register_blkdev(unsigned int major,struct blk_major_name **n, *p;int index, ret =0 ; -down (&block_subsys_sem); +mutex_lock (&block_subsys_lock);/* temporary */ if (major ==0 ) {
@@ -183,7 +184,7 @@ int register_blkdev(unsigned int major,kfree (p); } out: -up (&block_subsys_sem); +mutex_unlock (&block_subsys_lock);return ret; }
@@ -197,7 +198,7 @@ int unregister_blkdev(unsigned int majorint index =major_to_index (major);int ret =0 ; -down (&block_subsys_sem); +mutex_lock (&block_subsys_lock);for (n = &major_names[index]; *n; n = &(*n)->next)if ((*n)->major == major)break ;
@@ -207,7 +208,7 @@ int unregister_blkdev(unsigned int major p = *n; *n = p->next; } -up (&block_subsys_sem); +mutex_unlock (&block_subsys_lock);kfree (p);return ret;
@@ -301,7 +302,7 @@ static void *part_start(struct seq_filestruct list_head *p;loff_t l = *pos; -down (&block_subsys_sem); +mutex_lock (&block_subsys_lock);list_for_each (p, &block_subsys.kset.list)if (!l--) return list_entry(p, struct gendisk, kobj.entry);
@@ -318,7 +319,7 @@ static void *part_next(struct seq_file *static void part_stop (struct seq_file *part,void *v) { -up (&block_subsys_sem); +mutex_unlock (&block_subsys_lock); }static int show_partition (struct seq_file *part,void *v)
@@ -377,7 +378,7 @@ static struct kobject *base_probe(dev_tstatic int __init genhd_device_init (void ) { - bdev_map =kobj_map_init (base_probe, &block_subsys_sem); + bdev_map =kobj_map_init (base_probe, &block_subsys_lock);blk_dev_init ();subsystem_register (&block_subsys);return 0 ;
@@ -611,7 +612,7 @@ static void *diskstats_start(struct seq_loff_t k = *pos;struct list_head *p; -down (&block_subsys_sem); +mutex_lock (&block_subsys_lock);list_for_each (p, &block_subsys.kset.list)if (!k--) return list_entry(p, struct gendisk, kobj.entry);
@@ -628,7 +629,7 @@ static void *diskstats_next(struct seq_fstatic void diskstats_stop (struct seq_file *part,void *v) { -up (&block_subsys_sem); +mutex_unlock (&block_subsys_lock); }static int diskstats_show (struct seq_file *s,void *v)
drivers/base/map.c +11 -10
@@ -11,6 +11,7 @@#include <linux/module.h> #include <linux/slab.h> +#include <linux/mutex.h> #include <linux/kdev_t.h> #include <linux/kobject.h> #include <linux/kobj_map.h>
@@ -25,7 +26,7 @@ struct kobj_map {int (*lock)(dev_t ,void *);void *data; } *probes[255 ]; -struct semaphore *sem; +struct mutex *lock; }; int kobj_map(struct kobj_map *domain, dev_t dev, unsigned long range,
@@ -53,7 +54,7 @@ int kobj_map(struct kobj_map *domain, de p->range = range; p->data = data; } -down (domain->sem); +mutex_lock (domain->lock);for (i =0 , p -= n; i < n; i++, p++, index++) {struct probe **s = &domain->probes[index %255 ];while (*s && (*s)->range < range)
@@ -61,7 +62,7 @@ int kobj_map(struct kobj_map *domain, de p->next = *s; *s = p; } -up (domain->sem); +mutex_unlock (domain->lock);return 0 ; }
@@ -75,7 +76,7 @@ void kobj_unmap(struct kobj_map *domain,if (n >255 ) n =255 ; -down (domain->sem); +mutex_lock (domain->lock);for (i =0 ; i < n; i++, index++) {struct probe **s;for (s = &domain->probes[index %255 ]; *s; s = &(*s)->next) {
@@ -88,7 +89,7 @@ void kobj_unmap(struct kobj_map *domain, } } } -up (domain->sem); +mutex_unlock (domain->lock);kfree (found); }
@@ -99,7 +100,7 @@ struct kobject *kobj_lookup(struct kobj_unsigned long best = ~0UL ; retry: -down (domain->sem); +mutex_lock (domain->lock);for (p = domain->probes[MAJOR (dev) %255 ]; p; p = p->next) {struct kobject *(*probe)(dev_t ,int *,void *);struct module *owner;
@@ -120,7 +121,7 @@ retry:module_put (owner);continue ; } -up (domain->sem); +mutex_unlock (domain->lock); kobj =probe (dev, index, data);/* Currently ->owner protects _only_ ->probe() itself. */ module_put (owner);
@@ -128,11 +129,11 @@ retry:return kobj; goto retry; } -up (domain->sem); +mutex_unlock (domain->lock);return NULL ; } -struct kobj_map *kobj_map_init (kobj_probe_t *base_probe,struct semaphore *sem) +struct kobj_map *kobj_map_init (kobj_probe_t *base_probe,struct mutex *lock) {struct kobj_map *p =kmalloc (sizeof (struct kobj_map ),GFP_KERNEL );struct probe *base =kzalloc (sizeof (*base),GFP_KERNEL );
@@ -149,6 +150,6 @@ struct kobj_map *kobj_map_init(kobj_prob base->get = base_probe;for (i =0 ; i <255 ; i++) p->probes[i] = base; - p->sem = sem; + p->lock = lock;return p; }
fs/char_dev.c +9 -8
@@ -19,6 +19,7 @@#include <linux/kobject.h> #include <linux/kobj_map.h> #include <linux/cdev.h> +#include <linux/mutex.h> #ifdef CONFIG_KMOD #include <linux/kmod.h>
@@ -28,7 +29,7 @@ static struct kobj_map *cdev_map;#define MAX_PROBE_HASH 255/* random */ -static DECLARE_MUTEX (chrdevs_lock); +static DEFINE_MUTEX (chrdevs_lock); static struct char_device_struct { struct char_device_struct *next;
@@ -88,13 +89,13 @@ out:void *acquire_chrdev_list (void ) { -down (&chrdevs_lock); +mutex_lock (&chrdevs_lock);return get_next_chrdev (NULL ); }void release_chrdev_list (void *dev) { -up (&chrdevs_lock); +mutex_unlock (&chrdevs_lock);kfree (dev); }
@@ -151,7 +152,7 @@ __register_chrdev_region(unsigned int mamemset (cd,0 ,sizeof (struct char_device_struct )); -down (&chrdevs_lock); +mutex_lock (&chrdevs_lock);/* temporary */ if (major ==0 ) {
@@ -186,10 +187,10 @@ __register_chrdev_region(unsigned int ma } cd->next = *cp; *cp = cd; -up (&chrdevs_lock); +mutex_unlock (&chrdevs_lock);return cd; out: -up (&chrdevs_lock); +mutex_unlock (&chrdevs_lock);kfree (cd);return ERR_PTR (ret); }
@@ -200,7 +201,7 @@ __unregister_chrdev_region(unsigned majostruct char_device_struct *cd =NULL , **cp;int i =major_to_index (major); -down (&chrdevs_lock); +mutex_lock (&chrdevs_lock);for (cp = &chrdevs[i]; *cp; cp = &(*cp)->next) if ((*cp)->major == major && (*cp)->baseminor == baseminor &&
@@ -210,7 +211,7 @@ __unregister_chrdev_region(unsigned majo cd = *cp; *cp = cd->next; } -up (&chrdevs_lock); +mutex_unlock (&chrdevs_lock);return cd; }
include/linux/kobj_map.h +2 -2
@@ -1,6 +1,6 @@#ifdef __KERNEL__ -#include <asm/semaphore.h> +#include <linux/mutex.h> typedef struct kobject *kobj_probe_t (dev_t ,int *,void *);struct kobj_map ;
@@ -9,6 +9,6 @@ int kobj_map(struct kobj_map *, dev_t, u kobj_probe_t *, int (*)(dev_t, void *), void *);void kobj_unmap (struct kobj_map *,dev_t ,unsigned long );struct kobject *kobj_lookup (struct kobj_map *,dev_t ,int *); -struct kobj_map *kobj_map_init (kobj_probe_t *,struct semaphore *); +struct kobj_map *kobj_map_init (kobj_probe_t *,struct mutex *);#endif
Signature · 1 line
1.2.4
PATCH 8/23 Clean up module.c symbol searching logic
---
kernel/module.c | 73 +++++++++++++++++++++++++++++++------------------------ 1 files changed, 41 insertions(+), 32 deletions(-)
3fd6805f4dfb02bcfb5634972eabad0e790f119a
kernel/module.c +41 -32
@@ -135,6 +135,18 @@ extern const unsigned long __start___kcr#define symversion (base, idx) ((base) ? ((base) + (idx)) : NULL)#endif +/* lookup symbol in given range of kernel_symbols */ +static const struct kernel_symbol *lookup_symbol (const char *name, +const struct kernel_symbol *start, +const struct kernel_symbol *stop) +{ +const struct kernel_symbol *ks = start; +for (; ks < stop; ks++) +if (strcmp (ks->name, name) ==0 ) +return ks; +return NULL ; +} +/* Find a symbol, return value, crc and module which owns it */ static unsigned long __find_symbol(const char *name, struct module **owner,
@@ -142,39 +154,41 @@ static unsigned long __find_symbol(const int gplok) {struct module *mod; -unsigned int i; +const struct kernel_symbol *ks;/* Core kernel first. */ *owner =NULL ; -for (i =0 ; __start___ksymtab+i < __stop___ksymtab; i++) { -if (strcmp (__start___ksymtab[i].name, name) ==0 ) { - *crc =symversion (__start___kcrctab, i); -return __start___ksymtab[i].value; - } + ks =lookup_symbol (name, __start___ksymtab, __stop___ksymtab); +if (ks) { + *crc =symversion (__start___kcrctab, (ks - __start___ksymtab)); +return ks->value; }if (gplok) { -for (i =0 ; __start___ksymtab_gpl+i<__stop___ksymtab_gpl; i++) -if (strcmp (__start___ksymtab_gpl[i].name, name) ==0 ) { - *crc =symversion (__start___kcrctab_gpl, i); -return __start___ksymtab_gpl[i].value; - } + ks =lookup_symbol (name, __start___ksymtab_gpl, + __stop___ksymtab_gpl); +if (ks) { + *crc =symversion (__start___kcrctab_gpl, + (ks - __start___ksymtab_gpl)); +return ks->value; + } }/* Now try modules. */ list_for_each_entry (mod, &modules, list) { *owner = mod; -for (i =0 ; i < mod->num_syms; i++) -if (strcmp (mod->syms[i].name, name) ==0 ) { - *crc =symversion (mod->crcs, i); -return mod->syms[i].value; - } + ks =lookup_symbol (name, mod->syms, mod->syms + mod->num_syms); +if (ks) { + *crc =symversion (mod->crcs, (ks - mod->syms)); +return ks->value; + }if (gplok) { -for (i =0 ; i < mod->num_gpl_syms; i++) { -if (strcmp (mod->gpl_syms[i].name, name) ==0 ) { - *crc =symversion (mod->gpl_crcs, i); -return mod->gpl_syms[i].value; - } + ks =lookup_symbol (name, mod->gpl_syms, + mod->gpl_syms + mod->num_gpl_syms); +if (ks) { + *crc =symversion (mod->gpl_crcs, + (ks - mod->gpl_syms)); +return ks->value; } } }
@@ -1444,18 +1458,13 @@ static void setup_modinfo(struct module#ifdef CONFIG_KALLSYMS int is_exported (const char *name,const struct module *mod) { -unsigned int i; - -if (!mod) { -for (i =0 ; __start___ksymtab+i < __stop___ksymtab; i++) -if (strcmp (__start___ksymtab[i].name, name) ==0 ) -return 1 ; -return 0 ; - } -for (i =0 ; i < mod->num_syms; i++) -if (strcmp (mod->syms[i].name, name) ==0 ) +if (!mod &&lookup_symbol (name, __start___ksymtab, __stop___ksymtab)) +return 1 ; +else +if (lookup_symbol (name, mod->syms, mod->syms + mod->num_syms))return 1 ; -return 0 ; +else +return 0 ; }/* As per nm */
Signature · 1 line
1.2.4
PATCH 9/23 add EXPORT_SYMBOL_GPL_FUTURE()
This patch adds the ability to mark symbols that will be changed in the future, so that kernel modules that don't include MODULE_LICENSE("GPL") and use the symbols, will be flagged and printed out to the system log.
---
arch/m68knommu/kernel/vmlinux.lds.S | 10 ++++ arch/v850/kernel/vmlinux.lds.S | 8 +++ include/asm-generic/vmlinux.lds.h | 14 +++++ include/linux/module.h | 9 +++ kernel/module.c | 49 ++++++++++++++++++- scripts/genksyms/keywords.c_shipped | 91 ++++++++++++++++++----------------- scripts/genksyms/keywords.gperf | 1 7 files changed, 135 insertions(+), 47 deletions(-)
9f28bb7e1d0188a993403ab39b774785892805e1
arch/m68knommu/kernel/vmlinux.lds.S +10 -0
@@ -269,6 +269,11 @@ SECTIONS { *(__ksymtab_gpl) __stop___ksymtab_gpl = .; + /* Kernel symbol table: GPL-future symbols */ + __start___ksymtab_gpl_future = .; + *(__ksymtab_gpl_future) + __stop___ksymtab_gpl_future = .; + /* Kernel symbol table: Normal symbols */ __start___kcrctab = .; *(__kcrctab)
@@ -279,6 +284,11 @@ SECTIONS { *(__kcrctab_gpl) __stop___kcrctab_gpl = .; + /* Kernel symbol table: GPL-future symbols */ + __start___kcrctab_gpl_future = .; + *(__kcrctab_gpl_future) + __stop___kcrctab_gpl_future = .; + /* Kernel symbol table: strings */ *(__ksymtab_strings)
arch/v850/kernel/vmlinux.lds.S +8 -0
@@ -64,6 +64,10 @@ ___start___ksymtab_gpl = .; \ *(__ksymtab_gpl) \ ___stop___ksymtab_gpl = .; \ + /* Kernel symbol table: GPL-future symbols */ \ + ___start___ksymtab_gpl_future = .; \ + *(__ksymtab_gpl_future) \ + ___stop___ksymtab_gpl_future = .; \ /* Kernel symbol table: strings */ \ *(__ksymtab_strings) \ /* Kernel symbol table: Normal symbols */ \
@@ -74,6 +78,10 @@ ___start___kcrctab_gpl = .; \ *(__kcrctab_gpl) \ ___stop___kcrctab_gpl = .; \ + /* Kernel symbol table: GPL-future symbols */ \ + ___start___kcrctab_gpl_future = .; \ + *(__kcrctab_gpl_future) \ + ___stop___kcrctab_gpl_future = .; \ /* Built-in module parameters */ \ . = ALIGN (4) ; \ ___start___param = .; \
include/asm-generic/vmlinux.lds.h +14 -0
@@ -58,6 +58,13 @@ VMLINUX_SYMBOL(__stop___ksymtab_gpl) = .; \ } \ \ +/* Kernel symbol table: GPL-future-only symbols */ \ + __ksymtab_gpl_future : AT(ADDR(__ksymtab_gpl_future) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start___ksymtab_gpl_future) = .; \ + *(__ksymtab_gpl_future ) \ + VMLINUX_SYMBOL(__stop___ksymtab_gpl_future) = .; \ + } \ + \/* Kernel symbol table: Normal symbols */ \ __kcrctab : AT(ADDR(__kcrctab) - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__start___kcrctab) = .; \
@@ -72,6 +79,13 @@ VMLINUX_SYMBOL(__stop___kcrctab_gpl) = .; \ } \ \ +/* Kernel symbol table: GPL-future-only symbols */ \ + __kcrctab_gpl_future : AT(ADDR(__kcrctab_gpl_future) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start___kcrctab_gpl_future) = .; \ + *(__kcrctab_gpl_future ) \ + VMLINUX_SYMBOL(__stop___kcrctab_gpl_future) = .; \ + } \ + \/* Kernel symbol table: strings */ \ __ksymtab_strings : AT(ADDR(__ksymtab_strings) - LOAD_OFFSET) { \ *(__ksymtab_strings) \
include/linux/module.h +9 -0
@@ -198,6 +198,9 @@ void *__symbol_get_gpl(const char *symbo#define EXPORT_SYMBOL_GPL (sym) \ __EXPORT_SYMBOL(sym, "_gpl") +#define EXPORT_SYMBOL_GPL_FUTURE (sym) \ + __EXPORT_SYMBOL(sym, "_gpl_future") +#endif struct module_ref
@@ -255,6 +258,11 @@ struct moduleunsigned int num_gpl_syms;const unsigned long *gpl_crcs; +/* symbols that will be GPL-only in the near future. */ +const struct kernel_symbol *gpl_future_syms; +unsigned int num_gpl_future_syms; +const unsigned long *gpl_future_crcs; +/* Exception table */ unsigned int num_exentries;const struct exception_table_entry *extable;
@@ -441,6 +449,7 @@ void module_remove_driver(struct device_#else /* !CONFIG_MODULES... */ #define EXPORT_SYMBOL (sym)#define EXPORT_SYMBOL_GPL (sym) +#define EXPORT_SYMBOL_GPL_FUTURE (sym)/* Given an address, look for it in the exception tables. */ static inline const struct exception_table_entry *
kernel/module.c +47 -2
92 lines not on this page · raw message
scripts/genksyms/keywords.c_shipped +46 -45
168 lines not on this page · raw message
scripts/genksyms/keywords.gperf +1 -0
@@ -4,6 +4,7 @@ struct resword { const char *name; int t %% EXPORT_SYMBOL, EXPORT_SYMBOL_KEYW EXPORT_SYMBOL_GPL, EXPORT_SYMBOL_KEYW +EXPORT_SYMBOL_GPL_FUTURE, EXPORT_SYMBOL_KEYW __asm, ASM_KEYW __asm__, ASM_KEYW __attribute, ATTRIBUTE_KEYW
Signature · 1 line
1.2.4
PATCH 10/23 add EXPORT_SYMBOL_GPL_FUTURE() to RCU subsystem
As the RCU symbols are going to be changed to GPL in the near future, lets warn users that this is going to happen.
---
kernel/rcupdate.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-)
01ca70dca5c64cb774a8ac2f50bddff21d60169f
kernel/rcupdate.c +3 -3
@@ -609,7 +609,7 @@ module_param(qlowmark, int, 0);module_param (rsinterval, int,0 );#endif EXPORT_SYMBOL_GPL (rcu_batches_completed); -EXPORT_SYMBOL (call_rcu);/* WARNING: GPL-only in April 2006. */ -EXPORT_SYMBOL (call_rcu_bh);/* WARNING: GPL-only in April 2006. */ +EXPORT_SYMBOL_GPL_FUTURE (call_rcu);/* WARNING: GPL-only in April 2006. */ +EXPORT_SYMBOL_GPL_FUTURE (call_rcu_bh);/* WARNING: GPL-only in April 2006. */ EXPORT_SYMBOL_GPL (synchronize_rcu); -EXPORT_SYMBOL (synchronize_kernel);/* WARNING: GPL-only in April 2006. */ +EXPORT_SYMBOL_GPL_FUTURE (synchronize_kernel);/* WARNING: GPL-only in April 2006. */
Signature · 1 line
1.2.4
PATCH 12/23 fix module sysfs files reference counting
In reply to Greg Kroah-HartmanThe module files, refcnt, version, and srcversion did not properly increment the owner's module reference count, allowing the modules to be removed while the files were open, causing oopses.
This patch fixes this, and also fixes the problem that the version and srcversion files were not showing up, unless CONFIG_MODULE_UNLOAD was enabled, which is not correct.
---
include/linux/module.h | 1 + kernel/module.c | 77 +++++++++++++++++++----------------------------- kernel/params.c | 10 ------ 3 files changed, 32 insertions(+), 56 deletions(-)
03e88ae1b13dfdc8bbaa59b8198e1ca53aad12ac
include/linux/module.h +1 -0
@@ -245,6 +245,7 @@ struct module/* Sysfs stuff. */ struct module_kobject mkobj;struct module_param_attrs *param_attrs; +struct module_attribute *modinfo_attrs;const char *version;const char *srcversion;
kernel/module.c +31 -46
158 lines not on this page · raw message
kernel/params.c +0 -10
@@ -638,13 +638,8 @@ static ssize_t module_attr_show(struct kif (!attribute->show)return -EIO ; -if (!try_module_get (mk->mod)) -return -ENODEV ; - ret = attribute->show (attribute, mk->mod, buf); -module_put (mk->mod); -return ret; }
@@ -662,13 +657,8 @@ static ssize_t module_attr_store(structif (!attribute->store)return -EIO ; -if (!try_module_get (mk->mod)) -return -ENODEV ; - ret = attribute->store (attribute, mk->mod, buf, len); -module_put (mk->mod); -return ret; }
Signature · 1 line
1.2.4
PATCH 13/23 sysfs: kzalloc conversion
this converts fs/sysfs to kzalloc() usage.
compile tested with make allyesconfig
---
fs/sysfs/file.c | 3 +-- fs/sysfs/inode.c | 3 +-- 2 files changed, 2 insertions(+), 4 deletions(-)
58d49283b87751f7af75e021a629dcddb027e8eb
fs/sysfs/file.c +1 -2
@@ -301,9 +301,8 @@ static int check_perm(struct inode * ino/* No error? Great, allocate a buffer for the file, and store it * it in file->private_data for easy access. */ - buffer =kmalloc (sizeof (struct sysfs_buffer ),GFP_KERNEL ); + buffer =kzalloc (sizeof (struct sysfs_buffer ),GFP_KERNEL );if (buffer) { -memset (buffer,0 ,sizeof (struct sysfs_buffer ));init_MUTEX (&buffer->sem); buffer->needs_read_fill =1 ; buffer->ops = ops;
fs/sysfs/inode.c +1 -2
@@ -54,11 +54,10 @@ int sysfs_setattr(struct dentry * dentryif (!sd_iattr) {/* setting attributes for the first time, allocate now */ - sd_iattr =kmalloc (sizeof (struct iattr ),GFP_KERNEL ); + sd_iattr =kzalloc (sizeof (struct iattr ),GFP_KERNEL );if (!sd_iattr)return -ENOMEM ;/* assign default attributes */ -memset (sd_iattr,0 ,sizeof (struct iattr )); sd_iattr->ia_mode = sd->s_mode; sd_iattr->ia_uid =0 ; sd_iattr->ia_gid =0 ;
Signature · 1 line
1.2.4
PATCH 14/23 firmware: fix BUG: in fw_realloc_buffer
The fw_realloc_buffer routine does not handle an increase in buffer size of more than 4k. It's not clear to me why it expects that it will only get an extra 4k of data. The attached patch modifies fw_realloc_buffer to vmalloc as much memory as is requested, instead of what we previously had + 4k.
I've tested this on my laptop, which would crash occaisionally on boot without the patch. With the patch, it hasn't crashed, but I can't be certain that this code path is exercised.
---
drivers/base/firmware_class.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-)
30560ba6eda308c13a361d08eb5d4eaab94ab37e
drivers/base/firmware_class.c +4 -2
@@ -211,18 +211,20 @@ static int fw_realloc_buffer(struct firmware_priv *fw_priv, int min_size) {u8 *new_data; +int new_size = fw_priv->alloc_size;if (min_size <= fw_priv->alloc_size)return 0 ; - new_data =vmalloc (fw_priv->alloc_size +PAGE_SIZE ); + new_size =ALIGN (min_size,PAGE_SIZE ); + new_data =vmalloc (new_size);if (!new_data) {printk (KERN_ERR "%s: unable to alloc buffer\n" , __FUNCTION__);/* Make sure that we don't keep incomplete data */ fw_load_abort (fw_priv);return -ENOMEM ; } - fw_priv->alloc_size +=PAGE_SIZE ; + fw_priv->alloc_size = new_size;if (fw_priv->fw->data) {memcpy (new_data, fw_priv->fw->data, fw_priv->fw->size);vfree (fw_priv->fw->data);
Signature · 1 line
1.2.4
PATCH 15/23 Driver core: add macros notice(), dev_notice()
Both usb.h and device.h have collections of convenience macros for printk() with the KERN_ERR, KERN_WARNING, and KERN_NOTICE severity levels. This patch adds macros for the KERN_NOTICE level which was so far uncatered for.
These macros already exist privately in drivers/isdn/gigaset/gigaset.h (currently in the process of being submitted for the kernel tree) but they really belong with their brothers and sisters in include/linux/{device,usb}.h.
---
include/linux/device.h | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
4f2928d0a439553f0288d9483faf417430629635
include/linux/device.h +2 -0
@@ -424,6 +424,8 @@ extern void firmware_unregister(struct s dev_printk(KERN_INFO , dev , format , ## arg) #define dev_warn(dev, format, arg...) \ dev_printk(KERN_WARNING , dev , format , ## arg) +#define dev_notice(dev, format, arg...) \ + dev_printk(KERN_NOTICE , dev , format , ## arg) /* Create alias, so I can be autoloaded. */ #define MODULE_ALIAS_CHARDEV(major,minor) \
Signature · 1 line
1.2.4
PATCH 16/23 Kobject: provide better warning messages when people do stupid things
Now that kobject_add() is used more than kobject_register() the kernel wasn't always letting people know that they were doing something wrong. This change fixes this.
---
lib/kobject.c | 22 ++++++++++++++-------- 1 files changed, 14 insertions(+), 8 deletions(-)
dcd0da002122a70fe1c625c0ca9f58c95aa33ebe
lib/kobject.c +14 -8
@@ -194,6 +194,17 @@ int kobject_add(struct kobject * kobj)unlink (kobj);if (parent)kobject_put (parent); + +/* be noisy on error issues */ +if (error == -EEXIST ) +printk ("kobject_add failed for %s with -EEXIST, " +"don't try to register things with the " +"same name in the same directory.\n" , +kobject_name (kobj)); +else +printk ("kobject_add failed for %s (%d)\n" , +kobject_name (kobj), error); +dump_stack (); }return error;
@@ -207,18 +218,13 @@ int kobject_add(struct kobject * kobj)int kobject_register (struct kobject * kobj) { -int error =0 ; +int error = -EINVAL ;if (kobj) {kobject_init (kobj); error =kobject_add (kobj); -if (error) { -printk ("kobject_register failed for %s (%d)\n" , -kobject_name (kobj),error); -dump_stack (); - }else +if (!error)kobject_uevent (kobj,KOBJ_ADD ); - }else - error = -EINVAL ; + }return error; }
Signature · 1 line
1.2.4
PATCH 17/23 Kobject: kobject.h: fix a typo
It shouldn't cause real harm, but it hurts my eyes.
---
include/linux/kobject.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
22f98c0cd7e003b896ee52ded945081307118745
include/linux/kobject.h +1 -1
@@ -255,7 +255,7 @@ struct subsys_attribute {extern int subsys_create_file (struct subsystem * ,struct subsys_attribute *);extern void subsys_remove_file (struct subsystem * ,struct subsys_attribute *); -#if defined(CONFIG_HOTPLUG ) & defined(CONFIG_NET ) +#if defined(CONFIG_HOTPLUG ) && defined(CONFIG_NET )void kobject_uevent (struct kobject *kobj,enum kobject_action action); int add_uevent_var(char **envp, int num_envp, int *cur_index,
Signature · 1 line
1.2.4
PATCH 18/23 sysfs: fix problem with duplicate sysfs directories and files
The following patch checks for existing sysfs_dirent before preparing new one while creating sysfs directories and files.
---
fs/sysfs/dir.c | 32 +++++++++++++++++++++++++++++++- fs/sysfs/file.c | 6 ++++-- fs/sysfs/symlink.c | 5 +++-- fs/sysfs/sysfs.h | 1 + 4 files changed, 39 insertions(+), 5 deletions(-)
c516865cfbac0d862d4888df91793ad1e74ffd58
fs/sysfs/dir.c +31 -1
@@ -50,6 +50,32 @@ static struct sysfs_dirent * sysfs_new_dreturn sd; } +/** +* +* Return -EEXIST if there is already a sysfs element with the same name for +* the same parent. +* +* called with parent inode's i_mutex held +*/ +int sysfs_dirent_exist (struct sysfs_dirent *parent_sd, +const unsigned char *new) +{ +struct sysfs_dirent * sd; + +list_for_each_entry (sd, &parent_sd->s_children, s_sibling) { +if (sd->s_element) { +const unsigned char *existing =sysfs_get_name (sd); +if (strcmp (existing, new)) +continue ; +else +return -EEXIST ; + } + } + +return 0 ; +} + +int sysfs_make_dirent (struct sysfs_dirent * parent_sd,struct dentry * dentry,void * element,umode_t mode,int type) {
@@ -102,7 +128,11 @@ static int create_dir(struct kobject * kmutex_lock (&p->d_inode->i_mutex); *d =lookup_one_len (n, p,strlen (n));if (!IS_ERR (*d)) { - error =sysfs_make_dirent (p->d_fsdata, *d, k, mode,SYSFS_DIR ); +if (sysfs_dirent_exist (p->d_fsdata, n)) + error = -EEXIST ; +else + error =sysfs_make_dirent (p->d_fsdata, *d, k, mode, +SYSFS_DIR );if (!error) { error =sysfs_create (*d, mode, init_dir);if (!error) {
fs/sysfs/file.c +4 -2
@@ -361,10 +361,12 @@ int sysfs_add_file(struct dentry * dir, {struct sysfs_dirent * parent_sd = dir->d_fsdata;umode_t mode = (attr->mode &S_IALLUGO ) |S_IFREG ; -int error =0 ; +int error = -EEXIST ;mutex_lock (&dir->d_inode->i_mutex); - error =sysfs_make_dirent (parent_sd,NULL , (void *) attr, mode, type); +if (!sysfs_dirent_exist (parent_sd, attr->name)) + error =sysfs_make_dirent (parent_sd,NULL , (void *)attr, + mode, type);mutex_unlock (&dir->d_inode->i_mutex);return error;
fs/sysfs/symlink.c +3 -2
@@ -82,12 +82,13 @@ exit1:int sysfs_create_link (struct kobject * kobj,struct kobject * target,const char * name) {struct dentry * dentry = kobj->dentry; -int error =0 ; +int error = -EEXIST ;BUG_ON (!kobj || !kobj->dentry || !name);mutex_lock (&dentry->d_inode->i_mutex); - error =sysfs_add_link (dentry, name, target); +if (!sysfs_dirent_exist (dentry->d_fsdata, name)) + error =sysfs_add_link (dentry, name, target);mutex_unlock (&dentry->d_inode->i_mutex);return error; }
fs/sysfs/sysfs.h +1 -0
@@ -5,6 +5,7 @@ extern kmem_cache_t *sysfs_dir_cachep;extern struct inode *sysfs_new_inode (mode_t mode,struct sysfs_dirent *);extern int sysfs_create (struct dentry *,int mode,int (*init)(struct inode *)); +extern int sysfs_dirent_exist (struct sysfs_dirent *,const unsigned char *);extern int sysfs_make_dirent (struct sysfs_dirent *,struct dentry *,void *,umode_t ,int );
Signature · 1 line
1.2.4
PATCH 19/23 debugfs: Add debugfs_create_blob() helper for exporting binary data
I wanted to export a binary blob via debugfs, and although it was pretty easy it seems like it'd be easier if there was a helper for it. It's a pity we need the wrapper struct but I can't see a cleaner way to do it.
---
fs/debugfs/file.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/debugfs.h | 15 +++++++++++++++ 2 files changed, 61 insertions(+), 0 deletions(-)
dd308bc355a1aa4f202fe9a3133b6c676cb9606c
fs/debugfs/file.c +46 -0
@@ -251,3 +251,49 @@ struct dentry *debugfs_create_bool(const }EXPORT_SYMBOL_GPL (debugfs_create_bool); +static ssize_t read_file_blob (struct file *file,char __user *user_buf, +size_t count,loff_t *ppos) +{ +struct debugfs_blob_wrapper *blob = file->private_data; +return simple_read_from_buffer (user_buf, count, ppos, blob->data, + blob->size); +} + +static struct file_operations fops_blob = { + .read = read_file_blob, + .open = default_open, +}; + +/** +* debugfs_create_blob - create a file in the debugfs filesystem that is +* used to read and write a binary blob. +* +* @name: a pointer to a string containing the name of the file to create. +* @mode: the permission that the file should have +* @parent: a pointer to the parent dentry for this file. This should be a +* directory dentry if set. If this paramater is NULL, then the +* file will be created in the root of the debugfs filesystem. +* @blob: a pointer to a struct debugfs_blob_wrapper which contains a pointer +* to the blob data and the size of the data. +* +* This function creates a file in debugfs with the given name that exports +* @blob->data as a binary blob. If the @mode variable is so set it can be +* read from. Writing is not supported. +* +* This function will return a pointer to a dentry if it succeeds. This +* pointer must be passed to the debugfs_remove() function when the file is +* to be removed (no automatic cleanup happens if your module is unloaded, +* you are responsible here.) If an error occurs, NULL will be returned. +* +* If debugfs is not enabled in the kernel, the value -ENODEV will be +* returned. It is not wise to check for this value, but rather, check for +* NULL or !NULL instead as to eliminate the need for #ifdef in the calling +* code. +*/ +struct dentry *debugfs_create_blob (const char *name,mode_t mode, +struct dentry *parent, +struct debugfs_blob_wrapper *blob) +{ +return debugfs_create_file (name, mode, parent, blob, &fops_blob); +} +EXPORT_SYMBOL_GPL (debugfs_create_blob);
include/linux/debugfs.h +15 -0
@@ -21,6 +21,11 @@struct file_operations ; +struct debugfs_blob_wrapper { +void *data; +unsigned long size; +}; +#if defined(CONFIG_DEBUG_FS ) struct dentry *debugfs_create_file(const char *name, mode_t mode, struct dentry *parent, void *data,
@@ -39,6 +44,9 @@ struct dentry *debugfs_create_u32(conststruct dentry *debugfs_create_bool (const char *name,mode_t mode,struct dentry *parent,u32 *value); +struct dentry *debugfs_create_blob (const char *name,mode_t mode, +struct dentry *parent, +struct debugfs_blob_wrapper *blob);#else #include <linux/err.h>
@@ -94,6 +102,13 @@ static inline struct dentry *debugfs_crereturn ERR_PTR (-ENODEV ); } +static inline struct dentry *debugfs_create_blob (const char *name,mode_t mode, +struct dentry *parent, +struct debugfs_blob_wrapper *blob) +{ +return ERR_PTR (-ENODEV ); +} +#endif #endif
Signature · 1 line
1.2.4
PATCH 20/23 kobject_add_dir
Adding kobject_add_dir() function which creates a subdirectory for a given kobject.
---
include/linux/kobject.h | 2 ++ lib/kobject.c | 38 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 40 insertions(+), 0 deletions(-)
7423172a50968de1905a61413c52bb070a62f5ce
include/linux/kobject.h +2 -0
@@ -80,6 +80,8 @@ extern void kobject_unregister(struct koextern struct kobject *kobject_get (struct kobject *);extern void kobject_put (struct kobject *); +extern struct kobject *kobject_add_dir (struct kobject *,const char *); +extern char *kobject_get_path (struct kobject *,gfp_t );struct kobj_type {
lib/kobject.c +38 -0
@@ -385,6 +385,44 @@ void kobject_put(struct kobject * kobj) } +static void dir_release (struct kobject *kobj) +{ +kfree (kobj); +} + +static struct kobj_type dir_ktype = { + .release = dir_release, + .sysfs_ops =NULL , + .default_attrs =NULL , +}; + +/** +* kobject_add_dir - add sub directory of object. +* @parent: object in which a directory is created. +* @name: directory name. +* +* Add a plain directory object as child of given object. +*/ +struct kobject *kobject_add_dir (struct kobject *parent,const char *name) +{ +struct kobject *k; + +if (!parent) +return NULL ; + + k =kzalloc (sizeof (*k),GFP_KERNEL ); +if (!k) +return NULL ; + + k->parent = parent; + k->ktype = &dir_ktype; +kobject_set_name (k, name); +kobject_register (k); + +return k; +} +EXPORT_SYMBOL_GPL (kobject_add_dir); +/** * kset_init - initialize a kset for use * @k: kset
Signature · 1 line
1.2.4
PATCH 21/23 get_cpu_sysdev() signedness fix
Doing (int < NR_CPUS) doesn't dtrt if it's negative..
---
drivers/base/cpu.c | 2 +- include/linux/cpu.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
a29d642a4aa99c5234314ab2523281139226c231
drivers/base/cpu.c +1 -1
@@ -141,7 +141,7 @@ int __devinit register_cpu(struct cpu *creturn error; } -struct sys_device *get_cpu_sysdev (int cpu) +struct sys_device *get_cpu_sysdev (unsigned cpu) {if (cpu <NR_CPUS )return cpu_sys_devices[cpu];
include/linux/cpu.h +1 -1
@@ -32,7 +32,7 @@ struct cpu { };extern int register_cpu (struct cpu *,int ,struct node *); -extern struct sys_device *get_cpu_sysdev (int cpu); +extern struct sys_device *get_cpu_sysdev (unsigned cpu);#ifdef CONFIG_HOTPLUG_CPU extern void unregister_cpu (struct cpu *,struct node *);#endif
Signature · 1 line
1.2.4
PATCH 22/23 sysfs: don't export dir symbols
These functions should only be used by the kobject core, and if any driver tries to use them, bad things happen. Unexport them to try to prevent this from happening.
---
fs/sysfs/dir.c | 4 ---- 1 files changed, 0 insertions(+), 4 deletions(-)
832c57e9afa7a263bb2f8ee6d04d527ef6709aae
fs/sysfs/dir.c +0 -4
@@ -510,7 +510,3 @@ struct file_operations sysfs_dir_operati .read = generic_read_dir, .readdir = sysfs_readdir, }; - -EXPORT_SYMBOL_GPL (sysfs_create_dir); -EXPORT_SYMBOL_GPL (sysfs_remove_dir); -EXPORT_SYMBOL_GPL (sysfs_rename_dir);
Signature · 1 line
1.2.4
PATCH 23/23 sysfs: fix a kobject leak in sysfs_add_link on the error path
As pointed out by Oliver Neukum.
---
fs/sysfs/symlink.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)
b3229087c5e08589cea4f5040dab56f7dc11332a
fs/sysfs/symlink.c +1 -0
@@ -66,6 +66,7 @@ static int sysfs_add_link(struct dentryif (!error)return 0 ; +kobject_put (target);kfree (sl->link_name); exit2:kfree (sl);
Signature · 1 line
1.2.4
Re: [PATCH 16/23] Kobject: provide better warning messages when people do stupid things
In reply to Greg Kroah-Hartmanlib/kobject.c int kobject_add(struct kobject * kobj) · line 203
Quoted 7 lines · Greg Kroah-Hartman
lib/kobject.c +7
-0+ +/* be noisy on error issues */ +if (error == -EEXIST ) +printk ("kobject_add failed for %s with -EEXIST, " +"don't try to register things with the " +"same name in the same directory.\n" , +kobject_name (kobj));
This looks like an attempt to put documentation into kernel code. Isn't
it better to add good documentation to the header file just above the
prototype?
When I started using sysfs, I noticed a lack of good comments above prototypes of exported functions.
Signature · 4 lines
Best regards, Artem B. Bityutskiy Oktet Labs (St. Petersburg), Software Engineer. +7 812 4286709 (office) +7 911 2449030 (mobile) E-mail: [email protected], Web: www.oktetlabs.ru
Re: [PATCH 16/23] Kobject: provide better warning messages when people do stupid things
Quoted 12 lines · Artem B. Bityutskiy
lib/kobject.c int kobject_add(struct kobject * kobj) · line 203
Quoted 7 lines · Greg Kroah-Hartman
lib/kobject.c +7
-0+ +/* be noisy on error issues */ +if (error == -EEXIST ) +printk ("kobject_add failed for %s with -EEXIST, " +"don't try to register things with the " +"same name in the same directory.\n" , +kobject_name (kobj));This looks like an attempt to put documentation into kernel code. Isn't
it better to add good documentation to the header file just above the
prototype?
Normally I would agree with you, but it's users that hit this error, when they load a combination of drivers that no developer has in the past. And due to the noise in my inbox, people are still quite confused when this happens. That error message is to help people determine what the real problem is.
Quoted 2 lines
When I started using sysfs, I noticed a lack of good comments above prototypes of exported functions.
I agree, care to provide a patch that fixes this?
thanks,
greg k-h
Re: [PATCH 16/23] Kobject: provide better warning messages when people do stupid things
Quoted 4 lines · Greg KH
Quoted 1 line
When I started using sysfs, I noticed a lack of good comments above prototypes of exported functions.
I agree, care to provide a patch that fixes this?
I fear I have no enough knowledge for that.
Signature · 4 lines
Best regards, Artem B. Bityutskiy Oktet Labs (St. Petersburg), Software Engineer. +7 812 4286709 (office) +7 911 2449030 (mobile) E-mail: [email protected], Web: www.oktetlabs.ru
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
In reply to Greg Kroah-HartmanQuoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.
0 is NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.
Check http://lkml.org/lkml/2005/11/21/211
OG.
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
Quoted 10 lines · Olivier Galibert
Quoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.0 is
NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.
Regardless, it does make sense to make these functions to return a -ve error code as is standard.
David Vrabel
Signature · 4 lines
David Vrabel, Design Engineer Arcom, Clifton Road Tel: +44 (0)1223 411200 ext. 3233 Cambridge CB1 7EA, UK Web: http://www.arcom.com/
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
In reply to Olivier GalibertQuoted 8 lines · Olivier Galibert
Quoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.0 is
NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.
No. That's Linus' _opinion_, which is not applicable to systems without the obviously broken PCI or ISA busses. On such systems, IRQ0 has no special meaning what so ever.
Greg - please continue sending this patch.
Signature · 3 lines
Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: 2.6 Serial core
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
Quoted 5 lines · Russell King
Quoted 4 lines · Olivier Galibert
Quoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.
Quoted 8 lines
Quoted 4 lines
0 is
NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.No. That's Linus' _opinion_, which is not applicable to systems without the obviously broken PCI or ISA busses. On such systems, IRQ0 has no special meaning what so ever.
Do the drivers know?
OG.
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
Quoted 9 lines · Olivier Galibert
Quoted 8 lines · Russell King
Quoted 7 lines · Olivier Galibert
Quoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.0 is
NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
No. That's Linus' _opinion_, which is not applicable to systems without the obviously broken PCI or ISA busses. On such systems, IRQ0 has no special meaning what so ever.
Do the drivers know?
If you look at the following patch in the series, the users of this function have been updated. So the answer is "yes".
Signature · 3 lines
Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: 2.6 Serial core
Re: [PATCH 04/23] driver core: platform_get_irq*(): return -ENXIO on error
Quoted 12 lines · Russell King
Quoted 11 lines · Olivier Galibert
Quoted 10 lines · Russell King
Quoted 8 lines · Olivier Galibert
Quoted 2 lines · Greg Kroah-Hartman
platform_get_irq*() cannot return 0 on error as 0 is a valid IRQ on some platforms, return
-ENXIOinstead.0 is
NO_IRQ, and can not be a valid IRQ number, ever. A platform_get_irq*() returning 0 as a valid irq is buggy.
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
No. That's Linus' _opinion_, which is not applicable to systems without the obviously broken PCI or ISA busses. On such systems, IRQ0 has no special meaning what so ever.
Do the drivers know?
If you look at the following patch in the series, the users of this function have been updated. So the answer is "yes".
They've just been updated not to barf with -ENOENT, nothing more. That's nowhere near an audit of use of the return value.
But whatever, I'm not the one who will have to debug it.
OG.
Re: [PATCH 02/23] kobject: fix build error if CONFIG_SYSFS=n
In reply to Greg Kroah-HartmanHi Greg,
Moving them inside CONFIG_HOTPLUG && CONFIG_NET will break CONFIG_NET=n build, because kernel/ksysfs.c will use them #ifdef CONFIG_HOTPLUG.
Quoted 3 lines · Greg Kroah-Hartman
Moving
uevent_seqnumanduevent_helpertokobject_uevent.cbecause they are used even ifCONFIG_SYSFS=nwhilekernel/ksysfs.cis built only ifCONFIG_SYSFS=y,
...
lib/kobject_uevent.c line 32
Quoted 12 lines
lib/kobject_uevent.c +2
-0@@ -26,6 +26,8 @@#define NUM_ENVP 32/* number of env pointers */ #if defined(CONFIG_HOTPLUG ) && defined(CONFIG_NET ) +u64 uevent_seqnum; +char uevent_helper[UEVENT_HELPER_PATH_LEN ] ="/sbin/hotplug" ;static DEFINE_SPINLOCK (sequence_lock);static struct sock *uevent_sock;
Thanks,
Signature · 1 line
Jun'ichi Nomura, NEC Solutions (America), Inc.