From bd5d080ab99642e3245ef7cfa54490384c01d878 Mon Sep 17 00:00:00 2001 From: Richard Purdie Date: Sun, 13 Nov 2005 10:07:48 +0000 Subject: [ARM] 3160/1: SharpSL: Add driver for Akita specific GPIOs Patch from Richard Purdie Add a driver for the extra GPIOs found on the Sharp SL-C1000 (Akita). These GPIOs are found on a Maxim MAX7310 I2C i/o expander chip. A generic GPIO driver for the MAX7310 was attempted but this mini driver is a much simpler and much more effective solution avoiding several issues and complexity the generic driver had (as discussed on LKML). The platform device is required so the device parent can be set correctly which ensures the device is one of the last to suspend and first to resume. Whilst the i2c suspend/resume calls can be influenced, nothing guarantees this is easlier/later than the subsystems the gpios are used on which are all independent of i2c (sound, irda, video/backlight etc.). Signed-off-by: Richard Purdie Signed-off-by: Russell King --- include/linux/i2c-id.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/linux') diff --git a/include/linux/i2c-id.h b/include/linux/i2c-id.h index 74abaecdb572..1543daaa9c5e 100644 --- a/include/linux/i2c-id.h +++ b/include/linux/i2c-id.h @@ -107,6 +107,7 @@ #define I2C_DRIVERID_CX25840 71 /* cx2584x video encoder */ #define I2C_DRIVERID_SAA7127 72 /* saa7124 video encoder */ #define I2C_DRIVERID_SAA711X 73 /* saa711x video encoders */ +#define I2C_DRIVERID_AKITAIOEXP 74 /* IO Expander on Sharp SL-C1000 */ #define I2C_DRIVERID_EXP0 0xF0 /* experimental use id's */ #define I2C_DRIVERID_EXP1 0xF1 -- cgit v1.2.3-71-gd317 From 95e861db3eaba7bc99f8605db70103ec3d078203 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sun, 13 Nov 2005 16:06:24 -0800 Subject: [PATCH] reorder struct files_struct The file_lock spinlock sits close to mostly read fields of 'struct files_struct' In SMP (and NUMA) environments, each time a thread wants to open or close a file, it has to acquire the spinlock, thus invalidating the cache line containing this spinlock on other CPUS. So other threads doing read()/write()/... calls that use RCU to access the file table are going to ask further memory (possibly NUMA) transactions to read again this memory line. Move the spinlock to another cache line, so that concurrent threads can share the cache line containing 'count' and 'fdt' fields. It's worth up to 9% on a microbenchmark using a 4-thread 2-package x86 machine. See http://marc.theaimsgroup.com/?l=linux-kernel&m=112680448713342&w=2 Signed-off-by: Eric Dumazet Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/file.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/file.h b/include/linux/file.h index d3b1a15d5f21..418b6101b59a 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -33,13 +33,13 @@ struct fdtable { * Open file table structure */ struct files_struct { - atomic_t count; - spinlock_t file_lock; /* Protects all the below members. Nests inside tsk->alloc_lock */ + atomic_t count; struct fdtable *fdt; struct fdtable fdtab; - fd_set close_on_exec_init; - fd_set open_fds_init; - struct file * fd_array[NR_OPEN_DEFAULT]; + fd_set close_on_exec_init; + fd_set open_fds_init; + struct file * fd_array[NR_OPEN_DEFAULT]; + spinlock_t file_lock; /* Protects concurrent writers. Nests inside tsk->alloc_lock */ }; #define files_fdtable(files) (rcu_dereference((files)->fdt)) -- cgit v1.2.3-71-gd317 From bca73e4bf8563d83f7856164caa44d5f42e44cca Mon Sep 17 00:00:00 2001 From: Jeff Garzik Date: Sun, 13 Nov 2005 16:06:25 -0800 Subject: [PATCH] move pm_register/etc. to CONFIG_PM_LEGACY, pm_legacy.h Since few people need the support anymore, this moves the legacy pm_xxx functions to CONFIG_PM_LEGACY, and include/linux/pm_legacy.h. Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/arm/kernel/apm.c | 1 + arch/frv/kernel/pm.c | 1 + arch/i386/Kconfig | 2 +- arch/i386/kernel/apm.c | 1 + arch/mips/au1000/common/power.c | 1 + drivers/acpi/bus.c | 3 ++- drivers/net/3c509.c | 13 +++++----- drivers/net/irda/ali-ircc.c | 1 + drivers/net/irda/nsc-ircc.c | 1 + drivers/serial/68328serial.c | 7 +++--- include/linux/pm.h | 49 ------------------------------------ include/linux/pm_legacy.h | 56 +++++++++++++++++++++++++++++++++++++++++ kernel/power/Kconfig | 9 +++++++ kernel/power/Makefile | 3 ++- kernel/power/pm.c | 1 + sound/oss/ad1848.c | 1 + sound/oss/cs4281/cs4281m.c | 1 + sound/oss/maestro.c | 1 + sound/oss/nm256_audio.c | 1 + sound/oss/opl3sa2.c | 18 +++++++------ 20 files changed, 102 insertions(+), 69 deletions(-) create mode 100644 include/linux/pm_legacy.h (limited to 'include/linux') diff --git a/arch/arm/kernel/apm.c b/arch/arm/kernel/apm.c index b0bbd1e62ebb..a2843be05557 100644 --- a/arch/arm/kernel/apm.c +++ b/arch/arm/kernel/apm.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include #include diff --git a/arch/frv/kernel/pm.c b/arch/frv/kernel/pm.c index 1a1e8a119c3d..712c3c24c954 100644 --- a/arch/frv/kernel/pm.c +++ b/arch/frv/kernel/pm.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include diff --git a/arch/i386/Kconfig b/arch/i386/Kconfig index dbf90ad6eac3..6004bb0795e0 100644 --- a/arch/i386/Kconfig +++ b/arch/i386/Kconfig @@ -699,7 +699,7 @@ depends on PM && !X86_VISWS config APM tristate "APM (Advanced Power Management) BIOS support" - depends on PM + depends on PM && PM_LEGACY ---help--- APM is a BIOS specification for saving power using several different techniques. This is mostly useful for battery powered laptops with diff --git a/arch/i386/kernel/apm.c b/arch/i386/kernel/apm.c index 003548b8735f..1e60acbed3c1 100644 --- a/arch/i386/kernel/apm.c +++ b/arch/i386/kernel/apm.c @@ -218,6 +218,7 @@ #include #include #include +#include #include #include #include diff --git a/arch/mips/au1000/common/power.c b/arch/mips/au1000/common/power.c index f85093b8d54d..f4926315fb68 100644 --- a/arch/mips/au1000/common/power.c +++ b/arch/mips/au1000/common/power.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c index 6a4da417c16b..606f8733a776 100644 --- a/drivers/acpi/bus.c +++ b/drivers/acpi/bus.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #ifdef CONFIG_X86 @@ -754,7 +755,7 @@ static int __init acpi_init(void) result = acpi_bus_init(); if (!result) { -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY if (!PM_IS_ACTIVE()) pm_active = 1; else { diff --git a/drivers/net/3c509.c b/drivers/net/3c509.c index 977935a3d898..824e430486c2 100644 --- a/drivers/net/3c509.c +++ b/drivers/net/3c509.c @@ -84,6 +84,7 @@ static int max_interrupt_work = 10; #include #include #include +#include #include #include /* for udelay() */ #include @@ -173,7 +174,7 @@ struct el3_private { /* skb send-queue */ int head, size; struct sk_buff *queue[SKB_QUEUE_SIZE]; -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY struct pm_dev *pmdev; #endif enum { @@ -200,7 +201,7 @@ static void el3_tx_timeout (struct net_device *dev); static void el3_down(struct net_device *dev); static void el3_up(struct net_device *dev); static struct ethtool_ops ethtool_ops; -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY static int el3_suspend(struct pm_dev *pdev); static int el3_resume(struct pm_dev *pdev); static int el3_pm_callback(struct pm_dev *pdev, pm_request_t rqst, void *data); @@ -361,7 +362,7 @@ static void el3_common_remove (struct net_device *dev) struct el3_private *lp = netdev_priv(dev); (void) lp; /* Keep gcc quiet... */ -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY if (lp->pmdev) pm_unregister(lp->pmdev); #endif @@ -571,7 +572,7 @@ no_pnp: if (err) goto out1; -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY /* register power management */ lp->pmdev = pm_register(PM_ISA_DEV, card_idx, el3_pm_callback); if (lp->pmdev) { @@ -1479,7 +1480,7 @@ el3_up(struct net_device *dev) } /* Power Management support functions */ -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY static int el3_suspend(struct pm_dev *pdev) @@ -1548,7 +1549,7 @@ el3_pm_callback(struct pm_dev *pdev, pm_request_t rqst, void *data) return 0; } -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM_LEGACY */ /* Parameters that may be passed into the module. */ static int debug = -1; diff --git a/drivers/net/irda/ali-ircc.c b/drivers/net/irda/ali-ircc.c index 9bf34681d3df..2e7882eb7d6f 100644 --- a/drivers/net/irda/ali-ircc.c +++ b/drivers/net/irda/ali-ircc.c @@ -40,6 +40,7 @@ #include #include +#include #include #include diff --git a/drivers/net/irda/nsc-ircc.c b/drivers/net/irda/nsc-ircc.c index 805714ec9a8a..ee717d0e939e 100644 --- a/drivers/net/irda/nsc-ircc.c +++ b/drivers/net/irda/nsc-ircc.c @@ -59,6 +59,7 @@ #include #include +#include #include #include diff --git a/drivers/serial/68328serial.c b/drivers/serial/68328serial.c index 2efb317153ce..67e9afa000c1 100644 --- a/drivers/serial/68328serial.c +++ b/drivers/serial/68328serial.c @@ -34,6 +34,7 @@ #include #include #include +#include #include #include @@ -1343,7 +1344,7 @@ static void show_serial_version(void) printk("MC68328 serial driver version 1.00\n"); } -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY /* Serial Power management * The console (currently fixed at line 0) is a special case for power * management because the kernel is so chatty. The console will be @@ -1393,7 +1394,7 @@ void startup_console(void) struct m68k_serial *info = &m68k_soft[0]; startup(info); } -#endif +#endif /* CONFIG_PM_LEGACY */ static struct tty_operations rs_ops = { @@ -1486,7 +1487,7 @@ rs68328_init(void) IRQ_FLG_STD, "M68328_UART", NULL)) panic("Unable to attach 68328 serial interrupt\n"); -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY serial_pm[i] = pm_register(PM_SYS_DEV, PM_SYS_COM, serial_pm_callback); if (serial_pm[i]) serial_pm[i]->data = info; diff --git a/include/linux/pm.h b/include/linux/pm.h index 1514098d156d..5be87ba3b7ac 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -94,55 +94,6 @@ struct pm_dev struct list_head entry; }; -#ifdef CONFIG_PM - -extern int pm_active; - -#define PM_IS_ACTIVE() (pm_active != 0) - -/* - * Register a device with power management - */ -struct pm_dev __deprecated * -pm_register(pm_dev_t type, unsigned long id, pm_callback callback); - -/* - * Unregister a device with power management - */ -void __deprecated pm_unregister(struct pm_dev *dev); - -/* - * Unregister all devices with matching callback - */ -void __deprecated pm_unregister_all(pm_callback callback); - -/* - * Send a request to all devices - */ -int __deprecated pm_send_all(pm_request_t rqst, void *data); - -#else /* CONFIG_PM */ - -#define PM_IS_ACTIVE() 0 - -static inline struct pm_dev *pm_register(pm_dev_t type, - unsigned long id, - pm_callback callback) -{ - return NULL; -} - -static inline void pm_unregister(struct pm_dev *dev) {} - -static inline void pm_unregister_all(pm_callback callback) {} - -static inline int pm_send_all(pm_request_t rqst, void *data) -{ - return 0; -} - -#endif /* CONFIG_PM */ - /* Functions above this comment are list-based old-style power * managment. Please avoid using them. */ diff --git a/include/linux/pm_legacy.h b/include/linux/pm_legacy.h new file mode 100644 index 000000000000..1252b45face1 --- /dev/null +++ b/include/linux/pm_legacy.h @@ -0,0 +1,56 @@ +#ifndef __LINUX_PM_LEGACY_H__ +#define __LINUX_PM_LEGACY_H__ + +#include + +#ifdef CONFIG_PM_LEGACY + +extern int pm_active; + +#define PM_IS_ACTIVE() (pm_active != 0) + +/* + * Register a device with power management + */ +struct pm_dev __deprecated * +pm_register(pm_dev_t type, unsigned long id, pm_callback callback); + +/* + * Unregister a device with power management + */ +void __deprecated pm_unregister(struct pm_dev *dev); + +/* + * Unregister all devices with matching callback + */ +void __deprecated pm_unregister_all(pm_callback callback); + +/* + * Send a request to all devices + */ +int __deprecated pm_send_all(pm_request_t rqst, void *data); + +#else /* CONFIG_PM_LEGACY */ + +#define PM_IS_ACTIVE() 0 + +static inline struct pm_dev *pm_register(pm_dev_t type, + unsigned long id, + pm_callback callback) +{ + return NULL; +} + +static inline void pm_unregister(struct pm_dev *dev) {} + +static inline void pm_unregister_all(pm_callback callback) {} + +static inline int pm_send_all(pm_request_t rqst, void *data) +{ + return 0; +} + +#endif /* CONFIG_PM_LEGACY */ + +#endif /* __LINUX_PM_LEGACY_H__ */ + diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 46a5e5acff97..5ec248cb7f4a 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -19,6 +19,15 @@ config PM will issue the hlt instruction if nothing is to be done, thereby sending the processor to sleep and saving power. +config PM_LEGACY + bool "Legacy Power Management API" + depends on PM + default y + ---help--- + Support for pm_register() and friends. + + If unsure, say Y. + config PM_DEBUG bool "Power Management Debug Support" depends on PM diff --git a/kernel/power/Makefile b/kernel/power/Makefile index c71eb4579c07..04be7d0d96a7 100644 --- a/kernel/power/Makefile +++ b/kernel/power/Makefile @@ -3,7 +3,8 @@ ifeq ($(CONFIG_PM_DEBUG),y) EXTRA_CFLAGS += -DDEBUG endif -obj-y := main.o process.o console.o pm.o +obj-y := main.o process.o console.o +obj-$(CONFIG_PM_LEGACY) += pm.o obj-$(CONFIG_SOFTWARE_SUSPEND) += swsusp.o disk.o snapshot.o obj-$(CONFIG_SUSPEND_SMP) += smp.o diff --git a/kernel/power/pm.c b/kernel/power/pm.c index 159149321b3c..33c508e857dd 100644 --- a/kernel/power/pm.c +++ b/kernel/power/pm.c @@ -23,6 +23,7 @@ #include #include #include +#include #include int pm_active; diff --git a/sound/oss/ad1848.c b/sound/oss/ad1848.c index 7c835abd99bc..3f30c57676c1 100644 --- a/sound/oss/ad1848.c +++ b/sound/oss/ad1848.c @@ -47,6 +47,7 @@ #include #include #include +#include #include #include #include diff --git a/sound/oss/cs4281/cs4281m.c b/sound/oss/cs4281/cs4281m.c index d0d3963e1b83..adc689649fe1 100644 --- a/sound/oss/cs4281/cs4281m.c +++ b/sound/oss/cs4281/cs4281m.c @@ -298,6 +298,7 @@ struct cs4281_state { struct cs4281_pipeline pl[CS4281_NUMBER_OF_PIPELINES]; }; +#include #include "cs4281pm-24.c" #if CSDEBUG diff --git a/sound/oss/maestro.c b/sound/oss/maestro.c index 3dce504e6d6d..3abd3541cbc7 100644 --- a/sound/oss/maestro.c +++ b/sound/oss/maestro.c @@ -231,6 +231,7 @@ #include #include +#include static int maestro_pm_callback(struct pm_dev *dev, pm_request_t rqst, void *d); #include "maestro.h" diff --git a/sound/oss/nm256_audio.c b/sound/oss/nm256_audio.c index 66970062eb36..0ce2c404a730 100644 --- a/sound/oss/nm256_audio.c +++ b/sound/oss/nm256_audio.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include "sound_config.h" diff --git a/sound/oss/opl3sa2.c b/sound/oss/opl3sa2.c index 2efbd865109b..cd41d0e4706a 100644 --- a/sound/oss/opl3sa2.c +++ b/sound/oss/opl3sa2.c @@ -70,6 +70,7 @@ #include #include #include +#include #include "sound_config.h" #include "ad1848.h" @@ -138,7 +139,7 @@ typedef struct { struct pnp_dev* pdev; int activated; /* Whether said devices have been activated */ #endif -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY unsigned int in_suspend; struct pm_dev *pmdev; #endif @@ -341,7 +342,7 @@ static void opl3sa2_mixer_reset(opl3sa2_state_t* devc) } /* Currently only used for power management */ -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY static void opl3sa2_mixer_restore(opl3sa2_state_t* devc) { if (devc) { @@ -354,7 +355,7 @@ static void opl3sa2_mixer_restore(opl3sa2_state_t* devc) } } } -#endif +#endif /* CONFIG_PM_LEGACY */ static inline void arg_to_vol_mono(unsigned int vol, int* value) { @@ -831,7 +832,8 @@ static struct pnp_driver opl3sa2_driver = { /* End of component functions */ -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY + static DEFINE_SPINLOCK(opl3sa2_lock); /* Power Management support functions */ @@ -906,7 +908,7 @@ static int opl3sa2_pm_callback(struct pm_dev *pdev, pm_request_t rqst, void *dat } return 0; } -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM_LEGACY */ /* * Install OPL3-SA2 based card(s). @@ -1019,12 +1021,12 @@ static int __init init_opl3sa2(void) /* ewww =) */ opl3sa2_state[card].card = card; -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY /* register our power management capabilities */ opl3sa2_state[card].pmdev = pm_register(PM_ISA_DEV, card, opl3sa2_pm_callback); if (opl3sa2_state[card].pmdev) opl3sa2_state[card].pmdev->data = &opl3sa2_state[card]; -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM_LEGACY */ /* * Set the Yamaha 3D enhancement mode (aka Ymersion) if asked to and @@ -1081,7 +1083,7 @@ static void __exit cleanup_opl3sa2(void) int card; for(card = 0; card < opl3sa2_cards_num; card++) { -#ifdef CONFIG_PM +#ifdef CONFIG_PM_LEGACY if (opl3sa2_state[card].pmdev) pm_unregister(opl3sa2_state[card].pmdev); #endif -- cgit v1.2.3-71-gd317 From c1986ee9bea3d880bcf0d3f1a31e055778f306c7 Mon Sep 17 00:00:00 2001 From: Harald Welte Date: Sun, 13 Nov 2005 16:06:29 -0800 Subject: [PATCH] New Omnikey Cardman 4000 driver Add new Omnikey Cardman 4000 smartcard reader driver Signed-off-by: Harald Welte Cc: Dominik Brodowski Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- MAINTAINERS | 5 + drivers/char/pcmcia/Kconfig | 11 + drivers/char/pcmcia/Makefile | 1 + drivers/char/pcmcia/cm4000_cs.c | 2078 +++++++++++++++++++++++++++++++++++++++ include/linux/cm4000_cs.h | 66 ++ 5 files changed, 2161 insertions(+) create mode 100644 drivers/char/pcmcia/cm4000_cs.c create mode 100644 include/linux/cm4000_cs.h (limited to 'include/linux') diff --git a/MAINTAINERS b/MAINTAINERS index c1350d7e6789..cc924073f599 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1873,6 +1873,11 @@ L: linux-tr@linuxtr.net W: http://www.linuxtr.net S: Maintained +OMNIKEY CARDMAN 4000 DRIVER +P: Harald Welte +M: laforge@gnumonks.org +S: Maintained + OMNIKEY CARDMAN 4040 DRIVER P: Harald Welte M: laforge@gnumonks.org diff --git a/drivers/char/pcmcia/Kconfig b/drivers/char/pcmcia/Kconfig index e8d41f36635d..27c1179ee527 100644 --- a/drivers/char/pcmcia/Kconfig +++ b/drivers/char/pcmcia/Kconfig @@ -18,6 +18,17 @@ config SYNCLINK_CS The module will be called synclinkmp. If you want to do that, say M here. +config CARDMAN_4000 + tristate "Omnikey Cardman 4000 support" + depends on PCMCIA + help + Enable support for the Omnikey Cardman 4000 PCMCIA Smartcard + reader. + + This kernel driver requires additional userspace support, either + by the vendor-provided PC/SC ifd_handler (http://www.omnikey.com/), + or via the cm4000 backend of OpenCT (http://www.opensc.com/). + config CARDMAN_4040 tristate "Omnikey CardMan 4040 support" depends on PCMCIA diff --git a/drivers/char/pcmcia/Makefile b/drivers/char/pcmcia/Makefile index 35cd766a0a17..0aae20985d57 100644 --- a/drivers/char/pcmcia/Makefile +++ b/drivers/char/pcmcia/Makefile @@ -5,4 +5,5 @@ # obj-$(CONFIG_SYNCLINK_CS) += synclink_cs.o +obj-$(CONFIG_CARDMAN_4000) += cm4000_cs.o obj-$(CONFIG_CARDMAN_4040) += cm4040_cs.o diff --git a/drivers/char/pcmcia/cm4000_cs.c b/drivers/char/pcmcia/cm4000_cs.c new file mode 100644 index 000000000000..ef011ef5dc46 --- /dev/null +++ b/drivers/char/pcmcia/cm4000_cs.c @@ -0,0 +1,2078 @@ + /* + * A driver for the PCMCIA Smartcard Reader "Omnikey CardMan Mobile 4000" + * + * cm4000_cs.c support.linux@omnikey.com + * + * Tue Oct 23 11:32:43 GMT 2001 herp - cleaned up header files + * Sun Jan 20 10:11:15 MET 2002 herp - added modversion header files + * Thu Nov 14 16:34:11 GMT 2002 mh - added PPS functionality + * Tue Nov 19 16:36:27 GMT 2002 mh - added SUSPEND/RESUME functionailty + * Wed Jul 28 12:55:01 CEST 2004 mh - kernel 2.6 adjustments + * + * current version: 2.4.0gm4 + * + * (C) 2000,2001,2002,2003,2004 Omnikey AG + * + * (C) 2005 Harald Welte + * - Adhere to Kernel CodingStyle + * - Port to 2.6.13 "new" style PCMCIA + * - Check for copy_{from,to}_user return values + * - Use nonseekable_open() + * + * All rights reserved. Licensed under dual BSD/GPL license. + */ + +/* #define PCMCIA_DEBUG 6 */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include + +#include + +/* #define ATR_CSUM */ + +#ifdef PCMCIA_DEBUG +#define reader_to_dev(x) (&handle_to_dev(x->link.handle)) +static int pc_debug = PCMCIA_DEBUG; +module_param(pc_debug, int, 0600); +#define DEBUGP(n, rdr, x, args...) do { \ + if (pc_debug >= (n)) \ + dev_printk(KERN_DEBUG, reader_to_dev(rdr), "%s:" x, \ + __FUNCTION__ , ## args); \ + } while (0) +#else +#define DEBUGP(n, rdr, x, args...) +#endif +static char *version = "cm4000_cs.c v2.4.0gm5 - All bugs added by Harald Welte"; + +#define T_1SEC (HZ) +#define T_10MSEC msecs_to_jiffies(10) +#define T_20MSEC msecs_to_jiffies(20) +#define T_40MSEC msecs_to_jiffies(40) +#define T_50MSEC msecs_to_jiffies(50) +#define T_100MSEC msecs_to_jiffies(100) +#define T_500MSEC msecs_to_jiffies(500) + +static void cm4000_detach(dev_link_t *link); +static void cm4000_release(dev_link_t *link); + +static int major; /* major number we get from the kernel */ + +/* note: the first state has to have number 0 always */ + +#define M_FETCH_ATR 0 +#define M_TIMEOUT_WAIT 1 +#define M_READ_ATR_LEN 2 +#define M_READ_ATR 3 +#define M_ATR_PRESENT 4 +#define M_BAD_CARD 5 +#define M_CARDOFF 6 + +#define LOCK_IO 0 +#define LOCK_MONITOR 1 + +#define IS_AUTOPPS_ACT 6 +#define IS_PROCBYTE_PRESENT 7 +#define IS_INVREV 8 +#define IS_ANY_T0 9 +#define IS_ANY_T1 10 +#define IS_ATR_PRESENT 11 +#define IS_ATR_VALID 12 +#define IS_CMM_ABSENT 13 +#define IS_BAD_LENGTH 14 +#define IS_BAD_CSUM 15 +#define IS_BAD_CARD 16 + +#define REG_FLAGS0(x) (x + 0) +#define REG_FLAGS1(x) (x + 1) +#define REG_NUM_BYTES(x) (x + 2) +#define REG_BUF_ADDR(x) (x + 3) +#define REG_BUF_DATA(x) (x + 4) +#define REG_NUM_SEND(x) (x + 5) +#define REG_BAUDRATE(x) (x + 6) +#define REG_STOPBITS(x) (x + 7) + +struct cm4000_dev { + dev_link_t link; /* pcmcia link */ + dev_node_t node; /* OS node (major,minor) */ + + unsigned char atr[MAX_ATR]; + unsigned char rbuf[512]; + unsigned char sbuf[512]; + + wait_queue_head_t devq; /* when removing cardman must not be + zeroed! */ + + wait_queue_head_t ioq; /* if IO is locked, wait on this Q */ + wait_queue_head_t atrq; /* wait for ATR valid */ + wait_queue_head_t readq; /* used by write to wake blk.read */ + + /* warning: do not move this fields. + * initialising to zero depends on it - see ZERO_DEV below. */ + unsigned char atr_csum; + unsigned char atr_len_retry; + unsigned short atr_len; + unsigned short rlen; /* bytes avail. after write */ + unsigned short rpos; /* latest read pos. write zeroes */ + unsigned char procbyte; /* T=0 procedure byte */ + unsigned char mstate; /* state of card monitor */ + unsigned char cwarn; /* slow down warning */ + unsigned char flags0; /* cardman IO-flags 0 */ + unsigned char flags1; /* cardman IO-flags 1 */ + unsigned int mdelay; /* variable monitor speeds, in jiffies */ + + unsigned int baudv; /* baud value for speed */ + unsigned char ta1; + unsigned char proto; /* T=0, T=1, ... */ + unsigned long flags; /* lock+flags (MONITOR,IO,ATR) * for concurrent + access */ + + unsigned char pts[4]; + + struct timer_list timer; /* used to keep monitor running */ + int monitor_running; +}; + +#define ZERO_DEV(dev) \ + memset(&dev->atr_csum,0, \ + sizeof(struct cm4000_dev) - \ + /*link*/ sizeof(dev_link_t) - \ + /*node*/ sizeof(dev_node_t) - \ + /*atr*/ MAX_ATR*sizeof(char) - \ + /*rbuf*/ 512*sizeof(char) - \ + /*sbuf*/ 512*sizeof(char) - \ + /*queue*/ 4*sizeof(wait_queue_head_t)) + +static dev_info_t dev_info = MODULE_NAME; +static dev_link_t *dev_table[CM4000_MAX_DEV]; + +/* This table doesn't use spaces after the comma between fields and thus + * violates CodingStyle. However, I don't really think wrapping it around will + * make it any clearer to read -HW */ +static unsigned char fi_di_table[10][14] = { +/*FI 00 01 02 03 04 05 06 07 08 09 10 11 12 13 */ +/*DI */ +/* 0 */ {0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11}, +/* 1 */ {0x01,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x91,0x11,0x11,0x11,0x11}, +/* 2 */ {0x02,0x12,0x22,0x32,0x11,0x11,0x11,0x11,0x11,0x92,0xA2,0xB2,0x11,0x11}, +/* 3 */ {0x03,0x13,0x23,0x33,0x43,0x53,0x63,0x11,0x11,0x93,0xA3,0xB3,0xC3,0xD3}, +/* 4 */ {0x04,0x14,0x24,0x34,0x44,0x54,0x64,0x11,0x11,0x94,0xA4,0xB4,0xC4,0xD4}, +/* 5 */ {0x00,0x15,0x25,0x35,0x45,0x55,0x65,0x11,0x11,0x95,0xA5,0xB5,0xC5,0xD5}, +/* 6 */ {0x06,0x16,0x26,0x36,0x46,0x56,0x66,0x11,0x11,0x96,0xA6,0xB6,0xC6,0xD6}, +/* 7 */ {0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11,0x11}, +/* 8 */ {0x08,0x11,0x28,0x38,0x48,0x58,0x68,0x11,0x11,0x98,0xA8,0xB8,0xC8,0xD8}, +/* 9 */ {0x09,0x19,0x29,0x39,0x49,0x59,0x69,0x11,0x11,0x99,0xA9,0xB9,0xC9,0xD9} +}; + +#ifndef PCMCIA_DEBUG +#define xoutb outb +#define xinb inb +#else +static inline void xoutb(unsigned char val, unsigned short port) +{ + if (pc_debug >= 7) + printk(KERN_DEBUG "outb(val=%.2x,port=%.4x)\n", val, port); + outb(val, port); +} +static inline unsigned char xinb(unsigned short port) +{ + unsigned char val; + + val = inb(port); + if (pc_debug >= 7) + printk(KERN_DEBUG "%.2x=inb(%.4x)\n", val, port); + + return val; +} +#endif + +#define b_0000 15 +#define b_0001 14 +#define b_0010 13 +#define b_0011 12 +#define b_0100 11 +#define b_0101 10 +#define b_0110 9 +#define b_0111 8 +#define b_1000 7 +#define b_1001 6 +#define b_1010 5 +#define b_1011 4 +#define b_1100 3 +#define b_1101 2 +#define b_1110 1 +#define b_1111 0 + +static unsigned char irtab[16] = { + b_0000, b_1000, b_0100, b_1100, + b_0010, b_1010, b_0110, b_1110, + b_0001, b_1001, b_0101, b_1101, + b_0011, b_1011, b_0111, b_1111 +}; + +static void str_invert_revert(unsigned char *b, int len) +{ + int i; + + for (i = 0; i < len; i++) + b[i] = (irtab[b[i] & 0x0f] << 4) | irtab[b[i] >> 4]; +} + +static unsigned char invert_revert(unsigned char ch) +{ + return (irtab[ch & 0x0f] << 4) | irtab[ch >> 4]; +} + +#define ATRLENCK(dev,pos) \ + if (pos>=dev->atr_len || pos>=MAX_ATR) \ + goto return_0; + +static unsigned int calc_baudv(unsigned char fidi) +{ + unsigned int wcrcf, wbrcf, fi_rfu, di_rfu; + + fi_rfu = 372; + di_rfu = 1; + + /* FI */ + switch ((fidi >> 4) & 0x0F) { + case 0x00: + wcrcf = 372; + break; + case 0x01: + wcrcf = 372; + break; + case 0x02: + wcrcf = 558; + break; + case 0x03: + wcrcf = 744; + break; + case 0x04: + wcrcf = 1116; + break; + case 0x05: + wcrcf = 1488; + break; + case 0x06: + wcrcf = 1860; + break; + case 0x07: + wcrcf = fi_rfu; + break; + case 0x08: + wcrcf = fi_rfu; + break; + case 0x09: + wcrcf = 512; + break; + case 0x0A: + wcrcf = 768; + break; + case 0x0B: + wcrcf = 1024; + break; + case 0x0C: + wcrcf = 1536; + break; + case 0x0D: + wcrcf = 2048; + break; + default: + wcrcf = fi_rfu; + break; + } + + /* DI */ + switch (fidi & 0x0F) { + case 0x00: + wbrcf = di_rfu; + break; + case 0x01: + wbrcf = 1; + break; + case 0x02: + wbrcf = 2; + break; + case 0x03: + wbrcf = 4; + break; + case 0x04: + wbrcf = 8; + break; + case 0x05: + wbrcf = 16; + break; + case 0x06: + wbrcf = 32; + break; + case 0x07: + wbrcf = di_rfu; + break; + case 0x08: + wbrcf = 12; + break; + case 0x09: + wbrcf = 20; + break; + default: + wbrcf = di_rfu; + break; + } + + return (wcrcf / wbrcf); +} + +static unsigned short io_read_num_rec_bytes(ioaddr_t iobase, unsigned short *s) +{ + unsigned short tmp; + + tmp = *s = 0; + do { + *s = tmp; + tmp = inb(REG_NUM_BYTES(iobase)) | + (inb(REG_FLAGS0(iobase)) & 4 ? 0x100 : 0); + } while (tmp != *s); + + return *s; +} + +static int parse_atr(struct cm4000_dev *dev) +{ + unsigned char any_t1, any_t0; + unsigned char ch, ifno; + int ix, done; + + DEBUGP(3, dev, "-> parse_atr: dev->atr_len = %i\n", dev->atr_len); + + if (dev->atr_len < 3) { + DEBUGP(5, dev, "parse_atr: atr_len < 3\n"); + return 0; + } + + if (dev->atr[0] == 0x3f) + set_bit(IS_INVREV, &dev->flags); + else + clear_bit(IS_INVREV, &dev->flags); + ix = 1; + ifno = 1; + ch = dev->atr[1]; + dev->proto = 0; /* XXX PROTO */ + any_t1 = any_t0 = done = 0; + dev->ta1 = 0x11; /* defaults to 9600 baud */ + do { + if (ifno == 1 && (ch & 0x10)) { + /* read first interface byte and TA1 is present */ + dev->ta1 = dev->atr[2]; + DEBUGP(5, dev, "Card says FiDi is 0x%.2x\n", dev->ta1); + ifno++; + } else if ((ifno == 2) && (ch & 0x10)) { /* TA(2) */ + dev->ta1 = 0x11; + ifno++; + } + + DEBUGP(5, dev, "Yi=%.2x\n", ch & 0xf0); + ix += ((ch & 0x10) >> 4) /* no of int.face chars */ + +((ch & 0x20) >> 5) + + ((ch & 0x40) >> 6) + + ((ch & 0x80) >> 7); + /* ATRLENCK(dev,ix); */ + if (ch & 0x80) { /* TDi */ + ch = dev->atr[ix]; + if ((ch & 0x0f)) { + any_t1 = 1; + DEBUGP(5, dev, "card is capable of T=1\n"); + } else { + any_t0 = 1; + DEBUGP(5, dev, "card is capable of T=0\n"); + } + } else + done = 1; + } while (!done); + + DEBUGP(5, dev, "ix=%d noHist=%d any_t1=%d\n", + ix, dev->atr[1] & 15, any_t1); + if (ix + 1 + (dev->atr[1] & 0x0f) + any_t1 != dev->atr_len) { + DEBUGP(5, dev, "length error\n"); + return 0; + } + if (any_t0) + set_bit(IS_ANY_T0, &dev->flags); + + if (any_t1) { /* compute csum */ + dev->atr_csum = 0; +#ifdef ATR_CSUM + for (i = 1; i < dev->atr_len; i++) + dev->atr_csum ^= dev->atr[i]; + if (dev->atr_csum) { + set_bit(IS_BAD_CSUM, &dev->flags); + DEBUGP(5, dev, "bad checksum\n"); + goto return_0; + } +#endif + if (any_t0 == 0) + dev->proto = 1; /* XXX PROTO */ + set_bit(IS_ANY_T1, &dev->flags); + } + + return 1; +} + +struct card_fixup { + char atr[12]; + u_int8_t atr_len; + u_int8_t stopbits; +}; + +static struct card_fixup card_fixups[] = { + { /* ACOS */ + .atr = { 0x3b, 0xb3, 0x11, 0x00, 0x00, 0x41, 0x01 }, + .atr_len = 7, + .stopbits = 0x03, + }, + { /* Motorola */ + .atr = {0x3b, 0x76, 0x13, 0x00, 0x00, 0x80, 0x62, 0x07, + 0x41, 0x81, 0x81 }, + .atr_len = 11, + .stopbits = 0x04, + }, +}; + +static void set_cardparameter(struct cm4000_dev *dev) +{ + int i; + ioaddr_t iobase = dev->link.io.BasePort1; + u_int8_t stopbits = 0x02; /* ISO default */ + + DEBUGP(3, dev, "-> set_cardparameter\n"); + + dev->flags1 = dev->flags1 | (((dev->baudv - 1) & 0x0100) >> 8); + xoutb(dev->flags1, REG_FLAGS1(iobase)); + DEBUGP(5, dev, "flags1 = 0x%02x\n", dev->flags1); + + /* set baudrate */ + xoutb((unsigned char)((dev->baudv - 1) & 0xFF), REG_BAUDRATE(iobase)); + + DEBUGP(5, dev, "baudv = %i -> write 0x%02x\n", dev->baudv, + ((dev->baudv - 1) & 0xFF)); + + /* set stopbits */ + for (i = 0; i < ARRAY_SIZE(card_fixups); i++) { + if (!memcmp(dev->atr, card_fixups[i].atr, + card_fixups[i].atr_len)) + stopbits = card_fixups[i].stopbits; + } + xoutb(stopbits, REG_STOPBITS(iobase)); + + DEBUGP(3, dev, "<- set_cardparameter\n"); +} + +static int set_protocol(struct cm4000_dev *dev, struct ptsreq *ptsreq) +{ + + unsigned long tmp, i; + unsigned short num_bytes_read; + unsigned char pts_reply[4]; + ssize_t rc; + ioaddr_t iobase = dev->link.io.BasePort1; + + rc = 0; + + DEBUGP(3, dev, "-> set_protocol\n"); + DEBUGP(5, dev, "ptsreq->Protocol = 0x%.8x, ptsreq->Flags=0x%.8x, " + "ptsreq->pts1=0x%.2x, ptsreq->pts2=0x%.2x, " + "ptsreq->pts3=0x%.2x\n", (unsigned int)ptsreq->protocol, + (unsigned int)ptsreq->flags, ptsreq->pts1, ptsreq->pts2, + ptsreq->pts3); + + /* Fill PTS structure */ + dev->pts[0] = 0xff; + dev->pts[1] = 0x00; + tmp = ptsreq->protocol; + while ((tmp = (tmp >> 1)) > 0) + dev->pts[1]++; + dev->proto = dev->pts[1]; /* Set new protocol */ + dev->pts[1] = (0x01 << 4) | (dev->pts[1]); + + /* Correct Fi/Di according to CM4000 Fi/Di table */ + DEBUGP(5, dev, "Ta(1) from ATR is 0x%.2x\n", dev->ta1); + /* set Fi/Di according to ATR TA(1) */ + dev->pts[2] = fi_di_table[dev->ta1 & 0x0F][(dev->ta1 >> 4) & 0x0F]; + + /* Calculate PCK character */ + dev->pts[3] = dev->pts[0] ^ dev->pts[1] ^ dev->pts[2]; + + DEBUGP(5, dev, "pts0=%.2x, pts1=%.2x, pts2=%.2x, pts3=%.2x\n", + dev->pts[0], dev->pts[1], dev->pts[2], dev->pts[3]); + + /* check card convention */ + if (test_bit(IS_INVREV, &dev->flags)) + str_invert_revert(dev->pts, 4); + + /* reset SM */ + xoutb(0x80, REG_FLAGS0(iobase)); + + /* Enable access to the message buffer */ + DEBUGP(5, dev, "Enable access to the messages buffer\n"); + dev->flags1 = 0x20 /* T_Active */ + | (test_bit(IS_INVREV, &dev->flags) ? 0x02 : 0x00) /* inv parity */ + | ((dev->baudv >> 8) & 0x01); /* MSB-baud */ + xoutb(dev->flags1, REG_FLAGS1(iobase)); + + DEBUGP(5, dev, "Enable message buffer -> flags1 = 0x%.2x\n", + dev->flags1); + + /* write challenge to the buffer */ + DEBUGP(5, dev, "Write challenge to buffer: "); + for (i = 0; i < 4; i++) { + xoutb(i, REG_BUF_ADDR(iobase)); + xoutb(dev->pts[i], REG_BUF_DATA(iobase)); /* buf data */ +#ifdef PCMCIA_DEBUG + if (pc_debug >= 5) + printk("0x%.2x ", dev->pts[i]); + } + if (pc_debug >= 5) + printk("\n"); +#else + } +#endif + + /* set number of bytes to write */ + DEBUGP(5, dev, "Set number of bytes to write\n"); + xoutb(0x04, REG_NUM_SEND(iobase)); + + /* Trigger CARDMAN CONTROLLER */ + xoutb(0x50, REG_FLAGS0(iobase)); + + /* Monitor progress */ + /* wait for xmit done */ + DEBUGP(5, dev, "Waiting for NumRecBytes getting valid\n"); + + for (i = 0; i < 100; i++) { + if (inb(REG_FLAGS0(iobase)) & 0x08) { + DEBUGP(5, dev, "NumRecBytes is valid\n"); + break; + } + mdelay(10); + } + if (i == 100) { + DEBUGP(5, dev, "Timeout waiting for NumRecBytes getting " + "valid\n"); + rc = -EIO; + goto exit_setprotocol; + } + + DEBUGP(5, dev, "Reading NumRecBytes\n"); + for (i = 0; i < 100; i++) { + io_read_num_rec_bytes(iobase, &num_bytes_read); + if (num_bytes_read >= 4) { + DEBUGP(2, dev, "NumRecBytes = %i\n", num_bytes_read); + break; + } + mdelay(10); + } + + /* check whether it is a short PTS reply? */ + if (num_bytes_read == 3) + i = 0; + + if (i == 100) { + DEBUGP(5, dev, "Timeout reading num_bytes_read\n"); + rc = -EIO; + goto exit_setprotocol; + } + + DEBUGP(5, dev, "Reset the CARDMAN CONTROLLER\n"); + xoutb(0x80, REG_FLAGS0(iobase)); + + /* Read PPS reply */ + DEBUGP(5, dev, "Read PPS reply\n"); + for (i = 0; i < num_bytes_read; i++) { + xoutb(i, REG_BUF_ADDR(iobase)); + pts_reply[i] = inb(REG_BUF_DATA(iobase)); + } + +#ifdef PCMCIA_DEBUG + DEBUGP(2, dev, "PTSreply: "); + for (i = 0; i < num_bytes_read; i++) { + if (pc_debug >= 5) + printk("0x%.2x ", pts_reply[i]); + } + printk("\n"); +#endif /* PCMCIA_DEBUG */ + + DEBUGP(5, dev, "Clear Tactive in Flags1\n"); + xoutb(0x20, REG_FLAGS1(iobase)); + + /* Compare ptsreq and ptsreply */ + if ((dev->pts[0] == pts_reply[0]) && + (dev->pts[1] == pts_reply[1]) && + (dev->pts[2] == pts_reply[2]) && (dev->pts[3] == pts_reply[3])) { + /* setcardparameter according to PPS */ + dev->baudv = calc_baudv(dev->pts[2]); + set_cardparameter(dev); + } else if ((dev->pts[0] == pts_reply[0]) && + ((dev->pts[1] & 0xef) == pts_reply[1]) && + ((pts_reply[0] ^ pts_reply[1]) == pts_reply[2])) { + /* short PTS reply, set card parameter to default values */ + dev->baudv = calc_baudv(0x11); + set_cardparameter(dev); + } else + rc = -EIO; + +exit_setprotocol: + DEBUGP(3, dev, "<- set_protocol\n"); + return rc; +} + +static int io_detect_cm4000(ioaddr_t iobase, struct cm4000_dev *dev) +{ + + /* note: statemachine is assumed to be reset */ + if (inb(REG_FLAGS0(iobase)) & 8) { + clear_bit(IS_ATR_VALID, &dev->flags); + set_bit(IS_CMM_ABSENT, &dev->flags); + return 0; /* detect CMM = 1 -> failure */ + } + /* xoutb(0x40, REG_FLAGS1(iobase)); detectCMM */ + xoutb(dev->flags1 | 0x40, REG_FLAGS1(iobase)); + if ((inb(REG_FLAGS0(iobase)) & 8) == 0) { + clear_bit(IS_ATR_VALID, &dev->flags); + set_bit(IS_CMM_ABSENT, &dev->flags); + return 0; /* detect CMM=0 -> failure */ + } + /* clear detectCMM again by restoring original flags1 */ + xoutb(dev->flags1, REG_FLAGS1(iobase)); + return 1; +} + +static void terminate_monitor(struct cm4000_dev *dev) +{ + + /* tell the monitor to stop and wait until + * it terminates. + */ + DEBUGP(3, dev, "-> terminate_monitor\n"); + wait_event_interruptible(dev->devq, + test_and_set_bit(LOCK_MONITOR, + (void *)&dev->flags)); + + /* now, LOCK_MONITOR has been set. + * allow a last cycle in the monitor. + * the monitor will indicate that it has + * finished by clearing this bit. + */ + DEBUGP(5, dev, "Now allow last cycle of monitor!\n"); + while (test_bit(LOCK_MONITOR, (void *)&dev->flags)) + msleep(25); + + DEBUGP(5, dev, "Delete timer\n"); + del_timer_sync(&dev->timer); +#ifdef PCMCIA_DEBUG + dev->monitor_running = 0; +#endif + + DEBUGP(3, dev, "<- terminate_monitor\n"); +} + +/* + * monitor the card every 50msec. as a side-effect, retrieve the + * atr once a card is inserted. another side-effect of retrieving the + * atr is that the card will be powered on, so there is no need to + * power on the card explictely from the application: the driver + * is already doing that for you. + */ + +static void monitor_card(unsigned long p) +{ + struct cm4000_dev *dev = (struct cm4000_dev *) p; + ioaddr_t iobase = dev->link.io.BasePort1; + unsigned short s; + struct ptsreq ptsreq; + int i, atrc; + + DEBUGP(7, dev, "-> monitor_card\n"); + + /* if someone has set the lock for us: we're done! */ + if (test_and_set_bit(LOCK_MONITOR, &dev->flags)) { + DEBUGP(4, dev, "About to stop monitor\n"); + /* no */ + dev->rlen = + dev->rpos = + dev->atr_csum = dev->atr_len_retry = dev->cwarn = 0; + dev->mstate = M_FETCH_ATR; + clear_bit(LOCK_MONITOR, &dev->flags); + /* close et al. are sleeping on devq, so wake it */ + wake_up_interruptible(&dev->devq); + DEBUGP(2, dev, "<- monitor_card (we are done now)\n"); + return; + } + + /* try to lock io: if it is already locked, just add another timer */ + if (test_and_set_bit(LOCK_IO, (void *)&dev->flags)) { + DEBUGP(4, dev, "Couldn't get IO lock\n"); + goto return_with_timer; + } + + /* is a card/a reader inserted at all ? */ + dev->flags0 = xinb(REG_FLAGS0(iobase)); + DEBUGP(7, dev, "dev->flags0 = 0x%2x\n", dev->flags0); + DEBUGP(7, dev, "smartcard present: %s\n", + dev->flags0 & 1 ? "yes" : "no"); + DEBUGP(7, dev, "cardman present: %s\n", + dev->flags0 == 0xff ? "no" : "yes"); + + if ((dev->flags0 & 1) == 0 /* no smartcard inserted */ + || dev->flags0 == 0xff) { /* no cardman inserted */ + /* no */ + dev->rlen = + dev->rpos = + dev->atr_csum = dev->atr_len_retry = dev->cwarn = 0; + dev->mstate = M_FETCH_ATR; + + dev->flags &= 0x000000ff; /* only keep IO and MONITOR locks */ + + if (dev->flags0 == 0xff) { + DEBUGP(4, dev, "set IS_CMM_ABSENT bit\n"); + set_bit(IS_CMM_ABSENT, &dev->flags); + } else if (test_bit(IS_CMM_ABSENT, &dev->flags)) { + DEBUGP(4, dev, "clear IS_CMM_ABSENT bit " + "(card is removed)\n"); + clear_bit(IS_CMM_ABSENT, &dev->flags); + } + + goto release_io; + } else if ((dev->flags0 & 1) && test_bit(IS_CMM_ABSENT, &dev->flags)) { + /* cardman and card present but cardman was absent before + * (after suspend with inserted card) */ + DEBUGP(4, dev, "clear IS_CMM_ABSENT bit (card is inserted)\n"); + clear_bit(IS_CMM_ABSENT, &dev->flags); + } + + if (test_bit(IS_ATR_VALID, &dev->flags) == 1) { + DEBUGP(7, dev, "believe ATR is already valid (do nothing)\n"); + goto release_io; + } + + switch (dev->mstate) { + unsigned char flags0; + case M_CARDOFF: + DEBUGP(4, dev, "M_CARDOFF\n"); + flags0 = inb(REG_FLAGS0(iobase)); + if (flags0 & 0x02) { + /* wait until Flags0 indicate power is off */ + dev->mdelay = T_10MSEC; + } else { + /* Flags0 indicate power off and no card inserted now; + * Reset CARDMAN CONTROLLER */ + xoutb(0x80, REG_FLAGS0(iobase)); + + /* prepare for fetching ATR again: after card off ATR + * is read again automatically */ + dev->rlen = + dev->rpos = + dev->atr_csum = + dev->atr_len_retry = dev->cwarn = 0; + dev->mstate = M_FETCH_ATR; + + /* minimal gap between CARDOFF and read ATR is 50msec */ + dev->mdelay = T_50MSEC; + } + break; + case M_FETCH_ATR: + DEBUGP(4, dev, "M_FETCH_ATR\n"); + xoutb(0x80, REG_FLAGS0(iobase)); + DEBUGP(4, dev, "Reset BAUDV to 9600\n"); + dev->baudv = 0x173; /* 9600 */ + xoutb(0x02, REG_STOPBITS(iobase)); /* stopbits=2 */ + xoutb(0x73, REG_BAUDRATE(iobase)); /* baud value */ + xoutb(0x21, REG_FLAGS1(iobase)); /* T_Active=1, baud + value */ + /* warm start vs. power on: */ + xoutb(dev->flags0 & 2 ? 0x46 : 0x44, REG_FLAGS0(iobase)); + dev->mdelay = T_40MSEC; + dev->mstate = M_TIMEOUT_WAIT; + break; + case M_TIMEOUT_WAIT: + DEBUGP(4, dev, "M_TIMEOUT_WAIT\n"); + /* numRecBytes */ + io_read_num_rec_bytes(iobase, &dev->atr_len); + dev->mdelay = T_10MSEC; + dev->mstate = M_READ_ATR_LEN; + break; + case M_READ_ATR_LEN: + DEBUGP(4, dev, "M_READ_ATR_LEN\n"); + /* infinite loop possible, since there is no timeout */ + +#define MAX_ATR_LEN_RETRY 100 + + if (dev->atr_len == io_read_num_rec_bytes(iobase, &s)) { + if (dev->atr_len_retry++ >= MAX_ATR_LEN_RETRY) { /* + XX msec */ + dev->mdelay = T_10MSEC; + dev->mstate = M_READ_ATR; + } + } else { + dev->atr_len = s; + dev->atr_len_retry = 0; /* set new timeout */ + } + + DEBUGP(4, dev, "Current ATR_LEN = %i\n", dev->atr_len); + break; + case M_READ_ATR: + DEBUGP(4, dev, "M_READ_ATR\n"); + xoutb(0x80, REG_FLAGS0(iobase)); /* reset SM */ + for (i = 0; i < dev->atr_len; i++) { + xoutb(i, REG_BUF_ADDR(iobase)); + dev->atr[i] = inb(REG_BUF_DATA(iobase)); + } + /* Deactivate T_Active flags */ + DEBUGP(4, dev, "Deactivate T_Active flags\n"); + dev->flags1 = 0x01; + xoutb(dev->flags1, REG_FLAGS1(iobase)); + + /* atr is present (which doesnt mean it's valid) */ + set_bit(IS_ATR_PRESENT, &dev->flags); + if (dev->atr[0] == 0x03) + str_invert_revert(dev->atr, dev->atr_len); + atrc = parse_atr(dev); + if (atrc == 0) { /* atr invalid */ + dev->mdelay = 0; + dev->mstate = M_BAD_CARD; + } else { + dev->mdelay = T_50MSEC; + dev->mstate = M_ATR_PRESENT; + set_bit(IS_ATR_VALID, &dev->flags); + } + + if (test_bit(IS_ATR_VALID, &dev->flags) == 1) { + DEBUGP(4, dev, "monitor_card: ATR valid\n"); + /* if ta1 == 0x11, no PPS necessary (default values) */ + /* do not do PPS with multi protocol cards */ + if ((test_bit(IS_AUTOPPS_ACT, &dev->flags) == 0) && + (dev->ta1 != 0x11) && + !(test_bit(IS_ANY_T0, &dev->flags) && + test_bit(IS_ANY_T1, &dev->flags))) { + DEBUGP(4, dev, "Perform AUTOPPS\n"); + set_bit(IS_AUTOPPS_ACT, &dev->flags); + ptsreq.protocol = ptsreq.protocol = + (0x01 << dev->proto); + ptsreq.flags = 0x01; + ptsreq.pts1 = 0x00; + ptsreq.pts2 = 0x00; + ptsreq.pts3 = 0x00; + if (set_protocol(dev, &ptsreq) == 0) { + DEBUGP(4, dev, "AUTOPPS ret SUCC\n"); + clear_bit(IS_AUTOPPS_ACT, &dev->flags); + wake_up_interruptible(&dev->atrq); + } else { + DEBUGP(4, dev, "AUTOPPS failed: " + "repower using defaults\n"); + /* prepare for repowering */ + clear_bit(IS_ATR_PRESENT, &dev->flags); + clear_bit(IS_ATR_VALID, &dev->flags); + dev->rlen = + dev->rpos = + dev->atr_csum = + dev->atr_len_retry = dev->cwarn = 0; + dev->mstate = M_FETCH_ATR; + + dev->mdelay = T_50MSEC; + } + } else { + /* for cards which use slightly different + * params (extra guard time) */ + set_cardparameter(dev); + if (test_bit(IS_AUTOPPS_ACT, &dev->flags) == 1) + DEBUGP(4, dev, "AUTOPPS already active " + "2nd try:use default values\n"); + if (dev->ta1 == 0x11) + DEBUGP(4, dev, "No AUTOPPS necessary " + "TA(1)==0x11\n"); + if (test_bit(IS_ANY_T0, &dev->flags) + && test_bit(IS_ANY_T1, &dev->flags)) + DEBUGP(4, dev, "Do NOT perform AUTOPPS " + "with multiprotocol cards\n"); + clear_bit(IS_AUTOPPS_ACT, &dev->flags); + wake_up_interruptible(&dev->atrq); + } + } else { + DEBUGP(4, dev, "ATR invalid\n"); + wake_up_interruptible(&dev->atrq); + } + break; + case M_BAD_CARD: + DEBUGP(4, dev, "M_BAD_CARD\n"); + /* slow down warning, but prompt immediately after insertion */ + if (dev->cwarn == 0 || dev->cwarn == 10) { + set_bit(IS_BAD_CARD, &dev->flags); + printk(KERN_WARNING MODULE_NAME ": device %s: ", + dev->node.dev_name); + if (test_bit(IS_BAD_CSUM, &dev->flags)) { + DEBUGP(4, dev, "ATR checksum (0x%.2x, should " + "be zero) failed\n", dev->atr_csum); + } +#ifdef PCMCIA_DEBUG + else if (test_bit(IS_BAD_LENGTH, &dev->flags)) { + DEBUGP(4, dev, "ATR length error\n"); + } else { + DEBUGP(4, dev, "card damaged or wrong way " + "inserted\n"); + } +#endif + dev->cwarn = 0; + wake_up_interruptible(&dev->atrq); /* wake open */ + } + dev->cwarn++; + dev->mdelay = T_100MSEC; + dev->mstate = M_FETCH_ATR; + break; + default: + DEBUGP(7, dev, "Unknown action\n"); + break; /* nothing */ + } + +release_io: + DEBUGP(7, dev, "release_io\n"); + clear_bit(LOCK_IO, &dev->flags); + wake_up_interruptible(&dev->ioq); /* whoever needs IO */ + +return_with_timer: + DEBUGP(7, dev, "<- monitor_card (returns with timer)\n"); + dev->timer.expires = jiffies + dev->mdelay; + add_timer(&dev->timer); + clear_bit(LOCK_MONITOR, &dev->flags); +} + +/* Interface to userland (file_operations) */ + +static ssize_t cmm_read(struct file *filp, __user char *buf, size_t count, + loff_t *ppos) +{ + struct cm4000_dev *dev = filp->private_data; + ioaddr_t iobase = dev->link.io.BasePort1; + ssize_t rc; + int i, j, k; + + DEBUGP(2, dev, "-> cmm_read(%s,%d)\n", current->comm, current->pid); + + if (count == 0) /* according to manpage */ + return 0; + + if ((dev->link.state & DEV_PRESENT) == 0 || /* socket removed */ + test_bit(IS_CMM_ABSENT, &dev->flags)) + return -ENODEV; + + if (test_bit(IS_BAD_CSUM, &dev->flags)) + return -EIO; + + /* also see the note about this in cmm_write */ + if (wait_event_interruptible + (dev->atrq, + ((filp->f_flags & O_NONBLOCK) + || (test_bit(IS_ATR_PRESENT, (void *)&dev->flags) != 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + if (test_bit(IS_ATR_VALID, &dev->flags) == 0) + return -EIO; + + /* this one implements blocking IO */ + if (wait_event_interruptible + (dev->readq, + ((filp->f_flags & O_NONBLOCK) || (dev->rpos < dev->rlen)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + /* lock io */ + if (wait_event_interruptible + (dev->ioq, + ((filp->f_flags & O_NONBLOCK) + || (test_and_set_bit(LOCK_IO, (void *)&dev->flags) == 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + rc = 0; + dev->flags0 = inb(REG_FLAGS0(iobase)); + if ((dev->flags0 & 1) == 0 /* no smartcard inserted */ + || dev->flags0 == 0xff) { /* no cardman inserted */ + clear_bit(IS_ATR_VALID, &dev->flags); + if (dev->flags0 & 1) { + set_bit(IS_CMM_ABSENT, &dev->flags); + rc = -ENODEV; + } + rc = -EIO; + goto release_io; + } + + DEBUGP(4, dev, "begin read answer\n"); + j = min(count, (size_t)(dev->rlen - dev->rpos)); + k = dev->rpos; + if (k + j > 255) + j = 256 - k; + DEBUGP(4, dev, "read1 j=%d\n", j); + for (i = 0; i < j; i++) { + xoutb(k++, REG_BUF_ADDR(iobase)); + dev->rbuf[i] = xinb(REG_BUF_DATA(iobase)); + } + j = min(count, (size_t)(dev->rlen - dev->rpos)); + if (k + j > 255) { + DEBUGP(4, dev, "read2 j=%d\n", j); + dev->flags1 |= 0x10; /* MSB buf addr set */ + xoutb(dev->flags1, REG_FLAGS1(iobase)); + for (; i < j; i++) { + xoutb(k++, REG_BUF_ADDR(iobase)); + dev->rbuf[i] = xinb(REG_BUF_DATA(iobase)); + } + } + + if (dev->proto == 0 && count > dev->rlen - dev->rpos) { + DEBUGP(4, dev, "T=0 and count > buffer\n"); + dev->rbuf[i] = dev->rbuf[i - 1]; + dev->rbuf[i - 1] = dev->procbyte; + j++; + } + count = j; + + dev->rpos = dev->rlen + 1; + + /* Clear T1Active */ + DEBUGP(4, dev, "Clear T1Active\n"); + dev->flags1 &= 0xdf; + xoutb(dev->flags1, REG_FLAGS1(iobase)); + + xoutb(0, REG_FLAGS1(iobase)); /* clear detectCMM */ + /* last check before exit */ + if (!io_detect_cm4000(iobase, dev)) + count = -ENODEV; + + if (test_bit(IS_INVREV, &dev->flags) && count > 0) + str_invert_revert(dev->rbuf, count); + + if (copy_to_user(buf, dev->rbuf, count)) + return -EFAULT; + +release_io: + clear_bit(LOCK_IO, &dev->flags); + wake_up_interruptible(&dev->ioq); + + DEBUGP(2, dev, "<- cmm_read returns: rc = %Zi\n", + (rc < 0 ? rc : count)); + return rc < 0 ? rc : count; +} + +static ssize_t cmm_write(struct file *filp, const char __user *buf, + size_t count, loff_t *ppos) +{ + struct cm4000_dev *dev = (struct cm4000_dev *) filp->private_data; + ioaddr_t iobase = dev->link.io.BasePort1; + unsigned short s; + unsigned char tmp; + unsigned char infolen; + unsigned char sendT0; + unsigned short nsend; + unsigned short nr; + ssize_t rc; + int i; + + DEBUGP(2, dev, "-> cmm_write(%s,%d)\n", current->comm, current->pid); + + if (count == 0) /* according to manpage */ + return 0; + + if (dev->proto == 0 && count < 4) { + /* T0 must have at least 4 bytes */ + DEBUGP(4, dev, "T0 short write\n"); + return -EIO; + } + + nr = count & 0x1ff; /* max bytes to write */ + + sendT0 = dev->proto ? 0 : nr > 5 ? 0x08 : 0; + + if ((dev->link.state & DEV_PRESENT) == 0 || /* socket removed */ + test_bit(IS_CMM_ABSENT, &dev->flags)) + return -ENODEV; + + if (test_bit(IS_BAD_CSUM, &dev->flags)) { + DEBUGP(4, dev, "bad csum\n"); + return -EIO; + } + + /* + * wait for atr to become valid. + * note: it is important to lock this code. if we dont, the monitor + * could be run between test_bit and the the call the sleep on the + * atr-queue. if *then* the monitor detects atr valid, it will wake up + * any process on the atr-queue, *but* since we have been interrupted, + * we do not yet sleep on this queue. this would result in a missed + * wake_up and the calling process would sleep forever (until + * interrupted). also, do *not* restore_flags before sleep_on, because + * this could result in the same situation! + */ + if (wait_event_interruptible + (dev->atrq, + ((filp->f_flags & O_NONBLOCK) + || (test_bit(IS_ATR_PRESENT, (void *)&dev->flags) != 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + if (test_bit(IS_ATR_VALID, &dev->flags) == 0) { /* invalid atr */ + DEBUGP(4, dev, "invalid ATR\n"); + return -EIO; + } + + /* lock io */ + if (wait_event_interruptible + (dev->ioq, + ((filp->f_flags & O_NONBLOCK) + || (test_and_set_bit(LOCK_IO, (void *)&dev->flags) == 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + if (copy_from_user(dev->sbuf, buf, ((count > 512) ? 512 : count))) + return -EFAULT; + + rc = 0; + dev->flags0 = inb(REG_FLAGS0(iobase)); + if ((dev->flags0 & 1) == 0 /* no smartcard inserted */ + || dev->flags0 == 0xff) { /* no cardman inserted */ + clear_bit(IS_ATR_VALID, &dev->flags); + if (dev->flags0 & 1) { + set_bit(IS_CMM_ABSENT, &dev->flags); + rc = -ENODEV; + } else { + DEBUGP(4, dev, "IO error\n"); + rc = -EIO; + } + goto release_io; + } + + xoutb(0x80, REG_FLAGS0(iobase)); /* reset SM */ + + if (!io_detect_cm4000(iobase, dev)) { + rc = -ENODEV; + goto release_io; + } + + /* reflect T=0 send/read mode in flags1 */ + dev->flags1 |= (sendT0); + + set_cardparameter(dev); + + /* dummy read, reset flag procedure received */ + tmp = inb(REG_FLAGS1(iobase)); + + dev->flags1 = 0x20 /* T_Active */ + | (sendT0) + | (test_bit(IS_INVREV, &dev->flags) ? 2 : 0)/* inverse parity */ + | (((dev->baudv - 1) & 0x0100) >> 8); /* MSB-Baud */ + DEBUGP(1, dev, "set dev->flags1 = 0x%.2x\n", dev->flags1); + xoutb(dev->flags1, REG_FLAGS1(iobase)); + + /* xmit data */ + DEBUGP(4, dev, "Xmit data\n"); + for (i = 0; i < nr; i++) { + if (i >= 256) { + dev->flags1 = 0x20 /* T_Active */ + | (sendT0) /* SendT0 */ + /* inverse parity: */ + | (test_bit(IS_INVREV, &dev->flags) ? 2 : 0) + | (((dev->baudv - 1) & 0x0100) >> 8) /* MSB-Baud */ + | 0x10; /* set address high */ + DEBUGP(4, dev, "dev->flags = 0x%.2x - set address " + "high\n", dev->flags1); + xoutb(dev->flags1, REG_FLAGS1(iobase)); + } + if (test_bit(IS_INVREV, &dev->flags)) { + DEBUGP(4, dev, "Apply inverse convention for 0x%.2x " + "-> 0x%.2x\n", (unsigned char)dev->sbuf[i], + invert_revert(dev->sbuf[i])); + xoutb(i, REG_BUF_ADDR(iobase)); + xoutb(invert_revert(dev->sbuf[i]), + REG_BUF_DATA(iobase)); + } else { + xoutb(i, REG_BUF_ADDR(iobase)); + xoutb(dev->sbuf[i], REG_BUF_DATA(iobase)); + } + } + DEBUGP(4, dev, "Xmit done\n"); + + if (dev->proto == 0) { + /* T=0 proto: 0 byte reply */ + if (nr == 4) { + DEBUGP(4, dev, "T=0 assumes 0 byte reply\n"); + xoutb(i, REG_BUF_ADDR(iobase)); + if (test_bit(IS_INVREV, &dev->flags)) + xoutb(0xff, REG_BUF_DATA(iobase)); + else + xoutb(0x00, REG_BUF_DATA(iobase)); + } + + /* numSendBytes */ + if (sendT0) + nsend = nr; + else { + if (nr == 4) + nsend = 5; + else { + nsend = 5 + (unsigned char)dev->sbuf[4]; + if (dev->sbuf[4] == 0) + nsend += 0x100; + } + } + } else + nsend = nr; + + /* T0: output procedure byte */ + if (test_bit(IS_INVREV, &dev->flags)) { + DEBUGP(4, dev, "T=0 set Procedure byte (inverse-reverse) " + "0x%.2x\n", invert_revert(dev->sbuf[1])); + xoutb(invert_revert(dev->sbuf[1]), REG_NUM_BYTES(iobase)); + } else { + DEBUGP(4, dev, "T=0 set Procedure byte 0x%.2x\n", dev->sbuf[1]); + xoutb(dev->sbuf[1], REG_NUM_BYTES(iobase)); + } + + DEBUGP(1, dev, "set NumSendBytes = 0x%.2x\n", + (unsigned char)(nsend & 0xff)); + xoutb((unsigned char)(nsend & 0xff), REG_NUM_SEND(iobase)); + + DEBUGP(1, dev, "Trigger CARDMAN CONTROLLER (0x%.2x)\n", + 0x40 /* SM_Active */ + | (dev->flags0 & 2 ? 0 : 4) /* power on if needed */ + |(dev->proto ? 0x10 : 0x08) /* T=1/T=0 */ + |(nsend & 0x100) >> 8 /* MSB numSendBytes */ ); + xoutb(0x40 /* SM_Active */ + | (dev->flags0 & 2 ? 0 : 4) /* power on if needed */ + |(dev->proto ? 0x10 : 0x08) /* T=1/T=0 */ + |(nsend & 0x100) >> 8, /* MSB numSendBytes */ + REG_FLAGS0(iobase)); + + /* wait for xmit done */ + if (dev->proto == 1) { + DEBUGP(4, dev, "Wait for xmit done\n"); + for (i = 0; i < 1000; i++) { + if (inb(REG_FLAGS0(iobase)) & 0x08) + break; + msleep_interruptible(10); + } + if (i == 1000) { + DEBUGP(4, dev, "timeout waiting for xmit done\n"); + rc = -EIO; + goto release_io; + } + } + + /* T=1: wait for infoLen */ + + infolen = 0; + if (dev->proto) { + /* wait until infoLen is valid */ + for (i = 0; i < 6000; i++) { /* max waiting time of 1 min */ + io_read_num_rec_bytes(iobase, &s); + if (s >= 3) { + infolen = inb(REG_FLAGS1(iobase)); + DEBUGP(4, dev, "infolen=%d\n", infolen); + break; + } + msleep_interruptible(10); + } + if (i == 6000) { + DEBUGP(4, dev, "timeout waiting for infoLen\n"); + rc = -EIO; + goto release_io; + } + } else + clear_bit(IS_PROCBYTE_PRESENT, &dev->flags); + + /* numRecBytes | bit9 of numRecytes */ + io_read_num_rec_bytes(iobase, &dev->rlen); + for (i = 0; i < 600; i++) { /* max waiting time of 2 sec */ + if (dev->proto) { + if (dev->rlen >= infolen + 4) + break; + } + msleep_interruptible(10); + /* numRecBytes | bit9 of numRecytes */ + io_read_num_rec_bytes(iobase, &s); + if (s > dev->rlen) { + DEBUGP(1, dev, "NumRecBytes inc (reset timeout)\n"); + i = 0; /* reset timeout */ + dev->rlen = s; + } + /* T=0: we are done when numRecBytes doesn't + * increment any more and NoProcedureByte + * is set and numRecBytes == bytes sent + 6 + * (header bytes + data + 1 for sw2) + * except when the card replies an error + * which means, no data will be sent back. + */ + else if (dev->proto == 0) { + if ((inb(REG_BUF_ADDR(iobase)) & 0x80)) { + /* no procedure byte received since last read */ + DEBUGP(1, dev, "NoProcedure byte set\n"); + /* i=0; */ + } else { + /* procedure byte received since last read */ + DEBUGP(1, dev, "NoProcedure byte unset " + "(reset timeout)\n"); + dev->procbyte = inb(REG_FLAGS1(iobase)); + DEBUGP(1, dev, "Read procedure byte 0x%.2x\n", + dev->procbyte); + i = 0; /* resettimeout */ + } + if (inb(REG_FLAGS0(iobase)) & 0x08) { + DEBUGP(1, dev, "T0Done flag (read reply)\n"); + break; + } + } + if (dev->proto) + infolen = inb(REG_FLAGS1(iobase)); + } + if (i == 600) { + DEBUGP(1, dev, "timeout waiting for numRecBytes\n"); + rc = -EIO; + goto release_io; + } else { + if (dev->proto == 0) { + DEBUGP(1, dev, "Wait for T0Done bit to be set\n"); + for (i = 0; i < 1000; i++) { + if (inb(REG_FLAGS0(iobase)) & 0x08) + break; + msleep_interruptible(10); + } + if (i == 1000) { + DEBUGP(1, dev, "timeout waiting for T0Done\n"); + rc = -EIO; + goto release_io; + } + + dev->procbyte = inb(REG_FLAGS1(iobase)); + DEBUGP(4, dev, "Read procedure byte 0x%.2x\n", + dev->procbyte); + + io_read_num_rec_bytes(iobase, &dev->rlen); + DEBUGP(4, dev, "Read NumRecBytes = %i\n", dev->rlen); + + } + } + /* T=1: read offset=zero, T=0: read offset=after challenge */ + dev->rpos = dev->proto ? 0 : nr == 4 ? 5 : nr > dev->rlen ? 5 : nr; + DEBUGP(4, dev, "dev->rlen = %i, dev->rpos = %i, nr = %i\n", + dev->rlen, dev->rpos, nr); + +release_io: + DEBUGP(4, dev, "Reset SM\n"); + xoutb(0x80, REG_FLAGS0(iobase)); /* reset SM */ + + if (rc < 0) { + DEBUGP(4, dev, "Write failed but clear T_Active\n"); + dev->flags1 &= 0xdf; + xoutb(dev->flags1, REG_FLAGS1(iobase)); + } + + clear_bit(LOCK_IO, &dev->flags); + wake_up_interruptible(&dev->ioq); + wake_up_interruptible(&dev->readq); /* tell read we have data */ + + /* ITSEC E2: clear write buffer */ + memset((char *)dev->sbuf, 0, 512); + + /* return error or actually written bytes */ + DEBUGP(2, dev, "<- cmm_write\n"); + return rc < 0 ? rc : nr; +} + +static void start_monitor(struct cm4000_dev *dev) +{ + DEBUGP(3, dev, "-> start_monitor\n"); + if (!dev->monitor_running) { + DEBUGP(5, dev, "create, init and add timer\n"); + init_timer(&dev->timer); + dev->monitor_running = 1; + dev->timer.expires = jiffies; + dev->timer.data = (unsigned long) dev; + dev->timer.function = monitor_card; + add_timer(&dev->timer); + } else + DEBUGP(5, dev, "monitor already running\n"); + DEBUGP(3, dev, "<- start_monitor\n"); +} + +static void stop_monitor(struct cm4000_dev *dev) +{ + DEBUGP(3, dev, "-> stop_monitor\n"); + if (dev->monitor_running) { + DEBUGP(5, dev, "stopping monitor\n"); + terminate_monitor(dev); + /* reset monitor SM */ + clear_bit(IS_ATR_VALID, &dev->flags); + clear_bit(IS_ATR_PRESENT, &dev->flags); + } else + DEBUGP(5, dev, "monitor already stopped\n"); + DEBUGP(3, dev, "<- stop_monitor\n"); +} + +static int cmm_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, + unsigned long arg) +{ + struct cm4000_dev *dev = filp->private_data; + ioaddr_t iobase = dev->link.io.BasePort1; + dev_link_t *link; + int size; + int rc; +#ifdef PCMCIA_DEBUG + char *ioctl_names[CM_IOC_MAXNR + 1] = { + [_IOC_NR(CM_IOCGSTATUS)] "CM_IOCGSTATUS", + [_IOC_NR(CM_IOCGATR)] "CM_IOCGATR", + [_IOC_NR(CM_IOCARDOFF)] "CM_IOCARDOFF", + [_IOC_NR(CM_IOCSPTS)] "CM_IOCSPTS", + [_IOC_NR(CM_IOSDBGLVL)] "CM4000_DBGLVL", + }; +#endif + DEBUGP(3, dev, "cmm_ioctl(device=%d.%d) %s\n", imajor(inode), + iminor(inode), ioctl_names[_IOC_NR(cmd)]); + + link = dev_table[iminor(inode)]; + if (!(DEV_OK(link))) { + DEBUGP(4, dev, "DEV_OK false\n"); + return -ENODEV; + } + + if (test_bit(IS_CMM_ABSENT, &dev->flags)) { + DEBUGP(4, dev, "CMM_ABSENT flag set\n"); + return -ENODEV; + } + + if (_IOC_TYPE(cmd) != CM_IOC_MAGIC) { + DEBUGP(4, dev, "ioctype mismatch\n"); + return -EINVAL; + } + if (_IOC_NR(cmd) > CM_IOC_MAXNR) { + DEBUGP(4, dev, "iocnr mismatch\n"); + return -EINVAL; + } + size = _IOC_SIZE(cmd); + rc = 0; + DEBUGP(4, dev, "iocdir=%.4x iocr=%.4x iocw=%.4x iocsize=%d cmd=%.4x\n", + _IOC_DIR(cmd), _IOC_READ, _IOC_WRITE, size, cmd); + + if (_IOC_DIR(cmd) & _IOC_READ) { + if (!access_ok(VERIFY_WRITE, (void *)arg, size)) + return -EFAULT; + } + if (_IOC_DIR(cmd) & _IOC_WRITE) { + if (!access_ok(VERIFY_READ, (void *)arg, size)) + return -EFAULT; + } + + switch (cmd) { + case CM_IOCGSTATUS: + DEBUGP(4, dev, " ... in CM_IOCGSTATUS\n"); + { + int status; + + /* clear other bits, but leave inserted & powered as + * they are */ + status = dev->flags0 & 3; + if (test_bit(IS_ATR_PRESENT, &dev->flags)) + status |= CM_ATR_PRESENT; + if (test_bit(IS_ATR_VALID, &dev->flags)) + status |= CM_ATR_VALID; + if (test_bit(IS_CMM_ABSENT, &dev->flags)) + status |= CM_NO_READER; + if (test_bit(IS_BAD_CARD, &dev->flags)) + status |= CM_BAD_CARD; + if (copy_to_user((int *)arg, &status, sizeof(int))) + return -EFAULT; + } + return 0; + case CM_IOCGATR: + DEBUGP(4, dev, "... in CM_IOCGATR\n"); + { + struct atreq *atreq = (struct atreq *) arg; + int tmp; + /* allow nonblocking io and being interrupted */ + if (wait_event_interruptible + (dev->atrq, + ((filp->f_flags & O_NONBLOCK) + || (test_bit(IS_ATR_PRESENT, (void *)&dev->flags) + != 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + if (test_bit(IS_ATR_VALID, &dev->flags) == 0) { + tmp = -1; + if (copy_to_user(&(atreq->atr_len), &tmp, + sizeof(int))) + return -EFAULT; + } else { + if (copy_to_user(atreq->atr, dev->atr, + dev->atr_len)) + return -EFAULT; + + tmp = dev->atr_len; + if (copy_to_user(&(atreq->atr_len), &tmp, sizeof(int))) + return -EFAULT; + } + return 0; + } + case CM_IOCARDOFF: + +#ifdef PCMCIA_DEBUG + DEBUGP(4, dev, "... in CM_IOCARDOFF\n"); + if (dev->flags0 & 0x01) { + DEBUGP(4, dev, " Card inserted\n"); + } else { + DEBUGP(2, dev, " No card inserted\n"); + } + if (dev->flags0 & 0x02) { + DEBUGP(4, dev, " Card powered\n"); + } else { + DEBUGP(2, dev, " Card not powered\n"); + } +#endif + + /* is a card inserted and powered? */ + if ((dev->flags0 & 0x01) && (dev->flags0 & 0x02)) { + + /* get IO lock */ + if (wait_event_interruptible + (dev->ioq, + ((filp->f_flags & O_NONBLOCK) + || (test_and_set_bit(LOCK_IO, (void *)&dev->flags) + == 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + /* Set Flags0 = 0x42 */ + DEBUGP(4, dev, "Set Flags0=0x42 \n"); + xoutb(0x42, REG_FLAGS0(iobase)); + clear_bit(IS_ATR_PRESENT, &dev->flags); + clear_bit(IS_ATR_VALID, &dev->flags); + dev->mstate = M_CARDOFF; + clear_bit(LOCK_IO, &dev->flags); + if (wait_event_interruptible + (dev->atrq, + ((filp->f_flags & O_NONBLOCK) + || (test_bit(IS_ATR_VALID, (void *)&dev->flags) != + 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + } + /* release lock */ + clear_bit(LOCK_IO, &dev->flags); + wake_up_interruptible(&dev->ioq); + + return 0; + case CM_IOCSPTS: + { + struct ptsreq krnptsreq; + + if (copy_from_user(&krnptsreq, (struct ptsreq *) arg, + sizeof(struct ptsreq))) + return -EFAULT; + + rc = 0; + DEBUGP(4, dev, "... in CM_IOCSPTS\n"); + /* wait for ATR to get valid */ + if (wait_event_interruptible + (dev->atrq, + ((filp->f_flags & O_NONBLOCK) + || (test_bit(IS_ATR_PRESENT, (void *)&dev->flags) + != 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + /* get IO lock */ + if (wait_event_interruptible + (dev->ioq, + ((filp->f_flags & O_NONBLOCK) + || (test_and_set_bit(LOCK_IO, (void *)&dev->flags) + == 0)))) { + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + return -ERESTARTSYS; + } + + if ((rc = set_protocol(dev, &krnptsreq)) != 0) { + /* auto power_on again */ + dev->mstate = M_FETCH_ATR; + clear_bit(IS_ATR_VALID, &dev->flags); + } + /* release lock */ + clear_bit(LOCK_IO, &dev->flags); + wake_up_interruptible(&dev->ioq); + + } + return rc; +#ifdef PCMCIA_DEBUG + case CM_IOSDBGLVL: /* set debug log level */ + { + int old_pc_debug = 0; + + old_pc_debug = pc_debug; + if (copy_from_user(&pc_debug, (int *)arg, sizeof(int))) + return -EFAULT; + + if (old_pc_debug != pc_debug) + DEBUGP(0, dev, "Changed debug log level " + "to %i\n", pc_debug); + } + return rc; +#endif + default: + DEBUGP(4, dev, "... in default (unknown IOCTL code)\n"); + return -EINVAL; + } +} + +static int cmm_open(struct inode *inode, struct file *filp) +{ + struct cm4000_dev *dev; + dev_link_t *link; + int rc, minor = iminor(inode); + + if (minor >= CM4000_MAX_DEV) + return -ENODEV; + + link = dev_table[minor]; + if (link == NULL || !(DEV_OK(link))) + return -ENODEV; + + if (link->open) + return -EBUSY; + + dev = link->priv; + filp->private_data = dev; + + DEBUGP(2, dev, "-> cmm_open(device=%d.%d process=%s,%d)\n", + imajor(inode), minor, current->comm, current->pid); + + /* init device variables, they may be "polluted" after close + * or, the device may never have been closed (i.e. open failed) + */ + + ZERO_DEV(dev); + + /* opening will always block since the + * monitor will be started by open, which + * means we have to wait for ATR becoming + * vaild = block until valid (or card + * inserted) + */ + if (filp->f_flags & O_NONBLOCK) + return -EAGAIN; + + dev->mdelay = T_50MSEC; + + /* start monitoring the cardstatus */ + start_monitor(dev); + + link->open = 1; /* only one open per device */ + rc = 0; + + DEBUGP(2, dev, "<- cmm_open\n"); + return nonseekable_open(inode, filp); +} + +static int cmm_close(struct inode *inode, struct file *filp) +{ + struct cm4000_dev *dev; + dev_link_t *link; + int minor = iminor(inode); + + if (minor >= CM4000_MAX_DEV) + return -ENODEV; + + link = dev_table[minor]; + if (link == NULL) + return -ENODEV; + + dev = link->priv; + + DEBUGP(2, dev, "-> cmm_close(maj/min=%d.%d)\n", + imajor(inode), minor); + + stop_monitor(dev); + + ZERO_DEV(dev); + + link->open = 0; /* only one open per device */ + wake_up(&dev->devq); /* socket removed? */ + + DEBUGP(2, dev, "cmm_close\n"); + return 0; +} + +static void cmm_cm4000_release(dev_link_t * link) +{ + struct cm4000_dev *dev = link->priv; + + /* dont terminate the monitor, rather rely on + * close doing that for us. + */ + DEBUGP(3, dev, "-> cmm_cm4000_release\n"); + while (link->open) { + printk(KERN_INFO MODULE_NAME ": delaying release until " + "process has terminated\n"); + /* note: don't interrupt us: + * close the applications which own + * the devices _first_ ! + */ + wait_event(dev->devq, (link->open == 0)); + } + /* dev->devq=NULL; this cannot be zeroed earlier */ + DEBUGP(3, dev, "<- cmm_cm4000_release\n"); + return; +} + +/*==== Interface to PCMCIA Layer =======================================*/ + +static void cm4000_config(dev_link_t * link, int devno) +{ + client_handle_t handle = link->handle; + struct cm4000_dev *dev; + tuple_t tuple; + cisparse_t parse; + config_info_t conf; + u_char buf[64]; + int fail_fn, fail_rc; + int rc; + + /* read the config-tuples */ + tuple.DesiredTuple = CISTPL_CONFIG; + tuple.Attributes = 0; + tuple.TupleData = buf; + tuple.TupleDataMax = sizeof(buf); + tuple.TupleOffset = 0; + + if ((fail_rc = pcmcia_get_first_tuple(handle, &tuple)) != CS_SUCCESS) { + fail_fn = GetFirstTuple; + goto cs_failed; + } + if ((fail_rc = pcmcia_get_tuple_data(handle, &tuple)) != CS_SUCCESS) { + fail_fn = GetTupleData; + goto cs_failed; + } + if ((fail_rc = + pcmcia_parse_tuple(handle, &tuple, &parse)) != CS_SUCCESS) { + fail_fn = ParseTuple; + goto cs_failed; + } + if ((fail_rc = + pcmcia_get_configuration_info(handle, &conf)) != CS_SUCCESS) { + fail_fn = GetConfigurationInfo; + goto cs_failed; + } + + link->state |= DEV_CONFIG; + link->conf.ConfigBase = parse.config.base; + link->conf.Present = parse.config.rmask[0]; + link->conf.Vcc = conf.Vcc; + + link->io.BasePort2 = 0; + link->io.NumPorts2 = 0; + link->io.Attributes2 = 0; + tuple.DesiredTuple = CISTPL_CFTABLE_ENTRY; + for (rc = pcmcia_get_first_tuple(handle, &tuple); + rc == CS_SUCCESS; rc = pcmcia_get_next_tuple(handle, &tuple)) { + + rc = pcmcia_get_tuple_data(handle, &tuple); + if (rc != CS_SUCCESS) + continue; + rc = pcmcia_parse_tuple(handle, &tuple, &parse); + if (rc != CS_SUCCESS) + continue; + + link->conf.ConfigIndex = parse.cftable_entry.index; + + if (!parse.cftable_entry.io.nwin) + continue; + + /* Get the IOaddr */ + link->io.BasePort1 = parse.cftable_entry.io.win[0].base; + link->io.NumPorts1 = parse.cftable_entry.io.win[0].len; + link->io.Attributes1 = IO_DATA_PATH_WIDTH_AUTO; + if (!(parse.cftable_entry.io.flags & CISTPL_IO_8BIT)) + link->io.Attributes1 = IO_DATA_PATH_WIDTH_16; + if (!(parse.cftable_entry.io.flags & CISTPL_IO_16BIT)) + link->io.Attributes1 = IO_DATA_PATH_WIDTH_8; + link->io.IOAddrLines = parse.cftable_entry.io.flags + & CISTPL_IO_LINES_MASK; + + rc = pcmcia_request_io(handle, &link->io); + if (rc == CS_SUCCESS) + break; /* we are done */ + } + if (rc != CS_SUCCESS) + goto cs_release; + + link->conf.IntType = 00000002; + + if ((fail_rc = + pcmcia_request_configuration(handle, &link->conf)) != CS_SUCCESS) { + fail_fn = RequestConfiguration; + goto cs_release; + } + + dev = link->priv; + sprintf(dev->node.dev_name, DEVICE_NAME "%d", devno); + dev->node.major = major; + dev->node.minor = devno; + dev->node.next = NULL; + link->dev = &dev->node; + link->state &= ~DEV_CONFIG_PENDING; + + return; + +cs_failed: + cs_error(handle, fail_fn, fail_rc); +cs_release: + cm4000_release(link); + + link->state &= ~DEV_CONFIG_PENDING; +} + +static int cm4000_event(event_t event, int priority, + event_callback_args_t *args) +{ + dev_link_t *link; + struct cm4000_dev *dev; + int devno; + + link = args->client_data; + dev = link->priv; + + DEBUGP(3, dev, "-> cm4000_event\n"); + for (devno = 0; devno < CM4000_MAX_DEV; devno++) + if (dev_table[devno] == link) + break; + + if (devno == CM4000_MAX_DEV) + return CS_BAD_ADAPTER; + + switch (event) { + case CS_EVENT_CARD_INSERTION: + DEBUGP(5, dev, "CS_EVENT_CARD_INSERTION\n"); + link->state |= DEV_PRESENT | DEV_CONFIG_PENDING; + cm4000_config(link, devno); + break; + case CS_EVENT_CARD_REMOVAL: + DEBUGP(5, dev, "CS_EVENT_CARD_REMOVAL\n"); + link->state &= ~DEV_PRESENT; + stop_monitor(dev); + break; + case CS_EVENT_PM_SUSPEND: + DEBUGP(5, dev, "CS_EVENT_PM_SUSPEND " + "(fall-through to CS_EVENT_RESET_PHYSICAL)\n"); + link->state |= DEV_SUSPEND; + /* fall-through */ + case CS_EVENT_RESET_PHYSICAL: + DEBUGP(5, dev, "CS_EVENT_RESET_PHYSICAL\n"); + if (link->state & DEV_CONFIG) { + DEBUGP(5, dev, "ReleaseConfiguration\n"); + pcmcia_release_configuration(link->handle); + } + stop_monitor(dev); + break; + case CS_EVENT_PM_RESUME: + DEBUGP(5, dev, "CS_EVENT_PM_RESUME " + "(fall-through to CS_EVENT_CARD_RESET)\n"); + link->state &= ~DEV_SUSPEND; + /* fall-through */ + case CS_EVENT_CARD_RESET: + DEBUGP(5, dev, "CS_EVENT_CARD_RESET\n"); + if ((link->state & DEV_CONFIG)) { + DEBUGP(5, dev, "RequestConfiguration\n"); + pcmcia_request_configuration(link->handle, &link->conf); + } + if (link->open) + start_monitor(dev); + break; + default: + DEBUGP(5, dev, "unknown event %.2x\n", event); + break; + } + DEBUGP(3, dev, "<- cm4000_event\n"); + return CS_SUCCESS; +} + +static void cm4000_release(dev_link_t *link) +{ + cmm_cm4000_release(link->priv); /* delay release until device closed */ + pcmcia_release_configuration(link->handle); + pcmcia_release_io(link->handle, &link->io); +} + +static dev_link_t *cm4000_attach(void) +{ + struct cm4000_dev *dev; + dev_link_t *link; + client_reg_t client_reg; + int i; + + for (i = 0; i < CM4000_MAX_DEV; i++) + if (dev_table[i] == NULL) + break; + + if (i == CM4000_MAX_DEV) { + printk(KERN_NOTICE MODULE_NAME ": all devices in use\n"); + return NULL; + } + + /* create a new cm4000_cs device */ + dev = kzalloc(sizeof(struct cm4000_dev), GFP_KERNEL); + if (dev == NULL) + return NULL; + + link = &dev->link; + link->priv = dev; + link->conf.IntType = INT_MEMORY_AND_IO; + dev_table[i] = link; + + /* register with card services */ + client_reg.dev_info = &dev_info; + client_reg.EventMask = + CS_EVENT_CARD_INSERTION | CS_EVENT_CARD_REMOVAL | + CS_EVENT_RESET_PHYSICAL | CS_EVENT_CARD_RESET | + CS_EVENT_PM_SUSPEND | CS_EVENT_PM_RESUME; + client_reg.Version = 0x0210; + client_reg.event_callback_args.client_data = link; + + i = pcmcia_register_client(&link->handle, &client_reg); + if (i) { + cs_error(link->handle, RegisterClient, i); + cm4000_detach(link); + return NULL; + } + + init_waitqueue_head(&dev->devq); + init_waitqueue_head(&dev->ioq); + init_waitqueue_head(&dev->atrq); + init_waitqueue_head(&dev->readq); + + return link; +} + +static void cm4000_detach_by_devno(int devno, dev_link_t * link) +{ + struct cm4000_dev *dev = link->priv; + + DEBUGP(3, dev, "-> detach_by_devno(devno=%d)\n", devno); + + if (link->state & DEV_CONFIG) { + DEBUGP(5, dev, "device still configured (try to release it)\n"); + cm4000_release(link); + } + + if (link->handle) { + pcmcia_deregister_client(link->handle); + } + + dev_table[devno] = NULL; + kfree(dev); + return; +} + +static void cm4000_detach(dev_link_t * link) +{ + int i; + + /* find device */ + for (i = 0; i < CM4000_MAX_DEV; i++) + if (dev_table[i] == link) + break; + + if (i == CM4000_MAX_DEV) + return; + + cm4000_detach_by_devno(i, link); + return; +} + +static struct file_operations cm4000_fops = { + .owner = THIS_MODULE, + .read = cmm_read, + .write = cmm_write, + .ioctl = cmm_ioctl, + .open = cmm_open, + .release= cmm_close, +}; + +static struct pcmcia_device_id cm4000_ids[] = { + PCMCIA_DEVICE_MANF_CARD(0x0223, 0x0002), + PCMCIA_DEVICE_PROD_ID12("CardMan", "4000", 0x2FB368CA, 0xA2BD8C39), + PCMCIA_DEVICE_NULL, +}; +MODULE_DEVICE_TABLE(pcmcia, cm4000_ids); + +static struct pcmcia_driver cm4000_driver = { + .owner = THIS_MODULE, + .drv = { + .name = "cm4000_cs", + }, + .attach = cm4000_attach, + .detach = cm4000_detach, + .event = cm4000_event, + .id_table = cm4000_ids, +}; + +static int __init cmm_init(void) +{ + printk(KERN_INFO "%s\n", version); + pcmcia_register_driver(&cm4000_driver); + major = register_chrdev(0, DEVICE_NAME, &cm4000_fops); + if (major < 0) { + printk(KERN_WARNING MODULE_NAME + ": could not get major number\n"); + return -1; + } + + return 0; +} + +static void __exit cmm_exit(void) +{ + int i; + + printk(KERN_INFO MODULE_NAME ": unloading\n"); + pcmcia_unregister_driver(&cm4000_driver); + for (i = 0; i < CM4000_MAX_DEV; i++) + if (dev_table[i]) + cm4000_detach_by_devno(i, dev_table[i]); + unregister_chrdev(major, DEVICE_NAME); +}; + +module_init(cmm_init); +module_exit(cmm_exit); +MODULE_LICENSE("Dual BSD/GPL"); diff --git a/include/linux/cm4000_cs.h b/include/linux/cm4000_cs.h new file mode 100644 index 000000000000..605ebe24bb2e --- /dev/null +++ b/include/linux/cm4000_cs.h @@ -0,0 +1,66 @@ +#ifndef _CM4000_H_ +#define _CM4000_H_ + +#define MAX_ATR 33 + +#define CM4000_MAX_DEV 4 + +/* those two structures are passed via ioctl() from/to userspace. They are + * used by existing userspace programs, so I kepth the awkward "bIFSD" naming + * not to break compilation of userspace apps. -HW */ + +typedef struct atreq { + int32_t atr_len; + unsigned char atr[64]; + int32_t power_act; + unsigned char bIFSD; + unsigned char bIFSC; +} atreq_t; + + +/* what is particularly stupid in the original driver is the arch-dependant + * member sizes. This leads to CONFIG_COMPAT breakage, since 32bit userspace + * will lay out the structure members differently than the 64bit kernel. + * + * I've changed "ptsreq.protocol" from "unsigned long" to "u_int32_t". + * On 32bit this will make no difference. With 64bit kernels, it will make + * 32bit apps work, too. + */ + +typedef struct ptsreq { + u_int32_t protocol; /*T=0: 2^0, T=1: 2^1*/ + unsigned char flags; + unsigned char pts1; + unsigned char pts2; + unsigned char pts3; +} ptsreq_t; + +#define CM_IOC_MAGIC 'c' +#define CM_IOC_MAXNR 255 + +#define CM_IOCGSTATUS _IOR (CM_IOC_MAGIC, 0, unsigned char *) +#define CM_IOCGATR _IOWR(CM_IOC_MAGIC, 1, atreq_t *) +#define CM_IOCSPTS _IOW (CM_IOC_MAGIC, 2, ptsreq_t *) +#define CM_IOCSRDR _IO (CM_IOC_MAGIC, 3) +#define CM_IOCARDOFF _IO (CM_IOC_MAGIC, 4) + +#define CM_IOSDBGLVL _IOW(CM_IOC_MAGIC, 250, int*) + +/* card and device states */ +#define CM_CARD_INSERTED 0x01 +#define CM_CARD_POWERED 0x02 +#define CM_ATR_PRESENT 0x04 +#define CM_ATR_VALID 0x08 +#define CM_STATE_VALID 0x0f +/* extra info only from CM4000 */ +#define CM_NO_READER 0x10 +#define CM_BAD_CARD 0x20 + + +#ifdef __KERNEL__ + +#define DEVICE_NAME "cmm" +#define MODULE_NAME "cm4000_cs" + +#endif /* __KERNEL__ */ +#endif /* _CM4000_H_ */ -- cgit v1.2.3-71-gd317 From 4c8d3d997ef3c0594350fba716529905b314287e Mon Sep 17 00:00:00 2001 From: Kumar Gala Date: Sun, 13 Nov 2005 16:06:30 -0800 Subject: [PATCH] Update email address for Kumar Changed jobs and the Freescale address is no longer valid. Signed-off-by: Kumar Gala Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- CREDITS | 2 +- MAINTAINERS | 2 +- arch/powerpc/kernel/head_fsl_booke.S | 2 +- arch/powerpc/mm/fsl_booke_mmu.c | 2 +- arch/powerpc/oprofile/op_model_fsl_booke.c | 2 +- arch/ppc/kernel/head_fsl_booke.S | 2 +- arch/ppc/mm/fsl_booke_mmu.c | 2 +- arch/ppc/platforms/83xx/mpc834x_sys.c | 2 +- arch/ppc/platforms/83xx/mpc834x_sys.h | 2 +- arch/ppc/platforms/85xx/mpc8540_ads.c | 2 +- arch/ppc/platforms/85xx/mpc8540_ads.h | 2 +- arch/ppc/platforms/85xx/mpc8555_cds.h | 2 +- arch/ppc/platforms/85xx/mpc8560_ads.c | 2 +- arch/ppc/platforms/85xx/mpc8560_ads.h | 2 +- arch/ppc/platforms/85xx/mpc85xx_ads_common.c | 2 +- arch/ppc/platforms/85xx/mpc85xx_ads_common.h | 2 +- arch/ppc/platforms/85xx/mpc85xx_cds_common.c | 2 +- arch/ppc/platforms/85xx/mpc85xx_cds_common.h | 2 +- arch/ppc/platforms/85xx/sbc8560.c | 2 +- arch/ppc/platforms/pq2ads.c | 2 +- arch/ppc/syslib/ipic.h | 2 +- arch/ppc/syslib/mpc83xx_devices.c | 2 +- arch/ppc/syslib/mpc83xx_sys.c | 2 +- arch/ppc/syslib/mpc85xx_devices.c | 2 +- arch/ppc/syslib/mpc85xx_sys.c | 2 +- arch/ppc/syslib/mpc8xx_devices.c | 2 +- arch/ppc/syslib/mpc8xx_sys.c | 2 +- arch/ppc/syslib/ppc83xx_setup.c | 2 +- arch/ppc/syslib/ppc83xx_setup.h | 2 +- arch/ppc/syslib/ppc85xx_common.c | 2 +- arch/ppc/syslib/ppc85xx_common.h | 2 +- arch/ppc/syslib/ppc85xx_setup.c | 2 +- arch/ppc/syslib/ppc85xx_setup.h | 2 +- arch/ppc/syslib/ppc_sys.c | 2 +- arch/ppc/syslib/pq2_devices.c | 2 +- arch/ppc/syslib/pq2_sys.c | 2 +- drivers/char/watchdog/booke_wdt.c | 2 +- drivers/net/gianfar.c | 2 +- drivers/net/gianfar.h | 2 +- drivers/net/gianfar_ethtool.c | 2 +- drivers/net/gianfar_mii.c | 2 +- drivers/net/gianfar_mii.h | 2 +- drivers/serial/cpm_uart/cpm_uart_core.c | 2 +- drivers/serial/cpm_uart/cpm_uart_cpm1.c | 2 +- drivers/serial/cpm_uart/cpm_uart_cpm2.c | 2 +- include/asm-ppc/immap_85xx.h | 2 +- include/asm-ppc/ipic.h | 2 +- include/asm-ppc/mpc83xx.h | 2 +- include/asm-ppc/mpc85xx.h | 2 +- include/asm-ppc/ppc_sys.h | 2 +- include/linux/fsl_devices.h | 2 +- 51 files changed, 51 insertions(+), 51 deletions(-) (limited to 'include/linux') diff --git a/CREDITS b/CREDITS index 7fb4c73e0228..192f749eba25 100644 --- a/CREDITS +++ b/CREDITS @@ -1097,7 +1097,7 @@ S: 80050-430 - Curitiba - Paran S: Brazil N: Kumar Gala -E: kumar.gala@freescale.com +E: galak@kernel.crashing.org D: Embedded PowerPC 6xx/7xx/74xx/82xx/83xx/85xx support S: Austin, Texas 78729 S: USA diff --git a/MAINTAINERS b/MAINTAINERS index cc924073f599..509927e40bbb 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1565,7 +1565,7 @@ S: Maintained LINUX FOR POWERPC EMBEDDED PPC83XX AND PPC85XX P: Kumar Gala -M: kumar.gala@freescale.com +M: galak@kernel.crashing.org W: http://www.penguinppc.org/ L: linuxppc-embedded@ozlabs.org S: Maintained diff --git a/arch/powerpc/kernel/head_fsl_booke.S b/arch/powerpc/kernel/head_fsl_booke.S index 5063c603fad4..8d60fa99fc4b 100644 --- a/arch/powerpc/kernel/head_fsl_booke.S +++ b/arch/powerpc/kernel/head_fsl_booke.S @@ -24,7 +24,7 @@ * Copyright 2002-2004 MontaVista Software, Inc. * PowerPC 44x support, Matt Porter * Copyright 2004 Freescale Semiconductor, Inc - * PowerPC e500 modifications, Kumar Gala + * PowerPC e500 modifications, Kumar Gala * * This program is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License as published by the diff --git a/arch/powerpc/mm/fsl_booke_mmu.c b/arch/powerpc/mm/fsl_booke_mmu.c index af9ca0eb6d55..5d581bb3aa12 100644 --- a/arch/powerpc/mm/fsl_booke_mmu.c +++ b/arch/powerpc/mm/fsl_booke_mmu.c @@ -1,5 +1,5 @@ /* - * Modifications by Kumar Gala (kumar.gala@freescale.com) to support + * Modifications by Kumar Gala (galak@kernel.crashing.org) to support * E500 Book E processors. * * Copyright 2004 Freescale Semiconductor, Inc diff --git a/arch/powerpc/oprofile/op_model_fsl_booke.c b/arch/powerpc/oprofile/op_model_fsl_booke.c index 86124a94c9af..26539cda6023 100644 --- a/arch/powerpc/oprofile/op_model_fsl_booke.c +++ b/arch/powerpc/oprofile/op_model_fsl_booke.c @@ -7,7 +7,7 @@ * Copyright (c) 2004 Freescale Semiconductor, Inc * * Author: Andy Fleming - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License diff --git a/arch/ppc/kernel/head_fsl_booke.S b/arch/ppc/kernel/head_fsl_booke.S index 5063c603fad4..8d60fa99fc4b 100644 --- a/arch/ppc/kernel/head_fsl_booke.S +++ b/arch/ppc/kernel/head_fsl_booke.S @@ -24,7 +24,7 @@ * Copyright 2002-2004 MontaVista Software, Inc. * PowerPC 44x support, Matt Porter * Copyright 2004 Freescale Semiconductor, Inc - * PowerPC e500 modifications, Kumar Gala + * PowerPC e500 modifications, Kumar Gala * * This program is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License as published by the diff --git a/arch/ppc/mm/fsl_booke_mmu.c b/arch/ppc/mm/fsl_booke_mmu.c index af9ca0eb6d55..5d581bb3aa12 100644 --- a/arch/ppc/mm/fsl_booke_mmu.c +++ b/arch/ppc/mm/fsl_booke_mmu.c @@ -1,5 +1,5 @@ /* - * Modifications by Kumar Gala (kumar.gala@freescale.com) to support + * Modifications by Kumar Gala (galak@kernel.crashing.org) to support * E500 Book E processors. * * Copyright 2004 Freescale Semiconductor, Inc diff --git a/arch/ppc/platforms/83xx/mpc834x_sys.c b/arch/ppc/platforms/83xx/mpc834x_sys.c index 98edc75f4105..84efc0ced880 100644 --- a/arch/ppc/platforms/83xx/mpc834x_sys.c +++ b/arch/ppc/platforms/83xx/mpc834x_sys.c @@ -3,7 +3,7 @@ * * MPC834x SYS board specific routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/83xx/mpc834x_sys.h b/arch/ppc/platforms/83xx/mpc834x_sys.h index 58e44c042535..2e514d316fb8 100644 --- a/arch/ppc/platforms/83xx/mpc834x_sys.h +++ b/arch/ppc/platforms/83xx/mpc834x_sys.h @@ -3,7 +3,7 @@ * * MPC834X SYS common board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor, Inc. * diff --git a/arch/ppc/platforms/85xx/mpc8540_ads.c b/arch/ppc/platforms/85xx/mpc8540_ads.c index 7e952c1228cb..c5cde97c6ef0 100644 --- a/arch/ppc/platforms/85xx/mpc8540_ads.c +++ b/arch/ppc/platforms/85xx/mpc8540_ads.c @@ -3,7 +3,7 @@ * * MPC8540ADS board specific routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc8540_ads.h b/arch/ppc/platforms/85xx/mpc8540_ads.h index 3d05d7c4a938..e48ca3a97397 100644 --- a/arch/ppc/platforms/85xx/mpc8540_ads.h +++ b/arch/ppc/platforms/85xx/mpc8540_ads.h @@ -3,7 +3,7 @@ * * MPC8540ADS board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc8555_cds.h b/arch/ppc/platforms/85xx/mpc8555_cds.h index e0e75568bc57..1a8e6c67355d 100644 --- a/arch/ppc/platforms/85xx/mpc8555_cds.h +++ b/arch/ppc/platforms/85xx/mpc8555_cds.h @@ -3,7 +3,7 @@ * * MPC8555CDS board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc8560_ads.c b/arch/ppc/platforms/85xx/mpc8560_ads.c index 208433f1e93a..8e39a5517092 100644 --- a/arch/ppc/platforms/85xx/mpc8560_ads.c +++ b/arch/ppc/platforms/85xx/mpc8560_ads.c @@ -3,7 +3,7 @@ * * MPC8560ADS board specific routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc8560_ads.h b/arch/ppc/platforms/85xx/mpc8560_ads.h index 7df885d73e9d..143ae7eefa7c 100644 --- a/arch/ppc/platforms/85xx/mpc8560_ads.h +++ b/arch/ppc/platforms/85xx/mpc8560_ads.h @@ -3,7 +3,7 @@ * * MPC8540ADS board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc85xx_ads_common.c b/arch/ppc/platforms/85xx/mpc85xx_ads_common.c index 16ad092d8a06..17ce48fe3503 100644 --- a/arch/ppc/platforms/85xx/mpc85xx_ads_common.c +++ b/arch/ppc/platforms/85xx/mpc85xx_ads_common.c @@ -3,7 +3,7 @@ * * MPC85xx ADS board common routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc85xx_ads_common.h b/arch/ppc/platforms/85xx/mpc85xx_ads_common.h index 84acf6e8d45e..7b26bcc5d10d 100644 --- a/arch/ppc/platforms/85xx/mpc85xx_ads_common.h +++ b/arch/ppc/platforms/85xx/mpc85xx_ads_common.h @@ -3,7 +3,7 @@ * * MPC85XX ADS common board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/85xx/mpc85xx_cds_common.c b/arch/ppc/platforms/85xx/mpc85xx_cds_common.c index a21156967a5e..d8991b88dc9c 100644 --- a/arch/ppc/platforms/85xx/mpc85xx_cds_common.c +++ b/arch/ppc/platforms/85xx/mpc85xx_cds_common.c @@ -3,7 +3,7 @@ * * MPC85xx CDS board specific routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor, Inc * diff --git a/arch/ppc/platforms/85xx/mpc85xx_cds_common.h b/arch/ppc/platforms/85xx/mpc85xx_cds_common.h index 12b292c6ae32..5b588cfd0e41 100644 --- a/arch/ppc/platforms/85xx/mpc85xx_cds_common.h +++ b/arch/ppc/platforms/85xx/mpc85xx_cds_common.h @@ -3,7 +3,7 @@ * * MPC85xx CDS board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor, Inc * diff --git a/arch/ppc/platforms/85xx/sbc8560.c b/arch/ppc/platforms/85xx/sbc8560.c index b4ee1707a836..45a5b81b4ed1 100644 --- a/arch/ppc/platforms/85xx/sbc8560.c +++ b/arch/ppc/platforms/85xx/sbc8560.c @@ -3,7 +3,7 @@ * * Wind River SBC8560 board specific routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/platforms/pq2ads.c b/arch/ppc/platforms/pq2ads.c index 6a1475c1e128..71c9fca1fe9b 100644 --- a/arch/ppc/platforms/pq2ads.c +++ b/arch/ppc/platforms/pq2ads.c @@ -3,7 +3,7 @@ * * PQ2ADS platform support * - * Author: Kumar Gala + * Author: Kumar Gala * Derived from: est8260_setup.c by Allen Curtis * * Copyright 2004 Freescale Semiconductor, Inc. diff --git a/arch/ppc/syslib/ipic.h b/arch/ppc/syslib/ipic.h index 2b56a4fcf373..a7ce7da8785c 100644 --- a/arch/ppc/syslib/ipic.h +++ b/arch/ppc/syslib/ipic.h @@ -3,7 +3,7 @@ * * IPIC private definitions and structure. * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor, Inc * diff --git a/arch/ppc/syslib/mpc83xx_devices.c b/arch/ppc/syslib/mpc83xx_devices.c index f43fbf9a9389..847df4409982 100644 --- a/arch/ppc/syslib/mpc83xx_devices.c +++ b/arch/ppc/syslib/mpc83xx_devices.c @@ -3,7 +3,7 @@ * * MPC83xx Device descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/mpc83xx_sys.c b/arch/ppc/syslib/mpc83xx_sys.c index da743446789b..a1523989aff4 100644 --- a/arch/ppc/syslib/mpc83xx_sys.c +++ b/arch/ppc/syslib/mpc83xx_sys.c @@ -3,7 +3,7 @@ * * MPC83xx System descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/mpc85xx_devices.c b/arch/ppc/syslib/mpc85xx_devices.c index 2ede677a0a53..69949d255658 100644 --- a/arch/ppc/syslib/mpc85xx_devices.c +++ b/arch/ppc/syslib/mpc85xx_devices.c @@ -3,7 +3,7 @@ * * MPC85xx Device descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/mpc85xx_sys.c b/arch/ppc/syslib/mpc85xx_sys.c index cb68d8c58348..397cfbcce5ea 100644 --- a/arch/ppc/syslib/mpc85xx_sys.c +++ b/arch/ppc/syslib/mpc85xx_sys.c @@ -3,7 +3,7 @@ * * MPC85xx System descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/mpc8xx_devices.c b/arch/ppc/syslib/mpc8xx_devices.c index 2b5f0e701687..92dc98b36bde 100644 --- a/arch/ppc/syslib/mpc8xx_devices.c +++ b/arch/ppc/syslib/mpc8xx_devices.c @@ -3,7 +3,7 @@ * * MPC8xx Device descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 MontaVista Software, Inc. by Vitaly Bordug * diff --git a/arch/ppc/syslib/mpc8xx_sys.c b/arch/ppc/syslib/mpc8xx_sys.c index 3cc27d29e3af..d3c617521603 100644 --- a/arch/ppc/syslib/mpc8xx_sys.c +++ b/arch/ppc/syslib/mpc8xx_sys.c @@ -3,7 +3,7 @@ * * MPC8xx System descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 MontaVista Software, Inc. by Vitaly Bordug * diff --git a/arch/ppc/syslib/ppc83xx_setup.c b/arch/ppc/syslib/ppc83xx_setup.c index 4da168a6ad03..1b5fe9e398d4 100644 --- a/arch/ppc/syslib/ppc83xx_setup.c +++ b/arch/ppc/syslib/ppc83xx_setup.c @@ -3,7 +3,7 @@ * * MPC83XX common board code * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc83xx_setup.h b/arch/ppc/syslib/ppc83xx_setup.h index c766c1a5f786..a122a7322e5e 100644 --- a/arch/ppc/syslib/ppc83xx_setup.h +++ b/arch/ppc/syslib/ppc83xx_setup.h @@ -3,7 +3,7 @@ * * MPC83XX common board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc85xx_common.c b/arch/ppc/syslib/ppc85xx_common.c index da841dacdc13..19ad537225e4 100644 --- a/arch/ppc/syslib/ppc85xx_common.c +++ b/arch/ppc/syslib/ppc85xx_common.c @@ -3,7 +3,7 @@ * * MPC85xx support routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc85xx_common.h b/arch/ppc/syslib/ppc85xx_common.h index 2c8f304441bf..94edf32151dd 100644 --- a/arch/ppc/syslib/ppc85xx_common.h +++ b/arch/ppc/syslib/ppc85xx_common.h @@ -3,7 +3,7 @@ * * MPC85xx support routines * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc85xx_setup.c b/arch/ppc/syslib/ppc85xx_setup.c index de2f90576577..1a47ff4b831d 100644 --- a/arch/ppc/syslib/ppc85xx_setup.c +++ b/arch/ppc/syslib/ppc85xx_setup.c @@ -3,7 +3,7 @@ * * MPC85XX common board code * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc85xx_setup.h b/arch/ppc/syslib/ppc85xx_setup.h index 6e6cfe162faf..e340b0545fb5 100644 --- a/arch/ppc/syslib/ppc85xx_setup.h +++ b/arch/ppc/syslib/ppc85xx_setup.h @@ -3,7 +3,7 @@ * * MPC85XX common board definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor Inc. * diff --git a/arch/ppc/syslib/ppc_sys.c b/arch/ppc/syslib/ppc_sys.c index 603f01190816..c0b93c4191ee 100644 --- a/arch/ppc/syslib/ppc_sys.c +++ b/arch/ppc/syslib/ppc_sys.c @@ -3,7 +3,7 @@ * * PPC System library functions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * Copyright 2005 MontaVista, Inc. by Vitaly Bordug diff --git a/arch/ppc/syslib/pq2_devices.c b/arch/ppc/syslib/pq2_devices.c index e960fe935325..6ff3aab82fc3 100644 --- a/arch/ppc/syslib/pq2_devices.c +++ b/arch/ppc/syslib/pq2_devices.c @@ -3,7 +3,7 @@ * * PQ2 Device descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * This file is licensed under the terms of the GNU General Public License * version 2. This program is licensed "as is" without any warranty of any diff --git a/arch/ppc/syslib/pq2_sys.c b/arch/ppc/syslib/pq2_sys.c index 7b6c9ebdb9e3..36d6e2179940 100644 --- a/arch/ppc/syslib/pq2_sys.c +++ b/arch/ppc/syslib/pq2_sys.c @@ -3,7 +3,7 @@ * * PQ2 System descriptions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * This file is licensed under the terms of the GNU General Public License * version 2. This program is licensed "as is" without any warranty of any diff --git a/drivers/char/watchdog/booke_wdt.c b/drivers/char/watchdog/booke_wdt.c index abc30cca6645..65830ec71042 100644 --- a/drivers/char/watchdog/booke_wdt.c +++ b/drivers/char/watchdog/booke_wdt.c @@ -4,7 +4,7 @@ * Watchdog timer for PowerPC Book-E systems * * Author: Matthew McClintock - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor Inc. * diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c index e3a329539f1c..0f030b73cbb3 100644 --- a/drivers/net/gianfar.c +++ b/drivers/net/gianfar.c @@ -6,7 +6,7 @@ * Based on 8260_io/fcc_enet.c * * Author: Andy Fleming - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright (c) 2002-2004 Freescale Semiconductor, Inc. * diff --git a/drivers/net/gianfar.h b/drivers/net/gianfar.h index 220084e53341..5065ba82cb76 100644 --- a/drivers/net/gianfar.h +++ b/drivers/net/gianfar.h @@ -6,7 +6,7 @@ * Based on 8260_io/fcc_enet.c * * Author: Andy Fleming - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright (c) 2002-2004 Freescale Semiconductor, Inc. * diff --git a/drivers/net/gianfar_ethtool.c b/drivers/net/gianfar_ethtool.c index 5a2d810ce575..cfa3cd7c91a0 100644 --- a/drivers/net/gianfar_ethtool.c +++ b/drivers/net/gianfar_ethtool.c @@ -6,7 +6,7 @@ * Based on e1000 ethtool support * * Author: Andy Fleming - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright (c) 2003,2004 Freescale Semiconductor, Inc. * diff --git a/drivers/net/gianfar_mii.c b/drivers/net/gianfar_mii.c index 9544279e8bcd..04a462c2a5b7 100644 --- a/drivers/net/gianfar_mii.c +++ b/drivers/net/gianfar_mii.c @@ -5,7 +5,7 @@ * Provides Bus interface for MIIM regs * * Author: Andy Fleming - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright (c) 2002-2004 Freescale Semiconductor, Inc. * diff --git a/drivers/net/gianfar_mii.h b/drivers/net/gianfar_mii.h index 56e5665d5c9b..e85eb216fb5b 100644 --- a/drivers/net/gianfar_mii.h +++ b/drivers/net/gianfar_mii.h @@ -5,7 +5,7 @@ * Driver for the MDIO bus controller in the Gianfar register space * * Author: Andy Fleming - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright (c) 2002-2004 Freescale Semiconductor, Inc. * diff --git a/drivers/serial/cpm_uart/cpm_uart_core.c b/drivers/serial/cpm_uart/cpm_uart_core.c index 25825f2aba22..987d22b53c22 100644 --- a/drivers/serial/cpm_uart/cpm_uart_core.c +++ b/drivers/serial/cpm_uart/cpm_uart_core.c @@ -7,7 +7,7 @@ * Based on ppc8xx.c by Thomas Gleixner * Based on drivers/serial/amba.c by Russell King * - * Maintainer: Kumar Gala (kumar.gala@freescale.com) (CPM2) + * Maintainer: Kumar Gala (galak@kernel.crashing.org) (CPM2) * Pantelis Antoniou (panto@intracom.gr) (CPM1) * * Copyright (C) 2004 Freescale Semiconductor, Inc. diff --git a/drivers/serial/cpm_uart/cpm_uart_cpm1.c b/drivers/serial/cpm_uart/cpm_uart_cpm1.c index 4b0786e7eb7f..d789ee55cbb7 100644 --- a/drivers/serial/cpm_uart/cpm_uart_cpm1.c +++ b/drivers/serial/cpm_uart/cpm_uart_cpm1.c @@ -3,7 +3,7 @@ * * Driver for CPM (SCC/SMC) serial ports; CPM1 definitions * - * Maintainer: Kumar Gala (kumar.gala@freescale.com) (CPM2) + * Maintainer: Kumar Gala (galak@kernel.crashing.org) (CPM2) * Pantelis Antoniou (panto@intracom.gr) (CPM1) * * Copyright (C) 2004 Freescale Semiconductor, Inc. diff --git a/drivers/serial/cpm_uart/cpm_uart_cpm2.c b/drivers/serial/cpm_uart/cpm_uart_cpm2.c index 15ad58d94889..fd9e53ed3feb 100644 --- a/drivers/serial/cpm_uart/cpm_uart_cpm2.c +++ b/drivers/serial/cpm_uart/cpm_uart_cpm2.c @@ -3,7 +3,7 @@ * * Driver for CPM (SCC/SMC) serial ports; CPM2 definitions * - * Maintainer: Kumar Gala (kumar.gala@freescale.com) (CPM2) + * Maintainer: Kumar Gala (galak@kernel.crashing.org) (CPM2) * Pantelis Antoniou (panto@intracom.gr) (CPM1) * * Copyright (C) 2004 Freescale Semiconductor, Inc. diff --git a/include/asm-ppc/immap_85xx.h b/include/asm-ppc/immap_85xx.h index 50fb5e47094a..9383d0c13ff8 100644 --- a/include/asm-ppc/immap_85xx.h +++ b/include/asm-ppc/immap_85xx.h @@ -3,7 +3,7 @@ * * MPC85xx Internal Memory Map * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor, Inc * diff --git a/include/asm-ppc/ipic.h b/include/asm-ppc/ipic.h index 9092b920997a..0fe396a2b666 100644 --- a/include/asm-ppc/ipic.h +++ b/include/asm-ppc/ipic.h @@ -3,7 +3,7 @@ * * IPIC external definitions and structure. * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor, Inc * diff --git a/include/asm-ppc/mpc83xx.h b/include/asm-ppc/mpc83xx.h index ce212201db2a..7cdf60fa69b6 100644 --- a/include/asm-ppc/mpc83xx.h +++ b/include/asm-ppc/mpc83xx.h @@ -3,7 +3,7 @@ * * MPC83xx definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor, Inc * diff --git a/include/asm-ppc/mpc85xx.h b/include/asm-ppc/mpc85xx.h index d98db980cd49..9d14baea3d71 100644 --- a/include/asm-ppc/mpc85xx.h +++ b/include/asm-ppc/mpc85xx.h @@ -3,7 +3,7 @@ * * MPC85xx definitions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor, Inc * diff --git a/include/asm-ppc/ppc_sys.h b/include/asm-ppc/ppc_sys.h index bba5305c29ed..83d8c77c124d 100644 --- a/include/asm-ppc/ppc_sys.h +++ b/include/asm-ppc/ppc_sys.h @@ -3,7 +3,7 @@ * * PPC system definitions and library functions * - * Maintainer: Kumar Gala + * Maintainer: Kumar Gala * * Copyright 2005 Freescale Semiconductor, Inc * diff --git a/include/linux/fsl_devices.h b/include/linux/fsl_devices.h index 114d5d59f695..934aa9bda481 100644 --- a/include/linux/fsl_devices.h +++ b/include/linux/fsl_devices.h @@ -4,7 +4,7 @@ * Definitions for any platform device related flags or structures for * Freescale processor devices * - * Maintainer: Kumar Gala (kumar.gala@freescale.com) + * Maintainer: Kumar Gala * * Copyright 2004 Freescale Semiconductor, Inc * -- cgit v1.2.3-71-gd317 From 51c6f666fceb3184eeff045dad4432b602cd648e Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Sun, 13 Nov 2005 16:06:42 -0800 Subject: [PATCH] mm: ZAP_BLOCK causes redundant work The address based work estimate for unmapping (for lockbreak) is and always was horribly inefficient for sparse mappings. The problem is most simply explained with an example: If we find a pgd is clear, we still have to call into unmap_page_range PGDIR_SIZE / ZAP_BLOCK_SIZE times, each time checking the clear pgd, in order to progress the working address to the next pgd. The fundamental way to solve the problem is to keep track of the end address we've processed and pass it back to the higher layers. From: Nick Piggin Modification to completely get away from address based work estimate and instead use an abstract count, with a very small cost for empty entries as opposed to present pages. On 2.6.14-git2, ppc64, and CONFIG_PREEMPT=y, mapping and unmapping 1TB of virtual address space takes 1.69s; with the following patch applied, this operation can be done 1000 times in less than 0.01s From: Andrew Morton With CONFIG_HUTETLB_PAGE=n: mm/memory.c: In function `unmap_vmas': mm/memory.c:779: warning: division by zero Due to zap_work -= (end - start) / (HPAGE_SIZE / PAGE_SIZE); So make the dummy HPAGE_SIZE non-zero Signed-off-by: Robin Holt Signed-off-by: Nick Piggin Cc: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/hugetlb.h | 4 +-- mm/memory.c | 89 ++++++++++++++++++++++++++++++------------------- 2 files changed, 57 insertions(+), 36 deletions(-) (limited to 'include/linux') diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 0cea162b08c0..1056717ee501 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -102,8 +102,8 @@ static inline unsigned long hugetlb_total_pages(void) #define hugetlb_fault(mm, vma, addr, write) ({ BUG(); 0; }) #ifndef HPAGE_MASK -#define HPAGE_MASK 0 /* Keep the compiler happy */ -#define HPAGE_SIZE 0 +#define HPAGE_MASK PAGE_MASK /* Keep the compiler happy */ +#define HPAGE_SIZE PAGE_SIZE #endif #endif /* !CONFIG_HUGETLB_PAGE */ diff --git a/mm/memory.c b/mm/memory.c index 0f60baf6f69b..2998cfc12f5b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -549,10 +549,10 @@ int copy_page_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, return 0; } -static void zap_pte_range(struct mmu_gather *tlb, +static unsigned long zap_pte_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pmd_t *pmd, unsigned long addr, unsigned long end, - struct zap_details *details) + long *zap_work, struct zap_details *details) { struct mm_struct *mm = tlb->mm; pte_t *pte; @@ -563,10 +563,15 @@ static void zap_pte_range(struct mmu_gather *tlb, pte = pte_offset_map_lock(mm, pmd, addr, &ptl); do { pte_t ptent = *pte; - if (pte_none(ptent)) + if (pte_none(ptent)) { + (*zap_work)--; continue; + } if (pte_present(ptent)) { struct page *page = NULL; + + (*zap_work) -= PAGE_SIZE; + if (!(vma->vm_flags & VM_RESERVED)) { unsigned long pfn = pte_pfn(ptent); if (unlikely(!pfn_valid(pfn))) @@ -624,16 +629,18 @@ static void zap_pte_range(struct mmu_gather *tlb, if (!pte_file(ptent)) free_swap_and_cache(pte_to_swp_entry(ptent)); pte_clear_full(mm, addr, pte, tlb->fullmm); - } while (pte++, addr += PAGE_SIZE, addr != end); + } while (pte++, addr += PAGE_SIZE, (addr != end && *zap_work > 0)); add_mm_rss(mm, file_rss, anon_rss); pte_unmap_unlock(pte - 1, ptl); + + return addr; } -static inline void zap_pmd_range(struct mmu_gather *tlb, +static inline unsigned long zap_pmd_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pud_t *pud, unsigned long addr, unsigned long end, - struct zap_details *details) + long *zap_work, struct zap_details *details) { pmd_t *pmd; unsigned long next; @@ -641,16 +648,21 @@ static inline void zap_pmd_range(struct mmu_gather *tlb, pmd = pmd_offset(pud, addr); do { next = pmd_addr_end(addr, end); - if (pmd_none_or_clear_bad(pmd)) + if (pmd_none_or_clear_bad(pmd)) { + (*zap_work)--; continue; - zap_pte_range(tlb, vma, pmd, addr, next, details); - } while (pmd++, addr = next, addr != end); + } + next = zap_pte_range(tlb, vma, pmd, addr, next, + zap_work, details); + } while (pmd++, addr = next, (addr != end && *zap_work > 0)); + + return addr; } -static inline void zap_pud_range(struct mmu_gather *tlb, +static inline unsigned long zap_pud_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pgd_t *pgd, unsigned long addr, unsigned long end, - struct zap_details *details) + long *zap_work, struct zap_details *details) { pud_t *pud; unsigned long next; @@ -658,15 +670,21 @@ static inline void zap_pud_range(struct mmu_gather *tlb, pud = pud_offset(pgd, addr); do { next = pud_addr_end(addr, end); - if (pud_none_or_clear_bad(pud)) + if (pud_none_or_clear_bad(pud)) { + (*zap_work)--; continue; - zap_pmd_range(tlb, vma, pud, addr, next, details); - } while (pud++, addr = next, addr != end); + } + next = zap_pmd_range(tlb, vma, pud, addr, next, + zap_work, details); + } while (pud++, addr = next, (addr != end && *zap_work > 0)); + + return addr; } -static void unmap_page_range(struct mmu_gather *tlb, struct vm_area_struct *vma, +static unsigned long unmap_page_range(struct mmu_gather *tlb, + struct vm_area_struct *vma, unsigned long addr, unsigned long end, - struct zap_details *details) + long *zap_work, struct zap_details *details) { pgd_t *pgd; unsigned long next; @@ -679,11 +697,16 @@ static void unmap_page_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pgd = pgd_offset(vma->vm_mm, addr); do { next = pgd_addr_end(addr, end); - if (pgd_none_or_clear_bad(pgd)) + if (pgd_none_or_clear_bad(pgd)) { + (*zap_work)--; continue; - zap_pud_range(tlb, vma, pgd, addr, next, details); - } while (pgd++, addr = next, addr != end); + } + next = zap_pud_range(tlb, vma, pgd, addr, next, + zap_work, details); + } while (pgd++, addr = next, (addr != end && *zap_work > 0)); tlb_end_vma(tlb, vma); + + return addr; } #ifdef CONFIG_PREEMPT @@ -724,7 +747,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, unsigned long end_addr, unsigned long *nr_accounted, struct zap_details *details) { - unsigned long zap_bytes = ZAP_BLOCK_SIZE; + long zap_work = ZAP_BLOCK_SIZE; unsigned long tlb_start = 0; /* For tlb_finish_mmu */ int tlb_start_valid = 0; unsigned long start = start_addr; @@ -745,27 +768,25 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, *nr_accounted += (end - start) >> PAGE_SHIFT; while (start != end) { - unsigned long block; - if (!tlb_start_valid) { tlb_start = start; tlb_start_valid = 1; } - if (is_vm_hugetlb_page(vma)) { - block = end - start; + if (unlikely(is_vm_hugetlb_page(vma))) { unmap_hugepage_range(vma, start, end); - } else { - block = min(zap_bytes, end - start); - unmap_page_range(*tlbp, vma, start, - start + block, details); + zap_work -= (end - start) / + (HPAGE_SIZE / PAGE_SIZE); + start = end; + } else + start = unmap_page_range(*tlbp, vma, + start, end, &zap_work, details); + + if (zap_work > 0) { + BUG_ON(start != end); + break; } - start += block; - zap_bytes -= block; - if ((long)zap_bytes > 0) - continue; - tlb_finish_mmu(*tlbp, tlb_start, start); if (need_resched() || @@ -779,7 +800,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, *tlbp = tlb_gather_mmu(vma->vm_mm, fullmm); tlb_start_valid = 0; - zap_bytes = ZAP_BLOCK_SIZE; + zap_work = ZAP_BLOCK_SIZE; } } out: -- cgit v1.2.3-71-gd317 From 7fb1d9fca5c6e3b06773b69165a73f3fb786b8ee Mon Sep 17 00:00:00 2001 From: Rohit Seth Date: Sun, 13 Nov 2005 16:06:43 -0800 Subject: [PATCH] mm: __alloc_pages cleanup Clean up of __alloc_pages. Restoration of previous behaviour, plus further cleanups by introducing an 'alloc_flags', removing the last of should_reclaim_zone. Signed-off-by: Rohit Seth Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 2 +- mm/page_alloc.c | 195 +++++++++++++++++++++---------------------------- mm/vmscan.c | 6 +- 3 files changed, 89 insertions(+), 114 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index f5fa3082fd6a..6cfb114a0c34 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -329,7 +329,7 @@ void get_zone_counts(unsigned long *active, unsigned long *inactive, void build_all_zonelists(void); void wakeup_kswapd(struct zone *zone, int order); int zone_watermark_ok(struct zone *z, int order, unsigned long mark, - int alloc_type, int can_try_harder, gfp_t gfp_high); + int classzone_idx, int alloc_flags); #ifdef CONFIG_HAVE_MEMORY_PRESENT void memory_present(int nid, unsigned long start, unsigned long end); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index b37dc0f78d07..845b91749a42 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -732,9 +732,7 @@ buffered_rmqueue(struct zone *zone, int order, gfp_t gfp_flags) } local_irq_restore(flags); put_cpu(); - } - - if (page == NULL) { + } else { spin_lock_irqsave(&zone->lock, flags); page = __rmqueue(zone, order); spin_unlock_irqrestore(&zone->lock, flags); @@ -754,20 +752,25 @@ buffered_rmqueue(struct zone *zone, int order, gfp_t gfp_flags) return page; } +#define ALLOC_NO_WATERMARKS 0x01 /* don't check watermarks at all */ +#define ALLOC_HARDER 0x02 /* try to alloc harder */ +#define ALLOC_HIGH 0x04 /* __GFP_HIGH set */ +#define ALLOC_CPUSET 0x08 /* check for correct cpuset */ + /* * Return 1 if free pages are above 'mark'. This takes into account the order * of the allocation. */ int zone_watermark_ok(struct zone *z, int order, unsigned long mark, - int classzone_idx, int can_try_harder, gfp_t gfp_high) + int classzone_idx, int alloc_flags) { /* free_pages my go negative - that's OK */ long min = mark, free_pages = z->free_pages - (1 << order) + 1; int o; - if (gfp_high) + if (alloc_flags & ALLOC_HIGH) min -= min / 2; - if (can_try_harder) + if (alloc_flags & ALLOC_HARDER) min -= min / 4; if (free_pages <= min + z->lowmem_reserve[classzone_idx]) @@ -785,14 +788,40 @@ int zone_watermark_ok(struct zone *z, int order, unsigned long mark, return 1; } -static inline int -should_reclaim_zone(struct zone *z, gfp_t gfp_mask) +/* + * get_page_from_freeliest goes through the zonelist trying to allocate + * a page. + */ +static struct page * +get_page_from_freelist(gfp_t gfp_mask, unsigned int order, + struct zonelist *zonelist, int alloc_flags) { - if (!z->reclaim_pages) - return 0; - if (gfp_mask & __GFP_NORECLAIM) - return 0; - return 1; + struct zone **z = zonelist->zones; + struct page *page = NULL; + int classzone_idx = zone_idx(*z); + + /* + * Go through the zonelist once, looking for a zone with enough free. + * See also cpuset_zone_allowed() comment in kernel/cpuset.c. + */ + do { + if ((alloc_flags & ALLOC_CPUSET) && + !cpuset_zone_allowed(*z, gfp_mask)) + continue; + + if (!(alloc_flags & ALLOC_NO_WATERMARKS)) { + if (!zone_watermark_ok(*z, order, (*z)->pages_low, + classzone_idx, alloc_flags)) + continue; + } + + page = buffered_rmqueue(*z, order, gfp_mask); + if (page) { + zone_statistics(zonelist, *z); + break; + } + } while (*(++z) != NULL); + return page; } /* @@ -803,92 +832,60 @@ __alloc_pages(gfp_t gfp_mask, unsigned int order, struct zonelist *zonelist) { const gfp_t wait = gfp_mask & __GFP_WAIT; - struct zone **zones, *z; + struct zone **z; struct page *page; struct reclaim_state reclaim_state; struct task_struct *p = current; - int i; - int classzone_idx; int do_retry; - int can_try_harder; + int alloc_flags; int did_some_progress; might_sleep_if(wait); - /* - * The caller may dip into page reserves a bit more if the caller - * cannot run direct reclaim, or is the caller has realtime scheduling - * policy - */ - can_try_harder = (unlikely(rt_task(p)) && !in_interrupt()) || !wait; - - zones = zonelist->zones; /* the list of zones suitable for gfp_mask */ + z = zonelist->zones; /* the list of zones suitable for gfp_mask */ - if (unlikely(zones[0] == NULL)) { + if (unlikely(*z == NULL)) { /* Should this ever happen?? */ return NULL; } +restart: + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_CPUSET); + if (page) + goto got_pg; - classzone_idx = zone_idx(zones[0]); + do + wakeup_kswapd(*z, order); + while (*(++z)); -restart: /* - * Go through the zonelist once, looking for a zone with enough free. - * See also cpuset_zone_allowed() comment in kernel/cpuset.c. + * OK, we're below the kswapd watermark and have kicked background + * reclaim. Now things get more complex, so set up alloc_flags according + * to how we want to proceed. + * + * The caller may dip into page reserves a bit more if the caller + * cannot run direct reclaim, or if the caller has realtime scheduling + * policy. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - int do_reclaim = should_reclaim_zone(z, gfp_mask); - - if (!cpuset_zone_allowed(z, __GFP_HARDWALL)) - continue; - - /* - * If the zone is to attempt early page reclaim then this loop - * will try to reclaim pages and check the watermark a second - * time before giving up and falling back to the next zone. - */ -zone_reclaim_retry: - if (!zone_watermark_ok(z, order, z->pages_low, - classzone_idx, 0, 0)) { - if (!do_reclaim) - continue; - else { - zone_reclaim(z, gfp_mask, order); - /* Only try reclaim once */ - do_reclaim = 0; - goto zone_reclaim_retry; - } - } - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } - - for (i = 0; (z = zones[i]) != NULL; i++) - wakeup_kswapd(z, order); + alloc_flags = 0; + if ((unlikely(rt_task(p)) && !in_interrupt()) || !wait) + alloc_flags |= ALLOC_HARDER; + if (gfp_mask & __GFP_HIGH) + alloc_flags |= ALLOC_HIGH; + if (wait) + alloc_flags |= ALLOC_CPUSET; /* * Go through the zonelist again. Let __GFP_HIGH and allocations - * coming from realtime tasks to go deeper into reserves + * coming from realtime tasks go deeper into reserves. * * This is the last chance, in general, before the goto nopage. * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc. * See also cpuset_zone_allowed() comment in kernel/cpuset.c. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_min, - classzone_idx, can_try_harder, - gfp_mask & __GFP_HIGH)) - continue; - - if (wait && !cpuset_zone_allowed(z, gfp_mask)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, zonelist, alloc_flags); + if (page) + goto got_pg; /* This allocation should allow future memory freeing. */ @@ -897,13 +894,10 @@ zone_reclaim_retry: if (!(gfp_mask & __GFP_NOMEMALLOC)) { nofail_alloc: /* go through the zonelist yet again, ignoring mins */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!cpuset_zone_allowed(z, gfp_mask)) - continue; - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, + zonelist, ALLOC_NO_WATERMARKS|ALLOC_CPUSET); + if (page) + goto got_pg; if (gfp_mask & __GFP_NOFAIL) { blk_congestion_wait(WRITE, HZ/50); goto nofail_alloc; @@ -924,7 +918,7 @@ rebalance: reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; - did_some_progress = try_to_free_pages(zones, gfp_mask); + did_some_progress = try_to_free_pages(zonelist->zones, gfp_mask); p->reclaim_state = NULL; p->flags &= ~PF_MEMALLOC; @@ -932,19 +926,10 @@ rebalance: cond_resched(); if (likely(did_some_progress)) { - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_min, - classzone_idx, can_try_harder, - gfp_mask & __GFP_HIGH)) - continue; - - if (!cpuset_zone_allowed(z, gfp_mask)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, + zonelist, alloc_flags); + if (page) + goto got_pg; } else if ((gfp_mask & __GFP_FS) && !(gfp_mask & __GFP_NORETRY)) { /* * Go through the zonelist yet one more time, keep @@ -952,18 +937,10 @@ rebalance: * a parallel oom killing, we must fail if we're still * under heavy pressure. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_high, - classzone_idx, 0, 0)) - continue; - - if (!cpuset_zone_allowed(z, __GFP_HARDWALL)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_CPUSET); + if (page) + goto got_pg; out_of_memory(gfp_mask, order); goto restart; @@ -996,9 +973,7 @@ nopage: dump_stack(); show_mem(); } - return NULL; got_pg: - zone_statistics(zonelist, z); return page; } diff --git a/mm/vmscan.c b/mm/vmscan.c index 135bf8ca96ee..28130541270f 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1074,7 +1074,7 @@ loop_again: continue; if (!zone_watermark_ok(zone, order, - zone->pages_high, 0, 0, 0)) { + zone->pages_high, 0, 0)) { end_zone = i; goto scan; } @@ -1111,7 +1111,7 @@ scan: if (nr_pages == 0) { /* Not software suspend */ if (!zone_watermark_ok(zone, order, - zone->pages_high, end_zone, 0, 0)) + zone->pages_high, end_zone, 0)) all_zones_ok = 0; } zone->temp_priority = priority; @@ -1259,7 +1259,7 @@ void wakeup_kswapd(struct zone *zone, int order) return; pgdat = zone->zone_pgdat; - if (zone_watermark_ok(zone, order, zone->pages_low, 0, 0, 0)) + if (zone_watermark_ok(zone, order, zone->pages_low, 0, 0)) return; if (pgdat->kswapd_max_order < order) pgdat->kswapd_max_order = order; -- cgit v1.2.3-71-gd317 From 2d6c666e8704cf06267f29a4fa3d2cf823469c38 Mon Sep 17 00:00:00 2001 From: Paul Jackson Date: Sun, 13 Nov 2005 16:06:44 -0800 Subject: [PATCH] mm: gfp_noreclaim cleanup Remove last remnant of the defunct early reclaim page logic, the no longer used __GFP_NORECLAIM flag bit. Signed-off-by: Paul Jackson Acked-by: Martin Hicks Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/gfp.h | 5 ++--- include/linux/pagemap.h | 4 ++-- 2 files changed, 4 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/gfp.h b/include/linux/gfp.h index c3779432a723..23279d8f19b1 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -39,8 +39,7 @@ struct vm_area_struct; #define __GFP_COMP ((__force gfp_t)0x4000u)/* Add compound page metadata */ #define __GFP_ZERO ((__force gfp_t)0x8000u)/* Return zeroed page on success */ #define __GFP_NOMEMALLOC ((__force gfp_t)0x10000u) /* Don't use emergency reserves */ -#define __GFP_NORECLAIM ((__force gfp_t)0x20000u) /* No realy zone reclaim during allocation */ -#define __GFP_HARDWALL ((__force gfp_t)0x40000u) /* Enforce hardwall cpuset memory allocs */ +#define __GFP_HARDWALL ((__force gfp_t)0x20000u) /* Enforce hardwall cpuset memory allocs */ #define __GFP_BITS_SHIFT 20 /* Room for 20 __GFP_FOO bits */ #define __GFP_BITS_MASK ((__force gfp_t)((1 << __GFP_BITS_SHIFT) - 1)) @@ -49,7 +48,7 @@ struct vm_area_struct; #define GFP_LEVEL_MASK (__GFP_WAIT|__GFP_HIGH|__GFP_IO|__GFP_FS| \ __GFP_COLD|__GFP_NOWARN|__GFP_REPEAT| \ __GFP_NOFAIL|__GFP_NORETRY|__GFP_NO_GROW|__GFP_COMP| \ - __GFP_NOMEMALLOC|__GFP_NORECLAIM|__GFP_HARDWALL) + __GFP_NOMEMALLOC|__GFP_HARDWALL) #define GFP_ATOMIC (__GFP_HIGH) #define GFP_NOIO (__GFP_WAIT) diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index ba6c310a055f..ee700c6eb442 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -53,12 +53,12 @@ void release_pages(struct page **pages, int nr, int cold); static inline struct page *page_cache_alloc(struct address_space *x) { - return alloc_pages(mapping_gfp_mask(x)|__GFP_NORECLAIM, 0); + return alloc_pages(mapping_gfp_mask(x), 0); } static inline struct page *page_cache_alloc_cold(struct address_space *x) { - return alloc_pages(mapping_gfp_mask(x)|__GFP_COLD|__GFP_NORECLAIM, 0); + return alloc_pages(mapping_gfp_mask(x)|__GFP_COLD, 0); } typedef int filler_t(void *, struct page *); -- cgit v1.2.3-71-gd317 From 7fce260a6bf75080ef61408504add5618f90e41b Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Sun, 13 Nov 2005 16:06:48 -0800 Subject: [PATCH] ppc: add support for new powerbooks Enablement patch for the new PowerBooks (late 2005 edition). This enables the ATA controller, Gigabit ethernet and basic AGP setup. Bluetooth works out-of-the box after running hid2hci. Still remaining is to get the touchpad to work, the simple change of just adding the new USB ids isn't enough. Signed-off-by: Olof Johansson Acked-by: Benjamin Herrenschmidt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ppc/platforms/pmac_feature.c | 8 ++++++++ drivers/char/agp/uninorth-agp.c | 4 ++++ drivers/ide/ppc/pmac.c | 11 ++++++++--- drivers/net/sungem.c | 2 ++ include/linux/pci_ids.h | 4 ++++ 5 files changed, 26 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/arch/ppc/platforms/pmac_feature.c b/arch/ppc/platforms/pmac_feature.c index 58884a63ebdb..1e69b0593162 100644 --- a/arch/ppc/platforms/pmac_feature.c +++ b/arch/ppc/platforms/pmac_feature.c @@ -2317,6 +2317,14 @@ static struct pmac_mb_def pmac_mb_defs[] = { PMAC_TYPE_UNKNOWN_INTREPID, intrepid_features, PMAC_MB_MAY_SLEEP | PMAC_MB_HAS_FW_POWER | PMAC_MB_MOBILE, }, + { "PowerBook5,8", "PowerBook G4 15\"", + PMAC_TYPE_UNKNOWN_INTREPID, intrepid_features, + PMAC_MB_MAY_SLEEP | PMAC_MB_HAS_FW_POWER | PMAC_MB_MOBILE, + }, + { "PowerBook5,9", "PowerBook G4 17\"", + PMAC_TYPE_UNKNOWN_INTREPID, intrepid_features, + PMAC_MB_MAY_SLEEP | PMAC_MB_HAS_FW_POWER | PMAC_MB_MOBILE, + }, { "PowerBook6,1", "PowerBook G4 12\"", PMAC_TYPE_UNKNOWN_INTREPID, intrepid_features, PMAC_MB_MAY_SLEEP | PMAC_MB_HAS_FW_POWER | PMAC_MB_MOBILE, diff --git a/drivers/char/agp/uninorth-agp.c b/drivers/char/agp/uninorth-agp.c index c8255312b8c1..50947e38501a 100644 --- a/drivers/char/agp/uninorth-agp.c +++ b/drivers/char/agp/uninorth-agp.c @@ -557,6 +557,10 @@ static struct agp_device_ids uninorth_agp_device_ids[] __devinitdata = { .device_id = PCI_DEVICE_ID_APPLE_U3H_AGP, .chipset_name = "U3H", }, + { + .device_id = PCI_DEVICE_ID_APPLE_IPID2_AGP, + .chipset_name = "UniNorth/Intrepid2", + }, }; static int __devinit agp_uninorth_probe(struct pci_dev *pdev, diff --git a/drivers/ide/ppc/pmac.c b/drivers/ide/ppc/pmac.c index b3e65a65d202..136911a86e84 100644 --- a/drivers/ide/ppc/pmac.c +++ b/drivers/ide/ppc/pmac.c @@ -1667,11 +1667,16 @@ static struct macio_driver pmac_ide_macio_driver = }; static struct pci_device_id pmac_ide_pci_match[] = { - { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_UNI_N_ATA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, - { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_IPID_ATA100, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, - { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_K2_ATA100, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_UNI_N_ATA, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_IPID_ATA100, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_K2_ATA100, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_SH_ATA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_IPID2_ATA, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, }; static struct pci_driver pmac_ide_pci_driver = { diff --git a/drivers/net/sungem.c b/drivers/net/sungem.c index de399563a9db..081717d01374 100644 --- a/drivers/net/sungem.c +++ b/drivers/net/sungem.c @@ -128,6 +128,8 @@ static struct pci_device_id gem_pci_tbl[] = { PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL }, { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_SH_SUNGEM, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL }, + { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_IPID2_GMAC, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL }, {0, } }; diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index d00f8ba7f22b..d4c1c8fd2925 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -805,6 +805,10 @@ #define PCI_DEVICE_ID_APPLE_SH_SUNGEM 0x0051 #define PCI_DEVICE_ID_APPLE_U3L_AGP 0x0058 #define PCI_DEVICE_ID_APPLE_U3H_AGP 0x0059 +#define PCI_DEVICE_ID_APPLE_IPID2_AGP 0x0066 +#define PCI_DEVICE_ID_APPLE_IPID2_ATA 0x0069 +#define PCI_DEVICE_ID_APPLE_IPID2_FW 0x006a +#define PCI_DEVICE_ID_APPLE_IPID2_GMAC 0x006b #define PCI_DEVICE_ID_APPLE_TIGON3 0x1645 #define PCI_VENDOR_ID_YAMAHA 0x1073 -- cgit v1.2.3-71-gd317 From a1261f54611ec4ad6a7ab7080f86747e3ac3685b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 13 Nov 2005 16:06:55 -0800 Subject: [PATCH] m68k: introduce task_thread_info new helper - task_thread_info(task). On platforms that have thread_info allocated separately (i.e. in default case) it simply returns task->thread_info. m68k wants (and for good reasons) to embed its thread_info into task_struct. So it will (in later patch) have task_thread_info() of its own. For now we just add a macro for generic case and convert existing instances of its body in core kernel to uses of new macro. Obviously safe - all normal architectures get the same preprocessor output they used to get. Signed-off-by: Al Viro Signed-off-by: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/sched.h | 16 +++++++++------- kernel/exit.c | 2 +- kernel/fork.c | 4 ++-- kernel/sched.c | 6 +++--- 4 files changed, 15 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 2bbf968b23d9..f8650314ba2f 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1233,32 +1233,34 @@ static inline void task_unlock(struct task_struct *p) spin_unlock(&p->alloc_lock); } +#define task_thread_info(task) (task)->thread_info + /* set thread flags in other task's structures * - see asm/thread_info.h for TIF_xxxx flags available */ static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag) { - set_ti_thread_flag(tsk->thread_info,flag); + set_ti_thread_flag(task_thread_info(tsk), flag); } static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag) { - clear_ti_thread_flag(tsk->thread_info,flag); + clear_ti_thread_flag(task_thread_info(tsk), flag); } static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag) { - return test_and_set_ti_thread_flag(tsk->thread_info,flag); + return test_and_set_ti_thread_flag(task_thread_info(tsk), flag); } static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag) { - return test_and_clear_ti_thread_flag(tsk->thread_info,flag); + return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag); } static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag) { - return test_ti_thread_flag(tsk->thread_info,flag); + return test_ti_thread_flag(task_thread_info(tsk), flag); } static inline void set_tsk_need_resched(struct task_struct *tsk) @@ -1329,12 +1331,12 @@ extern void signal_wake_up(struct task_struct *t, int resume_stopped); static inline unsigned int task_cpu(const struct task_struct *p) { - return p->thread_info->cpu; + return task_thread_info(p)->cpu; } static inline void set_task_cpu(struct task_struct *p, unsigned int cpu) { - p->thread_info->cpu = cpu; + task_thread_info(p)->cpu = cpu; } #else diff --git a/kernel/exit.c b/kernel/exit.c index 452a1d116178..ee515683b92d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -859,7 +859,7 @@ fastcall NORET_TYPE void do_exit(long code) if (group_dead && tsk->signal->leader) disassociate_ctty(1); - module_put(tsk->thread_info->exec_domain->module); + module_put(task_thread_info(tsk)->exec_domain->module); if (tsk->binfmt) module_put(tsk->binfmt->module); diff --git a/kernel/fork.c b/kernel/fork.c index 158710d22566..7ef352ce347b 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -919,7 +919,7 @@ static task_t *copy_process(unsigned long clone_flags, if (nr_threads >= max_threads) goto bad_fork_cleanup_count; - if (!try_module_get(p->thread_info->exec_domain->module)) + if (!try_module_get(task_thread_info(p)->exec_domain->module)) goto bad_fork_cleanup_count; if (p->binfmt && !try_module_get(p->binfmt->module)) @@ -1180,7 +1180,7 @@ bad_fork_cleanup: if (p->binfmt) module_put(p->binfmt->module); bad_fork_cleanup_put_domain: - module_put(p->thread_info->exec_domain->module); + module_put(task_thread_info(p)->exec_domain->module); bad_fork_cleanup_count: put_group_info(p->group_info); atomic_dec(&p->user->processes); diff --git a/kernel/sched.c b/kernel/sched.c index b6506671b2be..831f7e9d8f1c 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1437,7 +1437,7 @@ void fastcall sched_fork(task_t *p, int clone_flags) #endif #ifdef CONFIG_PREEMPT /* Want to start with kernel preemption disabled. */ - p->thread_info->preempt_count = 1; + task_thread_info(p)->preempt_count = 1; #endif /* * Share the timeslice between parent and child, thus the @@ -4410,9 +4410,9 @@ void __devinit init_idle(task_t *idle, int cpu) /* Set the preempt count _outside_ the spinlocks! */ #if defined(CONFIG_PREEMPT) && !defined(CONFIG_PREEMPT_BKL) - idle->thread_info->preempt_count = (idle->lock_depth >= 0); + task_thread_info(idle)->preempt_count = (idle->lock_depth >= 0); #else - idle->thread_info->preempt_count = 0; + task_thread_info(idle)->preempt_count = 0; #endif } -- cgit v1.2.3-71-gd317 From 10ebffde3d3916026974352b7900e44afe2b243f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 13 Nov 2005 16:06:56 -0800 Subject: [PATCH] m68k: introduce setup_thread_stack() and end_of_stack() encapsulates the rest of arch-dependent operations with thread_info access. Two new helpers - setup_thread_stack() and end_of_stack(). For normal case the former consists of copying thread_info of parent to new thread_info and the latter returns pointer immediately past the end of thread_info. Signed-off-by: Al Viro Signed-off-by: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/sched.h | 11 +++++++++++ kernel/fork.c | 3 +-- kernel/sched.c | 4 ++-- 3 files changed, 14 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index f8650314ba2f..e4681256e43e 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1235,6 +1235,17 @@ static inline void task_unlock(struct task_struct *p) #define task_thread_info(task) (task)->thread_info +static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org) +{ + *task_thread_info(p) = *task_thread_info(org); + task_thread_info(p)->task = p; +} + +static inline unsigned long *end_of_stack(struct task_struct *p) +{ + return (unsigned long *)(p->thread_info + 1); +} + /* set thread flags in other task's structures * - see asm/thread_info.h for TIF_xxxx flags available */ diff --git a/kernel/fork.c b/kernel/fork.c index 7ef352ce347b..2c70c9cdf5dc 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -171,10 +171,9 @@ static struct task_struct *dup_task_struct(struct task_struct *orig) return NULL; } - *ti = *orig->thread_info; *tsk = *orig; tsk->thread_info = ti; - ti->task = tsk; + setup_thread_stack(tsk, orig); /* One for us, one for whoever does the "release_task()" (usually parent) */ atomic_set(&tsk->usage,2); diff --git a/kernel/sched.c b/kernel/sched.c index 831f7e9d8f1c..6f46c94cc29e 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4327,10 +4327,10 @@ static void show_task(task_t *p) #endif #ifdef CONFIG_DEBUG_STACK_USAGE { - unsigned long *n = (unsigned long *) (p->thread_info+1); + unsigned long *n = end_of_stack(p); while (!*n) n++; - free = (unsigned long) n - (unsigned long)(p->thread_info+1); + free = (unsigned long)n - (unsigned long)end_of_stack(p); } #endif printk("%5lu %5d %6d ", free, p->pid, p->parent->pid); -- cgit v1.2.3-71-gd317 From f037360f2ed111fe89a8f5cb6ba351f4e9934e53 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 13 Nov 2005 16:06:57 -0800 Subject: [PATCH] m68k: thread_info header cleanup a) in smp_lock.h #include of sched.h and spinlock.h moved under #ifdef CONFIG_LOCK_KERNEL. b) interrupt.h now explicitly pulls sched.h (not via smp_lock.h from hardirq.h as it used to) c) in three more places we need changes to compensate for (a) - one place in arch/sparc needs string.h now, hardirq.h needs forward declaration of task_struct and preempt.h needs direct include of thread_info.h. d) thread_info-related helpers in sched.h and thread_info.h put under ifndef __HAVE_THREAD_FUNCTIONS. Obviously safe. Signed-off-by: Al Viro Signed-off-by: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/sparc/lib/bitext.c | 1 + include/linux/hardirq.h | 2 ++ include/linux/interrupt.h | 1 + include/linux/preempt.h | 1 + include/linux/sched.h | 4 ++++ include/linux/smp_lock.h | 3 +-- 6 files changed, 10 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/arch/sparc/lib/bitext.c b/arch/sparc/lib/bitext.c index 94b05e8c906c..2e168d16547f 100644 --- a/arch/sparc/lib/bitext.c +++ b/arch/sparc/lib/bitext.c @@ -10,6 +10,7 @@ */ #include +#include #include #include diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 5912874ca83c..71d2b8a723b9 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -90,6 +90,8 @@ extern void synchronize_irq(unsigned int irq); #define nmi_enter() irq_enter() #define nmi_exit() sub_preempt_count(HARDIRQ_OFFSET) +struct task_struct; + #ifndef CONFIG_VIRT_CPU_ACCOUNTING static inline void account_user_vtime(struct task_struct *tsk) { diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 0a90205184b0..41f150a3d2dd 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -9,6 +9,7 @@ #include #include #include +#include #include #include #include diff --git a/include/linux/preempt.h b/include/linux/preempt.h index dd98c54a23b4..d9a2f5254a51 100644 --- a/include/linux/preempt.h +++ b/include/linux/preempt.h @@ -7,6 +7,7 @@ */ #include +#include #include #ifdef CONFIG_DEBUG_PREEMPT diff --git a/include/linux/sched.h b/include/linux/sched.h index e4681256e43e..41df81395719 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1233,6 +1233,8 @@ static inline void task_unlock(struct task_struct *p) spin_unlock(&p->alloc_lock); } +#ifndef __HAVE_THREAD_FUNCTIONS + #define task_thread_info(task) (task)->thread_info static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org) @@ -1246,6 +1248,8 @@ static inline unsigned long *end_of_stack(struct task_struct *p) return (unsigned long *)(p->thread_info + 1); } +#endif + /* set thread flags in other task's structures * - see asm/thread_info.h for TIF_xxxx flags available */ diff --git a/include/linux/smp_lock.h b/include/linux/smp_lock.h index b63ce7014093..fa1ff3b165fe 100644 --- a/include/linux/smp_lock.h +++ b/include/linux/smp_lock.h @@ -2,11 +2,10 @@ #define __LINUX_SMPLOCK_H #include +#ifdef CONFIG_LOCK_KERNEL #include #include -#ifdef CONFIG_LOCK_KERNEL - #define kernel_locked() (current->lock_depth >= 0) extern int __lockfunc __reacquire_kernel_lock(void); -- cgit v1.2.3-71-gd317 From 3b66a1edb01b82269a668a478625765b1fa4936f Mon Sep 17 00:00:00 2001 From: Roman Zippel Date: Sun, 13 Nov 2005 16:06:59 -0800 Subject: [PATCH] m68k: convert thread flags to use bit fields Remove task_work structure, use the standard thread flags functions and use shifts in entry.S to test the thread flags. Add a few local labels to entry.S to allow gas to generate short jumps. Finally it changes a number of inline functions in thread_info.h to macros to delay the current_thread_info() usage, which requires on m68k a structure (task_struct) not yet defined at this point. Signed-off-by: Roman Zippel Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/m68k/fpsp040/skeleton.S | 6 ++-- arch/m68k/ifpsp060/iskeleton.S | 6 ++-- arch/m68k/kernel/asm-offsets.c | 5 --- arch/m68k/kernel/entry.S | 78 ++++++++++++++++++++-------------------- arch/m68k/kernel/ptrace.c | 15 ++++---- include/asm-m68k/processor.h | 12 ------- include/asm-m68k/thread_info.h | 81 +++++------------------------------------- include/linux/thread_info.h | 47 +++++++----------------- 8 files changed, 72 insertions(+), 178 deletions(-) (limited to 'include/linux') diff --git a/arch/m68k/fpsp040/skeleton.S b/arch/m68k/fpsp040/skeleton.S index 9571a21d6ad4..a1629194e3fd 100644 --- a/arch/m68k/fpsp040/skeleton.S +++ b/arch/m68k/fpsp040/skeleton.S @@ -381,10 +381,8 @@ fpsp_done: .Lnotkern: SAVE_ALL_INT GET_CURRENT(%d0) - tstb %curptr@(TASK_NEEDRESCHED) - jne ret_from_exception | deliver signals, - | reschedule etc.. - RESTORE_ALL + | deliver signals, reschedule etc.. + jra ret_from_exception | | mem_write --- write to user or supervisor address space diff --git a/arch/m68k/ifpsp060/iskeleton.S b/arch/m68k/ifpsp060/iskeleton.S index 4ba2c74da93d..b2dbdf5ee309 100644 --- a/arch/m68k/ifpsp060/iskeleton.S +++ b/arch/m68k/ifpsp060/iskeleton.S @@ -75,10 +75,8 @@ _060_isp_done: .Lnotkern: SAVE_ALL_INT GET_CURRENT(%d0) - tstb %curptr@(TASK_NEEDRESCHED) - jne ret_from_exception | deliver signals, - | reschedule etc.. - RESTORE_ALL + | deliver signals, reschedule etc.. + jra ret_from_exception | | _060_real_chk(): diff --git a/arch/m68k/kernel/asm-offsets.c b/arch/m68k/kernel/asm-offsets.c index 30beacfef8fb..c787c5ba9513 100644 --- a/arch/m68k/kernel/asm-offsets.c +++ b/arch/m68k/kernel/asm-offsets.c @@ -25,11 +25,6 @@ int main(void) DEFINE(TASK_STATE, offsetof(struct task_struct, state)); DEFINE(TASK_FLAGS, offsetof(struct task_struct, flags)); DEFINE(TASK_PTRACE, offsetof(struct task_struct, ptrace)); - DEFINE(TASK_WORK, offsetof(struct task_struct, thread.work)); - DEFINE(TASK_NEEDRESCHED, offsetof(struct task_struct, thread.work.need_resched)); - DEFINE(TASK_SYSCALL_TRACE, offsetof(struct task_struct, thread.work.syscall_trace)); - DEFINE(TASK_SIGPENDING, offsetof(struct task_struct, thread.work.sigpending)); - DEFINE(TASK_NOTIFY_RESUME, offsetof(struct task_struct, thread.work.notify_resume)); DEFINE(TASK_THREAD, offsetof(struct task_struct, thread)); DEFINE(TASK_INFO, offsetof(struct task_struct, thread.info)); DEFINE(TASK_MM, offsetof(struct task_struct, mm)); diff --git a/arch/m68k/kernel/entry.S b/arch/m68k/kernel/entry.S index 23ca60a45552..320fde05dc63 100644 --- a/arch/m68k/kernel/entry.S +++ b/arch/m68k/kernel/entry.S @@ -44,9 +44,7 @@ #include -.globl system_call, buserr, trap -.globl resume, ret_from_exception -.globl ret_from_signal +.globl system_call, buserr, trap, resume .globl inthandler, sys_call_table .globl sys_fork, sys_clone, sys_vfork .globl ret_from_interrupt, bad_interrupt @@ -58,7 +56,7 @@ ENTRY(buserr) movel %sp,%sp@- | stack frame pointer argument bsrl buserr_c addql #4,%sp - jra ret_from_exception + jra .Lret_from_exception ENTRY(trap) SAVE_ALL_INT @@ -66,7 +64,7 @@ ENTRY(trap) movel %sp,%sp@- | stack frame pointer argument bsrl trap_c addql #4,%sp - jra ret_from_exception + jra .Lret_from_exception | After a fork we jump here directly from resume, | so that %d1 contains the previous task @@ -75,30 +73,31 @@ ENTRY(ret_from_fork) movel %d1,%sp@- jsr schedule_tail addql #4,%sp - jra ret_from_exception + jra .Lret_from_exception -badsys: - movel #-ENOSYS,%sp@(PT_D0) - jra ret_from_exception - -do_trace: +do_trace_entry: movel #-ENOSYS,%sp@(PT_D0) | needed for strace subql #4,%sp SAVE_SWITCH_STACK jbsr syscall_trace RESTORE_SWITCH_STACK addql #4,%sp - movel %sp@(PT_ORIG_D0),%d1 - movel #-ENOSYS,%d0 - cmpl #NR_syscalls,%d1 - jcc 1f - jbsr @(sys_call_table,%d1:l:4)@(0) -1: movel %d0,%sp@(PT_D0) | save the return value - subql #4,%sp | dummy return address + movel %sp@(PT_ORIG_D0),%d0 + cmpl #NR_syscalls,%d0 + jcs syscall +badsys: + movel #-ENOSYS,%sp@(PT_D0) + jra ret_from_syscall + +do_trace_exit: + subql #4,%sp SAVE_SWITCH_STACK jbsr syscall_trace + RESTORE_SWITCH_STACK + addql #4,%sp + jra .Lret_from_exception -ret_from_signal: +ENTRY(ret_from_signal) RESTORE_SWITCH_STACK addql #4,%sp /* on 68040 complete pending writebacks if any */ @@ -111,7 +110,7 @@ ret_from_signal: addql #4,%sp 1: #endif - jra ret_from_exception + jra .Lret_from_exception ENTRY(system_call) SAVE_ALL_SYS @@ -120,30 +119,34 @@ ENTRY(system_call) | save top of frame movel %sp,%curptr@(TASK_THREAD+THREAD_ESP0) - tstb %curptr@(TASK_SYSCALL_TRACE) - jne do_trace + | syscall trace? + tstb %curptr@(TASK_INFO+TINFO_FLAGS+2) + jmi do_trace_entry cmpl #NR_syscalls,%d0 jcc badsys +syscall: jbsr @(sys_call_table,%d0:l:4)@(0) movel %d0,%sp@(PT_D0) | save the return value - +ret_from_syscall: |oriw #0x0700,%sr - movel %curptr@(TASK_WORK),%d0 + movew %curptr@(TASK_INFO+TINFO_FLAGS+2),%d0 jne syscall_exit_work 1: RESTORE_ALL syscall_exit_work: btst #5,%sp@(PT_SR) | check if returning to kernel bnes 1b | if so, skip resched, signals - tstw %d0 - jeq do_signal_return - tstb %d0 - jne do_delayed_trace - + lslw #1,%d0 + jcs do_trace_exit + jmi do_delayed_trace + lslw #8,%d0 + jmi do_signal_return pea resume_userspace - jmp schedule + jra schedule + -ret_from_exception: +ENTRY(ret_from_exception) +.Lret_from_exception: btst #5,%sp@(PT_SR) | check if returning to kernel bnes 1f | if so, skip resched, signals | only allow interrupts when we are really the last one on the @@ -152,19 +155,18 @@ ret_from_exception: andw #ALLOWINT,%sr resume_userspace: - movel %curptr@(TASK_WORK),%d0 - lsrl #8,%d0 + moveb %curptr@(TASK_INFO+TINFO_FLAGS+3),%d0 jne exit_work 1: RESTORE_ALL exit_work: | save top of frame movel %sp,%curptr@(TASK_THREAD+THREAD_ESP0) - tstb %d0 - jeq do_signal_return - + lslb #1,%d0 + jmi do_signal_return pea resume_userspace - jmp schedule + jra schedule + do_signal_return: |andw #ALLOWINT,%sr @@ -254,7 +256,7 @@ ret_from_interrupt: /* check if we need to do software interrupts */ tstl irq_stat+CPUSTAT_SOFTIRQ_PENDING - jeq ret_from_exception + jeq .Lret_from_exception pea ret_from_exception jra do_softirq diff --git a/arch/m68k/kernel/ptrace.c b/arch/m68k/kernel/ptrace.c index 7e54422685cf..540638ca81f9 100644 --- a/arch/m68k/kernel/ptrace.c +++ b/arch/m68k/kernel/ptrace.c @@ -109,7 +109,7 @@ static inline void singlestep_disable(struct task_struct *child) { unsigned long tmp = get_reg(child, PT_SR) & ~(TRACE_BITS << 16); put_reg(child, PT_SR, tmp); - child->thread.work.delayed_trace = 0; + clear_tsk_thread_flag(child, TIF_DELAYED_TRACE); } /* @@ -118,7 +118,7 @@ static inline void singlestep_disable(struct task_struct *child) void ptrace_disable(struct task_struct *child) { singlestep_disable(child); - child->thread.work.syscall_trace = 0; + clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); } long arch_ptrace(struct task_struct *child, long request, long addr, long data) @@ -198,9 +198,9 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) goto out_eio; if (request == PTRACE_SYSCALL) - child->thread.work.syscall_trace = ~0; + set_tsk_thread_flag(child, TIF_SYSCALL_TRACE); else - child->thread.work.syscall_trace = 0; + clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); child->exit_code = data; singlestep_disable(child); wake_up_process(child); @@ -223,10 +223,10 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) if (!valid_signal(data)) goto out_eio; - child->thread.work.syscall_trace = 0; + clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); tmp = get_reg(child, PT_SR) | (TRACE_BITS << 16); put_reg(child, PT_SR, tmp); - child->thread.work.delayed_trace = 1; + set_tsk_thread_flag(child, TIF_DELAYED_TRACE); child->exit_code = data; /* give it a chance to run. */ @@ -288,9 +288,6 @@ out_eio: asmlinkage void syscall_trace(void) { - if (!current->thread.work.delayed_trace && - !current->thread.work.syscall_trace) - return; ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD) ? 0x80 : 0)); /* diff --git a/include/asm-m68k/processor.h b/include/asm-m68k/processor.h index 84b4b26df04c..7982285e84ed 100644 --- a/include/asm-m68k/processor.h +++ b/include/asm-m68k/processor.h @@ -56,17 +56,6 @@ static inline void wrusp(unsigned long usp) #endif #define TASK_UNMAPPED_ALIGN(addr, off) PAGE_ALIGN(addr) -struct task_work { - unsigned char sigpending; - unsigned char notify_resume; /* request for notification on - userspace execution resumption */ - char need_resched; - unsigned char delayed_trace; /* single step a syscall */ - unsigned char syscall_trace; /* count of syscall interceptors */ - unsigned char memdie; /* task was selected to be killed */ - unsigned char pad[2]; -}; - struct thread_struct { unsigned long ksp; /* kernel stack pointer */ unsigned long usp; /* user stack pointer */ @@ -79,7 +68,6 @@ struct thread_struct { unsigned long fp[8*3]; unsigned long fpcntl[3]; /* fp control regs */ unsigned char fpstate[FPSTATESIZE]; /* floating point state */ - struct task_work work; struct thread_info info; }; diff --git a/include/asm-m68k/thread_info.h b/include/asm-m68k/thread_info.h index 4fdbf55f95e5..9532ca3c45cb 100644 --- a/include/asm-m68k/thread_info.h +++ b/include/asm-m68k/thread_info.h @@ -6,12 +6,11 @@ struct thread_info { struct task_struct *task; /* main task structure */ + unsigned long flags; struct exec_domain *exec_domain; /* execution domain */ int preempt_count; /* 0 => preemptable, <0 => BUG */ __u32 cpu; /* should always be 0 on m68k */ struct restart_block restart_block; - - __u8 supervisor_stack[0]; }; #define PREEMPT_ACTIVE 0x4000000 @@ -49,76 +48,14 @@ struct thread_info { #define end_of_stack(p) ((unsigned long *)(p)->thread_info + 1) -#define TIF_SYSCALL_TRACE 0 /* syscall trace active */ -#define TIF_DELAYED_TRACE 1 /* single step a syscall */ -#define TIF_NOTIFY_RESUME 2 /* resumption notification requested */ -#define TIF_SIGPENDING 3 /* signal pending */ -#define TIF_NEED_RESCHED 4 /* rescheduling necessary */ -#define TIF_MEMDIE 5 - -extern int thread_flag_fixme(void); - -/* - * flag set/clear/test wrappers - * - pass TIF_xxxx constants to these functions +/* entry.S relies on these definitions! + * bits 0-7 are tested at every exception exit + * bits 8-15 are also tested at syscall exit */ - -#define __set_tsk_thread_flag(tsk, flag, val) ({ \ - switch (flag) { \ - case TIF_SIGPENDING: \ - tsk->thread.work.sigpending = val; \ - break; \ - case TIF_NEED_RESCHED: \ - tsk->thread.work.need_resched = val; \ - break; \ - case TIF_SYSCALL_TRACE: \ - tsk->thread.work.syscall_trace = val; \ - break; \ - case TIF_MEMDIE: \ - tsk->thread.work.memdie = val; \ - break; \ - default: \ - thread_flag_fixme(); \ - } \ -}) - -#define __get_tsk_thread_flag(tsk, flag) ({ \ - int ___res; \ - switch (flag) { \ - case TIF_SIGPENDING: \ - ___res = tsk->thread.work.sigpending; \ - break; \ - case TIF_NEED_RESCHED: \ - ___res = tsk->thread.work.need_resched; \ - break; \ - case TIF_SYSCALL_TRACE: \ - ___res = tsk->thread.work.syscall_trace;\ - break; \ - case TIF_MEMDIE: \ - ___res = tsk->thread.work.memdie;\ - break; \ - default: \ - ___res = thread_flag_fixme(); \ - } \ - ___res; \ -}) - -#define __get_set_tsk_thread_flag(tsk, flag, val) ({ \ - int __res = __get_tsk_thread_flag(tsk, flag); \ - __set_tsk_thread_flag(tsk, flag, val); \ - __res; \ -}) - -#define set_tsk_thread_flag(tsk, flag) __set_tsk_thread_flag(tsk, flag, ~0) -#define clear_tsk_thread_flag(tsk, flag) __set_tsk_thread_flag(tsk, flag, 0) -#define test_and_set_tsk_thread_flag(tsk, flag) __get_set_tsk_thread_flag(tsk, flag, ~0) -#define test_tsk_thread_flag(tsk, flag) __get_tsk_thread_flag(tsk, flag) - -#define set_thread_flag(flag) set_tsk_thread_flag(current, flag) -#define clear_thread_flag(flag) clear_tsk_thread_flag(current, flag) -#define test_thread_flag(flag) test_tsk_thread_flag(current, flag) - -#define set_need_resched() set_thread_flag(TIF_NEED_RESCHED) -#define clear_need_resched() clear_thread_flag(TIF_NEED_RESCHED) +#define TIF_SIGPENDING 6 /* signal pending */ +#define TIF_NEED_RESCHED 7 /* rescheduling necessary */ +#define TIF_DELAYED_TRACE 14 /* single step a syscall */ +#define TIF_SYSCALL_TRACE 15 /* syscall trace active */ +#define TIF_MEMDIE 16 #endif /* _ASM_M68K_THREAD_INFO_H */ diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h index d252f45a0f9b..1c4eb41dbd89 100644 --- a/include/linux/thread_info.h +++ b/include/linux/thread_info.h @@ -27,31 +27,6 @@ extern long do_no_restart_syscall(struct restart_block *parm); * - pass TIF_xxxx constants to these functions */ -static inline void set_thread_flag(int flag) -{ - set_bit(flag,¤t_thread_info()->flags); -} - -static inline void clear_thread_flag(int flag) -{ - clear_bit(flag,¤t_thread_info()->flags); -} - -static inline int test_and_set_thread_flag(int flag) -{ - return test_and_set_bit(flag,¤t_thread_info()->flags); -} - -static inline int test_and_clear_thread_flag(int flag) -{ - return test_and_clear_bit(flag,¤t_thread_info()->flags); -} - -static inline int test_thread_flag(int flag) -{ - return test_bit(flag,¤t_thread_info()->flags); -} - static inline void set_ti_thread_flag(struct thread_info *ti, int flag) { set_bit(flag,&ti->flags); @@ -77,15 +52,19 @@ static inline int test_ti_thread_flag(struct thread_info *ti, int flag) return test_bit(flag,&ti->flags); } -static inline void set_need_resched(void) -{ - set_thread_flag(TIF_NEED_RESCHED); -} - -static inline void clear_need_resched(void) -{ - clear_thread_flag(TIF_NEED_RESCHED); -} +#define set_thread_flag(flag) \ + set_ti_thread_flag(current_thread_info(), flag) +#define clear_thread_flag(flag) \ + clear_ti_thread_flag(current_thread_info(), flag) +#define test_and_set_thread_flag(flag) \ + test_and_set_ti_thread_flag(current_thread_info(), flag) +#define test_and_clear_thread_flag(flag) \ + test_and_clear_ti_thread_flag(current_thread_info(), flag) +#define test_thread_flag(flag) \ + test_ti_thread_flag(current_thread_info(), flag) + +#define set_need_resched() set_thread_flag(TIF_NEED_RESCHED) +#define clear_need_resched() clear_thread_flag(TIF_NEED_RESCHED) #endif -- cgit v1.2.3-71-gd317 From 66341a905ef5b3e7aea65b5d9bd1b0361b0ccc61 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Sun, 13 Nov 2005 16:07:21 -0800 Subject: [PATCH] Shut up per_cpu_ptr() on UP Currently per_cpu_ptr() doesn't really do anything with 'cpu' in the UP case. This is problematic in the cases where this is the only place the variable is referenced: CC kernel/workqueue.o kernel/workqueue.c: In function `current_is_keventd': kernel/workqueue.c:460: warning: unused variable `cpu' Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/percpu.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 5451eb1e781d..fb8d2d24e4bb 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -38,7 +38,7 @@ extern void free_percpu(const void *); #else /* CONFIG_SMP */ -#define per_cpu_ptr(ptr, cpu) (ptr) +#define per_cpu_ptr(ptr, cpu) ({ (void)(cpu); (ptr); }) static inline void *__alloc_percpu(size_t size, size_t align) { -- cgit v1.2.3-71-gd317 From 20dcae32439384b6863c626bb3b2a09bed65b33e Mon Sep 17 00:00:00 2001 From: Zach Brown Date: Sun, 13 Nov 2005 16:07:33 -0800 Subject: [PATCH] aio: remove kioctx from mm_struct Sync iocbs have a life cycle that don't need a kioctx. Their retrying, if any, is done in the context of their owner who has allocated them on the stack. The sole user of a sync iocb's ctx reference was aio_complete() checking for an elevated iocb ref count that could never happen. No path which grabs an iocb ref has access to sync iocbs. If we were to implement sync iocb cancelation it would be done by the owner of the iocb using its on-stack reference. Removing this chunk from aio_complete allows us to remove the entire kioctx instance from mm_struct, reducing its size by a third. On a i386 testing box the slab size went from 768 to 504 bytes and from 5 to 8 per page. Signed-off-by: Zach Brown Acked-by: Benjamin LaHaise Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/aio.c | 27 +++++++++------------------ include/linux/aio.h | 2 +- include/linux/init_task.h | 1 - include/linux/sched.h | 1 - kernel/fork.c | 1 - 5 files changed, 10 insertions(+), 22 deletions(-) (limited to 'include/linux') diff --git a/fs/aio.c b/fs/aio.c index 20bb919eb195..e7cd40b626b7 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -937,28 +937,19 @@ int fastcall aio_complete(struct kiocb *iocb, long res, long res2) unsigned long tail; int ret; - /* Special case handling for sync iocbs: events go directly - * into the iocb for fast handling. Note that this will not - * work if we allow sync kiocbs to be cancelled. in which - * case the usage count checks will have to move under ctx_lock - * for all cases. + /* + * Special case handling for sync iocbs: + * - events go directly into the iocb for fast handling + * - the sync task with the iocb in its stack holds the single iocb + * ref, no other paths have a way to get another ref + * - the sync task helpfully left a reference to itself in the iocb */ if (is_sync_kiocb(iocb)) { - int ret; - + BUG_ON(iocb->ki_users != 1); iocb->ki_user_data = res; - if (iocb->ki_users == 1) { - iocb->ki_users = 0; - ret = 1; - } else { - spin_lock_irq(&ctx->ctx_lock); - iocb->ki_users--; - ret = (0 == iocb->ki_users); - spin_unlock_irq(&ctx->ctx_lock); - } - /* sync iocbs put the task here for us */ + iocb->ki_users = 0; wake_up_process(iocb->ki_obj.tsk); - return ret; + return 1; } info = &ctx->ring_info; diff --git a/include/linux/aio.h b/include/linux/aio.h index 403d71dcb7c8..9e0ae8711276 100644 --- a/include/linux/aio.h +++ b/include/linux/aio.h @@ -124,7 +124,7 @@ struct kiocb { (x)->ki_users = 1; \ (x)->ki_key = KIOCB_SYNC_KEY; \ (x)->ki_filp = (filp); \ - (x)->ki_ctx = &tsk->active_mm->default_kioctx; \ + (x)->ki_ctx = NULL; \ (x)->ki_cancel = NULL; \ (x)->ki_dtor = NULL; \ (x)->ki_obj.tsk = tsk; \ diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 68ab5f2ab9cd..dcfd2ecccb5d 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -51,7 +51,6 @@ .page_table_lock = SPIN_LOCK_UNLOCKED, \ .mmlist = LIST_HEAD_INIT(name.mmlist), \ .cpu_vm_mask = CPU_MASK_ALL, \ - .default_kioctx = INIT_KIOCTX(name.default_kioctx, name), \ } #define INIT_SIGNALS(sig) { \ diff --git a/include/linux/sched.h b/include/linux/sched.h index 41df81395719..2038bd27b041 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -357,7 +357,6 @@ struct mm_struct { /* aio bits */ rwlock_t ioctx_list_lock; struct kioctx *ioctx_list; - struct kioctx default_kioctx; }; struct sighand_struct { diff --git a/kernel/fork.c b/kernel/fork.c index 2c70c9cdf5dc..e0d0b77343f8 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -323,7 +323,6 @@ static struct mm_struct * mm_init(struct mm_struct * mm) spin_lock_init(&mm->page_table_lock); rwlock_init(&mm->ioctx_list_lock); mm->ioctx_list = NULL; - mm->default_kioctx = (struct kioctx)INIT_KIOCTX(mm->default_kioctx, *mm); mm->free_area_cache = TASK_UNMAPPED_BASE; mm->cached_hole_size = ~0UL; -- cgit v1.2.3-71-gd317 From 5ef1c49f8f9f0d6b5b8d57bb4b66c605a3d65876 Mon Sep 17 00:00:00 2001 From: Zach Brown Date: Sun, 13 Nov 2005 16:07:35 -0800 Subject: [PATCH] aio: don't ref kioctx after decref in put_ioctx put_ioctx's refcount debugging was doing an atomic_read after dropping its reference when it wasn't the last ref, leaving a tiny race for another freeing thread to sneak into. This shifts the debugging before the ops, uses BUG_ON, and reformats the defines a little. Sadly, moving to inlines increased the code size but this change decreases the code size by a whole 9 bytes :) Signed-off-by: Zach Brown Cc: Benjamin LaHaise Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/aio.h | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/aio.h b/include/linux/aio.h index 9e0ae8711276..49fd37629ee4 100644 --- a/include/linux/aio.h +++ b/include/linux/aio.h @@ -210,8 +210,15 @@ struct kioctx *lookup_ioctx(unsigned long ctx_id); int FASTCALL(io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb, struct iocb *iocb)); -#define get_ioctx(kioctx) do { if (unlikely(atomic_read(&(kioctx)->users) <= 0)) BUG(); atomic_inc(&(kioctx)->users); } while (0) -#define put_ioctx(kioctx) do { if (unlikely(atomic_dec_and_test(&(kioctx)->users))) __put_ioctx(kioctx); else if (unlikely(atomic_read(&(kioctx)->users) < 0)) BUG(); } while (0) +#define get_ioctx(kioctx) do { \ + BUG_ON(unlikely(atomic_read(&(kioctx)->users) <= 0)); \ + atomic_inc(&(kioctx)->users); \ +} while (0) +#define put_ioctx(kioctx) do { \ + BUG_ON(unlikely(atomic_read(&(kioctx)->users) <= 0)); \ + if (unlikely(atomic_dec_and_test(&(kioctx)->users))) \ + __put_ioctx(kioctx); \ +} while (0) #define in_aio() !is_sync_wait(current->io_wait) /* may be used for debugging */ -- cgit v1.2.3-71-gd317 From ff6ed4063da39e6a30ce904005e4ed17385e2739 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Sun, 13 Nov 2005 16:07:38 -0800 Subject: [PATCH] acct.h needs jiffies.h allnoconfig: In file included from fs/super.c:28: include/linux/acct.h:173: warning: `TICK_NSEC' is not defined Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/acct.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/acct.h b/include/linux/acct.h index 93c5b3cdf951..9a66401073fc 100644 --- a/include/linux/acct.h +++ b/include/linux/acct.h @@ -16,6 +16,8 @@ #define _LINUX_ACCT_H #include +#include + #include #include -- cgit v1.2.3-71-gd317 From 3f39894d1b5c253b10fcb8fbbbcf65a330f6cdc7 Mon Sep 17 00:00:00 2001 From: George Anzinger Date: Sun, 13 Nov 2005 16:07:44 -0800 Subject: [PATCH] timespec: normalize off by one errors It would appear that the timespec normalize code has an off by one error. Found in three places. Thanks to Ben for spotting. Signed-off-by: George Anzinger Cc: Benjamin Herrenschmidt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/time.h | 2 +- kernel/posix-timers.c | 10 +++------- 2 files changed, 4 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/include/linux/time.h b/include/linux/time.h index 8e83f4e778bb..bfbe92d0767c 100644 --- a/include/linux/time.h +++ b/include/linux/time.h @@ -101,7 +101,7 @@ extern struct timespec timespec_trunc(struct timespec t, unsigned gran); static inline void set_normalized_timespec (struct timespec *ts, time_t sec, long nsec) { - while (nsec > NSEC_PER_SEC) { + while (nsec >= NSEC_PER_SEC) { nsec -= NSEC_PER_SEC; ++sec; } diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index ea55c7a1cd75..5870efb3e200 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -270,7 +270,7 @@ static void tstojiffie(struct timespec *tp, int res, u64 *jiff) long sec = tp->tv_sec; long nsec = tp->tv_nsec + res - 1; - if (nsec > NSEC_PER_SEC) { + if (nsec >= NSEC_PER_SEC) { sec++; nsec -= NSEC_PER_SEC; } @@ -1209,13 +1209,9 @@ static int do_posix_clock_monotonic_get(clockid_t clock, struct timespec *tp) do_posix_clock_monotonic_gettime_parts(tp, &wall_to_mono); - tp->tv_sec += wall_to_mono.tv_sec; - tp->tv_nsec += wall_to_mono.tv_nsec; + set_normalized_timespec(tp, tp->tv_sec + wall_to_mono.tv_sec, + tp->tv_nsec + wall_to_mono.tv_nsec); - if ((tp->tv_nsec - NSEC_PER_SEC) > 0) { - tp->tv_nsec -= NSEC_PER_SEC; - tp->tv_sec++; - } return 0; } -- cgit v1.2.3-71-gd317 From 800d3c6f90b61cc82b09db635b59c00b1c460728 Mon Sep 17 00:00:00 2001 From: Mauro Carvalho Chehab Date: Sun, 13 Nov 2005 16:07:48 -0800 Subject: [PATCH] v4l: (943) added secam l video standard - Added SECAM L' video standard - SECAM L' is a Secam variant that requires special config. This patch adds support on V4L core. Requires aditional patches on tuners to support. Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/media/video/tuner-core.c | 11 ++++++++--- include/linux/videodev2.h | 1 + 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/drivers/media/video/tuner-core.c b/drivers/media/video/tuner-core.c index 73c4041c35d7..e58abdfcaab8 100644 --- a/drivers/media/video/tuner-core.c +++ b/drivers/media/video/tuner-core.c @@ -251,7 +251,7 @@ static inline int check_mode(struct tuner *t, char *cmd) static char pal[] = "-"; module_param_string(pal, pal, sizeof(pal), 0644); -static char secam[] = "-"; +static char secam[] = "--"; module_param_string(secam, secam, sizeof(secam), 0644); /* get more precise norm info from insmod option */ @@ -307,8 +307,13 @@ static int tuner_fixup_std(struct tuner *t) break; case 'l': case 'L': - tuner_dbg ("insmod fixup: SECAM => SECAM-L\n"); - t->std = V4L2_STD_SECAM_L; + if ((secam[1]=='C')||(secam[1]=='c')) { + tuner_dbg ("insmod fixup: SECAM => SECAM-L'\n"); + t->std = V4L2_STD_SECAM_LC; + } else { + tuner_dbg ("insmod fixup: SECAM => SECAM-L\n"); + t->std = V4L2_STD_SECAM_L; + } break; case '-': /* default parameter, do nothing */ diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h index a114fff6568b..1cded681eb6d 100644 --- a/include/linux/videodev2.h +++ b/include/linux/videodev2.h @@ -636,6 +636,7 @@ typedef __u64 v4l2_std_id; #define V4L2_STD_SECAM_K ((v4l2_std_id)0x00100000) #define V4L2_STD_SECAM_K1 ((v4l2_std_id)0x00200000) #define V4L2_STD_SECAM_L ((v4l2_std_id)0x00400000) +#define V4L2_STD_SECAM_LC ((v4l2_std_id)0x00800000) /* ATSC/HDTV */ #define V4L2_STD_ATSC_8_VSB ((v4l2_std_id)0x01000000) -- cgit v1.2.3-71-gd317 From aeec46b97a7975fd983219177980c58ed4fd607c Mon Sep 17 00:00:00 2001 From: Martin Waitz Date: Sun, 13 Nov 2005 16:08:13 -0800 Subject: [PATCH] DocBook: allow to mark structure members private Many structures contain both an internal part and one which is part of the API to other modules. With this patch it is possible to only include these public members in the kernel documentation. Signed-off-by: Martin Waitz Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/usb.h | 6 +++--- scripts/kernel-doc | 13 +++++++++++-- 2 files changed, 14 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/usb.h b/include/linux/usb.h index 748d04385256..856d232c7562 100644 --- a/include/linux/usb.h +++ b/include/linux/usb.h @@ -819,7 +819,7 @@ typedef void (*usb_complete_t)(struct urb *, struct pt_regs *); */ struct urb { - /* private, usb core and host controller only fields in the urb */ + /* private: usb core and host controller only fields in the urb */ struct kref kref; /* reference count of the URB */ spinlock_t lock; /* lock for the URB */ void *hcpriv; /* private data for host controller */ @@ -827,7 +827,7 @@ struct urb atomic_t use_count; /* concurrent submissions counter */ u8 reject; /* submissions will fail */ - /* public, documented fields in the urb that can be used by drivers */ + /* public: documented fields in the urb that can be used by drivers */ struct list_head urb_list; /* list head for use by the urb's * current owner */ struct usb_device *dev; /* (in) pointer to associated device */ @@ -1045,7 +1045,7 @@ struct usb_sg_request { size_t bytes; /* - * members below are private to usbcore, + * members below are private: to usbcore, * and are not provided for driver access! */ spinlock_t lock; diff --git a/scripts/kernel-doc b/scripts/kernel-doc index 8aaf74e64183..2f45fd2969d0 100755 --- a/scripts/kernel-doc +++ b/scripts/kernel-doc @@ -117,6 +117,8 @@ use strict; # struct my_struct { # int a; # int b; +# /* private: */ +# int c; # }; # # All descriptions can be multiline, except the short function description. @@ -1304,6 +1306,12 @@ sub dump_struct($$) { # ignore embedded structs or unions $members =~ s/{.*?}//g; + # ignore members marked private: + $members =~ s/\/\*.*?private:.*?public:.*?\*\///gos; + $members =~ s/\/\*.*?private:.*//gos; + # strip comments: + $members =~ s/\/\*.*?\*\///gos; + create_parameterlist($members, ';', $file); output_declaration($declaration_name, @@ -1329,6 +1337,7 @@ sub dump_enum($$) { my $x = shift; my $file = shift; + $x =~ s@/\*.*?\*/@@gos; # strip comments. if ($x =~ /enum\s+(\w+)\s*{(.*)}/) { $declaration_name = $1; my $members = $2; @@ -1365,6 +1374,7 @@ sub dump_typedef($$) { my $x = shift; my $file = shift; + $x =~ s@/\*.*?\*/@@gos; # strip comments. while (($x =~ /\(*.\)\s*;$/) || ($x =~ /\[*.\]\s*;$/)) { $x =~ s/\(*.\)\s*;$/;/; $x =~ s/\[*.\]\s*;$/;/; @@ -1420,7 +1430,7 @@ sub create_parameterlist($$$) { $type = $arg; $type =~ s/([^\(]+\(\*)$param/$1/; push_parameter($param, $type, $file); - } else { + } elsif ($arg) { $arg =~ s/\s*:\s*/:/g; $arg =~ s/\s*\[/\[/g; @@ -1628,7 +1638,6 @@ sub process_state3_type($$) { my $x = shift; my $file = shift; - $x =~ s@/\*.*?\*/@@gos; # strip comments. $x =~ s@[\r\n]+@ @gos; # strip newlines/cr's. $x =~ s@^\s+@@gos; # strip leading spaces $x =~ s@\s+$@@gos; # strip trailing spaces -- cgit v1.2.3-71-gd317 From 2c13b7cee045af689b36349c2bc6a9ed6e3d73fa Mon Sep 17 00:00:00 2001 From: Jeff Garzik Date: Mon, 14 Nov 2005 14:14:16 -0500 Subject: [libata] minor fixes, new helpers - in ata_dev_identify(), don't assume that all devices are either ATA or ATAPI. In the future, this code will see port multipliers and other devices. - make a debugging printk less verbose - add new helper ata_qc_reinit() - add new helper BPRINTK() and port flag ATA_FLAG_DEBUGMSG, for fine-grained debugging use. --- drivers/scsi/libata-core.c | 11 +++-------- include/linux/libata.h | 14 ++++++++++++++ 2 files changed, 17 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/drivers/scsi/libata-core.c b/drivers/scsi/libata-core.c index 1ccaf467d516..3855bfa8e8d4 100644 --- a/drivers/scsi/libata-core.c +++ b/drivers/scsi/libata-core.c @@ -1263,7 +1263,7 @@ retry: } /* ATAPI-specific feature tests */ - else { + else if (dev->class == ATA_DEV_ATAPI) { if (ata_id_is_ata(dev->id)) /* sanity check */ goto err_out_nosup; @@ -2399,7 +2399,7 @@ static void ata_sg_clean(struct ata_queued_cmd *qc) if (qc->flags & ATA_QCFLAG_SINGLE) assert(qc->n_elem == 1); - DPRINTK("unmapping %u sg elements\n", qc->n_elem); + VPRINTK("unmapping %u sg elements\n", qc->n_elem); /* if we padded the buffer out to 32-bit bound, and data * xfer direction is from-device, we must copy from the @@ -3432,16 +3432,11 @@ struct ata_queued_cmd *ata_qc_new_init(struct ata_port *ap, qc = ata_qc_new(ap); if (qc) { - qc->__sg = NULL; - qc->flags = 0; qc->scsicmd = NULL; qc->ap = ap; qc->dev = dev; - qc->cursect = qc->cursg = qc->cursg_ofs = 0; - qc->nsect = 0; - qc->nbytes = qc->curbytes = 0; - ata_tf_init(ap, &qc->tf, dev->devno); + ata_qc_reinit(qc); } return qc; diff --git a/include/linux/libata.h b/include/linux/libata.h index ad5996183ec2..f2dbb684ce9e 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h @@ -59,6 +59,8 @@ #define VPRINTK(fmt, args...) #endif /* ATA_DEBUG */ +#define BPRINTK(fmt, args...) if (ap->flags & ATA_FLAG_DEBUGMSG) printk(KERN_ERR "%s: " fmt, __FUNCTION__, ## args) + #ifdef ATA_NDEBUG #define assert(expr) #else @@ -119,6 +121,7 @@ enum { ATA_FLAG_PIO_DMA = (1 << 8), /* PIO cmds via DMA */ ATA_FLAG_NOINTR = (1 << 9), /* FIXME: Remove this once * proper HSM is in place. */ + ATA_FLAG_DEBUGMSG = (1 << 10), ATA_QCFLAG_ACTIVE = (1 << 1), /* cmd not yet ack'd to scsi lyer */ ATA_QCFLAG_SG = (1 << 3), /* have s/g table? */ @@ -659,6 +662,17 @@ static inline void ata_tf_init(struct ata_port *ap, struct ata_taskfile *tf, uns tf->device = ATA_DEVICE_OBS | ATA_DEV1; } +static inline void ata_qc_reinit(struct ata_queued_cmd *qc) +{ + qc->__sg = NULL; + qc->flags = 0; + qc->cursect = qc->cursg = qc->cursg_ofs = 0; + qc->nsect = 0; + qc->nbytes = qc->curbytes = 0; + + ata_tf_init(qc->ap, &qc->tf, qc->dev->devno); +} + /** * ata_irq_on - Enable interrupts on a port. -- cgit v1.2.3-71-gd317 From c0400c4f5a08cfd1c657f7f616fcf1dfbd76a4d7 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Mon, 14 Nov 2005 15:21:41 -0800 Subject: [NETFILTER] nfnetlink: skip size check if size not specified (== 0) Skip sizecheck if the size of the attribute wasn't specified, ie. zero. Signed-off-by: Pablo Neira Ayuso Signed-off-by: Harald Welte Signed-off-by: David S. Miller --- include/linux/netfilter/nfnetlink.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/netfilter/nfnetlink.h b/include/linux/netfilter/nfnetlink.h index 72975fa8795d..8be2f840a557 100644 --- a/include/linux/netfilter/nfnetlink.h +++ b/include/linux/netfilter/nfnetlink.h @@ -154,11 +154,14 @@ extern void nfattr_parse(struct nfattr *tb[], int maxattr, #define nfattr_bad_size(tb, max, cta_min) \ ({ int __i, __res = 0; \ - for (__i=0; __i Date: Mon, 14 Nov 2005 15:24:59 -0800 Subject: [NETFILTER] nfnetlink: unconditionally require CAP_NET_ADMIN This patch unconditionally requires CAP_NET_ADMIN for all nfnetlink messages. It also removes the per-message cap_required field, since all existing subsystems use CAP_NET_ADMIN for all their messages anyway. Patrick McHardy owes me a beer if we ever need to re-introduce this. Signed-off-by: Harald Welte Signed-off-by: David S. Miller --- include/linux/netfilter/nfnetlink.h | 1 - net/ipv4/netfilter/ip_conntrack_netlink.c | 21 +++++++-------------- net/netfilter/nfnetlink.c | 28 ++++++++++++---------------- net/netfilter/nfnetlink_log.c | 6 ++---- net/netfilter/nfnetlink_queue.c | 9 +++------ 5 files changed, 24 insertions(+), 41 deletions(-) (limited to 'include/linux') diff --git a/include/linux/netfilter/nfnetlink.h b/include/linux/netfilter/nfnetlink.h index 8be2f840a557..934a2479f160 100644 --- a/include/linux/netfilter/nfnetlink.h +++ b/include/linux/netfilter/nfnetlink.h @@ -112,7 +112,6 @@ struct nfnl_callback { int (*call)(struct sock *nl, struct sk_buff *skb, struct nlmsghdr *nlh, struct nfattr *cda[], int *errp); - kernel_cap_t cap_required; /* capabilities required for this msg */ u_int16_t attr_count; /* number of nfattr's */ }; diff --git a/net/ipv4/netfilter/ip_conntrack_netlink.c b/net/ipv4/netfilter/ip_conntrack_netlink.c index f5e5e3158670..de9f4464438d 100644 --- a/net/ipv4/netfilter/ip_conntrack_netlink.c +++ b/net/ipv4/netfilter/ip_conntrack_netlink.c @@ -1543,29 +1543,22 @@ static struct notifier_block ctnl_notifier_exp = { static struct nfnl_callback ctnl_cb[IPCTNL_MSG_MAX] = { [IPCTNL_MSG_CT_NEW] = { .call = ctnetlink_new_conntrack, - .attr_count = CTA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_MAX, }, [IPCTNL_MSG_CT_GET] = { .call = ctnetlink_get_conntrack, - .attr_count = CTA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_MAX, }, [IPCTNL_MSG_CT_DELETE] = { .call = ctnetlink_del_conntrack, - .attr_count = CTA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_MAX, }, [IPCTNL_MSG_CT_GET_CTRZERO] = { .call = ctnetlink_get_conntrack, - .attr_count = CTA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_MAX, }, }; static struct nfnl_callback ctnl_exp_cb[IPCTNL_MSG_EXP_MAX] = { [IPCTNL_MSG_EXP_GET] = { .call = ctnetlink_get_expect, - .attr_count = CTA_EXPECT_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_EXPECT_MAX, }, [IPCTNL_MSG_EXP_NEW] = { .call = ctnetlink_new_expect, - .attr_count = CTA_EXPECT_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_EXPECT_MAX, }, [IPCTNL_MSG_EXP_DELETE] = { .call = ctnetlink_del_expect, - .attr_count = CTA_EXPECT_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = CTA_EXPECT_MAX, }, }; static struct nfnetlink_subsystem ctnl_subsys = { diff --git a/net/netfilter/nfnetlink.c b/net/netfilter/nfnetlink.c index 83f4c53030fc..a60c59b97631 100644 --- a/net/netfilter/nfnetlink.c +++ b/net/netfilter/nfnetlink.c @@ -223,6 +223,12 @@ static inline int nfnetlink_rcv_msg(struct sk_buff *skb, NFNL_SUBSYS_ID(nlh->nlmsg_type), NFNL_MSG_TYPE(nlh->nlmsg_type)); + if (!cap_raised(NETLINK_CB(skb).eff_cap, CAP_NET_ADMIN)) { + DEBUGP("missing CAP_NET_ADMIN\n"); + *errp = -EPERM; + return -1; + } + /* Only requests are handled by kernel now. */ if (!(nlh->nlmsg_flags & NLM_F_REQUEST)) { DEBUGP("received non-request message\n"); @@ -240,15 +246,12 @@ static inline int nfnetlink_rcv_msg(struct sk_buff *skb, ss = nfnetlink_get_subsys(type); if (!ss) { #ifdef CONFIG_KMOD - if (cap_raised(NETLINK_CB(skb).eff_cap, CAP_NET_ADMIN)) { - /* don't call nfnl_shunlock, since it would reenter - * with further packet processing */ - up(&nfnl_sem); - request_module("nfnetlink-subsys-%d", - NFNL_SUBSYS_ID(type)); - nfnl_shlock(); - ss = nfnetlink_get_subsys(type); - } + /* don't call nfnl_shunlock, since it would reenter + * with further packet processing */ + up(&nfnl_sem); + request_module("nfnetlink-subsys-%d", NFNL_SUBSYS_ID(type)); + nfnl_shlock(); + ss = nfnetlink_get_subsys(type); if (!ss) #endif goto err_inval; @@ -260,13 +263,6 @@ static inline int nfnetlink_rcv_msg(struct sk_buff *skb, goto err_inval; } - if (nc->cap_required && - !cap_raised(NETLINK_CB(skb).eff_cap, nc->cap_required)) { - DEBUGP("permission denied for type %d\n", type); - *errp = -EPERM; - return -1; - } - { u_int16_t attr_count = ss->cb[NFNL_MSG_TYPE(nlh->nlmsg_type)].attr_count; diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c index d194676f3655..cba63729313d 100644 --- a/net/netfilter/nfnetlink_log.c +++ b/net/netfilter/nfnetlink_log.c @@ -862,11 +862,9 @@ out_put: static struct nfnl_callback nfulnl_cb[NFULNL_MSG_MAX] = { [NFULNL_MSG_PACKET] = { .call = nfulnl_recv_unsupp, - .attr_count = NFULA_MAX, - .cap_required = CAP_NET_ADMIN, }, + .attr_count = NFULA_MAX, }, [NFULNL_MSG_CONFIG] = { .call = nfulnl_recv_config, - .attr_count = NFULA_CFG_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = NFULA_CFG_MAX, }, }; static struct nfnetlink_subsystem nfulnl_subsys = { diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c index f065a6c94953..f28460b61e47 100644 --- a/net/netfilter/nfnetlink_queue.c +++ b/net/netfilter/nfnetlink_queue.c @@ -931,14 +931,11 @@ out_put: static struct nfnl_callback nfqnl_cb[NFQNL_MSG_MAX] = { [NFQNL_MSG_PACKET] = { .call = nfqnl_recv_unsupp, - .attr_count = NFQA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = NFQA_MAX, }, [NFQNL_MSG_VERDICT] = { .call = nfqnl_recv_verdict, - .attr_count = NFQA_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = NFQA_MAX, }, [NFQNL_MSG_CONFIG] = { .call = nfqnl_recv_config, - .attr_count = NFQA_CFG_MAX, - .cap_required = CAP_NET_ADMIN }, + .attr_count = NFQA_CFG_MAX, }, }; static struct nfnetlink_subsystem nfqnl_subsys = { -- cgit v1.2.3-71-gd317 From a2f1b424900715ed9d1699c3bb88a434a2b42bc0 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: Add 4GB DMA32 zone Add a new 4GB GFP_DMA32 zone between the GFP_DMA and GFP_NORMAL zones. As a bit of historical background: when the x86-64 port was originally designed we had some discussion if we should use a 16MB DMA zone like i386 or a 4GB DMA zone like IA64 or both. Both was ruled out at this point because it was in early 2.4 when VM is still quite shakey and had bad troubles even dealing with one DMA zone. We settled on the 16MB DMA zone mainly because we worried about older soundcards and the floppy. But this has always caused problems since then because device drivers had trouble getting enough DMA able memory. These days the VM works much better and the wide use of NUMA has proven it can deal with many zones successfully. So this patch adds both zones. This helps drivers who need a lot of memory below 4GB because their hardware is not accessing more (graphic drivers - proprietary and free ones, video frame buffer drivers, sound drivers etc.). Previously they could only use IOMMU+16MB GFP_DMA, which was not enough memory. Another common problem is that hardware who has full memory addressing for >4GB misses it for some control structures in memory (like transmit rings or other metadata). They tended to allocate memory in the 16MB GFP_DMA or the IOMMU/swiotlb then using pci_alloc_consistent, but that can tie up a lot of precious 16MB GFPDMA/IOMMU/swiotlb memory (even on AMD systems the IOMMU tends to be quite small) especially if you have many devices. With the new zone pci_alloc_consistent can just put this stuff into memory below 4GB which works better. One argument was still if the zone should be 4GB or 2GB. The main motivation for 2GB would be an unnamed not so unpopular hardware raid controller (mostly found in older machines from a particular four letter company) who has a strange 2GB restriction in firmware. But that one works ok with swiotlb/IOMMU anyways, so it doesn't really need GFP_DMA32. I chose 4GB to be compatible with IA64 and because it seems to be the most common restriction. The new zone is so far added only for x86-64. For other architectures who don't set up this new zone nothing changes. Architectures can set a compatibility define in Kconfig CONFIG_DMA_IS_DMA32 that will define GFP_DMA32 as GFP_DMA. Otherwise it's a nop because on 32bit architectures it's normally not needed because GFP_NORMAL (=0) is DMA able enough. One problem is still that GFP_DMA means different things on different architectures. e.g. some drivers used to have #ifdef ia64 use GFP_DMA (trusting it to be 4GB) #elif __x86_64__ (use other hacks like the swiotlb because 16MB is not enough) ... . This was quite ugly and is now obsolete. These should be now converted to use GFP_DMA32 unconditionally. I haven't done this yet. Or best only use pci_alloc_consistent/dma_alloc_coherent which will use GFP_DMA32 transparently. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/mm/init.c | 65 ++++++++++++++++++++++++++++++---------------- arch/x86_64/mm/numa.c | 25 ++++-------------- include/asm-x86_64/dma.h | 11 ++++++-- include/asm-x86_64/proto.h | 2 ++ include/linux/gfp.h | 11 ++++++++ include/linux/mmzone.h | 16 +++++++----- mm/page_alloc.c | 15 ++++++++--- 7 files changed, 90 insertions(+), 55 deletions(-) (limited to 'include/linux') diff --git a/arch/x86_64/mm/init.c b/arch/x86_64/mm/init.c index e60a1a848de8..a1ad4cc423a7 100644 --- a/arch/x86_64/mm/init.c +++ b/arch/x86_64/mm/init.c @@ -318,32 +318,51 @@ void zap_low_mappings(void) flush_tlb_all(); } +/* Compute zone sizes for the DMA and DMA32 zones in a node. */ +__init void +size_zones(unsigned long *z, unsigned long *h, + unsigned long start_pfn, unsigned long end_pfn) +{ + int i; + unsigned long w; + + for (i = 0; i < MAX_NR_ZONES; i++) + z[i] = 0; + + if (start_pfn < MAX_DMA_PFN) + z[ZONE_DMA] = MAX_DMA_PFN - start_pfn; + if (start_pfn < MAX_DMA32_PFN) { + unsigned long dma32_pfn = MAX_DMA32_PFN; + if (dma32_pfn > end_pfn) + dma32_pfn = end_pfn; + z[ZONE_DMA32] = dma32_pfn - start_pfn; + } + z[ZONE_NORMAL] = end_pfn - start_pfn; + + /* Remove lower zones from higher ones. */ + w = 0; + for (i = 0; i < MAX_NR_ZONES; i++) { + if (z[i]) + z[i] -= w; + w += z[i]; + } + + /* Compute holes */ + w = 0; + for (i = 0; i < MAX_NR_ZONES; i++) { + unsigned long s = w; + w += z[i]; + h[i] = e820_hole_size(s, w); + } +} + #ifndef CONFIG_NUMA void __init paging_init(void) { - { - unsigned long zones_size[MAX_NR_ZONES]; - unsigned long holes[MAX_NR_ZONES]; - unsigned int max_dma; - - memset(zones_size, 0, sizeof(zones_size)); - memset(holes, 0, sizeof(holes)); - - max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; - - if (end_pfn < max_dma) { - zones_size[ZONE_DMA] = end_pfn; - holes[ZONE_DMA] = e820_hole_size(0, end_pfn); - } else { - zones_size[ZONE_DMA] = max_dma; - holes[ZONE_DMA] = e820_hole_size(0, max_dma); - zones_size[ZONE_NORMAL] = end_pfn - max_dma; - holes[ZONE_NORMAL] = e820_hole_size(max_dma, end_pfn); - } - free_area_init_node(0, NODE_DATA(0), zones_size, - __pa(PAGE_OFFSET) >> PAGE_SHIFT, holes); - } - return; + unsigned long zones[MAX_NR_ZONES], holes[MAX_NR_ZONES]; + size_zones(zones, holes, 0, end_pfn); + free_area_init_node(0, NODE_DATA(0), zones, + __pa(PAGE_OFFSET) >> PAGE_SHIFT, holes); } #endif diff --git a/arch/x86_64/mm/numa.c b/arch/x86_64/mm/numa.c index 214803821001..18e86e2eac2d 100644 --- a/arch/x86_64/mm/numa.c +++ b/arch/x86_64/mm/numa.c @@ -132,29 +132,14 @@ void __init setup_node_zones(int nodeid) unsigned long start_pfn, end_pfn; unsigned long zones[MAX_NR_ZONES]; unsigned long holes[MAX_NR_ZONES]; - unsigned long dma_end_pfn; - memset(zones, 0, sizeof(unsigned long) * MAX_NR_ZONES); - memset(holes, 0, sizeof(unsigned long) * MAX_NR_ZONES); + start_pfn = node_start_pfn(nodeid); + end_pfn = node_end_pfn(nodeid); - start_pfn = node_start_pfn(nodeid); - end_pfn = node_end_pfn(nodeid); + Dprintk(KERN_INFO "setting up node %d %lx-%lx\n", + nodeid, start_pfn, end_pfn); - Dprintk(KERN_INFO "setting up node %d %lx-%lx\n", nodeid, start_pfn, end_pfn); - - /* All nodes > 0 have a zero length zone DMA */ - dma_end_pfn = __pa(MAX_DMA_ADDRESS) >> PAGE_SHIFT; - if (start_pfn < dma_end_pfn) { - zones[ZONE_DMA] = dma_end_pfn - start_pfn; - holes[ZONE_DMA] = e820_hole_size(start_pfn, dma_end_pfn); - zones[ZONE_NORMAL] = end_pfn - dma_end_pfn; - holes[ZONE_NORMAL] = e820_hole_size(dma_end_pfn, end_pfn); - - } else { - zones[ZONE_NORMAL] = end_pfn - start_pfn; - holes[ZONE_NORMAL] = e820_hole_size(start_pfn, end_pfn); - } - + size_zones(zones, holes, start_pfn, end_pfn); free_area_init_node(nodeid, NODE_DATA(nodeid), zones, start_pfn, holes); } diff --git a/include/asm-x86_64/dma.h b/include/asm-x86_64/dma.h index 16fa3a064d0c..6f2a817b6a7c 100644 --- a/include/asm-x86_64/dma.h +++ b/include/asm-x86_64/dma.h @@ -72,8 +72,15 @@ #define MAX_DMA_CHANNELS 8 -/* The maximum address that we can perform a DMA transfer to on this platform */ -#define MAX_DMA_ADDRESS (PAGE_OFFSET+0x1000000) + +/* 16MB ISA DMA zone */ +#define MAX_DMA_PFN ((16*1024*1024) >> PAGE_SHIFT) + +/* 4GB broken PCI/AGP hardware bus master zone */ +#define MAX_DMA32_PFN ((4UL*1024*1024*1024) >> PAGE_SHIFT) + +/* Compat define for old dma zone */ +#define MAX_DMA_ADDRESS ((unsigned long)__va(MAX_DMA_PFN << PAGE_SHIFT)) /* 8237 DMA controllers */ #define IO_DMA1_BASE 0x00 /* 8 bit slave DMA, channels 0..3 */ diff --git a/include/asm-x86_64/proto.h b/include/asm-x86_64/proto.h index dbb37b0adb43..c251152a0658 100644 --- a/include/asm-x86_64/proto.h +++ b/include/asm-x86_64/proto.h @@ -22,6 +22,8 @@ extern void mtrr_bp_init(void); #define mtrr_bp_init() do {} while (0) #endif extern void init_memory_mapping(unsigned long start, unsigned long end); +extern void size_zones(unsigned long *z, unsigned long *h, + unsigned long start_pfn, unsigned long end_pfn); extern void system_call(void); extern int kernel_syscall(void); diff --git a/include/linux/gfp.h b/include/linux/gfp.h index c3779432a723..4351e6bb5a79 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -14,6 +14,13 @@ struct vm_area_struct; /* Zone modifiers in GFP_ZONEMASK (see linux/mmzone.h - low two bits) */ #define __GFP_DMA ((__force gfp_t)0x01u) #define __GFP_HIGHMEM ((__force gfp_t)0x02u) +#ifdef CONFIG_DMA_IS_DMA32 +#define __GFP_DMA32 ((__force gfp_t)0x01) /* ZONE_DMA is ZONE_DMA32 */ +#elif BITS_PER_LONG < 64 +#define __GFP_DMA32 ((__force gfp_t)0x00) /* ZONE_NORMAL is ZONE_DMA32 */ +#else +#define __GFP_DMA32 ((__force gfp_t)0x04) /* Has own ZONE_DMA32 */ +#endif /* * Action modifiers - doesn't change the zoning @@ -64,6 +71,10 @@ struct vm_area_struct; #define GFP_DMA __GFP_DMA +/* 4GB DMA on some platforms */ +#define GFP_DMA32 __GFP_DMA32 + + #define gfp_zone(mask) ((__force int)((mask) & (__force gfp_t)GFP_ZONEMASK)) /* diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index f5fa3082fd6a..da7a829f8561 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -71,10 +71,11 @@ struct per_cpu_pageset { #endif #define ZONE_DMA 0 -#define ZONE_NORMAL 1 -#define ZONE_HIGHMEM 2 +#define ZONE_DMA32 1 +#define ZONE_NORMAL 2 +#define ZONE_HIGHMEM 3 -#define MAX_NR_ZONES 3 /* Sync this with ZONES_SHIFT */ +#define MAX_NR_ZONES 4 /* Sync this with ZONES_SHIFT */ #define ZONES_SHIFT 2 /* ceil(log2(MAX_NR_ZONES)) */ @@ -108,9 +109,10 @@ struct per_cpu_pageset { /* * On machines where it is needed (eg PCs) we divide physical memory - * into multiple physical zones. On a PC we have 3 zones: + * into multiple physical zones. On a PC we have 4 zones: * * ZONE_DMA < 16 MB ISA DMA capable memory + * ZONE_DMA32 0 MB Empty * ZONE_NORMAL 16-896 MB direct mapped by the kernel * ZONE_HIGHMEM > 896 MB only page cache and user processes */ @@ -455,10 +457,10 @@ extern struct pglist_data contig_page_data; #if BITS_PER_LONG == 32 || defined(ARCH_HAS_ATOMIC_UNSIGNED) /* - * with 32 bit page->flags field, we reserve 8 bits for node/zone info. - * there are 3 zones (2 bits) and this leaves 8-2=6 bits for nodes. + * with 32 bit page->flags field, we reserve 9 bits for node/zone info. + * there are 4 zones (3 bits) and this leaves 9-3=6 bits for nodes. */ -#define FLAGS_RESERVED 8 +#define FLAGS_RESERVED 9 #elif BITS_PER_LONG == 64 /* diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2dbdd98426fd..9b43511dbefd 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -60,8 +60,11 @@ long nr_swap_pages; * NORMAL allocation will leave 784M/256 of ram reserved in the ZONE_DMA * HIGHMEM allocation will leave 224M/32 of ram reserved in ZONE_NORMAL * HIGHMEM allocation will (224M+784M)/256 of ram reserved in ZONE_DMA + * + * TBD: should special case ZONE_DMA32 machines here - in those we normally + * don't need any ZONE_NORMAL reservation */ -int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 32 }; +int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 256, 32 }; EXPORT_SYMBOL(totalram_pages); EXPORT_SYMBOL(nr_swap_pages); @@ -73,7 +76,7 @@ EXPORT_SYMBOL(nr_swap_pages); struct zone *zone_table[1 << ZONETABLE_SHIFT] __read_mostly; EXPORT_SYMBOL(zone_table); -static char *zone_names[MAX_NR_ZONES] = { "DMA", "Normal", "HighMem" }; +static char *zone_names[MAX_NR_ZONES] = { "DMA", "DMA32", "Normal", "HighMem" }; int min_free_kbytes = 1024; unsigned long __initdata nr_kernel_pages; @@ -1442,6 +1445,10 @@ static int __init build_zonelists_node(pg_data_t *pgdat, struct zonelist *zoneli zone = pgdat->node_zones + ZONE_NORMAL; if (zone->present_pages) zonelist->zones[j++] = zone; + case ZONE_DMA32: + zone = pgdat->node_zones + ZONE_DMA32; + if (zone->present_pages) + zonelist->zones[j++] = zone; case ZONE_DMA: zone = pgdat->node_zones + ZONE_DMA; if (zone->present_pages) @@ -1456,6 +1463,8 @@ static inline int highest_zone(int zone_bits) int res = ZONE_NORMAL; if (zone_bits & (__force int)__GFP_HIGHMEM) res = ZONE_HIGHMEM; + if (zone_bits & (__force int)__GFP_DMA32) + res = ZONE_DMA32; if (zone_bits & (__force int)__GFP_DMA) res = ZONE_DMA; return res; @@ -1976,7 +1985,7 @@ static void __init free_area_init_core(struct pglist_data *pgdat, if (zholes_size) realsize -= zholes_size[j]; - if (j == ZONE_DMA || j == ZONE_NORMAL) + if (j < ZONE_HIGHMEM) nr_kernel_pages += realsize; nr_all_pages += realsize; -- cgit v1.2.3-71-gd317 From 07808b74e7dab1aa385e698795875337d72daf7d Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: Remove obsolete ARCH_HAS_ATOMIC_UNSIGNED and page_flags_t Has been introduced for x86-64 at some point to save memory in struct page, but has been obsolete for some time. Just remove it. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- include/linux/mm.h | 10 ++-------- include/linux/mmzone.h | 2 +- mm/filemap.c | 2 +- mm/page_alloc.c | 2 +- 4 files changed, 5 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 5c1fb0a2e806..23fad4dae23c 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -206,12 +206,6 @@ struct vm_operations_struct { struct mmu_gather; struct inode; -#ifdef ARCH_HAS_ATOMIC_UNSIGNED -typedef unsigned page_flags_t; -#else -typedef unsigned long page_flags_t; -#endif - /* * Each physical page in the system has a struct page associated with * it to keep track of whatever it is we are using the page for at the @@ -219,7 +213,7 @@ typedef unsigned long page_flags_t; * a page. */ struct page { - page_flags_t flags; /* Atomic flags, some possibly + unsigned long flags; /* Atomic flags, some possibly * updated asynchronously */ atomic_t _count; /* Usage count, see below. */ atomic_t _mapcount; /* Count of ptes mapped in mms, @@ -435,7 +429,7 @@ static inline void put_page(struct page *page) #endif /* Page flags: | [SECTION] | [NODE] | ZONE | ... | FLAGS | */ -#define SECTIONS_PGOFF ((sizeof(page_flags_t)*8) - SECTIONS_WIDTH) +#define SECTIONS_PGOFF ((sizeof(unsigned long)*8) - SECTIONS_WIDTH) #define NODES_PGOFF (SECTIONS_PGOFF - NODES_WIDTH) #define ZONES_PGOFF (NODES_PGOFF - ZONES_WIDTH) diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index da7a829f8561..57fc99c67c31 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -455,7 +455,7 @@ extern struct pglist_data contig_page_data; #include #endif -#if BITS_PER_LONG == 32 || defined(ARCH_HAS_ATOMIC_UNSIGNED) +#if BITS_PER_LONG == 32 /* * with 32 bit page->flags field, we reserve 9 bits for node/zone info. * there are 4 zones (3 bits) and this leaves 9-3=6 bits for nodes. diff --git a/mm/filemap.c b/mm/filemap.c index 5d6e4c2000dc..33a28bfde158 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -134,7 +134,7 @@ static int sync_page(void *word) struct address_space *mapping; struct page *page; - page = container_of((page_flags_t *)word, struct page, flags); + page = container_of((unsigned long *)word, struct page, flags); /* * page_mapping() is being called without PG_locked held. diff --git a/mm/page_alloc.c b/mm/page_alloc.c index e8810a7e0264..259a71bacca4 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -128,7 +128,7 @@ static void bad_page(const char *function, struct page *page) printk(KERN_EMERG "Bad page state at %s (in process '%s', page %p)\n", function, current->comm, page); printk(KERN_EMERG "flags:0x%0*lx mapping:%p mapcount:%d count:%d\n", - (int)(2*sizeof(page_flags_t)), (unsigned long)page->flags, + (int)(2*sizeof(unsigned long)), (unsigned long)page->flags, page->mapping, page_mapcount(page), page_count(page)); printk(KERN_EMERG "Backtrace:\n"); dump_stack(); -- cgit v1.2.3-71-gd317 From 69d81fcde7797342417591ba7affb372b9c86eae Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: Speed up numa_node_id by putting it directly into the PDA Not go from the CPU number to an mapping array. Mode number is often used now in fast paths. This also adds a generic numa_node_id to all the topology includes Suggested by Eric Dumazet Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/setup.c | 4 ++-- arch/x86_64/mm/numa.c | 10 ++++++++-- arch/x86_64/mm/srat.c | 2 +- include/asm-x86_64/numa.h | 2 ++ include/asm-x86_64/pda.h | 1 + include/asm-x86_64/topology.h | 2 ++ include/linux/mmzone.h | 2 ++ 7 files changed, 18 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index f27731ac95c5..99cfa751949a 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -823,7 +823,7 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c) if (!node_online(node)) node = nearby_node(apicid); } - cpu_to_node[cpu] = node; + numa_set_node(cpu, node); printk(KERN_INFO "CPU %d(%d) -> Node %d -> Core %d\n", cpu, c->x86_num_cores, node, cpu_core_id[cpu]); @@ -975,7 +975,7 @@ static void srat_detect_node(void) node = apicid_to_node[hard_smp_processor_id()]; if (node == NUMA_NO_NODE) node = 0; - cpu_to_node[cpu] = node; + numa_set_node(cpu, node); if (acpi_numa > 0) printk(KERN_INFO "CPU %d -> Node %d\n", cpu, node); diff --git a/arch/x86_64/mm/numa.c b/arch/x86_64/mm/numa.c index 18e86e2eac2d..4bf64583ba3b 100644 --- a/arch/x86_64/mm/numa.c +++ b/arch/x86_64/mm/numa.c @@ -156,7 +156,7 @@ void __init numa_init_array(void) for (i = 0; i < NR_CPUS; i++) { if (cpu_to_node[i] != NUMA_NO_NODE) continue; - cpu_to_node[i] = rr; + numa_set_node(i, rr); rr = next_node(rr, node_online_map); if (rr == MAX_NUMNODES) rr = first_node(node_online_map); @@ -242,7 +242,7 @@ void __init numa_initmem_init(unsigned long start_pfn, unsigned long end_pfn) nodes_clear(node_online_map); node_set_online(0); for (i = 0; i < NR_CPUS; i++) - cpu_to_node[i] = 0; + numa_set_node(i, 0); node_to_cpumask[0] = cpumask_of_cpu(0); setup_node_bootmem(0, start_pfn << PAGE_SHIFT, end_pfn << PAGE_SHIFT); } @@ -252,6 +252,12 @@ __cpuinit void numa_add_cpu(int cpu) set_bit(cpu, &node_to_cpumask[cpu_to_node(cpu)]); } +void __cpuinit numa_set_node(int cpu, int node) +{ + cpu_pda[cpu].nodenumber = node; + cpu_to_node[cpu] = node; +} + unsigned long __init numa_free_all_bootmem(void) { int i; diff --git a/arch/x86_64/mm/srat.c b/arch/x86_64/mm/srat.c index 4b2e844c15a7..c7aa08a58041 100644 --- a/arch/x86_64/mm/srat.c +++ b/arch/x86_64/mm/srat.c @@ -203,7 +203,7 @@ int __init acpi_scan_nodes(unsigned long start, unsigned long end) if (cpu_to_node[i] == NUMA_NO_NODE) continue; if (!node_isset(cpu_to_node[i], nodes_parsed)) - cpu_to_node[i] = NUMA_NO_NODE; + numa_set_node(i, NUMA_NO_NODE); } numa_init_array(); return 0; diff --git a/include/asm-x86_64/numa.h b/include/asm-x86_64/numa.h index bcf55c3f7f7f..d51e56fdc3da 100644 --- a/include/asm-x86_64/numa.h +++ b/include/asm-x86_64/numa.h @@ -17,6 +17,8 @@ extern void numa_add_cpu(int cpu); extern void numa_init_array(void); extern int numa_off; +extern void numa_set_node(int cpu, int node); + extern unsigned char apicid_to_node[256]; #define NUMA_NO_NODE 0xff diff --git a/include/asm-x86_64/pda.h b/include/asm-x86_64/pda.h index bbf89aa8a1af..8733ccfa442e 100644 --- a/include/asm-x86_64/pda.h +++ b/include/asm-x86_64/pda.h @@ -15,6 +15,7 @@ struct x8664_pda { int irqcount; /* Irq nesting counter. Starts with -1 */ int cpunumber; /* Logical CPU number */ char *irqstackptr; /* top of irqstack */ + int nodenumber; /* number of current node */ unsigned int __softirq_pending; unsigned int __nmi_count; /* number of NMI on this CPUs */ struct mm_struct *active_mm; diff --git a/include/asm-x86_64/topology.h b/include/asm-x86_64/topology.h index 1c603cd7e4d0..d39ebd5263ed 100644 --- a/include/asm-x86_64/topology.h +++ b/include/asm-x86_64/topology.h @@ -28,6 +28,8 @@ extern int __node_distance(int, int); #define pcibus_to_node(bus) ((long)(bus->sysdata)) #define pcibus_to_cpumask(bus) node_to_cpumask(pcibus_to_node(bus)); +#define numa_node_id() read_pda(nodenumber) + /* sched_domains SD_NODE_INIT for x86_64 machines */ #define SD_NODE_INIT (struct sched_domain) { \ .span = CPU_MASK_NONE, \ diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 57fc99c67c31..f3cffc354dea 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -435,7 +435,9 @@ int lowmem_reserve_ratio_sysctl_handler(struct ctl_table *, int, struct file *, #include /* Returns the number of the current Node. */ +#ifndef numa_node_id #define numa_node_id() (cpu_to_node(raw_smp_processor_id())) +#endif #ifndef CONFIG_NEED_MULTIPLE_NODES -- cgit v1.2.3-71-gd317 From 94605eff572b727aaad9b4b29bc358b919096503 Mon Sep 17 00:00:00 2001 From: "Siddha, Suresh B" Date: Sat, 5 Nov 2005 17:25:54 +0100 Subject: [PATCH] x86-64/i386: Intel HT, Multi core detection fixes Fields obtained through cpuid vector 0x1(ebx[16:23]) and vector 0x4(eax[14:25], eax[26:31]) indicate the maximum values and might not always be the same as what is available and what OS sees. So make sure "siblings" and "cpu cores" values in /proc/cpuinfo reflect the values as seen by OS instead of what cpuid instruction says. This will also fix the buggy BIOS cases (for example where cpuid on a single core cpu says there are "2" siblings, even when HT is disabled in the BIOS. http://bugzilla.kernel.org/show_bug.cgi?id=4359) Signed-off-by: Suresh Siddha Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/i386/kernel/cpu/amd.c | 12 +++--- arch/i386/kernel/cpu/common.c | 36 +++++++---------- arch/i386/kernel/cpu/intel.c | 2 +- arch/i386/kernel/cpu/intel_cacheinfo.c | 2 +- arch/i386/kernel/cpu/proc.c | 7 ++-- arch/i386/kernel/smpboot.c | 73 ++++++++++++++++++++++++---------- arch/x86_64/kernel/setup.c | 69 ++++++++++++++------------------ arch/x86_64/kernel/smpboot.c | 69 +++++++++++++++++++++++++------- include/asm-i386/processor.h | 4 +- include/asm-x86_64/processor.h | 4 +- include/linux/bitops.h | 10 +++++ 11 files changed, 178 insertions(+), 110 deletions(-) (limited to 'include/linux') diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c index 53a1681cd964..e344ef88cfcd 100644 --- a/arch/i386/kernel/cpu/amd.c +++ b/arch/i386/kernel/cpu/amd.c @@ -206,9 +206,9 @@ static void __init init_amd(struct cpuinfo_x86 *c) display_cacheinfo(c); if (cpuid_eax(0x80000000) >= 0x80000008) { - c->x86_num_cores = (cpuid_ecx(0x80000008) & 0xff) + 1; - if (c->x86_num_cores & (c->x86_num_cores - 1)) - c->x86_num_cores = 1; + c->x86_max_cores = (cpuid_ecx(0x80000008) & 0xff) + 1; + if (c->x86_max_cores & (c->x86_max_cores - 1)) + c->x86_max_cores = 1; } #ifdef CONFIG_X86_HT @@ -217,15 +217,15 @@ static void __init init_amd(struct cpuinfo_x86 *c) * distingush the cores. Assumes number of cores is a power * of two. */ - if (c->x86_num_cores > 1) { + if (c->x86_max_cores > 1) { int cpu = smp_processor_id(); unsigned bits = 0; - while ((1 << bits) < c->x86_num_cores) + while ((1 << bits) < c->x86_max_cores) bits++; cpu_core_id[cpu] = phys_proc_id[cpu] & ((1<>= bits; printk(KERN_INFO "CPU %d(%d) -> Core %d\n", - cpu, c->x86_num_cores, cpu_core_id[cpu]); + cpu, c->x86_max_cores, cpu_core_id[cpu]); } #endif } diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c index 35a67dab4a94..4e9c2e99b0a5 100644 --- a/arch/i386/kernel/cpu/common.c +++ b/arch/i386/kernel/cpu/common.c @@ -335,7 +335,7 @@ void __devinit identify_cpu(struct cpuinfo_x86 *c) c->x86_model = c->x86_mask = 0; /* So far unknown... */ c->x86_vendor_id[0] = '\0'; /* Unset */ c->x86_model_id[0] = '\0'; /* Unset */ - c->x86_num_cores = 1; + c->x86_max_cores = 1; memset(&c->x86_capability, 0, sizeof c->x86_capability); if (!have_cpuid_p()) { @@ -446,52 +446,44 @@ void __devinit identify_cpu(struct cpuinfo_x86 *c) void __devinit detect_ht(struct cpuinfo_x86 *c) { u32 eax, ebx, ecx, edx; - int index_msb, tmp; + int index_msb, core_bits; int cpu = smp_processor_id(); + cpuid(1, &eax, &ebx, &ecx, &edx); + + c->apicid = phys_pkg_id((ebx >> 24) & 0xFF, 0); + if (!cpu_has(c, X86_FEATURE_HT) || cpu_has(c, X86_FEATURE_CMP_LEGACY)) return; - cpuid(1, &eax, &ebx, &ecx, &edx); smp_num_siblings = (ebx & 0xff0000) >> 16; if (smp_num_siblings == 1) { printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); } else if (smp_num_siblings > 1 ) { - index_msb = 31; if (smp_num_siblings > NR_CPUS) { printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings); smp_num_siblings = 1; return; } - tmp = smp_num_siblings; - while ((tmp & 0x80000000 ) == 0) { - tmp <<=1 ; - index_msb--; - } - if (smp_num_siblings & (smp_num_siblings - 1)) - index_msb++; + + index_msb = get_count_order(smp_num_siblings); phys_proc_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb); printk(KERN_INFO "CPU: Physical Processor ID: %d\n", phys_proc_id[cpu]); - smp_num_siblings = smp_num_siblings / c->x86_num_cores; + smp_num_siblings = smp_num_siblings / c->x86_max_cores; - tmp = smp_num_siblings; - index_msb = 31; - while ((tmp & 0x80000000) == 0) { - tmp <<=1 ; - index_msb--; - } + index_msb = get_count_order(smp_num_siblings) ; - if (smp_num_siblings & (smp_num_siblings - 1)) - index_msb++; + core_bits = get_count_order(c->x86_max_cores); - cpu_core_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb); + cpu_core_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb) & + ((1 << core_bits) - 1); - if (c->x86_num_cores > 1) + if (c->x86_max_cores > 1) printk(KERN_INFO "CPU: Processor Core ID: %d\n", cpu_core_id[cpu]); } diff --git a/arch/i386/kernel/cpu/intel.c b/arch/i386/kernel/cpu/intel.c index 43601de0f633..8d603ba28126 100644 --- a/arch/i386/kernel/cpu/intel.c +++ b/arch/i386/kernel/cpu/intel.c @@ -157,7 +157,7 @@ static void __devinit init_intel(struct cpuinfo_x86 *c) if ( p ) strcpy(c->x86_model_id, p); - c->x86_num_cores = num_cpu_cores(c); + c->x86_max_cores = num_cpu_cores(c); detect_ht(c); diff --git a/arch/i386/kernel/cpu/intel_cacheinfo.c b/arch/i386/kernel/cpu/intel_cacheinfo.c index 4dc42a189ae5..e66d14099564 100644 --- a/arch/i386/kernel/cpu/intel_cacheinfo.c +++ b/arch/i386/kernel/cpu/intel_cacheinfo.c @@ -307,7 +307,7 @@ static void __cpuinit cache_shared_cpu_map_setup(unsigned int cpu, int index) #ifdef CONFIG_X86_HT else if (num_threads_sharing == smp_num_siblings) this_leaf->shared_cpu_map = cpu_sibling_map[cpu]; - else if (num_threads_sharing == (c->x86_num_cores * smp_num_siblings)) + else if (num_threads_sharing == (c->x86_max_cores * smp_num_siblings)) this_leaf->shared_cpu_map = cpu_core_map[cpu]; else printk(KERN_DEBUG "Number of CPUs sharing cache didn't match " diff --git a/arch/i386/kernel/cpu/proc.c b/arch/i386/kernel/cpu/proc.c index 41b871ecf4b3..e7921315ae9d 100644 --- a/arch/i386/kernel/cpu/proc.c +++ b/arch/i386/kernel/cpu/proc.c @@ -94,12 +94,11 @@ static int show_cpuinfo(struct seq_file *m, void *v) if (c->x86_cache_size >= 0) seq_printf(m, "cache size\t: %d KB\n", c->x86_cache_size); #ifdef CONFIG_X86_HT - if (c->x86_num_cores * smp_num_siblings > 1) { + if (c->x86_max_cores * smp_num_siblings > 1) { seq_printf(m, "physical id\t: %d\n", phys_proc_id[n]); - seq_printf(m, "siblings\t: %d\n", - c->x86_num_cores * smp_num_siblings); + seq_printf(m, "siblings\t: %d\n", cpus_weight(cpu_core_map[n])); seq_printf(m, "core id\t\t: %d\n", cpu_core_id[n]); - seq_printf(m, "cpu cores\t: %d\n", c->x86_num_cores); + seq_printf(m, "cpu cores\t: %d\n", c->booted_cores); } #endif diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c index 01b618e73ecd..0a9c64655236 100644 --- a/arch/i386/kernel/smpboot.c +++ b/arch/i386/kernel/smpboot.c @@ -74,9 +74,11 @@ EXPORT_SYMBOL(phys_proc_id); int cpu_core_id[NR_CPUS] __read_mostly = {[0 ... NR_CPUS-1] = BAD_APICID}; EXPORT_SYMBOL(cpu_core_id); +/* representing HT siblings of each logical CPU */ cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly; EXPORT_SYMBOL(cpu_sibling_map); +/* representing HT and core siblings of each logical CPU */ cpumask_t cpu_core_map[NR_CPUS] __read_mostly; EXPORT_SYMBOL(cpu_core_map); @@ -444,35 +446,60 @@ static void __devinit smp_callin(void) static int cpucount; +/* representing cpus for which sibling maps can be computed */ +static cpumask_t cpu_sibling_setup_map; + static inline void set_cpu_sibling_map(int cpu) { int i; + struct cpuinfo_x86 *c = cpu_data; + + cpu_set(cpu, cpu_sibling_setup_map); if (smp_num_siblings > 1) { - for (i = 0; i < NR_CPUS; i++) { - if (!cpu_isset(i, cpu_callout_map)) - continue; - if (cpu_core_id[cpu] == cpu_core_id[i]) { + for_each_cpu_mask(i, cpu_sibling_setup_map) { + if (phys_proc_id[cpu] == phys_proc_id[i] && + cpu_core_id[cpu] == cpu_core_id[i]) { cpu_set(i, cpu_sibling_map[cpu]); cpu_set(cpu, cpu_sibling_map[i]); + cpu_set(i, cpu_core_map[cpu]); + cpu_set(cpu, cpu_core_map[i]); } } } else { cpu_set(cpu, cpu_sibling_map[cpu]); } - if (current_cpu_data.x86_num_cores > 1) { - for (i = 0; i < NR_CPUS; i++) { - if (!cpu_isset(i, cpu_callout_map)) - continue; - if (phys_proc_id[cpu] == phys_proc_id[i]) { - cpu_set(i, cpu_core_map[cpu]); - cpu_set(cpu, cpu_core_map[i]); - } - } - } else { + if (current_cpu_data.x86_max_cores == 1) { cpu_core_map[cpu] = cpu_sibling_map[cpu]; + c[cpu].booted_cores = 1; + return; + } + + for_each_cpu_mask(i, cpu_sibling_setup_map) { + if (phys_proc_id[cpu] == phys_proc_id[i]) { + cpu_set(i, cpu_core_map[cpu]); + cpu_set(cpu, cpu_core_map[i]); + /* + * Does this new cpu bringup a new core? + */ + if (cpus_weight(cpu_sibling_map[cpu]) == 1) { + /* + * for each core in package, increment + * the booted_cores for this new cpu + */ + if (first_cpu(cpu_sibling_map[i]) == i) + c[cpu].booted_cores++; + /* + * increment the core count for all + * the other cpus in this package + */ + if (i != cpu) + c[i].booted_cores++; + } else if (i != cpu && !c[cpu].booted_cores) + c[cpu].booted_cores = c[i].booted_cores; + } } } @@ -1096,11 +1123,8 @@ static void __init smp_boot_cpus(unsigned int max_cpus) current_thread_info()->cpu = 0; smp_tune_scheduling(); - cpus_clear(cpu_sibling_map[0]); - cpu_set(0, cpu_sibling_map[0]); - cpus_clear(cpu_core_map[0]); - cpu_set(0, cpu_core_map[0]); + set_cpu_sibling_map(0); /* * If we couldn't find an SMP configuration at boot time, @@ -1279,15 +1303,24 @@ static void remove_siblinginfo(int cpu) { int sibling; + struct cpuinfo_x86 *c = cpu_data; + for_each_cpu_mask(sibling, cpu_core_map[cpu]) { + cpu_clear(cpu, cpu_core_map[sibling]); + /* + * last thread sibling in this cpu core going down + */ + if (cpus_weight(cpu_sibling_map[cpu]) == 1) + c[sibling].booted_cores--; + } + for_each_cpu_mask(sibling, cpu_sibling_map[cpu]) cpu_clear(cpu, cpu_sibling_map[sibling]); - for_each_cpu_mask(sibling, cpu_core_map[cpu]) - cpu_clear(cpu, cpu_core_map[sibling]); cpus_clear(cpu_sibling_map[cpu]); cpus_clear(cpu_core_map[cpu]); phys_proc_id[cpu] = BAD_APICID; cpu_core_id[cpu] = BAD_APICID; + cpu_clear(cpu, cpu_sibling_setup_map); } int __cpu_disable(void) diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 476ee034fca2..40c77f6fe4b0 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -795,7 +795,7 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c) #endif bits = 0; - while ((1 << bits) < c->x86_num_cores) + while ((1 << bits) < c->x86_max_cores) bits++; /* Low order bits define the core id (index of core in socket) */ @@ -828,7 +828,7 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c) numa_set_node(cpu, node); printk(KERN_INFO "CPU %d(%d) -> Node %d -> Core %d\n", - cpu, c->x86_num_cores, node, cpu_core_id[cpu]); + cpu, c->x86_max_cores, node, cpu_core_id[cpu]); #endif #endif } @@ -877,9 +877,9 @@ static int __init init_amd(struct cpuinfo_x86 *c) display_cacheinfo(c); if (c->extended_cpuid_level >= 0x80000008) { - c->x86_num_cores = (cpuid_ecx(0x80000008) & 0xff) + 1; - if (c->x86_num_cores & (c->x86_num_cores - 1)) - c->x86_num_cores = 1; + c->x86_max_cores = (cpuid_ecx(0x80000008) & 0xff) + 1; + if (c->x86_max_cores & (c->x86_max_cores - 1)) + c->x86_max_cores = 1; amd_detect_cmp(c); } @@ -891,54 +891,44 @@ static void __cpuinit detect_ht(struct cpuinfo_x86 *c) { #ifdef CONFIG_SMP u32 eax, ebx, ecx, edx; - int index_msb, tmp; + int index_msb, core_bits; int cpu = smp_processor_id(); - + + cpuid(1, &eax, &ebx, &ecx, &edx); + + c->apicid = phys_pkg_id(0); + if (!cpu_has(c, X86_FEATURE_HT) || cpu_has(c, X86_FEATURE_CMP_LEGACY)) return; - cpuid(1, &eax, &ebx, &ecx, &edx); smp_num_siblings = (ebx & 0xff0000) >> 16; - + if (smp_num_siblings == 1) { printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); - } else if (smp_num_siblings > 1) { - index_msb = 31; - /* - * At this point we only support two siblings per - * processor package. - */ + } else if (smp_num_siblings > 1 ) { + if (smp_num_siblings > NR_CPUS) { printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings); smp_num_siblings = 1; return; } - tmp = smp_num_siblings; - while ((tmp & 0x80000000 ) == 0) { - tmp <<=1 ; - index_msb--; - } - if (smp_num_siblings & (smp_num_siblings - 1)) - index_msb++; + + index_msb = get_count_order(smp_num_siblings); phys_proc_id[cpu] = phys_pkg_id(index_msb); - + printk(KERN_INFO "CPU: Physical Processor ID: %d\n", phys_proc_id[cpu]); - smp_num_siblings = smp_num_siblings / c->x86_num_cores; + smp_num_siblings = smp_num_siblings / c->x86_max_cores; - tmp = smp_num_siblings; - index_msb = 31; - while ((tmp & 0x80000000) == 0) { - tmp <<=1 ; - index_msb--; - } - if (smp_num_siblings & (smp_num_siblings - 1)) - index_msb++; + index_msb = get_count_order(smp_num_siblings) ; + + core_bits = get_count_order(c->x86_max_cores); - cpu_core_id[cpu] = phys_pkg_id(index_msb); + cpu_core_id[cpu] = phys_pkg_id(index_msb) & + ((1 << core_bits) - 1); - if (c->x86_num_cores > 1) + if (c->x86_max_cores > 1) printk(KERN_INFO "CPU: Processor Core ID: %d\n", cpu_core_id[cpu]); } @@ -1006,7 +996,7 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) c->x86_cache_alignment = c->x86_clflush_size * 2; if (c->x86 >= 15) set_bit(X86_FEATURE_CONSTANT_TSC, &c->x86_capability); - c->x86_num_cores = intel_num_cpu_cores(c); + c->x86_max_cores = intel_num_cpu_cores(c); srat_detect_node(); } @@ -1044,7 +1034,7 @@ void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c) c->x86_model_id[0] = '\0'; /* Unset */ c->x86_clflush_size = 64; c->x86_cache_alignment = c->x86_clflush_size; - c->x86_num_cores = 1; + c->x86_max_cores = 1; c->extended_cpuid_level = 0; memset(&c->x86_capability, 0, sizeof c->x86_capability); @@ -1278,13 +1268,12 @@ static int show_cpuinfo(struct seq_file *m, void *v) seq_printf(m, "cache size\t: %d KB\n", c->x86_cache_size); #ifdef CONFIG_SMP - if (smp_num_siblings * c->x86_num_cores > 1) { + if (smp_num_siblings * c->x86_max_cores > 1) { int cpu = c - cpu_data; seq_printf(m, "physical id\t: %d\n", phys_proc_id[cpu]); - seq_printf(m, "siblings\t: %d\n", - c->x86_num_cores * smp_num_siblings); + seq_printf(m, "siblings\t: %d\n", cpus_weight(cpu_core_map[cpu])); seq_printf(m, "core id\t\t: %d\n", cpu_core_id[cpu]); - seq_printf(m, "cpu cores\t: %d\n", c->x86_num_cores); + seq_printf(m, "cpu cores\t: %d\n", c->booted_cores); } #endif diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index f74319a80659..2b9ddba61b37 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -64,6 +64,7 @@ int smp_num_siblings = 1; /* Package ID of each logical CPU */ u8 phys_proc_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID }; +/* core ID of each logical CPU */ u8 cpu_core_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID }; EXPORT_SYMBOL(phys_proc_id); EXPORT_SYMBOL(cpu_core_id); @@ -89,7 +90,10 @@ struct cpuinfo_x86 cpu_data[NR_CPUS] __cacheline_aligned; /* Set when the idlers are all forked */ int smp_threads_ready; +/* representing HT siblings of each logical CPU */ cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly; + +/* representing HT and core siblings of each logical CPU */ cpumask_t cpu_core_map[NR_CPUS] __read_mostly; EXPORT_SYMBOL(cpu_core_map); @@ -436,30 +440,59 @@ void __cpuinit smp_callin(void) cpu_set(cpuid, cpu_callin_map); } +/* representing cpus for which sibling maps can be computed */ +static cpumask_t cpu_sibling_setup_map; + static inline void set_cpu_sibling_map(int cpu) { int i; + struct cpuinfo_x86 *c = cpu_data; + + cpu_set(cpu, cpu_sibling_setup_map); if (smp_num_siblings > 1) { - for_each_cpu(i) { - if (cpu_core_id[cpu] == cpu_core_id[i]) { + for_each_cpu_mask(i, cpu_sibling_setup_map) { + if (phys_proc_id[cpu] == phys_proc_id[i] && + cpu_core_id[cpu] == cpu_core_id[i]) { cpu_set(i, cpu_sibling_map[cpu]); cpu_set(cpu, cpu_sibling_map[i]); + cpu_set(i, cpu_core_map[cpu]); + cpu_set(cpu, cpu_core_map[i]); } } } else { cpu_set(cpu, cpu_sibling_map[cpu]); } - if (current_cpu_data.x86_num_cores > 1) { - for_each_cpu(i) { - if (phys_proc_id[cpu] == phys_proc_id[i]) { - cpu_set(i, cpu_core_map[cpu]); - cpu_set(cpu, cpu_core_map[i]); - } - } - } else { + if (current_cpu_data.x86_max_cores == 1) { cpu_core_map[cpu] = cpu_sibling_map[cpu]; + c[cpu].booted_cores = 1; + return; + } + + for_each_cpu_mask(i, cpu_sibling_setup_map) { + if (phys_proc_id[cpu] == phys_proc_id[i]) { + cpu_set(i, cpu_core_map[cpu]); + cpu_set(cpu, cpu_core_map[i]); + /* + * Does this new cpu bringup a new core? + */ + if (cpus_weight(cpu_sibling_map[cpu]) == 1) { + /* + * for each core in package, increment + * the booted_cores for this new cpu + */ + if (first_cpu(cpu_sibling_map[i]) == i) + c[cpu].booted_cores++; + /* + * increment the core count for all + * the other cpus in this package + */ + if (i != cpu) + c[i].booted_cores++; + } else if (i != cpu && !c[cpu].booted_cores) + c[cpu].booted_cores = c[i].booted_cores; + } } } @@ -993,6 +1026,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus) nmi_watchdog_default(); current_cpu_data = boot_cpu_data; current_thread_info()->cpu = 0; /* needed? */ + set_cpu_sibling_map(0); if (smp_sanity_check(max_cpus) < 0) { printk(KERN_INFO "SMP disabled\n"); @@ -1036,8 +1070,6 @@ void __init smp_prepare_boot_cpu(void) int me = smp_processor_id(); cpu_set(me, cpu_online_map); cpu_set(me, cpu_callout_map); - cpu_set(0, cpu_sibling_map[0]); - cpu_set(0, cpu_core_map[0]); per_cpu(cpu_state, me) = CPU_ONLINE; } @@ -1106,15 +1138,24 @@ void __init smp_cpus_done(unsigned int max_cpus) static void remove_siblinginfo(int cpu) { int sibling; + struct cpuinfo_x86 *c = cpu_data; + for_each_cpu_mask(sibling, cpu_core_map[cpu]) { + cpu_clear(cpu, cpu_core_map[sibling]); + /* + * last thread sibling in this cpu core going down + */ + if (cpus_weight(cpu_sibling_map[cpu]) == 1) + c[sibling].booted_cores--; + } + for_each_cpu_mask(sibling, cpu_sibling_map[cpu]) cpu_clear(cpu, cpu_sibling_map[sibling]); - for_each_cpu_mask(sibling, cpu_core_map[cpu]) - cpu_clear(cpu, cpu_core_map[sibling]); cpus_clear(cpu_sibling_map[cpu]); cpus_clear(cpu_core_map[cpu]); phys_proc_id[cpu] = BAD_APICID; cpu_core_id[cpu] = BAD_APICID; + cpu_clear(cpu, cpu_sibling_setup_map); } void remove_cpu_from_maps(void) diff --git a/include/asm-i386/processor.h b/include/asm-i386/processor.h index 0a4ec764377c..9cd4a05234a1 100644 --- a/include/asm-i386/processor.h +++ b/include/asm-i386/processor.h @@ -65,7 +65,9 @@ struct cpuinfo_x86 { int f00f_bug; int coma_bug; unsigned long loops_per_jiffy; - unsigned char x86_num_cores; + unsigned char x86_max_cores; /* cpuid returned max cores value */ + unsigned char booted_cores; /* number of cores as seen by OS */ + unsigned char apicid; } __attribute__((__aligned__(SMP_CACHE_BYTES))); #define X86_VENDOR_INTEL 0 diff --git a/include/asm-x86_64/processor.h b/include/asm-x86_64/processor.h index 03837d34fba0..4861246548f7 100644 --- a/include/asm-x86_64/processor.h +++ b/include/asm-x86_64/processor.h @@ -61,10 +61,12 @@ struct cpuinfo_x86 { int x86_cache_alignment; int x86_tlbsize; /* number of 4K pages in DTLB/ITLB combined(in pages)*/ __u8 x86_virt_bits, x86_phys_bits; - __u8 x86_num_cores; + __u8 x86_max_cores; /* cpuid returned max cores value */ __u32 x86_power; __u32 extended_cpuid_level; /* Max extended CPUID function supported */ unsigned long loops_per_jiffy; + __u8 apicid; + __u8 booted_cores; /* number of cores as seen by OS */ } ____cacheline_aligned; #define X86_VENDOR_INTEL 0 diff --git a/include/linux/bitops.h b/include/linux/bitops.h index cb3c3ef50f50..38c2fb7ebe09 100644 --- a/include/linux/bitops.h +++ b/include/linux/bitops.h @@ -84,6 +84,16 @@ static __inline__ int get_bitmask_order(unsigned int count) return order; /* We could be slightly more clever with -1 here... */ } +static __inline__ int get_count_order(unsigned int count) +{ + int order; + + order = fls(count) - 1; + if (count & (count - 1)) + order++; + return order; +} + /* * hweightN: returns the hamming weight (i.e. the number * of bits set) of a N-bit word -- cgit v1.2.3-71-gd317 From d4ed803c564701eae9534ab26a86ddb06acaf49c Mon Sep 17 00:00:00 2001 From: Harald Welte Date: Tue, 15 Nov 2005 00:09:06 -0800 Subject: [PATCH] Make sysctl.h (again) usable from userspace Make sysctl.h (again) useable from userspace Signed-off-by: Harald Welte Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/sysctl.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h index 64f203c45378..6bc03c911a83 100644 --- a/include/linux/sysctl.h +++ b/include/linux/sysctl.h @@ -20,7 +20,6 @@ #include #include -#include #include struct file; @@ -859,6 +858,7 @@ enum }; #ifdef __KERNEL__ +#include extern void sysctl_init(void); -- cgit v1.2.3-71-gd317