aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--Documentation/lockstat.txt120
-rw-r--r--Documentation/sysrq.txt2
-rw-r--r--arch/i386/kernel/cpu/cpufreq/longhaul.c60
-rw-r--r--drivers/char/vt_ioctl.c4
-rw-r--r--drivers/serial/serial_cs.c1
-rw-r--r--include/linux/sched.h2
-rw-r--r--kernel/signal.c22
-rw-r--r--kernel/time/timer_stats.c5
-rw-r--r--lib/Kconfig.debug2
-rw-r--r--lib/Makefile4
10 files changed, 200 insertions, 22 deletions
diff --git a/Documentation/lockstat.txt b/Documentation/lockstat.txt
new file mode 100644
index 000000000000..4ba4664ce5c3
--- /dev/null
+++ b/Documentation/lockstat.txt
@@ -0,0 +1,120 @@
1
2LOCK STATISTICS
3
4- WHAT
5
6As the name suggests, it provides statistics on locks.
7
8- WHY
9
10Because things like lock contention can severely impact performance.
11
12- HOW
13
14Lockdep already has hooks in the lock functions and maps lock instances to
15lock classes. We build on that. The graph below shows the relation between
16the lock functions and the various hooks therein.
17
18 __acquire
19 |
20 lock _____
21 | \
22 | __contended
23 | |
24 | <wait>
25 | _______/
26 |/
27 |
28 __acquired
29 |
30 .
31 <hold>
32 .
33 |
34 __release
35 |
36 unlock
37
38lock, unlock - the regular lock functions
39__* - the hooks
40<> - states
41
42With these hooks we provide the following statistics:
43
44 con-bounces - number of lock contention that involved x-cpu data
45 contentions - number of lock acquisitions that had to wait
46 wait time min - shortest (non-0) time we ever had to wait for a lock
47 max - longest time we ever had to wait for a lock
48 total - total time we spend waiting on this lock
49 acq-bounces - number of lock acquisitions that involved x-cpu data
50 acquisitions - number of times we took the lock
51 hold time min - shortest (non-0) time we ever held the lock
52 max - longest time we ever held the lock
53 total - total time this lock was held
54
55From these number various other statistics can be derived, such as:
56
57 hold time average = hold time total / acquisitions
58
59These numbers are gathered per lock class, per read/write state (when
60applicable).
61
62It also tracks 4 contention points per class. A contention point is a call site
63that had to wait on lock acquisition.
64
65 - USAGE
66
67Look at the current lock statistics:
68
69( line numbers not part of actual output, done for clarity in the explanation
70 below )
71
72# less /proc/lock_stat
73
7401 lock_stat version 0.2
7502 -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
7603 class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total
7704 -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
7805
7906 &inode->i_data.tree_lock-W: 15 21657 0.18 1093295.30 11547131054.85 58 10415 0.16 87.51 6387.60
8007 &inode->i_data.tree_lock-R: 0 0 0.00 0.00 0.00 23302 231198 0.25 8.45 98023.38
8108 --------------------------
8209 &inode->i_data.tree_lock 0 [<ffffffff8027c08f>] add_to_page_cache+0x5f/0x190
8310
8411 ...............................................................................................................................................................................................
8512
8613 dcache_lock: 1037 1161 0.38 45.32 774.51 6611 243371 0.15 306.48 77387.24
8714 -----------
8815 dcache_lock 180 [<ffffffff802c0d7e>] sys_getcwd+0x11e/0x230
8916 dcache_lock 165 [<ffffffff802c002a>] d_alloc+0x15a/0x210
9017 dcache_lock 33 [<ffffffff8035818d>] _atomic_dec_and_lock+0x4d/0x70
9118 dcache_lock 1 [<ffffffff802beef8>] shrink_dcache_parent+0x18/0x130
92
93This excerpt shows the first two lock class statistics. Line 01 shows the
94output version - each time the format changes this will be updated. Line 02-04
95show the header with column descriptions. Lines 05-10 and 13-18 show the actual
96statistics. These statistics come in two parts; the actual stats separated by a
97short separator (line 08, 14) from the contention points.
98
99The first lock (05-10) is a read/write lock, and shows two lines above the
100short separator. The contention points don't match the column descriptors,
101they have two: contentions and [<IP>] symbol.
102
103
104View the top contending locks:
105
106# grep : /proc/lock_stat | head
107 &inode->i_data.tree_lock-W: 15 21657 0.18 1093295.30 11547131054.85 58 10415 0.16 87.51 6387.60
108 &inode->i_data.tree_lock-R: 0 0 0.00 0.00 0.00 23302 231198 0.25 8.45 98023.38
109 dcache_lock: 1037 1161 0.38 45.32 774.51 6611 243371 0.15 306.48 77387.24
110 &inode->i_mutex: 161 286 18446744073709 62882.54 1244614.55 3653 20598 18446744073709 62318.60 1693822.74
111 &zone->lru_lock: 94 94 0.53 7.33 92.10 4366 32690 0.29 59.81 16350.06
112 &inode->i_data.i_mmap_lock: 79 79 0.40 3.77 53.03 11779 87755 0.28 116.93 29898.44
113 &q->__queue_lock: 48 50 0.52 31.62 86.31 774 13131 0.17 113.08 12277.52
114 &rq->rq_lock_key: 43 47 0.74 68.50 170.63 3706 33929 0.22 107.99 17460.62
115 &rq->rq_lock_key#2: 39 46 0.75 6.68 49.03 2979 32292 0.17 125.17 17137.63
116 tasklist_lock-W: 15 15 1.45 10.87 32.70 1201 7390 0.58 62.55 13648.47
117
118Clear the statistics:
119
120# echo 0 > /proc/lock_stat
diff --git a/Documentation/sysrq.txt b/Documentation/sysrq.txt
index ef19142896ca..10c8f6922ef4 100644
--- a/Documentation/sysrq.txt
+++ b/Documentation/sysrq.txt
@@ -43,7 +43,7 @@ On x86 - You press the key combo 'ALT-SysRq-<command key>'. Note - Some
43 keyboards may not have a key labeled 'SysRq'. The 'SysRq' key is 43 keyboards may not have a key labeled 'SysRq'. The 'SysRq' key is
44 also known as the 'Print Screen' key. Also some keyboards cannot 44 also known as the 'Print Screen' key. Also some keyboards cannot
45 handle so many keys being pressed at the same time, so you might 45 handle so many keys being pressed at the same time, so you might
46 have better luck with "press Alt", "press SysRq", "release Alt", 46 have better luck with "press Alt", "press SysRq", "release SysRq",
47 "press <command key>", release everything. 47 "press <command key>", release everything.
48 48
49On SPARC - You press 'ALT-STOP-<command key>', I believe. 49On SPARC - You press 'ALT-STOP-<command key>', I believe.
diff --git a/arch/i386/kernel/cpu/cpufreq/longhaul.c b/arch/i386/kernel/cpu/cpufreq/longhaul.c
index ef8f0bc3fc71..f0cce3c2dc3a 100644
--- a/arch/i386/kernel/cpu/cpufreq/longhaul.c
+++ b/arch/i386/kernel/cpu/cpufreq/longhaul.c
@@ -76,6 +76,7 @@ static unsigned int longhaul_index;
76/* Module parameters */ 76/* Module parameters */
77static int scale_voltage; 77static int scale_voltage;
78static int disable_acpi_c3; 78static int disable_acpi_c3;
79static int revid_errata;
79 80
80#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "longhaul", msg) 81#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "longhaul", msg)
81 82
@@ -168,7 +169,10 @@ static void do_powersaver(int cx_address, unsigned int clock_ratio_index,
168 169
169 rdmsrl(MSR_VIA_LONGHAUL, longhaul.val); 170 rdmsrl(MSR_VIA_LONGHAUL, longhaul.val);
170 /* Setup new frequency */ 171 /* Setup new frequency */
171 longhaul.bits.RevisionKey = longhaul.bits.RevisionID; 172 if (!revid_errata)
173 longhaul.bits.RevisionKey = longhaul.bits.RevisionID;
174 else
175 longhaul.bits.RevisionKey = 0;
172 longhaul.bits.SoftBusRatio = clock_ratio_index & 0xf; 176 longhaul.bits.SoftBusRatio = clock_ratio_index & 0xf;
173 longhaul.bits.SoftBusRatio4 = (clock_ratio_index & 0x10) >> 4; 177 longhaul.bits.SoftBusRatio4 = (clock_ratio_index & 0x10) >> 4;
174 /* Setup new voltage */ 178 /* Setup new voltage */
@@ -272,7 +276,7 @@ static void longhaul_setstate(unsigned int table_index)
272 276
273 dprintk ("Setting to FSB:%dMHz Mult:%d.%dx (%s)\n", 277 dprintk ("Setting to FSB:%dMHz Mult:%d.%dx (%s)\n",
274 fsb, mult/10, mult%10, print_speed(speed/1000)); 278 fsb, mult/10, mult%10, print_speed(speed/1000));
275 279retry_loop:
276 preempt_disable(); 280 preempt_disable();
277 local_irq_save(flags); 281 local_irq_save(flags);
278 282
@@ -344,6 +348,47 @@ static void longhaul_setstate(unsigned int table_index)
344 preempt_enable(); 348 preempt_enable();
345 349
346 freqs.new = calc_speed(longhaul_get_cpu_mult()); 350 freqs.new = calc_speed(longhaul_get_cpu_mult());
351 /* Check if requested frequency is set. */
352 if (unlikely(freqs.new != speed)) {
353 printk(KERN_INFO PFX "Failed to set requested frequency!\n");
354 /* Revision ID = 1 but processor is expecting revision key
355 * equal to 0. Jumpers at the bottom of processor will change
356 * multiplier and FSB, but will not change bits in Longhaul
357 * MSR nor enable voltage scaling. */
358 if (!revid_errata) {
359 printk(KERN_INFO PFX "Enabling \"Ignore Revision ID\" "
360 "option.\n");
361 revid_errata = 1;
362 msleep(200);
363 goto retry_loop;
364 }
365 /* Why ACPI C3 sometimes doesn't work is a mystery for me.
366 * But it does happen. Processor is entering ACPI C3 state,
367 * but it doesn't change frequency. I tried poking various
368 * bits in northbridge registers, but without success. */
369 if (longhaul_flags & USE_ACPI_C3) {
370 printk(KERN_INFO PFX "Disabling ACPI C3 support.\n");
371 longhaul_flags &= ~USE_ACPI_C3;
372 if (revid_errata) {
373 printk(KERN_INFO PFX "Disabling \"Ignore "
374 "Revision ID\" option.\n");
375 revid_errata = 0;
376 }
377 msleep(200);
378 goto retry_loop;
379 }
380 /* This shouldn't happen. Longhaul ver. 2 was reported not
381 * working on processors without voltage scaling, but with
382 * RevID = 1. RevID errata will make things right. Just
383 * to be 100% sure. */
384 if (longhaul_version == TYPE_LONGHAUL_V2) {
385 printk(KERN_INFO PFX "Switching to Longhaul ver. 1\n");
386 longhaul_version = TYPE_LONGHAUL_V1;
387 msleep(200);
388 goto retry_loop;
389 }
390 }
391 /* Report true CPU frequency */
347 cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); 392 cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
348 393
349 if (!bm_timeout) 394 if (!bm_timeout)
@@ -956,11 +1001,20 @@ static void __exit longhaul_exit(void)
956 kfree(longhaul_table); 1001 kfree(longhaul_table);
957} 1002}
958 1003
1004/* Even if BIOS is exporting ACPI C3 state, and it is used
1005 * with success when CPU is idle, this state doesn't
1006 * trigger frequency transition in some cases. */
959module_param (disable_acpi_c3, int, 0644); 1007module_param (disable_acpi_c3, int, 0644);
960MODULE_PARM_DESC(disable_acpi_c3, "Don't use ACPI C3 support"); 1008MODULE_PARM_DESC(disable_acpi_c3, "Don't use ACPI C3 support");
961 1009/* Change CPU voltage with frequency. Very usefull to save
1010 * power, but most VIA C3 processors aren't supporting it. */
962module_param (scale_voltage, int, 0644); 1011module_param (scale_voltage, int, 0644);
963MODULE_PARM_DESC(scale_voltage, "Scale voltage of processor"); 1012MODULE_PARM_DESC(scale_voltage, "Scale voltage of processor");
1013/* Force revision key to 0 for processors which doesn't
1014 * support voltage scaling, but are introducing itself as
1015 * such. */
1016module_param(revid_errata, int, 0644);
1017MODULE_PARM_DESC(revid_errata, "Ignore CPU Revision ID");
964 1018
965MODULE_AUTHOR ("Dave Jones <davej@codemonkey.org.uk>"); 1019MODULE_AUTHOR ("Dave Jones <davej@codemonkey.org.uk>");
966MODULE_DESCRIPTION ("Longhaul driver for VIA Cyrix processors."); 1020MODULE_DESCRIPTION ("Longhaul driver for VIA Cyrix processors.");
diff --git a/drivers/char/vt_ioctl.c b/drivers/char/vt_ioctl.c
index c799b7f7bbb3..7a61a2a9aafe 100644
--- a/drivers/char/vt_ioctl.c
+++ b/drivers/char/vt_ioctl.c
@@ -1032,7 +1032,7 @@ static DECLARE_WAIT_QUEUE_HEAD(vt_activate_queue);
1032 1032
1033/* 1033/*
1034 * Sleeps until a vt is activated, or the task is interrupted. Returns 1034 * Sleeps until a vt is activated, or the task is interrupted. Returns
1035 * 0 if activation, -EINTR if interrupted. 1035 * 0 if activation, -EINTR if interrupted by a signal handler.
1036 */ 1036 */
1037int vt_waitactive(int vt) 1037int vt_waitactive(int vt)
1038{ 1038{
@@ -1057,7 +1057,7 @@ int vt_waitactive(int vt)
1057 break; 1057 break;
1058 } 1058 }
1059 release_console_sem(); 1059 release_console_sem();
1060 retval = -EINTR; 1060 retval = -ERESTARTNOHAND;
1061 if (signal_pending(current)) 1061 if (signal_pending(current))
1062 break; 1062 break;
1063 schedule(); 1063 schedule();
diff --git a/drivers/serial/serial_cs.c b/drivers/serial/serial_cs.c
index a0ea43598515..7c8d78fbbbfb 100644
--- a/drivers/serial/serial_cs.c
+++ b/drivers/serial/serial_cs.c
@@ -943,6 +943,7 @@ static struct pcmcia_device_id serial_ids[] = {
943 PCMCIA_MFC_DEVICE_PROD_ID12(1,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4), 943 PCMCIA_MFC_DEVICE_PROD_ID12(1,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4),
944 PCMCIA_MFC_DEVICE_PROD_ID12(2,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4), 944 PCMCIA_MFC_DEVICE_PROD_ID12(2,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4),
945 PCMCIA_MFC_DEVICE_PROD_ID12(3,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4), 945 PCMCIA_MFC_DEVICE_PROD_ID12(3,"Elan","Serial Port: SL432",0x3beb8cf2,0x1cce7ac4),
946 PCMCIA_DEVICE_MANF_CARD(0x0279, 0x950b),
946 /* too generic */ 947 /* too generic */
947 /* PCMCIA_MFC_DEVICE_MANF_CARD(0, 0x0160, 0x0002), */ 948 /* PCMCIA_MFC_DEVICE_MANF_CARD(0, 0x0160, 0x0002), */
948 /* PCMCIA_MFC_DEVICE_MANF_CARD(1, 0x0160, 0x0002), */ 949 /* PCMCIA_MFC_DEVICE_MANF_CARD(1, 0x0160, 0x0002), */
diff --git a/include/linux/sched.h b/include/linux/sched.h
index a01ac6dd5f5e..313c6b6e774f 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -113,7 +113,7 @@ extern unsigned long avenrun[]; /* Load averages */
113 113
114#define FSHIFT 11 /* nr of bits of precision */ 114#define FSHIFT 11 /* nr of bits of precision */
115#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */ 115#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
116#define LOAD_FREQ (5*HZ) /* 5 sec intervals */ 116#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
117#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */ 117#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
118#define EXP_5 2014 /* 1/exp(5sec/5min) */ 118#define EXP_5 2014 /* 1/exp(5sec/5min) */
119#define EXP_15 2037 /* 1/exp(5sec/15min) */ 119#define EXP_15 2037 /* 1/exp(5sec/15min) */
diff --git a/kernel/signal.c b/kernel/signal.c
index 9fb91a32edda..792952381092 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -531,18 +531,18 @@ static int check_kill_permission(int sig, struct siginfo *info,
531 if (!valid_signal(sig)) 531 if (!valid_signal(sig))
532 return error; 532 return error;
533 533
534 error = audit_signal_info(sig, t); /* Let audit system see the signal */ 534 if (info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) {
535 if (error) 535 error = audit_signal_info(sig, t); /* Let audit system see the signal */
536 return error; 536 if (error)
537 537 return error;
538 error = -EPERM; 538 error = -EPERM;
539 if ((info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) 539 if (((sig != SIGCONT) ||
540 && ((sig != SIGCONT) || 540 (process_session(current) != process_session(t)))
541 (process_session(current) != process_session(t))) 541 && (current->euid ^ t->suid) && (current->euid ^ t->uid)
542 && (current->euid ^ t->suid) && (current->euid ^ t->uid) 542 && (current->uid ^ t->suid) && (current->uid ^ t->uid)
543 && (current->uid ^ t->suid) && (current->uid ^ t->uid) 543 && !capable(CAP_KILL))
544 && !capable(CAP_KILL))
545 return error; 544 return error;
545 }
546 546
547 return security_task_kill(t, info, sig, 0); 547 return security_task_kill(t, info, sig, 0);
548} 548}
diff --git a/kernel/time/timer_stats.c b/kernel/time/timer_stats.c
index 3c38fb5eae1b..c36bb7ed0301 100644
--- a/kernel/time/timer_stats.c
+++ b/kernel/time/timer_stats.c
@@ -327,8 +327,9 @@ static int tstats_show(struct seq_file *m, void *v)
327 ms = 1; 327 ms = 1;
328 328
329 if (events && period.tv_sec) 329 if (events && period.tv_sec)
330 seq_printf(m, "%ld total events, %ld.%ld events/sec\n", events, 330 seq_printf(m, "%ld total events, %ld.%03ld events/sec\n",
331 events / period.tv_sec, events * 1000 / ms); 331 events, events * 1000 / ms,
332 (events * 1000000 / ms) % 1000);
332 else 333 else
333 seq_printf(m, "%ld total events\n", events); 334 seq_printf(m, "%ld total events\n", events);
334 335
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 495863a500cd..cdc9b099e620 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -294,6 +294,8 @@ config LOCK_STAT
294 help 294 help
295 This feature enables tracking lock contention points 295 This feature enables tracking lock contention points
296 296
297 For more details, see Documentation/lockstat.txt
298
297config DEBUG_LOCKDEP 299config DEBUG_LOCKDEP
298 bool "Lock dependency engine debugging" 300 bool "Lock dependency engine debugging"
299 depends on DEBUG_KERNEL && LOCKDEP 301 depends on DEBUG_KERNEL && LOCKDEP
diff --git a/lib/Makefile b/lib/Makefile
index 6b0ba8cf4e5f..4f3f3e256501 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -2,7 +2,7 @@
2# Makefile for some libs needed in the kernel. 2# Makefile for some libs needed in the kernel.
3# 3#
4 4
5lib-y := ctype.o string.o vsprintf.o kasprintf.o cmdline.o \ 5lib-y := ctype.o string.o vsprintf.o cmdline.o \
6 rbtree.o radix-tree.o dump_stack.o \ 6 rbtree.o radix-tree.o dump_stack.o \
7 idr.o int_sqrt.o bitmap.o extable.o prio_tree.o \ 7 idr.o int_sqrt.o bitmap.o extable.o prio_tree.o \
8 sha1.o irq_regs.o reciprocal_div.o argv_split.o 8 sha1.o irq_regs.o reciprocal_div.o argv_split.o
@@ -13,7 +13,7 @@ lib-$(CONFIG_SMP) += cpumask.o
13lib-y += kobject.o kref.o kobject_uevent.o klist.o 13lib-y += kobject.o kref.o kobject_uevent.o klist.o
14 14
15obj-y += div64.o sort.o parser.o halfmd4.o debug_locks.o random32.o \ 15obj-y += div64.o sort.o parser.o halfmd4.o debug_locks.o random32.o \
16 bust_spinlocks.o hexdump.o 16 bust_spinlocks.o hexdump.o kasprintf.o
17 17
18ifeq ($(CONFIG_DEBUG_KOBJECT),y) 18ifeq ($(CONFIG_DEBUG_KOBJECT),y)
19CFLAGS_kobject.o += -DDEBUG 19CFLAGS_kobject.o += -DDEBUG