2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* arch/s390/mm/fault.c
|
|
|
|
*
|
|
|
|
* S390 version
|
|
|
|
* Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
|
|
|
|
* Author(s): Hartmut Penner (hp@de.ibm.com)
|
|
|
|
* Ulrich Weigand (uweigand@de.ibm.com)
|
|
|
|
*
|
|
|
|
* Derived from "arch/i386/mm/fault.c"
|
|
|
|
* Copyright (C) 1995 Linus Torvalds
|
|
|
|
*/
|
|
|
|
|
perf: Do the big rename: Performance Counters -> Performance Events
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 12:02:48 +02:00
|
|
|
#include <linux/perf_event.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <linux/signal.h>
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/ptrace.h>
|
|
|
|
#include <linux/mman.h>
|
|
|
|
#include <linux/mm.h>
|
2009-06-12 10:26:25 +02:00
|
|
|
#include <linux/compat.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <linux/smp.h>
|
2007-05-08 09:27:03 +02:00
|
|
|
#include <linux/kdebug.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/console.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/hardirq.h>
|
2006-09-20 15:58:39 +02:00
|
|
|
#include <linux/kprobes.h>
|
2007-04-27 16:01:44 +02:00
|
|
|
#include <linux/uaccess.h>
|
2008-04-30 13:38:46 +02:00
|
|
|
#include <linux/hugetlb.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <asm/system.h>
|
|
|
|
#include <asm/pgtable.h>
|
2006-12-04 15:40:40 +01:00
|
|
|
#include <asm/s390_ext.h>
|
2008-02-09 18:24:37 +01:00
|
|
|
#include <asm/mmu_context.h>
|
2008-04-17 07:46:26 +02:00
|
|
|
#include "../kernel/entry.h"
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-01-06 09:19:28 +01:00
|
|
|
#ifndef CONFIG_64BIT
|
2005-04-17 00:20:36 +02:00
|
|
|
#define __FAIL_ADDR_MASK 0x7ffff000
|
|
|
|
#define __FIXUP_MASK 0x7fffffff
|
|
|
|
#define __SUBCODE_MASK 0x0200
|
|
|
|
#define __PF_RES_FIELD 0ULL
|
2006-01-06 09:19:28 +01:00
|
|
|
#else /* CONFIG_64BIT */
|
2005-04-17 00:20:36 +02:00
|
|
|
#define __FAIL_ADDR_MASK -4096L
|
|
|
|
#define __FIXUP_MASK ~0L
|
|
|
|
#define __SUBCODE_MASK 0x0600
|
|
|
|
#define __PF_RES_FIELD 0x8000000000000000ULL
|
2006-01-06 09:19:28 +01:00
|
|
|
#endif /* CONFIG_64BIT */
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
#ifdef CONFIG_SYSCTL
|
|
|
|
extern int sysctl_userprocess_debug;
|
|
|
|
#endif
|
|
|
|
|
2006-09-20 15:58:39 +02:00
|
|
|
#ifdef CONFIG_KPROBES
|
2007-04-27 16:01:43 +02:00
|
|
|
static inline int notify_page_fault(struct pt_regs *regs, long err)
|
|
|
|
{
|
2007-05-04 18:47:46 +02:00
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
/* kprobe_running() needs smp_processor_id() */
|
|
|
|
if (!user_mode(regs)) {
|
|
|
|
preempt_disable();
|
|
|
|
if (kprobe_running() && kprobe_fault_handler(regs, 14))
|
|
|
|
ret = 1;
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
2006-09-20 15:58:39 +02:00
|
|
|
}
|
|
|
|
#else
|
2007-04-27 16:01:43 +02:00
|
|
|
static inline int notify_page_fault(struct pt_regs *regs, long err)
|
2006-09-20 15:58:39 +02:00
|
|
|
{
|
2007-05-04 18:47:46 +02:00
|
|
|
return 0;
|
2006-09-20 15:58:39 +02:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Unlock any spinlocks which will prevent us from getting the
|
2007-02-10 10:46:18 +01:00
|
|
|
* message out.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
|
|
|
void bust_spinlocks(int yes)
|
|
|
|
{
|
|
|
|
if (yes) {
|
|
|
|
oops_in_progress = 1;
|
|
|
|
} else {
|
|
|
|
int loglevel_save = console_loglevel;
|
|
|
|
console_unblank();
|
|
|
|
oops_in_progress = 0;
|
|
|
|
/*
|
|
|
|
* OK, the message is on the console. Now we call printk()
|
|
|
|
* without oops_in_progress set so that printk will give klogd
|
|
|
|
* a poke. Hold onto your hats...
|
|
|
|
*/
|
|
|
|
console_loglevel = 15;
|
|
|
|
printk(" ");
|
|
|
|
console_loglevel = loglevel_save;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2007-03-05 23:35:54 +01:00
|
|
|
* Returns the address space associated with the fault.
|
|
|
|
* Returns 0 for kernel space, 1 for user space and
|
|
|
|
* 2 for code execution in user space with noexec=on.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2007-03-05 23:35:54 +01:00
|
|
|
static inline int check_space(struct task_struct *tsk)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
/*
|
2007-03-05 23:35:54 +01:00
|
|
|
* The lowest two bits of S390_lowcore.trans_exc_code
|
|
|
|
* indicate which paging table was used.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2007-03-05 23:35:54 +01:00
|
|
|
int desc = S390_lowcore.trans_exc_code & 3;
|
|
|
|
|
|
|
|
if (desc == 3) /* Home Segment Table Descriptor */
|
|
|
|
return switch_amode == 0;
|
|
|
|
if (desc == 2) /* Secondary Segment Table Descriptor */
|
|
|
|
return tsk->thread.mm_segment.ar4;
|
|
|
|
#ifdef CONFIG_S390_SWITCH_AMODE
|
|
|
|
if (unlikely(desc == 1)) { /* STD determined via access register */
|
|
|
|
/* %a0 always indicates primary space. */
|
|
|
|
if (S390_lowcore.exc_access_id != 0) {
|
|
|
|
save_access_regs(tsk->thread.acrs);
|
|
|
|
/*
|
|
|
|
* An alet of 0 indicates primary space.
|
|
|
|
* An alet of 1 indicates secondary space.
|
|
|
|
* Any other alet values generate an
|
|
|
|
* alen-translation exception.
|
|
|
|
*/
|
|
|
|
if (tsk->thread.acrs[S390_lowcore.exc_access_id])
|
|
|
|
return tsk->thread.mm_segment.ar4;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
/* Primary Segment Table Descriptor */
|
|
|
|
return switch_amode << s390_noexec;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Send SIGSEGV to task. This is an external routine
|
|
|
|
* to keep the stack usage of do_page_fault small.
|
|
|
|
*/
|
|
|
|
static void do_sigsegv(struct pt_regs *regs, unsigned long error_code,
|
|
|
|
int si_code, unsigned long address)
|
|
|
|
{
|
|
|
|
struct siginfo si;
|
|
|
|
|
|
|
|
#if defined(CONFIG_SYSCTL) || defined(CONFIG_PROCESS_DEBUG)
|
|
|
|
#if defined(CONFIG_SYSCTL)
|
|
|
|
if (sysctl_userprocess_debug)
|
|
|
|
#endif
|
|
|
|
{
|
|
|
|
printk("User process fault: interruption code 0x%lX\n",
|
|
|
|
error_code);
|
|
|
|
printk("failing address: %lX\n", address);
|
|
|
|
show_regs(regs);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
si.si_signo = SIGSEGV;
|
|
|
|
si.si_code = si_code;
|
2006-07-12 16:41:55 +02:00
|
|
|
si.si_addr = (void __user *) address;
|
2005-04-17 00:20:36 +02:00
|
|
|
force_sig_info(SIGSEGV, &si, current);
|
|
|
|
}
|
|
|
|
|
2007-04-27 16:01:43 +02:00
|
|
|
static void do_no_context(struct pt_regs *regs, unsigned long error_code,
|
|
|
|
unsigned long address)
|
|
|
|
{
|
|
|
|
const struct exception_table_entry *fixup;
|
|
|
|
|
|
|
|
/* Are we prepared to handle this kernel fault? */
|
|
|
|
fixup = search_exception_tables(regs->psw.addr & __FIXUP_MASK);
|
|
|
|
if (fixup) {
|
|
|
|
regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Oops. The kernel tried to access some bad page. We'll have to
|
|
|
|
* terminate things with extreme prejudice.
|
|
|
|
*/
|
|
|
|
if (check_space(current) == 0)
|
|
|
|
printk(KERN_ALERT "Unable to handle kernel pointer dereference"
|
|
|
|
" at virtual kernel address %p\n", (void *)address);
|
|
|
|
else
|
|
|
|
printk(KERN_ALERT "Unable to handle kernel paging request"
|
|
|
|
" at virtual user address %p\n", (void *)address);
|
|
|
|
|
|
|
|
die("Oops", regs, error_code);
|
|
|
|
do_exit(SIGKILL);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void do_low_address(struct pt_regs *regs, unsigned long error_code)
|
|
|
|
{
|
|
|
|
/* Low-address protection hit in kernel mode means
|
|
|
|
NULL pointer write access in kernel mode. */
|
|
|
|
if (regs->psw.mask & PSW_MASK_PSTATE) {
|
|
|
|
/* Low-address protection hit in user mode 'cannot happen'. */
|
|
|
|
die ("Low-address protection", regs, error_code);
|
|
|
|
do_exit(SIGKILL);
|
|
|
|
}
|
|
|
|
|
|
|
|
do_no_context(regs, error_code, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void do_sigbus(struct pt_regs *regs, unsigned long error_code,
|
|
|
|
unsigned long address)
|
|
|
|
{
|
|
|
|
struct task_struct *tsk = current;
|
|
|
|
struct mm_struct *mm = tsk->mm;
|
|
|
|
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
/*
|
|
|
|
* Send a sigbus, regardless of whether we were in kernel
|
|
|
|
* or user mode.
|
|
|
|
*/
|
|
|
|
tsk->thread.prot_addr = address;
|
|
|
|
tsk->thread.trap_no = error_code;
|
|
|
|
force_sig(SIGBUS, tsk);
|
|
|
|
|
|
|
|
/* Kernel mode? Handle exceptions or die */
|
|
|
|
if (!(regs->psw.mask & PSW_MASK_PSTATE))
|
|
|
|
do_no_context(regs, error_code, address);
|
|
|
|
}
|
|
|
|
|
2007-02-05 21:18:17 +01:00
|
|
|
#ifdef CONFIG_S390_EXEC_PROTECT
|
2007-04-27 16:01:44 +02:00
|
|
|
static int signal_return(struct mm_struct *mm, struct pt_regs *regs,
|
|
|
|
unsigned long address, unsigned long error_code)
|
2007-02-05 21:18:17 +01:00
|
|
|
{
|
2007-04-27 16:01:44 +02:00
|
|
|
u16 instruction;
|
2007-05-10 15:45:48 +02:00
|
|
|
int rc;
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
int compat;
|
|
|
|
#endif
|
2007-04-27 16:01:44 +02:00
|
|
|
|
|
|
|
pagefault_disable();
|
|
|
|
rc = __get_user(instruction, (u16 __user *) regs->psw.addr);
|
|
|
|
pagefault_enable();
|
|
|
|
if (rc)
|
|
|
|
return -EFAULT;
|
|
|
|
|
2007-02-05 21:18:17 +01:00
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
clear_tsk_thread_flag(current, TIF_SINGLE_STEP);
|
|
|
|
#ifdef CONFIG_COMPAT
|
2009-06-12 10:26:25 +02:00
|
|
|
compat = is_compat_task();
|
2007-04-27 16:01:44 +02:00
|
|
|
if (compat && instruction == 0x0a77)
|
2008-04-17 07:46:07 +02:00
|
|
|
sys32_sigreturn();
|
2007-04-27 16:01:44 +02:00
|
|
|
else if (compat && instruction == 0x0aad)
|
2008-04-17 07:46:07 +02:00
|
|
|
sys32_rt_sigreturn();
|
2007-02-05 21:18:17 +01:00
|
|
|
else
|
2007-04-27 16:01:44 +02:00
|
|
|
#endif
|
|
|
|
if (instruction == 0x0a77)
|
2008-04-17 07:46:07 +02:00
|
|
|
sys_sigreturn();
|
2007-04-27 16:01:44 +02:00
|
|
|
else if (instruction == 0x0aad)
|
2008-04-17 07:46:07 +02:00
|
|
|
sys_rt_sigreturn();
|
2007-02-05 21:18:17 +01:00
|
|
|
else {
|
|
|
|
current->thread.prot_addr = address;
|
|
|
|
current->thread.trap_no = error_code;
|
|
|
|
do_sigsegv(regs, error_code, SEGV_MAPERR, address);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_S390_EXEC_PROTECT */
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* This routine handles page faults. It determines the address,
|
|
|
|
* and the problem, and then passes it off to one of the appropriate
|
|
|
|
* routines.
|
|
|
|
*
|
|
|
|
* error_code:
|
|
|
|
* 04 Protection -> Write-Protection (suprression)
|
|
|
|
* 10 Segment translation -> Not present (nullification)
|
|
|
|
* 11 Page translation -> Not present (nullification)
|
|
|
|
* 3b Region third trans. -> Not present (nullification)
|
|
|
|
*/
|
2007-03-05 23:35:54 +01:00
|
|
|
static inline void
|
2007-04-27 16:01:43 +02:00
|
|
|
do_exception(struct pt_regs *regs, unsigned long error_code, int write)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-04-27 16:01:43 +02:00
|
|
|
struct task_struct *tsk;
|
|
|
|
struct mm_struct *mm;
|
|
|
|
struct vm_area_struct *vma;
|
|
|
|
unsigned long address;
|
2007-03-05 23:35:54 +01:00
|
|
|
int space;
|
2007-04-27 16:01:43 +02:00
|
|
|
int si_code;
|
2007-07-19 10:47:05 +02:00
|
|
|
int fault;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-05-04 18:47:46 +02:00
|
|
|
if (notify_page_fault(regs, error_code))
|
2006-09-20 15:58:39 +02:00
|
|
|
return;
|
|
|
|
|
2007-04-27 16:01:43 +02:00
|
|
|
tsk = current;
|
|
|
|
mm = tsk->mm;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-04-27 16:01:43 +02:00
|
|
|
/* get the failing address and the affected space */
|
|
|
|
address = S390_lowcore.trans_exc_code & __FAIL_ADDR_MASK;
|
2007-03-05 23:35:54 +01:00
|
|
|
space = check_space(tsk);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Verify that the fault happened in user space, that
|
|
|
|
* we are not in an interrupt and that there is a
|
|
|
|
* user context.
|
|
|
|
*/
|
2007-03-05 23:35:54 +01:00
|
|
|
if (unlikely(space == 0 || in_atomic() || !mm))
|
|
|
|
goto no_context;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* When we get here, the fault happened in the current
|
|
|
|
* task's user address space, so we can switch on the
|
|
|
|
* interrupts again and then search the VMAs
|
|
|
|
*/
|
|
|
|
local_irq_enable();
|
perf: Do the big rename: Performance Counters -> Performance Events
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 12:02:48 +02:00
|
|
|
perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
|
2007-04-27 16:01:43 +02:00
|
|
|
down_read(&mm->mmap_sem);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-03-05 23:35:54 +01:00
|
|
|
si_code = SEGV_MAPERR;
|
|
|
|
vma = find_vma(mm, address);
|
|
|
|
if (!vma)
|
|
|
|
goto bad_area;
|
2007-02-05 21:18:17 +01:00
|
|
|
|
|
|
|
#ifdef CONFIG_S390_EXEC_PROTECT
|
2007-03-05 23:35:54 +01:00
|
|
|
if (unlikely((space == 2) && !(vma->vm_flags & VM_EXEC)))
|
2007-02-05 21:18:17 +01:00
|
|
|
if (!signal_return(mm, regs, address, error_code))
|
|
|
|
/*
|
|
|
|
* signal_return() has done an up_read(&mm->mmap_sem)
|
|
|
|
* if it returns 0.
|
|
|
|
*/
|
|
|
|
return;
|
|
|
|
#endif
|
|
|
|
|
2007-04-27 16:01:43 +02:00
|
|
|
if (vma->vm_start <= address)
|
|
|
|
goto good_area;
|
|
|
|
if (!(vma->vm_flags & VM_GROWSDOWN))
|
|
|
|
goto bad_area;
|
|
|
|
if (expand_stack(vma, address))
|
|
|
|
goto bad_area;
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* Ok, we have a good vm_area for this memory access, so
|
|
|
|
* we can handle it..
|
|
|
|
*/
|
|
|
|
good_area:
|
|
|
|
si_code = SEGV_ACCERR;
|
2007-04-27 16:01:43 +02:00
|
|
|
if (!write) {
|
2005-04-17 00:20:36 +02:00
|
|
|
/* page not present, check vm flags */
|
|
|
|
if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))
|
|
|
|
goto bad_area;
|
|
|
|
} else {
|
|
|
|
if (!(vma->vm_flags & VM_WRITE))
|
|
|
|
goto bad_area;
|
|
|
|
}
|
|
|
|
|
2008-04-30 13:38:46 +02:00
|
|
|
if (is_vm_hugetlb_page(vma))
|
|
|
|
address &= HPAGE_MASK;
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* If for any reason at all we couldn't handle the fault,
|
|
|
|
* make sure we exit gracefully rather than endlessly redo
|
|
|
|
* the fault.
|
|
|
|
*/
|
2009-04-10 18:01:23 +02:00
|
|
|
fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0);
|
2007-07-19 10:47:05 +02:00
|
|
|
if (unlikely(fault & VM_FAULT_ERROR)) {
|
|
|
|
if (fault & VM_FAULT_OOM) {
|
2009-03-26 15:23:44 +01:00
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
pagefault_out_of_memory();
|
2007-07-19 10:47:05 +02:00
|
|
|
return;
|
|
|
|
} else if (fault & VM_FAULT_SIGBUS) {
|
|
|
|
do_sigbus(regs, error_code, address);
|
|
|
|
return;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
BUG();
|
|
|
|
}
|
2009-09-11 10:29:06 +02:00
|
|
|
if (fault & VM_FAULT_MAJOR) {
|
2007-07-19 10:47:05 +02:00
|
|
|
tsk->maj_flt++;
|
perf: Do the big rename: Performance Counters -> Performance Events
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 12:02:48 +02:00
|
|
|
perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
|
2009-09-11 10:29:06 +02:00
|
|
|
regs, address);
|
|
|
|
} else {
|
2007-07-19 10:47:05 +02:00
|
|
|
tsk->min_flt++;
|
perf: Do the big rename: Performance Counters -> Performance Events
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 12:02:48 +02:00
|
|
|
perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
|
2009-09-11 10:29:06 +02:00
|
|
|
regs, address);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
/*
|
|
|
|
* The instruction that caused the program check will
|
|
|
|
* be repeated. Don't signal single step via SIGTRAP.
|
|
|
|
*/
|
2007-03-05 23:35:54 +01:00
|
|
|
clear_tsk_thread_flag(tsk, TIF_SINGLE_STEP);
|
2005-04-17 00:20:36 +02:00
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Something tried to access memory that isn't in our memory map..
|
|
|
|
* Fix it, but check if it's kernel or user first..
|
|
|
|
*/
|
|
|
|
bad_area:
|
2007-04-27 16:01:43 +02:00
|
|
|
up_read(&mm->mmap_sem);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-04-27 16:01:43 +02:00
|
|
|
/* User mode accesses just cause a SIGSEGV */
|
|
|
|
if (regs->psw.mask & PSW_MASK_PSTATE) {
|
|
|
|
tsk->thread.prot_addr = address;
|
|
|
|
tsk->thread.trap_no = error_code;
|
2005-04-17 00:20:36 +02:00
|
|
|
do_sigsegv(regs, error_code, si_code, address);
|
2007-04-27 16:01:43 +02:00
|
|
|
return;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
no_context:
|
2007-04-27 16:01:43 +02:00
|
|
|
do_no_context(regs, error_code, address);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-03-05 23:35:54 +01:00
|
|
|
void __kprobes do_protection_exception(struct pt_regs *regs,
|
2008-04-17 07:46:26 +02:00
|
|
|
long error_code)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-04-27 16:01:43 +02:00
|
|
|
/* Protection exception is supressing, decrement psw address. */
|
2005-04-17 00:20:36 +02:00
|
|
|
regs->psw.addr -= (error_code >> 16);
|
2007-04-27 16:01:43 +02:00
|
|
|
/*
|
|
|
|
* Check for low-address protection. This needs to be treated
|
|
|
|
* as a special case because the translation exception code
|
|
|
|
* field is not guaranteed to contain valid data in this case.
|
|
|
|
*/
|
|
|
|
if (unlikely(!(S390_lowcore.trans_exc_code & 4))) {
|
|
|
|
do_low_address(regs, error_code);
|
|
|
|
return;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
do_exception(regs, 4, 1);
|
|
|
|
}
|
|
|
|
|
2008-04-17 07:46:26 +02:00
|
|
|
void __kprobes do_dat_exception(struct pt_regs *regs, long error_code)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
do_exception(regs, error_code & 0xff, 0);
|
|
|
|
}
|
|
|
|
|
2008-02-09 18:24:37 +01:00
|
|
|
#ifdef CONFIG_64BIT
|
|
|
|
void __kprobes do_asce_exception(struct pt_regs *regs, unsigned long error_code)
|
|
|
|
{
|
|
|
|
struct mm_struct *mm;
|
|
|
|
struct vm_area_struct *vma;
|
|
|
|
unsigned long address;
|
|
|
|
int space;
|
|
|
|
|
|
|
|
mm = current->mm;
|
|
|
|
address = S390_lowcore.trans_exc_code & __FAIL_ADDR_MASK;
|
|
|
|
space = check_space(current);
|
|
|
|
|
|
|
|
if (unlikely(space == 0 || in_atomic() || !mm))
|
|
|
|
goto no_context;
|
|
|
|
|
|
|
|
local_irq_enable();
|
|
|
|
|
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
vma = find_vma(mm, address);
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
|
|
|
|
if (vma) {
|
|
|
|
update_mm(mm, current);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* User mode accesses just cause a SIGSEGV */
|
|
|
|
if (regs->psw.mask & PSW_MASK_PSTATE) {
|
|
|
|
current->thread.prot_addr = address;
|
|
|
|
current->thread.trap_no = error_code;
|
|
|
|
do_sigsegv(regs, error_code, SEGV_MAPERR, address);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
no_context:
|
|
|
|
do_no_context(regs, error_code, address);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
#ifdef CONFIG_PFAULT
|
|
|
|
/*
|
|
|
|
* 'pfault' pseudo page faults routines.
|
|
|
|
*/
|
2006-12-04 15:40:40 +01:00
|
|
|
static ext_int_info_t ext_int_pfault;
|
2005-04-17 00:20:36 +02:00
|
|
|
static int pfault_disable = 0;
|
|
|
|
|
|
|
|
static int __init nopfault(char *str)
|
|
|
|
{
|
|
|
|
pfault_disable = 1;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
__setup("nopfault", nopfault);
|
|
|
|
|
|
|
|
typedef struct {
|
|
|
|
__u16 refdiagc;
|
|
|
|
__u16 reffcode;
|
|
|
|
__u16 refdwlen;
|
|
|
|
__u16 refversn;
|
|
|
|
__u64 refgaddr;
|
|
|
|
__u64 refselmk;
|
|
|
|
__u64 refcmpmk;
|
|
|
|
__u64 reserved;
|
2007-10-12 16:11:51 +02:00
|
|
|
} __attribute__ ((packed, aligned(8))) pfault_refbk_t;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
int pfault_init(void)
|
|
|
|
{
|
|
|
|
pfault_refbk_t refbk =
|
|
|
|
{ 0x258, 0, 5, 2, __LC_CURRENT, 1ULL << 48, 1ULL << 48,
|
|
|
|
__PF_RES_FIELD };
|
|
|
|
int rc;
|
|
|
|
|
2006-12-04 15:40:40 +01:00
|
|
|
if (!MACHINE_IS_VM || pfault_disable)
|
2005-04-17 00:20:36 +02:00
|
|
|
return -1;
|
2006-09-28 16:56:43 +02:00
|
|
|
asm volatile(
|
|
|
|
" diag %1,%0,0x258\n"
|
|
|
|
"0: j 2f\n"
|
|
|
|
"1: la %0,8\n"
|
2005-04-17 00:20:36 +02:00
|
|
|
"2:\n"
|
2006-09-28 16:56:43 +02:00
|
|
|
EX_TABLE(0b,1b)
|
|
|
|
: "=d" (rc) : "a" (&refbk), "m" (refbk) : "cc");
|
2005-04-17 00:20:36 +02:00
|
|
|
__ctl_set_bit(0, 9);
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
void pfault_fini(void)
|
|
|
|
{
|
|
|
|
pfault_refbk_t refbk =
|
|
|
|
{ 0x258, 1, 5, 2, 0ULL, 0ULL, 0ULL, 0ULL };
|
|
|
|
|
2006-12-04 15:40:40 +01:00
|
|
|
if (!MACHINE_IS_VM || pfault_disable)
|
2005-04-17 00:20:36 +02:00
|
|
|
return;
|
|
|
|
__ctl_clear_bit(0,9);
|
2006-09-28 16:56:43 +02:00
|
|
|
asm volatile(
|
|
|
|
" diag %0,0,0x258\n"
|
2005-04-17 00:20:36 +02:00
|
|
|
"0:\n"
|
2006-09-28 16:56:43 +02:00
|
|
|
EX_TABLE(0b,0b)
|
|
|
|
: : "a" (&refbk), "m" (refbk) : "cc");
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-02-05 21:16:47 +01:00
|
|
|
static void pfault_interrupt(__u16 error_code)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct task_struct *tsk;
|
|
|
|
__u16 subcode;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get the external interruption subcode & pfault
|
|
|
|
* initial/completion signal bit. VM stores this
|
|
|
|
* in the 'cpu address' field associated with the
|
|
|
|
* external interrupt.
|
|
|
|
*/
|
|
|
|
subcode = S390_lowcore.cpu_addr;
|
|
|
|
if ((subcode & 0xff00) != __SUBCODE_MASK)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get the token (= address of the task structure of the affected task).
|
|
|
|
*/
|
|
|
|
tsk = *(struct task_struct **) __LC_PFAULT_INTPARM;
|
|
|
|
|
|
|
|
if (subcode & 0x0080) {
|
|
|
|
/* signal bit is set -> a page has been swapped in by VM */
|
|
|
|
if (xchg(&tsk->thread.pfault_wait, -1) != 0) {
|
|
|
|
/* Initial interrupt was faster than the completion
|
|
|
|
* interrupt. pfault_wait is valid. Set pfault_wait
|
|
|
|
* back to zero and wake up the process. This can
|
|
|
|
* safely be done because the task is still sleeping
|
2005-09-04 00:58:02 +02:00
|
|
|
* and can't produce new pfaults. */
|
2005-04-17 00:20:36 +02:00
|
|
|
tsk->thread.pfault_wait = 0;
|
|
|
|
wake_up_process(tsk);
|
2005-09-04 00:58:02 +02:00
|
|
|
put_task_struct(tsk);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* signal bit not set -> a real page is missing. */
|
2005-09-04 00:58:02 +02:00
|
|
|
get_task_struct(tsk);
|
2005-04-17 00:20:36 +02:00
|
|
|
set_task_state(tsk, TASK_UNINTERRUPTIBLE);
|
|
|
|
if (xchg(&tsk->thread.pfault_wait, 1) != 0) {
|
|
|
|
/* Completion interrupt was faster than the initial
|
|
|
|
* interrupt (swapped in a -1 for pfault_wait). Set
|
|
|
|
* pfault_wait back to zero and exit. This can be
|
|
|
|
* done safely because tsk is running in kernel
|
|
|
|
* mode and can't produce new pfaults. */
|
|
|
|
tsk->thread.pfault_wait = 0;
|
|
|
|
set_task_state(tsk, TASK_RUNNING);
|
2005-09-04 00:58:02 +02:00
|
|
|
put_task_struct(tsk);
|
2005-04-17 00:20:36 +02:00
|
|
|
} else
|
|
|
|
set_tsk_need_resched(tsk);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-12-04 15:40:40 +01:00
|
|
|
void __init pfault_irq_init(void)
|
|
|
|
{
|
|
|
|
if (!MACHINE_IS_VM)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Try to get pfault pseudo page faults going.
|
|
|
|
*/
|
|
|
|
if (register_early_external_interrupt(0x2603, pfault_interrupt,
|
|
|
|
&ext_int_pfault) != 0)
|
|
|
|
panic("Couldn't request external interrupt 0x2603");
|
|
|
|
|
|
|
|
if (pfault_init() == 0)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/* Tough luck, no pfault. */
|
|
|
|
pfault_disable = 1;
|
|
|
|
unregister_early_external_interrupt(0x2603, pfault_interrupt,
|
|
|
|
&ext_int_pfault);
|
|
|
|
}
|
|
|
|
#endif
|