redonkable
/
alistair23-linux
1
0
Fork 0
Code Releases Activity
alistair23-linux/arch/sparc/kernel/perf_event.c

1382 lines
33 KiB
C
Raw Normal View History

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 04:02:48 -06:00
/* Performance event support for sparc64.
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
*
sparc64: Add perf callchain support. Pretty straightforward, and it should be easy to add accurate walk through of signal stack frames in userspace. Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Jens Axboe <jens.axboe@oracle.com>
2010-01-19 01:26:13 -07:00
* Copyright (C) 2009, 2010 David S. Miller <davem@davemloft.net>
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
*
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 04:02:48 -06:00
* This code is based almost entirely upon the x86 perf event
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
* code, which is:
*
* Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
* Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar
* Copyright (C) 2009 Jaswinder Singh Rajput
* Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter
* Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
*/
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 04:02:48 -06:00
#include <linux/perf_event.h>
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
#include <linux/kprobes.h>
#include <linux/kernel.h>
#include <linux/kdebug.h>
#include <linux/mutex.h>
sparc64: Add perf callchain support. Pretty straightforward, and it should be easy to add accurate walk through of signal stack frames in userspace. Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Jens Axboe <jens.axboe@oracle.com>
2010-01-19 01:26:13 -07:00
#include <asm/stacktrace.h>
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
#include <asm/cpudata.h>
sparc64: Add perf callchain support. Pretty straightforward, and it should be easy to add accurate walk through of signal stack frames in userspace. Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Jens Axboe <jens.axboe@oracle.com>
2010-01-19 01:26:13 -07:00
#include <asm/uaccess.h>
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
#include <asm/atomic.h>
#include <asm/nmi.h>
#include <asm/pcr.h>
sparc64: Add perf callchain support. Pretty straightforward, and it should be easy to add accurate walk through of signal stack frames in userspace. Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Jens Axboe <jens.axboe@oracle.com>
2010-01-19 01:26:13 -07:00
#include "kstack.h"
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
/* Sparc64 chips have two performance counters, 32-bits each, with
* overflow interrupts generated on transition from 0xffffffff to 0.
* The counters are accessed in one go using a 64-bit register.
*
* Both counters are controlled using a single control register. The
* only way to stop all sampling is to clear all of the context (user,
* supervisor, hypervisor) sampling enable bits. But these bits apply
* to both counters, thus the two counters can't be enabled/disabled
* individually.
*
* The control register has two event fields, one for each of the two
* counters. It's thus nearly impossible to have one counter going
* while keeping the other one stopped. Therefore it is possible to
* get overflow interrupts for counters not currently "in use" and
* that condition must be checked in the overflow interrupt handler.
*
* So we use a hack, in that we program inactive counters with the
* "sw_count0" and "sw_count1" events. These count how many times
* the instruction "sethi %hi(0xfc000), %g0" is executed. It's an
* unusual way to encode a NOP and therefore will not trigger in
* normal code.
*/
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 04:02:48 -06:00
#define MAX_HWEVENTS 2
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
#define MAX_PERIOD ((1UL << 32) - 1)
#define PIC_UPPER_INDEX 0
#define PIC_LOWER_INDEX 1
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
#define PIC_NO_INDEX -1
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
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 04:02:48 -06:00
struct cpu_hw_events {
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* Number of events currently scheduled onto this cpu.
* This tells how many entries in the arrays below
* are valid.
*/
int n_events;
/* Number of new events added since the last hw_perf_disable().
* This works because the perf event layer always adds new
* events inside of a perf_{disable,enable}() sequence.
*/
int n_added;
/* Array of events current scheduled on this cpu. */
struct perf_event *event[MAX_HWEVENTS];
/* Array of encoded longs, specifying the %pcr register
* encoding and the mask of PIC counters this even can
* be scheduled on. See perf_event_encode() et al.
*/
unsigned long events[MAX_HWEVENTS];
/* The current counter index assigned to an event. When the
* event hasn't been programmed into the cpu yet, this will
* hold PIC_NO_INDEX. The event->hw.idx value tells us where
* we ought to schedule the event.
*/
int current_idx[MAX_HWEVENTS];
/* Software copy of %pcr register on this cpu. */
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
u64 pcr;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* Enabled/disable state. */
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
int enabled;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
};
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 04:02:48 -06:00
DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { .enabled = 1, };
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* An event map describes the characteristics of a performance
* counter event. In particular it gives the encoding as well as
* a mask telling which counters the event can be measured on.
*/
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
struct perf_event_map {
u16 encoding;
u8 pic_mask;
#define PIC_NONE 0x00
#define PIC_UPPER 0x01
#define PIC_LOWER 0x02
};
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* Encode a perf_event_map entry into a long. */
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
static unsigned long perf_event_encode(const struct perf_event_map *pmap)
{
return ((unsigned long) pmap->encoding << 16) | pmap->pic_mask;
}
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
static u8 perf_event_get_msk(unsigned long val)
{
return val & 0xff;
}
static u64 perf_event_get_enc(unsigned long val)
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
{
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
return val >> 16;
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
}
sparc: Support HW cache events. First supported chip for HW cache events is Ultra-IIIi. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:42:10 -06:00
#define C(x) PERF_COUNT_HW_CACHE_##x
#define CACHE_OP_UNSUPPORTED 0xfffe
#define CACHE_OP_NONSENSE 0xffff
typedef struct perf_event_map cache_map_t
[PERF_COUNT_HW_CACHE_MAX]
[PERF_COUNT_HW_CACHE_OP_MAX]
[PERF_COUNT_HW_CACHE_RESULT_MAX];
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
struct sparc_pmu {
const struct perf_event_map *(*event_map)(int);
sparc: Support HW cache events. First supported chip for HW cache events is Ultra-IIIi. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:42:10 -06:00
const cache_map_t *cache_map;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
int max_events;
int upper_shift;
int lower_shift;
int event_mask;
sparc64: Provide hypervisor tracing bit support for perf counters. A PMU need only specify which bit in the PCR enabled hypervisor tracing in order to enable this. This will be used in Niagara-2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:09:06 -06:00
int hv_bit;
sparc64: Provide a way to specify a perf counter overflow IRQ enable bit. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:10:59 -06:00
int irq_bit;
sparc64: Perf counter 'nop' event is not constant. On Niagara-2, for example, it's going to be different. So make it something specified in sparc_pmu. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:13:26 -06:00
int upper_nop;
int lower_nop;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
};
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
static const struct perf_event_map ultra3_perfmon_event_map[] = {
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
[PERF_COUNT_HW_CPU_CYCLES] = { 0x0000, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_INSTRUCTIONS] = { 0x0001, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_CACHE_REFERENCES] = { 0x0009, PIC_LOWER },
[PERF_COUNT_HW_CACHE_MISSES] = { 0x0009, PIC_UPPER },
};
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
static const struct perf_event_map *ultra3_event_map(int event_id)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
return &ultra3_perfmon_event_map[event_id];
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
static const cache_map_t ultra3_cache_map = {
sparc: Support HW cache events. First supported chip for HW cache events is Ultra-IIIi. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:42:10 -06:00
[C(L1D)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x09, PIC_LOWER, },
[C(RESULT_MISS)] = { 0x09, PIC_UPPER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { 0x0a, PIC_LOWER },
[C(RESULT_MISS)] = { 0x0a, PIC_UPPER },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(L1I)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x09, PIC_LOWER, },
[C(RESULT_MISS)] = { 0x09, PIC_UPPER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_NONSENSE },
[ C(RESULT_MISS) ] = { CACHE_OP_NONSENSE },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(LL)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x0c, PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0c, PIC_UPPER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { 0x0c, PIC_LOWER },
[C(RESULT_MISS)] = { 0x0c, PIC_UPPER },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(DTLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x12, PIC_UPPER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(ITLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x11, PIC_UPPER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(BPU)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
};
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
static const struct sparc_pmu ultra3_pmu = {
.event_map = ultra3_event_map,
.cache_map = &ultra3_cache_map,
.max_events = ARRAY_SIZE(ultra3_perfmon_event_map),
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
.upper_shift = 11,
.lower_shift = 4,
.event_mask = 0x3f,
sparc64: Perf counter 'nop' event is not constant. On Niagara-2, for example, it's going to be different. So make it something specified in sparc_pmu. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:13:26 -06:00
.upper_nop = 0x1c,
.lower_nop = 0x14,
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
};
sparc: Niagara1 perf event support. This chip is extremely limited, and many of the events supported are approximations at best. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 22:23:41 -06:00
/* Niagara1 is very limited. The upper PIC is hard-locked to count
* only instructions, so it is free running which creates all kinds of
sparc64: Fix comment typo in perf_event.c Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 16:10:23 -06:00
* problems. Some hardware designs make one wonder if the creator
sparc: Niagara1 perf event support. This chip is extremely limited, and many of the events supported are approximations at best. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 22:23:41 -06:00
* even looked at how this stuff gets used by software.
*/
static const struct perf_event_map niagara1_perfmon_event_map[] = {
[PERF_COUNT_HW_CPU_CYCLES] = { 0x00, PIC_UPPER },
[PERF_COUNT_HW_INSTRUCTIONS] = { 0x00, PIC_UPPER },
[PERF_COUNT_HW_CACHE_REFERENCES] = { 0, PIC_NONE },
[PERF_COUNT_HW_CACHE_MISSES] = { 0x03, PIC_LOWER },
};
static const struct perf_event_map *niagara1_event_map(int event_id)
{
return &niagara1_perfmon_event_map[event_id];
}
static const cache_map_t niagara1_cache_map = {
[C(L1D)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x03, PIC_LOWER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x03, PIC_LOWER, },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(L1I)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x00, PIC_UPPER },
[C(RESULT_MISS)] = { 0x02, PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_NONSENSE },
[ C(RESULT_MISS) ] = { CACHE_OP_NONSENSE },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(LL)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x07, PIC_LOWER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x07, PIC_LOWER, },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(DTLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x05, PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(ITLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x04, PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(BPU)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
};
static const struct sparc_pmu niagara1_pmu = {
.event_map = niagara1_event_map,
.cache_map = &niagara1_cache_map,
.max_events = ARRAY_SIZE(niagara1_perfmon_event_map),
.upper_shift = 0,
.lower_shift = 4,
.event_mask = 0x7,
.upper_nop = 0x0,
.lower_nop = 0x0,
};
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
static const struct perf_event_map niagara2_perfmon_event_map[] = {
[PERF_COUNT_HW_CPU_CYCLES] = { 0x02ff, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_INSTRUCTIONS] = { 0x02ff, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_CACHE_REFERENCES] = { 0x0208, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_CACHE_MISSES] = { 0x0302, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = { 0x0201, PIC_UPPER | PIC_LOWER },
[PERF_COUNT_HW_BRANCH_MISSES] = { 0x0202, PIC_UPPER | PIC_LOWER },
};
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 04:02:48 -06:00
static const struct perf_event_map *niagara2_event_map(int event_id)
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
{
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 04:02:48 -06:00
return &niagara2_perfmon_event_map[event_id];
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
}
sparc: Add Niagara2 HW cache event support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 22:04:16 -06:00
static const cache_map_t niagara2_cache_map = {
[C(L1D)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x0208, PIC_UPPER | PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0302, PIC_UPPER | PIC_LOWER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { 0x0210, PIC_UPPER | PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0302, PIC_UPPER | PIC_LOWER, },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(L1I)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x02ff, PIC_UPPER | PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0301, PIC_UPPER | PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_NONSENSE },
[ C(RESULT_MISS) ] = { CACHE_OP_NONSENSE },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(LL)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { 0x0208, PIC_UPPER | PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0330, PIC_UPPER | PIC_LOWER, },
},
[C(OP_WRITE)] = {
[C(RESULT_ACCESS)] = { 0x0210, PIC_UPPER | PIC_LOWER, },
[C(RESULT_MISS)] = { 0x0320, PIC_UPPER | PIC_LOWER, },
},
[C(OP_PREFETCH)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
},
[C(DTLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0x0b08, PIC_UPPER | PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(ITLB)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { 0xb04, PIC_UPPER | PIC_LOWER, },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
[C(BPU)] = {
[C(OP_READ)] = {
[C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED },
[C(RESULT_MISS)] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED },
[ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED },
},
},
};
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
static const struct sparc_pmu niagara2_pmu = {
.event_map = niagara2_event_map,
sparc: Add Niagara2 HW cache event support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 22:04:16 -06:00
.cache_map = &niagara2_cache_map,
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
.max_events = ARRAY_SIZE(niagara2_perfmon_event_map),
.upper_shift = 19,
.lower_shift = 6,
.event_mask = 0xfff,
.hv_bit = 0x8,
sparc64: Fix niagara2 perf IRQ bits. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-10-09 01:42:40 -06:00
.irq_bit = 0x30,
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
.upper_nop = 0x220,
.lower_nop = 0x220,
};
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
static const struct sparc_pmu *sparc_pmu __read_mostly;
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 04:02:48 -06:00
static u64 event_encoding(u64 event_id, int idx)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
if (idx == PIC_UPPER_INDEX)
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 04:02:48 -06:00
event_id <<= sparc_pmu->upper_shift;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
else
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 04:02:48 -06:00
event_id <<= sparc_pmu->lower_shift;
return event_id;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
static u64 mask_for_index(int idx)
{
return event_encoding(sparc_pmu->event_mask, idx);
}
static u64 nop_for_index(int idx)
{
return event_encoding(idx == PIC_UPPER_INDEX ?
sparc64: Perf counter 'nop' event is not constant. On Niagara-2, for example, it's going to be different. So make it something specified in sparc_pmu. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:13:26 -06:00
sparc_pmu->upper_nop :
sparc_pmu->lower_nop, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
static inline void sparc_pmu_enable_event(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc, int idx)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
u64 val, mask = mask_for_index(idx);
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
val = cpuc->pcr;
val &= ~mask;
val |= hwc->config;
cpuc->pcr = val;
pcr_ops->write(cpuc->pcr);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
static inline void sparc_pmu_disable_event(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc, int idx)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
u64 mask = mask_for_index(idx);
u64 nop = nop_for_index(idx);
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
u64 val;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
val = cpuc->pcr;
val &= ~mask;
val |= nop;
cpuc->pcr = val;
pcr_ops->write(cpuc->pcr);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
static u32 read_pmc(int idx)
{
u64 val;
read_pic(val);
if (idx == PIC_UPPER_INDEX)
val >>= 32;
return val & 0xffffffff;
}
static void write_pmc(int idx, u64 val)
{
u64 shift, mask, pic;
shift = 0;
if (idx == PIC_UPPER_INDEX)
shift = 32;
mask = ((u64) 0xffffffff) << shift;
val <<= shift;
read_pic(pic);
pic &= ~mask;
pic |= val;
write_pic(pic);
}
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
static u64 sparc_perf_event_update(struct perf_event *event,
struct hw_perf_event *hwc, int idx)
{
int shift = 64 - 32;
u64 prev_raw_count, new_raw_count;
s64 delta;
again:
prev_raw_count = atomic64_read(&hwc->prev_count);
new_raw_count = read_pmc(idx);
if (atomic64_cmpxchg(&hwc->prev_count, prev_raw_count,
new_raw_count) != prev_raw_count)
goto again;
delta = (new_raw_count << shift) - (prev_raw_count << shift);
delta >>= shift;
atomic64_add(delta, &event->count);
atomic64_sub(delta, &hwc->period_left);
return new_raw_count;
}
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 04:02:48 -06:00
static int sparc_perf_event_set_period(struct perf_event *event,
sparc64: Minor coding style fixups in perf code. These got introduced during the counter --> event tree-wide renaming. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:37:12 -06:00
struct hw_perf_event *hwc, int idx)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
s64 left = atomic64_read(&hwc->period_left);
s64 period = hwc->sample_period;
int ret = 0;
if (unlikely(left <= -period)) {
left = period;
atomic64_set(&hwc->period_left, left);
hwc->last_period = period;
ret = 1;
}
if (unlikely(left <= 0)) {
left += period;
atomic64_set(&hwc->period_left, left);
hwc->last_period = period;
ret = 1;
}
if (left > MAX_PERIOD)
left = MAX_PERIOD;
atomic64_set(&hwc->prev_count, (u64)-left);
write_pmc(idx, (u64)(-left) & 0xffffffff);
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 04:02:48 -06:00
perf_event_update_userpage(event);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return ret;
}
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* If performance event entries have been added, move existing
* events around (if necessary) and then assign new entries to
* counters.
*/
static u64 maybe_change_configuration(struct cpu_hw_events *cpuc, u64 pcr)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
int i;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
if (!cpuc->n_added)
goto out;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* Read in the counters which are moving. */
for (i = 0; i < cpuc->n_events; i++) {
struct perf_event *cp = cpuc->event[i];
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
if (cpuc->current_idx[i] != PIC_NO_INDEX &&
cpuc->current_idx[i] != cp->hw.idx) {
sparc_perf_event_update(cp, &cp->hw,
cpuc->current_idx[i]);
cpuc->current_idx[i] = PIC_NO_INDEX;
}
}
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* Assign to counters all unassigned events. */
for (i = 0; i < cpuc->n_events; i++) {
struct perf_event *cp = cpuc->event[i];
struct hw_perf_event *hwc = &cp->hw;
int idx = hwc->idx;
u64 enc;
if (cpuc->current_idx[i] != PIC_NO_INDEX)
continue;
sparc_perf_event_set_period(cp, hwc, idx);
cpuc->current_idx[i] = idx;
enc = perf_event_get_enc(cpuc->events[i]);
pcr |= event_encoding(enc, idx);
}
out:
return pcr;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
void hw_perf_enable(void)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
u64 pcr;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
if (cpuc->enabled)
return;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
cpuc->enabled = 1;
barrier();
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
pcr = cpuc->pcr;
if (!cpuc->n_events) {
pcr = 0;
} else {
pcr = maybe_change_configuration(cpuc, pcr);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* We require that all of the events have the same
* configuration, so just fetch the settings from the
* first entry.
*/
cpuc->pcr = pcr | cpuc->event[0]->hw.config_base;
}
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
pcr_ops->write(cpuc->pcr);
}
void hw_perf_disable(void)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
u64 val;
if (!cpuc->enabled)
return;
cpuc->enabled = 0;
cpuc->n_added = 0;
val = cpuc->pcr;
val &= ~(PCR_UTRACE | PCR_STRACE |
sparc_pmu->hv_bit | sparc_pmu->irq_bit);
cpuc->pcr = val;
pcr_ops->write(cpuc->pcr);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
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 04:02:48 -06:00
static void sparc_pmu_disable(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
struct hw_perf_event *hwc = &event->hw;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
unsigned long flags;
int i;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
local_irq_save(flags);
perf_disable();
for (i = 0; i < cpuc->n_events; i++) {
if (event == cpuc->event[i]) {
int idx = cpuc->current_idx[i];
/* Shift remaining entries down into
* the existing slot.
*/
while (++i < cpuc->n_events) {
cpuc->event[i - 1] = cpuc->event[i];
cpuc->events[i - 1] = cpuc->events[i];
cpuc->current_idx[i - 1] =
cpuc->current_idx[i];
}
/* Absorb the final count and turn off the
* event.
*/
sparc_pmu_disable_event(cpuc, hwc, idx);
barrier();
sparc_perf_event_update(event, hwc, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
perf_event_update_userpage(event);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
cpuc->n_events--;
break;
}
}
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
perf_enable();
local_irq_restore(flags);
}
static int active_event_index(struct cpu_hw_events *cpuc,
struct perf_event *event)
{
int i;
for (i = 0; i < cpuc->n_events; i++) {
if (cpuc->event[i] == event)
break;
}
BUG_ON(i == cpuc->n_events);
return cpuc->current_idx[i];
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
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 04:02:48 -06:00
static void sparc_pmu_read(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
int idx = active_event_index(cpuc, event);
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 04:02:48 -06:00
struct hw_perf_event *hwc = &event->hw;
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
sparc_perf_event_update(event, hwc, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
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 04:02:48 -06:00
static void sparc_pmu_unthrottle(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
int idx = active_event_index(cpuc, event);
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 04:02:48 -06:00
struct hw_perf_event *hwc = &event->hw;
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
sparc_pmu_enable_event(cpuc, hwc, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
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 04:02:48 -06:00
static atomic_t active_events = ATOMIC_INIT(0);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
static DEFINE_MUTEX(pmc_grab_mutex);
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
static void perf_stop_nmi_watchdog(void *unused)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
stop_nmi_watchdog(NULL);
cpuc->pcr = pcr_ops->read();
}
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 04:02:48 -06:00
void perf_event_grab_pmc(void)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
if (atomic_inc_not_zero(&active_events))
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return;
mutex_lock(&pmc_grab_mutex);
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 04:02:48 -06:00
if (atomic_read(&active_events) == 0) {
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (atomic_read(&nmi_active) > 0) {
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
on_each_cpu(perf_stop_nmi_watchdog, NULL, 1);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
BUG_ON(atomic_read(&nmi_active) != 0);
}
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 04:02:48 -06:00
atomic_inc(&active_events);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
mutex_unlock(&pmc_grab_mutex);
}
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 04:02:48 -06:00
void perf_event_release_pmc(void)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
if (atomic_dec_and_mutex_lock(&active_events, &pmc_grab_mutex)) {
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (atomic_read(&nmi_active) == 0)
on_each_cpu(start_nmi_watchdog, NULL, 1);
mutex_unlock(&pmc_grab_mutex);
}
}
sparc: Support HW cache events. First supported chip for HW cache events is Ultra-IIIi. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:42:10 -06:00
static const struct perf_event_map *sparc_map_cache_event(u64 config)
{
unsigned int cache_type, cache_op, cache_result;
const struct perf_event_map *pmap;
if (!sparc_pmu->cache_map)
return ERR_PTR(-ENOENT);
cache_type = (config >> 0) & 0xff;
if (cache_type >= PERF_COUNT_HW_CACHE_MAX)
return ERR_PTR(-EINVAL);
cache_op = (config >> 8) & 0xff;
if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX)
return ERR_PTR(-EINVAL);
cache_result = (config >> 16) & 0xff;
if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
return ERR_PTR(-EINVAL);
pmap = &((*sparc_pmu->cache_map)[cache_type][cache_op][cache_result]);
if (pmap->encoding == CACHE_OP_UNSUPPORTED)
return ERR_PTR(-ENOENT);
if (pmap->encoding == CACHE_OP_NONSENSE)
return ERR_PTR(-EINVAL);
return pmap;
}
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 04:02:48 -06:00
static void hw_perf_event_destroy(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
perf_event_release_pmc();
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
/* Make sure all events can be scheduled into the hardware at
* the same time. This is simplified by the fact that we only
* need to support 2 simultaneous HW events.
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
*
* As a side effect, the evts[]->hw.idx values will be assigned
* on success. These are pending indexes. When the events are
* actually programmed into the chip, these values will propagate
* to the per-cpu cpuc->current_idx[] slots, see the code in
* maybe_change_configuration() for details.
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
*/
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
static int sparc_check_constraints(struct perf_event **evts,
unsigned long *events, int n_ev)
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
{
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
u8 msk0 = 0, msk1 = 0;
int idx0 = 0;
/* This case is possible when we are invoked from
* hw_perf_group_sched_in().
*/
if (!n_ev)
return 0;
if (n_ev > perf_max_events)
return -1;
msk0 = perf_event_get_msk(events[0]);
if (n_ev == 1) {
if (msk0 & PIC_LOWER)
idx0 = 1;
goto success;
}
BUG_ON(n_ev != 2);
msk1 = perf_event_get_msk(events[1]);
/* If both events can go on any counter, OK. */
if (msk0 == (PIC_UPPER | PIC_LOWER) &&
msk1 == (PIC_UPPER | PIC_LOWER))
goto success;
/* If one event is limited to a specific counter,
* and the other can go on both, OK.
*/
if ((msk0 == PIC_UPPER || msk0 == PIC_LOWER) &&
msk1 == (PIC_UPPER | PIC_LOWER)) {
if (msk0 & PIC_LOWER)
idx0 = 1;
goto success;
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
}
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
if ((msk1 == PIC_UPPER || msk1 == PIC_LOWER) &&
msk0 == (PIC_UPPER | PIC_LOWER)) {
if (msk1 & PIC_UPPER)
idx0 = 1;
goto success;
}
/* If the events are fixed to different counters, OK. */
if ((msk0 == PIC_UPPER && msk1 == PIC_LOWER) ||
(msk0 == PIC_LOWER && msk1 == PIC_UPPER)) {
if (msk0 & PIC_LOWER)
idx0 = 1;
goto success;
}
/* Otherwise, there is a conflict. */
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
return -1;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
success:
evts[0]->hw.idx = idx0;
if (n_ev == 2)
evts[1]->hw.idx = idx0 ^ 1;
return 0;
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
}
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
static int check_excludes(struct perf_event **evts, int n_prev, int n_new)
{
int eu = 0, ek = 0, eh = 0;
struct perf_event *event;
int i, n, first;
n = n_prev + n_new;
if (n <= 1)
return 0;
first = 1;
for (i = 0; i < n; i++) {
event = evts[i];
if (first) {
eu = event->attr.exclude_user;
ek = event->attr.exclude_kernel;
eh = event->attr.exclude_hv;
first = 0;
} else if (event->attr.exclude_user != eu ||
event->attr.exclude_kernel != ek ||
event->attr.exclude_hv != eh) {
return -EAGAIN;
}
}
return 0;
}
static int collect_events(struct perf_event *group, int max_count,
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
struct perf_event *evts[], unsigned long *events,
int *current_idx)
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
{
struct perf_event *event;
int n = 0;
if (!is_software_event(group)) {
if (n >= max_count)
return -1;
evts[n] = group;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
events[n] = group->hw.event_base;
current_idx[n++] = PIC_NO_INDEX;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
}
list_for_each_entry(event, &group->sibling_list, group_entry) {
if (!is_software_event(event) &&
event->state != PERF_EVENT_STATE_OFF) {
if (n >= max_count)
return -1;
evts[n] = event;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
events[n] = event->hw.event_base;
current_idx[n++] = PIC_NO_INDEX;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
}
}
return n;
}
perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Miller <davem@davemloft.net> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-02-11 05:21:58 -07:00
static void event_sched_in(struct perf_event *event)
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
{
event->state = PERF_EVENT_STATE_ACTIVE;
perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Miller <davem@davemloft.net> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-02-11 05:21:58 -07:00
event->oncpu = smp_processor_id();
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
event->tstamp_running += event->ctx->time - event->tstamp_stopped;
if (is_software_event(event))
event->pmu->enable(event);
}
int hw_perf_group_sched_in(struct perf_event *group_leader,
struct perf_cpu_context *cpuctx,
perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Miller <davem@davemloft.net> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-02-11 05:21:58 -07:00
struct perf_event_context *ctx)
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
struct perf_event *sub;
int n0, n;
if (!sparc_pmu)
return 0;
n0 = cpuc->n_events;
n = collect_events(group_leader, perf_max_events - n0,
&cpuc->event[n0], &cpuc->events[n0],
&cpuc->current_idx[n0]);
if (n < 0)
return -EAGAIN;
if (check_excludes(cpuc->event, n0, n))
return -EINVAL;
if (sparc_check_constraints(cpuc->event, cpuc->events, n + n0))
return -EAGAIN;
cpuc->n_events = n0 + n;
cpuc->n_added += n;
cpuctx->active_oncpu += n;
n = 1;
perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Miller <davem@davemloft.net> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-02-11 05:21:58 -07:00
event_sched_in(group_leader);
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
list_for_each_entry(sub, &group_leader->sibling_list, group_entry) {
if (sub->state != PERF_EVENT_STATE_OFF) {
perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Miller <davem@davemloft.net> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-02-11 05:21:58 -07:00
event_sched_in(sub);
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
n++;
}
}
ctx->nr_active += n;
return 1;
}
static int sparc_pmu_enable(struct perf_event *event)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
int n0, ret = -EAGAIN;
unsigned long flags;
local_irq_save(flags);
perf_disable();
n0 = cpuc->n_events;
if (n0 >= perf_max_events)
goto out;
cpuc->event[n0] = event;
cpuc->events[n0] = event->hw.event_base;
cpuc->current_idx[n0] = PIC_NO_INDEX;
if (check_excludes(cpuc->event, n0, 1))
goto out;
if (sparc_check_constraints(cpuc->event, cpuc->events, n0 + 1))
goto out;
cpuc->n_events++;
cpuc->n_added++;
ret = 0;
out:
perf_enable();
local_irq_restore(flags);
return ret;
}
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 04:02:48 -06:00
static int __hw_perf_event_init(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
struct perf_event_attr *attr = &event->attr;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
struct perf_event *evts[MAX_HWEVENTS];
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 04:02:48 -06:00
struct hw_perf_event *hwc = &event->hw;
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
unsigned long events[MAX_HWEVENTS];
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
int current_idx_dmy[MAX_HWEVENTS];
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
const struct perf_event_map *pmap;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
int n;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (atomic_read(&nmi_active) < 0)
return -ENODEV;
sparc: Support HW cache events. First supported chip for HW cache events is Ultra-IIIi. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:42:10 -06:00
if (attr->type == PERF_TYPE_HARDWARE) {
if (attr->config >= sparc_pmu->max_events)
return -EINVAL;
pmap = sparc_pmu->event_map(attr->config);
} else if (attr->type == PERF_TYPE_HW_CACHE) {
pmap = sparc_map_cache_event(attr->config);
if (IS_ERR(pmap))
return PTR_ERR(pmap);
} else
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return -EOPNOTSUPP;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* We save the enable bits in the config_base. */
sparc64: Provide a way to specify a perf counter overflow IRQ enable bit. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:10:59 -06:00
hwc->config_base = sparc_pmu->irq_bit;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (!attr->exclude_user)
hwc->config_base |= PCR_UTRACE;
if (!attr->exclude_kernel)
hwc->config_base |= PCR_STRACE;
sparc64: Provide hypervisor tracing bit support for perf counters. A PMU need only specify which bit in the PCR enabled hypervisor tracing in order to enable this. This will be used in Niagara-2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:09:06 -06:00
if (!attr->exclude_hv)
hwc->config_base |= sparc_pmu->hv_bit;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
hwc->event_base = perf_event_encode(pmap);
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
n = 0;
if (event->group_leader != event) {
n = collect_events(event->group_leader,
perf_max_events - 1,
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
evts, events, current_idx_dmy);
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
if (n < 0)
return -EINVAL;
}
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
events[n] = hwc->event_base;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
evts[n] = event;
if (check_excludes(evts, n, 1))
return -EINVAL;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
if (sparc_check_constraints(evts, events, n + 1))
sparc64: Add a basic conflict engine in preparation for multi-counter support. The hardware counter ->event_base state records and encoding of the "struct perf_event_map" entry used for the event. We use this to make sure that when we have more than 1 event, both can be scheduled into the hardware at the same time. As usual, structure of code is largely cribbed from powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:35:20 -06:00
return -EINVAL;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
hwc->idx = PIC_NO_INDEX;
sparc64: Add initial perf event conflict resolution and checks. Cribbed from powerpc code, as usual. :-) Currently it is only used to validate that all counters have the same user/kernel/hv attributes. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-27 21:43:07 -06:00
/* Try to do all error checking before this point, as unwinding
* state after grabbing the PMC is difficult.
*/
perf_event_grab_pmc();
event->destroy = hw_perf_event_destroy;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (!hwc->sample_period) {
hwc->sample_period = MAX_PERIOD;
hwc->last_period = hwc->sample_period;
atomic64_set(&hwc->period_left, hwc->sample_period);
}
return 0;
}
static const struct pmu pmu = {
.enable = sparc_pmu_enable,
.disable = sparc_pmu_disable,
.read = sparc_pmu_read,
.unthrottle = sparc_pmu_unthrottle,
};
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 04:02:48 -06:00
const struct pmu *hw_perf_event_init(struct perf_event *event)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
int err = __hw_perf_event_init(event);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (err)
return ERR_PTR(err);
return &pmu;
}
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 04:02:48 -06:00
void perf_event_print_debug(void)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
unsigned long flags;
u64 pcr, pic;
int cpu;
if (!sparc_pmu)
return;
local_irq_save(flags);
cpu = smp_processor_id();
pcr = pcr_ops->read();
read_pic(pic);
pr_info("\n");
pr_info("CPU#%d: PCR[%016llx] PIC[%016llx]\n",
cpu, pcr, pic);
local_irq_restore(flags);
}
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 04:02:48 -06:00
static int __kprobes perf_event_nmi_handler(struct notifier_block *self,
sparc64: Minor coding style fixups in perf code. These got introduced during the counter --> event tree-wide renaming. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-28 18:37:12 -06:00
unsigned long cmd, void *__args)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
struct die_args *args = __args;
struct perf_sample_data data;
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 04:02:48 -06:00
struct cpu_hw_events *cpuc;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
struct pt_regs *regs;
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
int i;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
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 04:02:48 -06:00
if (!atomic_read(&active_events))
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return NOTIFY_DONE;
switch (cmd) {
case DIE_NMI:
break;
default:
return NOTIFY_DONE;
}
regs = args->regs;
data.addr = 0;
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 04:02:48 -06:00
cpuc = &__get_cpu_var(cpu_hw_events);
sparc64: Fix Niagara2 perf event handling. For chips like Niagara2 that have true overflow indications in the %pcr (which we don't actually need and don't use) the interrupt signal persists until the overflow bits are cleared by an explicit %pcr write. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-05 00:16:03 -07:00
/* If the PMU has the TOE IRQ enable bits, we need to do a
* dummy write to the %pcr to clear the overflow bits and thus
* the interrupt.
*
* Do this before we peek at the counters to determine
* overflow so we don't lose any events.
*/
if (sparc_pmu->irq_bit)
pcr_ops->write(cpuc->pcr);
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
for (i = 0; i < cpuc->n_events; i++) {
struct perf_event *event = cpuc->event[i];
int idx = cpuc->current_idx[i];
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 04:02:48 -06:00
struct hw_perf_event *hwc;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
u64 val;
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 04:02:48 -06:00
hwc = &event->hw;
val = sparc_perf_event_update(event, hwc, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (val & (1ULL << 31))
continue;
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 04:02:48 -06:00
data.period = event->hw.last_period;
if (!sparc_perf_event_set_period(event, hwc, idx))
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
continue;
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 04:02:48 -06:00
if (perf_event_overflow(event, 1, &data, regs))
sparc64: Cache per-cpu %pcr register value in perf code. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-29 22:27:06 -06:00
sparc_pmu_disable_event(cpuc, hwc, idx);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
return NOTIFY_STOP;
}
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 04:02:48 -06:00
static __read_mostly struct notifier_block perf_event_nmi_notifier = {
.notifier_call = perf_event_nmi_handler,
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
};
static bool __init supported_pmu(void)
{
sparc: Support all ultra3 and ultra4 derivatives. For the generic events we support, all of these chips have the same encodings as ultra3i. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 21:54:22 -06:00
if (!strcmp(sparc_pmu_type, "ultra3") ||
!strcmp(sparc_pmu_type, "ultra3+") ||
!strcmp(sparc_pmu_type, "ultra3i") ||
!strcmp(sparc_pmu_type, "ultra4+")) {
sparc_pmu = &ultra3_pmu;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return true;
}
sparc: Niagara1 perf event support. This chip is extremely limited, and many of the events supported are approximations at best. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-26 22:23:41 -06:00
if (!strcmp(sparc_pmu_type, "niagara")) {
sparc_pmu = &niagara1_pmu;
return true;
}
sparc64: Initial niagara2 perf counter support. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 08:22:18 -06:00
if (!strcmp(sparc_pmu_type, "niagara2")) {
sparc_pmu = &niagara2_pmu;
return true;
}
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
return false;
}
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 04:02:48 -06:00
void __init init_hw_perf_events(void)
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
{
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 04:02:48 -06:00
pr_info("Performance events: ");
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
if (!supported_pmu()) {
pr_cont("No support for PMU type '%s'\n", sparc_pmu_type);
return;
}
pr_cont("Supported PMU type is '%s'\n", sparc_pmu_type);
sparc64: Fully support both performance counters. Add the rest of the conflict detection and resolution logic necessary to support more than one counter at a time on sparc64. The structure and implementation closely mimicks that of powerpc. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-01-20 03:59:47 -07:00
/* All sparc64 PMUs currently have 2 events. */
perf_max_events = 2;
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
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 04:02:48 -06:00
register_die_notifier(&perf_event_nmi_notifier);
sparc64: Initial hw perf counter support. Only supports one simple counter and only UltraSPARC-IIIi chips. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-10 07:28:20 -06:00
}
sparc64: Add perf callchain support. Pretty straightforward, and it should be easy to add accurate walk through of signal stack frames in userspace. Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Jens Axboe <jens.axboe@oracle.com>
2010-01-19 01:26:13 -07:00
static inline void callchain_store(struct perf_callchain_entry *entry, u64 ip)
{
if (entry->nr < PERF_MAX_STACK_DEPTH)
entry->ip[entry->nr++] = ip;
}
static void perf_callchain_kernel(struct pt_regs *regs,
struct perf_callchain_entry *entry)
{
unsigned long ksp, fp;
callchain_store(entry, PERF_CONTEXT_KERNEL);
callchain_store(entry, regs->tpc);
ksp = regs->u_regs[UREG_I6];
fp = ksp + STACK_BIAS;
do {
struct sparc_stackf *sf;
struct pt_regs *regs;
unsigned long pc;
if (!kstack_valid(current_thread_info(), fp))
break;
sf = (struct sparc_stackf *) fp;
regs = (struct pt_regs *) (sf + 1);
if (kstack_is_trap_frame(current_thread_info(), regs)) {
if (user_mode(regs))
break;
pc = regs->tpc;
fp = regs->u_regs[UREG_I6] + STACK_BIAS;
} else {
pc = sf->callers_pc;
fp = (unsigned long)sf->fp + STACK_BIAS;
}
callchain_store(entry, pc);
} while (entry->nr < PERF_MAX_STACK_DEPTH);
}
static void perf_callchain_user_64(struct pt_regs *regs,
struct perf_callchain_entry *entry)
{
unsigned long ufp;
callchain_store(entry, PERF_CONTEXT_USER);
callchain_store(entry, regs->tpc);
ufp = regs->u_regs[UREG_I6] + STACK_BIAS;
do {
struct sparc_stackf *usf, sf;
unsigned long pc;
usf = (struct sparc_stackf *) ufp;
if (__copy_from_user_inatomic(&sf, usf, sizeof(sf)))
break;
pc = sf.callers_pc;
ufp = (unsigned long)sf.fp + STACK_BIAS;
callchain_store(entry, pc);
} while (entry->nr < PERF_MAX_STACK_DEPTH);
}
static void perf_callchain_user_32(struct pt_regs *regs,
struct perf_callchain_entry *entry)
{
unsigned long ufp;
callchain_store(entry, PERF_CONTEXT_USER);
callchain_store(entry, regs->tpc);
ufp = regs->u_regs[UREG_I6];
do {
struct sparc_stackf32 *usf, sf;
unsigned long pc;
usf = (struct sparc_stackf32 *) ufp;
if (__copy_from_user_inatomic(&sf, usf, sizeof(sf)))
break;
pc = sf.callers_pc;
ufp = (unsigned long)sf.fp;
callchain_store(entry, pc);
} while (entry->nr < PERF_MAX_STACK_DEPTH);
}
/* Like powerpc we can't get PMU interrupts within the PMU handler,
* so no need for seperate NMI and IRQ chains as on x86.
*/
static DEFINE_PER_CPU(struct perf_callchain_entry, callchain);
struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
{
struct perf_callchain_entry *entry = &__get_cpu_var(callchain);
entry->nr = 0;
if (!user_mode(regs)) {
stack_trace_flush();
perf_callchain_kernel(regs, entry);
if (current->mm)
regs = task_pt_regs(current);
else
regs = NULL;
}
if (regs) {
flushw_user();
if (test_thread_flag(TIF_32BIT))
perf_callchain_user_32(regs, entry);
else
perf_callchain_user_64(regs, entry);
}
return entry;
}