linux_dsm_epyc7002/arch/powerpc/kernel/mpc7450-pmu.c

418 lines
10 KiB
C
Raw Normal View History

perf_counter: powerpc: Add processor back-end for MPC7450 family This adds support for the performance monitor hardware on the MPC7450 family of processors (7450, 7451, 7455, 7447/7457, 7447A, 7448), used in the later Apple G4 powermacs/powerbooks and other machines. These machines have 6 hardware counters with a unique set of events which can be counted on each counter, with some events being available on multiple counters. Raw event codes for these processors are (PMC << 8) + PMCSEL. If PMC is non-zero then the event is that selected by the given PMCSEL value for that PMC (hardware counter). If PMC is zero then the event selected is one of the low-numbered ones that are common to several PMCs. In this case PMCSEL must be <= 22 and the event is what that PMCSEL value would select on PMC1 (but it may be placed any other PMC that has the same event for that PMCSEL value). For events that count cycles or occurrences that exceed a threshold, the threshold requested can be specified in the 0x3f000 bits of the raw event codes. If the event uses the threshold multiplier bit and that bit should be set, that is indicated with the 0x40000 bit of the raw event code. This fills in some of the generic cache events. Unfortunately there are quite a few blank spaces in the table, partly because these processors tend to count cache hits rather than cache accesses. Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: linuxppc-dev@ozlabs.org Cc: benh@kernel.crashing.org LKML-Reference: <19000.55631.802122.696927@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-17 18:53:51 +07:00
/*
* Performance counter support for MPC7450-family processors.
*
* Copyright 2008-2009 Paul Mackerras, IBM Corporation.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
#include <linux/string.h>
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 17:02:48 +07:00
#include <linux/perf_event.h>
perf_counter: powerpc: Add processor back-end for MPC7450 family This adds support for the performance monitor hardware on the MPC7450 family of processors (7450, 7451, 7455, 7447/7457, 7447A, 7448), used in the later Apple G4 powermacs/powerbooks and other machines. These machines have 6 hardware counters with a unique set of events which can be counted on each counter, with some events being available on multiple counters. Raw event codes for these processors are (PMC << 8) + PMCSEL. If PMC is non-zero then the event is that selected by the given PMCSEL value for that PMC (hardware counter). If PMC is zero then the event selected is one of the low-numbered ones that are common to several PMCs. In this case PMCSEL must be <= 22 and the event is what that PMCSEL value would select on PMC1 (but it may be placed any other PMC that has the same event for that PMCSEL value). For events that count cycles or occurrences that exceed a threshold, the threshold requested can be specified in the 0x3f000 bits of the raw event codes. If the event uses the threshold multiplier bit and that bit should be set, that is indicated with the 0x40000 bit of the raw event code. This fills in some of the generic cache events. Unfortunately there are quite a few blank spaces in the table, partly because these processors tend to count cache hits rather than cache accesses. Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: linuxppc-dev@ozlabs.org Cc: benh@kernel.crashing.org LKML-Reference: <19000.55631.802122.696927@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-17 18:53:51 +07:00
#include <asm/reg.h>
#include <asm/cputable.h>
#define N_COUNTER 6 /* Number of hardware counters */
#define MAX_ALT 3 /* Maximum number of event alternative codes */
/*
* Bits in event code for MPC7450 family
*/
#define PM_THRMULT_MSKS 0x40000
#define PM_THRESH_SH 12
#define PM_THRESH_MSK 0x3f
#define PM_PMC_SH 8
#define PM_PMC_MSK 7
#define PM_PMCSEL_MSK 0x7f
/*
* Classify events according to how specific their PMC requirements are.
* Result is:
* 0: can go on any PMC
* 1: can go on PMCs 1-4
* 2: can go on PMCs 1,2,4
* 3: can go on PMCs 1 or 2
* 4: can only go on one PMC
* -1: event code is invalid
*/
#define N_CLASSES 5
static int mpc7450_classify_event(u32 event)
{
int pmc;
pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
if (pmc) {
if (pmc > N_COUNTER)
return -1;
return 4;
}
event &= PM_PMCSEL_MSK;
if (event <= 1)
return 0;
if (event <= 7)
return 1;
if (event <= 13)
return 2;
if (event <= 22)
return 3;
return -1;
}
/*
* Events using threshold and possible threshold scale:
* code scale? name
* 11e N PM_INSTQ_EXCEED_CYC
* 11f N PM_ALTV_IQ_EXCEED_CYC
* 128 Y PM_DTLB_SEARCH_EXCEED_CYC
* 12b Y PM_LD_MISS_EXCEED_L1_CYC
* 220 N PM_CQ_EXCEED_CYC
* 30c N PM_GPR_RB_EXCEED_CYC
* 30d ? PM_FPR_IQ_EXCEED_CYC ?
* 311 Y PM_ITLB_SEARCH_EXCEED
* 410 N PM_GPR_IQ_EXCEED_CYC
*/
/*
* Return use of threshold and threshold scale bits:
* 0 = uses neither, 1 = uses threshold, 2 = uses both
*/
static int mpc7450_threshold_use(u32 event)
{
int pmc, sel;
pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
sel = event & PM_PMCSEL_MSK;
switch (pmc) {
case 1:
if (sel == 0x1e || sel == 0x1f)
return 1;
if (sel == 0x28 || sel == 0x2b)
return 2;
break;
case 2:
if (sel == 0x20)
return 1;
break;
case 3:
if (sel == 0xc || sel == 0xd)
return 1;
if (sel == 0x11)
return 2;
break;
case 4:
if (sel == 0x10)
return 1;
break;
}
return 0;
}
/*
* Layout of constraint bits:
* 33222222222211111111110000000000
* 10987654321098765432109876543210
* |< >< > < > < ><><><><><><>
* TS TV G4 G3 G2P6P5P4P3P2P1
*
* P1 - P6
* 0 - 11: Count of events needing PMC1 .. PMC6
*
* G2
* 12 - 14: Count of events needing PMC1 or PMC2
*
* G3
* 16 - 18: Count of events needing PMC1, PMC2 or PMC4
*
* G4
* 20 - 23: Count of events needing PMC1, PMC2, PMC3 or PMC4
*
* TV
* 24 - 29: Threshold value requested
*
* TS
* 30: Threshold scale value requested
*/
static u32 pmcbits[N_COUNTER][2] = {
{ 0x00844002, 0x00111001 }, /* PMC1 mask, value: P1,G2,G3,G4 */
{ 0x00844008, 0x00111004 }, /* PMC2: P2,G2,G3,G4 */
{ 0x00800020, 0x00100010 }, /* PMC3: P3,G4 */
{ 0x00840080, 0x00110040 }, /* PMC4: P4,G3,G4 */
{ 0x00000200, 0x00000100 }, /* PMC5: P5 */
{ 0x00000800, 0x00000400 } /* PMC6: P6 */
};
static u32 classbits[N_CLASSES - 1][2] = {
{ 0x00000000, 0x00000000 }, /* class 0: no constraint */
{ 0x00800000, 0x00100000 }, /* class 1: G4 */
{ 0x00040000, 0x00010000 }, /* class 2: G3 */
{ 0x00004000, 0x00001000 }, /* class 3: G2 */
};
static int mpc7450_get_constraint(u64 event, unsigned long *maskp,
unsigned long *valp)
{
int pmc, class;
u32 mask, value;
int thresh, tuse;
class = mpc7450_classify_event(event);
if (class < 0)
return -1;
if (class == 4) {
pmc = ((unsigned int)event >> PM_PMC_SH) & PM_PMC_MSK;
mask = pmcbits[pmc - 1][0];
value = pmcbits[pmc - 1][1];
} else {
mask = classbits[class][0];
value = classbits[class][1];
}
tuse = mpc7450_threshold_use(event);
if (tuse) {
thresh = ((unsigned int)event >> PM_THRESH_SH) & PM_THRESH_MSK;
mask |= 0x3f << 24;
value |= thresh << 24;
if (tuse == 2) {
mask |= 0x40000000;
if ((unsigned int)event & PM_THRMULT_MSKS)
value |= 0x40000000;
}
}
*maskp = mask;
*valp = value;
return 0;
}
static const unsigned int event_alternatives[][MAX_ALT] = {
{ 0x217, 0x317 }, /* PM_L1_DCACHE_MISS */
{ 0x418, 0x50f, 0x60f }, /* PM_SNOOP_RETRY */
{ 0x502, 0x602 }, /* PM_L2_HIT */
{ 0x503, 0x603 }, /* PM_L3_HIT */
{ 0x504, 0x604 }, /* PM_L2_ICACHE_MISS */
{ 0x505, 0x605 }, /* PM_L3_ICACHE_MISS */
{ 0x506, 0x606 }, /* PM_L2_DCACHE_MISS */
{ 0x507, 0x607 }, /* PM_L3_DCACHE_MISS */
{ 0x50a, 0x623 }, /* PM_LD_HIT_L3 */
{ 0x50b, 0x624 }, /* PM_ST_HIT_L3 */
{ 0x50d, 0x60d }, /* PM_L2_TOUCH_HIT */
{ 0x50e, 0x60e }, /* PM_L3_TOUCH_HIT */
{ 0x512, 0x612 }, /* PM_INT_LOCAL */
{ 0x513, 0x61d }, /* PM_L2_MISS */
{ 0x514, 0x61e }, /* PM_L3_MISS */
};
/*
* Scan the alternatives table for a match and return the
* index into the alternatives table if found, else -1.
*/
static int find_alternative(u32 event)
{
int i, j;
for (i = 0; i < ARRAY_SIZE(event_alternatives); ++i) {
if (event < event_alternatives[i][0])
break;
for (j = 0; j < MAX_ALT && event_alternatives[i][j]; ++j)
if (event == event_alternatives[i][j])
return i;
}
return -1;
}
static int mpc7450_get_alternatives(u64 event, unsigned int flags, u64 alt[])
{
int i, j, nalt = 1;
u32 ae;
alt[0] = event;
nalt = 1;
i = find_alternative((u32)event);
if (i >= 0) {
for (j = 0; j < MAX_ALT; ++j) {
ae = event_alternatives[i][j];
if (ae && ae != (u32)event)
alt[nalt++] = ae;
}
}
return nalt;
}
/*
* Bitmaps of which PMCs each class can use for classes 0 - 3.
* Bit i is set if PMC i+1 is usable.
*/
static const u8 classmap[N_CLASSES] = {
0x3f, 0x0f, 0x0b, 0x03, 0
};
/* Bit position and width of each PMCSEL field */
static const int pmcsel_shift[N_COUNTER] = {
6, 0, 27, 22, 17, 11
};
static const u32 pmcsel_mask[N_COUNTER] = {
0x7f, 0x3f, 0x1f, 0x1f, 0x1f, 0x3f
};
/*
* Compute MMCR0/1/2 values for a set of events.
*/
static int mpc7450_compute_mmcr(u64 event[], int n_ev,
unsigned int hwc[], unsigned long mmcr[])
{
u8 event_index[N_CLASSES][N_COUNTER];
int n_classevent[N_CLASSES];
int i, j, class, tuse;
u32 pmc_inuse = 0, pmc_avail;
u32 mmcr0 = 0, mmcr1 = 0, mmcr2 = 0;
u32 ev, pmc, thresh;
if (n_ev > N_COUNTER)
return -1;
/* First pass: count usage in each class */
for (i = 0; i < N_CLASSES; ++i)
n_classevent[i] = 0;
for (i = 0; i < n_ev; ++i) {
class = mpc7450_classify_event(event[i]);
if (class < 0)
return -1;
j = n_classevent[class]++;
event_index[class][j] = i;
}
/* Second pass: allocate PMCs from most specific event to least */
for (class = N_CLASSES - 1; class >= 0; --class) {
for (i = 0; i < n_classevent[class]; ++i) {
ev = event[event_index[class][i]];
if (class == 4) {
pmc = (ev >> PM_PMC_SH) & PM_PMC_MSK;
if (pmc_inuse & (1 << (pmc - 1)))
return -1;
} else {
/* Find a suitable PMC */
pmc_avail = classmap[class] & ~pmc_inuse;
if (!pmc_avail)
return -1;
pmc = ffs(pmc_avail);
}
pmc_inuse |= 1 << (pmc - 1);
tuse = mpc7450_threshold_use(ev);
if (tuse) {
thresh = (ev >> PM_THRESH_SH) & PM_THRESH_MSK;
mmcr0 |= thresh << 16;
if (tuse == 2 && (ev & PM_THRMULT_MSKS))
mmcr2 = 0x80000000;
}
ev &= pmcsel_mask[pmc - 1];
ev <<= pmcsel_shift[pmc - 1];
if (pmc <= 2)
mmcr0 |= ev;
else
mmcr1 |= ev;
hwc[event_index[class][i]] = pmc - 1;
}
}
if (pmc_inuse & 1)
mmcr0 |= MMCR0_PMC1CE;
if (pmc_inuse & 0x3e)
mmcr0 |= MMCR0_PMCnCE;
/* Return MMCRx values */
mmcr[0] = mmcr0;
mmcr[1] = mmcr1;
mmcr[2] = mmcr2;
return 0;
}
/*
* Disable counting by a PMC.
* Note that the pmc argument is 0-based here, not 1-based.
*/
static void mpc7450_disable_pmc(unsigned int pmc, unsigned long mmcr[])
{
if (pmc <= 1)
mmcr[0] &= ~(pmcsel_mask[pmc] << pmcsel_shift[pmc]);
else
mmcr[1] &= ~(pmcsel_mask[pmc] << pmcsel_shift[pmc]);
}
static int mpc7450_generic_events[] = {
[PERF_COUNT_HW_CPU_CYCLES] = 1,
[PERF_COUNT_HW_INSTRUCTIONS] = 2,
[PERF_COUNT_HW_CACHE_MISSES] = 0x217, /* PM_L1_DCACHE_MISS */
[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x122, /* PM_BR_CMPL */
[PERF_COUNT_HW_BRANCH_MISSES] = 0x41c, /* PM_BR_MPRED */
};
#define C(x) PERF_COUNT_HW_CACHE_##x
/*
* Table of generalized cache-related events.
* 0 means not supported, -1 means nonsensical, other values
* are event codes.
*/
static int mpc7450_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = {
[C(L1D)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x225 },
[C(OP_WRITE)] = { 0, 0x227 },
[C(OP_PREFETCH)] = { 0, 0 },
},
[C(L1I)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0x129, 0x115 },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { 0x634, 0 },
},
[C(LL)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0 },
[C(OP_WRITE)] = { 0, 0 },
[C(OP_PREFETCH)] = { 0, 0 },
},
[C(DTLB)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x312 },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
[C(ITLB)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x223 },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
[C(BPU)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0x122, 0x41c },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
};
struct power_pmu mpc7450_pmu = {
.name = "MPC7450 family",
.n_counter = N_COUNTER,
.max_alternatives = MAX_ALT,
.add_fields = 0x00111555ul,
.test_adder = 0x00301000ul,
.compute_mmcr = mpc7450_compute_mmcr,
.get_constraint = mpc7450_get_constraint,
.get_alternatives = mpc7450_get_alternatives,
.disable_pmc = mpc7450_disable_pmc,
.n_generic = ARRAY_SIZE(mpc7450_generic_events),
.generic_events = mpc7450_generic_events,
.cache_events = &mpc7450_cache_events,
};
static int init_mpc7450_pmu(void)
{
if (!cur_cpu_spec->oprofile_cpu_type ||
strcmp(cur_cpu_spec->oprofile_cpu_type, "ppc/7450"))
perf_counter: powerpc: Add processor back-end for MPC7450 family This adds support for the performance monitor hardware on the MPC7450 family of processors (7450, 7451, 7455, 7447/7457, 7447A, 7448), used in the later Apple G4 powermacs/powerbooks and other machines. These machines have 6 hardware counters with a unique set of events which can be counted on each counter, with some events being available on multiple counters. Raw event codes for these processors are (PMC << 8) + PMCSEL. If PMC is non-zero then the event is that selected by the given PMCSEL value for that PMC (hardware counter). If PMC is zero then the event selected is one of the low-numbered ones that are common to several PMCs. In this case PMCSEL must be <= 22 and the event is what that PMCSEL value would select on PMC1 (but it may be placed any other PMC that has the same event for that PMCSEL value). For events that count cycles or occurrences that exceed a threshold, the threshold requested can be specified in the 0x3f000 bits of the raw event codes. If the event uses the threshold multiplier bit and that bit should be set, that is indicated with the 0x40000 bit of the raw event code. This fills in some of the generic cache events. Unfortunately there are quite a few blank spaces in the table, partly because these processors tend to count cache hits rather than cache accesses. Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: linuxppc-dev@ozlabs.org Cc: benh@kernel.crashing.org LKML-Reference: <19000.55631.802122.696927@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-06-17 18:53:51 +07:00
return -ENODEV;
return register_power_pmu(&mpc7450_pmu);
}
arch_initcall(init_mpc7450_pmu);