Add hardware performance counter support to kernel "perf" code for
BMIPS5000. The BMIPS5000 performance counters are similar to MIPS
MTI cores, so the changes were mostly made in perf_event_mipsxx.c
which is typically for MTI cores.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/4109/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Split the Kconfig option CONFIG_MIPS_MT_SMP into CONFIG_MIPS_MT_SMP
and CONFIG_MIPS_PERF_SHARED_TC_COUNTERS so some of the code used
for performance counters that are shared between threads can be used
for MIPS cores that are not MT_SMP.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/4108/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The #ifdef for CONFIG_HW_PERF_EVENTS is not needed because the
Makefile will only compile the module if this config option is set.
This means that the code under #else would never be compiled. This
may have been done to leave the original broken code around for
reference, but the FIXME comment above the code already shows the
broken code.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/4107/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The PCI (Program Counter Interrupt) bit in the "cause" register
is mandatory for MIPS32R2 cores, but has also been added to some R1
cores (BMIPS5000). This change adds a cpu feature bit to make it
easier to check for and use this feature.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/4106/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Change the indicator from 0xffffffff in the "event_id" member to
zero in the "cntr_mask" member. This removes the need to initialize
entries that are unsupported. This also solves a problem where the
number of entries in the table was increased based on a globel enum
used for all platforms, but the new unsupported entries were not added
for mips. This was leaving new table entries of all zeros that we not
marked UNSUPPORTED.
Signed-off-by: NAl Cooper <alcooperx@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/4110/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Loongson 1B is a 32-bit SoC designed by Institute of Computing Technology
(ICT) and the Chinese Academy of Sciences (CAS), which implements the
MIPS32 release 2 instruction set.

[ralf@linux-mips.org: But which is not strictly a MIPS32 compliant device
which also is why it identifies itself with the Legacy Vendor ID in the
PrID register.  When applying the patch I shoveled some code around to
keep things in alphabetical order and avoid forward declarations.]
Signed-off-by: NKelvin Cheung <keguang.zhang@gmail.com>
Cc: To: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: wuzhangjin@gmail.com
Cc: zhzhl555@gmail.com
Cc: Kelvin Cheung <keguang.zhang@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/3976/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

cc1: warnings being treated as errors
arch/mips/kernel/perf_event_mipsxx.c:166: error: 'counters_per_cpu_to_total' defined but not used
make[2]: *** [arch/mips/kernel/perf_event_mipsxx.o] Error 1
make[2]: *** Waiting for unfinished jobs....

It was first introduced by 82091564 [MIPS:
perf: Add support for 64-bit perf counters.] in 3.2.
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
Cc: linux-mips@linux-mips.org
Cc: david.daney@cavium.com
Patchwork: https://patchwork.linux-mips.org/patch/3357/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Make the oprofile code use the performance counters irq.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn Crispin <blogic@openwrt.org>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/3723/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

We always need to pass the last sample period to
perf_sample_data_init(), otherwise the event distribution will be
wrong. Thus, modifiyng the function interface with the required period
as argument. So basically a pattern like this:

        perf_sample_data_init(&data, ~0ULL);
        data.period = event->hw.last_period;

will now be like that:

        perf_sample_data_init(&data, ~0ULL, event->hw.last_period);

Avoids unininitialized data.period and simplifies code.
Signed-off-by: NRobert Richter <robert.richter@amd.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1333390758-10893-3-git-send-email-robert.richter@amd.comSigned-off-by: NIngo Molnar <mingo@kernel.org>

PERF_SAMPLE_BRANCH_* is disabled for:

 - SW events (sw counters, tracepoints)
 - HW breakpoints
 - ALL but Intel x86 architecture
 - AMD64 processors
Signed-off-by: NStephane Eranian <eranian@google.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328826068-11713-10-git-send-email-eranian@google.comSigned-off-by: NIngo Molnar <mingo@elte.hu>

Simplify the code by changing the place of event->destroy().
Signed-off-by: NDeng-Cheng Zhu <dczhu@mips.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: David Daney <david.daney@cavium.com>
Cc: Eyal Barzilay <eyal@mips.com>
Cc: Zenon Fortuna <zenon@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/3109/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Why removing pmu checking:
Since 3.2-rc1, when arch level event init is called, the event is already
connected to its PMU. Also, validate_event() is _only_ called by
validate_group() in event init, so there is no need of checking or
temporarily assigning event pmu during validate_group().

Why removing event state checking:
Events could be created in PERF_EVENT_STATE_OFF (attr->disabled == 1), when
these events go through this checking, validate_group() does dummy work.
But we do need to do group scheduling emulation for them in event init.
Again, validate_event() is _only_ called by validate_group().

Reference: http://www.spinics.net/lists/mips/msg42190.htmlSigned-off-by: NDeng-Cheng Zhu <dczhu@mips.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: David Daney <david.daney@cavium.com>
Cc: Eyal Barzilay <eyal@mips.com>
Cc: Zenon Fortuna <zenon@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/3108/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Port the following patch for ARM by Mark Rutland:

- 57ce9bb3
    ARM: 6902/1: perf: Remove erroneous check on active_events

    When initialising a PMU, there is a check to protect against races with
    other CPUs filling all of the available event slots. Since armpmu_add
    checks that an event can be scheduled, we do not need to do this at
    initialisation time. Furthermore the current code is broken because it
    assumes that atomic_inc_not_zero will unconditionally increment
    active_counts and then tries to decrement it again on failure.

    This patch removes the broken, redundant code.
Signed-off-by: NDeng-Cheng Zhu <dczhu@mips.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: David Daney <david.daney@cavium.com>
Cc: Eyal Barzilay <eyal@mips.com>
Cc: Zenon Fortuna <zenon@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/3106/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

MIPS licensees may want to modify performance counters to count extra
events. Also, now that the user is working on raw events, the manual is
being used for sure. And feeding unsupported events shouldn't cause
hardware failure and the like.

[ralf@linux-mips.org: performance events also being used in internal
performance evaluation and have a tendency to change as the micro-
architecture evolves, even for minor revisions that may not be
distinguishable by PrID.  It's not very practicable to maintain a list
of all events and there is no real benefit.]
Signed-off-by: NDeng-Cheng Zhu <dczhu@mips.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: David Daney <david.daney@cavium.com>
Cc: Eyal Barzilay <eyal@mips.com>
Cc: Zenon Fortuna <zenon@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/3107/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

People (Linus) objected to using -ENOSPC to signal not having enough
resources on the PMU to satisfy the request. Use -EINVAL.
Requested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
Cc: David Daney <david.daney@cavium.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-xv8geaz2zpbjhlx0svmpp28n@git.kernel.org
[ merged to newer kernel, fixed up MIPS impact ]
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Enable hardware counters for Octeon, and add the corresponding event
mappings.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/2790/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The hard coded constants are moved to struct mips_pmu.  All counter
register access move to the read_counter and write_counter function
pointers, which are set to either 32-bit or 64-bit access methods at
initialization time.

Many of the function pointers in struct mips_pmu were not needed as
there was only a single implementation, these were removed.

I couldn't figure out what made struct cpu_hw_events.msbs[] at all
useful, so I removed it too.

Some functions and other declarations were reordered to reduce the
need for forward declarations.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/2792/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The contents of arch/mips/kernel/perf_event.c and
arch/mips/kernel/perf_event_mipsxx.c were divided in a seemingly ad
hoc manner, with the first including the second.

I moved all the hardware counter support code to perf_event_mipsxx.c
and removed the gating #ifdefs to the Kconfig and Makefile.

Now perf_event.c contains only the callchain support, everything else
is in perf_event_mipsxx.c

There are no code changes, only moving of functions from one file to
the other, or removing empty unneeded functions.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Dezhong Diao <dediao@cisco.com>
Cc: Gabor Juhos <juhosg@openwrt.org>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/2791/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Get rid of a bunch of useless inline declarations, and join a bunch of
improperly split lines.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/2793/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Add a NODE level to the generic cache events which is used to measure
local vs remote memory accesses. Like all other cache events, an
ACCESS is HIT+MISS, if there is no way to distinguish between reads
and writes do reads only etc..

The below needs filling out for !x86 (which I filled out with
unsupported events).

I'm fairly sure ARM can leave it like that since it doesn't strike me as
an architecture that even has NUMA support. SH might have something since
it does appear to have some NUMA bits.

Sparc64, PowerPC and MIPS certainly want a good look there since they
clearly are NUMA capable.
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: David Miller <davem@davemloft.net>
Cc: Anton Blanchard <anton@samba.org>
Cc: David Daney <ddaney@caviumnetworks.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1303508226.4865.8.camel@laptopSigned-off-by: NIngo Molnar <mingo@elte.hu>

Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: NLucas De Marchi <lucas.demarchi@profusion.mobi>

This is the MIPS part of the following commits by Peter Zijlstra:

- a4eaf7f1
    perf: Rework the PMU methods

    Replace pmu::{enable,disable,start,stop,unthrottle} with
    pmu::{add,del,start,stop}, all of which take a flags argument.

    The new interface extends the capability to stop a counter while
    keeping it scheduled on the PMU. We replace the throttled state with
    the generic stopped state.

    This also allows us to efficiently stop/start counters over certain
    code paths (like IRQ handlers).

    It also allows scheduling a counter without it starting, allowing for
    a generic frozen state (useful for rotating stopped counters).

    The stopped state is implemented in two different ways, depending on
    how the architecture implemented the throttled state:

     1) We disable the counter:
        a) the pmu has per-counter enable bits, we flip that
        b) we program a NOP event, preserving the counter state

     2) We store the counter state and ignore all read/overflow events

For MIPSXX, the stopped state is implemented in the way of 1.b as above.

- 33696fc0
    perf: Per PMU disable

    Changes perf_disable() into perf_pmu_disable().

- 24cd7f54
    perf: Reduce perf_disable() usage

    Since the current perf_disable() usage is only an optimization,
    remove it for now. This eases the removal of the __weak
    hw_perf_enable() interface.

- b0a873eb
    perf: Register PMU implementations

    Simple registration interface for struct pmu, this provides the
    infrastructure for removing all the weak functions.

- 51b0fe39
    perf: Deconstify struct pmu

    sed -ie 's/const struct pmu\>/struct pmu/g' `git grep -l "const struct pmu\>"`
Reported-by: NWu Zhangjin <wuzhangjin@gmail.com>
Acked-by: NDavid Daney <ddaney@caviumnetworks.com>
Signed-off-by: NDeng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: a.p.zijlstra@chello.nl
To: fweisbec@gmail.com
To: will.deacon@arm.com
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: wuzhangjin@gmail.com
Cc: paulus@samba.org
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: dengcheng.zhu@gmail.com
Cc: matt@console-pimps.org
Cc: sshtylyov@mvista.com
Cc: ddaney@caviumnetworks.com
Patchwork: http://patchwork.linux-mips.org/patch/2012/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This is the MIPS part of the following commit by Peter Zijlstra:

- e360adbe
    irq_work: Add generic hardirq context callbacks

    Provide a mechanism that allows running code in IRQ context. It is
    most useful for NMI code that needs to interact with the rest of the
    system -- like wakeup a task to drain buffers.

    Perf currently has such a mechanism, so extract that and provide it as
    a generic feature, independent of perf so that others may also
    benefit.

    The IRQ context callback is generated through self-IPIs where
    possible, or on architectures like powerpc the decrementer (the
    built-in timer facility) is set to generate an interrupt immediately.

    Architectures that don't have anything like this get to do with a
    callback from the timer tick. These architectures can call
    irq_work_run() at the tail of any IRQ handlers that might enqueue such
    work (like the perf IRQ handler) to avoid undue latencies in
    processing the work.

For MIPSXX, we need to call irq_work_run() at the tail of the perf IRQ
handler as described above.
Reported-by: NWu Zhangjin <wuzhangjin@gmail.com>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: NDavid Daney <ddaney@caviumnetworks.com>
Signed-off-by: NDeng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: fweisbec@gmail.com
To: will.deacon@arm.com
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: paulus@samba.org
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: matt@console-pimps.org
Cc: sshtylyov@mvista.com,
Patchwork: http://patchwork.linux-mips.org/patch/2011/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The perf hardware pmu got initialized at various points in the boot,
some before early_initcall() some after (notably arch_initcall).

The problem is that the NMI lockup detector is ran from early_initcall()
and expects the hardware pmu to be present.

Sanitize this by moving all architecture hardware pmu implementations to
initialize at early_initcall() and move the lockup detector to an explicit
initcall right after that.

Cc: paulus <paulus@samba.org>
Cc: davem <davem@davemloft.net>
Cc: Michael Cree <mcree@orcon.net.nz>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
Acked-by: NPaul Mundt <lethal@linux-sh.org>
Acked-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1290707759.2145.119.camel@laptop>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This patch adds the mipsxx Perf-events support based on the skeleton.
Generic hardware events and cache events are now fully implemented for
the 24K/34K/74K/1004K cores. To support other cores in mipsxx (such as
R10000/SB1), the generic hardware event tables and cache event tables
need to be filled out. To support other CPUs which have different PMU
than mipsxx, such as RM9000 and LOONGSON2, the additional files
perf_event_$cpu.c need to be created.

Raw event is an important part of Perf-events. It helps the user collect
performance data for events that are not listed as the generic hardware
events and cache events but ARE supported by the CPU's PMU.

This patch also adds this feature for mipsxx 24K/34K/74K/1004K. For how to
use it, please refer to processor core software user's manual and the
comments for mipsxx_pmu_map_raw_event() for more details.

Please note that this is a "precise" implementation, which means the
kernel will check whether the requested raw events are supported by this
CPU and which hardware counters can be assigned for them.

To test the functionality of Perf-event, you may want to compile the tool
"perf" for your MIPS platform. You can refer to the following URL:
http://www.linux-mips.org/archives/linux-mips/2010-10/msg00126.html

You also need to customize the CFLAGS and LDFLAGS in tools/perf/Makefile
for your libs, includes, etc.

In case you encounter the boot failure in SMVP kernel on multi-threading
CPUs, you may take a look at:
http://www.linux-mips.org/git?p=linux-mti.git;a=commitdiff;h=5460815027d802697b879644c74f0e8365254020Signed-off-by: NDeng-Cheng Zhu <dengcheng.zhu@gmail.com>
To: linux-mips@linux-mips.org
Cc: a.p.zijlstra@chello.nl
Cc: paulus@samba.org
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: jamie.iles@picochip.com
Cc: ddaney@caviumnetworks.com
Cc: matt@console-pimps.org
Patchwork: https://patchwork.linux-mips.org/patch/1689/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

 create mode 100644 arch/mips/kernel/perf_event_mipsxx.c