x86 Intel updates:
- Add Jasper Lake support
- Add support for TopDown metrics on Ice Lake
- Fix Ice Lake & Tiger Lake uncore support, add Snow Ridge support
- Add a PCI sub driver to support uncore PMUs where the PCI resources
have been claimed already - extending the range of supported systems.
x86 AMD updates:
- Restore 'perf stat -a' behaviour to program the uncore PMU
to count all CPU threads.
- Fix setting the proper count when sampling Large Increment
per Cycle events / 'paired' events.
- Fix IBS Fetch sampling on F17h and some other IBS fine tuning,
greatly reducing the number of interrupts when large sample
periods are specified.
- Extends Family 17h RAPL support to also work on compatible
F19h machines.
Core code updates:
- Fix race in perf_mmap_close()
- Add PERF_EV_CAP_SIBLING, to denote that sibling events should be
closed if the leader is removed.
- Smaller fixes and updates.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=2fuY
-----END PGP SIGNATURE-----
Merge tag 'perf-core-2020-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull performance events updates from Ingo Molnar:
"x86 Intel updates:
- Add Jasper Lake support
- Add support for TopDown metrics on Ice Lake
- Fix Ice Lake & Tiger Lake uncore support, add Snow Ridge support
- Add a PCI sub driver to support uncore PMUs where the PCI resources
have been claimed already - extending the range of supported
systems.
x86 AMD updates:
- Restore 'perf stat -a' behaviour to program the uncore PMU to count
all CPU threads.
- Fix setting the proper count when sampling Large Increment per
Cycle events / 'paired' events.
- Fix IBS Fetch sampling on F17h and some other IBS fine tuning,
greatly reducing the number of interrupts when large sample periods
are specified.
- Extends Family 17h RAPL support to also work on compatible F19h
machines.
Core code updates:
- Fix race in perf_mmap_close()
- Add PERF_EV_CAP_SIBLING, to denote that sibling events should be
closed if the leader is removed.
- Smaller fixes and updates"
* tag 'perf-core-2020-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (45 commits)
perf/core: Fix race in the perf_mmap_close() function
perf/x86: Fix n_metric for cancelled txn
perf/x86: Fix n_pair for cancelled txn
x86/events/amd/iommu: Fix sizeof mismatch
perf/x86/intel: Check perf metrics feature for each CPU
perf/x86/intel: Fix Ice Lake event constraint table
perf/x86/intel/uncore: Fix the scale of the IMC free-running events
perf/x86/intel/uncore: Fix for iio mapping on Skylake Server
perf/x86/msr: Add Jasper Lake support
perf/x86/intel: Add Jasper Lake support
perf/x86/intel/uncore: Reduce the number of CBOX counters
perf/x86/intel/uncore: Update Ice Lake uncore units
perf/x86/intel/uncore: Split the Ice Lake and Tiger Lake MSR uncore support
perf/x86/intel/uncore: Support PCIe3 unit on Snow Ridge
perf/x86/intel/uncore: Generic support for the PCI sub driver
perf/x86/intel/uncore: Factor out uncore_pci_pmu_unregister()
perf/x86/intel/uncore: Factor out uncore_pci_pmu_register()
perf/x86/intel/uncore: Factor out uncore_pci_find_dev_pmu()
perf/x86/intel/uncore: Factor out uncore_pci_get_dev_die_info()
perf/amd/uncore: Inform the user how many counters each uncore PMU has
...
When a group that has TopDown members is failed to be scheduled, any
later TopDown groups will not return valid values.
Here is an example.
A background perf that occupies all the GP counters and the fixed
counter 1.
$perf stat -e "{cycles,cycles,cycles,cycles,cycles,cycles,cycles,
cycles,cycles}:D" -a
A user monitors a TopDown group. It works well, because the fixed
counter 3 and the PERF_METRICS are available.
$perf stat -x, --topdown -- ./workload
retiring,bad speculation,frontend bound,backend bound,
18.0,16.1,40.4,25.5,
Then the user tries to monitor a group that has TopDown members.
Because of the cycles event, the group is failed to be scheduled.
$perf stat -x, -e '{slots,topdown-retiring,topdown-be-bound,
topdown-fe-bound,topdown-bad-spec,cycles}'
-- ./workload
<not counted>,,slots,0,0.00,,
<not counted>,,topdown-retiring,0,0.00,,
<not counted>,,topdown-be-bound,0,0.00,,
<not counted>,,topdown-fe-bound,0,0.00,,
<not counted>,,topdown-bad-spec,0,0.00,,
<not counted>,,cycles,0,0.00,,
The user tries to monitor a TopDown group again. It doesn't work anymore.
$perf stat -x, --topdown -- ./workload
,,,,,
In a txn, cancel_txn() is to truncate the event_list for a canceled
group and update the number of events added in this transaction.
However, the number of TopDown events added in this transaction is not
updated. The kernel will probably fail to add new Topdown events.
Fixes: 7b2c05a15d ("perf/x86/intel: Generic support for hardware TopDown metrics")
Reported-by: Andi Kleen <ak@linux.intel.com>
Reported-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lkml.kernel.org/r/20201005082611.GH2628@hirez.programming.kicks-ass.net
Kan reported that n_metric gets corrupted for cancelled transactions;
a similar issue exists for n_pair for AMD's Large Increment thing.
The problem was confirmed and confirmed fixed by Kim using:
sudo perf stat -e "{cycles,cycles,cycles,cycles}:D" -a sleep 10 &
# should succeed:
sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload
# should fail:
sudo perf stat -e "{fp_ret_sse_avx_ops.all,fp_ret_sse_avx_ops.all,cycles}:D" -a workload
# previously failed, now succeeds with this patch:
sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload
Fixes: 5738891229 ("perf/x86/amd: Add support for Large Increment per Cycle Events")
Reported-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Kim Phillips <kim.phillips@amd.com>
Link: https://lkml.kernel.org/r/20201005082516.GG2628@hirez.programming.kicks-ass.net
Commit 5738891229 ("perf/x86/amd: Add support for Large Increment
per Cycle Events") mistakenly zeroes the upper 16 bits of the count
in set_period(). That's fine for counting with perf stat, but not
sampling with perf record when only Large Increment events are being
sampled. To enable sampling, we sign extend the upper 16 bits of the
merged counter pair as described in the Family 17h PPRs:
"Software wanting to preload a value to a merged counter pair writes the
high-order 16-bit value to the low-order 16 bits of the odd counter and
then writes the low-order 48-bit value to the even counter. Reading the
even counter of the merged counter pair returns the full 64-bit value."
Fixes: 5738891229 ("perf/x86/amd: Add support for Large Increment per Cycle Events")
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537
Replace many of the indirect calls with static_call().
The average PMI time, as measured by perf_sample_event_took()*:
PRE: 3283.03 [ns]
POST: 3145.12 [ns]
Which is a ~138 [ns] win per PMI, or a ~4.2% decrease.
[*] on an IVB-EP, using: 'perf record -a -e cycles -- make O=defconfig-build/ -j80'
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20200818135805.338001015@infradead.org
Starts from Ice Lake, the TopDown metrics are directly available as
fixed counters and do not require generic counters. Also, the TopDown
metrics can be collected per thread. Extend the RDPMC usage to support
per-thread TopDown metrics.
The RDPMC index of the PERF_METRICS will be output if RDPMC users ask
for the RDPMC index of the metrics events.
To support per thread RDPMC TopDown, the metrics and slots counters have
to be saved/restored during the context switching.
The last_period and period_left are not used in the counting mode. Use
the fields for saved_metric and saved_slots.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-12-kan.liang@linux.intel.com
The RDPMC base offset of fixed counters is hard-code. Use a meaningful
name to replace the magic number to improve the readability of the code.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-10-kan.liang@linux.intel.com
Intro
=====
The TopDown Microarchitecture Analysis (TMA) Method is a structured
analysis methodology to identify critical performance bottlenecks in
out-of-order processors. Current perf has supported the method.
The method works well, but there is one problem. To collect the TopDown
events, several GP counters have to be used. If a user wants to collect
other events at the same time, the multiplexing probably be triggered,
which impacts the accuracy.
To free up the scarce GP counters, the hardware TopDown metrics feature
is introduced from Ice Lake. The hardware implements an additional
"metrics" register and a new Fixed Counter 3 that measures pipeline
"slots". The TopDown events can be calculated from them instead.
Events
======
The level 1 TopDown has four metrics. There is no event-code assigned to
the TopDown metrics. Four metric events are exported as separate perf
events, which map to the internal "metrics" counter register. Those
events do not exist in hardware, but can be allocated by the scheduler.
For the event mapping, a special 0x00 event code is used, which is
reserved for fake events. The metric events start from umask 0x10.
When setting up the metric events, they point to the Fixed Counter 3.
They have to be specially handled.
- Add the update_topdown_event() callback to read the additional metrics
MSR and generate the metrics.
- Add the set_topdown_event_period() callback to initialize metrics MSR
and the fixed counter 3.
- Add a variable n_metric_event to track the number of the accepted
metrics events. The sharing between multiple users of the same metric
without multiplexing is not allowed.
- Only enable/disable the fixed counter 3 when there are no other active
TopDown events, which avoid the unnecessary writing of the fixed
control register.
- Disable the PMU when reading the metrics event. The metrics MSR and
the fixed counter 3 are read separately. The values may be modified by
an NMI.
All four metric events don't support sampling. Since they will be
handled specially for event update, a flag PERF_X86_EVENT_TOPDOWN is
introduced to indicate this case.
The slots event can support both sampling and counting.
For counting, the flag is also applied.
For sampling, it will be handled normally as other normal events.
Groups
======
The slots event is required in a Topdown group.
To avoid reading the METRICS register multiple times, the metrics and
slots value can only be updated by slots event in a group.
All active slots and metrics events will be updated one time.
Therefore, the slots event must be before any metric events in a Topdown
group.
NMI
======
The METRICS related register may be overflow. The bit 48 of the STATUS
register will be set. If so, PERF_METRICS and Fixed counter 3 are
required to be reset. The patch also update all active slots and
metrics events in the NMI handler.
The update_topdown_event() has to read two registers separately. The
values may be modified by an NMI. PMU has to be disabled before calling
the function.
RDPMC
======
RDPMC is temporarily disabled. A later patch will enable it.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-9-kan.liang@linux.intel.com
The RDPMC index is always re-calculated for the RDPMC userspace support,
which is unnecessary.
The RDPMC index value is stored in the variable event_base_rdpmc for
the kernel usage, which can be used for RDPMC userspace support as well.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-2-kan.liang@linux.intel.com
Reading LBR registers in a perf NMI handler for a non-PEBS event
causes a high overhead because the number of LBR registers is huge.
To reduce the overhead, the XSAVES instruction should be used to replace
the LBR registers' reading method.
The XSAVES buffer used for LBR read has to be per-CPU because the NMI
handler invoked the lbr_read(). The existing task_ctx_data buffer
cannot be used which is per-task and only be allocated for the LBR call
stack mode. A new lbr_xsave pointer is introduced in the cpu_hw_events
as an XSAVES buffer for LBR read.
The XSAVES buffer should be allocated only when LBR is used by a
non-PEBS event on the CPU because the total size of the lbr_xsave is
not small (~1.4KB).
The XSAVES buffer is allocated when a non-PEBS event is added, but it
is lazily released in x86_release_hardware() when perf releases the
entire PMU hardware resource, because perf may frequently schedule the
event, e.g. high context switch. The lazy release method reduces the
overhead of frequently allocate/free the buffer.
If the lbr_xsave fails to be allocated, roll back to normal Arch LBR
lbr_read().
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dave Hansen <dave.hansen@intel.com>
Link: https://lkml.kernel.org/r/1593780569-62993-24-git-send-email-kan.liang@linux.intel.com
The hypervisor may request the perf subsystem to schedule a time window
to directly access the LBR records msrs for its own use. Normally, it would
create a guest LBR event with callstack mode enabled, which is scheduled
along with other ordinary LBR events on the host but in an exclusive way.
To avoid wasting a counter for the guest LBR event, the perf tracks its
hw->idx via INTEL_PMC_IDX_FIXED_VLBR and assigns it with a fake VLBR
counter with the help of new vlbr_constraint. As with the BTS event,
there is actually no hardware counter assigned for the guest LBR event.
Signed-off-by: Like Xu <like.xu@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200514083054.62538-5-like.xu@linux.intel.com
For intel_pmu_en/disable_event(), reorder the branches checks for hw->idx
and make them sorted by probability: gp,fixed,bts,others.
Clean up the x86_assign_hw_event() by converting multiple if-else
statements to a switch statement.
To skip x86_perf_event_update() and x86_perf_event_set_period(),
it's generic to replace "idx == INTEL_PMC_IDX_FIXED_BTS" check with
'!hwc->event_base' because that should be 0 for all non-gp/fixed cases.
Wrap related bit operations into intel_set/clear_masks() and make the main
path more cleaner and readable.
No functional changes.
Signed-off-by: Like Xu <like.xu@linux.intel.com>
Original-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200613080958.132489-3-like.xu@linux.intel.com
Add new APIs to assert that mmap_sem is held.
Using this instead of rwsem_is_locked and lockdep_assert_held[_write]
makes the assertions more tolerant of future changes to the lock type.
Signed-off-by: Michel Lespinasse <walken@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Davidlohr Bueso <dbueso@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Laurent Dufour <ldufour@linux.ibm.com>
Cc: Liam Howlett <Liam.Howlett@oracle.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ying Han <yinghan@google.com>
Link: http://lkml.kernel.org/r/20200520052908.204642-10-walken@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
- Unexport various PAT primitives
- Unexport per-CPU tlbstate
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=RXKK
-----END PGP SIGNATURE-----
Merge tag 'x86-mm-2020-06-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 mm updates from Ingo Molnar:
"Misc changes:
- Unexport various PAT primitives
- Unexport per-CPU tlbstate and uninline TLB helpers"
* tag 'x86-mm-2020-06-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (23 commits)
x86/tlb/uv: Add a forward declaration for struct flush_tlb_info
x86/cpu: Export native_write_cr4() only when CONFIG_LKTDM=m
x86/tlb: Restrict access to tlbstate
xen/privcmd: Remove unneeded asm/tlb.h include
x86/tlb: Move PCID helpers where they are used
x86/tlb: Uninline nmi_uaccess_okay()
x86/tlb: Move cr4_set_bits_and_update_boot() to the usage site
x86/tlb: Move paravirt_tlb_remove_table() to the usage site
x86/tlb: Move __flush_tlb_all() out of line
x86/tlb: Move flush_tlb_others() out of line
x86/tlb: Move __flush_tlb_one_kernel() out of line
x86/tlb: Move __flush_tlb_one_user() out of line
x86/tlb: Move __flush_tlb_global() out of line
x86/tlb: Move __flush_tlb() out of line
x86/alternatives: Move temporary_mm helpers into C
x86/cr4: Sanitize CR4.PCE update
x86/cpu: Uninline CR4 accessors
x86/tlb: Uninline __get_current_cr3_fast()
x86/mm: Use pgprotval_t in protval_4k_2_large() and protval_large_2_4k()
x86/mm: Unexport __cachemode2pte_tbl
...
Zhaoxin CPU has provided facilities for monitoring performance
via PMU (Performance Monitor Unit), but the functionality is unused so far.
Therefore, add support for zhaoxin pmu to make performance related
hardware events available.
The PMU is mostly an Intel Architectural PerfMon-v2 with a novel
errata for the ZXC line. It supports the following events:
-----------------------------------------------------------------------------------------------------------------------------------
Event | Event | Umask | Description
| Select | |
-----------------------------------------------------------------------------------------------------------------------------------
cpu-cycles | 82h | 00h | unhalt core clock
instructions | 00h | 00h | number of instructions at retirement.
cache-references | 15h | 05h | number of fillq pushs at the current cycle.
cache-misses | 1ah | 05h | number of l2 miss pushed by fillq.
branch-instructions | 28h | 00h | counts the number of branch instructions retired.
branch-misses | 29h | 00h | mispredicted branch instructions at retirement.
bus-cycles | 83h | 00h | unhalt bus clock
stalled-cycles-frontend | 01h | 01h | Increments each cycle the # of Uops issued by the RAT to RS.
stalled-cycles-backend | 0fh | 04h | RS0/1/2/3/45 empty
L1-dcache-loads | 68h | 05h | number of retire/commit load.
L1-dcache-load-misses | 4bh | 05h | retired load uops whose data source followed an L1 miss.
L1-dcache-stores | 69h | 06h | number of retire/commit Store,no LEA
L1-dcache-store-misses | 62h | 05h | cache lines in M state evicted out of L1D due to Snoop HitM or dirty line replacement.
L1-icache-loads | 00h | 03h | number of l1i cache access for valid normal fetch,including un-cacheable access.
L1-icache-load-misses | 01h | 03h | number of l1i cache miss for valid normal fetch,including un-cacheable miss.
L1-icache-prefetches | 0ah | 03h | number of prefetch.
L1-icache-prefetch-misses | 0bh | 03h | number of prefetch miss.
dTLB-loads | 68h | 05h | number of retire/commit load
dTLB-load-misses | 2ch | 05h | number of load operations miss all level tlbs and cause a tablewalk.
dTLB-stores | 69h | 06h | number of retire/commit Store,no LEA
dTLB-store-misses | 30h | 05h | number of store operations miss all level tlbs and cause a tablewalk.
dTLB-prefetches | 64h | 05h | number of hardware pte prefetch requests dispatched out of the prefetch FIFO.
dTLB-prefetch-misses | 65h | 05h | number of hardware pte prefetch requests miss the l1d data cache.
iTLB-load | 00h | 00h | actually counter instructions.
iTLB-load-misses | 34h | 05h | number of code operations miss all level tlbs and cause a tablewalk.
-----------------------------------------------------------------------------------------------------------------------------------
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: CodyYao-oc <CodyYao-oc@zhaoxin.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1586747669-4827-1-git-send-email-CodyYao-oc@zhaoxin.com
load_mm_cr4_irqsoff() is really a strange name for a function which has
only one purpose: Update the CR4.PCE bit depending on the perf state.
Rename it to update_cr4_pce_mm(), move it into the tlb code and provide a
function which can be invoked by the perf smp function calls.
Another step to remove exposure of cpu_tlbstate.
No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200421092559.049499158@linutronix.de
Description of hardware operation
---------------------------------
The core AMD PMU has a 4-bit wide per-cycle increment for each
performance monitor counter. That works for most events, but
now with AMD Family 17h and above processors, some events can
occur more than 15 times in a cycle. Those events are called
"Large Increment per Cycle" events. In order to count these
events, two adjacent h/w PMCs get their count signals merged
to form 8 bits per cycle total. In addition, the PERF_CTR count
registers are merged to be able to count up to 64 bits.
Normally, events like instructions retired, get programmed on a single
counter like so:
PERF_CTL0 (MSR 0xc0010200) 0x000000000053ff0c # event 0x0c, umask 0xff
PERF_CTR0 (MSR 0xc0010201) 0x0000800000000001 # r/w 48-bit count
The next counter at MSRs 0xc0010202-3 remains unused, or can be used
independently to count something else.
When counting Large Increment per Cycle events, such as FLOPs,
however, we now have to reserve the next counter and program the
PERF_CTL (config) register with the Merge event (0xFFF), like so:
PERF_CTL0 (msr 0xc0010200) 0x000000000053ff03 # FLOPs event, umask 0xff
PERF_CTR0 (msr 0xc0010201) 0x0000800000000001 # rd 64-bit cnt, wr lo 48b
PERF_CTL1 (msr 0xc0010202) 0x0000000f004000ff # Merge event, enable bit
PERF_CTR1 (msr 0xc0010203) 0x0000000000000000 # wr hi 16-bits count
The count is widened from the normal 48-bits to 64 bits by having the
second counter carry the higher 16 bits of the count in its lower 16
bits of its counter register.
The odd counter, e.g., PERF_CTL1, is programmed with the enabled Merge
event before the even counter, PERF_CTL0.
The Large Increment feature is available starting with Family 17h.
For more details, search any Family 17h PPR for the "Large Increment
per Cycle Events" section, e.g., section 2.1.15.3 on p. 173 in this
version:
https://www.amd.com/system/files/TechDocs/56176_ppr_Family_17h_Model_71h_B0_pub_Rev_3.06.zip
Description of software operation
---------------------------------
The following steps are taken in order to support reserving and
enabling the extra counter for Large Increment per Cycle events:
1. In the main x86 scheduler, we reduce the number of available
counters by the number of Large Increment per Cycle events being
scheduled, tracked by a new cpuc variable 'n_pair' and a new
amd_put_event_constraints_f17h(). This improves the counter
scheduler success rate.
2. In perf_assign_events(), if a counter is assigned to a Large
Increment event, we increment the current counter variable, so the
counter used for the Merge event is removed from assignment
consideration by upcoming event assignments.
3. In find_counter(), if a counter has been found for the Large
Increment event, we set the next counter as used, to prevent other
events from using it.
4. We perform steps 2 & 3 also in the x86 scheduler fastpath, i.e.,
we add Merge event accounting to the existing used_mask logic.
5. Finally, we add on the programming of Merge event to the
neighbouring PMC counters in the counter enable/disable{_all}
code paths.
Currently, software does not support a single PMU with mixed 48- and
64-bit counting, so Large increment event counts are limited to 48
bits. In set_period, we zero-out the upper 16 bits of the count, so
the hardware doesn't copy them to the even counter's higher bits.
Simple invocation example showing counting 8 FLOPs per 256-bit/%ymm
vaddps instruction executed in a loop 100 million times:
perf stat -e cpu/fp_ret_sse_avx_ops.all/,cpu/instructions/ <workload>
Performance counter stats for '<workload>':
800,000,000 cpu/fp_ret_sse_avx_ops.all/u
300,042,101 cpu/instructions/u
Prior to this patch, the reported SSE/AVX FLOPs retired count would
be wrong.
[peterz: lots of renames and edits to the code]
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Commit:
ccbebba4c6 ("perf/x86/intel/pt: Bypass PT vs. LBR exclusivity if the core supports it")
skips the PT/LBR exclusivity check on CPUs where PT and LBRs coexist, but
also inadvertently skips the active_events bump for PT in that case, which
is a bug. If there aren't any hardware events at the same time as PT, the
PMI handler will ignore PT PMIs, as active_events reads zero in that case,
resulting in the "Uhhuh" spurious NMI warning and PT data loss.
Fix this by always increasing active_events for PT events.
Fixes: ccbebba4c6 ("perf/x86/intel/pt: Bypass PT vs. LBR exclusivity if the core supports it")
Reported-by: Vitaly Slobodskoy <vitaly.slobodskoy@intel.com>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Alexey Budankov <alexey.budankov@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: https://lkml.kernel.org/r/20191210105101.77210-1-alexander.shishkin@linux.intel.com
UBSAN reported out-of-bound accesses for x86_pmu.event_map(), it's
arguments should be < x86_pmu.max_events. Make sure all users observe
this constraint.
Reported-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Meelis Roos <mroos@linux.ee>
When you successfully write 0 to /sys/devices/cpu/rdpmc, the RDPMC
instruction should be disabled unconditionally and immediately (after you
close the SYSFS file) by the documentation.
Instead, in the current implementation the PMU must be reloaded which
happens only eventually some time in the future. Only after that the RDPMC
instruction becomes disabled (on ring 3) on the respective core.
This change makes the treatment of the 0 value as blocking and as
unconditional as the current treatment of the 2 value, only the CR4.PCE
bit is naturally set to false instead of true.
Signed-off-by: Anthony Steinhauser <asteinhauser@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: acme@kernel.org
Link: https://lkml.kernel.org/r/20191125054838.137615-1-asteinhauser@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86 mm updates from Ingo Molnar:
- Make cpumask_of_node() more robust against invalid node IDs
- Simplify and speed up load_mm_cr4()
- Unexport and remove various unused set_memory_*() APIs
- Misc cleanups
* 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/mm: Fix cpumask_of_node() error condition
x86/mm: Remove the unused set_memory_wt() function
x86/mm: Remove set_pages_x() and set_pages_nx()
x86/mm: Remove the unused set_memory_array_*() functions
x86/mm: Unexport set_memory_x() and set_memory_nx()
x86/fixmap: Cleanup outdated comments
x86/kconfig: Remove X86_DIRECT_GBPAGES dependency on !DEBUG_PAGEALLOC
x86/mm: Avoid redundant interrupt disable in load_mm_cr4()
If PEBS declares ability to output its data to Intel PT stream, use the
aux_output attribute bit to enable PEBS data output to PT. This requires
a PT event to be present and scheduled in the same context. Unlike the
DS area, the kernel does not extract PEBS records from the PT stream to
generate corresponding records in the perf stream, because that would
require real time in-kernel PT decoding, which is not feasible. The PMI,
however, can still be used.
The output setting is per-CPU, so all PEBS events must be either writing
to PT or to the DS area, therefore, in case of conflict, the conflicting
event will fail to schedule, allowing the rotation logic to alternate
between the PEBS->PT and PEBS->DS events.
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: kan.liang@linux.intel.com
Link: https://lkml.kernel.org/r/20190806084606.4021-3-alexander.shishkin@linux.intel.com
load_mm_cr4() is always called with interrupts disabled from:
- switch_mm_irqs_off()
- refresh_pce(), which is a on_each_cpu() callback
Thus, disabling interrupts in cr4_set/clear_bits() is redundant.
Implement cr4_set/clear_bits_irqsoff() helpers, rename load_mm_cr4() to
load_mm_cr4_irqsoff() and use the new helpers. The new helpers do not need
a lockdep assert as __cr4_set() has one already.
The renaming in combination with the checks in __cr4_set() ensure that any
changes in the boundary conditions at the call sites will be detected.
[ tglx: Massaged change log ]
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/0fbbcb64-5f26-4ffb-1bb9-4f5f48426893@siemens.com
Pull perf updates from Ingo Molnar:
"The main changes in this cycle on the kernel side were:
- CPU PMU and uncore driver updates to Intel Snow Ridge, IceLake,
KabyLake, AmberLake and WhiskeyLake CPUs.
- Rework the MSR probing infrastructure to make it more robust, make
it work better on virtualized systems and to better expose it on
sysfs.
- Rework PMU attributes group support based on the feedback from
Greg. The core sysfs patch that adds sysfs_update_groups() was
acked by Greg.
There's a lot of perf tooling changes as well, all around the place:
- vendor updates to Intel, cs-etm (ARM), ARM64, s390,
- various enhancements to Intel PT tooling support:
- Improve CBR (Core to Bus Ratio) packets support.
- Export power and ptwrite events to sqlite and postgresql.
- Add support for decoding PEBS via PT packets.
- Add support for samples to contain IPC ratio, collecting cycles
information from CYC packets, showing the IPC info periodically
- Allow using time ranges
- lots of updates to perf pmu, perf stat, perf trace, eBPF support,
perf record, perf diff, etc. - please see the shortlog and Git log
for details"
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (252 commits)
tools arch x86: Sync asm/cpufeatures.h with the with the kernel
tools build: Check if gettid() is available before providing helper
perf jvmti: Address gcc string overflow warning for strncpy()
perf python: Remove -fstack-protector-strong if clang doesn't have it
perf annotate TUI browser: Do not use member from variable within its own initialization
perf tests: Fix record+probe_libc_inet_pton.sh for powerpc64
perf evsel: Do not rely on errno values for precise_ip fallback
perf thread: Allow references to thread objects after machine__exit()
perf header: Assign proper ff->ph in perf_event__synthesize_features()
tools arch kvm: Sync kvm headers with the kernel sources
perf script: Allow specifying the files to process guest samples
perf tools metric: Don't include duration_time in group
perf list: Avoid extra : for --raw metrics
perf vendor events intel: Metric fixes for SKX/CLX
perf tools: Fix typos / broken sentences
perf jevents: Add support for Hisi hip08 L3C PMU aliasing
perf jevents: Add support for Hisi hip08 HHA PMU aliasing
perf jevents: Add support for Hisi hip08 DDRC PMU aliasing
perf pmu: Support more complex PMU event aliasing
perf diff: Documentation -c cycles option
...
Pull locking updates from Ingo Molnar:
"The main changes in this cycle are:
- rwsem scalability improvements, phase #2, by Waiman Long, which are
rather impressive:
"On a 2-socket 40-core 80-thread Skylake system with 40 reader
and writer locking threads, the min/mean/max locking operations
done in a 5-second testing window before the patchset were:
40 readers, Iterations Min/Mean/Max = 1,807/1,808/1,810
40 writers, Iterations Min/Mean/Max = 1,807/50,344/151,255
After the patchset, they became:
40 readers, Iterations Min/Mean/Max = 30,057/31,359/32,741
40 writers, Iterations Min/Mean/Max = 94,466/95,845/97,098"
There's a lot of changes to the locking implementation that makes
it similar to qrwlock, including owner handoff for more fair
locking.
Another microbenchmark shows how across the spectrum the
improvements are:
"With a locking microbenchmark running on 5.1 based kernel, the
total locking rates (in kops/s) on a 2-socket Skylake system
with equal numbers of readers and writers (mixed) before and
after this patchset were:
# of Threads Before Patch After Patch
------------ ------------ -----------
2 2,618 4,193
4 1,202 3,726
8 802 3,622
16 729 3,359
32 319 2,826
64 102 2,744"
The changes are extensive and the patch-set has been through
several iterations addressing various locking workloads. There
might be more regressions, but unless they are pathological I
believe we want to use this new implementation as the baseline
going forward.
- jump-label optimizations by Daniel Bristot de Oliveira: the primary
motivation was to remove IPI disturbance of isolated RT-workload
CPUs, which resulted in the implementation of batched jump-label
updates. Beyond the improvement of the real-time characteristics
kernel, in one test this patchset improved static key update
overhead from 57 msecs to just 1.4 msecs - which is a nice speedup
as well.
- atomic64_t cross-arch type cleanups by Mark Rutland: over the last
~10 years of atomic64_t existence the various types used by the
APIs only had to be self-consistent within each architecture -
which means they became wildly inconsistent across architectures.
Mark puts and end to this by reworking all the atomic64
implementations to use 's64' as the base type for atomic64_t, and
to ensure that this type is consistently used for parameters and
return values in the API, avoiding further problems in this area.
- A large set of small improvements to lockdep by Yuyang Du: type
cleanups, output cleanups, function return type and othr cleanups
all around the place.
- A set of percpu ops cleanups and fixes by Peter Zijlstra.
- Misc other changes - please see the Git log for more details"
* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (82 commits)
locking/lockdep: increase size of counters for lockdep statistics
locking/atomics: Use sed(1) instead of non-standard head(1) option
locking/lockdep: Move mark_lock() inside CONFIG_TRACE_IRQFLAGS && CONFIG_PROVE_LOCKING
x86/jump_label: Make tp_vec_nr static
x86/percpu: Optimize raw_cpu_xchg()
x86/percpu, sched/fair: Avoid local_clock()
x86/percpu, x86/irq: Relax {set,get}_irq_regs()
x86/percpu: Relax smp_processor_id()
x86/percpu: Differentiate this_cpu_{}() and __this_cpu_{}()
locking/rwsem: Guard against making count negative
locking/rwsem: Adaptive disabling of reader optimistic spinning
locking/rwsem: Enable time-based spinning on reader-owned rwsem
locking/rwsem: Make rwsem->owner an atomic_long_t
locking/rwsem: Enable readers spinning on writer
locking/rwsem: Clarify usage of owner's nonspinaable bit
locking/rwsem: Wake up almost all readers in wait queue
locking/rwsem: More optimal RT task handling of null owner
locking/rwsem: Always release wait_lock before waking up tasks
locking/rwsem: Implement lock handoff to prevent lock starvation
locking/rwsem: Make rwsem_spin_on_owner() return owner state
...
Pull x86 fixes from Ingo Molnar:
"Misc fixes all over the place:
- might_sleep() atomicity fix in the microcode loader
- resctrl boundary condition fix
- APIC arithmethics bug fix for frequencies >= 4.2 GHz
- three 5-level paging crash fixes
- two speculation fixes
- a perf/stacktrace fix"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/unwind/orc: Fall back to using frame pointers for generated code
perf/x86: Always store regs->ip in perf_callchain_kernel()
x86/speculation: Allow guests to use SSBD even if host does not
x86/mm: Handle physical-virtual alignment mismatch in phys_p4d_init()
x86/boot/64: Add missing fixup_pointer() for next_early_pgt access
x86/boot/64: Fix crash if kernel image crosses page table boundary
x86/apic: Fix integer overflow on 10 bit left shift of cpu_khz
x86/resctrl: Prevent possible overrun during bitmap operations
x86/microcode: Fix the microcode load on CPU hotplug for real
The stacktrace_map_raw_tp BPF selftest is failing because the RIP saved by
perf_arch_fetch_caller_regs() isn't getting saved by perf_callchain_kernel().
This was broken by the following commit:
d15d356887 ("perf/x86: Make perf callchains work without CONFIG_FRAME_POINTER")
With that change, when starting with non-HW regs, the unwinder starts
with the current stack frame and unwinds until it passes up the frame
which called perf_arch_fetch_caller_regs(). So regs->ip needs to be
saved deliberately.
Fixes: d15d356887 ("perf/x86: Make perf callchains work without CONFIG_FRAME_POINTER")
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Kairui Song <kasong@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Borislav Petkov <bp@alien8.de>
Link: https://lkml.kernel.org/r/3975a298fa52b506fea32666d8ff6a13467eee6d.1561595111.git.jpoimboe@redhat.com
We don't need pmu->pebs_no_xmm_regs anymore, the capabilities
PERF_PMU_CAP_EXTENDED_REGS can be used to check if XMM registers
collection is supported.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: https://lkml.kernel.org/r/1559081314-9714-4-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
All callers of lockdep_assert_held_exclusive() use it to verify the
correct locking state of either a semaphore (ldisc_sem in tty,
mmap_sem for perf events, i_rwsem of inode for dax) or rwlock by
apparmor. Thus it makes sense to rename _exclusive to _write since
that's the semantics callers care. Additionally there is already
lockdep_assert_held_read(), which this new naming is more consistent with.
No functional changes.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190531100651.3969-1-nborisov@suse.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Using the new pmu::update_attrs attribute group for default
attributes - freeze_on_smi, allow_tsx_force_abort.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190512155518.21468-10-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Using the new pmu::update_attrs attribute group for
"caps" directory.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190512155518.21468-7-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We dont need to pre-filter out unsupported base events,
we can just use its group's is_visible function to do this.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190512155518.21468-6-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Using the new pmu::update_attrs attribute group to
create detected events for x86_pmu.
Moving the topdown/memory/tsx attributes to separate
attribute groups with specific is_visible functions.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190512155518.21468-5-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently perf callchain doesn't work well with ORC unwinder
when sampling from trace point. We'll get useless in kernel callchain
like this:
perf 6429 [000] 22.498450: kmem:mm_page_alloc: page=0x176a17 pfn=1534487 order=0 migratetype=0 gfp_flags=GFP_KERNEL
ffffffffbe23e32e __alloc_pages_nodemask+0x22e (/lib/modules/5.1.0-rc3+/build/vmlinux)
7efdf7f7d3e8 __poll+0x18 (/usr/lib64/libc-2.28.so)
5651468729c1 [unknown] (/usr/bin/perf)
5651467ee82a main+0x69a (/usr/bin/perf)
7efdf7eaf413 __libc_start_main+0xf3 (/usr/lib64/libc-2.28.so)
5541f689495641d7 [unknown] ([unknown])
The root cause is that, for trace point events, it doesn't provide a
real snapshot of the hardware registers. Instead perf tries to get
required caller's registers and compose a fake register snapshot
which suppose to contain enough information for start a unwinding.
However without CONFIG_FRAME_POINTER, if failed to get caller's BP as the
frame pointer, so current frame pointer is returned instead. We get
a invalid register combination which confuse the unwinder, and end the
stacktrace early.
So in such case just don't try dump BP, and let the unwinder start
directly when the register is not a real snapshot. Use SP
as the skip mark, unwinder will skip all the frames until it meet
the frame of the trace point caller.
Tested with frame pointer unwinder and ORC unwinder, this makes perf
callchain get the full kernel space stacktrace again like this:
perf 6503 [000] 1567.570191: kmem:mm_page_alloc: page=0x16c904 pfn=1493252 order=0 migratetype=0 gfp_flags=GFP_KERNEL
ffffffffb523e2ae __alloc_pages_nodemask+0x22e (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb52383bd __get_free_pages+0xd (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb52fd28a __pollwait+0x8a (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb521426f perf_poll+0x2f (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb52fe3e2 do_sys_poll+0x252 (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb52ff027 __x64_sys_poll+0x37 (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb500418b do_syscall_64+0x5b (/lib/modules/5.1.0-rc3+/build/vmlinux)
ffffffffb5a0008c entry_SYSCALL_64_after_hwframe+0x44 (/lib/modules/5.1.0-rc3+/build/vmlinux)
7f71e92d03e8 __poll+0x18 (/usr/lib64/libc-2.28.so)
55a22960d9c1 [unknown] (/usr/bin/perf)
55a22958982a main+0x69a (/usr/bin/perf)
7f71e9202413 __libc_start_main+0xf3 (/usr/lib64/libc-2.28.so)
5541f689495641d7 [unknown] ([unknown])
Co-developed-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kairui Song <kasong@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Young <dyoung@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190422162652.15483-1-kasong@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Starting from Icelake, XMM registers can be collected in PEBS record.
But current code only output the pt_regs.
Add a new struct x86_perf_regs for both pt_regs and xmm_regs. The
xmm_regs will be used later to keep a pointer to PEBS record which has
XMM information.
XMM registers are 128 bit. To simplify the code, they are handled like
two different registers, which means setting two bits in the register
bitmap. This also allows only sampling the lower 64bit bits in XMM.
The index of XMM registers starts from 32. There are 16 XMM registers.
So all reserved space for regs are used. Remove REG_RESERVED.
Add PERF_REG_X86_XMM_MAX, which stands for the max number of all x86
regs including both GPRs and XMM.
Add REG_NOSUPPORT for 32bit to exclude unsupported registers.
Previous platforms can not collect XMM information in PEBS record.
Adding pebs_no_xmm_regs to indicate the unsupported platforms.
The common code still validates the supported registers. However, it
cannot check model specific registers, e.g. XMM. Add extra check in
x86_pmu_hw_config() to reject invalid config of regs_user and regs_intr.
The regs_user never supports XMM collection.
The regs_intr only supports XMM collection when sampling PEBS event on
icelake and later platforms.
Originally-by: Andi Kleen <ak@linux.intel.com>
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: acme@kernel.org
Cc: jolsa@kernel.org
Link: https://lkml.kernel.org/r/20190402194509.2832-3-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Spurious interrupt support was added to perf in the following commit, almost
a decade ago:
63e6be6d98 ("perf, x86: Catch spurious interrupts after disabling counters")
The two previous patches (resolving the race condition when disabling a
PMC and NMI latency mitigation) allow for the removal of this older
spurious interrupt support.
Currently in x86_pmu_stop(), the bit for the PMC in the active_mask bitmap
is cleared before disabling the PMC, which sets up a race condition. This
race condition was mitigated by introducing the running bitmap. That race
condition can be eliminated by first disabling the PMC, waiting for PMC
reset on overflow and then clearing the bit for the PMC in the active_mask
bitmap. The NMI handler will not re-enable a disabled counter.
If x86_pmu_stop() is called from the perf NMI handler, the NMI latency
mitigation support will guard against any unhandled NMI messages.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: <stable@vger.kernel.org> # 4.14.x-
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: https://lkml.kernel.org/r/Message-ID:
Signed-off-by: Ingo Molnar <mingo@kernel.org>
By computing the 'committed' index earlier, we can use it to validate
the cached constraint state.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Now that cpuc->event_constraint[] is retained, we can avoid calling
get_event_constraints() over and over again.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The current code unconditionally clears cpuc->event_constraint[i]
before calling get_event_constraints(.idx=i). The only site that cares
is intel_get_event_constraints() where the c1 load will always be
NULL.
However, always calling get_event_constraints() on all events is
wastefull, most times it will return the exact same result. Therefore
retain the logic in intel_get_event_constraints() and change the
generic code to only clear the constraint on put.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Stephane Eranian <eranian@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
