Nothing really major this release, despite quite a lot of activity. Just lots of
things all over the place.
Some things of note include:
- Access via perf to a new type of PMU (IMC) on Power9, which can count both
core events as well as nest unit events (Memory controller etc).
- Optimisations to the radix MMU TLB flushing, mostly to avoid unnecessary Page
Walk Cache (PWC) flushes when the structure of the tree is not changing.
- Reworks/cleanups of do_page_fault() to modernise it and bring it closer to
other architectures where possible.
- Rework of our page table walking so that THP updates only need to send IPIs
to CPUs where the affected mm has run, rather than all CPUs.
- The size of our vmalloc area is increased to 56T on 64-bit hash MMU systems.
This avoids problems with the percpu allocator on systems with very sparse
NUMA layouts.
- STRICT_KERNEL_RWX support on PPC32.
- A new sched domain topology for Power9, to capture the fact that pairs of
cores may share an L2 cache.
- Power9 support for VAS, which is a new mechanism for accessing coprocessors,
and initial support for using it with the NX compression accelerator.
- Major work on the instruction emulation support, adding support for many new
instructions, and reworking it so it can be used to implement the emulation
needed to fixup alignment faults.
- Support for guests under PowerVM to use the Power9 XIVE interrupt controller.
And probably that many things again that are almost as interesting, but I had to
keep the list short. Plus the usual fixes and cleanups as always.
Thanks to:
Alexey Kardashevskiy, Alistair Popple, Andreas Schwab, Aneesh Kumar K.V, Anju
T Sudhakar, Arvind Yadav, Balbir Singh, Benjamin Herrenschmidt, Bhumika Goyal,
Breno Leitao, Bryant G. Ly, Christophe Leroy, Cédric Le Goater, Dan Carpenter,
Dou Liyang, Frederic Barrat, Gautham R. Shenoy, Geliang Tang, Geoff Levand,
Hannes Reinecke, Haren Myneni, Ivan Mikhaylov, John Allen, Julia Lawall, LABBE
Corentin, Laurentiu Tudor, Madhavan Srinivasan, Markus Elfring, Masahiro
Yamada, Matt Brown, Michael Neuling, Murilo Opsfelder Araujo, Nathan Fontenot,
Naveen N. Rao, Nicholas Piggin, Oliver O'Halloran, Paul Mackerras, Rashmica
Gupta, Rob Herring, Rui Teng, Sam Bobroff, Santosh Sivaraj, Scott Wood,
Shilpasri G Bhat, Sukadev Bhattiprolu, Suraj Jitindar Singh, Tobin C. Harding,
Victor Aoqui.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=UJtt
-----END PGP SIGNATURE-----
Merge tag 'powerpc-4.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc updates from Michael Ellerman:
"Nothing really major this release, despite quite a lot of activity.
Just lots of things all over the place.
Some things of note include:
- Access via perf to a new type of PMU (IMC) on Power9, which can
count both core events as well as nest unit events (Memory
controller etc).
- Optimisations to the radix MMU TLB flushing, mostly to avoid
unnecessary Page Walk Cache (PWC) flushes when the structure of the
tree is not changing.
- Reworks/cleanups of do_page_fault() to modernise it and bring it
closer to other architectures where possible.
- Rework of our page table walking so that THP updates only need to
send IPIs to CPUs where the affected mm has run, rather than all
CPUs.
- The size of our vmalloc area is increased to 56T on 64-bit hash MMU
systems. This avoids problems with the percpu allocator on systems
with very sparse NUMA layouts.
- STRICT_KERNEL_RWX support on PPC32.
- A new sched domain topology for Power9, to capture the fact that
pairs of cores may share an L2 cache.
- Power9 support for VAS, which is a new mechanism for accessing
coprocessors, and initial support for using it with the NX
compression accelerator.
- Major work on the instruction emulation support, adding support for
many new instructions, and reworking it so it can be used to
implement the emulation needed to fixup alignment faults.
- Support for guests under PowerVM to use the Power9 XIVE interrupt
controller.
And probably that many things again that are almost as interesting,
but I had to keep the list short. Plus the usual fixes and cleanups as
always.
Thanks to: Alexey Kardashevskiy, Alistair Popple, Andreas Schwab,
Aneesh Kumar K.V, Anju T Sudhakar, Arvind Yadav, Balbir Singh,
Benjamin Herrenschmidt, Bhumika Goyal, Breno Leitao, Bryant G. Ly,
Christophe Leroy, Cédric Le Goater, Dan Carpenter, Dou Liyang,
Frederic Barrat, Gautham R. Shenoy, Geliang Tang, Geoff Levand, Hannes
Reinecke, Haren Myneni, Ivan Mikhaylov, John Allen, Julia Lawall,
LABBE Corentin, Laurentiu Tudor, Madhavan Srinivasan, Markus Elfring,
Masahiro Yamada, Matt Brown, Michael Neuling, Murilo Opsfelder Araujo,
Nathan Fontenot, Naveen N. Rao, Nicholas Piggin, Oliver O'Halloran,
Paul Mackerras, Rashmica Gupta, Rob Herring, Rui Teng, Sam Bobroff,
Santosh Sivaraj, Scott Wood, Shilpasri G Bhat, Sukadev Bhattiprolu,
Suraj Jitindar Singh, Tobin C. Harding, Victor Aoqui"
* tag 'powerpc-4.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (321 commits)
powerpc/xive: Fix section __init warning
powerpc: Fix kernel crash in emulation of vector loads and stores
powerpc/xive: improve debugging macros
powerpc/xive: add XIVE Exploitation Mode to CAS
powerpc/xive: introduce H_INT_ESB hcall
powerpc/xive: add the HW IRQ number under xive_irq_data
powerpc/xive: introduce xive_esb_write()
powerpc/xive: rename xive_poke_esb() in xive_esb_read()
powerpc/xive: guest exploitation of the XIVE interrupt controller
powerpc/xive: introduce a common routine xive_queue_page_alloc()
powerpc/sstep: Avoid used uninitialized error
axonram: Return directly after a failed kzalloc() in axon_ram_probe()
axonram: Improve a size determination in axon_ram_probe()
axonram: Delete an error message for a failed memory allocation in axon_ram_probe()
powerpc/powernv/npu: Move tlb flush before launching ATSD
powerpc/macintosh: constify wf_sensor_ops structures
powerpc/iommu: Use permission-specific DEVICE_ATTR variants
powerpc/eeh: Delete an error out of memory message at init time
powerpc/mm: Use seq_putc() in two functions
macintosh: Convert to using %pOF instead of full_name
...
For understanding how the workload maps to memory channels and hardware
behavior, it's very important to collect address maps with physical
addresses. For example, 3D XPoint access can only be found by filtering
the physical address.
Add a new sample type for physical address.
perf already has a facility to collect data virtual address. This patch
introduces a function to convert the virtual address to physical address.
The function is quite generic and can be extended to any architecture as
long as a virtual address is provided.
- For kernel direct mapping addresses, virt_to_phys is used to convert
the virtual addresses to physical address.
- For user virtual addresses, __get_user_pages_fast is used to walk the
pages tables for user physical address.
- This does not work for vmalloc addresses right now. These are not
resolved, but code to do that could be added.
The new sample type requires collecting the virtual address. The
virtual address will not be output unless SAMPLE_ADDR is applied.
For security, the physical address can only be exposed to root or
privileged user.
Tested-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Kan Liang <kan.liang@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: mpe@ellerman.id.au
Link: http://lkml.kernel.org/r/1503967969-48278-1-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Bring in the commit to rename find_linux_pte_or_hugepte() which touches
arch and KVM code, and might need to be merged with the kvmppc tree to
avoid conflicts.
Add newer helpers to make the function usage simpler. It is always
recommended to use find_current_mm_pte() for walking the page table.
If we cannot use find_current_mm_pte(), it should be documented why
the said usage of __find_linux_pte() is safe against a parallel THP
split.
For now we have KVM code using __find_linux_pte(). This is because kvm
code ends up calling __find_linux_pte() in real mode with MSR_EE=0 but
with PACA soft_enabled = 1. We may want to fix that later and make
sure we keep the MSR_EE and PACA soft_enabled in sync. When we do that
we can switch kvm to use find_linux_pte().
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
nest_imc_refc is a reference count struct, used to track number of
active perf sessions using the nest units.
Currently the code accesses nest_imc_refc using node_id, which is
incorrect, the array is indexed by node number. Meaning in the case of
sparse node ids we index off the end of the array.
Fix it to use get_nest_pmu_ref() which uses the existing per-cpu
variable local_nest_imc_refc.
Fixes: 885dcd709b ('powerpc/perf: Add nest IMC PMU support')
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
[mpe: Tweak change log]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In a multi node system with discontiguous node ids, nest event values
are not showing up properly. eg. lscpu output:
NUMA node0 CPU(s): 0-15
NUMA node8 CPU(s): 16-31
Nest event values on such systems can be counted on CPUs <= 15:
$./perf stat -e 'nest_powerbus0_imc/PM_PB_CYC/' -C 0-14 -I 1000 sleep 1000
# time counts unit events
1.000294577 30,17,24,42,880 nest_powerbus0_imc/PM_PB_CYC/
But not on CPUs >= 16:
$./perf stat -e 'nest_powerbus0_imc/PM_PB_CYC/' -C 16-28 -I 1000 sleep 1000
# time counts unit events
1.000049902 <not supported> nest_powerbus0_imc/PM_PB_CYC/
This is because, when fetching the reference count, the node id (which
may be sparse) is used as the array index, not the node number (which
is 0 based and contiguous).
Fix it by using the node number as the array index.
$./perf stat -e 'nest_powerbus0_imc/PM_PB_CYC/' -C 16-28 -I 1000 sleep 1000
# time counts unit events
1.000241961 26,12,35,28,704 nest_powerbus0_imc/PM_PB_CYC/
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
[mpe: Change log tweaks for clarity and brevity]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This function is not called with the nest_init_lock held, and it also
unlocks the nest_init_lock immediately below, so it's fairly clear
that this is a typo and should be locking the lock.
Fixes: 885dcd709b ("powerpc/perf: Add nest IMC PMU support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add couple of more events (PM_LD_MISS_L1 and PM_BR_2PATH) to
power9 event list and power9_event_alternatives array (these
events can be counted in more than one PMC).
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
There are some hardware events on Power systems which only count when
the processor is not idle, and there are some fixed-function counters
which count such events. For example, the "run cycles" event counts
cycles when the processor is not idle. If the user asks to count
cycles, we can use "run cycles" if this is a per-task event, since the
processor is running when the task is running, by definition. We can't
use "run cycles" if the user asks for "cycles" on a system-wide
counter.
Currently in power8 this check is done using PPMU_ONLY_COUNT_RUN flag
in power8_get_alternatives() function. Based on the flag, events are
switched if needed. This function should also be enabled in power9, so
factor out the code to isa207_get_alternatives().
Fixes: efe881afdd ('powerpc/perf: Factor out event_alternative function')
Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Commit 20dd4c624d ('powerpc/perf: Fix SDAR_MODE value for continous
sampling on Power9') set the default sdar_mode value in MMCRA[SDAR_MODE]
to be used as 0b01 (Update on TLB miss). And this value is set if sdar_mode
from event is zero, or we are in continous sampling mode in power9 dd1.
But it is preferred to have the sdar_mode value for power9 as
0b10 (Update on dcache miss) for better sampling updates instead
of 0b01 (Update on TLB miss).
From Anton:
Using a bandwidth test case with a 1MB footprint, I profiled cycles and
chose TLB updates of the SDAR:
$ perf record -d -e r000400000000001E:u ./bw2001 1M
^
SDAR TLB
$ perf report -D | grep PERF_RECORD_SAMPLE | sed 's/.*addr: //' | sort -u | wc -l
4
I get 4 unique addresses. If I ran with dcache misses:
$ perf record -d -e r000800000000001E:u ./bw2001 1M
^
SDAR dcache miss
$ perf report -D|grep PERF_RECORD_SAMPLE| sed 's/.*addr: //'|sort -u | wc -l
5217
I get 5217 unique addresses. No surprises here, but it does show why
TLB misses is the wrong event to default to - we get very little useful
information out of it.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Acked-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add support to register Thread In-Memory Collection PMU counters.
Patch adds thread IMC specific data structures, along with memory
init functions and CPU hotplug support.
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
Signed-off-by: Hemant Kumar <hemant@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add support to register Core In-Memory Collection PMU counters.
Patch adds core IMC specific data structures, along with memory
init functions and CPU hotplug support.
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
Signed-off-by: Hemant Kumar <hemant@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add support to register Nest In-Memory Collection PMU counters.
Patch adds a new device file called "imc-pmu.c" under powerpc/perf
folder to contain all the device PMU functions.
Device tree parser code added to parse the PMU events information
and create sysfs event attributes for the PMU.
Cpumask attribute added along with Cpu hotplug online/offline functions
specific for nest PMU. A new state "CPUHP_AP_PERF_POWERPC_NEST_IMC_ONLINE"
added for the cpu hotplug callbacks. Error handle path frees the memory
and unregisters the CPU hotplug callbacks.
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
Signed-off-by: Hemant Kumar <hemant@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Similar to POWER8, POWER9 can count run cycles and run instructions
completed on more than one PMU.
Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In case of continous sampling (non-marked), the code currently
sets MMCRA[SDAR_MODE] to 0b01 (Update on TLB miss) for Power9 DD1.
On DD2 and later it copies the sdar_mode value from the event code,
which for most events is 0b00 (No updates).
However we must set a non-zero value for SDAR_MODE when doing
continuous sampling, so honor the event code, unless it's zero, in
which case we use use 0b01 (Update on TLB miss).
Fixes: 78b4416aa2 ("powerpc/perf: Handle sdar_mode for marked event in power9")
Cc: stable@vger.kernel.org # v4.11+
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Merge our fixes branch, a few of them are tripping people up while
working on top of next, and we also have a dependency between the CXL
fixes and new CXL code we want to merge into next.
On POWER9 SMT8 the 24x7 API returns two result elements for physical core
and virtual CPU events and we need to add their counts to get the final
result.
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
POWER9 introduces a new version of the hypervisor API to access the 24x7
perf counters. The new version changed some of the structures used for
requests and results.
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
There's an H24x7_DATA_BUFFER_SIZE constant, so use it in init_24x7_request.
There's also an HV_PERF_DOMAIN_MAX constant, so use it in
h_24x7_event_init. This makes the comment above the check redundant,
so remove it.
In add_event_to_24x7_request, a statement is terminated with a comma
instead of a semicolon. Fix it.
In hv-24x7.h, improve comments in struct hv_24x7_result.
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The H_GET_24X7_CATALOG_PAGE hcall can return a signed error code, so fix
this in the code.
The H_GET_24X7_DATA hcall can return a signed error code, so fix this in
the code. Also, don't truncate it to 32 bit to use as return value for
make_24x7_request. In case of error h_24x7_event_commit_txn passes that
return value to generic code, so it should be a proper errno. The other
caller of make_24x7_request is single_24x7_request, whose callers don't
actually care which error code is returned so they are not affected by this
change.
Finally, h_24x7_get_value doesn't use the error code from
single_24x7_request, so there's no need to store it.
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
make_24x7_request already calls log_24x7_hcall if it fails, so callers
don't have to do it again.
In fact, since the latter is now only called from the former, there's no
need for a separate log_24x7_hcall anymore so remove it.
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
hv-24x7.h has a comment mentioning that result_buffer->results can't be
indexed as a normal array because it may contain results of variable sizes,
so fix the loop in h_24x7_event_commit_txn to take the variation into
account when iterating through results.
Another problem in that loop is that it sets h24x7hw->events[i] to NULL.
This assumes that only the i'th result maps to the i'th request, but that
is not guaranteed to be true. We need to leave the event in the array so
that we don't dereference a NULL pointer in case more than one result maps
to one request.
We still assume that each result has only one result element, so warn if
that assumption is violated.
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
request_buffer can hold 254 requests, so if it already has that number of
entries we can't add a new one.
Also, define constant to show where the number comes from.
Fixes: e3ee15dc5d ("powerpc/perf/hv-24x7: Define add_event_to_24x7_request()")
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
H_GET_24X7_CATALOG_PAGE needs to be passed the version number obtained from
the first catalog page obtained previously. This is a 64 bit number, but
create_events_from_catalog truncates it to 32-bit.
This worked on POWER8, but POWER9 actually uses the upper bits so the call
fails with H_P3 because the hypervisor doesn't recognize the version.
This patch also adds the hcall return code to the error message, which is
helpful when debugging the problem.
Fixes: 5c5cd7b502 ("powerpc/perf/hv-24x7: parse catalog and populate sysfs with events")
Reviewed-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Correct "branch" event code of Power9 is "r4d05e". Replace the current
"branch" event code with "r4d05e" and add a hack to use "r10012" as
event code for Power9 DD1.
Fixes: d89f473ff6 ("powerpc/perf: Fix PM_BRU_CMPL event code for power9")
Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When a kthread calls call_usermodehelper() the steps are:
1. allocate current->mm
2. load_elf_binary()
3. populate current->thread.regs
While doing this, interrupts are not disabled. If there is a perf
interrupt in the middle of this process (i.e. step 1 has completed
but not yet reached to step 3) and if perf tries to read userspace
regs, kernel oops with following log:
Unable to handle kernel paging request for data at address 0x00000000
Faulting instruction address: 0xc0000000000da0fc
...
Call Trace:
perf_output_sample_regs+0x6c/0xd0
perf_output_sample+0x4e4/0x830
perf_event_output_forward+0x64/0x90
__perf_event_overflow+0x8c/0x1e0
record_and_restart+0x220/0x5c0
perf_event_interrupt+0x2d8/0x4d0
performance_monitor_exception+0x54/0x70
performance_monitor_common+0x158/0x160
--- interrupt: f01 at avtab_search_node+0x150/0x1a0
LR = avtab_search_node+0x100/0x1a0
...
load_elf_binary+0x6e8/0x15a0
search_binary_handler+0xe8/0x290
do_execveat_common.isra.14+0x5f4/0x840
call_usermodehelper_exec_async+0x170/0x210
ret_from_kernel_thread+0x5c/0x7c
Fix it by setting abi to PERF_SAMPLE_REGS_ABI_NONE when userspace
pt_regs are not set.
Fixes: ed4a4ef85c ("powerpc/perf: Add support for sampling interrupt register state")
Cc: stable@vger.kernel.org # v4.7+
Signed-off-by: Ravi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Commit 8d911904f3 ('powerpc/perf: Add restrictions to PMC5 in power9 DD1')
was added to restrict the use of PMC5 in Power9 DD1. Intention was to disable
the use of PMC5 using raw event code. But instead of updating the
power9_isa207_pmu structure (used on DD1), the commit incorrectly updated the
power9_pmu structure. Fix it.
Fixes: 8d911904f3 ("powerpc/perf: Add restrictions to PMC5 in power9 DD1")
Reported-by: Shriya <shriyak@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Tested-by: Shriya <shriyak@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Patch add "mem_access" event to sysfs. This as-is not a raw event
supported by Power8 pmu. Instead, it is formed based on
raw event encoding specificed in isa207-common.h.
Primary PMU event used here is PM_MRK_INST_CMPL.
This event tracks only the completed marked instructions.
Random sampling mode (MMCRA[SM]) with Random Instruction
Sampling (RIS) is enabled to mark type of instructions.
With Random sampling in RLS mode with PM_MRK_INST_CMPL event,
the LDST /DATA_SRC fields in SIER identifies the memory
hierarchy level (eg: L1, L2 etc) statisfied a data-cache
miss for a marked instruction.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Patch to export SIER bits to userspace via
perf_mem_data_src and perf_sample_data struct.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Patch to export SIER bits to userspace via
perf_mem_data_src and perf_sample_data struct.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Threshold feature when used with MMCRA [Threshold Event Counter Event],
MMCRA[Threshold Start event] and MMCRA[Threshold End event] will update
MMCRA[Threashold Event Counter Exponent] and MMCRA[Threshold Event
Counter Multiplier] with the corresponding threshold event count values.
Patch to export MMCRA[TECX/TECM] to userspace in 'weight' field of
struct perf_sample_data.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The LDST field and DATA_SRC in SIER identifies the memory hierarchy level
(eg: L1, L2 etc), from which a data-cache miss for a marked instruction
was satisfied. Use the 'perf_mem_data_src' object to export this
hierarchy level to user space.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
MMCRA[SDAR_MODE] specifices how the SDAR should be updated in
continous sampling mode. On P9 it must be set to 0b00 when
MMCRA[63] is set.
Fixes: c7c3f568be ('powerpc/perf: macros for power9 format encoding')
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Power9 DD1 do not support PMU_HAS_SIER flag and sdsync in
perf_get_data_addr() defaults to MMCRA_SDSYNC which is wrong. Since
power9 MMCRA does not support SDSYNC bit, patch includes PPMU_NO_SIAR
flag to the check and set the sdsync with MMCRA_SAMPLE_ENABLE;
Fixes: 27593d72c4 ("powerpc/perf: Use MSR to report privilege level on P9 DD1")
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We are going to split <linux/sched/task_stack.h> out of <linux/sched.h>, which
will have to be picked up from other headers and a couple of .c files.
Create a trivial placeholder <linux/sched/task_stack.h> file that just
maps to <linux/sched.h> to make this patch obviously correct and
bisectable.
Include the new header in the files that are going to need it.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cleanup to use is_kernel_addr macro.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Since power9 does not support FAB_*_MATCH bits in MMCR1,
avoid these checks for power9. For this, patch factor out
code in isa207_get_constraint() to retain these checks
only for power8.
Patch also updates the comment in power9-pmu raw event
encode layout to remove FAB_*_MATCH.
Finally for power9, patch adds additional check for
threshold events when adding the thresh mask and value in
isa207_get_constraint().
fixes: 7ffd948fae ('powerpc/perf: factor out power8 pmu functions')
fixes: 18201b2042 ('powerpc/perf: power9 raw event format encoding')
Signed-off-by: Ravi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
PMC5 on POWER9 DD1 may not provide right counts in all
sampling scenarios, hence use PM_INST_DISP event instead
in PMC2 or PMC3 in preference.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Since PM_INST_DISP include speculative instruction,
based on the workload the dispatch count could vary
considerably. Hence as an alternative, for completed
instruction counting, program the PM_INST_DISP event
to the MMCR* but use Instruction Counter register value.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Since PM_INST_CMPL may not provide right counts in all
sampling scenarios in power9 DD1, instead use PM_INST_DISP.
Patch also update generic instruction sampling with the same.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Factor out the power8 event_alternative function to share
the code with power9.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch has been reworked since RFC version. In the RFC, this patch
was preceded by a patch clearing MSR RI for all PPC32 at all time at
exception prologs. Now MSR RI clearing is done only when this 8xx perf
events functionality is compiled in, it is therefore limited to 8xx
and merged inside this patch.
Other main changes have been to take into account detailed review from
Peter Zijlstra. The instructions counter has been reworked to behave
as a free running counter like the three other counters.
The 8xx has no PMU, however some events can be emulated by other means.
This patch implements the following events (as reported by 'perf list'):
cpu-cycles OR cycles [Hardware event]
instructions [Hardware event]
dTLB-load-misses [Hardware cache event]
iTLB-load-misses [Hardware cache event]
'cycles' event is implemented using the timebase clock. Timebase clock
corresponds to CPU clock divided by 16, so number of cycles is
approximatly 16 times the number of TB ticks
On the 8xx, TLB misses are handled by software. It is therefore
easy to count all TLB misses each time the TLB miss exception is
called.
'instructions' is calculated by using instruction watchpoint counter.
This patch sets counter A to count instructions at address greater
than 0, hence we count all instructions executed while MSR RI bit is
set. The counter is set to the maximum which is 0xffff. Every 65535
instructions, debug instruction breakpoint exception fires. The
exception handler increments a counter in memory which then
represent the upper part of the instruction counter. We therefore
end up with a 48 bits counter. In order to avoid unnecessary overhead
while no perf event is active, this counter is started when the first
event referring to this counter is added, and the counter is stopped
when the last event referring to it is deleted. In order to properly
support breakpoint exceptions, MSR RI bit has to be unset in exception
epilogs in order to avoid breakpoint exceptions during critical
sections during changes to SRR0 and SRR1 would be problematic.
All counters are handled as free running counters.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
SIER and SIAR are not updated correctly for some samples, so force the
use of MSR and regs->nip instead for misc_flag updates. This is done by
adding a new ppmu flag and updating the use_siar logic in
perf_read_regs() to use it, and dropping the PPMU_HAS_SIER flag.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
[mpe: Rename flag to PPMU_NO_SIAR, and also drop PPMU_HAS_SIER]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Use 0x10012 event code for PM_BRU_CMPL event in power9 event list
instead of current 0x40060.
Fixes: 34922527a2 ('powerpc/perf: Add power9 event list macros for generic and cache events')
Cc: stable@vger.kernel.org # v4.9+
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When the state names got added a script was used to add the extra argument
to the calls. The script basically converted the state constant to a
string, but the cleanup to convert these strings into meaningful ones did
not happen.
Replace all the useless strings with 'subsys/xxx/yyy:state' strings which
are used in all the other places already.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Link: http://lkml.kernel.org/r/20161221192112.085444152@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Patch to add macros and contants to support the power9 raw
event encoding format. Couple of functions added since some of the
bits fields like PMCxCOMB and THRESH_CMP has different width and location
within MMCR* in power9.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
