Pull x86 cpu handling changes from Ingo Molnar:
"Bigger changes:
- Intel CPU hardware-enablement: new vector instructions support
(AVX-512), by Fenghua Yu.
- Support the clflushopt instruction and use it in appropriate
places. clflushopt is similar to clflush but with more relaxed
ordering, by Ross Zwisler.
- MSR accessor cleanups, by Borislav Petkov.
- 'forcepae' boot flag for those who have way too much time to spend
on way too old Pentium-M systems and want to live way too
dangerously, by Chris Bainbridge"
* 'x86-cpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, cpu: Add forcepae parameter for booting PAE kernels on PAE-disabled Pentium M
Rename TAINT_UNSAFE_SMP to TAINT_CPU_OUT_OF_SPEC
x86, intel: Make MSR_IA32_MISC_ENABLE bit constants systematic
x86, Intel: Convert to the new bit access MSR accessors
x86, AMD: Convert to the new bit access MSR accessors
x86: Add another set of MSR accessor functions
x86: Use clflushopt in drm_clflush_virt_range
x86: Use clflushopt in drm_clflush_page
x86: Use clflushopt in clflush_cache_range
x86: Add support for the clflushopt instruction
x86, AVX-512: Enable AVX-512 States Context Switch
x86, AVX-512: AVX-512 Feature Detection
Pull perf changes from Ingo Molnar:
"Main changes:
Kernel side changes:
- Add SNB/IVB/HSW client uncore memory controller support (Stephane
Eranian)
- Fix various x86/P4 PMU driver bugs (Don Zickus)
Tooling, user visible changes:
- Add several futex 'perf bench' microbenchmarks (Davidlohr Bueso)
- Speed up thread map generation (Don Zickus)
- Introduce 'perf kvm --list-cmds' command line option for use by
scripts (Ramkumar Ramachandra)
- Print the evsel name in the annotate stdio output, prep to fix
support outputting annotation for multiple events, not just for the
first one (Arnaldo Carvalho de Melo)
- Allow setting preferred callchain method in .perfconfig (Jiri Olsa)
- Show in what binaries/modules 'perf probe's are set (Masami
Hiramatsu)
- Support distro-style debuginfo for uprobe in 'perf probe' (Masami
Hiramatsu)
Tooling, internal changes and fixes:
- Use tid in mmap/mmap2 events to find maps (Don Zickus)
- Record the reason for filtering an address_location (Namhyung Kim)
- Apply all filters to an addr_location (Namhyung Kim)
- Merge al->filtered with hist_entry->filtered in report/hists
(Namhyung Kim)
- Fix memory leak when synthesizing thread records (Namhyung Kim)
- Use ui__has_annotation() in 'report' (Namhyung Kim)
- hists browser refactorings to reuse code accross UIs (Namhyung Kim)
- Add support for the new DWARF unwinder library in elfutils (Jiri
Olsa)
- Fix build race in the generation of bison files (Jiri Olsa)
- Further streamline the feature detection display, trimming it a bit
to show just the libraries detected, using VF=1 gets a more verbose
output, showing the less interesting feature checks as well (Jiri
Olsa).
- Check compatible symtab type before loading dso (Namhyung Kim)
- Check return value of filename__read_debuglink() (Stephane Eranian)
- Move some hashing and fs related code from tools/perf/util/ to
tools/lib/ so that it can be used by more tools/ living utilities
(Borislav Petkov)
- Prepare DWARF unwinding code for using an elfutils alternative
unwinding library (Jiri Olsa)
- Fix DWARF unwind max_stack processing (Jiri Olsa)
- Add dwarf unwind 'perf test' entry (Jiri Olsa)
- 'perf probe' improvements including memory leak fixes, sharing the
intlist class with other tools, uprobes/kprobes code sharing and
use of ref_reloc_sym (Masami Hiramatsu)
- Shorten sample symbol resolving by adding cpumode to struct
addr_location (Arnaldo Carvalho de Melo)
- Fix synthesizing mmaps for threads (Don Zickus)
- Fix invalid output on event group stdio report (Namhyung Kim)
- Fixup header alignment in 'perf sched latency' output (Ramkumar
Ramachandra)
- Fix off-by-one error in 'perf timechart record' argv handling
(Ramkumar Ramachandra)
Tooling, cleanups:
- Remove unused thread__find_map function (Jiri Olsa)
- Remove unused simple_strtoul() function (Ramkumar Ramachandra)
Tooling, documentation updates:
- Update function names in debug messages (Ramkumar Ramachandra)
- Update some code references in design.txt (Ramkumar Ramachandra)
- Clarify load-latency information in the 'perf mem' docs (Andi
Kleen)
- Clarify x86 register naming in 'perf probe' docs (Andi Kleen)"
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (96 commits)
perf tools: Remove unused simple_strtoul() function
perf tools: Update some code references in design.txt
perf evsel: Update function names in debug messages
perf tools: Remove thread__find_map function
perf annotate: Print the evsel name in the stdio output
perf report: Use ui__has_annotation()
perf tools: Fix memory leak when synthesizing thread records
perf tools: Use tid in mmap/mmap2 events to find maps
perf report: Merge al->filtered with hist_entry->filtered
perf symbols: Apply all filters to an addr_location
perf symbols: Record the reason for filtering an address_location
perf sched: Fixup header alignment in 'latency' output
perf timechart: Fix off-by-one error in 'record' argv handling
perf machine: Factor machine__find_thread to take tid argument
perf tools: Speed up thread map generation
perf kvm: introduce --list-cmds for use by scripts
perf ui hists: Pass evsel to hpp->header/width functions explicitly
perf symbols: Introduce thread__find_cpumode_addr_location
perf session: Change header.misc dump from decimal to hex
perf ui/tui: Reuse generic __hpp__fmt() code
...
When CMCI storm persists for a long time(at least beyond predefined
threshold. It's 30 seconds for now), we can watch CMCI storm is
detected immediately after it subsides.
...
Dec 10 22:04:29 kernel: CMCI storm detected: switching to poll mode
Dec 10 22:04:59 kernel: CMCI storm subsided: switching to interrupt mode
Dec 10 22:04:59 kernel: CMCI storm detected: switching to poll mode
Dec 10 22:05:29 kernel: CMCI storm subsided: switching to interrupt mode
...
The problem is that our logic that determines that the storm has
ended is incorrect. We announce the end, re-enable interrupts and
realize that the storm is still going on, so we switch back to
polling mode. Rinse, repeat.
When a storm happens we disable signaling of errors via CMCI and begin
polling machine check banks instead. If we find any logged errors,
then we need to set a per-cpu flag so that our per-cpu tests that
check whether the storm is ongoing will see that errors are still
being logged independently of whether mce_notify_irq() says that the
error has been fully processed.
cmci_clear() is not the right tool to disable a bank. It disables the
interrupt for the bank as desired, but it also clears the bit for
this bank in "mce_banks_owned" so we will skip the bank when polling
(so we fail to see that the storm continues because we stop looking).
New cmci_storm_disable_banks() just disables the interrupt while
allowing polling to continue.
Reported-by: William Dauchy <wdauchy@gmail.com>
Signed-off-by: Chen, Gong <gong.chen@linux.intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch bypass the timer_irq_works() check for hyperv guest since:
- It was guaranteed to work.
- timer_irq_works() may fail sometime due to the lpj calibration were inaccurate
in a hyperv guest or a buggy host.
In the future, we should get the tsc frequency from hypervisor and use preset
lpj instead.
[ hpa: I would prefer to not defer things to "the future" in the future... ]
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: Haiyang Zhang <haiyangz@microsoft.com>
Cc: <stable@vger.kernel.org>
Acked-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Link: http://lkml.kernel.org/r/1393558229-14755-1-git-send-email-jasowang@redhat.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Many Pentium M systems disable PAE but may have a functionally usable PAE
implementation. This adds the "forcepae" parameter which bypasses the boot
check for PAE, and sets the CPU as being PAE capable. Using this parameter
will taint the kernel with TAINT_CPU_OUT_OF_SPEC.
Signed-off-by: Chris Bainbridge <chris.bainbridge@gmail.com>
Link: http://lkml.kernel.org/r/20140307114040.GA4997@localhost
Acked-by: Borislav Petkov <bp@suse.de>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Rename TAINT_UNSAFE_SMP to TAINT_CPU_OUT_OF_SPEC, so we can repurpose
the flag to encompass a wider range of pushing the CPU beyond its
warrany.
Signed-off-by: Dave Jones <davej@fedoraproject.org>
Link: http://lkml.kernel.org/r/20140226154949.GA770@redhat.com
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the amd-uncore code in x86 by using this latter form of callback
registration.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the intel rapl code in x86 by using this latter form of callback
registration.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the intel cacheinfo code in x86 by using this latter form of callback
registration.
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the amd-ibs code in x86 by using this latter form of callback
registration.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
After fixing the CPU hotplug callback registration code, the callbacks
invoked for each online CPU, during the initialization phase in
thermal_throttle_init_device(), can no longer race with the actual CPU
hotplug notifier callbacks (in thermal_throttle_cpu_callback). Hence the
therm_cpu_lock is unnecessary now. Remove it.
Cc: Tony Luck <tony.luck@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the thermal throttle code in x86 by using this latter form of callback
registration.
Cc: Tony Luck <tony.luck@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the mce code in x86 by using this latter form of callback registration.
Cc: Tony Luck <tony.luck@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Instead, the correct and race-free way of performing the callback
registration is:
cpu_notifier_register_begin();
for_each_online_cpu(cpu)
init_cpu(cpu);
/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);
cpu_notifier_register_done();
Fix the uncore code in intel-x86 by using this latter form of callback
registration.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Pull perf fixes from Ingo Molnar:
"Misc smaller fixes"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86: Fix leak in uncore_type_init failure paths
perf machine: Use map as success in ip__resolve_ams
perf symbols: Fix crash in elf_section_by_name
perf trace: Decode architecture-specific signal numbers
Replace somewhat arbitrary constants for bits in MSR_IA32_MISC_ENABLE
with verbose but systematic ones. Add _BIT defines for all the rest
of them, too.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Merge the request/release callbacks which are in a separate branch for
consumption by the gpio folks.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch fixes a compilation problem (unused variable) with the
new SNB/IVB/HSW uncore IMC code.
[ In -v2 we simplify the fix as suggested by Peter Zjilstra. ]
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140311235329.GA28624@quad
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This was an optimization that made memcpy type benchmarks a little
faster on ancient (Circa 1998) IDT Winchip CPUs. In real-life
workloads, it wasn't even noticable, and I doubt anyone is running
benchmarks on 16 year old silicon any more.
Given this code has likely seen very little use over the last decade,
let's just remove it.
Signed-off-by: Dave Jones <davej@fedoraproject.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The error path of uncore_type_init() frees up any allocations
that were made along the way, but it relies upon type->pmus
being set, which only happens if the function succeeds. As
type->pmus remains null in this case, the call to
uncore_type_exit will do nothing.
Moving the assignment earlier will allow us to actually free
those allocations should something go awry.
Signed-off-by: Dave Jones <davej@fedoraproject.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140306172028.GA552@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit:
411cf180fa perf/x86/uncore: fix initialization of cpumask
introduced the function uncore_cpumask_init(), which is only
called in __init intel_uncore_init(). But it is not marked
with __init, which produces the following warning:
WARNING: vmlinux.o(.text+0x2464a): Section mismatch in reference from the function uncore_cpumask_init() to the function .init.text:uncore_cpu_setup()
The function uncore_cpumask_init() references
the function __init uncore_cpu_setup().
This is often because uncore_cpumask_init lacks a __init
annotation or the annotation of uncore_cpu_setup is wrong.
This patch marks uncore_cpumask_init() with __init.
Signed-off-by: Dongsheng Yang <yangds.fnst@cn.fujitsu.com>
Acked-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Link: http://lkml.kernel.org/r/1394013516-4964-1-git-send-email-yangds.fnst@cn.fujitsu.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
x86_64 uses a per_cpu variable kernel_stack to always point to
the thread stack of current. This is where the thread_info is stored
and is accessed from this location even when the irq or exception stack
is in use. This removes the complexity of having to maintain the
thread info on the stack when interrupts are running and having to
copy the preempt_count and other fields to the interrupt stack.
x86_32 uses the old method of copying the thread_info from the thread
stack to the exception stack just before executing the exception.
Having the two different requires #ifdefs and also the x86_32 way
is a bit of a pain to maintain. By converting x86_32 to the same
method of x86_64, we can remove #ifdefs, clean up the x86_32 code
a little, and remove the overhead of the copy.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20110806012354.263834829@goodmis.org
Link: http://lkml.kernel.org/r/20140206144321.852942014@goodmis.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Compiling last minute changes without setting the proper config
options is not really clever.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reported-by: fengguang.wu@intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>
Commit 1aec16967 (x86: Hyperv: Cleanup the irq mess) removed the
ability to build the hyperv stuff as a module. Bring it back.
Reported-by: fengguang.wu@intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>
The vmbus/hyperv interrupt handling is another complete trainwreck and
probably the worst of all currently in tree.
If CONFIG_HYPERV=y then the interrupt delivery to the vmbus happens
via the direct HYPERVISOR_CALLBACK_VECTOR. So far so good, but:
The driver requests first a normal device interrupt. The only reason
to do so is to increment the interrupt stats of that device
interrupt. For no reason it also installs a private flow handler.
We have proper accounting mechanisms for direct vectors, but of
course it's too much effort to add that 5 lines of code.
Aside of that the alloc_intr_gate() is not protected against
reallocation which makes module reload impossible.
Solution to the problem is simple to rip out the whole mess and
implement it correctly.
First of all move all that code to arch/x86/kernel/cpu/mshyperv.c and
merily install the HYPERVISOR_CALLBACK_VECTOR with proper reallocation
protection and use the proper direct vector accounting mechanism.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>
Link: http://lkml.kernel.org/r/20140223212739.028307673@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
We call this "clflush" in /proc/cpuinfo, and have
cpu_has_clflush()... let's be consistent and just call it that.
Cc: Gleb Natapov <gleb@kernel.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Alan Cox <alan@linux.intel.com>
Link: http://lkml.kernel.org/n/tip-mlytfzjkvuf739okyn40p8a5@git.kernel.org
The NUMAQ support seems to be unmaintained, remove it.
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: David Rientjes <rientjes@google.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Link: http://lkml.kernel.org/r/n/530CFD6C.7040705@zytor.com
Add a few comments on the ->add(), ->del() and ->*_txn()
implementation.
Requested-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-he3819318c245j7t5e1e22tr@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Vince "Super Tester" Weaver reported a new round of syscall fuzzing (Trinity) failures,
with perf WARN_ON()s triggering. He also provided traces of the failures.
This is I think the relevant bit:
> pec_1076_warn-2804 [000] d... 147.926153: x86_pmu_disable: x86_pmu_disable
> pec_1076_warn-2804 [000] d... 147.926153: x86_pmu_state: Events: {
> pec_1076_warn-2804 [000] d... 147.926156: x86_pmu_state: 0: state: .R config: ffffffffffffffff ( (null))
> pec_1076_warn-2804 [000] d... 147.926158: x86_pmu_state: 33: state: AR config: 0 (ffff88011ac99800)
> pec_1076_warn-2804 [000] d... 147.926159: x86_pmu_state: }
> pec_1076_warn-2804 [000] d... 147.926160: x86_pmu_state: n_events: 1, n_added: 0, n_txn: 1
> pec_1076_warn-2804 [000] d... 147.926161: x86_pmu_state: Assignment: {
> pec_1076_warn-2804 [000] d... 147.926162: x86_pmu_state: 0->33 tag: 1 config: 0 (ffff88011ac99800)
> pec_1076_warn-2804 [000] d... 147.926163: x86_pmu_state: }
> pec_1076_warn-2804 [000] d... 147.926166: collect_events: Adding event: 1 (ffff880119ec8800)
So we add the insn:p event (fd[23]).
At this point we should have:
n_events = 2, n_added = 1, n_txn = 1
> pec_1076_warn-2804 [000] d... 147.926170: collect_events: Adding event: 0 (ffff8800c9e01800)
> pec_1076_warn-2804 [000] d... 147.926172: collect_events: Adding event: 4 (ffff8800cbab2c00)
We try and add the {BP,cycles,br_insn} group (fd[3], fd[4], fd[15]).
These events are 0:cycles and 4:br_insn, the BP event isn't x86_pmu so
that's not visible.
group_sched_in()
pmu->start_txn() /* nop - BP pmu */
event_sched_in()
event->pmu->add()
So here we should end up with:
0: n_events = 3, n_added = 2, n_txn = 2
4: n_events = 4, n_added = 3, n_txn = 3
But seeing the below state on x86_pmu_enable(), the must have failed,
because the 0 and 4 events aren't there anymore.
Looking at group_sched_in(), since the BP is the leader, its
event_sched_in() must have succeeded, for otherwise we would not have
seen the sibling adds.
But since neither 0 or 4 are in the below state; their event_sched_in()
must have failed; but I don't see why, the complete state: 0,0,1:p,4
fits perfectly fine on a core2.
However, since we try and schedule 4 it means the 0 event must have
succeeded! Therefore the 4 event must have failed, its failure will
have put group_sched_in() into the fail path, which will call:
event_sched_out()
event->pmu->del()
on 0 and the BP event.
Now x86_pmu_del() will reduce n_events; but it will not reduce n_added;
giving what we see below:
n_event = 2, n_added = 2, n_txn = 2
> pec_1076_warn-2804 [000] d... 147.926177: x86_pmu_enable: x86_pmu_enable
> pec_1076_warn-2804 [000] d... 147.926177: x86_pmu_state: Events: {
> pec_1076_warn-2804 [000] d... 147.926179: x86_pmu_state: 0: state: .R config: ffffffffffffffff ( (null))
> pec_1076_warn-2804 [000] d... 147.926181: x86_pmu_state: 33: state: AR config: 0 (ffff88011ac99800)
> pec_1076_warn-2804 [000] d... 147.926182: x86_pmu_state: }
> pec_1076_warn-2804 [000] d... 147.926184: x86_pmu_state: n_events: 2, n_added: 2, n_txn: 2
> pec_1076_warn-2804 [000] d... 147.926184: x86_pmu_state: Assignment: {
> pec_1076_warn-2804 [000] d... 147.926186: x86_pmu_state: 0->33 tag: 1 config: 0 (ffff88011ac99800)
> pec_1076_warn-2804 [000] d... 147.926188: x86_pmu_state: 1->0 tag: 1 config: 1 (ffff880119ec8800)
> pec_1076_warn-2804 [000] d... 147.926188: x86_pmu_state: }
> pec_1076_warn-2804 [000] d... 147.926190: x86_pmu_enable: S0: hwc->idx: 33, hwc->last_cpu: 0, hwc->last_tag: 1 hwc->state: 0
So the problem is that x86_pmu_del(), when called from a
group_sched_in() that fails (for whatever reason), and without x86_pmu
TXN support (because the leader is !x86_pmu), will corrupt the n_added
state.
Reported-and-Tested-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Dave Jones <davej@redhat.com>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/20140221150312.GF3104@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch updates the CBOX PMU filters mapping tables for SNB-EP
and IVT (model 45 and 62 respectively).
The NID umask always comes in addition to another umask.
When set, the NID filter is applied.
The current mapping tables were missing some code/umask
combinations to account for the NID umask. This patch
fixes that.
Cc: mingo@elte.hu
Cc: ak@linux.intel.com
Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140219131018.GA24475@quad
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The current code simply assumes Intel Arch PerfMon v2+ to have
the IA32_PERF_CAPABILITIES MSR; the SDM specifies that we should check
CPUID[1].ECX[15] (aka, FEATURE_PDCM) instead.
This was found by KVM which implements v2+ but didn't provide the
capabilities MSR. Change the code to DTRT; KVM will also implement the
MSR and return 0.
Cc: pbonzini@redhat.com
Reported-by: "Michael S. Tsirkin" <mst@redhat.com>
Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140203132903.GI8874@twins.programming.kicks-ass.net
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
When using BTS on Core i7-4*, I get the below kernel warning.
$ perf record -c 1 -e branches:u ls
Message from syslogd@labpc1501 at Nov 11 15:49:25 ...
kernel:[ 438.317893] Uhhuh. NMI received for unknown reason 31 on CPU 2.
Message from syslogd@labpc1501 at Nov 11 15:49:25 ...
kernel:[ 438.317920] Do you have a strange power saving mode enabled?
Message from syslogd@labpc1501 at Nov 11 15:49:25 ...
kernel:[ 438.317945] Dazed and confused, but trying to continue
Make intel_pmu_handle_irq() take the full exit path when returning early.
Cc: eranian@google.com
Cc: peterz@infradead.org
Cc: mingo@kernel.org
Signed-off-by: Markus Metzger <markus.t.metzger@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1392425048-5309-1-git-send-email-andi@firstfloor.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch is needed because that PMU uses 32-bit free
running counters with no interrupt capabilities.
On SNB/IVB/HSW, we used 20GB/s theoretical peak to calculate
the hrtimer timeout necessary to avoid missing an overflow.
That delay is set to 5s to be on the cautious side.
The SNB IMC uses free running counters, which are handled
via pseudo fixed counters. The SNB IMC PMU implementation
supports an arbitrary number of events, because the counters
are read-only. Therefore it is not possible to track active
counters. Instead we put active events on a linked list which
is then used by the hrtimer handler to update the SW counts.
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: ak@linux.intel.com
Cc: zheng.z.yan@intel.com
Cc: peterz@infradead.org
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1392132015-14521-8-git-send-email-eranian@google.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch makes the hrtimer timeout configurable per PMU
box. Not all counters have necessarily the same width and
rate, thus the default timeout of 60s may need to be adjusted.
This patch adds box->hrtimer_duration. It is set to default
when the box is allocated. It can be overriden when the box
is initialized.
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: ak@linux.intel.com
Cc: zheng.z.yan@intel.com
Cc: peterz@infradead.org
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1392132015-14521-5-git-send-email-eranian@google.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
On certain processors, the uncore PMU boxes may only be
msr-bsed or PCI-based. But in both cases, the cpumask,
suggesting on which CPUs to monitor to get full coverage
of the particular PMU, must be created.
However with the current code base, the cpumask was only
created on processor which had at least one MSR-based
uncore PMU. This patch removes that restriction and
ensures the cpumask is created even when there is no
msr-based PMU. For instance, on SNB client where only
a PCI-based memory controller PMU is supported.
Cc: mingo@elte.hu
Cc: acme@redhat.com
Cc: ak@linux.intel.com
Cc: zheng.z.yan@intel.com
Cc: peterz@infradead.org
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1392132015-14521-2-git-send-email-eranian@google.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The x86 CPU feature modalias handling existed before it was reimplemented
generically. This patch aligns the x86 handling so that it
(a) reuses some more code that is now generic;
(b) uses the generic format for the modalias module metadata entry, i.e., it
now uses 'cpu:type:x86,venVVVVfamFFFFmodMMMM:feature:,XXXX,YYYY' instead of
the 'x86cpu:vendor:VVVV👪FFFF:model:MMMM:feature:,XXXX,YYYY' that was
used before.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If SMAP support is not compiled into the kernel, don't enable SMAP in
CR4 -- in fact, we should clear it, because the kernel doesn't contain
the proper STAC/CLAC instructions for SMAP support.
Found by Fengguang Wu's test system.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Link: http://lkml.kernel.org/r/20140213124550.GA30497@localhost
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Cc: <stable@vger.kernel.org> # v3.7+
A bunch of unknown NMIs have popped up on a Pentium4 recently when booting
into a kdump kernel. This was exposed because the watchdog timer went
from 60 seconds down to 10 seconds (increasing the ability to reproduce
this problem).
What is happening is on boot up of the second kernel (the kdump one),
the previous nmi_watchdogs were enabled on thread 0 and thread 1. The
second kernel only initializes one cpu but the perf counter on thread 1
still counts.
Normally in a kdump scenario, the other cpus are blocking in an NMI loop,
but more importantly their local apics have the performance counters disabled
(iow LVTPC is masked). So any counters that fire are masked and never get
through to the second kernel.
However, on a P4 the local apic is shared by both threads and thread1's PMI
(despite being configured to only interrupt thread1) will generate an NMI on
thread0. Because thread0 knows nothing about this NMI, it is seen as an
unknown NMI.
This would be fine because it is a kdump kernel, strange things happen
what is the big deal about a single unknown NMI.
Unfortunately, the P4 comes with another quirk: clearing the overflow bit
to prevent a stream of NMIs. This is the problem.
The kdump kernel can not execute because of the endless NMIs that happen.
To solve this, I instrumented the p4 perf init code, to walk all the counters
and zero them out (just like a normal reset would).
Now when the counters go off, they do not generate anything and no unknown
NMIs are seen.
I tested this on a P4 we have in our lab. After two or three crashes, I could
normally reproduce the problem. Now after 10 crashes, everything continues
to boot correctly.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140120154115.GZ25953@redhat.com
[ Fixed a stylistic detail. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
On a P4 box stressing perf with:
./perf record -o perf.data ./perf stat -v ./perf bench all
it was noticed that a slew of unknown NMIs would pop out rather quickly.
Painfully debugging this ancient platform, led me to notice cross cpu counter
corruption.
The P4 machine is special in that it has 18 counters, half are used for cpu0
and the other half is for cpu1 (or all 18 if hyperthreading is disabled). But
the splitting of the counters has to be actively managed by the software.
In this particular bug, one of the cpu0 specific counters was being used by
cpu1 and caused all sorts of random unknown nmis.
I am not entirely sure on the corruption path, but what happens is:
o perf schedules a group with p4_pmu_schedule_events()
o inside p4_pmu_schedule_events(), it notices an hwc pointer is being reused
but for a different cpu, so it 'swaps' the config bits and returns the
updated 'assign' array with a _new_ index.
o perf schedules another group with p4_pmu_schedule_events()
o inside p4_pmu_schedule_events(), it notices an hwc pointer is being reused
(the same one as above) but for the _same_ cpu [BUG!!], so it updates the
'assign' array to use the _old_ (wrong cpu) index because the _new_ index is in
an earlier part of the 'assign' array (and hasn't been committed yet).
o perf commits the transaction using the wrong index and corrupts the other cpu
The [BUG!!] is because the 'hwc->config' is updated but not the 'hwc->idx'. So
the check for 'p4_should_swap_ts()' is correct the first time around but
incorrect the second time around (because hwc->config was updated in between).
I think the spirit of perf was to not modify anything until all the
transactions had a chance to 'test' if they would succeed, and if so, commit
atomically. However, P4 breaks this spirit by touching the hwc->config
element.
So my fix is to continue the un-perf like breakage, by assigning hwc->idx to -1
on swap to tell follow up group scheduling to find a new index.
Of course if the transaction fails rolling this back will be difficult, but
that is not different than how the current code works. :-) And I wasn't sure
how much effort to cleanup the code I should do for a platform that is almost
10 years old by now.
Hence the lazy fix.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1391024270-19469-1-git-send-email-dzickus@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The current code forgets to change the CR4 state on the current CPU.
Use on_each_cpu() instead of smp_call_function().
Reported-by: Mark Davies <junk@eslaf.co.uk>
Suggested-by: Mark Davies <junk@eslaf.co.uk>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: fweisbec@gmail.com
Link: http://lkml.kernel.org/n/tip-69efsat90ibhnd577zy3z9gh@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
For additional coverage, BorisO and friends unknowlingly did swap AMD
microcode with Intel microcode blobs in order to see what happens. What
did happen on 32-bit was
[ 5.722656] BUG: unable to handle kernel paging request at be3a6008
[ 5.722693] IP: [<c106d6b4>] load_microcode_amd+0x24/0x3f0
[ 5.722716] *pdpt = 0000000000000000 *pde = 0000000000000000
because there was a valid initrd there but without valid microcode in it
and the container check happened *after* the relocated ramdisk handling
on 32-bit, which was clearly wrong.
While at it, take care of the ramdisk relocation on both 32- and 64-bit
as it is done on both. Also, comment what we're doing because this code
is a bit tricky.
Reported-and-tested-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/1391460104-7261-1-git-send-email-bp@alien8.de
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
There was a large ebizzy performance regression that was
bisected to commit 611ae8e3 (x86/tlb: enable tlb flush range
support for x86). The problem was related to the
tlb_flushall_shift tuning for IvyBridge which was altered. The
problem is that it is not clear if the tuning values for each
CPU family is correct as the methodology used to tune the values
is unclear.
This patch uses a conservative tlb_flushall_shift value for all
CPU families except IvyBridge so the decision can be revisited
if any regression is found as a result of this change.
IvyBridge is an exception as testing with one methodology
determined that the value of 2 is acceptable. Details are in
the changelog for the patch "x86: mm: Change tlb_flushall_shift
for IvyBridge".
One important aspect of this to watch out for is Xen. The
original commit log mentioned large performance gains on Xen.
It's possible Xen is more sensitive to this value if it flushes
small ranges of pages more frequently than workloads on bare
metal typically do.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alex Shi <alex.shi@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-dyzMww3fqugnhbhgo6Gxmtkw@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There was a large performance regression that was bisected to
commit 611ae8e3 ("x86/tlb: enable tlb flush range support for
x86"). This patch simply changes the default balance point
between a local and global flush for IvyBridge.
In the interest of allowing the tests to be reproduced, this
patch was tested using mmtests 0.15 with the following
configurations
configs/config-global-dhp__tlbflush-performance
configs/config-global-dhp__scheduler-performance
configs/config-global-dhp__network-performance
Results are from two machines
Ivybridge 4 threads: Intel(R) Core(TM) i3-3240 CPU @ 3.40GHz
Ivybridge 8 threads: Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz
Page fault microbenchmark showed nothing interesting.
Ebizzy was configured to run multiple iterations and threads.
Thread counts ranged from 1 to NR_CPUS*2. For each thread count,
it ran 100 iterations and each iteration lasted 10 seconds.
Ivybridge 4 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 6395.44 ( 0.00%) 6789.09 ( 6.16%)
Mean 2 7012.85 ( 0.00%) 8052.16 ( 14.82%)
Mean 3 6403.04 ( 0.00%) 6973.74 ( 8.91%)
Mean 4 6135.32 ( 0.00%) 6582.33 ( 7.29%)
Mean 5 6095.69 ( 0.00%) 6526.68 ( 7.07%)
Mean 6 6114.33 ( 0.00%) 6416.64 ( 4.94%)
Mean 7 6085.10 ( 0.00%) 6448.51 ( 5.97%)
Mean 8 6120.62 ( 0.00%) 6462.97 ( 5.59%)
Ivybridge 8 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 7336.65 ( 0.00%) 7787.02 ( 6.14%)
Mean 2 8218.41 ( 0.00%) 9484.13 ( 15.40%)
Mean 3 7973.62 ( 0.00%) 8922.01 ( 11.89%)
Mean 4 7798.33 ( 0.00%) 8567.03 ( 9.86%)
Mean 5 7158.72 ( 0.00%) 8214.23 ( 14.74%)
Mean 6 6852.27 ( 0.00%) 7952.45 ( 16.06%)
Mean 7 6774.65 ( 0.00%) 7536.35 ( 11.24%)
Mean 8 6510.50 ( 0.00%) 6894.05 ( 5.89%)
Mean 12 6182.90 ( 0.00%) 6661.29 ( 7.74%)
Mean 16 6100.09 ( 0.00%) 6608.69 ( 8.34%)
Ebizzy hits the worst case scenario for TLB range flushing every
time and it shows for these Ivybridge CPUs at least that the
default choice is a poor on. The patch addresses the problem.
Next was a tlbflush microbenchmark written by Alex Shi at
http://marc.info/?l=linux-kernel&m=133727348217113 . It
measures access costs while the TLB is being flushed. The
expectation is that if there are always full TLB flushes that
the benchmark would suffer and it benefits from range flushing
There are 320 iterations of the test per thread count. The
number of entries is randomly selected with a min of 1 and max
of 512. To ensure a reasonably even spread of entries, the full
range is broken up into 8 sections and a random number selected
within that section.
iteration 1, random number between 0-64
iteration 2, random number between 64-128 etc
This is still a very weak methodology. When you do not know
what are typical ranges, random is a reasonable choice but it
can be easily argued that the opimisation was for smaller ranges
and an even spread is not representative of any workload that
matters. To improve this, we'd need to know the probability
distribution of TLB flush range sizes for a set of workloads
that are considered "common", build a synthetic trace and feed
that into this benchmark. Even that is not perfect because it
would not account for the time between flushes but there are
limits of what can be reasonably done and still be doing
something useful. If a representative synthetic trace is
provided then this benchmark could be revisited and the shift values retuned.
Ivybridge 4 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 10.50 ( 0.00%) 10.50 ( 0.03%)
Mean 2 17.59 ( 0.00%) 17.18 ( 2.34%)
Mean 3 22.98 ( 0.00%) 21.74 ( 5.41%)
Mean 5 47.13 ( 0.00%) 46.23 ( 1.92%)
Mean 8 43.30 ( 0.00%) 42.56 ( 1.72%)
Ivybridge 8 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 9.45 ( 0.00%) 9.36 ( 0.93%)
Mean 2 9.37 ( 0.00%) 9.70 ( -3.54%)
Mean 3 9.36 ( 0.00%) 9.29 ( 0.70%)
Mean 5 14.49 ( 0.00%) 15.04 ( -3.75%)
Mean 8 41.08 ( 0.00%) 38.73 ( 5.71%)
Mean 13 32.04 ( 0.00%) 31.24 ( 2.49%)
Mean 16 40.05 ( 0.00%) 39.04 ( 2.51%)
For both CPUs, average access time is reduced which is good as
this is the benchmark that was used to tune the shift values in
the first place albeit it is now known *how* the benchmark was
used.
The scheduler benchmarks were somewhat inconclusive. They
showed gains and losses and makes me reconsider how stable those
benchmarks really are or if something else might be interfering
with the test results recently.
Network benchmarks were inconclusive. Almost all results were
flat except for netperf-udp tests on the 4 thread machine.
These results were unstable and showed large variations between
reboots. It is unknown if this is a recent problems but I've
noticed before that netperf-udp results tend to vary.
Based on these results, changing the default for Ivybridge seems
like a logical choice.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Alex Shi <alex.shi@linaro.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-cqnadffh1tiqrshthRj3Esge@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Bisection between 3.11 and 3.12 fingered commit 9824cf97 ("mm:
vmstats: tlb flush counters") to cause overhead problems.
The counters are undeniably useful but how often do we really
need to debug TLB flush related issues? It does not justify
taking the penalty everywhere so make it a debugging option.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Alex Shi <alex.shi@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-XzxjntugxuwpxXhcrxqqh53b@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Here's the big driver core and sysfs patch set for 3.14-rc1.
There's a lot of work here moving sysfs logic out into a "kernfs" to
allow other subsystems to also have a virtual filesystem with the same
attributes of sysfs (handle device disconnect, dynamic creation /
removal as needed / unneeded, etc. This is primarily being done for
the cgroups filesystem, but the goal is to also move debugfs to it when
it is ready, solving all of the known issues in that filesystem as well.
The code isn't completed yet, but all should be stable now (there is a
big section that was reverted due to problems found when testing.)
There's also some other smaller fixes, and a driver core addition that
allows for a "collection" of objects, that the DRM people will be using
soon (it's in this tree to make merges after -rc1 easier.)
All of this has been in linux-next with no reported issues.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iEYEABECAAYFAlLdh0cACgkQMUfUDdst+ylv4QCfeDKDgLo4LsaBIIrFSxLoH/c7
UUsAoMPRwA0h8wy+BQcJAg4H4J4maKj3
=0pc0
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core / sysfs patches from Greg KH:
"Here's the big driver core and sysfs patch set for 3.14-rc1.
There's a lot of work here moving sysfs logic out into a "kernfs" to
allow other subsystems to also have a virtual filesystem with the same
attributes of sysfs (handle device disconnect, dynamic creation /
removal as needed / unneeded, etc)
This is primarily being done for the cgroups filesystem, but the goal
is to also move debugfs to it when it is ready, solving all of the
known issues in that filesystem as well. The code isn't completed
yet, but all should be stable now (there is a big section that was
reverted due to problems found when testing)
There's also some other smaller fixes, and a driver core addition that
allows for a "collection" of objects, that the DRM people will be
using soon (it's in this tree to make merges after -rc1 easier)
All of this has been in linux-next with no reported issues"
* tag 'driver-core-3.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (113 commits)
kernfs: associate a new kernfs_node with its parent on creation
kernfs: add struct dentry declaration in kernfs.h
kernfs: fix get_active failure handling in kernfs_seq_*()
Revert "kernfs: fix get_active failure handling in kernfs_seq_*()"
Revert "kernfs: replace kernfs_node->u.completion with kernfs_root->deactivate_waitq"
Revert "kernfs: remove KERNFS_ACTIVE_REF and add kernfs_lockdep()"
Revert "kernfs: remove KERNFS_REMOVED"
Revert "kernfs: restructure removal path to fix possible premature return"
Revert "kernfs: invoke kernfs_unmap_bin_file() directly from __kernfs_remove()"
Revert "kernfs: remove kernfs_addrm_cxt"
Revert "kernfs: make kernfs_get_active() block if the node is deactivated but not removed"
Revert "kernfs: implement kernfs_{de|re}activate[_self]()"
Revert "kernfs, sysfs, driver-core: implement kernfs_remove_self() and its wrappers"
Revert "pci: use device_remove_file_self() instead of device_schedule_callback()"
Revert "scsi: use device_remove_file_self() instead of device_schedule_callback()"
Revert "s390: use device_remove_file_self() instead of device_schedule_callback()"
Revert "sysfs, driver-core: remove unused {sysfs|device}_schedule_callback_owner()"
Revert "kernfs: remove unnecessary NULL check in __kernfs_remove()"
kernfs: remove unnecessary NULL check in __kernfs_remove()
drivers/base: provide an infrastructure for componentised subsystems
...
Pull x86 kernel address space randomization support from Peter Anvin:
"This enables kernel address space randomization for x86"
* 'x86-kaslr-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, kaslr: Clarify RANDOMIZE_BASE_MAX_OFFSET
x86, kaslr: Remove unused including <linux/version.h>
x86, kaslr: Use char array to gain sizeof sanity
x86, kaslr: Add a circular multiply for better bit diffusion
x86, kaslr: Mix entropy sources together as needed
x86/relocs: Add percpu fixup for GNU ld 2.23
x86, boot: Rename get_flags() and check_flags() to *_cpuflags()
x86, kaslr: Raise the maximum virtual address to -1 GiB on x86_64
x86, kaslr: Report kernel offset on panic
x86, kaslr: Select random position from e820 maps
x86, kaslr: Provide randomness functions
x86, kaslr: Return location from decompress_kernel
x86, boot: Move CPU flags out of cpucheck
x86, relocs: Add more per-cpu gold special cases
Pull leftover x86 fixes from Ingo Molnar:
"Two leftover fixes that did not make it into v3.13"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86: Add check for number of available vectors before CPU down
x86, cpu, amd: Add workaround for family 16h, erratum 793
Pull x86 RAS changes from Ingo Molnar:
- SCI reporting for other error types not only correctable ones
- GHES cleanups
- Add the functionality to override error reporting agents as some
machines are sporting a new extended error logging capability which,
if done properly in the BIOS, makes a corresponding EDAC module
redundant
- PCIe AER tracepoint severity levels fix
- Error path correction for the mce device init
- MCE timer fix
- Add more flexibility to the error injection (EINJ) debugfs interface
* 'x86-ras-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, mce: Fix mce_start_timer semantics
ACPI, APEI, GHES: Cleanup ghes memory error handling
ACPI, APEI: Cleanup alignment-aware accesses
ACPI, APEI, GHES: Do not report only correctable errors with SCI
ACPI, APEI, EINJ: Changes to the ACPI/APEI/EINJ debugfs interface
ACPI, eMCA: Combine eMCA/EDAC event reporting priority
EDAC, sb_edac: Modify H/W event reporting policy
EDAC: Add an edac_report parameter to EDAC
PCI, AER: Fix severity usage in aer trace event
x86, mce: Call put_device on device_register failure
Pull x86 microcode loader updates from Ingo Molnar:
"There are two main changes in this tree:
- AMD microcode early loading fixes
- some microcode loader source files reorganization"
* 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, microcode: Move to a proper location
x86, microcode, AMD: Fix early ucode loading
x86, microcode: Share native MSR accessing variants
x86, ramdisk: Export relocated ramdisk VA
Pull x86 TLB detection update from Ingo Molnar:
"A single change that extends our TLB cache size detection+reporting
code"
* 'x86-cpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, cpu: Detect more TLB configuration
Pull x86 cleanups from Ingo Molnar:
"Misc cleanups"
* 'x86-cleanups-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, cpu, amd: Fix a shadowed variable situation
um, x86: Fix vDSO build
x86: Delete non-required instances of include <linux/init.h>
x86, realmode: Pointer walk cleanups, pull out invariant use of __pa()
x86/traps: Clean up error exception handler definitions
Pull scheduler changes from Ingo Molnar:
- Add the initial implementation of SCHED_DEADLINE support: a real-time
scheduling policy where tasks that meet their deadlines and
periodically execute their instances in less than their runtime quota
see real-time scheduling and won't miss any of their deadlines.
Tasks that go over their quota get delayed (Available to privileged
users for now)
- Clean up and fix preempt_enable_no_resched() abuse all around the
tree
- Do sched_clock() performance optimizations on x86 and elsewhere
- Fix and improve auto-NUMA balancing
- Fix and clean up the idle loop
- Apply various cleanups and fixes
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (60 commits)
sched: Fix __sched_setscheduler() nice test
sched: Move SCHED_RESET_ON_FORK into attr::sched_flags
sched: Fix up attr::sched_priority warning
sched: Fix up scheduler syscall LTP fails
sched: Preserve the nice level over sched_setscheduler() and sched_setparam() calls
sched/core: Fix htmldocs warnings
sched/deadline: No need to check p if dl_se is valid
sched/deadline: Remove unused variables
sched/deadline: Fix sparse static warnings
m68k: Fix build warning in mac_via.h
sched, thermal: Clean up preempt_enable_no_resched() abuse
sched, net: Fixup busy_loop_us_clock()
sched, net: Clean up preempt_enable_no_resched() abuse
sched/preempt: Fix up missed PREEMPT_NEED_RESCHED folding
sched/preempt, locking: Rework local_bh_{dis,en}able()
sched/clock, x86: Avoid a runtime condition in native_sched_clock()
sched/clock: Fix up clear_sched_clock_stable()
sched/clock, x86: Use a static_key for sched_clock_stable
sched/clock: Remove local_irq_disable() from the clocks
sched/clock, x86: Rewrite cyc2ns() to avoid the need to disable IRQs
...
On AMD family 10h we see following error messages while waking up from
S3 for all non-boot CPUs leading to a failed IBS initialization:
Enabling non-boot CPUs ...
smpboot: Booting Node 0 Processor 1 APIC 0x1
[Firmware Bug]: cpu 1, try to use APIC500 (LVT offset 0) for vector 0x400, but the register is already in use for vector 0xf9 on another cpu
perf: IBS APIC setup failed on cpu #1
process: Switch to broadcast mode on CPU1
CPU1 is up
...
ACPI: Waking up from system sleep state S3
Reason for this is that during suspend the LVT offset for the IBS
vector gets lost and needs to be reinialized while resuming.
The offset is read from the IBSCTL msr. On family 10h the offset needs
to be 1 as offset 0 is used for the MCE threshold interrupt, but
firmware assings it for IBS to 0 too. The kernel needs to reprogram
the vector. The msr is a readonly node msr, but a new value can be
written via pci config space access. The reinitialization is
implemented for family 10h in setup_ibs_ctl() which is forced during
IBS setup.
This patch fixes IBS setup after waking up from S3 by adding
resume/supend hooks for the boot cpu which does the offset
reinitialization.
Marking it as stable to let distros pick up this fix.
Signed-off-by: Robert Richter <rric@kernel.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: <stable@vger.kernel.org> v3.2..
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1389797849-5565-1-git-send-email-rric.net@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Having u32 and struct cpuinfo_x86 * by the same name is not very smart,
although it was ok in this case due to the limited scope of u32 c and it
being used only once in there.
Fix this.
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/1389786735-16751-1-git-send-email-bp@alien8.de
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
This adds the workaround for erratum 793 as a precaution in case not
every BIOS implements it. This addresses CVE-2013-6885.
Erratum text:
[Revision Guide for AMD Family 16h Models 00h-0Fh Processors,
document 51810 Rev. 3.04 November 2013]
793 Specific Combination of Writes to Write Combined Memory Types and
Locked Instructions May Cause Core Hang
Description
Under a highly specific and detailed set of internal timing
conditions, a locked instruction may trigger a timing sequence whereby
the write to a write combined memory type is not flushed, causing the
locked instruction to stall indefinitely.
Potential Effect on System
Processor core hang.
Suggested Workaround
BIOS should set MSR
C001_1020[15] = 1b.
Fix Planned
No fix planned
[ hpa: updated description, fixed typo in MSR name ]
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/20140114230711.GS29865@pd.tnic
Tested-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@amd.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
We've grown a bunch of microcode loader files all prefixed with
"microcode_". They should be under cpu/ because this is strictly
CPU-related functionality so do that and drop the prefix since they're
in their own directory now which gives that prefix. :)
While at it, drop MICROCODE_INTEL_LIB config item and stash the
functionality under CONFIG_MICROCODE_INTEL as it was its only user.
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Aravind Gopalakrishnan <Aravind.Gopalakrishnan@amd.com>
Use a ring-buffer like multi-version object structure which allows
always having a coherent object; we use this to avoid having to
disable IRQs while reading sched_clock() and avoids a problem when
getting an NMI while changing the cyc2ns data.
MAINLINE PRE POST
sched_clock_stable: 1 1 1
(cold) sched_clock: 329841 331312 257223
(cold) local_clock: 301773 310296 309889
(warm) sched_clock: 38375 38247 25280
(warm) local_clock: 100371 102713 85268
(warm) rdtsc: 27340 27289 24247
sched_clock_stable: 0 0 0
(cold) sched_clock: 382634 372706 301224
(cold) local_clock: 396890 399275 399870
(warm) sched_clock: 38194 38124 25630
(warm) local_clock: 143452 148698 129629
(warm) rdtsc: 27345 27365 24307
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/n/tip-s567in1e5ekq2nlyhn8f987r@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
So mce_start_timer() has a 'cpu' argument which is supposed to mean to
start a timer on that cpu. However, the code currently starts a timer on
the *current* cpu the function runs on and causes the sanity-check in
mce_timer_fn to fire:
WARNING: CPU: 0 PID: 0 at arch/x86/kernel/cpu/mcheck/mce.c:1286 mce_timer_fn
because it is running on the wrong cpu.
This was triggered by Prarit Bhargava <prarit@redhat.com> by offlining
all the cpus in succession.
Then, we were fiddling with the CMCI storm settings when starting the
timer whereas there's no need for that - if there's storm happening
on this newly restarted cpu, we're going to be in normal CMCI mode
initially and then when the CMCI interrupt starts firing, we're going to
go to the polling mode with the timer real soon.
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Prarit Bhargava <prarit@redhat.com>
Cc: Tony Luck <tony.luck@intel.com>
Reviewed-by: Chen, Gong <gong.chen@linux.intel.com>
Link: http://lkml.kernel.org/r/1387722156-5511-1-git-send-email-prarit@redhat.com
This patch adds support for the Intel RAPL energy counter
PP1 (Power Plane 1).
On client processors, it usually corresponds to the
energy consumption of the builtin graphic card. That
is why the sysfs event is called energy-gpu.
New event:
- name: power/energy-gpu/
- code: event=0x4
- unit: 2^-32 Joules
On processors without graphics, this should count 0.
The patch only enables this event on client processors.
Reviewed-by: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: ak@linux.intel.com
Cc: acme@redhat.com
Cc: jolsa@redhat.com
Cc: zheng.z.yan@intel.com
Cc: bp@alien8.de
Cc: vincent.weaver@maine.edu
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1389176153-3128-3-git-send-email-eranian@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
None of these files are actually using any __init type directives
and hence don't need to include <linux/init.h>. Most are just a
left over from __devinit and __cpuinit removal, or simply due to
code getting copied from one driver to the next.
[ hpa: undid incorrect removal from arch/x86/kernel/head_32.S ]
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Link: http://lkml.kernel.org/r/1389054026-12947-1-git-send-email-paul.gortmaker@windriver.com
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Pull x86 fixes from Peter Anvin:
"There is a small EFI fix and a big power regression fix in this batch.
My queue also had a fix for downing a CPU when there are insufficient
number of IRQ vectors available, but I'm holding that one for now due
to recent bug reports"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/efi: Don't select EFI from certain special ACPI drivers
x86 idle: Repair large-server 50-watt idle-power regression
Currently SCI is employed to handle corrected errors - memory corrected
errors, more specifically but in fact SCI still can be used to handle
any errors, e.g. uncorrected or even fatal ones if enabled by the BIOS.
Enable logging for those kinds of errors too.
Signed-off-by: Chen, Gong <gong.chen@linux.intel.com>
Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: http://lkml.kernel.org/r/1385363701-12387-1-git-send-email-gong.chen@linux.intel.com
[ Boris: massage commit message, rename function arg. ]
Signed-off-by: Borislav Petkov <bp@suse.de>
Linux 3.10 changed the timing of how thread_info->flags is touched:
x86: Use generic idle loop
(7d1a941731)
This caused Intel NHM-EX and WSM-EX servers to experience a large number
of immediate MONITOR/MWAIT break wakeups, which caused cpuidle to demote
from deep C-states to shallow C-states, which caused these platforms
to experience a significant increase in idle power.
Note that this issue was already present before the commit above,
however, it wasn't seen often enough to be noticed in power measurements.
Here we extend an errata workaround from the Core2 EX "Dunnington"
to extend to NHM-EX and WSM-EX, to prevent these immediate
returns from MWAIT, reducing idle power on these platforms.
While only acpi_idle ran on Dunnington, intel_idle
may also run on these two newer systems.
As of today, there are no other models that are known
to need this tweak.
Link: http://lkml.kernel.org/r/CAJvTdK=%2BaNN66mYpCGgbHGCHhYQAKx-vB0kJSWjVpsNb_hOAtQ@mail.gmail.com
Signed-off-by: Len Brown <len.brown@intel.com>
Link: http://lkml.kernel.org/r/baff264285f6e585df757d58b17788feabc68918.1387403066.git.len.brown@intel.com
Cc: <stable@vger.kernel.org> # 3.12.x, 3.11.x, 3.10.x
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
The EVENT_CONSTRAINT_END() macro defines the end marker as
a constraint with a weight of zero. This was all fine
until we blacklisted the corrupting memory events on
Intel IvyBridge. These events are blacklisted by using
a counter bitmask of zero. Thus, they also get a constraint
weight of zero.
The iteration macro: for_each_constraint tests the weight==0.
Therefore, it was stopping at the first blacklisted event, i.e.,
0xd0. The corrupting events were therefore considered as
unconstrained and were scheduled on any of the generic counters.
This patch fixes the end marker to have a weight of -1. With
this, the blacklisted events get an empty constraint and cannot
be scheduled which is what we want for now.
Signed-off-by: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>
Reviewed-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: ak@linux.intel.com
Cc: jolsa@redhat.com
Cc: zheng.z.yan@intel.com
Link: http://lkml.kernel.org/r/20131204232437.GA10689@starlight
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch adds a call to put_device() when the device_register() call
has failed. This is required so that the last reference to the device is
given up.
Signed-off-by: Levente Kurusa <levex@linux.com>
Link: http://lkml.kernel.org/r/5298F900.9000208@linux.com
Signed-off-by: Borislav Petkov <bp@suse.de>
The RAPL PMU counters do not interrupt on overflow.
Therefore, the kernel needs to poll the counters
to avoid missing an overflow. This patch adds
the hrtimer code to do this.
The timer interval is calculated at boot time
based on the power unit used by the HW.
There is one hrtimer per-cpu to handle the case
of multiple simultaneous use across cores on
the same package + hotplug CPU.
Thanks to Maria Dimakopoulou for her contributions
to this patch especially on the math aspects.
Signed-off-by: Stephane Eranian <eranian@google.com>
Reviewed-by: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
[ Applied 32-bit build fix. ]
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: acme@redhat.com
Cc: jolsa@redhat.com
Cc: zheng.z.yan@intel.com
Cc: bp@alien8.de
Cc: maria.n.dimakopoulou@gmail.com
Link: http://lkml.kernel.org/r/1384275531-10892-5-git-send-email-eranian@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch adds a new uncore PMU to expose the Intel
RAPL energy consumption counters. Up to 3 counters,
each counting a particular RAPL event are exposed.
The RAPL counters are available on Intel SandyBridge,
IvyBridge, Haswell. The server skus add a 3rd counter.
The following events are available and exposed in sysfs:
- power/energy-cores: power consumption of all cores on socket
- power/energy-pkg: power consumption of all cores + LLc cache
- power/energy-dram: power consumption of DRAM (servers only)
For each event both the unit (Joules) and scale (2^-32 J)
is exposed in sysfs for use by perf stat and other tools.
The files are:
/sys/devices/power/events/energy-*.unit
/sys/devices/power/events/energy-*.scale
The RAPL PMU is uncore by nature and is implemented such
that it only works in system-wide mode. Measuring only
one CPU per socket is sufficient. The /sys/devices/power/cpumask
file can be used by tools to figure out which CPUs to monitor
by default. For instance, on a 2-socket system, 2 CPUs
(one on each socket) will be shown.
All the counters measure in the same unit (exposed via sysfs).
The perf_events API exposes all RAPL counters as 64-bit integers
counting in unit of 1/2^32 Joules (about 0.23 nJ). User level tools
must convert the counts by multiplying them by 2^-32 to obtain
Joules. The reason for this is that the kernel avoids
doing floating point math whenever possible because it is
expensive (user floating-point state must be saved). The method
used avoids kernel floating-point usage. There is no loss of
precision. Thanks to PeterZ for suggesting this approach.
To convert the raw count in Watt:
W = C * 2.3 / (1e10 * time)
or ldexp(C, -32).
RAPL PMU is a new standalone PMU which registers with the
perf_event core subsystem. The PMU type (attr->type) is
dynamically allocated and is available from /sys/device/power/type.
Sampling is not supported by the RAPL PMU. There is no
privilege level filtering either.
Signed-off-by: Stephane Eranian <eranian@google.com>
Reviewed-by: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: acme@redhat.com
Cc: jolsa@redhat.com
Cc: zheng.z.yan@intel.com
Cc: bp@alien8.de
Link: http://lkml.kernel.org/r/1384275531-10892-4-git-send-email-eranian@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull trivial tree updates from Jiri Kosina:
"Usual earth-shaking, news-breaking, rocket science pile from
trivial.git"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (23 commits)
doc: usb: Fix typo in Documentation/usb/gadget_configs.txt
doc: add missing files to timers/00-INDEX
timekeeping: Fix some trivial typos in comments
mm: Fix some trivial typos in comments
irq: Fix some trivial typos in comments
NUMA: fix typos in Kconfig help text
mm: update 00-INDEX
doc: Documentation/DMA-attributes.txt fix typo
DRM: comment: `halve' -> `half'
Docs: Kconfig: `devlopers' -> `developers'
doc: typo on word accounting in kprobes.c in mutliple architectures
treewide: fix "usefull" typo
treewide: fix "distingush" typo
mm/Kconfig: Grammar s/an/a/
kexec: Typo s/the/then/
Documentation/kvm: Update cpuid documentation for steal time and pv eoi
treewide: Fix common typo in "identify"
__page_to_pfn: Fix typo in comment
Correct some typos for word frequency
clk: fixed-factor: Fix a trivial typo
...
Pull x86 RAS changes from Ingo Molnar:
"The biggest change adds support for Intel 'CPER' (UEFI Common Platform
Error Record) error logging, which builds upon an enhanced error
logging mechanism available on Xeon processors.
Full description is here:
http://www.intel.com/content/www/us/en/architecture-and-technology/enhanced-mca-logging-xeon-paper.html
This change provides a module (and support code) to check for an
extended error log and prints extra details about the error on the
console"
* 'x86-mce-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
ACPI, x86: Fix extended error log driver to depend on CONFIG_X86_LOCAL_APIC
dmi: Avoid unaligned memory access in save_mem_devices()
Move cper.c from drivers/acpi/apei to drivers/firmware/efi
EDAC, GHES: Update ghes error record info
ACPI, APEI, CPER: Cleanup CPER memory error output format
ACPI, APEI, CPER: Enhance memory reporting capability
ACPI, APEI, CPER: Add UEFI 2.4 support for memory error
DMI: Parse memory device (type 17) in SMBIOS
ACPI, x86: Extended error log driver for x86 platform
bitops: Introduce a more generic BITMASK macro
ACPI, CPER: Update cper info
ACPI, APEI, CPER: Fix status check during error printing
Pull x86/hyperv changes from Ingo Molnar:
"These changes enable Linux guests to boot as 'Modern VM' guest kernels
on MS-Hyperv hosts"
* 'x86-hyperv-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, hyperv: Move a variable to avoid an unused variable warning
x86, hyperv: Fix build error due to missing <asm/apic.h> include
x86, hyperv: Correctly guard the local APIC calibration code
x86, hyperv: Get the local APIC timer frequency from the hypervisor
Pull x86 cpu changes from Ingo Molnar:
"The biggest change that stands out is the increase of the
CONFIG_NR_CPUS range from 4096 to 8192 - as real hardware out there
already went beyond 4k CPUs ...
We only allow more than 512 CPUs if offstack cpumasks are enabled.
CONFIG_MAXSMP=y remains to be the 'you are nuts!' extreme testcase,
which now means a max of 8192 CPUs"
* 'x86-cpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/cpu: Increase max CPU count to 8192
x86/cpu: Allow higher NR_CPUS values
x86/cpu: Always print SMP information in /proc/cpuinfo
x86/cpu: Track legacy CPU model data only on 32-bit kernels
Pull scheduler changes from Ingo Molnar:
"The main changes in this cycle are:
- (much) improved CONFIG_NUMA_BALANCING support from Mel Gorman, Rik
van Riel, Peter Zijlstra et al. Yay!
- optimize preemption counter handling: merge the NEED_RESCHED flag
into the preempt_count variable, by Peter Zijlstra.
- wait.h fixes and code reorganization from Peter Zijlstra
- cfs_bandwidth fixes from Ben Segall
- SMP load-balancer cleanups from Peter Zijstra
- idle balancer improvements from Jason Low
- other fixes and cleanups"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (129 commits)
ftrace, sched: Add TRACE_FLAG_PREEMPT_RESCHED
stop_machine: Fix race between stop_two_cpus() and stop_cpus()
sched: Remove unnecessary iteration over sched domains to update nr_busy_cpus
sched: Fix asymmetric scheduling for POWER7
sched: Move completion code from core.c to completion.c
sched: Move wait code from core.c to wait.c
sched: Move wait.c into kernel/sched/
sched/wait: Fix __wait_event_interruptible_lock_irq_timeout()
sched: Avoid throttle_cfs_rq() racing with period_timer stopping
sched: Guarantee new group-entities always have weight
sched: Fix hrtimer_cancel()/rq->lock deadlock
sched: Fix cfs_bandwidth misuse of hrtimer_expires_remaining
sched: Fix race on toggling cfs_bandwidth_used
sched: Remove extra put_online_cpus() inside sched_setaffinity()
sched/rt: Fix task_tick_rt() comment
sched/wait: Fix build breakage
sched/wait: Introduce prepare_to_wait_event()
sched/wait: Add ___wait_cond_timeout() to wait_event*_timeout() too
sched: Remove get_online_cpus() usage
sched: Fix race in migrate_swap_stop()
...
The variable hv_lapic_frequency causes an unused variable warning if
CONFIG_X86_LOCAL_APIC is disabled. Since the variable is only used
inside a small if statement, move the declaration of that variable
into the if statement itself.
Cc: K. Y. Srinivasan <kys@microsoft.com>
Link: http://lkml.kernel.org/r/1381444224-3303-1-git-send-email-kys@microsoft.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Unlike other uncore boxes, IRP boxes live in PCI buses with no UBOX
device. For PCI bus without UBOX device, we find the next bus that
has UBOX device and use its 'bus to socket' mapping.
Besides the counter/control registers in IRP boxes are not properly
aligned.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: eranian@google.com
Cc: "Yan Zheng" <zheng.z.yan@intel.com>
Link: http://lkml.kernel.org/r/1383197815-17706-2-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The encoding for filter registers of IvyBridge-EP uncore QPI boxes is
completely the same as SandyBridge-EP.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: eranian@google.com
Cc: "Yan Zheng" <zheng.z.yan@intel.com>
Link: http://lkml.kernel.org/r/1383197815-17706-1-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The arch_perf_output_copy_user() default of
__copy_from_user_inatomic() returns bytes not copied, while all other
argument functions given DEFINE_OUTPUT_COPY() return bytes copied.
Since copy_from_user_nmi() is the odd duck out by returning bytes
copied where all other *copy_{to,from}* functions return bytes not
copied, change it over and ammend DEFINE_OUTPUT_COPY() to expect bytes
not copied.
Oddly enough DEFINE_OUTPUT_COPY() already returned bytes not copied
while expecting its worker functions to return bytes copied.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: will.deacon@arm.com
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/r/20131030201622.GR16117@laptop.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently show_cpuinfo_core() displays cpu core information only if
the number of threads per a whole cores is 2 or larger.
However, this condition doesn't care about the number of
sockets. For example, this condition doesn't hold on systems
with two logical cpus consisting of two sockets and a single
core on each socket - yet the topology information would be
interesting to see in that case as well.
I don't know whether or not there are processors in real world
by which such configurations are possible, but at least on
vitual machine environments, such configuration can occur,
typically when no explicit SMP information is provided in
advance.
For example, on qemu/KVM, SMP information is specified via -smp
command-line option, more specifically, its syntax is:
-smp n[,cores=cores][,threads=threads][,sockets=sockets][,maxcpus=maxcpus]
If this is not specified, qemu tells configuration with
n-sockets, 1-core and 1-thread to the guest machine, on which
guest, MP information is not displayed in /proc/cpuinfo.
I saw this situation on VMWare guest environment, too.
To fix this issue, this patch simply removes the condition
because this information is useful even if there's only 1
thread.
Signed-off-by: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/5277D644.4090707@jp.fujitsu.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.15 (GNU/Linux)
iQEcBAABAgAGBQJSdt9HAAoJEHm+PkMAQRiGnzEH/345Keg5dp+oKACnokBfzOtp
V0p3g5EBsGtzEVnV+1B96trczDUtWdDFFr5GfGSj565NBQpFyc+iZC1mC99RDJCs
WUquGFqlLMK2aV0SbKwCO4K1rJ5A0TRVj0ZRJOUJUY7jwNf5Qahny0WBVjO/8qAY
UvJK1rktBClhKdH53YtpDHHgXBeZ2LOrzt1fQ/AMpujGbZauGvnLdNOli5r2kCFK
jzoOgFLvX+PHU/5/d4/QyJPeQNPva5hjk5Ho9UuSJYhnFtPO3EkD4XZLcpcbNEJb
LqBvbnZWm6CS435lfU1l93RqQa5xMO9ITk0oe4h69syTSHwWk9aJ+ZTc/4Up+t8=
=57MC
-----END PGP SIGNATURE-----
Merge tag 'v3.12' into x86/cpu, to refresh the branch before queueing up more changes
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9e7827b5ea ("x86, hyperv: Get the local APIC timer frequency from the
hypervisor") breaks the build with some configs because apic.h isn't
directly included:
arch/x86/kernel/cpu/mshyperv.c: In function 'ms_hyperv_init_platform':
arch/x86/kernel/cpu/mshyperv.c:90:3: error: 'lapic_timer_frequency' undeclared (first use in this function)
arch/x86/kernel/cpu/mshyperv.c:90:3: note: each undeclared identifier is reported only once for each function it appears in
Fix it by including asm/apic.h.
Signed-off-by: David Rientjes <rientjes@google.com>
Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1310111604160.31170@chino.kir.corp.google.com
Acked-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
OK, so what I'm actually seeing on my WSM is that sched/clock.c is
'broken' for the purpose we're using it for.
What triggered it is that my WSM-EP is broken :-(
[ 0.001000] tsc: Fast TSC calibration using PIT
[ 0.002000] tsc: Detected 2533.715 MHz processor
[ 0.500180] TSC synchronization [CPU#0 -> CPU#6]:
[ 0.505197] Measured 3 cycles TSC warp between CPUs, turning off TSC clock.
[ 0.004000] tsc: Marking TSC unstable due to check_tsc_sync_source failed
For some reason it consistently detects TSC skew, even though NHM+
should have a single clock domain for 'reasonable' systems.
This marks sched_clock_stable=0, which means that we do fancy stuff to
try and get a 'sane' clock. Part of this fancy stuff relies on the tick,
clearly that's gone when NOHZ=y. So for idle cpus time gets stuck, until
it either wakes up or gets kicked by another cpu.
While this is perfectly fine for the scheduler -- it only cares about
actually running stuff, and when we're running stuff we're obviously not
idle. This does somewhat break down for perf which can trigger events
just fine on an otherwise idle cpu.
So I've got NMIs get get 'measured' as taking ~1ms, which actually
don't last nearly that long:
<idle>-0 [013] d.h. 886.311970: rcu_nmi_enter <-do_nmi
...
<idle>-0 [013] d.h. 886.311997: perf_sample_event_took: HERE!!! : 1040990
So ftrace (which uses sched_clock(), not the fancy bits) only sees
~27us, but we measure ~1ms !!
Now since all this measurement stuff lives in x86 code, we can actually
fix it.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: mingo@kernel.org
Cc: dave.hansen@linux.intel.com
Cc: eranian@google.com
Cc: Don Zickus <dzickus@redhat.com>
Cc: jmario@redhat.com
Cc: acme@infradead.org
Link: http://lkml.kernel.org/r/20131017133350.GG3364@laptop.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
struct cpu_dev's c_models is only ever set inside CONFIG_X86_32
conditionals (or code that's being built for 32-bit only), so
there's no use of reserving the (empty) space for the model
names in a 64-bit kernel.
Similarly, c_size_cache is only used in the #else of a
CONFIG_X86_64 conditional, so reserving space for (and in one
case even initializing) that field is pointless for 64-bit
kernels too.
While moving both fields to the end of the structure, I also
noticed that:
- the c_models array size was one too small, potentially causing
table_lookup_model() to return garbage on Intel CPUs (intel.c's
instance was lacking the sentinel with family being zero), so the
patch bumps that by one,
- c_models' vendor sub-field was unused (and anyway redundant
with the base structure's c_x86_vendor field), so the patch deletes it.
Also rename the legacy fields so that their legacy nature stands out
and comment their declarations.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Link: http://lkml.kernel.org/r/5265036802000078000FC4DB@nat28.tlf.novell.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In latest UEFI spec(by now it is 2.4) memory error definition
for CPER (UEFI 2.4 Appendix N Common Platform Error Record)
adds some new fields. These fields help people to locate
memory error to an actual DIMM location.
Original-author: Tony Luck <tony.luck@intel.com>
Signed-off-by: Chen, Gong <gong.chen@linux.intel.com>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
There's been reports of high NMI handler overhead, highlighted by
such kernel messages:
[ 3697.380195] perf samples too long (10009 > 10000), lowering kernel.perf_event_max_sample_rate to 13000
[ 3697.389509] INFO: NMI handler (perf_event_nmi_handler) took too long to run: 9.331 msecs
Don Zickus analyzed the source of the overhead and reported:
> While there are a few places that are causing latencies, for now I focused on
> the longest one first. It seems to be 'copy_user_from_nmi'
>
> intel_pmu_handle_irq ->
> intel_pmu_drain_pebs_nhm ->
> __intel_pmu_drain_pebs_nhm ->
> __intel_pmu_pebs_event ->
> intel_pmu_pebs_fixup_ip ->
> copy_from_user_nmi
>
> In intel_pmu_pebs_fixup_ip(), if the while-loop goes over 50, the sum of
> all the copy_from_user_nmi latencies seems to go over 1,000,000 cycles
> (there are some cases where only 10 iterations are needed to go that high
> too, but in generall over 50 or so). At this point copy_user_from_nmi
> seems to account for over 90% of the nmi latency.
The solution to that is to avoid having to call copy_from_user_nmi() for
every instruction.
Since we already limit the max basic block size, we can easily
pre-allocate a piece of memory to copy the entire thing into in one
go.
Don reported this test result:
> Your patch made a huge difference in improvement. The
> copy_from_user_nmi() no longer hits the million of cycles. I still
> have a batch of 100,000-300,000 cycles. My longest NMI paths used
> to be dominated by copy_from_user_nmi, now it is not (I have to dig
> up the new hot path).
Reported-and-tested-by: Don Zickus <dzickus@redhat.com>
Cc: jmario@redhat.com
Cc: acme@infradead.org
Cc: dave.hansen@linux.intel.com
Cc: eranian@google.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20131016105755.GX10651@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Correct common misspelling of "identify" as "indentify" throughout
the kernel
Signed-off-by: Maxime Jayat <maxime@artisandeveloppeur.fr>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Adds potential sources of randomness: RDRAND, RDTSC, or the i8254.
This moves the pre-alternatives inline rdrand function into the header so
both pieces of code can use it. Availability of RDRAND is then controlled
by CONFIG_ARCH_RANDOM, if someone wants to disable it even for kASLR.
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: http://lkml.kernel.org/r/1381450698-28710-4-git-send-email-keescook@chromium.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Hyper-V supports a mechanism for retrieving the local APIC frequency.
Use this and bypass the calibration code in the kernel . This would
allow us to boot the Linux kernel as a "modern VM" on Hyper-V where
many of the legacy devices (such as PIT) are not emulated.
I would like to thank Olaf Hering <olaf@aepfle.de>, Jan Beulich <JBeulich@suse.com> and
H. Peter Anvin <h.peter.anvin@intel.com> for their help in this effort.
In this version of the patch, I have addressed Jan's comments.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Link: http://lkml.kernel.org/r/1380554932-9888-1-git-send-email-olaf@aepfle.de
Tested-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
iQEcBAABAgAGBQJSUc9zAAoJEHm+PkMAQRiG9DMH/AtpuAF6LlMRPjrCeuJQ1pyh
T0IUO+CsLKO6qtM5IyweP8V6zaasNjIuW1+B6IwVIl8aOrM+M7CwRiKvpey26ldM
I8G2ron7hqSOSQqSQs20jN2yGAqQGpYIbTmpdGLAjQ350NNNvEKthbP5SZR5PAmE
UuIx5OGEkaOyZXvCZJXU9AZkCxbihlMSt2zFVxybq2pwnGezRUYgCigE81aeyE0I
QLwzzMVdkCxtZEpkdJMpLILAz22jN4RoVDbXRa2XC7dA9I2PEEXI9CcLzqCsx2Ii
8eYS+no2K5N2rrpER7JFUB2B/2X8FaVDE+aJBCkfbtwaYTV9UYLq3a/sKVpo1Cs=
=xSFJ
-----END PGP SIGNATURE-----
Merge tag 'v3.12-rc4' into sched/core
Merge Linux v3.12-rc4 to fix a conflict and also to refresh the tree
before applying more scheduler patches.
Conflicts:
arch/avr32/include/asm/Kbuild
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Haswell always give an extra LBR record after every TSX abort.
Suppress the extra record.
This only works when the abort is visible in the LBR
If the original abort has already left the 16 LBR entries
the extra entry will will stay.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379688044-14173-7-git-send-email-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In the PEBS handler report the transaction flags using the new
generic transaction flags facility. Most of them come from
the "tsx_tuning" field in PEBSv2, but the abort code is derived
from the RAX register reported in the PEBS record.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379688044-14173-3-git-send-email-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently the cap_user_time_zero capability has different tests than
cap_user_time; even though they expose the exact same data.
Switch from CONSTANT && NONSTOP to sched_clock_stable to also deal
with multi cabinet machines and drop the tsc_disabled() check.. non of
this will work sanely without tsc anyway.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-nmgn0j0muo1r4c94vlfh23xy@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf fixes from Ingo Molnar:
"A couple of tooling fixlets and a PMU detection printout fix"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86: Fix PMU detection printout when no PMU is detected
perf symbols: Demangle cloned functions
perf machine: Fix path unpopulated in machine__create_modules()
perf tools: Explicitly add libdl dependency
perf probe: Fix probing symbols with optimization suffix
perf trace: Add mmap2 handler
perf kmem: Make it work again on non NUMA machines
Ran into this cryptic PMU bootup log recently:
[ 0.124047] Performance Events:
[ 0.125000] smpboot: ...
Turns out we print this if no PMU is detected. Fall back to
the right condition so that the following is printed:
[ 0.122381] Performance Events: no PMU driver, software events only.
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: http://lkml.kernel.org/n/tip-u2fwaUffakjp0qkpRfqljgsn@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf fixes from Ingo Molnar:
"Assorted standalone fixes"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Add model number for Avoton Silvermont
perf: Fix capabilities bitfield compatibility in 'struct perf_event_mmap_page'
perf/x86/intel/uncore: Don't use smp_processor_id() in validate_group()
perf: Update ABI comment
tools lib lk: Uninclude linux/magic.h in debugfs.c
perf tools: Fix old GCC build error in trace-event-parse.c:parse_proc_kallsyms()
perf probe: Fix finder to find lines of given function
perf session: Check for SIGINT in more loops
perf tools: Fix compile with libelf without get_phdrnum
perf tools: Fix buildid cache handling of kallsyms with kcore
perf annotate: Fix objdump line parsing offset validation
perf tools: Fill in new definitions for madvise()/mmap() flags
perf tools: Sharpen the libaudit dependencies test
Convert x86 to use a per-cpu preemption count. The reason for doing so
is that accessing per-cpu variables is a lot cheaper than accessing
thread_info variables.
We still need to save/restore the actual preemption count due to
PREEMPT_ACTIVE so we place the per-cpu __preempt_count variable in the
same cache-line as the other hot __switch_to() variables such as
current_task.
NOTE: this save/restore is required even for !PREEMPT kernels as
cond_resched() also relies on preempt_count's PREEMPT_ACTIVE to ignore
task_struct::state.
Also rename thread_info::preempt_count to ensure nobody is
'accidentally' still poking at it.
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-gzn5rfsf8trgjoqx8hyayy3q@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fengguang Wu reported this build warning:
arch/x86/kernel/cpu/perf_event_intel_ds.c: In function 'intel_pmu_drain_pebs_nhm':
arch/x86/kernel/cpu/perf_event_intel_ds.c:964:2: warning: format '%ld' expects argument of type 'long int', but argument 4 has type 'int'
Because pointer arithmetics result type is bitness dependent there's no natural
type to use here, cast it to long.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-jbpauwxJqtf24luewcsdFith@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Solve the problems around the broken definition of perf_event_mmap_page::
cap_usr_time and cap_usr_rdpmc fields which used to overlap, partially
fixed by:
860f085b74 ("perf: Fix broken union in 'struct perf_event_mmap_page'")
The problem with the fix (merged in v3.12-rc1 and not yet released
officially), noticed by Vince Weaver is that the new behavior is
not detectable by new user-space, and that due to the reuse of the
field names it's easy to mis-compile a binary if old headers are used
on a new kernel or new headers are used on an old kernel.
To solve all that make this change explicit, detectable and self-contained,
by iterating the ABI the following way:
- Always clear bit 0, and rename it to usrpage->cap_bit0, to at least not
confuse old user-space binaries. RDPMC will be marked as unavailable
to old binaries but that's within the ABI, this is a capability bit.
- Rename bit 1 to ->cap_bit0_is_deprecated and always set it to 1, so new
libraries can reliably detect that bit 0 is deprecated and perma-zero
without having to check the kernel version.
- Use bits 2, 3, 4 for the newly defined, correct functionality:
cap_user_rdpmc : 1, /* The RDPMC instruction can be used to read counts */
cap_user_time : 1, /* The time_* fields are used */
cap_user_time_zero : 1, /* The time_zero field is used */
- Rename all the bitfield names in perf_event.h to be different from the
old names, to make sure it's not possible to mis-compile it
accidentally with old assumptions.
The 'size' field can then be used in the future to add new fields and it
will act as a natural ABI version indicator as well.
Also adjust tools/perf/ userspace for the new definitions, noticed by
Adrian Hunter.
Reported-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Also-Fixed-by: Adrian Hunter <adrian.hunter@intel.com>
Link: http://lkml.kernel.org/n/tip-zr03yxjrpXesOzzupszqglbv@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
uncore_validate_group() can't call smp_processor_id() because it is
in preemptible context. Pass NUMA_NO_NODE to the allocator instead.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379400493-11505-1-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86 fixes from Ingo Molnar:
"Misc fixes"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/intel/lpss: Add pin control support to Intel low power subsystem
perf/x86/intel: Mark MEM_LOAD_UOPS_MISS_RETIRED as precise on SNB
x86: Remove now-unused save_rest()
x86/smpboot: Fix announce_cpu() to printk() the last "OK" properly
On Intel SNB (SNB, SNB-EP), the event MEM_LOAD_UOPS_MISS_RETIRED
supports PEBS. It was missing for the SNB PEBS event constraint
table thereby preventing any measurement with PEBS for it.
This patch adds the event to the PEBS table for SNB.
WARNING: it should be noted that this event like a few others
are subject to the erratum BT241 for Xeon E5 (SNB-EP). As such,
the event may undercount when used with PEBS unless the
workaround is implemented. But without this patch and just the
workaround, the kernel would not allow precise sampling on this
event. BT241 is documented in:
http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/xeon-e5-family-spec-update.pdf
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: ak@linux.intel.com
Cc: zheng.z.yan@intel.com
Link: http://lkml.kernel.org/r/20130913201646.GA23981@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf fixes from Ingo Molnar:
"Various fixes.
The -g perf report lockup you reported is only partially addressed,
patches that fix the excessive runtime are still being worked on"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86: Fix uncore PCI fixed counter handling
uprobes: Fix utask->depth accounting in handle_trampoline()
perf/x86: Add constraint for IVB CYCLE_ACTIVITY:CYCLES_LDM_PENDING
perf: Fix up MMAP2 buffer space reservation
perf tools: Add attr->mmap2 support
perf kvm: Fix sample_type manipulation
perf evlist: Fix id pos in perf_evlist__open()
perf trace: Handle perf.data files with no tracepoints
perf session: Separate progress bar update when processing events
perf trace: Check if MAP_32BIT is defined
perf hists: Fix formatting of long symbol names
perf evlist: Fix parsing with no sample_id_all bit set
perf tools: Add test for parsing with no sample_id_all bit
perf trace: Check control+C more often
Make the code a bit more readable by removing stray whitespaces et al.
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: http://lkml.kernel.org/n/tip-lzEnychz1ylqy8zjenxOmeht@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Clean up the weird CP interrupt exception code by keeping a CP mask.
Andi suggested this implementation but weirdly didn't actually
implement it himself, do so now because it removes the conditional in
the interrupt handler and avoids the assumption its only on cnt2.
Suggested-by: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-dvb4q0rydkfp00kqat4p5bah@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Add TSX event aliases, and export them from the kernel to perf.
These are used by perf stat -T and to allow
more user friendly access to events. The events are designed to
be fairly generic and may also apply to other architectures
implementing HTM. They all cover common situations that
happens during tuning of transactional code.
For Haswell we have to separate the HLE and RTM events,
as they are separate in the PMU.
This adds the following events:
tx-start Count start transaction (used by perf stat -T)
tx-commit Count commit of transaction
tx-abort Count all aborts
tx-conflict Count aborts due to conflict with another CPU.
tx-capacity Count capacity aborts (transaction too large)
Then matching el-* events for HLE
cycles-t Transactional cycles (used by perf stat -T)
* also exists on POWER8
cycles-ct Transactional cycles commited (used by perf stat -T)
* according to Michael Ellerman POWER8 has a cycles-transactional-committed,
* perf stat -T handles both cases
Note for useful abort profiling often precise has to be set,
as Haswell can only report the point inside the transaction
with precise=2.
For some classes of aborts, like conflicts, this is not needed,
as it makes more sense to look at the complete critical section.
This gives a clean set of generalized events to examine transaction
success and aborts. Haswell has additional events for TSX, but those
are more specialized for very specific situations.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1378438661-24765-4-git-send-email-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Use the existing weight reporting facility to report the transaction
abort cost, that is the number of cycles wasted in aborts.
Haswell reports this in the PEBS record.
This was in fact the original user for weight.
This is a very useful sort key to concentrate on the most
costly aborts and a good metric for TSX tuning.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1378438661-24765-3-git-send-email-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
With checkpointed counters there can be a situation where the counter
is overflowing, aborts the transaction, is set back to a non overflowing
checkpoint, causes interupt. The interrupt doesn't see the overflow
because it has been checkpointed. This is then a spurious PMI, typically with
a ugly NMI message. It can also lead to excessive aborts.
Avoid this problem by:
- Using the full counter width for counting counters (earlier patch)
- Forbid sampling for checkpointed counters. It's not too useful anyways,
checkpointing is mainly for counting. The check is approximate
(to still handle KVM), but should catch the majority of cases.
- On a PMI always set back checkpointed counters to zero.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1378438661-24765-2-git-send-email-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There was a bug in the handling of SNB-EP/IVB-EP uncore PCI
fixed counters, e.g., IMC.
It would cause erratic values to be returned for the IMC
clockticks event. This was due to a bogus hwc->config value
which was then written to PCI config space.
The erratic values can be seen via:
$ perf stat -a -C 0 -e uncore_imc_0/clockticks/ -I 1000 sleep 10
The fixed counter has most fields marked as reserved with
hw reset values of 0. Yet the kernel was defaulting to a
hwc->config = ~0 and that was causing the issues.
This patch sets the hwc->config values for fixed uncore event
to 0. Now, the values of IMC clockticks is correct.
Signed-off-by: Stephane Eranian <eranian@google.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: peterz@infradead.org
Cc: zheng.z.yan@intel.com
Link: http://lkml.kernel.org/r/20130909195350.GA17643@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The IvyBridge event CYCLE_ACTIVITY:CYCLES_LDM_PENDING can only
be measured on counters 0-3 when HT is off. When HT is on, you
only have counters 0-3.
If you program it on the eight counters for 1s on a 3GHz
IVB laptop running a noploop, you see:
2 747 527 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
2 747 527 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
2 747 527 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
2 747 527 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
3 280 563 608 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
3 280 563 608 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
3 280 563 608 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
3 280 563 608 CYCLE_ACTIVITY:CYCLES_LDM_PENDING
Clearly the last 4 values are bogus.
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: ak@linux.intel.com
Cc: zheng.z.yan@intel.com
Cc: dhsharp@google.com
Link: http://lkml.kernel.org/r/20130911152222.GA28761@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The previous patch doing vmstats for TLB flushes ("mm: vmstats: tlb flush
counters") effectively missed UP since arch/x86/mm/tlb.c is only compiled
for SMP.
UP systems do not do remote TLB flushes, so compile those counters out on
UP.
arch/x86/kernel/cpu/mtrr/generic.c calls __flush_tlb() directly. This is
probably an optimization since both the mtrr code and __flush_tlb() write
cr4. It would probably be safe to make that a flush_tlb_all() (and then
get these statistics), but the mtrr code is ancient and I'm hesitant to
touch it other than to just stick in the counters.
[akpm@linux-foundation.org: tweak comments]
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull x86 RAS changes from Ingo Molnar:
"[ The reason for drivers/ updates is that Boris asked for the
drivers/edac/ changes to go via x86/ras in this cycle ]
Main changes:
- AMD CPUs:
. Add ECC event decoding support for new F15h models
. Various erratum fixes
. Fix single-channel on dual-channel-controllers bug.
- Intel CPUs:
. UC uncorrectable memory error parsing fix
. Add support for CMC (Corrected Machine Check) 'FF' (Firmware
First) flag in the APEI HEST
- Various cleanups and fixes"
* 'x86-ras-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
amd64_edac: Fix incorrect wraparounds
amd64_edac: Correct erratum 505 range
cpc925_edac: Use proper array termination
x86/mce, acpi/apei: Only disable banks listed in HEST if mce is configured
amd64_edac: Get rid of boot_cpu_data accesses
amd64_edac: Add ECC decoding support for newer F15h models
x86, amd_nb: Clarify F15h, model 30h GART and L3 support
pci_ids: Add PCI device ID functions 3 and 4 for newer F15h models.
x38_edac: Make a local function static
i3200_edac: Make a local function static
x86/mce: Pay no attention to 'F' bit in MCACOD when parsing 'UC' errors
APEI/ERST: Fix error message formatting
amd64_edac: Fix single-channel setups
EDAC: Replace strict_strtol() with kstrtol()
mce: acpi/apei: Soft-offline a page on firmware GHES notification
mce: acpi/apei: Add a boot option to disable ff mode for corrected errors
mce: acpi/apei: Honour Firmware First for MCA banks listed in APEI HEST CMC
Pull x86 paravirt changes from Ingo Molnar:
"Hypervisor signature detection cleanup and fixes - the goal is to make
KVM guests run better on MS/Hyperv and to generalize and factor out
the code a bit"
* 'x86-paravirt-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86: Correctly detect hypervisor
x86, kvm: Switch to use hypervisor_cpuid_base()
xen: Switch to use hypervisor_cpuid_base()
x86: Introduce hypervisor_cpuid_base()
Pull x86/asmlinkage changes from Ingo Molnar:
"As a preparation for Andi Kleen's LTO patchset (link time
optimizations using GCC's -flto which build time optimization has
steadily increased in quality over the past few years and might
eventually be usable for the kernel too) this tree includes a handful
of preparatory patches that make function calling convention
annotations consistent again:
- Mark every function without arguments (or 64bit only) that is used
by assembly code with asmlinkage()
- Mark every function with parameters or variables that is used by
assembly code as __visible.
For the vanilla kernel this has documentation, consistency and
debuggability advantages, for the time being"
* 'x86-asmlinkage-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/asmlinkage: Fix warning in xen asmlinkage change
x86, asmlinkage, vdso: Mark vdso variables __visible
x86, asmlinkage, power: Make various symbols used by the suspend asm code visible
x86, asmlinkage: Make dump_stack visible
x86, asmlinkage: Make 64bit checksum functions visible
x86, asmlinkage, paravirt: Add __visible/asmlinkage to xen paravirt ops
x86, asmlinkage, apm: Make APM data structure used from assembler visible
x86, asmlinkage: Make syscall tables visible
x86, asmlinkage: Make several variables used from assembler/linker script visible
x86, asmlinkage: Make kprobes code visible and fix assembler code
x86, asmlinkage: Make various syscalls asmlinkage
x86, asmlinkage: Make 32bit/64bit __switch_to visible
x86, asmlinkage: Make _*_start_kernel visible
x86, asmlinkage: Make all interrupt handlers asmlinkage / __visible
x86, asmlinkage: Change dotraplinkage into __visible on 32bit
x86: Fix sys_call_table type in asm/syscall.h
Compared to old atom, Silvermont has offcore and has more events
that support PEBS.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Reviewed-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1374138144-17278-2-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Silvermont (22nm Atom) has two offcore response configuration MSRs,
unlike other Intel CPU, its event code for MSR_OFFCORE_RSP_1 is 0x02b7.
To avoid complicating intel_fixup_er(), use INTEL_UEVENT_EXTRA_REG to
define MSR_OFFCORE_RSP_X. So intel_fixup_er() can find the event code
for OFFCORE_RSP_N by x86_pmu.extra_regs[N].event.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1374138144-17278-1-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86 fixes from Ingo Molnar:
"Two AMD microcode loader fixes and an OLPC firmware support fix"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, microcode, AMD: Fix early microcode loading
x86, microcode, AMD: Make cpu_has_amd_erratum() use the correct struct cpuinfo_x86
x86: Don't clear olpc_ofw_header when sentinel is detected
This patch adds support for the SNB-EP PCU uncore PMU extra_sel_bit
(bit 21) which is missing from the documentation in Table-2.75 of
Intel Xeon Processor E5-2600 Product Family Uncore Performance
Monitoring Guide. It is referred to later in Table-2.81. Without
this selection bit explicitly enabled by the kernel, some events
such as COREx_TRANSITION_CYCLES do not count correctly.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Reviewed-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1376375382-21350-4-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The QPI uncore boxes have two pairs of MATCH/MASK registers that
user to filter packet traffic serviced by QPI link layer. These
registers are in auxiliary PCI devices.
This patch adds the auxiliary PCI devices to snbep_uncore_pci_ids
and adds field definitions for the MATCH/MASK registers.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1375856245-10717-2-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
