1. 12 Jul, 2016 16 commits
  2. 10 Jul, 2016 1 commit
    • Ingo Molnar's avatar
      Revert "perf/x86/intel, watchdog: Switch NMI watchdog to ref cycles on x86" · 44530d58
      Ingo Molnar authored
      This reverts commit 2c95afc1.
      
      Stephane reported the following regression:
      
       > Since Andi added:
       >
       > commit 2c95afc1
       > Author: Andi Kleen <ak@linux.intel.com>
       > Date:   Thu Jun 9 06:14:38 2016 -0700
       >
       >    perf/x86/intel, watchdog: Switch NMI watchdog to ref cycles on x86
       >
       > $ perf stat -e ref-cycles ls
       >   <not counted> ....
       >
       > fails systematically because the ref-cycles is now used by the
       > watchdog and given this is a system-wide pinned event, it monopolizes
       > the fixed counter 2 which is the only counter able to measure this event.
      
      Since the next merge window is near, fix the regression for now
      by reverting the commit.
      Reported-by: default avatarStephane Eranian <eranian@google.com>
      Cc: Andi Kleen <ak@linux.intel.com>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Vince Weaver <vincent.weaver@maine.edu>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: linux-kernel@vger.kernel.org
      Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
      44530d58
  3. 07 Jul, 2016 4 commits
    • Kan Liang's avatar
      perf/x86/intel/uncore: Add support for the Intel Skylake client uncore PMU · 46866b59
      Kan Liang authored
      This patch adds full support for Intel SKL client uncore PMU:
      
       - Add support for SKL client CPU uncore PMU, which is similar to the
         BDW client PMU driver. (There are some differences in CBOX numbering
         and uncore control MSR.)
       - Add new support for SkyLake Mobile uncore PMUs, for both CPU and PCI
         uncore functionality.
      Signed-off-by: default avatarKan Liang <kan.liang@intel.com>
      Signed-off-by: default avatarPeter Zijlstra (Intel) <peterz@infradead.org>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Stephane Eranian <eranian@google.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Vince Weaver <vincent.weaver@maine.edu>
      Link: http://lkml.kernel.org/r/1467208912-8179-1-git-send-email-kan.liang@intel.comSigned-off-by: default avatarIngo Molnar <mingo@kernel.org>
      46866b59
    • Peter Zijlstra's avatar
      perf/x86/intel: Fix rdlbr_to() MSR reading typo · aefbc4d0
      Peter Zijlstra authored
      It helps to actually read the right MSR..
      Signed-off-by: default avatarPeter Zijlstra (Intel) <peterz@infradead.org>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Stephane Eranian <eranian@google.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Vince Weaver <vincent.weaver@maine.edu>
      Cc: linux-kernel@vger.kernel.org
      Fixes: d4cf1949 ("perf/x86/intel: Add {rd,wr}lbr_{to,from} wrappers")
      Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
      aefbc4d0
    • Ingo Molnar's avatar
    • Mark Rutland's avatar
      perf/core: Fix pmu::filter_match for SW-led groups · 2c81a647
      Mark Rutland authored
      The following commit:
      
        66eb579e ("perf: allow for PMU-specific event filtering")
      
      added the pmu::filter_match() callback. This was intended to
      avoid HW constraints on events from resulting in extremely
      pessimistic scheduling.
      
      However, pmu::filter_match() is only called for the leader of each event
      group. When the leader is a SW event, we do not filter the groups, and
      may fail at pmu::add() time, and when this happens we'll give up on
      scheduling any event groups later in the list until they are rotated
      ahead of the failing group.
      
      This can result in extremely sub-optimal event scheduling behaviour,
      e.g. if running the following on a big.LITTLE platform:
      
      $ taskset -c 0 ./perf stat \
       -e 'a57{context-switches,armv8_cortex_a57/config=0x11/}' \
       -e 'a53{context-switches,armv8_cortex_a53/config=0x11/}' \
       ls
      
           <not counted>      context-switches                                              (0.00%)
           <not counted>      armv8_cortex_a57/config=0x11/                                 (0.00%)
                      24      context-switches                                              (37.36%)
                57589154      armv8_cortex_a53/config=0x11/                                 (37.36%)
      
      Here the 'a53' event group was always eligible to be scheduled, but
      the 'a57' group never eligible to be scheduled, as the task was always
      affine to a Cortex-A53 CPU. The SW (group leader) event in the 'a57'
      group was eligible, but the HW event failed at pmu::add() time,
      resulting in ctx_flexible_sched_in giving up on scheduling further
      groups with HW events.
      
      One way of avoiding this is to check pmu::filter_match() on siblings
      as well as the group leader. If any of these fail their
      pmu::filter_match() call, we must skip the entire group before
      attempting to add any events.
      Signed-off-by: default avatarMark Rutland <mark.rutland@arm.com>
      Signed-off-by: default avatarPeter Zijlstra (Intel) <peterz@infradead.org>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Will Deacon <will.deacon@arm.com>
      Fixes: 66eb579e ("perf: allow for PMU-specific event filtering")
      Link: http://lkml.kernel.org/r/1465917041-15339-1-git-send-email-mark.rutland@arm.com
      [ Small readability edits. ]
      Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
      2c81a647
  4. 06 Jul, 2016 1 commit
  5. 05 Jul, 2016 1 commit
    • Ingo Molnar's avatar
      Merge tag 'perf-core-for-mingo-20160704' of... · c50f6245
      Ingo Molnar authored
      Merge tag 'perf-core-for-mingo-20160704' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core
      
      Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:
      
      Documentation changes:
      
       - Update android build documentation (Chris Phlipot)
      
      Infrastructure changes:
      
       - Respect WERROR=0 in libapi and libsubcmd, to allow building on Android (Chris Phlipot)
      
       - Prep work to support SDT events in probe cache (Masami Hiramatsu)
      
       - ELF support for SDT (Hemant Kumar)
      
       - Add feature detection for libelf's elf_getshdrstrndx function (Arnaldo Carvalho de Melo)
      
       - Fix hist accumulation test (Jiri Olsa)
      
       - Unwind callchain fixes (Jiri Olsa)
      
       - Change internal representation of numa nodes obtained from
         perf.data header (Jiri Olsa)
      
       - Sync copy of syscall_64.tbl with the kernel (Arnaldo Carvalho de Melo)
      
       - Add LGPL 2.1 license header to libbpf source files (Wang Nan)
      Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@redhat.com>
      Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
      c50f6245
  6. 04 Jul, 2016 13 commits
  7. 03 Jul, 2016 4 commits
    • Josh Poimboeuf's avatar
      perf/x86: Fix 32-bit perf user callgraph collection · fc188225
      Josh Poimboeuf authored
      A basic perf callgraph record operation causes an immediate panic on a
      32-bit kernel compiled with CONFIG_CC_STACKPROTECTOR=y:
      
        $ perf record -g ls
        Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: c0404fbd
      
        CPU: 0 PID: 998 Comm: ls Not tainted 4.7.0-rc5+ #1
        Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.1-1.fc24 04/01/2014
         c0dd5967 ff7afe1c 00000086 f41dbc2c c07445a0 464c457f f41dbca8 f41dbc44
         c05646f4 f41dbca8 464c457f f41dbca8 464c457f f41dbc54 c04625be c0ce56fc
         c0404fbd f41dbc88 c0404fbd b74668f0 f41dc000 00000000 c0000000 00000000
        Call Trace:
         [<c07445a0>] dump_stack+0x58/0x78
         [<c05646f4>] panic+0x8e/0x1c6
         [<c04625be>] __stack_chk_fail+0x1e/0x30
         [<c0404fbd>] ? perf_callchain_user+0x22d/0x230
         [<c0404fbd>] perf_callchain_user+0x22d/0x230
         [<c055f89f>] get_perf_callchain+0x1ff/0x270
         [<c055f988>] perf_callchain+0x78/0x90
         [<c055c7eb>] perf_prepare_sample+0x24b/0x370
         [<c055c934>] perf_event_output_forward+0x24/0x70
         [<c05531c0>] __perf_event_overflow+0xa0/0x210
         [<c0550a93>] ? cpu_clock_event_read+0x43/0x50
         [<c0553431>] perf_swevent_hrtimer+0x101/0x180
         [<c0456235>] ? kmap_atomic_prot+0x35/0x140
         [<c056dc69>] ? get_page_from_freelist+0x279/0x950
         [<c058fdd8>] ? vma_interval_tree_remove+0x158/0x230
         [<c05939f4>] ? wp_page_copy.isra.82+0x2f4/0x630
         [<c05a050d>] ? page_add_file_rmap+0x1d/0x50
         [<c0565611>] ? unlock_page+0x61/0x80
         [<c0566755>] ? filemap_map_pages+0x305/0x320
         [<c059769f>] ? handle_mm_fault+0xb7f/0x1560
         [<c074cbeb>] ? timerqueue_del+0x1b/0x70
         [<c04cfefe>] ? __remove_hrtimer+0x2e/0x60
         [<c04d017b>] __hrtimer_run_queues+0xcb/0x2a0
         [<c0553330>] ? __perf_event_overflow+0x210/0x210
         [<c04d0a2a>] hrtimer_interrupt+0x8a/0x180
         [<c043ecc2>] local_apic_timer_interrupt+0x32/0x60
         [<c043f643>] smp_apic_timer_interrupt+0x33/0x50
         [<c0b0cd38>] apic_timer_interrupt+0x34/0x3c
        Kernel Offset: disabled
        ---[ end Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: c0404fbd
      
      The panic is caused by the fact that perf_callchain_user() mistakenly
      assumes it's 64-bit only and ends up corrupting the stack.
      Signed-off-by: default avatarJosh Poimboeuf <jpoimboe@redhat.com>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Stephane Eranian <eranian@google.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Vince Weaver <vincent.weaver@maine.edu>
      Cc: stable@vger.kernel.org # v4.5+
      Fixes: 75925e1a ("perf/x86: Optimize stack walk user accesses")
      Link: http://lkml.kernel.org/r/1a547f5077ec30f75f9b57074837c3c80df86e5e.1467432113.git.jpoimboe@redhat.comSigned-off-by: default avatarIngo Molnar <mingo@kernel.org>
      fc188225
    • Stephane Eranian's avatar
      perf/x86/intel: Update event constraints when HT is off · 9010ae4a
      Stephane Eranian authored
      This patch updates the event constraints for non-PEBS mode for
      Intel Broadwell and Skylake processors. When HT is off, each
      CPU gets 8 generic counters. However, not all events can be
      programmed on any of the 8 counters.  This patch adds the
      constraints for the MEM_* events which can only be measured on the
      bottom 4 counters. The constraints are also valid when HT is off
      because, then, there are only 4 generic counters and they are the
      bottom counters.
      Signed-off-by: default avatarStephane Eranian <eranian@google.com>
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Vince Weaver <vincent.weaver@maine.edu>
      Cc: kan.liang@intel.com
      Link: http://lkml.kernel.org/r/1467411742-13245-1-git-send-email-eranian@google.comSigned-off-by: default avatarIngo Molnar <mingo@kernel.org>
      9010ae4a
    • Linus Torvalds's avatar
      Merge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus · 4f302921
      Linus Torvalds authored
      Pull MIPS fix from Ralf Baechle:
       "Only a single fix for 4.7 pending at this point.  It fixes an issue
        that may lead to corruption of the cache mode bits in the page table"
      
      * 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
        MIPS: Fix possible corruption of cache mode by mprotect.
      4f302921
    • Linus Torvalds's avatar
      Merge tag 'powerpc-4.7-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux · 70bd68d7
      Linus Torvalds authored
      Pull powerpc fixes from Michael Ellerman:
      
       - tm: Always reclaim in start_thread() for exec() class syscalls from
         Cyril Bur
      
       - tm: Avoid SLB faults in treclaim/trecheckpoint when RI=0 from Michael
         Neuling
      
       - eeh: Fix wrong argument passed to eeh_rmv_device() from Gavin Shan
      
       - Initialise pci_io_base as early as possible from Darren Stevens
      
      * tag 'powerpc-4.7-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
        powerpc: Initialise pci_io_base as early as possible
        powerpc/tm: Avoid SLB faults in treclaim/trecheckpoint when RI=0
        powerpc/eeh: Fix wrong argument passed to eeh_rmv_device()
        powerpc/tm: Always reclaim in start_thread() for exec() class syscalls
      70bd68d7