- 27 Mar, 2015 21 commits
-
-
Ingo Molnar authored
Signed-off-by:Ingo Molnar <mingo@kernel.org>
-
Ingo Molnar authored
An upcoming patch will depend on tai_ns() and NMI-safe ktime_get_raw_fast(), so merge timers/core here in a separate topic branch until it's all cooked and timers/core is merged upstream. Signed-off-by: -
Peter Zijlstra authored
While looking at some fuzzer output I noticed that we do not hold any locks on leader->ctx and therefore the sibling_list iteration is unsafe. Acquire the relevant ctx->mutex before calling into the pmu specific code. Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Vince Weaver <vincent.weaver@maine.edu> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Sasha Levin <sasha.levin@oracle.com> Link: http://lkml.kernel.org/r/20150225151639.GL5029@twins.programming.kicks-ass.netSigned-off-by:
-
David Ahern authored
perf_pmu_disable() is called before pmu->add() and perf_pmu_enable() is called afterwards. No need to call these inside of x86_pmu_add() as well. Signed-off-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
Ingo Molnar authored
Signed-off-by: -
Ingo Molnar authored
Signed-off-by: -
Peter Zijlstra authored
Because it was the only clock for which we didn't have a _ns() accessor yet. Signed-off-by:
-
Peter Zijlstra authored
Add the NMI safe CLOCK_MONOTONIC_RAW accessor.. Signed-off-by:
John Stultz <john.stultz@linaro.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20150319093400.562746929@infradead.orgSigned-off-by:
-
Peter Zijlstra authored
In preparation for more tk_fast instances, remove all hard-coded tk_fast_mono references. Signed-off-by:
-
Peter Zijlstra authored
Introduce tkr_raw and make use of it. base_raw -> tkr_raw.base clock->{mult,shift} -> tkr_raw.{mult.shift} Kill timekeeping_get_ns_raw() in favour of timekeeping_get_ns(&tkr_raw), this removes all mono_raw special casing. Duplicate the updates to tkr_mono.cycle_last into tkr_raw.cycle_last, both need the same value. Signed-off-by: -
Peter Zijlstra authored
In preparation of adding another tkr field, rename this one to tkr_mono. Also rename tk_read_base::base_mono to tk_read_base::base, since the structure is not specific to CLOCK_MONOTONIC and the mono name got added to the tk_read_base instance. Lots of trivial churn. Signed-off-by:
-
Andi Kleen authored
On Broadwell INST_RETIRED.ALL cannot be used with any period that doesn't have the lowest 6 bits cleared. And the period should not be smaller than 128. This is erratum BDM11 and BDM55: http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/5th-gen-core-family-spec-update.pdf BDM11: When using a period < 100; we may get incorrect PEBS/PMI interrupts and/or an invalid counter state. BDM55: When bit0-5 of the period are !0 we may get redundant PEBS records on overflow. Add a new callback to enforce this, and set it for Broadwell. How does this handle the case when an app requests a specific period with some of the bottom bits set? Short answer: Any useful instruction sampling period needs to be 4-6 orders of magnitude larger than 128, as an PMI every 128 instructions would instantly overwhelm the system and be throttled. So the +-64 error from this is really small compared to the period, much smaller than normal system jitter. Long answer (by Peterz): IFF we guarantee perf_event_attr::sample_period >= 128. Suppose we start out with sample_period=192; then we'll set period_left to 192, we'll end up with left = 128 (we truncate the lower bits). We get an interrupt, find that period_left = 64 (>0 so we return 0 and don't get an overflow handler), up that to 128. Then we trigger again, at n=256. Then we find period_left = -64 (<=0 so we return 1 and do get an overflow). We increment with sample_period so we get left = 128. We fire again, at n=384, period_left = 0 (<=0 so we return 1 and get an overflow). And on and on. So while the individual interrupts are 'wrong' we get then with interval=256,128 in exactly the right ratio to average out at 192. And this works for everything >=128. So the num_samples*fixed_period thing is still entirely correct +- 127, which is good enough I'd say, as you already have that error anyhow. So no need to 'fix' the tools, al we need to do is refuse to create INST_RETIRED:ALL events with sample_period < 128. Signed-off-by:
Andi Kleen <ak@linux.intel.com> [ Updated comments and changelog a bit. ] Signed-off-by:
-
Andi Kleen authored
Add Broadwell support for Broadwell to perf. The basic support is very similar to Haswell. We use the new cache event list added for Haswell earlier. The only differences are a few bits related to remote nodes. To avoid an extra, mostly identical, table these are patched up in the initialization code. The constraint list has one new event that needs to be handled over Haswell. Includes code and testing from Kan Liang. Signed-off-by:
-
Andi Kleen authored
Haswell offcore events are quite different from Sandy Bridge. Add a new table to handle Haswell properly. Note that the offcore bits listed in the SDM are not quite correct (this is currently being fixed). An uptodate list of bits is in the patch. The basic setup is similar to Sandy Bridge. The prefetch columns have been removed, as prefetch counting is not very reliable on Haswell. One L1 event that is not in the event list anymore has been also removed. - data reads do not include code reads (comparable to earlier Sandy Bridge tables) - data counts include speculative execution (except L1 write, dtlb, bpu) - remote node access includes both remote memory, remote cache, remote mmio. - prefetches are not included in the counts for consistency (different from Sandy Bridge, which includes prefetches in the remote node) Signed-off-by:
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' 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: User visible changes: - Show the first event with an invalid filter (David Ahern, Arnaldo Carvalho de Melo) - Fix garbage output when intermixing syscalls from different threads in 'perf trace' (Arnaldo Carvalho de Melo) - Fix 'perf timechart' SIBGUS error on sparc64 (David Ahern) Infrastructure changes: - Set JOBS based on CPU or processor, making it work on SPARC, where /proc/cpuinfo has "CPU", not "processor" (David Ahern) - Zero should not be considered "not found" in libtraceevent's eval_flag() (Steven Rostedt) Signed-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by:
-
Ingo Molnar authored
Trivial cleanups, to improve the readability of the generic sched_clock() code: - Improve and standardize comments - Standardize the coding style - Use vertical spacing where appropriate - etc. No code changed: md5: 19a053b31e0c54feaeff1492012b019a sched_clock.o.before.asm 19a053b31e0c54feaeff1492012b019a sched_clock.o.after.asm Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Daniel Thompson <daniel.thompson@linaro.org> Cc: John Stultz <john.stultz@linaro.org> Cc: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Will Deacon <will.deacon@arm.com> Signed-off-by: -
Daniel Thompson authored
Currently it is possible for an NMI (or FIQ on ARM) to come in and read sched_clock() whilst update_sched_clock() has locked the seqcount for writing. This results in the NMI handler locking up when it calls raw_read_seqcount_begin(). This patch fixes the NMI safety issues by providing banked clock data. This is a similar approach to the one used in Thomas Gleixner's 4396e058("timekeeping: Provide fast and NMI safe access to CLOCK_MONOTONIC"). Suggested-by:
Stephen Boyd <sboyd@codeaurora.org> Signed-off-by:
Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by:
-
Daniel Thompson authored
Currently update_sched_clock() is marked as notrace but this function is not called by ftrace. This is trivially fixed by removing the mark up. Signed-off-by:
-
Daniel Thompson authored
Currently cd.read_data.suspended is read by the hotpath function sched_clock(). This variable need not be accessed on the hotpath. In fact, once it is removed, we can remove the conditional branches from sched_clock() and install a dummy read_sched_clock function to suspend the clock. The new master copy of the function pointer (actual_read_sched_clock) is introduced and is used for all reads of the clock hardware except those within sched_clock itself. Suggested-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Daniel Thompson authored
Currently sched_clock(), a very hot code path, is not optimized to minimise its cache profile. In particular: 1. cd is not ____cacheline_aligned, 2. struct clock_data does not distinguish between hotpath and coldpath data, reducing locality of reference in the hotpath, 3. Some hotpath data is missing from struct clock_data and is marked __read_mostly (which more or less guarantees it will not share a cache line with cd). This patch corrects these problems by extracting all hotpath data into a separate structure and using ____cacheline_aligned to ensure the hotpath uses a single (64 byte) cache line. Signed-off-by: -
Daniel Thompson authored
Currently the scope of the raw_write_seqcount_begin/end() in sched_clock_register() far exceeds the scope of the read section in sched_clock(). This gives the impression of safety during cursory review but achieves little. Note that this is likely to be a latent issue at present because sched_clock_register() is typically called before we enable interrupts, however the issue does risk bugs being needlessly introduced as the code evolves. This patch fixes the problem by increasing the scope of the read locking performed by sched_clock() to cover all data modified by sched_clock_register. We also improve clarity by moving writes to struct clock_data that do not impact sched_clock() outside of the critical section. Signed-off-by:
-
- 26 Mar, 2015 14 commits
-
-
git://people.freedesktop.org/~airlied/linuxLinus Torvalds authored
Pull drm refcounting fixes from Dave Airlie: "Here is the complete set of i915 bug/warn/refcounting fixes" * 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: drm/i915: Fixup legacy plane->crtc link for initial fb config drm/i915: Fix atomic state when reusing the firmware fb drm/i915: Keep ring->active_list and ring->requests_list consistent drm/i915: Don't try to reference the fb in get_initial_plane_config() drm: Fixup racy refcounting in plane_force_disable
-
git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dmLinus Torvalds authored
Pull device mapper fix from Mike Snitzer: "Fix DM core device cleanup regression -- due to a latent race that was exposed by the bdi changes that were introduced during the 4.0 merge" * tag 'dm-4.0-fix-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: dm: fix add_disk() NULL pointer due to race with free_dev()
-
Linus Torvalds authored
Merge tag 'linux-kselftest-4.0-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest Pull kselftest fix from Shuah Khan. * tag 'linux-kselftest-4.0-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest: selftests: Fix build failures when invoked from kselftest target
-
git://anongit.freedesktop.org/drm-intelDave Airlie authored
This should cover the final warnings in -rc5 with two more backports from our development branch (drm-intel-next-queued). They're the ones from Daniel and Damien, with references to the reports. This is on top of drm-fixes because of the dependency on the two earlier fixes not yet in Linus' tree. There's an additional regression fix from Chris. * tag 'drm-intel-fixes-2015-03-26' of git://anongit.freedesktop.org/drm-intel: drm/i915: Fixup legacy plane->crtc link for initial fb config drm/i915: Fix atomic state when reusing the firmware fb drm/i915: Keep ring->active_list and ring->requests_list consistent
-
git://git.kernel.org/pub/scm/linux/kernel/git/s390/linuxLinus Torvalds authored
Pull s390 fixes from Martin Schwidefsky: "A couple of bug fixes for s390. The ftrace comile fix is quite large for a -rc6 release, but it would be nice to have it in 4.0" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: s390/smp: reenable smt after resume s390/mm: limit STACK_RND_MASK for compat tasks s390/ftrace: fix compile error if CONFIG_KPROBES is disabled s390/cpum_sf: add diagnostic sampling event only if it is authorized
-
Steven Rostedt authored
Guilherme Cox found that: There is, however, a potential bug if there is an item with code zero that is not the first one in the symbol list, since eval_flag(..) returns 0 when it doesn't find anything. That is, if you have the following enums: enum { FOO_START = 0, FOO_GO = 1, FOO_END = 2 } and then have: __print_symbolic(foo, FOO_GO, "go", FOO_START, "start", FOO_END, "end") If none of the enums are known to pevent, then eval_flag() will return zero, and it will match it to the first item in the list, which would be FOO_GO, which is not zero. Luckily, in most cases, the first element would be zero, and the parsing would match out of sheer luck. Reported-by:Guilherme Cox <cox@computer.org> Signed-off-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/r/20150324145813.0bfe95ba@gandalf.local.homeSigned-off-by:
-
Arnaldo Carvalho de Melo authored
commit e596663e Author: Arnaldo Carvalho de Melo <acme@redhat.com> Date: Fri Feb 13 13:22:21 2015 -0300 perf trace: Handle multiple threads better wrt syscalls being intermixed Introduced a bug where it considered the number of bytes output directly to the output file when formatting the syscall entry buffer that is stored to be finally printed at syscall exit, ending up leaving garbage at the start of syscalls that appeared while another syscall was being processed, in another thread. Fix it. Example of garbage in the output before this patch: 4280.102 ( 0.000 ms): lsmd/763 ... [continued]: select()) = 0 Timeout 4280.107 (275.250 ms): tuned/852 select(tvp: 0x7f41f7ffde50 ) ... 4280.109 ( 0.002 ms): lsmd/763 Xl�� ) = -10 4639.197 ( 0.000 ms): systemd-journa/542 ... [continued]: epoll_wait()) = 1 4639.202 (359.088 ms): lsmd/763 select(n: 6, inp: 0x7ffff21daad0, tvp: 0x7ffff21daac0) ... 4639.207 ( 0.005 ms): systemd-journa/542 Hn�� ) = 106 4639.221 ( 0.002 ms): systemd-journa/542 uname(name: 0x7ffdbaed8e00) = 0 4639.271 ( 0.008 ms): systemd-journa/542 ftruncate(fd: 11</run/log/journal/60cd52417cf440a4a80107518bbd3c20/system.journal>, length: 50331648) = 0 Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Borislav Petkov <bp@suse.de> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-9ckfe8mvsedgkg6y80gz1ul8@git.kernel.orgSigned-off-by:
-
David Ahern authored
Number of JOBS to use is set automatically to the number of processors found in /proc/cpuinfo. SPARC uses 'CPU' lines rather than 'processor'. Update the check in perf's Makefile to work for SPARC. Signed-off-by:
David Ahern <david.ahern@oracle.com> Acked-by:
Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/r/1427213455-127249-1-git-send-email-david.ahern@oracle.comSigned-off-by:
-
David Ahern authored
SPARC based systems currently support up to 1024 cpus (e.g. T5-8). Allow perf to work on those systems. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
Use of a bad filter currently generates the message: Error: failed to set filter with 22 (Invalid argument) Add the event name to make it clear to which event the filter failed to apply: Error: Failed to set filter "foo" on event sched:sg_lb_stats: 22: Invalid argument To test it use something like: # perf record -e sched:sched_switch -e sched:*fork --filter parent_pid==1 -e sched:*wait* --filter bla usleep 1 Error: failed to set filter "bla" on event sched:sched_stat_iowait with 22 (Invalid argument) # Based-on-a-patch-by:
-
David Ahern authored
perf timechart -T on sparc64 is terminating due to SIGBUS. Backtrace: Program received signal SIGBUS, Bus error. 0x0000000000173d7c in perf_evsel__intval (evsel=<value optimized out>, sample=0x7feffffda28, name=0x289b28 "prev_state") at util/evsel.c:1918 1918 util/evsel.c: No such file or directory. in util/evsel.c Missing separate debuginfos, use: debuginfo-install audit-libs-2.3.7-1.0.1.el6.sparc64 bzip2-libs-1.0.5-7.el6_0.sparc64 elfutils-libelf-0.155-2.0.3.el6.sparc64 elfutils-libs-0.155-2.0.3.el6.sparc64 glibc-2.12-1.132.0.8.el6_5.sparc64 numactl-2.0.7-8.el6.sparc64 python-libs-2.6.6-52.0.2.el6.sparc64 slang-2.2.1-1.el6.sparc64 xz-libs-4.999.9-0.3.beta.20091007git.el6.sparc64 zlib-1.2.3-29.el6.sparc64 (gdb) bt 0 0x0000000000173d7c in perf_evsel__intval (evsel=<value optimized out>, sample=0x7feffffda28, name=0x289b28 "prev_state") at util/evsel.c:1918 1 0x0000000000123b94 in process_sample_sched_switch (tchart=0x7feffffe040, evsel=0x4ca850, sample=0x7feffffda28, backtrace=0xc39010 "") at builtin-timechart.c:627 2 0x0000000000122828 in process_sample_event (tool=0x7feffffe040, event=<value optimized out>, sample=0x7feffffda28, evsel=0x4ca850, machine=0x4c9c88) at builtin-timechart.c:569 Another extended load on unaligned pointer. As before fix by copying to a temporary variable using memcpy. Signed-off-by: -
Daniel Vetter authored
This is a very similar bug in the load detect code fixed in commit 9128b040 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Mar 3 17:31:21 2015 +0100 drm/i915: Fix modeset state confusion in the load detect code But this time around it was the initial fb code that forgot to update the plane->crtc pointer. Otherwise it's the exact same bug, with the exact same restrains (any set_config call/ioctl that doesn't disable the pipe papers over the bug for free, so fairly hard to hit in normal testing). So if you want the full explanation just go read that one over there - it's rather long ... Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Josh Boyer <jwboyer@fedoraproject.org> Cc: Jani Nikula <jani.nikula@linux.intel.com> Reported-and-tested-by:
Josh Boyer <jwboyer@fedoraproject.org> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> [Jani: backported to drm-intel-fixes for v4.0-rc] Reference: http://mid.gmane.org/CA+5PVA7ChbtJrknqws1qvZcbrg1CW2pQAFkSMURWWgyASRyGXg@mail.gmail.comSigned-off-by:
Jani Nikula <jani.nikula@intel.com>
-
Damien Lespiau authored
Right now, we get a warning when taking over the firmware fb: [drm:drm_atomic_plane_check] FB set but no CRTC with the following backtrace: [<ffffffffa010339d>] drm_atomic_check_only+0x35d/0x510 [drm] [<ffffffffa0103567>] drm_atomic_commit+0x17/0x60 [drm] [<ffffffffa00a6ccd>] drm_atomic_helper_plane_set_property+0x8d/0xd0 [drm_kms_helper] [<ffffffffa00f1fed>] drm_mode_plane_set_obj_prop+0x2d/0x90 [drm] [<ffffffffa00a8a1b>] restore_fbdev_mode+0x6b/0xf0 [drm_kms_helper] [<ffffffffa00aa969>] drm_fb_helper_restore_fbdev_mode_unlocked+0x29/0x80 [drm_kms_helper] [<ffffffffa00aa9e2>] drm_fb_helper_set_par+0x22/0x50 [drm_kms_helper] [<ffffffffa050a71a>] intel_fbdev_set_par+0x1a/0x60 [i915] [<ffffffff813ad444>] fbcon_init+0x4f4/0x580 That's because we update the plane state with the fb from the firmware, but we never associate the plane to that CRTC. We don't quite have the full DRM take over from HW state just yet, so fake enough of the plane atomic state to pass the checks. v2: Fix the state on which we set the CRTC in the case we're sharing the initial fb with another pipe. (Matt) Signed-off-by:Damien Lespiau <damien.lespiau@intel.com> Reviewed-by:
Matt Roper <matthew.d.roper@intel.com> Signed-off-by:
-
Chris Wilson authored
If we retire requests last, we may use a later seqno and so clear the requests lists without clearing the active list, leading to confusion. Hence we should retire requests first for consistency with the early return. The order used to be important as the lifecycle for the object on the active list was determined by request->seqno. However, the requests themselves are now reference counted removing the constraint from the order of retirement. Fixes regression from commit 1b5a433a Author: John Harrison <John.C.Harrison@Intel.com> Date: Mon Nov 24 18:49:42 2014 +0000 drm/i915: Convert 'i915_seqno_passed' calls into 'i915_gem_request_completed ' and a WARNING: CPU: 0 PID: 1383 at drivers/gpu/drm/i915/i915_gem_evict.c:279 i915_gem_evict_vm+0x10c/0x140() WARN_ON(!list_empty(&vm->active_list)) Identified by updating WATCH_LISTS: [drm:i915_verify_lists] *ERROR* blitter ring: active list not empty, but no requests WARNING: CPU: 0 PID: 681 at drivers/gpu/drm/i915/i915_gem.c:2751 i915_gem_retire_requests_ring+0x149/0x230() WARN_ON(i915_verify_lists(ring->dev)) Note that this is only a problem in evict_vm where the following happens after a retire_request has cleaned out all requests, but not all active bo: - intel_ring_idle called from i915_gpu_idle notices that no requests are outstanding and immediately returns. - i915_gem_retire_requests_ring called from i915_gem_retire_requests also immediately returns when there's no request, still leaving the bo on the active list. - evict_vm hits the WARN_ON(!list_empty(&vm->active_list)) after evicting all active objects that there's still stuff left that shouldn't be there. Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: John Harrison <John.C.Harrison@Intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by:
-
- 25 Mar, 2015 5 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/metagLinus Torvalds authored
Pull arch/metag fix from James Hogan: "Another metag architecture fix for v4.0 This is another single fix, for an include dependency problem when using ioremap_wc() from asm/io.h without also including asm/pgtable.h" * tag 'metag-fixes-v4.0-2' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/metag: metag: Fix ioremap_wc/ioremap_cached build errors
-
Linus Torvalds authored
Merge misc fixes from Andrew Morton: "15 fixes" * emailed patches from Andrew Morton <akpm@linux-foundation.org>: mm: numa: mark huge PTEs young when clearing NUMA hinting faults mm: numa: slow PTE scan rate if migration failures occur mm: numa: preserve PTE write permissions across a NUMA hinting fault mm: numa: group related processes based on VMA flags instead of page table flags hfsplus: fix B-tree corruption after insertion at position 0 MAINTAINERS: add Jan as DMI/SMBIOS support maintainer fs/affs/file.c: unlock/release page on error mm/page_alloc.c: call kernel_map_pages in unset_migrateype_isolate mm/slub: fix lockups on PREEMPT && !SMP kernels mm/memory hotplug: postpone the reset of obsolete pgdat MAINTAINERS: correct rtc armada38x pattern entry mm/pagewalk.c: prevent positive return value of walk_page_test() from being passed to callers mm: fix anon_vma->degree underflow in anon_vma endless growing prevention drivers/rtc/rtc-mrst: fix suspend/resume aoe: update aoe maintainer information
-
Mel Gorman authored
Base PTEs are marked young when the NUMA hinting information is cleared but the same does not happen for huge pages which this patch addresses. Note that migrated pages are not marked young as the base page migration code does not assume that migrated pages have been referenced. This could be addressed but beyond the scope of this series which is aimed at Dave Chinners shrink workload that is unlikely to be affected by this issue. Signed-off-by:
Mel Gorman <mgorman@suse.de> Cc: Dave Chinner <david@fromorbit.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Aneesh Kumar <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Mel Gorman authored
Dave Chinner reported the following on https://lkml.org/lkml/2015/3/1/226 Across the board the 4.0-rc1 numbers are much slower, and the degradation is far worse when using the large memory footprint configs. Perf points straight at the cause - this is from 4.0-rc1 on the "-o bhash=101073" config: - 56.07% 56.07% [kernel] [k] default_send_IPI_mask_sequence_phys - default_send_IPI_mask_sequence_phys - 99.99% physflat_send_IPI_mask - 99.37% native_send_call_func_ipi smp_call_function_many - native_flush_tlb_others - 99.85% flush_tlb_page ptep_clear_flush try_to_unmap_one rmap_walk try_to_unmap migrate_pages migrate_misplaced_page - handle_mm_fault - 99.73% __do_page_fault trace_do_page_fault do_async_page_fault + async_page_fault 0.63% native_send_call_func_single_ipi generic_exec_single smp_call_function_single This is showing excessive migration activity even though excessive migrations are meant to get throttled. Normally, the scan rate is tuned on a per-task basis depending on the locality of faults. However, if migrations fail for any reason then the PTE scanner may scan faster if the faults continue to be remote. This means there is higher system CPU overhead and fault trapping at exactly the time we know that migrations cannot happen. This patch tracks when migration failures occur and slows the PTE scanner. Signed-off-by:
Dave Chinner <david@fromorbit.com> Tested-by:
-
Mel Gorman authored
Protecting a PTE to trap a NUMA hinting fault clears the writable bit and further faults are needed after trapping a NUMA hinting fault to set the writable bit again. This patch preserves the writable bit when trapping NUMA hinting faults. The impact is obvious from the number of minor faults trapped during the basis balancing benchmark and the system CPU usage; autonumabench 4.0.0-rc4 4.0.0-rc4 baseline preserve Time System-NUMA01 107.13 ( 0.00%) 103.13 ( 3.73%) Time System-NUMA01_THEADLOCAL 131.87 ( 0.00%) 83.30 ( 36.83%) Time System-NUMA02 8.95 ( 0.00%) 10.72 (-19.78%) Time System-NUMA02_SMT 4.57 ( 0.00%) 3.99 ( 12.69%) Time Elapsed-NUMA01 515.78 ( 0.00%) 517.26 ( -0.29%) Time Elapsed-NUMA01_THEADLOCAL 384.10 ( 0.00%) 384.31 ( -0.05%) Time Elapsed-NUMA02 48.86 ( 0.00%) 48.78 ( 0.16%) Time Elapsed-NUMA02_SMT 47.98 ( 0.00%) 48.12 ( -0.29%) 4.0.0-rc4 4.0.0-rc4 baseline preserve User 44383.95 43971.89 System 252.61 201.24 Elapsed 998.68 1000.94 Minor Faults 2597249 1981230 Major Faults 365 364 There is a similar drop in system CPU usage using Dave Chinner's xfsrepair workload 4.0.0-rc4 4.0.0-rc4 baseline preserve Amean real-xfsrepair 454.14 ( 0.00%) 442.36 ( 2.60%) Amean syst-xfsrepair 277.20 ( 0.00%) 204.68 ( 26.16%) The patch looks hacky but the alternatives looked worse. The tidest was to rewalk the page tables after a hinting fault but it was more complex than this approach and the performance was worse. It's not generally safe to just mark the page writable during the fault if it's a write fault as it may have been read-only for COW so that approach was discarded. Signed-off-by:
-
