perf changes
Subsystems (11)
Linus,
Please pull the latest perf-core-for-linus git tree from:
git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git perf-core-for-linus
Most of these changes are fixes and dead code removal, but also a set of optimizations and a series of tracing size-optimization bits that we kept pending because they conflicted with the previous push - and some TUI enhancements.
Thanks,
Ingo
------------------>
Akinobu Mita (1):
x86/mmiotrace: Remove redundant instruction prefix checks
Arnaldo Carvalho de Melo (15):
perf tui: Fix build problem with slang <= 2.0.6
perf tools: Remove some unused functions
perf probe: Fix some error exit paths
perf probe: Don't call die()
perf tools: remove xstrndup, xmalloc, xzalloc
perf symbols: Don't try to read the build-id twice
perf session: Make read_build_id routines look at the host_machine too perf TUI: Make 'space' be an alias to 'PgDn'
perf annotate: Use build-ids to find the right DSO
perf tui: Allow disabling the TUI on a per command basis in ~/.perfconfig perf report: Don't start the TUI if -D is used
perf tui: Remove annotate from popup menu after failure
perf annotate: Add TUI interface
perf annotate: Fix up usage of the build id cache
perf report: Support multiple events on the TUI
Cyrill Gorcunov (5):
perf, x86: P4 PMU -- handle unflagged events
perf, x86: P4 PMU -- fix typo in unflagged NMI handling perf, x86: P4 PMU -- do a real check for ESCR address being in hash perf, x86: P4_pmu_schedule_events -- use smp_processor_id instead of raw_ perf, x86: P4 PMU -- add missing bit in CCCR mask
Frederic Weisbecker (4):
perf: Comply with new rcu checks API
perf: Fix unaligned accesses while fetching trace values
perf: Fix forgotten preempt_enable by nested writers
perf: Fix getline undeclared
Li Zefan (1):
tracing: Fix function declarations if !CONFIG_STACKTRACE
Lin Ming (1):
perf, sparc: Implement group scheduling transactional APIs
Masami Hiramatsu (1):
perf probe: Don't compile CFI related code if elfutils is old
Mathieu Desnoyers (1):
tracepoints: Add check trace callback type
Peter Zijlstra (16):
perf/ftrace: Optimize perf/tracepoint interaction for single events
perf: Optimize buffer placement by allocating buffers NUMA aware
perf: Disallow mmap() on per-task inherited events
perf: Optimize the perf_output() path by removing IRQ-disables
perf: Optimize the hotpath by converting the perf output buffer to local_t
perf: Optimize perf_output_*() by avoiding local_xchg()
perf, trace: Optimize tracepoints by removing IRQ-disable from perf/tracepoint interaction perf, trace: Optimize tracepoints by using per-tracepoint-per-cpu hlist to track events
perf: Ensure that IOC_OUTPUT isn't used to create multi-writer buffers
perf-record: Remove -M
perf-record: Share per-cpu buffers
perf: Fix wakeup storm for RO mmap()s
perf: Optimize perf_output_copy()
perf: Optimize the !vmalloc backed buffer
perf: Remove more code from the fastpath
perf: Optimize perf_tp_event_match()
Russ Anderson (1):
perf record: remove unneeded gettimeofday() call
Stephane Eranian (2):
perf: Fix errors path in perf_output_begin() perf stat: add perf stat -B to pretty print large numbers
Steven Rostedt (10):
tracing: Create class struct for events
tracing: Let tracepoints have data passed to tracepoint callbacks
tracing: Remove per event trace registering
tracing: Move fields from event to class structure
tracing: Move raw_init from events to class
tracing: Allow events to share their print functions
tracing: Move print functions into event class
tracing: Remove duplicate id information in event structure
tracing: Combine event filter_active and enable into single flags field
tracing: Comment the use of event_mutex with trace event flags
Tom Zanussi (1):
perf: Use read() instead of lseek() in trace_event_read.c:skip()
arch/sparc/kernel/perf_event.c | 108 ++++--- arch/x86/include/asm/perf_event_p4.h | 3 +- arch/x86/kernel/cpu/perf_event_p4.c | 41 ++- arch/x86/mm/pf_in.c | 2 +- include/linux/ftrace_event.h | 103 ++++-- include/linux/perf_event.h | 28 +- include/linux/syscalls.h | 57 +-- include/linux/tracepoint.h | 98 ++++-- include/trace/ftrace.h | 247 +++++-------- include/trace/syscall.h | 10 +- kernel/perf_event.c | 389 +++++++++++--------- kernel/trace/blktrace.c | 138 ++++--- kernel/trace/ftrace.c | 7 +- kernel/trace/kmemtrace.c | 70 +++-- kernel/trace/trace.c | 9 +- kernel/trace/trace.h | 9 +- kernel/trace/trace_branch.c | 8 +- kernel/trace/trace_event_perf.c | 183 +++++----- kernel/trace/trace_events.c | 139 +++++--- kernel/trace/trace_events_filter.c | 28 +- kernel/trace/trace_export.c | 16 +- kernel/trace/trace_functions_graph.c | 13 +- kernel/trace/trace_kprobe.c | 113 ++++--- kernel/trace/trace_output.c | 137 +++++--- kernel/trace/trace_output.h | 2 +- kernel/trace/trace_sched_switch.c | 20 +- kernel/trace/trace_sched_wakeup.c | 28 +- kernel/trace/trace_syscalls.c | 146 ++++++-- kernel/trace/trace_workqueue.c | 26 +- kernel/tracepoint.c | 91 +++-- net/core/drop_monitor.c | 12 +- samples/tracepoints/tp-samples-trace.h | 4 +- samples/tracepoints/tracepoint-probe-sample.c | 13 +- samples/tracepoints/tracepoint-probe-sample2.c | 7 +- tools/perf/Documentation/perf-stat.txt | 3 + tools/perf/builtin-annotate.c | 61 +++- tools/perf/builtin-probe.c | 10 +- tools/perf/builtin-record.c | 72 ++--- tools/perf/builtin-report.c | 64 ++-- tools/perf/builtin-stat.c | 18 +- tools/perf/perf.c | 25 ++- tools/perf/util/abspath.c | 81 ----- tools/perf/util/build-id.c | 22 ++ tools/perf/util/build-id.h | 2 + tools/perf/util/cache.h | 57 +--- tools/perf/util/callchain.c | 1 + tools/perf/util/callchain.h | 1 - tools/perf/util/config.c | 461 +----------------------- tools/perf/util/exec_cmd.c | 6 +- tools/perf/util/exec_cmd.h | 1 - tools/perf/util/header.c | 84 +++-- tools/perf/util/help.c | 30 +-- tools/perf/util/hist.c | 42 ++- tools/perf/util/hist.h | 24 ++- tools/perf/util/newt.c | 150 ++++++-- tools/perf/util/path.c | 204 +----------- tools/perf/util/probe-finder.c | 33 ++- tools/perf/util/probe-finder.h | 3 + tools/perf/util/quote.c | 433 +---------------------- tools/perf/util/quote.h | 39 -- tools/perf/util/run-command.c | 90 ----- tools/perf/util/run-command.h | 30 -- tools/perf/util/session.c | 8 + tools/perf/util/session.h | 8 +- tools/perf/util/sigchain.c | 2 +- tools/perf/util/sigchain.h | 1 - tools/perf/util/strbuf.c | 229 +------------ tools/perf/util/strbuf.h | 45 --- tools/perf/util/symbol.c | 25 +- tools/perf/util/symbol.h | 1 + tools/perf/util/trace-event-read.c | 19 +- tools/perf/util/trace-event.h | 7 +- tools/perf/util/util.h | 177 +--------- tools/perf/util/wrapper.c | 110 ------ 74 files changed, 1839 insertions(+), 3145 deletions(-)
arch/sparc/kernel/perf_event.c +61 -47
@@ -91,6 +91,8 @@ struct cpu_hw_events {/* Enabled/disable state. */ int enabled; + +unsigned int group_flag; }; DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { .enabled = 1, };
@@ -980,53 +982,6 @@ static int collect_events(struct perf_event *group, int max_count,return n; } -static void event_sched_in (struct perf_event *event) -{ - event->state =PERF_EVENT_STATE_ACTIVE ; - event->oncpu =smp_processor_id (); - event->tstamp_running += event->ctx->time - event->tstamp_stopped; -if (is_software_event (event)) - event->pmu->enable (event); -} - -int hw_perf_group_sched_in (struct perf_event *group_leader, -struct perf_cpu_context *cpuctx, -struct perf_event_context *ctx) -{ -struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events); -struct perf_event *sub; -int n0, n; - -if (!sparc_pmu) -return 0 ; - - n0 = cpuc->n_events; - n =collect_events (group_leader, perf_max_events - n0, - &cpuc->event[n0], &cpuc->events[n0], - &cpuc->current_idx[n0]); -if (n <0 ) -return -EAGAIN ; -if (check_excludes (cpuc->event, n0, n)) -return -EINVAL ; -if (sparc_check_constraints (cpuc->event, cpuc->events, n + n0)) -return -EAGAIN ; - cpuc->n_events = n0 + n; - cpuc->n_added += n; - - cpuctx->active_oncpu += n; - n =1 ; -event_sched_in (group_leader); -list_for_each_entry (sub, &group_leader->sibling_list, group_entry) { -if (sub->state !=PERF_EVENT_STATE_OFF ) { -event_sched_in (sub); - n++; - } - } - ctx->nr_active += n; - -return 1 ; -} -static int sparc_pmu_enable (struct perf_event *event) {struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events);
@@ -1044,11 +999,20 @@ static int sparc_pmu_enable(struct perf_event *event) cpuc->events[n0] = event->hw.event_base; cpuc->current_idx[n0] =PIC_NO_INDEX ; +/* +* If group events scheduling transaction was started, +* skip the schedulability test here, it will be peformed +* at commit time(->commit_txn) as a whole +*/ +if (cpuc->group_flag &PERF_EVENT_TXN_STARTED ) + goto nocheck; +if (check_excludes (cpuc->event, n0,1 )) goto out;if (sparc_check_constraints (cpuc->event, cpuc->events, n0 +1 )) goto out; +nocheck: cpuc->n_events++; cpuc->n_added++;
@@ -1128,11 +1092,61 @@ static int __hw_perf_event_init(struct perf_event *event)return 0 ; } +/* +* Start group events scheduling transaction +* Set the flag to make pmu::enable() not perform the +* schedulability test, it will be performed at commit time +*/ +static void sparc_pmu_start_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuhw = &__get_cpu_var (cpu_hw_events); + + cpuhw->group_flag |=PERF_EVENT_TXN_STARTED ; +} + +/* +* Stop group events scheduling transaction +* Clear the flag and pmu::enable() will perform the +* schedulability test. +*/ +static void sparc_pmu_cancel_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuhw = &__get_cpu_var (cpu_hw_events); + + cpuhw->group_flag &= ~PERF_EVENT_TXN_STARTED ; +} + +/* +* Commit group events scheduling transaction +* Perform the group schedulability test as a whole +* Return 0 if success +*/ +static int sparc_pmu_commit_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events); +int n; + +if (!sparc_pmu) +return -EINVAL ; + + cpuc = &__get_cpu_var (cpu_hw_events); + n = cpuc->n_events; +if (check_excludes (cpuc->event,0 , n)) +return -EINVAL ; +if (sparc_check_constraints (cpuc->event, cpuc->events, n)) +return -EAGAIN ; + +return 0 ; +} +static const struct pmu pmu = { .enable = sparc_pmu_enable, .disable = sparc_pmu_disable, .read = sparc_pmu_read, .unthrottle = sparc_pmu_unthrottle, + .start_txn = sparc_pmu_start_txn, + .cancel_txn = sparc_pmu_cancel_txn, + .commit_txn = sparc_pmu_commit_txn, };const struct pmu *hw_perf_event_init (struct perf_event *event)
arch/x86/include/asm/perf_event_p4.h +2 -1
@@ -89,7 +89,8 @@ P4_CCCR_ENABLE)/* HT mask */ -#define P4_CCCR_MASK_HT (P4_CCCR_MASK | P4_CCCR_THREAD_ANY) +#define P4_CCCR_MASK_HT \ + (P4_CCCR_MASK | P4_CCCR_OVF_PMI_T1 | P4_CCCR_THREAD_ANY)#define P4_GEN_ESCR_EMASK (class, name, bit) \ class##__##name = ((1 << bit) << P4_ESCR_EVENTMASK_SHIFT)
arch/x86/kernel/cpu/perf_event_p4.c +21 -20
79 lines not on this page · raw message
arch/x86/mm/pf_in.c +1 -1
@@ -34,7 +34,7 @@/* IA32 Manual 3, 2-1 */ static unsigned char prefix_codes[] = {0xF0 ,0xF2 ,0xF3 ,0x2E ,0x36 ,0x3E ,0x26 ,0x64 , -0x65 ,0x2E ,0x3E ,0x66 ,0x67 +0x65 ,0x66 ,0x67 };/* IA32 Manual 3, 3-432*/ static unsigned int reg_rop[] = {
include/linux/ftrace_event.h +75 -28
145 lines not on this page · raw message
include/linux/perf_event.h +15 -13
@@ -485,6 +485,7 @@ struct perf_guest_info_callbacks {#include <linux/ftrace.h> #include <linux/cpu.h> #include <asm/atomic.h> +#include <asm/local.h> #define PERF_MAX_STACK_DEPTH 255
@@ -587,21 +588,19 @@ struct perf_mmap_data {struct rcu_head rcu_head;#ifdef CONFIG_PERF_USE_VMALLOC struct work_struct work; +int page_order;/* allocation order */ #endif -int data_order;int nr_pages;/* nr of data pages */ int writable;/* are we writable */ int nr_locked;/* nr pages mlocked */ atomic_t poll;/* POLL_ for wakeups */ -atomic_t events;/* event_id limit */ -atomic_long_t head;/* write position */ -atomic_long_t done_head;/* completed head */ - -atomic_t lock;/* concurrent writes */ -atomic_t wakeup;/* needs a wakeup */ -atomic_t lost;/* nr records lost */ +local_t head;/* write position */ +local_t nest;/* nested writers */ +local_t events;/* event limit */ +local_t wakeup;/* wakeup stamp */ +local_t lost;/* nr records lost */ long watermark;/* wakeup watermark */
@@ -728,6 +727,7 @@ struct perf_event {perf_overflow_handler_t overflow_handler;#ifdef CONFIG_EVENT_TRACING +struct ftrace_event_call *tp_event;struct event_filter *filter;#endif
@@ -803,11 +803,12 @@ struct perf_cpu_context {struct perf_output_handle {struct perf_event *event;struct perf_mmap_data *data; -unsigned long head; -unsigned long offset; +unsigned long wakeup; +unsigned long size; +void *addr; +int page;int nmi;int sample; -int locked; };#ifdef CONFIG_PERF_EVENTS
@@ -993,8 +994,9 @@ static inline bool perf_paranoid_kernel(void) }extern void perf_event_init (void ); -extern void perf_tp_event (int event_id,u64 addr,u64 count,void *record, -int entry_size,struct pt_regs *regs); +extern void perf_tp_event (u64 addr,u64 count,void *record, +int entry_size,struct pt_regs *regs, +struct hlist_head *head);extern void perf_bp_event (struct perf_event *event,void *data);#ifndef perf_misc_flags
include/linux/syscalls.h +17 -40
111 lines not on this page · raw message
include/linux/tracepoint.h +76 -22
169 lines not on this page · raw message
include/trace/ftrace.h +90 -157
429 lines not on this page · raw message
include/trace/syscall.h +6 -4
@@ -25,6 +25,8 @@ struct syscall_metadata {int nb_args;const char **types;const char **args; +struct list_head enter_fields; +struct list_head exit_fields;struct ftrace_event_call *enter_event;struct ftrace_event_call *exit_event;
@@ -34,16 +36,16 @@ struct syscall_metadata {extern unsigned long arch_syscall_addr (int nr);extern int init_syscall_trace (struct ftrace_event_call *call); -extern int syscall_enter_define_fields (struct ftrace_event_call *call); -extern int syscall_exit_define_fields (struct ftrace_event_call *call);extern int reg_event_syscall_enter (struct ftrace_event_call *call);extern void unreg_event_syscall_enter (struct ftrace_event_call *call);extern int reg_event_syscall_exit (struct ftrace_event_call *call);extern void unreg_event_syscall_exit (struct ftrace_event_call *call);extern int ftrace_format_syscall (struct ftrace_event_call *call,struct trace_seq *s); -enum print_line_t print_syscall_enter (struct trace_iterator *iter,int flags); -enum print_line_t print_syscall_exit (struct trace_iterator *iter,int flags); +enum print_line_t print_syscall_enter (struct trace_iterator *iter,int flags, +struct trace_event *event); +enum print_line_t print_syscall_exit (struct trace_iterator *iter,int flags, +struct trace_event *event);#endif #ifdef CONFIG_PERF_EVENTS
kernel/perf_event.c +217 -172
699 lines not on this page · raw message
kernel/trace/blktrace.c +79 -59
281 lines not on this page · raw message
kernel/trace/ftrace.c +4 -3
@@ -3234,7 +3234,8 @@ free: }static void -ftrace_graph_probe_sched_switch (struct task_struct *prev,struct task_struct *next) +ftrace_graph_probe_sched_switch (void *ignore, +struct task_struct *prev,struct task_struct *next) {unsigned long long timestamp;int index;
@@ -3288,7 +3289,7 @@ static int start_graph_tracing(void) }while (ret == -EAGAIN );if (!ret) { - ret =register_trace_sched_switch (ftrace_graph_probe_sched_switch); + ret =register_trace_sched_switch (ftrace_graph_probe_sched_switch,NULL );if (ret)pr_info ("ftrace_graph: Couldn't activate tracepoint" " probe to kernel_sched_switch\n" );
@@ -3364,7 +3365,7 @@ void unregister_ftrace_graph(void) ftrace_graph_entry = ftrace_graph_entry_stub;ftrace_shutdown (FTRACE_STOP_FUNC_RET );unregister_pm_notifier (&ftrace_suspend_notifier); -unregister_trace_sched_switch (ftrace_graph_probe_sched_switch); +unregister_trace_sched_switch (ftrace_graph_probe_sched_switch,NULL ); out:mutex_unlock (&ftrace_lock);
kernel/trace/kmemtrace.c +44 -26
164 lines not on this page · raw message
kernel/trace/trace.c +5 -4
@@ -1936,7 +1936,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) }if (event) -return event->trace (iter, sym_flags); +return event->funcs->trace (iter, sym_flags, event);if (!trace_seq_printf (s,"Unknown type %d\n" , entry->type)) goto partial;
@@ -1962,7 +1962,7 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) event =ftrace_find_event (entry->type);if (event) -return event->raw (iter,0 ); +return event->funcs->raw (iter,0 , event);if (!trace_seq_printf (s,"%d ?\n" , entry->type)) goto partial;
@@ -1989,7 +1989,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) event =ftrace_find_event (entry->type);if (event) { -enum print_line_t ret = event->hex (iter,0 ); +enum print_line_t ret = event->funcs->hex (iter,0 , event);if (ret !=TRACE_TYPE_HANDLED )return ret; }
@@ -2014,7 +2014,8 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) } event =ftrace_find_event (entry->type); -return event ? event->binary (iter,0 ) :TRACE_TYPE_HANDLED ; +return event ? event->funcs->binary (iter,0 , event) : +TRACE_TYPE_HANDLED ; }int trace_empty (struct trace_iterator *iter)
kernel/trace/trace.h +6 -3
@@ -405,12 +405,12 @@ void ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags,void __trace_stack (struct trace_array *tr,unsigned long flags,int skip,int pc);#else -static inline void ftrace_trace_stack (struct trace_array *tr, +static inline void ftrace_trace_stack (struct ring_buffer *buffer,unsigned long flags,int skip,int pc) { } -static inline void ftrace_trace_userstack (struct trace_array *tr, +static inline void ftrace_trace_userstack (struct ring_buffer *buffer,unsigned long flags,int pc) { }
@@ -778,12 +778,15 @@ extern void print_subsystem_event_filter(struct event_subsystem *system, struct trace_seq *s);extern int filter_assign_type (const char *type); +struct list_head * +trace_get_fields (struct ftrace_event_call *event_call); +static inline int filter_check_discard (struct ftrace_event_call *call,void *rec,struct ring_buffer *buffer,struct ring_buffer_event *event) { -if (unlikely (call->filter_active) && +if (unlikely (call->flags &TRACE_EVENT_FL_FILTERED ) && !filter_match_preds (call->filter, rec)) {ring_buffer_discard_commit (buffer, event);return 1 ;
kernel/trace/trace_branch.c +6 -2
@@ -143,7 +143,7 @@ static void branch_trace_reset(struct trace_array *tr) }static enum print_line_t trace_branch_print (struct trace_iterator *iter, -int flags) +int flags,struct trace_event *event) {struct trace_branch *field;
@@ -167,9 +167,13 @@ static void branch_print_header(struct seq_file *s) " |\n"); } +static struct trace_event_functions trace_branch_funcs = { + .trace = trace_branch_print, +}; +static struct trace_event trace_branch_event = { .type =TRACE_BRANCH , - .trace = trace_branch_print, + .funcs = &trace_branch_funcs, }; static struct tracer branch_trace __read_mostly =
kernel/trace/trace_event_perf.c +95 -88
255 lines not on this page · raw message
kernel/trace/trace_events.c +88 -51
304 lines not on this page · raw message
kernel/trace/trace_kprobe.c +67 -46
113 lines not on this page · raw message
kernel/trace/trace_output.c +92 -45
137 lines not on this page · raw message
kernel/trace/trace_syscalls.c +120 -26
146 lines not on this page · raw message
kernel/tracepoint.c +51 -40
91 lines not on this page · raw message
samples/tracepoints/tracepoint-probe-sample.c +7 -6
Lines of this file are not shown. View the raw message.
samples/tracepoints/tracepoint-probe-sample2.c +4 -3
Lines of this file are not shown. View the raw message.
tools/perf/builtin-annotate.c +45 -16
61 lines not on this page · raw message
tools/perf/builtin-record.c +25 -47
72 lines not on this page · raw message
tools/perf/builtin-report.c +38 -26
64 lines not on this page · raw message
tools/perf/util/abspath.c +0 -81
81 lines not on this page · raw message
tools/perf/util/cache.h +2 -55
57 lines not on this page · raw message
tools/perf/util/config.c +6 -455
461 lines not on this page · raw message
tools/perf/util/header.c +52 -32
84 lines not on this page · raw message
tools/perf/util/hist.c +33 -9
42 lines not on this page · raw message
tools/perf/util/newt.c +116 -34
150 lines not on this page · raw message
tools/perf/util/path.c +1 -203
204 lines not on this page · raw message
tools/perf/util/quote.c +1 -432
433 lines not on this page · raw message
tools/perf/util/run-command.c +0 -90
90 lines not on this page · raw message
tools/perf/util/strbuf.c +1 -228
229 lines not on this page · raw message
tools/perf/util/strbuf.h +0 -45
45 lines not on this page · raw message
tools/perf/util/util.h +14 -163
177 lines not on this page · raw message
tools/perf/util/wrapper.c +0 -110
110 lines not on this page · raw message
The rest of this diff is not shown. View the raw message.
Part of the diff is not shown here. View the raw message.
Quoted 9 lines · Ingo Molnar
Please pull the latest perf-core-for-linus git tree from:
git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.gitperf-core-for-linusMost of these changes are fixes and dead code removal, but also a set of optimizations and a series of tracing size-optimization bits that we kept pending because they conflicted with the previous push - and some TUI enhancements.
...
Quoted 8 lines
Peter Zijlstra (16):
perf/ftrace: Optimize perf/tracepoint interaction for single events
perf: Optimize buffer placement by allocating buffers NUMA aware
perf: Disallowmmap()on per-task inherited events
perf: Optimize theperf_output()path by removing IRQ-disables
perf: Optimize the hotpath by converting the perf output buffer tolocal_t
perf: Optimize perf_output_*() by avoidinglocal_xchg()
perf, trace: Optimize tracepoints by removing IRQ-disable from perf/tracepoint interaction
This commit breaks all non-x86 architectures by adding a reference to regs->flags in generic code. I mentioned this on lkml (with cc to Ingo and Peter), and Peter has a further patch that fixes the problem, which isn't in this pull.
So Linus, please don't do this pull just yet, until we get this extra patch merged.
Paul.
Quoted 10 lines · Paul Mackerras
Quoted 8 lines
Peter Zijlstra (16):
perf/ftrace: Optimize perf/tracepoint interaction for single events
perf: Optimize buffer placement by allocating buffers NUMA aware
perf: Disallowmmap()on per-task inherited events
perf: Optimize theperf_output()path by removing IRQ-disables
perf: Optimize the hotpath by converting the perf output buffer tolocal_t
perf: Optimize perf_output_*() by avoidinglocal_xchg()
perf, trace: Optimize tracepoints by removing IRQ-disable from perf/tracepoint interactionThis commit breaks all non-x86 architectures [...]
Indeed. I have merged and tested the fix, Linus, please pull the updated perf-core-for-linus git tree from:
git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git perf-core-for-linus
Thanks,
Ingo
------------------>
Akinobu Mita (1):
x86/mmiotrace: Remove redundant instruction prefix checks
Arnaldo Carvalho de Melo (15):
perf tui: Fix build problem with slang <= 2.0.6
perf tools: Remove some unused functions
perf probe: Fix some error exit paths
perf probe: Don't call die()
perf tools: remove xstrndup, xmalloc, xzalloc
perf symbols: Don't try to read the build-id twice
perf session: Make read_build_id routines look at the host_machine too perf TUI: Make 'space' be an alias to 'PgDn'
perf annotate: Use build-ids to find the right DSO
perf tui: Allow disabling the TUI on a per command basis in ~/.perfconfig perf report: Don't start the TUI if -D is used
perf tui: Remove annotate from popup menu after failure
perf annotate: Add TUI interface
perf annotate: Fix up usage of the build id cache
perf report: Support multiple events on the TUI
Cyrill Gorcunov (5):
perf, x86: P4 PMU -- handle unflagged events
perf, x86: P4 PMU -- fix typo in unflagged NMI handling perf, x86: P4 PMU -- do a real check for ESCR address being in hash perf, x86: P4_pmu_schedule_events -- use smp_processor_id instead of raw_ perf, x86: P4 PMU -- add missing bit in CCCR mask
Frederic Weisbecker (4):
perf: Comply with new rcu checks API
perf: Fix unaligned accesses while fetching trace values
perf: Fix forgotten preempt_enable by nested writers
perf: Fix getline undeclared
Li Zefan (1):
tracing: Fix function declarations if !CONFIG_STACKTRACE
Lin Ming (1):
perf, sparc: Implement group scheduling transactional APIs
Masami Hiramatsu (1):
perf probe: Don't compile CFI related code if elfutils is old
Mathieu Desnoyers (1):
tracepoints: Add check trace callback type
Peter Zijlstra (17):
perf/ftrace: Optimize perf/tracepoint interaction for single events
perf: Optimize buffer placement by allocating buffers NUMA aware
perf: Disallow mmap() on per-task inherited events
perf: Optimize the perf_output() path by removing IRQ-disables
perf: Optimize the hotpath by converting the perf output buffer to local_t
perf: Optimize perf_output_*() by avoiding local_xchg()
perf, trace: Optimize tracepoints by removing IRQ-disable from perf/tracepoint interaction perf, trace: Optimize tracepoints by using per-tracepoint-per-cpu hlist to track events
perf: Ensure that IOC_OUTPUT isn't used to create multi-writer buffers
perf-record: Remove -M
perf-record: Share per-cpu buffers
perf: Fix wakeup storm for RO mmap()s
perf: Optimize perf_output_copy()
perf: Optimize the !vmalloc backed buffer
perf: Remove more code from the fastpath
perf: Optimize perf_tp_event_match()
perf, trace: Fix !x86 build bug
Russ Anderson (1):
perf record: remove unneeded gettimeofday() call
Stephane Eranian (2):
perf: Fix errors path in perf_output_begin() perf stat: add perf stat -B to pretty print large numbers
Steven Rostedt (11):
tracing: Create class struct for events
tracing: Let tracepoints have data passed to tracepoint callbacks
tracing: Remove per event trace registering
tracing: Move fields from event to class structure
tracing: Move raw_init from events to class
tracing: Allow events to share their print functions
tracing: Move print functions into event class
tracing: Remove duplicate id information in event structure
tracing: Combine event filter_active and enable into single flags field
tracing: Comment the use of event_mutex with trace event flags
tracing: Add __used annotation to event variable
Tom Zanussi (1):
perf: Use read() instead of lseek() in trace_event_read.c:skip()
arch/sparc/kernel/perf_event.c | 108 ++++--- arch/x86/include/asm/perf_event_p4.h | 3 +- arch/x86/kernel/cpu/perf_event.c | 6 +- arch/x86/kernel/cpu/perf_event_p4.c | 41 ++- arch/x86/mm/pf_in.c | 2 +- include/linux/ftrace_event.h | 103 ++++-- include/linux/perf_event.h | 28 +- include/linux/syscalls.h | 57 +-- include/linux/tracepoint.h | 98 ++++-- include/trace/ftrace.h | 249 +++++-------- include/trace/syscall.h | 10 +- kernel/perf_event.c | 389 +++++++++++--------- kernel/trace/blktrace.c | 138 ++++--- kernel/trace/ftrace.c | 7 +- kernel/trace/kmemtrace.c | 70 +++-- kernel/trace/trace.c | 9 +- kernel/trace/trace.h | 9 +- kernel/trace/trace_branch.c | 8 +- kernel/trace/trace_event_perf.c | 185 +++++----- kernel/trace/trace_events.c | 139 +++++--- kernel/trace/trace_events_filter.c | 28 +- kernel/trace/trace_export.c | 16 +- kernel/trace/trace_functions_graph.c | 13 +- kernel/trace/trace_kprobe.c | 113 ++++--- kernel/trace/trace_output.c | 137 +++++--- kernel/trace/trace_output.h | 2 +- kernel/trace/trace_sched_switch.c | 20 +- kernel/trace/trace_sched_wakeup.c | 28 +- kernel/trace/trace_syscalls.c | 146 ++++++-- kernel/trace/trace_workqueue.c | 26 +- kernel/tracepoint.c | 91 +++-- net/core/drop_monitor.c | 12 +- samples/tracepoints/tp-samples-trace.h | 4 +- samples/tracepoints/tracepoint-probe-sample.c | 13 +- samples/tracepoints/tracepoint-probe-sample2.c | 7 +- tools/perf/Documentation/perf-stat.txt | 3 + tools/perf/builtin-annotate.c | 61 +++- tools/perf/builtin-probe.c | 10 +- tools/perf/builtin-record.c | 72 ++--- tools/perf/builtin-report.c | 64 ++-- tools/perf/builtin-stat.c | 18 +- tools/perf/perf.c | 25 ++- tools/perf/util/abspath.c | 81 ----- tools/perf/util/build-id.c | 22 ++ tools/perf/util/build-id.h | 2 + tools/perf/util/cache.h | 57 +--- tools/perf/util/callchain.c | 1 + tools/perf/util/callchain.h | 1 - tools/perf/util/config.c | 461 +----------------------- tools/perf/util/exec_cmd.c | 6 +- tools/perf/util/exec_cmd.h | 1 - tools/perf/util/header.c | 84 +++-- tools/perf/util/help.c | 30 +-- tools/perf/util/hist.c | 42 ++- tools/perf/util/hist.h | 24 ++- tools/perf/util/newt.c | 150 ++++++-- tools/perf/util/path.c | 204 +----------- tools/perf/util/probe-finder.c | 33 ++- tools/perf/util/probe-finder.h | 3 + tools/perf/util/quote.c | 433 +---------------------- tools/perf/util/quote.h | 39 -- tools/perf/util/run-command.c | 90 ----- tools/perf/util/run-command.h | 30 -- tools/perf/util/session.c | 8 + tools/perf/util/session.h | 8 +- tools/perf/util/sigchain.c | 2 +- tools/perf/util/sigchain.h | 1 - tools/perf/util/strbuf.c | 229 +------------ tools/perf/util/strbuf.h | 45 --- tools/perf/util/symbol.c | 25 +- tools/perf/util/symbol.h | 1 + tools/perf/util/trace-event-read.c | 19 +- tools/perf/util/trace-event.h | 7 +- tools/perf/util/util.h | 177 +--------- tools/perf/util/wrapper.c | 110 ------ 75 files changed, 1847 insertions(+), 3147 deletions(-)
arch/sparc/kernel/perf_event.c +61 -47
@@ -91,6 +91,8 @@ struct cpu_hw_events {/* Enabled/disable state. */ int enabled; + +unsigned int group_flag; }; DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { .enabled = 1, };
@@ -980,53 +982,6 @@ static int collect_events(struct perf_event *group, int max_count,return n; } -static void event_sched_in (struct perf_event *event) -{ - event->state =PERF_EVENT_STATE_ACTIVE ; - event->oncpu =smp_processor_id (); - event->tstamp_running += event->ctx->time - event->tstamp_stopped; -if (is_software_event (event)) - event->pmu->enable (event); -} - -int hw_perf_group_sched_in (struct perf_event *group_leader, -struct perf_cpu_context *cpuctx, -struct perf_event_context *ctx) -{ -struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events); -struct perf_event *sub; -int n0, n; - -if (!sparc_pmu) -return 0 ; - - n0 = cpuc->n_events; - n =collect_events (group_leader, perf_max_events - n0, - &cpuc->event[n0], &cpuc->events[n0], - &cpuc->current_idx[n0]); -if (n <0 ) -return -EAGAIN ; -if (check_excludes (cpuc->event, n0, n)) -return -EINVAL ; -if (sparc_check_constraints (cpuc->event, cpuc->events, n + n0)) -return -EAGAIN ; - cpuc->n_events = n0 + n; - cpuc->n_added += n; - - cpuctx->active_oncpu += n; - n =1 ; -event_sched_in (group_leader); -list_for_each_entry (sub, &group_leader->sibling_list, group_entry) { -if (sub->state !=PERF_EVENT_STATE_OFF ) { -event_sched_in (sub); - n++; - } - } - ctx->nr_active += n; - -return 1 ; -} -static int sparc_pmu_enable (struct perf_event *event) {struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events);
@@ -1044,11 +999,20 @@ static int sparc_pmu_enable(struct perf_event *event) cpuc->events[n0] = event->hw.event_base; cpuc->current_idx[n0] =PIC_NO_INDEX ; +/* +* If group events scheduling transaction was started, +* skip the schedulability test here, it will be peformed +* at commit time(->commit_txn) as a whole +*/ +if (cpuc->group_flag &PERF_EVENT_TXN_STARTED ) + goto nocheck; +if (check_excludes (cpuc->event, n0,1 )) goto out;if (sparc_check_constraints (cpuc->event, cpuc->events, n0 +1 )) goto out; +nocheck: cpuc->n_events++; cpuc->n_added++;
@@ -1128,11 +1092,61 @@ static int __hw_perf_event_init(struct perf_event *event)return 0 ; } +/* +* Start group events scheduling transaction +* Set the flag to make pmu::enable() not perform the +* schedulability test, it will be performed at commit time +*/ +static void sparc_pmu_start_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuhw = &__get_cpu_var (cpu_hw_events); + + cpuhw->group_flag |=PERF_EVENT_TXN_STARTED ; +} + +/* +* Stop group events scheduling transaction +* Clear the flag and pmu::enable() will perform the +* schedulability test. +*/ +static void sparc_pmu_cancel_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuhw = &__get_cpu_var (cpu_hw_events); + + cpuhw->group_flag &= ~PERF_EVENT_TXN_STARTED ; +} + +/* +* Commit group events scheduling transaction +* Perform the group schedulability test as a whole +* Return 0 if success +*/ +static int sparc_pmu_commit_txn (const struct pmu *pmu) +{ +struct cpu_hw_events *cpuc = &__get_cpu_var (cpu_hw_events); +int n; + +if (!sparc_pmu) +return -EINVAL ; + + cpuc = &__get_cpu_var (cpu_hw_events); + n = cpuc->n_events; +if (check_excludes (cpuc->event,0 , n)) +return -EINVAL ; +if (sparc_check_constraints (cpuc->event, cpuc->events, n)) +return -EAGAIN ; + +return 0 ; +} +static const struct pmu pmu = { .enable = sparc_pmu_enable, .disable = sparc_pmu_disable, .read = sparc_pmu_read, .unthrottle = sparc_pmu_unthrottle, + .start_txn = sparc_pmu_start_txn, + .cancel_txn = sparc_pmu_cancel_txn, + .commit_txn = sparc_pmu_commit_txn, };const struct pmu *hw_perf_event_init (struct perf_event *event)
arch/x86/include/asm/perf_event_p4.h +2 -1
@@ -89,7 +89,8 @@ P4_CCCR_ENABLE)/* HT mask */ -#define P4_CCCR_MASK_HT (P4_CCCR_MASK | P4_CCCR_THREAD_ANY) +#define P4_CCCR_MASK_HT \ + (P4_CCCR_MASK | P4_CCCR_OVF_PMI_T1 | P4_CCCR_THREAD_ANY)#define P4_GEN_ESCR_EMASK (class, name, bit) \ class##__##name = ((1 << bit) << P4_ESCR_EVENTMASK_SHIFT)
arch/x86/kernel/cpu/perf_event.c +5 -1
@@ -1717,7 +1717,11 @@ void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int ski*/ regs->bp =rewind_frame_pointer (skip +1 ); regs->cs = __KERNEL_CS; -local_save_flags (regs->flags); +/* +* We abuse bit 3 to pass exact information, see perf_misc_flags +* and the comment with PERF_EFLAGS_EXACT. +*/ + regs->flags =0 ; }unsigned long perf_instruction_pointer (struct pt_regs *regs)
arch/x86/kernel/cpu/perf_event_p4.c +21 -20
79 lines not on this page · raw message
arch/x86/mm/pf_in.c +1 -1
@@ -34,7 +34,7 @@/* IA32 Manual 3, 2-1 */ static unsigned char prefix_codes[] = {0xF0 ,0xF2 ,0xF3 ,0x2E ,0x36 ,0x3E ,0x26 ,0x64 , -0x65 ,0x2E ,0x3E ,0x66 ,0x67 +0x65 ,0x66 ,0x67 };/* IA32 Manual 3, 3-432*/ static unsigned int reg_rop[] = {
include/linux/ftrace_event.h +75 -28
145 lines not on this page · raw message
include/linux/perf_event.h +15 -13
@@ -485,6 +485,7 @@ struct perf_guest_info_callbacks {#include <linux/ftrace.h> #include <linux/cpu.h> #include <asm/atomic.h> +#include <asm/local.h> #define PERF_MAX_STACK_DEPTH 255
@@ -587,21 +588,19 @@ struct perf_mmap_data {struct rcu_head rcu_head;#ifdef CONFIG_PERF_USE_VMALLOC struct work_struct work; +int page_order;/* allocation order */ #endif -int data_order;int nr_pages;/* nr of data pages */ int writable;/* are we writable */ int nr_locked;/* nr pages mlocked */ atomic_t poll;/* POLL_ for wakeups */ -atomic_t events;/* event_id limit */ -atomic_long_t head;/* write position */ -atomic_long_t done_head;/* completed head */ - -atomic_t lock;/* concurrent writes */ -atomic_t wakeup;/* needs a wakeup */ -atomic_t lost;/* nr records lost */ +local_t head;/* write position */ +local_t nest;/* nested writers */ +local_t events;/* event limit */ +local_t wakeup;/* wakeup stamp */ +local_t lost;/* nr records lost */ long watermark;/* wakeup watermark */
@@ -728,6 +727,7 @@ struct perf_event {perf_overflow_handler_t overflow_handler;#ifdef CONFIG_EVENT_TRACING +struct ftrace_event_call *tp_event;struct event_filter *filter;#endif
@@ -803,11 +803,12 @@ struct perf_cpu_context {struct perf_output_handle {struct perf_event *event;struct perf_mmap_data *data; -unsigned long head; -unsigned long offset; +unsigned long wakeup; +unsigned long size; +void *addr; +int page;int nmi;int sample; -int locked; };#ifdef CONFIG_PERF_EVENTS
@@ -993,8 +994,9 @@ static inline bool perf_paranoid_kernel(void) }extern void perf_event_init (void ); -extern void perf_tp_event (int event_id,u64 addr,u64 count,void *record, -int entry_size,struct pt_regs *regs); +extern void perf_tp_event (u64 addr,u64 count,void *record, +int entry_size,struct pt_regs *regs, +struct hlist_head *head);extern void perf_bp_event (struct perf_event *event,void *data);#ifndef perf_misc_flags
include/linux/syscalls.h +17 -40
111 lines not on this page · raw message
include/linux/tracepoint.h +76 -22
169 lines not on this page · raw message
include/trace/ftrace.h +91 -158
433 lines not on this page · raw message
include/trace/syscall.h +6 -4
@@ -25,6 +25,8 @@ struct syscall_metadata {int nb_args;const char **types;const char **args; +struct list_head enter_fields; +struct list_head exit_fields;struct ftrace_event_call *enter_event;struct ftrace_event_call *exit_event;
@@ -34,16 +36,16 @@ struct syscall_metadata {extern unsigned long arch_syscall_addr (int nr);extern int init_syscall_trace (struct ftrace_event_call *call); -extern int syscall_enter_define_fields (struct ftrace_event_call *call); -extern int syscall_exit_define_fields (struct ftrace_event_call *call);extern int reg_event_syscall_enter (struct ftrace_event_call *call);extern void unreg_event_syscall_enter (struct ftrace_event_call *call);extern int reg_event_syscall_exit (struct ftrace_event_call *call);extern void unreg_event_syscall_exit (struct ftrace_event_call *call);extern int ftrace_format_syscall (struct ftrace_event_call *call,struct trace_seq *s); -enum print_line_t print_syscall_enter (struct trace_iterator *iter,int flags); -enum print_line_t print_syscall_exit (struct trace_iterator *iter,int flags); +enum print_line_t print_syscall_enter (struct trace_iterator *iter,int flags, +struct trace_event *event); +enum print_line_t print_syscall_exit (struct trace_iterator *iter,int flags, +struct trace_event *event);#endif #ifdef CONFIG_PERF_EVENTS
kernel/perf_event.c +217 -172
699 lines not on this page · raw message
kernel/trace/blktrace.c +79 -59
281 lines not on this page · raw message
kernel/trace/ftrace.c +4 -3
@@ -3234,7 +3234,8 @@ free: }static void -ftrace_graph_probe_sched_switch (struct task_struct *prev,struct task_struct *next) +ftrace_graph_probe_sched_switch (void *ignore, +struct task_struct *prev,struct task_struct *next) {unsigned long long timestamp;int index;
@@ -3288,7 +3289,7 @@ static int start_graph_tracing(void) }while (ret == -EAGAIN );if (!ret) { - ret =register_trace_sched_switch (ftrace_graph_probe_sched_switch); + ret =register_trace_sched_switch (ftrace_graph_probe_sched_switch,NULL );if (ret)pr_info ("ftrace_graph: Couldn't activate tracepoint" " probe to kernel_sched_switch\n" );
@@ -3364,7 +3365,7 @@ void unregister_ftrace_graph(void) ftrace_graph_entry = ftrace_graph_entry_stub;ftrace_shutdown (FTRACE_STOP_FUNC_RET );unregister_pm_notifier (&ftrace_suspend_notifier); -unregister_trace_sched_switch (ftrace_graph_probe_sched_switch); +unregister_trace_sched_switch (ftrace_graph_probe_sched_switch,NULL ); out:mutex_unlock (&ftrace_lock);
kernel/trace/kmemtrace.c +44 -26
164 lines not on this page · raw message
kernel/trace/trace.c +5 -4
@@ -1936,7 +1936,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) }if (event) -return event->trace (iter, sym_flags); +return event->funcs->trace (iter, sym_flags, event);if (!trace_seq_printf (s,"Unknown type %d\n" , entry->type)) goto partial;
@@ -1962,7 +1962,7 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) event =ftrace_find_event (entry->type);if (event) -return event->raw (iter,0 ); +return event->funcs->raw (iter,0 , event);if (!trace_seq_printf (s,"%d ?\n" , entry->type)) goto partial;
@@ -1989,7 +1989,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) event =ftrace_find_event (entry->type);if (event) { -enum print_line_t ret = event->hex (iter,0 ); +enum print_line_t ret = event->funcs->hex (iter,0 , event);if (ret !=TRACE_TYPE_HANDLED )return ret; }
@@ -2014,7 +2014,8 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) } event =ftrace_find_event (entry->type); -return event ? event->binary (iter,0 ) :TRACE_TYPE_HANDLED ; +return event ? event->funcs->binary (iter,0 , event) : +TRACE_TYPE_HANDLED ; }int trace_empty (struct trace_iterator *iter)
kernel/trace/trace.h +6 -3
@@ -405,12 +405,12 @@ void ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags,void __trace_stack (struct trace_array *tr,unsigned long flags,int skip,int pc);#else -static inline void ftrace_trace_stack (struct trace_array *tr, +static inline void ftrace_trace_stack (struct ring_buffer *buffer,unsigned long flags,int skip,int pc) { } -static inline void ftrace_trace_userstack (struct trace_array *tr, +static inline void ftrace_trace_userstack (struct ring_buffer *buffer,unsigned long flags,int pc) { }
@@ -778,12 +778,15 @@ extern void print_subsystem_event_filter(struct event_subsystem *system, struct trace_seq *s);extern int filter_assign_type (const char *type); +struct list_head * +trace_get_fields (struct ftrace_event_call *event_call); +static inline int filter_check_discard (struct ftrace_event_call *call,void *rec,struct ring_buffer *buffer,struct ring_buffer_event *event) { -if (unlikely (call->filter_active) && +if (unlikely (call->flags &TRACE_EVENT_FL_FILTERED ) && !filter_match_preds (call->filter, rec)) {ring_buffer_discard_commit (buffer, event);return 1 ;
kernel/trace/trace_branch.c +6 -2
@@ -143,7 +143,7 @@ static void branch_trace_reset(struct trace_array *tr) }static enum print_line_t trace_branch_print (struct trace_iterator *iter, -int flags) +int flags,struct trace_event *event) {struct trace_branch *field;
@@ -167,9 +167,13 @@ static void branch_print_header(struct seq_file *s) " |\n"); } +static struct trace_event_functions trace_branch_funcs = { + .trace = trace_branch_print, +}; +static struct trace_event trace_branch_event = { .type =TRACE_BRANCH , - .trace = trace_branch_print, + .funcs = &trace_branch_funcs, }; static struct tracer branch_trace __read_mostly =
kernel/trace/trace_event_perf.c +97 -88
257 lines not on this page · raw message
kernel/trace/trace_events.c +88 -51
286 lines not on this page · raw message
kernel/trace/trace_kprobe.c +67 -46
113 lines not on this page · raw message
kernel/trace/trace_output.c +92 -45
137 lines not on this page · raw message
kernel/trace/trace_syscalls.c +120 -26
146 lines not on this page · raw message
kernel/tracepoint.c +51 -40
91 lines not on this page · raw message
samples/tracepoints/tracepoint-probe-sample.c +7 -6
Lines of this file are not shown. View the raw message.
samples/tracepoints/tracepoint-probe-sample2.c +4 -3
Lines of this file are not shown. View the raw message.
tools/perf/builtin-annotate.c +45 -16
61 lines not on this page · raw message
tools/perf/builtin-record.c +25 -47
72 lines not on this page · raw message
tools/perf/builtin-report.c +38 -26
64 lines not on this page · raw message
tools/perf/util/abspath.c +0 -81
81 lines not on this page · raw message
tools/perf/util/cache.h +2 -55
57 lines not on this page · raw message
tools/perf/util/config.c +6 -455
461 lines not on this page · raw message
tools/perf/util/header.c +52 -32
84 lines not on this page · raw message
tools/perf/util/hist.c +33 -9
42 lines not on this page · raw message
tools/perf/util/newt.c +116 -34
150 lines not on this page · raw message
tools/perf/util/path.c +1 -203
204 lines not on this page · raw message
tools/perf/util/quote.c +1 -432
433 lines not on this page · raw message
tools/perf/util/run-command.c +0 -90
90 lines not on this page · raw message
tools/perf/util/strbuf.c +1 -228
229 lines not on this page · raw message
tools/perf/util/strbuf.h +0 -45
45 lines not on this page · raw message
tools/perf/util/util.h +14 -163
177 lines not on this page · raw message
tools/perf/util/wrapper.c +0 -110
110 lines not on this page · raw message
The rest of this diff is not shown. View the raw message.
Part of the diff is not shown here. View the raw message.
Quoted 2 lines · Ingo Molnar
75 files changed, 1847 insertions(+), 3147 deletions(-)
I was excited by this ("A code _reduction_! Will wonders never cease?"), but it turns out that all of the reduction was due to removing some unused utility functions from the user-level tools.
The actual kernel still grew:
32 files changed, 1327 insertions(+), 964 deletions(-)
oh well. Merged anyway.
Linus
Em Thu, May 27, 2010 at 03:34:51PM -0700, Linus Torvalds escreveu:
Quoted 9 lines · Ingo Molnar
Quoted 2 lines
75 files changed, 1847 insertions(+), 3147 deletions(-)I was excited by this ("A code _reduction_! Will wonders never cease?"), but it turns out that all of the reduction was due to removing some unused utility functions from the user-level tools.
That was Ingo's forward looking investments, which I decided to take off the balance sheet not to get the IMF on our backs, glad you noticed the trick, never the less, it was at least a step on the right direction, I guess.
Quoted 5 lines
The actual kernel still grew:
32 files changed, 1327 insertions(+), 964 deletions(-)oh well. Merged anyway.
Thanks!
Quoted 1 line
Linus
- Arnaldo
Steven Rostedt
maintainerQuoted 3 lines · Linus Torvalds
The actual kernel still grew:
32 files changed, 1327 insertions(+), 964 deletions(-)
Well, if it's any consolation, that patch pull did include my 100K reduction in trace events.
-- Steve
Em Thu, May 27, 2010 at 08:32:37PM -0400, Steven Rostedt escreveu:
Quoted 8 lines · Linus Torvalds
Quoted 4 lines
The actual kernel still grew:
32 files changed, 1327 insertions(+), 964 deletions(-)Well, if it's any consolation, that patch pull did include my 100K reduction in trace events.
Bah, these investors, who would believe in them? :-P
- Arnaldo
Quoted 7 lines · Linus Torvalds
Quoted 2 lines · Ingo Molnar
75 files changed, 1847 insertions(+), 3147 deletions(-)I was excited by this ("A code _reduction_! Will wonders never cease?"), but it turns out that all of the reduction was due to removing some unused utility functions from the user-level tools.
Simple things first and maybe it turns into a habit!
Quoted 5 lines
The actual kernel still grew:
32 files changed, 1327 insertions(+), 964 deletions(-)oh well. Merged anyway.
Hey, while i have to admit to being a member of the infinitely large group of kernel developers working hard to ensure that the kernel's size follows Moore's Law - this time i beg to differ, as we _did_ manage to shrink the kernel! ;-)
Here's the before/after vmlinux comparison:
(x86 defconfig, 64-bit): vmlinux:
text data bss dec hex filename
.......................................................
cad719d (before): 8441843 1281100 983876 10706819 a35f83 vmlinux
c5617b2 ( after): 8417488 1277724 983876 10679088 a2f330 vmlinux
.......................................................
-24.35K -3.37K
-0.28% -0.26%Ob'Suggestion: we could help debloating efforts by officially declaring to take provably-debloating patches up to -rc4? (with a vmlinux comparison mandated in the changelog or such)
Debloating patches tend to reduce complexity, and hence are more regression-resistent. (They are also rare, so it's not like we would be risking a flood of patches.)
That would be a clever way to redirect the creative energies of kernel developers who are otherwise bored by -rc1's legendary stability and always try to sneak further features upstream post -rc1.
Such a policy would mean that while the merge window for bloating patches is a strict ~1.5 weeks, the merge window for debloating patches would be a generous ~1.5 months.
Maybe such a mild hint would help propagate the idea some more.
Dunno - just an idea,
Ingo