From 1f208ea67821703fd4de056ea6f0baa81f4ad4a5 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:44 +1000 Subject: perf report: Fix -z option Fix a copy and paste error, -z was setting the group option. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230140.714204656@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index cf0d21f..5c29655 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -675,7 +675,7 @@ static const struct option options[] = { "put the counters into a counter group"), OPT_STRING('s', "sym-filter", &sym_filter, "pattern", "only display symbols matchig this pattern"), - OPT_BOOLEAN('z', "zero", &group, + OPT_BOOLEAN('z', "zero", &zero, "zero history across updates"), OPT_INTEGER('F', "freq", &freq, "profile at this frequency"), -- cgit v0.10.2 From 6717534ddc328ae5cdf89f1ef802db83fc451f19 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:45 +1000 Subject: perf_counter tools: Remove zlib dependency The zlib devel libraries may not be installed and since we aren't using zlib we may as well remove it. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230140.802078956@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 9c6d0ae3..f572c90 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -381,12 +381,6 @@ ifndef CC_LD_DYNPATH endif endif -ifdef ZLIB_PATH - BASIC_CFLAGS += -I$(ZLIB_PATH)/include - EXTLIBS += -L$(ZLIB_PATH)/$(lib) $(CC_LD_DYNPATH)$(ZLIB_PATH)/$(lib) -endif -EXTLIBS += -lz - ifdef NEEDS_SOCKET EXTLIBS += -lsocket endif -- cgit v0.10.2 From 2ab52083ffc057014e502cf3473adc41436922fa Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:46 +1000 Subject: perf top: Move skip symbols to an array Move the list of symbols we skip into an array, making it easier to add new ones. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230140.904782938@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 5c29655..731ec6d 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -286,11 +286,22 @@ static void *display_thread(void *arg) return NULL; } +/* Tag samples to be skipped. */ +char *skip_symbols[] = { + "default_idle", + "cpu_idle", + "enter_idle", + "exit_idle", + "mwait_idle", + NULL +}; + static int symbol_filter(struct dso *self, struct symbol *sym) { static int filter_match; struct sym_entry *syme; const char *name = sym->name; + int i; if (!strcmp(name, "_text") || !strcmp(name, "_etext") || @@ -302,13 +313,12 @@ static int symbol_filter(struct dso *self, struct symbol *sym) return 1; syme = dso__sym_priv(self, sym); - /* Tag samples to be skipped. */ - if (!strcmp("default_idle", name) || - !strcmp("cpu_idle", name) || - !strcmp("enter_idle", name) || - !strcmp("exit_idle", name) || - !strcmp("mwait_idle", name)) - syme->skip = 1; + for (i = 0; skip_symbols[i]; i++) { + if (!strcmp(skip_symbols[i], name)) { + syme->skip = 1; + break; + } + } if (filter_match == 1) { filter_end = sym->start; -- cgit v0.10.2 From 3a3393ef75a14ae259a82f3f38624efa17884168 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:47 +1000 Subject: perf top: Add ppc64 specific skip symbols and strip ppc64 . prefix Filter out some ppc64 specific idle loop functions and remove leading '.' on ppc64 text symbols. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230140.995643441@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 731ec6d..0506cd6 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -293,6 +293,8 @@ char *skip_symbols[] = { "enter_idle", "exit_idle", "mwait_idle", + "ppc64_runlatch_off", + "pseries_dedicated_idle_sleep", NULL }; @@ -303,6 +305,13 @@ static int symbol_filter(struct dso *self, struct symbol *sym) const char *name = sym->name; int i; + /* + * ppc64 uses function descriptors and appends a '.' to the + * start of every instruction address. Remove it. + */ + if (name[0] == '.') + name++; + if (!strcmp(name, "_text") || !strcmp(name, "_etext") || !strcmp(name, "_sinittext") || -- cgit v0.10.2 From d8db1b57d31a6b30ea2f0df318eab50fc92b38d6 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:48 +1000 Subject: perf report: Fix reporting of hypervisor PERF_EVENT_MISC_* is not a bitmask, so we have to mask and compare. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230141.088394681@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 135b783..88e88c5 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -1213,6 +1213,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) struct map *map = NULL; void *more_data = event->ip.__more_data; struct ip_callchain *chain = NULL; + int cpumode; if (sample_type & PERF_SAMPLE_PERIOD) { period = *(u64 *)more_data; @@ -1256,7 +1257,9 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) if (comm_list && !strlist__has_entry(comm_list, thread->comm)) return 0; - if (event->header.misc & PERF_EVENT_MISC_KERNEL) { + cpumode = event->header.misc & PERF_EVENT_MISC_CPUMODE_MASK; + + if (cpumode == PERF_EVENT_MISC_KERNEL) { show = SHOW_KERNEL; level = 'k'; @@ -1264,7 +1267,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) dprintf(" ...... dso: %s\n", dso->name); - } else if (event->header.misc & PERF_EVENT_MISC_USER) { + } else if (cpumode == PERF_EVENT_MISC_USER) { show = SHOW_USER; level = '.'; -- cgit v0.10.2 From fb9c818873a788c5c01c9868cc6050df96e2c7df Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 09:00:49 +1000 Subject: perf report: Add hypervisor dso Add a dso for hypervisor samples. We don't get any symbol information on the ppc64 hypervisor but this at least gives us a high level summary of the time spent in there. Signed-off-by: Anton Blanchard Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <20090630230141.182536873@samba.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 88e88c5..3f5d8ea 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -121,6 +121,7 @@ typedef union event_union { static LIST_HEAD(dsos); static struct dso *kernel_dso; static struct dso *vdso; +static struct dso *hypervisor_dso; static void dsos__add(struct dso *dso) { @@ -202,6 +203,11 @@ static int load_kernel(void) dsos__add(vdso); + hypervisor_dso = dso__new("[hypervisor]", 0); + if (!hypervisor_dso) + return -1; + dsos__add(hypervisor_dso); + return err; } @@ -640,7 +646,8 @@ sort__sym_print(FILE *fp, struct hist_entry *self) if (self->sym) { ret += fprintf(fp, "[%c] %s", - self->dso == kernel_dso ? 'k' : '.', self->sym->name); + self->dso == kernel_dso ? 'k' : + self->dso == hypervisor_dso ? 'h' : '.', self->sym->name); } else { ret += fprintf(fp, "%#016llx", (u64)self->ip); } @@ -963,6 +970,9 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, } switch (context) { + case PERF_CONTEXT_HV: + dso = hypervisor_dso; + break; case PERF_CONTEXT_KERNEL: dso = kernel_dso; break; @@ -1275,6 +1285,9 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) } else { show = SHOW_HV; level = 'H'; + + dso = hypervisor_dso; + dprintf(" ...... dso: [hypervisor]\n"); } -- cgit v0.10.2 From 9198aa77b69647d1d91207f8075763abe7dc0bf4 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 1 Jul 2009 05:35:13 +0200 Subject: perf_counter tools: Fix storage size allocation of callchain list Fix a confusion while giving the size of a callchain list during its allocation. We are using the wrong structure size. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246419315-9968-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index ad3c285..bbf7813 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -74,7 +74,7 @@ fill_node(struct callchain_node *node, struct ip_callchain *chain, int start) for (i = start; i < chain->nr; i++) { struct callchain_list *call; - call = malloc(sizeof(*chain)); + call = malloc(sizeof(*call)); if (!call) { perror("not enough memory for the code path tree"); return; -- cgit v0.10.2 From 4424961ad6621a02c6b4c9093e801002c1bb9f65 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 1 Jul 2009 05:35:14 +0200 Subject: perf_counter tools: Resolve symbols in callchains This patch resolves the names, when possible, of each ip present in the callchains while using the -c option with perf report. Example: 5.40% [k] __d_lookup 5.37% perf_callchain perf_counter_overflow intel_pmu_handle_irq perf_counter_nmi_handler notifier_call_chain atomic_notifier_call_chain notify_die do_nmi nmi do_lookup __link_path_walk path_walk do_path_lookup user_path_at sys_faccessat sys_access system_call_fastpath 0x7fb609846f77 0.01% perf_callchain perf_counter_overflow intel_pmu_handle_irq perf_counter_nmi_handler notifier_call_chain atomic_notifier_call_chain notify_die do_nmi nmi do_lookup __link_path_walk path_walk do_path_lookup user_path_at sys_faccessat Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246419315-9968-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 3f5d8ea..1977930 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -794,8 +794,15 @@ callchain__fprintf(FILE *fp, struct callchain_node *self, u64 total_samples) ret += callchain__fprintf(fp, self->parent, total_samples); - list_for_each_entry(chain, &self->val, list) - ret += fprintf(fp, " %p\n", (void *)chain->ip); + list_for_each_entry(chain, &self->val, list) { + if (chain->ip >= PERF_CONTEXT_MAX) + continue; + if (chain->sym) + ret += fprintf(fp, " %s\n", chain->sym->name); + else + ret += fprintf(fp, " %p\n", + (void *)chain->ip); + } return ret; } @@ -930,6 +937,55 @@ static int call__match(struct symbol *sym) return 0; } +static struct symbol ** +resolve_callchain(struct thread *thread, struct map *map, + struct ip_callchain *chain, struct hist_entry *entry) +{ + int i; + struct symbol **syms; + u64 context = PERF_CONTEXT_MAX; + + if (callchain) { + syms = calloc(chain->nr, sizeof(*syms)); + if (!syms) { + fprintf(stderr, "Can't allocate memory for symbols\n"); + exit(-1); + } + } + + for (i = 0; i < chain->nr; i++) { + u64 ip = chain->ips[i]; + struct dso *dso = NULL; + struct symbol *sym; + + if (ip >= PERF_CONTEXT_MAX) { + context = ip; + continue; + } + + switch (context) { + case PERF_CONTEXT_KERNEL: + dso = kernel_dso; + break; + default: + break; + } + + sym = resolve_symbol(thread, NULL, &dso, &ip); + + if (sym) { + if (sort__has_parent && call__match(sym) && + !entry->parent) + entry->parent = sym; + if (!callchain) + break; + syms[i] = sym; + } + } + + return syms; +} + /* * collect histogram counts */ @@ -942,6 +998,7 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, struct rb_node **p = &hist.rb_node; struct rb_node *parent = NULL; struct hist_entry *he; + struct symbol **syms = NULL; struct hist_entry entry = { .thread = thread, .map = map, @@ -955,39 +1012,11 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, }; int cmp; - if (sort__has_parent && chain) { - u64 context = PERF_CONTEXT_MAX; - int i; - - for (i = 0; i < chain->nr; i++) { - u64 ip = chain->ips[i]; - struct dso *dso = NULL; - struct symbol *sym; - - if (ip >= PERF_CONTEXT_MAX) { - context = ip; - continue; - } - - switch (context) { case PERF_CONTEXT_HV: dso = hypervisor_dso; break; - case PERF_CONTEXT_KERNEL: - dso = kernel_dso; - break; - default: - break; - } - - sym = resolve_symbol(thread, NULL, &dso, &ip); - - if (sym && call__match(sym)) { - entry.parent = sym; - break; - } - } - } + if ((sort__has_parent || callchain) && chain) + syms = resolve_callchain(thread, map, chain, &entry); while (*p != NULL) { parent = *p; @@ -997,8 +1026,10 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, if (!cmp) { he->count += count; - if (callchain) - append_chain(&he->callchain, chain); + if (callchain) { + append_chain(&he->callchain, chain, syms); + free(syms); + } return 0; } @@ -1014,7 +1045,8 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, *he = entry; if (callchain) { callchain_init(&he->callchain); - append_chain(&he->callchain, chain); + append_chain(&he->callchain, chain, syms); + free(syms); } rb_link_node(&he->rb_node, parent, p); rb_insert_color(&he->rb_node, &hist); diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index bbf7813..6568cb1 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -67,7 +67,8 @@ static struct callchain_node *create_child(struct callchain_node *parent) } static void -fill_node(struct callchain_node *node, struct ip_callchain *chain, int start) +fill_node(struct callchain_node *node, struct ip_callchain *chain, int start, + struct symbol **syms) { int i; @@ -80,24 +81,26 @@ fill_node(struct callchain_node *node, struct ip_callchain *chain, int start) return; } call->ip = chain->ips[i]; + call->sym = syms[i]; list_add_tail(&call->list, &node->val); } node->val_nr = i - start; } -static void add_child(struct callchain_node *parent, struct ip_callchain *chain) +static void add_child(struct callchain_node *parent, struct ip_callchain *chain, + struct symbol **syms) { struct callchain_node *new; new = create_child(parent); - fill_node(new, chain, parent->val_nr); + fill_node(new, chain, parent->val_nr, syms); new->hit = 1; } static void split_add_child(struct callchain_node *parent, struct ip_callchain *chain, - struct callchain_list *to_split, int idx) + struct callchain_list *to_split, int idx, struct symbol **syms) { struct callchain_node *new; @@ -109,21 +112,22 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, parent->val_nr = idx; /* create the new one */ - add_child(parent, chain); + add_child(parent, chain, syms); } static int __append_chain(struct callchain_node *root, struct ip_callchain *chain, - int start); + int start, struct symbol **syms); static int -__append_chain_children(struct callchain_node *root, struct ip_callchain *chain) +__append_chain_children(struct callchain_node *root, struct ip_callchain *chain, + struct symbol **syms) { struct callchain_node *rnode; /* lookup in childrens */ list_for_each_entry(rnode, &root->children, brothers) { - int ret = __append_chain(rnode, chain, root->val_nr); + int ret = __append_chain(rnode, chain, root->val_nr, syms); if (!ret) return 0; } @@ -132,7 +136,7 @@ __append_chain_children(struct callchain_node *root, struct ip_callchain *chain) static int __append_chain(struct callchain_node *root, struct ip_callchain *chain, - int start) + int start, struct symbol **syms) { struct callchain_list *cnode; int i = start; @@ -154,7 +158,7 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, /* we match only a part of the node. Split it and add the new chain */ if (i < root->val_nr) { - split_add_child(root, chain, cnode, i); + split_add_child(root, chain, cnode, i, syms); return 0; } @@ -164,11 +168,12 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, return 0; } - return __append_chain_children(root, chain); + return __append_chain_children(root, chain, syms); } -void append_chain(struct callchain_node *root, struct ip_callchain *chain) +void append_chain(struct callchain_node *root, struct ip_callchain *chain, + struct symbol **syms) { - if (__append_chain_children(root, chain) == -1) - add_child(root, chain); + if (__append_chain_children(root, chain, syms) == -1) + add_child(root, chain, syms); } diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index fa1cd2f..c942daa 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -4,6 +4,7 @@ #include "../perf.h" #include "list.h" #include "rbtree.h" +#include "symbol.h" struct callchain_node { @@ -18,6 +19,7 @@ struct callchain_node { struct callchain_list { unsigned long ip; + struct symbol *sym; struct list_head list; }; @@ -28,6 +30,7 @@ static inline void callchain_init(struct callchain_node *node) INIT_LIST_HEAD(&node->val); } -void append_chain(struct callchain_node *root, struct ip_callchain *chain); +void append_chain(struct callchain_node *root, struct ip_callchain *chain, + struct symbol **syms); void sort_chain_to_rbtree(struct rb_root *rb_root, struct callchain_node *node); #endif -- cgit v0.10.2 From deac911cbdcb124fa0cee47c588e0cb0400b23b7 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 1 Jul 2009 05:35:15 +0200 Subject: perf_counter tools: Various fixes for callchains The symbol resolving has of course revealed some bugs in the callchain tree handling. This patch fixes some of them, including: - inherit the children from the parents while splitting a node - fix list range moving - fix indexes setting in callchains - create a child on the current node if the path doesn't match in the existent children (was only done on the root) - compare using symbols when possible so that we can match a function using any ip inside by referring to its start address. The practical effects are: - remove double callchains - fix upside down or any random order of callchains - fix wrong paths - fix bad hits and percentage accounts Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246419315-9968-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 6568cb1..440db12 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -4,6 +4,9 @@ * Handle the callchains from the stream in an ad-hoc radix tree and then * sort them in an rbtree. * + * Using a radix for code path provides a fast retrieval and factorizes + * memory use. Also that lets us use the paths in a hierarchical graph view. + * */ #include @@ -14,7 +17,8 @@ #include "callchain.h" -static void rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) +static void +rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) { struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; @@ -49,7 +53,12 @@ void sort_chain_to_rbtree(struct rb_root *rb_root, struct callchain_node *node) rb_insert_callchain(rb_root, node); } -static struct callchain_node *create_child(struct callchain_node *parent) +/* + * Create a child for a parent. If inherit_children, then the new child + * will become the new parent of it's parent children + */ +static struct callchain_node * +create_child(struct callchain_node *parent, bool inherit_children) { struct callchain_node *new; @@ -61,14 +70,27 @@ static struct callchain_node *create_child(struct callchain_node *parent) new->parent = parent; INIT_LIST_HEAD(&new->children); INIT_LIST_HEAD(&new->val); + + if (inherit_children) { + struct callchain_node *next; + + list_splice(&parent->children, &new->children); + INIT_LIST_HEAD(&parent->children); + + list_for_each_entry(next, &new->children, brothers) + next->parent = new; + } list_add_tail(&new->brothers, &parent->children); return new; } +/* + * Fill the node with callchain values + */ static void -fill_node(struct callchain_node *node, struct ip_callchain *chain, int start, - struct symbol **syms) +fill_node(struct callchain_node *node, struct ip_callchain *chain, + int start, struct symbol **syms) { int i; @@ -84,54 +106,80 @@ fill_node(struct callchain_node *node, struct ip_callchain *chain, int start, call->sym = syms[i]; list_add_tail(&call->list, &node->val); } - node->val_nr = i - start; + node->val_nr = chain->nr - start; + if (!node->val_nr) + printf("Warning: empty node in callchain tree\n"); } -static void add_child(struct callchain_node *parent, struct ip_callchain *chain, - struct symbol **syms) +static void +add_child(struct callchain_node *parent, struct ip_callchain *chain, + int start, struct symbol **syms) { struct callchain_node *new; - new = create_child(parent); - fill_node(new, chain, parent->val_nr, syms); + new = create_child(parent, false); + fill_node(new, chain, start, syms); new->hit = 1; } +/* + * Split the parent in two parts (a new child is created) and + * give a part of its callchain to the created child. + * Then create another child to host the given callchain of new branch + */ static void split_add_child(struct callchain_node *parent, struct ip_callchain *chain, - struct callchain_list *to_split, int idx, struct symbol **syms) + struct callchain_list *to_split, int idx_parents, int idx_local, + struct symbol **syms) { struct callchain_node *new; + struct list_head *old_tail; + int idx_total = idx_parents + idx_local; /* split */ - new = create_child(parent); - list_move_tail(&to_split->list, &new->val); - new->hit = parent->hit; - parent->hit = 0; - parent->val_nr = idx; + new = create_child(parent, true); + + /* split the callchain and move a part to the new child */ + old_tail = parent->val.prev; + list_del_range(&to_split->list, old_tail); + new->val.next = &to_split->list; + new->val.prev = old_tail; + to_split->list.prev = &new->val; + old_tail->next = &new->val; - /* create the new one */ - add_child(parent, chain, syms); + /* split the hits */ + new->hit = parent->hit; + new->val_nr = parent->val_nr - idx_local; + parent->val_nr = idx_local; + + /* create a new child for the new branch if any */ + if (idx_total < chain->nr) { + parent->hit = 0; + add_child(parent, chain, idx_total, syms); + } else { + parent->hit = 1; + } } static int __append_chain(struct callchain_node *root, struct ip_callchain *chain, int start, struct symbol **syms); -static int +static void __append_chain_children(struct callchain_node *root, struct ip_callchain *chain, - struct symbol **syms) + struct symbol **syms, int start) { struct callchain_node *rnode; /* lookup in childrens */ list_for_each_entry(rnode, &root->children, brothers) { - int ret = __append_chain(rnode, chain, root->val_nr, syms); + int ret = __append_chain(rnode, chain, start, syms); if (!ret) - return 0; + return; } - return -1; + /* nothing in children, add to the current node */ + add_child(root, chain, start, syms); } static int @@ -142,14 +190,22 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, int i = start; bool found = false; - /* lookup in the current node */ + /* + * Lookup in the current node + * If we have a symbol, then compare the start to match + * anywhere inside a function. + */ list_for_each_entry(cnode, &root->val, list) { - if (cnode->ip != chain->ips[i++]) + if (i == chain->nr) + break; + if (cnode->sym && syms[i]) { + if (cnode->sym->start != syms[i]->start) + break; + } else if (cnode->ip != chain->ips[i]) break; if (!found) found = true; - if (i == chain->nr) - break; + i++; } /* matches not, relay on the parent */ @@ -157,23 +213,25 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, return -1; /* we match only a part of the node. Split it and add the new chain */ - if (i < root->val_nr) { - split_add_child(root, chain, cnode, i, syms); + if (i - start < root->val_nr) { + split_add_child(root, chain, cnode, start, i - start, syms); return 0; } /* we match 100% of the path, increment the hit */ - if (i == root->val_nr) { + if (i - start == root->val_nr && i == chain->nr) { root->hit++; return 0; } - return __append_chain_children(root, chain, syms); + /* We match the node and still have a part remaining */ + __append_chain_children(root, chain, syms, i); + + return 0; } void append_chain(struct callchain_node *root, struct ip_callchain *chain, struct symbol **syms) { - if (__append_chain_children(root, chain, syms) == -1) - add_child(root, chain, syms); + __append_chain_children(root, chain, syms, 0); } -- cgit v0.10.2 From 0a456fc58fb8ef3c53d18297ab5cd5d2a70d146b Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 1 Jul 2009 13:07:01 +1000 Subject: powerpc/perf_counter: Enable alternate PR/HV bits for POWER7 POWER7 has the same PR/HV bit layout as POWER6, so set the flag. Signed-off-by: Anton Blanchard Acked-by: Paul Mackerras Cc: a.p.zijlstra@chello.nl Cc: benh@kernel.crashing.org LKML-Reference: <20090701030701.GI3563@kryten> Signed-off-by: Ingo Molnar diff --git a/arch/powerpc/kernel/power7-pmu.c b/arch/powerpc/kernel/power7-pmu.c index 5d755ef..5a9f5cb 100644 --- a/arch/powerpc/kernel/power7-pmu.c +++ b/arch/powerpc/kernel/power7-pmu.c @@ -358,6 +358,7 @@ static struct power_pmu power7_pmu = { .get_constraint = power7_get_constraint, .get_alternatives = power7_get_alternatives, .disable_pmc = power7_disable_pmc, + .flags = PPMU_ALT_SIPR, .n_generic = ARRAY_SIZE(power7_generic_events), .generic_events = power7_generic_events, .cache_events = &power7_cache_events, -- cgit v0.10.2 From 61c45981ddbd718136d49464f00d2f11938aaa6e Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 1 Jul 2009 13:04:34 +1000 Subject: perf_counter tools: Rework event string parsing/syntax This reworks the parser for event descriptors to make it more consistent in what it accepts. It is now structured as a recursive descent parser for the following grammar: events ::= event ( ("," | space) space* event )* event ::= ( raw_event | numeric_event | symbolic_event | generic_hw_event ) [ event_modifier ] raw_event ::= "r" hex_number numeric_event ::= number ":" number number ::= decimal_number | "0x" hex_number | "0" octal_number symbolic_event ::= string_from_event_symbols_array generic_hw_event::= cache_type ( "-" ( cache_op | cache_result ) )* event_modifier ::= ":" ( "u" | "k" | "h" )+ with the extra restriction that you can have at most one cache_op and at most one cache_result. We pass the current string pointer by reference (i.e. as a const char **) to the various parsing functions so that they can advance the pointer to indicate how much they consumed. They return 0 if they didn't recognize the thing at the pointer or 1 if they did (and advance the pointer past it). This also fixes parse_aliases to take the longest matching alias from the table, not the first one. Otherwise "l1-data" would match the "l1-d" alias and the "ata" would not be consumed. This allows event modifiers indicating what processor modes to count in to be applied to any event, not just numeric events, and adds a ":h" modifier to indicate counting in hypervisor mode. Specifying ":u" now sets both exclude_kernel and exclude_hv, and so on. Multiple modes can be specified, e.g. ":uk" will count in user or hypervisor mode (i.e. only exclude_kernel will be set). Signed-off-by: Paul Mackerras Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <19018.53826.843815.189847@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 4d042f1..e6b83a3 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -184,16 +184,20 @@ char *event_name(int counter) return "unknown"; } -static int parse_aliases(const char *str, char *names[][MAX_ALIASES], int size) +static int parse_aliases(const char **str, char *names[][MAX_ALIASES], int size) { int i, j; + int n, longest = -1; for (i = 0; i < size; i++) { - for (j = 0; j < MAX_ALIASES; j++) { - if (!names[i][j]) - break; - if (strcasestr(str, names[i][j])) - return i; + for (j = 0; j < MAX_ALIASES && names[i][j]; j++) { + n = strlen(names[i][j]); + if (n > longest && !strncasecmp(*str, names[i][j], n)) + longest = n; + } + if (longest > 0) { + *str += longest; + return i; } } @@ -201,30 +205,53 @@ static int parse_aliases(const char *str, char *names[][MAX_ALIASES], int size) } static int -parse_generic_hw_symbols(const char *str, struct perf_counter_attr *attr) +parse_generic_hw_event(const char **str, struct perf_counter_attr *attr) { - int cache_type = -1, cache_op = 0, cache_result = 0; + const char *s = *str; + int cache_type = -1, cache_op = -1, cache_result = -1; - cache_type = parse_aliases(str, hw_cache, PERF_COUNT_HW_CACHE_MAX); + cache_type = parse_aliases(&s, hw_cache, PERF_COUNT_HW_CACHE_MAX); /* * No fallback - if we cannot get a clear cache type * then bail out: */ if (cache_type == -1) - return -EINVAL; + return 0; + + while ((cache_op == -1 || cache_result == -1) && *s == '-') { + ++s; + + if (cache_op == -1) { + cache_op = parse_aliases(&s, hw_cache_op, + PERF_COUNT_HW_CACHE_OP_MAX); + if (cache_op >= 0) { + if (!is_cache_op_valid(cache_type, cache_op)) + return 0; + continue; + } + } + + if (cache_result == -1) { + cache_result = parse_aliases(&s, hw_cache_result, + PERF_COUNT_HW_CACHE_RESULT_MAX); + if (cache_result >= 0) + continue; + } + + /* + * Can't parse this as a cache op or result, so back up + * to the '-'. + */ + --s; + break; + } - cache_op = parse_aliases(str, hw_cache_op, PERF_COUNT_HW_CACHE_OP_MAX); /* * Fall back to reads: */ if (cache_op == -1) cache_op = PERF_COUNT_HW_CACHE_OP_READ; - if (!is_cache_op_valid(cache_type, cache_op)) - return -EINVAL; - - cache_result = parse_aliases(str, hw_cache_result, - PERF_COUNT_HW_CACHE_RESULT_MAX); /* * Fall back to accesses: */ @@ -234,93 +261,154 @@ parse_generic_hw_symbols(const char *str, struct perf_counter_attr *attr) attr->config = cache_type | (cache_op << 8) | (cache_result << 16); attr->type = PERF_TYPE_HW_CACHE; - return 0; + *str = s; + return 1; } static int check_events(const char *str, unsigned int i) { - if (!strncmp(str, event_symbols[i].symbol, - strlen(event_symbols[i].symbol))) - return 1; + int n; - if (strlen(event_symbols[i].alias)) - if (!strncmp(str, event_symbols[i].alias, - strlen(event_symbols[i].alias))) - return 1; + n = strlen(event_symbols[i].symbol); + if (!strncmp(str, event_symbols[i].symbol, n)) + return n; + + n = strlen(event_symbols[i].alias); + if (n) + if (!strncmp(str, event_symbols[i].alias, n)) + return n; return 0; } -/* - * Each event can have multiple symbolic names. - * Symbolic names are (almost) exactly matched. - */ -static int parse_event_symbols(const char *str, struct perf_counter_attr *attr) +static int +parse_symbolic_event(const char **strp, struct perf_counter_attr *attr) { - u64 config, id; - int type; + const char *str = *strp; unsigned int i; - const char *sep, *pstr; + int n; - if (str[0] == 'r' && hex2u64(str + 1, &config) > 0) { - attr->type = PERF_TYPE_RAW; - attr->config = config; + for (i = 0; i < ARRAY_SIZE(event_symbols); i++) { + n = check_events(str, i); + if (n > 0) { + attr->type = event_symbols[i].type; + attr->config = event_symbols[i].config; + *strp = str + n; + return 1; + } + } + return 0; +} + +static int parse_raw_event(const char **strp, struct perf_counter_attr *attr) +{ + const char *str = *strp; + u64 config; + int n; + if (*str != 'r') return 0; + n = hex2u64(str + 1, &config); + if (n > 0) { + *strp = str + n + 1; + attr->type = PERF_TYPE_RAW; + attr->config = config; + return 1; } + return 0; +} - pstr = str; - sep = strchr(pstr, ':'); - if (sep) { - type = atoi(pstr); - pstr = sep + 1; - id = atoi(pstr); - sep = strchr(pstr, ':'); - if (sep) { - pstr = sep + 1; - if (strchr(pstr, 'k')) - attr->exclude_user = 1; - if (strchr(pstr, 'u')) - attr->exclude_kernel = 1; +static int +parse_numeric_event(const char **strp, struct perf_counter_attr *attr) +{ + const char *str = *strp; + char *endp; + unsigned long type; + u64 config; + + type = strtoul(str, &endp, 0); + if (endp > str && type < PERF_TYPE_MAX && *endp == ':') { + str = endp + 1; + config = strtoul(str, &endp, 0); + if (endp > str) { + attr->type = type; + attr->config = config; + *strp = endp; + return 1; } - attr->type = type; - attr->config = id; + } + return 0; +} + +static int +parse_event_modifier(const char **strp, struct perf_counter_attr *attr) +{ + const char *str = *strp; + int eu = 1, ek = 1, eh = 1; + if (*str++ != ':') return 0; + while (*str) { + if (*str == 'u') + eu = 0; + else if (*str == 'k') + ek = 0; + else if (*str == 'h') + eh = 0; + else + break; + ++str; } + if (str >= *strp + 2) { + *strp = str; + attr->exclude_user = eu; + attr->exclude_kernel = ek; + attr->exclude_hv = eh; + return 1; + } + return 0; +} - for (i = 0; i < ARRAY_SIZE(event_symbols); i++) { - if (check_events(str, i)) { - attr->type = event_symbols[i].type; - attr->config = event_symbols[i].config; +/* + * Each event can have multiple symbolic names. + * Symbolic names are (almost) exactly matched. + */ +static int parse_event_symbols(const char **str, struct perf_counter_attr *attr) +{ + if (!(parse_raw_event(str, attr) || + parse_numeric_event(str, attr) || + parse_symbolic_event(str, attr) || + parse_generic_hw_event(str, attr))) + return 0; - return 0; - } - } + parse_event_modifier(str, attr); - return parse_generic_hw_symbols(str, attr); + return 1; } int parse_events(const struct option *opt, const char *str, int unset) { struct perf_counter_attr attr; - int ret; - memset(&attr, 0, sizeof(attr)); -again: - if (nr_counters == MAX_COUNTERS) - return -1; + for (;;) { + if (nr_counters == MAX_COUNTERS) + return -1; + + memset(&attr, 0, sizeof(attr)); + if (!parse_event_symbols(&str, &attr)) + return -1; - ret = parse_event_symbols(str, &attr); - if (ret < 0) - return ret; + if (!(*str == 0 || *str == ',' || isspace(*str))) + return -1; - attrs[nr_counters] = attr; - nr_counters++; + attrs[nr_counters] = attr; + nr_counters++; - str = strstr(str, ","); - if (str) { - str++; - goto again; + if (*str == 0) + break; + if (*str == ',') + ++str; + while (isspace(*str)) + ++str; } return 0; -- cgit v0.10.2 From 88a69dfbc6ab1e3b51bba8c9103055e21089ebb9 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 1 Jul 2009 11:17:20 +0200 Subject: perf report: Fix HV bit mismerge MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix: builtin-report.c: In function ‘hist_entry__add’: builtin-report.c:1015: error: case label not within a switch statement builtin-report.c:1017: error: break statement not within loop or switch Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 1977930..7d2b49a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -964,6 +964,9 @@ resolve_callchain(struct thread *thread, struct map *map, } switch (context) { + case PERF_CONTEXT_HV: + dso = hypervisor_dso; + break; case PERF_CONTEXT_KERNEL: dso = kernel_dso; break; @@ -1012,9 +1015,6 @@ hist_entry__add(struct thread *thread, struct map *map, struct dso *dso, }; int cmp; - case PERF_CONTEXT_HV: - dso = hypervisor_dso; - break; if ((sort__has_parent || callchain) && chain) syms = resolve_callchain(thread, map, chain, &entry); -- cgit v0.10.2 From f37a291c527c954df4da568de718ebb36b8261c0 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 1 Jul 2009 12:37:06 +0200 Subject: perf_counter tools: Add more warnings and fix/annotate them Enable -Wextra. This found a few real bugs plus a number of signed/unsigned type mismatches/uncleanlinesses. It also required a few annotations All things considered it was still worth it so lets try with this enabled for now. Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index f572c90..eddf076 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -164,7 +164,7 @@ endif # CFLAGS and LDFLAGS are for the users to override from the command line. -CFLAGS = $(M64) -ggdb3 -Wall -Wstrict-prototypes -Wmissing-declarations -Wmissing-prototypes -std=gnu99 -Wdeclaration-after-statement -Werror -O6 +CFLAGS = $(M64) -ggdb3 -Wall -Wextra -Wstrict-prototypes -Wmissing-declarations -Wmissing-prototypes -std=gnu99 -Wdeclaration-after-statement -Werror -O6 LDFLAGS = -lpthread -lrt -lelf -lm ALL_CFLAGS = $(CFLAGS) ALL_LDFLAGS = $(LDFLAGS) diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 722c0f5..6cba70d 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -160,7 +160,7 @@ static void dsos__fprintf(FILE *fp) static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip) { - return dso__find_symbol(kernel_dso, ip); + return dso__find_symbol(dso, ip); } static int load_kernel(void) @@ -203,7 +203,7 @@ static u64 map__map_ip(struct map *map, u64 ip) return ip - map->start + map->pgoff; } -static u64 vdso__map_ip(struct map *map, u64 ip) +static u64 vdso__map_ip(struct map *map __used, u64 ip) { return ip; } @@ -600,7 +600,7 @@ static LIST_HEAD(hist_entry__sort_list); static int sort_dimension__add(char *tok) { - int i; + unsigned int i; for (i = 0; i < ARRAY_SIZE(sort_dimensions); i++) { struct sort_dimension *sd = &sort_dimensions[i]; @@ -1069,7 +1069,7 @@ parse_line(FILE *file, struct symbol *sym, u64 start, u64 len) static const char *prev_color; unsigned int offset; size_t line_len; - u64 line_ip; + s64 line_ip; int ret; char *c; @@ -1428,7 +1428,7 @@ more: head += size; - if (offset + head < stat.st_size) + if (offset + head < (unsigned long)stat.st_size) goto more; rc = EXIT_SUCCESS; @@ -1492,7 +1492,7 @@ static void setup_sorting(void) free(str); } -int cmd_annotate(int argc, const char **argv, const char *prefix) +int cmd_annotate(int argc, const char **argv, const char *prefix __used) { symbol__init(); diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index 0f32dc3..2599d86 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -3,6 +3,7 @@ * * Builtin help command */ +#include "perf.h" #include "util/cache.h" #include "builtin.h" #include "util/exec_cmd.h" @@ -277,7 +278,7 @@ static struct cmdnames main_cmds, other_cmds; void list_common_cmds_help(void) { - int i, longest = 0; + unsigned int i, longest = 0; for (i = 0; i < ARRAY_SIZE(common_cmds); i++) { if (longest < strlen(common_cmds[i].name)) @@ -415,9 +416,10 @@ static void show_html_page(const char *perf_cmd) open_html(page_path.buf); } -int cmd_help(int argc, const char **argv, const char *prefix) +int cmd_help(int argc, const char **argv, const char *prefix __used) { const char *alias; + load_command_list("perf-", &main_cmds, &other_cmds); perf_config(perf_help_config, NULL); diff --git a/tools/perf/builtin-list.c b/tools/perf/builtin-list.c index fe60e37..f990fa8 100644 --- a/tools/perf/builtin-list.c +++ b/tools/perf/builtin-list.c @@ -13,7 +13,7 @@ #include "util/parse-options.h" #include "util/parse-events.h" -int cmd_list(int argc, const char **argv, const char *prefix) +int cmd_list(int argc __used, const char **argv __used, const char *prefix __used) { print_events(); return 0; diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index d18546f..4ef78a5 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -294,7 +294,7 @@ static void pid_synthesize_mmap_samples(pid_t pid) while (1) { char bf[BUFSIZ], *pbf = bf; struct mmap_event mmap_ev = { - .header.type = PERF_EVENT_MMAP, + .header = { .type = PERF_EVENT_MMAP }, }; int n; size_t size; @@ -650,7 +650,7 @@ static const struct option options[] = { OPT_END() }; -int cmd_record(int argc, const char **argv, const char *prefix) +int cmd_record(int argc, const char **argv, const char *prefix __used) { int counter; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 7d2b49a..007363d 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -177,7 +177,7 @@ static void dsos__fprintf(FILE *fp) static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip) { - return dso__find_symbol(kernel_dso, ip); + return dso__find_symbol(dso, ip); } static int load_kernel(void) @@ -239,7 +239,7 @@ static u64 map__map_ip(struct map *map, u64 ip) return ip - map->start + map->pgoff; } -static u64 vdso__map_ip(struct map *map, u64 ip) +static u64 vdso__map_ip(struct map *map __used, u64 ip) { return ip; } @@ -712,7 +712,7 @@ static LIST_HEAD(hist_entry__sort_list); static int sort_dimension__add(char *tok) { - int i; + unsigned int i; for (i = 0; i < ARRAY_SIZE(sort_dimensions); i++) { struct sort_dimension *sd = &sort_dimensions[i]; @@ -801,7 +801,7 @@ callchain__fprintf(FILE *fp, struct callchain_node *self, u64 total_samples) ret += fprintf(fp, " %s\n", chain->sym->name); else ret += fprintf(fp, " %p\n", - (void *)chain->ip); + (void *)(long)chain->ip); } return ret; @@ -938,12 +938,12 @@ static int call__match(struct symbol *sym) } static struct symbol ** -resolve_callchain(struct thread *thread, struct map *map, +resolve_callchain(struct thread *thread, struct map *map __used, struct ip_callchain *chain, struct hist_entry *entry) { - int i; - struct symbol **syms; u64 context = PERF_CONTEXT_MAX; + struct symbol **syms; + unsigned int i; if (callchain) { syms = calloc(chain->nr, sizeof(*syms)); @@ -1183,7 +1183,7 @@ static size_t output__fprintf(FILE *fp, u64 total_samples) fprintf(fp, "# ........"); list_for_each_entry(se, &hist_entry__sort_list, list) { - int i; + unsigned int i; if (exclude_other && (se == &sort_parent)) continue; @@ -1271,7 +1271,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) (long long)period); if (sample_type & PERF_SAMPLE_CALLCHAIN) { - int i; + unsigned int i; chain = (void *)more_data; @@ -1667,7 +1667,7 @@ more: if (offset + head >= header->data_offset + header->data_size) goto done; - if (offset + head < stat.st_size) + if (offset + head < (unsigned long)stat.st_size) goto more; done: @@ -1756,7 +1756,7 @@ static void setup_list(struct strlist **list, const char *list_str, } } -int cmd_report(int argc, const char **argv, const char *prefix) +int cmd_report(int argc, const char **argv, const char *prefix __used) { symbol__init(); diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 2e03524..095a90e 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -64,7 +64,7 @@ static struct perf_counter_attr default_attrs[] = { static int system_wide = 0; static int verbose = 0; -static int nr_cpus = 0; +static unsigned int nr_cpus = 0; static int run_idx = 0; static int run_count = 1; @@ -108,7 +108,8 @@ static void create_perf_stat_counter(int counter, int pid) PERF_FORMAT_TOTAL_TIME_RUNNING; if (system_wide) { - int cpu; + unsigned int cpu; + for (cpu = 0; cpu < nr_cpus; cpu++) { fd[cpu][counter] = sys_perf_counter_open(attr, -1, cpu, -1, 0); if (fd[cpu][counter] < 0 && verbose) @@ -150,8 +151,8 @@ static inline int nsec_counter(int counter) static void read_counter(int counter) { u64 *count, single_count[3]; - ssize_t res; - int cpu, nv; + unsigned int cpu; + size_t res, nv; int scaled; count = event_res[run_idx][counter]; @@ -165,6 +166,7 @@ static void read_counter(int counter) res = read(fd[cpu][counter], single_count, nv * sizeof(u64)); assert(res == nv * sizeof(u64)); + close(fd[cpu][counter]); fd[cpu][counter] = -1; @@ -200,7 +202,7 @@ static void read_counter(int counter) runtime_cycles[run_idx] = count[0]; } -static int run_perf_stat(int argc, const char **argv) +static int run_perf_stat(int argc __used, const char **argv) { unsigned long long t0, t1; int status = 0; @@ -390,7 +392,7 @@ static void calc_avg(void) event_res_avg[j]+1, event_res[i][j]+1); update_avg("counter/2", j, event_res_avg[j]+2, event_res[i][j]+2); - if (event_scaled[i][j] != -1) + if (event_scaled[i][j] != (u64)-1) update_avg("scaled", j, event_scaled_avg + j, event_scaled[i]+j); else @@ -510,7 +512,7 @@ static const struct option options[] = { OPT_END() }; -int cmd_stat(int argc, const char **argv, const char *prefix) +int cmd_stat(int argc, const char **argv, const char *prefix __used) { int status; @@ -528,7 +530,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix) nr_cpus = sysconf(_SC_NPROCESSORS_ONLN); assert(nr_cpus <= MAX_NR_CPUS); - assert(nr_cpus >= 0); + assert((int)nr_cpus >= 0); /* * We dont want to block the signals - that would cause diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 0506cd6..5f5e7df 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -269,7 +269,7 @@ static void print_sym_table(void) } } -static void *display_thread(void *arg) +static void *display_thread(void *arg __used) { struct pollfd stdin_poll = { .fd = 0, .events = POLLIN }; int delay_msecs = delay_secs * 1000; @@ -287,7 +287,7 @@ static void *display_thread(void *arg) } /* Tag samples to be skipped. */ -char *skip_symbols[] = { +static const char *skip_symbols[] = { "default_idle", "cpu_idle", "enter_idle", @@ -426,7 +426,7 @@ static void process_event(u64 ip, int counter, int user) struct mmap_data { int counter; void *base; - unsigned int mask; + int mask; unsigned int prev; }; @@ -705,7 +705,7 @@ static const struct option options[] = { OPT_END() }; -int cmd_top(int argc, const char **argv, const char *prefix) +int cmd_top(int argc, const char **argv, const char *prefix __used) { int counter; diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 4eb7259..c565678 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -229,9 +229,6 @@ static int run_builtin(struct cmd_struct *p, int argc, const char **argv) use_pager = 1; commit_pager_choice(); - if (p->option & NEED_WORK_TREE) - /* setup_work_tree() */; - status = p->fn(argc, argv, prefix); if (status) return status & 0xff; @@ -266,7 +263,7 @@ static void handle_internal_command(int argc, const char **argv) { "annotate", cmd_annotate, 0 }, { "version", cmd_version, 0 }, }; - int i; + unsigned int i; static const char ext[] = STRIP_EXTENSION; if (sizeof(ext) > 1) { diff --git a/tools/perf/perf.h b/tools/perf/perf.h index ce39419..27887c9 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -52,6 +52,8 @@ static inline unsigned long long rdclock(void) #define __user #define asmlinkage +#define __used __attribute__((__unused__)) + #define unlikely(x) __builtin_expect(!!(x), 0) #define min(x, y) ({ \ typeof(x) _min1 = (x); \ diff --git a/tools/perf/util/alias.c b/tools/perf/util/alias.c index 9b3dd2b..b8144e8 100644 --- a/tools/perf/util/alias.c +++ b/tools/perf/util/alias.c @@ -3,7 +3,7 @@ static const char *alias_key; static char *alias_val; -static int alias_lookup_cb(const char *k, const char *v, void *cb) +static int alias_lookup_cb(const char *k, const char *v, void *cb __used) { if (!prefixcmp(k, "alias.") && !strcmp(k+6, alias_key)) { if (!v) diff --git a/tools/perf/util/cache.h b/tools/perf/util/cache.h index 393d614..161d5f4 100644 --- a/tools/perf/util/cache.h +++ b/tools/perf/util/cache.h @@ -3,6 +3,7 @@ #include "util.h" #include "strbuf.h" +#include "../perf.h" #define PERF_DIR_ENVIRONMENT "PERF_DIR" #define PERF_WORK_TREE_ENVIRONMENT "PERF_WORK_TREE" diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 440db12..3dceabd 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -92,7 +92,7 @@ static void fill_node(struct callchain_node *node, struct ip_callchain *chain, int start, struct symbol **syms) { - int i; + unsigned int i; for (i = start; i < chain->nr; i++) { struct callchain_list *call; @@ -135,7 +135,7 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, { struct callchain_node *new; struct list_head *old_tail; - int idx_total = idx_parents + idx_local; + unsigned int idx_total = idx_parents + idx_local; /* split */ new = create_child(parent, true); @@ -164,17 +164,18 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, static int __append_chain(struct callchain_node *root, struct ip_callchain *chain, - int start, struct symbol **syms); + unsigned int start, struct symbol **syms); static void __append_chain_children(struct callchain_node *root, struct ip_callchain *chain, - struct symbol **syms, int start) + struct symbol **syms, unsigned int start) { struct callchain_node *rnode; /* lookup in childrens */ list_for_each_entry(rnode, &root->children, brothers) { - int ret = __append_chain(rnode, chain, start, syms); + unsigned int ret = __append_chain(rnode, chain, start, syms); + if (!ret) return; } @@ -184,10 +185,10 @@ __append_chain_children(struct callchain_node *root, struct ip_callchain *chain, static int __append_chain(struct callchain_node *root, struct ip_callchain *chain, - int start, struct symbol **syms) + unsigned int start, struct symbol **syms) { struct callchain_list *cnode; - int i = start; + unsigned int i = start; bool found = false; /* diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index c942daa..251d99e 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -10,15 +10,15 @@ struct callchain_node { struct callchain_node *parent; struct list_head brothers; - struct list_head children; - struct list_head val; + struct list_head children; + struct list_head val; struct rb_node rb_node; - int val_nr; - int hit; + unsigned int val_nr; + u64 hit; }; struct callchain_list { - unsigned long ip; + u64 ip; struct symbol *sym; struct list_head list; }; diff --git a/tools/perf/util/color.c b/tools/perf/util/color.c index 9a8c20c..26f8231 100644 --- a/tools/perf/util/color.c +++ b/tools/perf/util/color.c @@ -11,7 +11,8 @@ static int parse_color(const char *name, int len) }; char *end; int i; - for (i = 0; i < ARRAY_SIZE(color_names); i++) { + + for (i = 0; i < (int)ARRAY_SIZE(color_names); i++) { const char *str = color_names[i]; if (!strncasecmp(name, str, len) && !str[len]) return i - 1; @@ -28,7 +29,8 @@ static int parse_attr(const char *name, int len) static const char * const attr_names[] = { "bold", "dim", "ul", "blink", "reverse" }; - int i; + unsigned int i; + for (i = 0; i < ARRAY_SIZE(attr_names); i++) { const char *str = attr_names[i]; if (!strncasecmp(name, str, len) && !str[len]) @@ -222,10 +224,12 @@ int color_fwrite_lines(FILE *fp, const char *color, { if (!*color) return fwrite(buf, count, 1, fp) != 1; + while (count) { char *p = memchr(buf, '\n', count); + if (p != buf && (fputs(color, fp) < 0 || - fwrite(buf, p ? p - buf : count, 1, fp) != 1 || + fwrite(buf, p ? (size_t)(p - buf) : count, 1, fp) != 1 || fputs(PERF_COLOR_RESET, fp) < 0)) return -1; if (!p) diff --git a/tools/perf/util/config.c b/tools/perf/util/config.c index 3dd13fa..780df54 100644 --- a/tools/perf/util/config.c +++ b/tools/perf/util/config.c @@ -47,10 +47,12 @@ static int get_next_char(void) static char *parse_value(void) { static char value[1024]; - int quote = 0, comment = 0, len = 0, space = 0; + int quote = 0, comment = 0, space = 0; + size_t len = 0; for (;;) { int c = get_next_char(); + if (len >= sizeof(value) - 1) return NULL; if (c == '\n') { @@ -353,13 +355,13 @@ int perf_config_string(const char **dest, const char *var, const char *value) return 0; } -static int perf_default_core_config(const char *var, const char *value) +static int perf_default_core_config(const char *var __used, const char *value __used) { /* Add other config variables here and to Documentation/config.txt. */ return 0; } -int perf_default_config(const char *var, const char *value, void *dummy) +int perf_default_config(const char *var, const char *value, void *dummy __used) { if (!prefixcmp(var, "core.")) return perf_default_core_config(var, value); @@ -471,10 +473,10 @@ static int matches(const char* key, const char* value) !regexec(store.value_regex, value, 0, NULL, 0))); } -static int store_aux(const char* key, const char* value, void *cb) +static int store_aux(const char* key, const char* value, void *cb __used) { + int section_len; const char *ep; - size_t section_len; switch (store.state) { case KEY_SEEN: @@ -551,7 +553,7 @@ static int store_write_section(int fd, const char* key) strbuf_addf(&sb, "[%.*s]\n", store.baselen, key); } - success = write_in_full(fd, sb.buf, sb.len) == sb.len; + success = (write_in_full(fd, sb.buf, sb.len) == (ssize_t)sb.len); strbuf_release(&sb); return success; @@ -599,7 +601,7 @@ static int store_write_pair(int fd, const char* key, const char* value) } strbuf_addf(&sb, "%s\n", quote); - success = write_in_full(fd, sb.buf, sb.len) == sb.len; + success = (write_in_full(fd, sb.buf, sb.len) == (ssize_t)sb.len); strbuf_release(&sb); return success; @@ -741,7 +743,7 @@ int perf_config_set_multivar(const char* key, const char* value, } else { struct stat st; char* contents; - size_t contents_sz, copy_begin, copy_end; + ssize_t contents_sz, copy_begin, copy_end; int i, new_line = 0; if (value_regex == NULL) diff --git a/tools/perf/util/exec_cmd.c b/tools/perf/util/exec_cmd.c index d392922..34a3528 100644 --- a/tools/perf/util/exec_cmd.c +++ b/tools/perf/util/exec_cmd.c @@ -1,6 +1,9 @@ #include "cache.h" #include "exec_cmd.h" #include "quote.h" + +#include + #define MAX_ARGS 32 extern char **environ; @@ -51,7 +54,7 @@ const char *perf_extract_argv0_path(const char *argv0) slash--; if (slash >= argv0) { - argv0_path = strndup(argv0, slash - argv0); + argv0_path = xstrndup(argv0, slash - argv0); return slash + 1; } diff --git a/tools/perf/util/help.c b/tools/perf/util/help.c index 17a00e0..fbb0097 100644 --- a/tools/perf/util/help.c +++ b/tools/perf/util/help.c @@ -26,7 +26,7 @@ static int term_columns(void) return 80; } -void add_cmdname(struct cmdnames *cmds, const char *name, int len) +void add_cmdname(struct cmdnames *cmds, const char *name, size_t len) { struct cmdname *ent = malloc(sizeof(*ent) + len + 1); @@ -40,7 +40,8 @@ void add_cmdname(struct cmdnames *cmds, const char *name, int len) static void clean_cmdnames(struct cmdnames *cmds) { - int i; + unsigned int i; + for (i = 0; i < cmds->cnt; ++i) free(cmds->names[i]); free(cmds->names); @@ -57,7 +58,7 @@ static int cmdname_compare(const void *a_, const void *b_) static void uniq(struct cmdnames *cmds) { - int i, j; + unsigned int i, j; if (!cmds->cnt) return; @@ -71,7 +72,7 @@ static void uniq(struct cmdnames *cmds) void exclude_cmds(struct cmdnames *cmds, struct cmdnames *excludes) { - int ci, cj, ei; + size_t ci, cj, ei; int cmp; ci = cj = ei = 0; @@ -106,8 +107,9 @@ static void pretty_print_string_list(struct cmdnames *cmds, int longest) printf(" "); for (j = 0; j < cols; j++) { - int n = j * rows + i; - int size = space; + unsigned int n = j * rows + i; + unsigned int size = space; + if (n >= cmds->cnt) break; if (j == cols-1 || n + rows >= cmds->cnt) @@ -208,7 +210,7 @@ void load_command_list(const char *prefix, void list_commands(const char *title, struct cmdnames *main_cmds, struct cmdnames *other_cmds) { - int i, longest = 0; + unsigned int i, longest = 0; for (i = 0; i < main_cmds->cnt; i++) if (longest < main_cmds->names[i]->len) @@ -239,7 +241,8 @@ void list_commands(const char *title, struct cmdnames *main_cmds, int is_in_cmdlist(struct cmdnames *c, const char *s) { - int i; + unsigned int i; + for (i = 0; i < c->cnt; i++) if (!strcmp(s, c->names[i]->name)) return 1; @@ -271,7 +274,8 @@ static int levenshtein_compare(const void *p1, const void *p2) static void add_cmd_list(struct cmdnames *cmds, struct cmdnames *old) { - int i; + unsigned int i; + ALLOC_GROW(cmds->names, cmds->cnt + old->cnt, cmds->alloc); for (i = 0; i < old->cnt; i++) @@ -283,7 +287,7 @@ static void add_cmd_list(struct cmdnames *cmds, struct cmdnames *old) const char *help_unknown_cmd(const char *cmd) { - int i, n = 0, best_similarity = 0; + unsigned int i, n = 0, best_similarity = 0; struct cmdnames main_cmds, other_cmds; memset(&main_cmds, 0, sizeof(main_cmds)); @@ -345,7 +349,7 @@ const char *help_unknown_cmd(const char *cmd) exit(1); } -int cmd_version(int argc, const char **argv, const char *prefix) +int cmd_version(int argc __used, const char **argv __used, const char *prefix __used) { printf("perf version %s\n", perf_version_string); return 0; diff --git a/tools/perf/util/help.h b/tools/perf/util/help.h index 56bc154..7128783 100644 --- a/tools/perf/util/help.h +++ b/tools/perf/util/help.h @@ -2,8 +2,8 @@ #define HELP_H struct cmdnames { - int alloc; - int cnt; + size_t alloc; + size_t cnt; struct cmdname { size_t len; /* also used for similarity index in help.c */ char name[FLEX_ARRAY]; @@ -19,7 +19,7 @@ static inline void mput_char(char c, unsigned int num) void load_command_list(const char *prefix, struct cmdnames *main_cmds, struct cmdnames *other_cmds); -void add_cmdname(struct cmdnames *cmds, const char *name, int len); +void add_cmdname(struct cmdnames *cmds, const char *name, size_t len); /* Here we require that excludes is a sorted list. */ void exclude_cmds(struct cmdnames *cmds, struct cmdnames *excludes); int is_in_cmdlist(struct cmdnames *c, const char *s); diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index e6b83a3..aed7090 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -385,7 +385,7 @@ static int parse_event_symbols(const char **str, struct perf_counter_attr *attr) return 1; } -int parse_events(const struct option *opt, const char *str, int unset) +int parse_events(const struct option *opt __used, const char *str, int unset __used) { struct perf_counter_attr attr; diff --git a/tools/perf/util/parse-options.c b/tools/perf/util/parse-options.c index b3affb1..9a897b7 100644 --- a/tools/perf/util/parse-options.c +++ b/tools/perf/util/parse-options.c @@ -485,7 +485,7 @@ int parse_options_usage(const char * const *usagestr, } -int parse_opt_verbosity_cb(const struct option *opt, const char *arg, +int parse_opt_verbosity_cb(const struct option *opt, const char *arg __used, int unset) { int *target = opt->value; diff --git a/tools/perf/util/parse-options.h b/tools/perf/util/parse-options.h index a1039a6..15c8aba 100644 --- a/tools/perf/util/parse-options.h +++ b/tools/perf/util/parse-options.h @@ -90,21 +90,20 @@ struct option { intptr_t defval; }; -#define OPT_END() { OPTION_END } -#define OPT_ARGUMENT(l, h) { OPTION_ARGUMENT, 0, (l), NULL, NULL, (h) } -#define OPT_GROUP(h) { OPTION_GROUP, 0, NULL, NULL, NULL, (h) } -#define OPT_BIT(s, l, v, h, b) { OPTION_BIT, (s), (l), (v), NULL, (h), 0, NULL, (b) } -#define OPT_BOOLEAN(s, l, v, h) { OPTION_BOOLEAN, (s), (l), (v), NULL, (h) } -#define OPT_SET_INT(s, l, v, h, i) { OPTION_SET_INT, (s), (l), (v), NULL, (h), 0, NULL, (i) } -#define OPT_SET_PTR(s, l, v, h, p) { OPTION_SET_PTR, (s), (l), (v), NULL, (h), 0, NULL, (p) } -#define OPT_INTEGER(s, l, v, h) { OPTION_INTEGER, (s), (l), (v), NULL, (h) } -#define OPT_LONG(s, l, v, h) { OPTION_LONG, (s), (l), (v), NULL, (h) } -#define OPT_STRING(s, l, v, a, h) { OPTION_STRING, (s), (l), (v), (a), (h) } +#define OPT_END() { .type = OPTION_END } +#define OPT_ARGUMENT(l, h) { .type = OPTION_ARGUMENT, .long_name = (l), .help = (h) } +#define OPT_GROUP(h) { .type = OPTION_GROUP, .help = (h) } +#define OPT_BIT(s, l, v, h, b) { .type = OPTION_BIT, .short_name = (s), .long_name = (l), .value = (v), .help = (h), .defval = (b) } +#define OPT_BOOLEAN(s, l, v, h) { .type = OPTION_BOOLEAN, .short_name = (s), .long_name = (l), .value = (v), .help = (h) } +#define OPT_SET_INT(s, l, v, h, i) { .type = OPTION_SET_INT, .short_name = (s), .long_name = (l), .value = (v), .help = (h), .defval = (i) } +#define OPT_SET_PTR(s, l, v, h, p) { .type = OPTION_SET_PTR, .short_name = (s), .long_name = (l), .value = (v), .help = (h), .defval = (p) } +#define OPT_INTEGER(s, l, v, h) { .type = OPTION_INTEGER, .short_name = (s), .long_name = (l), .value = (v), .help = (h) } +#define OPT_LONG(s, l, v, h) { .type = OPTION_LONG, .short_name = (s), .long_name = (l), .value = (v), .help = (h) } +#define OPT_STRING(s, l, v, a, h) { .type = OPTION_STRING, .short_name = (s), .long_name = (l), .value = (v), (a), .help = (h) } #define OPT_DATE(s, l, v, h) \ - { OPTION_CALLBACK, (s), (l), (v), "time",(h), 0, \ - parse_opt_approxidate_cb } + { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), .value = (v), .argh = "time", .help = (h), .callback = parse_opt_approxidate_cb } #define OPT_CALLBACK(s, l, v, a, h, f) \ - { OPTION_CALLBACK, (s), (l), (v), (a), (h), 0, (f) } + { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), .value = (v), (a), .help = (h), .callback = (f) } /* parse_options() will filter out the processed options and leave the * non-option argments in argv[]. diff --git a/tools/perf/util/quote.c b/tools/perf/util/quote.c index f18c521..c6e5dc0 100644 --- a/tools/perf/util/quote.c +++ b/tools/perf/util/quote.c @@ -162,12 +162,16 @@ static inline int sq_must_quote(char c) return sq_lookup[(unsigned char)c] + quote_path_fully > 0; } -/* returns the longest prefix not needing a quote up to maxlen if positive. - This stops at the first \0 because it's marked as a character needing an - escape */ -static size_t next_quote_pos(const char *s, ssize_t maxlen) +/* + * Returns the longest prefix not needing a quote up to maxlen if + * positive. + * This stops at the first \0 because it's marked as a character + * needing an escape. + */ +static ssize_t next_quote_pos(const char *s, ssize_t maxlen) { - size_t len; + ssize_t len; + if (maxlen < 0) { for (len = 0; !sq_must_quote(s[len]); len++); } else { @@ -192,22 +196,22 @@ static size_t next_quote_pos(const char *s, ssize_t maxlen) static size_t quote_c_style_counted(const char *name, ssize_t maxlen, struct strbuf *sb, FILE *fp, int no_dq) { -#undef EMIT -#define EMIT(c) \ - do { \ - if (sb) strbuf_addch(sb, (c)); \ - if (fp) fputc((c), fp); \ - count++; \ +#define EMIT(c) \ + do { \ + if (sb) strbuf_addch(sb, (c)); \ + if (fp) fputc((c), fp); \ + count++; \ } while (0) -#define EMITBUF(s, l) \ - do { \ - int __ret; \ - if (sb) strbuf_add(sb, (s), (l)); \ - if (fp) __ret = fwrite((s), (l), 1, fp); \ - count += (l); \ + +#define EMITBUF(s, l) \ + do { \ + int __ret; \ + if (sb) strbuf_add(sb, (s), (l)); \ + if (fp) __ret = fwrite((s), (l), 1, fp); \ + count += (l); \ } while (0) - size_t len, count = 0; + ssize_t len, count = 0; const char *p = name; for (;;) { @@ -273,8 +277,8 @@ void write_name_quoted(const char *name, FILE *fp, int terminator) fputc(terminator, fp); } -extern void write_name_quotedpfx(const char *pfx, size_t pfxlen, - const char *name, FILE *fp, int terminator) +void write_name_quotedpfx(const char *pfx, ssize_t pfxlen, + const char *name, FILE *fp, int terminator) { int needquote = 0; @@ -306,7 +310,7 @@ char *quote_path_relative(const char *in, int len, len = strlen(in); /* "../" prefix itself does not need quoting, but "in" might. */ - needquote = next_quote_pos(in, len) < len; + needquote = (next_quote_pos(in, len) < len); strbuf_setlen(out, 0); strbuf_grow(out, len); diff --git a/tools/perf/util/quote.h b/tools/perf/util/quote.h index 5dfad89..a5454a1 100644 --- a/tools/perf/util/quote.h +++ b/tools/perf/util/quote.h @@ -53,7 +53,7 @@ extern size_t quote_c_style(const char *name, struct strbuf *, FILE *, int no_dq extern void quote_two_c_style(struct strbuf *, const char *, const char *, int); extern void write_name_quoted(const char *name, FILE *, int terminator); -extern void write_name_quotedpfx(const char *pfx, size_t pfxlen, +extern void write_name_quotedpfx(const char *pfx, ssize_t pfxlen, const char *name, FILE *, int terminator); /* quote path as relative to the given prefix */ diff --git a/tools/perf/util/strbuf.c b/tools/perf/util/strbuf.c index 464e7ca..5249d5a 100644 --- a/tools/perf/util/strbuf.c +++ b/tools/perf/util/strbuf.c @@ -16,7 +16,7 @@ int prefixcmp(const char *str, const char *prefix) */ char strbuf_slopbuf[1]; -void strbuf_init(struct strbuf *sb, size_t hint) +void strbuf_init(struct strbuf *sb, ssize_t hint) { sb->alloc = sb->len = 0; sb->buf = strbuf_slopbuf; @@ -92,7 +92,8 @@ void strbuf_ltrim(struct strbuf *sb) void strbuf_tolower(struct strbuf *sb) { - int i; + unsigned int i; + for (i = 0; i < sb->len; i++) sb->buf[i] = tolower(sb->buf[i]); } @@ -264,7 +265,7 @@ size_t strbuf_fread(struct strbuf *sb, size_t size, FILE *f) return res; } -ssize_t strbuf_read(struct strbuf *sb, int fd, size_t hint) +ssize_t strbuf_read(struct strbuf *sb, int fd, ssize_t hint) { size_t oldlen = sb->len; size_t oldalloc = sb->alloc; @@ -293,7 +294,7 @@ ssize_t strbuf_read(struct strbuf *sb, int fd, size_t hint) #define STRBUF_MAXLINK (2*PATH_MAX) -int strbuf_readlink(struct strbuf *sb, const char *path, size_t hint) +int strbuf_readlink(struct strbuf *sb, const char *path, ssize_t hint) { size_t oldalloc = sb->alloc; @@ -301,7 +302,7 @@ int strbuf_readlink(struct strbuf *sb, const char *path, size_t hint) hint = 32; while (hint < STRBUF_MAXLINK) { - int len; + ssize_t len; strbuf_grow(sb, hint); len = readlink(path, sb->buf, hint); @@ -343,7 +344,7 @@ int strbuf_getline(struct strbuf *sb, FILE *fp, int term) return 0; } -int strbuf_read_file(struct strbuf *sb, const char *path, size_t hint) +int strbuf_read_file(struct strbuf *sb, const char *path, ssize_t hint) { int fd, len; diff --git a/tools/perf/util/strbuf.h b/tools/perf/util/strbuf.h index 9ee908a..d2aa86c 100644 --- a/tools/perf/util/strbuf.h +++ b/tools/perf/util/strbuf.h @@ -50,7 +50,7 @@ struct strbuf { #define STRBUF_INIT { 0, 0, strbuf_slopbuf } /*----- strbuf life cycle -----*/ -extern void strbuf_init(struct strbuf *, size_t); +extern void strbuf_init(struct strbuf *buf, ssize_t hint); extern void strbuf_release(struct strbuf *); extern char *strbuf_detach(struct strbuf *, size_t *); extern void strbuf_attach(struct strbuf *, void *, size_t, size_t); @@ -61,7 +61,7 @@ static inline void strbuf_swap(struct strbuf *a, struct strbuf *b) { } /*----- strbuf size related -----*/ -static inline size_t strbuf_avail(const struct strbuf *sb) { +static inline ssize_t strbuf_avail(const struct strbuf *sb) { return sb->alloc ? sb->alloc - sb->len - 1 : 0; } @@ -122,9 +122,9 @@ extern void strbuf_addf(struct strbuf *sb, const char *fmt, ...); extern size_t strbuf_fread(struct strbuf *, size_t, FILE *); /* XXX: if read fails, any partial read is undone */ -extern ssize_t strbuf_read(struct strbuf *, int fd, size_t hint); -extern int strbuf_read_file(struct strbuf *sb, const char *path, size_t hint); -extern int strbuf_readlink(struct strbuf *sb, const char *path, size_t hint); +extern ssize_t strbuf_read(struct strbuf *, int fd, ssize_t hint); +extern int strbuf_read_file(struct strbuf *sb, const char *path, ssize_t hint); +extern int strbuf_readlink(struct strbuf *sb, const char *path, ssize_t hint); extern int strbuf_getline(struct strbuf *, FILE *, int); diff --git a/tools/perf/util/wrapper.c b/tools/perf/util/wrapper.c index 6350d65..4574ac2 100644 --- a/tools/perf/util/wrapper.c +++ b/tools/perf/util/wrapper.c @@ -7,7 +7,7 @@ * There's no pack memory to release - but stay close to the Git * version so wrap this away: */ -static inline void release_pack_memory(size_t size, int flag) +static inline void release_pack_memory(size_t size __used, int flag __used) { } @@ -59,7 +59,8 @@ void *xmemdupz(const void *data, size_t len) char *xstrndup(const char *str, size_t len) { char *p = memchr(str, '\0', len); - return xmemdupz(str, p ? p - str : len); + + return xmemdupz(str, p ? (size_t)(p - str) : len); } void *xrealloc(void *ptr, size_t size) -- cgit v0.10.2 From b9ebdcc0ce1c676ebf5dc4f6df6b440d8fcf88ab Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 1 Jul 2009 15:05:09 +0530 Subject: perf stat: Define MATCH_EVENT for easy attr checking MATCH_EVENT is useful: 1. for multiple attrs checking 2. avoid repetition of PERF_TYPE_ and PERF_COUNT_ and save space 3. avoids line breakage Signed-off-by: Jaswinder Singh Rajput Cc: Peter Zijlstra LKML-Reference: <1246440909.3403.5.camel@hpdv5.satnam> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 095a90e..01cc07e 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -96,6 +96,10 @@ static u64 walltime_nsecs_noise; static u64 runtime_cycles_avg; static u64 runtime_cycles_noise; +#define MATCH_EVENT(t, c, counter) \ + (attrs[counter].type == PERF_TYPE_##t && \ + attrs[counter].config == PERF_COUNT_##c) + #define ERR_PERF_OPEN \ "Error: counter %d, sys_perf_counter_open() syscall returned with %d (%s)\n" @@ -133,13 +137,8 @@ static void create_perf_stat_counter(int counter, int pid) */ static inline int nsec_counter(int counter) { - if (attrs[counter].type != PERF_TYPE_SOFTWARE) - return 0; - - if (attrs[counter].config == PERF_COUNT_SW_CPU_CLOCK) - return 1; - - if (attrs[counter].config == PERF_COUNT_SW_TASK_CLOCK) + if (MATCH_EVENT(SOFTWARE, SW_CPU_CLOCK, counter) || + MATCH_EVENT(SOFTWARE, SW_TASK_CLOCK, counter)) return 1; return 0; @@ -194,11 +193,9 @@ static void read_counter(int counter) /* * Save the full runtime - to allow normalization during printout: */ - if (attrs[counter].type == PERF_TYPE_SOFTWARE && - attrs[counter].config == PERF_COUNT_SW_TASK_CLOCK) + if (MATCH_EVENT(SOFTWARE, SW_TASK_CLOCK, counter)) runtime_nsecs[run_idx] = count[0]; - if (attrs[counter].type == PERF_TYPE_HARDWARE && - attrs[counter].config == PERF_COUNT_HW_CPU_CYCLES) + if (MATCH_EVENT(HARDWARE, HW_CPU_CYCLES, counter)) runtime_cycles[run_idx] = count[0]; } @@ -292,9 +289,7 @@ static void nsec_printout(int counter, u64 *count, u64 *noise) fprintf(stderr, " %14.6f %-24s", msecs, event_name(counter)); - if (attrs[counter].type == PERF_TYPE_SOFTWARE && - attrs[counter].config == PERF_COUNT_SW_TASK_CLOCK) { - + if (MATCH_EVENT(SOFTWARE, SW_TASK_CLOCK, counter)) { if (walltime_nsecs_avg) fprintf(stderr, " # %10.3f CPUs ", (double)count[0] / (double)walltime_nsecs_avg); @@ -307,9 +302,7 @@ static void abs_printout(int counter, u64 *count, u64 *noise) fprintf(stderr, " %14Ld %-24s", count[0], event_name(counter)); if (runtime_cycles_avg && - attrs[counter].type == PERF_TYPE_HARDWARE && - attrs[counter].config == PERF_COUNT_HW_INSTRUCTIONS) { - + MATCH_EVENT(HARDWARE, HW_INSTRUCTIONS, counter)) { fprintf(stderr, " # %10.3f IPC ", (double)count[0] / (double)runtime_cycles_avg); } else { -- cgit v0.10.2 From 73c24cb86c51ff6445b292d9914d31236204393b Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 1 Jul 2009 18:36:18 +0530 Subject: perf list: Add cache events After: $ ./perf list List of pre-defined events (to be used in -e): cpu-cycles OR cycles [Hardware event] instructions [Hardware event] cache-references [Hardware event] cache-misses [Hardware event] branch-instructions OR branches [Hardware event] branch-misses [Hardware event] bus-cycles [Hardware event] cpu-clock [Software event] task-clock [Software event] page-faults OR faults [Software event] minor-faults [Software event] major-faults [Software event] context-switches OR cs [Software event] cpu-migrations OR migrations [Software event] L1-d$-loads [Hardware cache event] L1-d$-load-misses [Hardware cache event] L1-d$-stores [Hardware cache event] L1-d$-store-misses [Hardware cache event] L1-d$-prefetches [Hardware cache event] L1-d$-prefetch-misses [Hardware cache event] L1-i$-loads [Hardware cache event] L1-i$-load-misses [Hardware cache event] L1-i$-prefetches [Hardware cache event] L1-i$-prefetch-misses [Hardware cache event] LLC-loads [Hardware cache event] LLC-load-misses [Hardware cache event] LLC-stores [Hardware cache event] LLC-store-misses [Hardware cache event] LLC-prefetches [Hardware cache event] LLC-prefetch-misses [Hardware cache event] dTLB-loads [Hardware cache event] dTLB-load-misses [Hardware cache event] dTLB-stores [Hardware cache event] dTLB-store-misses [Hardware cache event] dTLB-prefetches [Hardware cache event] dTLB-prefetch-misses [Hardware cache event] iTLB-loads [Hardware cache event] iTLB-load-misses [Hardware cache event] branch-loads [Hardware cache event] branch-load-misses [Hardware cache event] rNNN [raw hardware event descriptor] Signed-off-by: Jaswinder Singh Rajput Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <1246453578.3072.1.camel@ht.satnam> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index aed7090..5184959 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -428,7 +428,7 @@ static const char * const event_type_descriptors[] = { void print_events(void) { struct event_symbol *syms = event_symbols; - unsigned int i, type, prev_type = -1; + unsigned int i, type, op, prev_type = -1; char name[40]; fprintf(stderr, "\n"); @@ -453,6 +453,21 @@ void print_events(void) } fprintf(stderr, "\n"); + for (type = 0; type < PERF_COUNT_HW_CACHE_MAX; type++) { + for (op = 0; op < PERF_COUNT_HW_CACHE_OP_MAX; op++) { + /* skip invalid cache type */ + if (!is_cache_op_valid(type, op)) + continue; + + for (i = 0; i < PERF_COUNT_HW_CACHE_RESULT_MAX; i++) { + fprintf(stderr, " %-40s [%s]\n", + event_cache_name(type, op, i), + event_type_descriptors[4]); + } + } + } + + fprintf(stderr, "\n"); fprintf(stderr, " %-40s [raw hardware event descriptor]\n", "rNNN"); fprintf(stderr, "\n"); -- cgit v0.10.2 From 43cbcd8acb4c992cbd22d1ec8a08c0591be5d719 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 1 Jul 2009 12:28:37 -0300 Subject: perf_counter tools: Share rbtree.with the kernel The tools/perf/util/rbtree.c copy already drifted by three csets: 4b324126e0c6c3a5080ca3ec0981e8766ed6f1ee 4c60117811171d867d4f27f17ea07d7419d45dae 16c047add3ceaf0ab882e3e094d1ec904d02312d So remove the copy and use the lib/rbtree.c directly, sharing the source code while still generating a separate object file, since tools/perf uses a far more agressive -O6 switch. Signed-off-by: Arnaldo Carvalho de Melo Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <20090701152837.GG15682@ghostprotocols.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index eddf076..c4f4882 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -223,7 +223,7 @@ SPARSE_FLAGS = -D__BIG_ENDIAN__ -D__powerpc__ # Those must not be GNU-specific; they are shared with perl/ which may # be built by a different compiler. (Note that this is an artifact now # but it still might be nice to keep that distinction.) -BASIC_CFLAGS = +BASIC_CFLAGS = -Iutil/include BASIC_LDFLAGS = # Guard against environment variables @@ -289,10 +289,10 @@ export PERL_PATH LIB_FILE=libperf.a LIB_H += ../../include/linux/perf_counter.h +LIB_H += ../../include/linux/rbtree.h LIB_H += perf.h LIB_H += util/types.h LIB_H += util/list.h -LIB_H += util/rbtree.h LIB_H += util/levenshtein.h LIB_H += util/parse-options.h LIB_H += util/parse-events.h @@ -691,6 +691,9 @@ builtin-init-db.o: builtin-init-db.c PERF-CFLAGS util/config.o: util/config.c PERF-CFLAGS $(QUIET_CC)$(CC) -o $*.o -c $(ALL_CFLAGS) -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $< +util/rbtree.o: ../../lib/rbtree.c PERF-CFLAGS + $(QUIET_CC)$(CC) -o util/rbtree.o -c $(ALL_CFLAGS) -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $< + perf-%$X: %.o $(PERFLIBS) $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $(filter %.o,$^) $(LIBS) diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 6cba70d..1960c22 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -12,7 +12,7 @@ #include "util/color.h" #include "util/list.h" #include "util/cache.h" -#include "util/rbtree.h" +#include #include "util/symbol.h" #include "util/string.h" diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 007363d..0d35e2b 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -12,7 +12,7 @@ #include "util/color.h" #include "util/list.h" #include "util/cache.h" -#include "util/rbtree.h" +#include #include "util/symbol.h" #include "util/string.h" #include "util/callchain.h" diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 5f5e7df..cdc74cf 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -23,7 +23,7 @@ #include "util/symbol.h" #include "util/color.h" #include "util/util.h" -#include "util/rbtree.h" +#include #include "util/parse-options.h" #include "util/parse-events.h" diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index 251d99e..0606b8f 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -3,7 +3,7 @@ #include "../perf.h" #include "list.h" -#include "rbtree.h" +#include #include "symbol.h" diff --git a/tools/perf/util/include/linux/kernel.h b/tools/perf/util/include/linux/kernel.h new file mode 100644 index 0000000..99c1b3d --- /dev/null +++ b/tools/perf/util/include/linux/kernel.h @@ -0,0 +1,21 @@ +#ifndef PERF_LINUX_KERNEL_H_ +#define PERF_LINUX_KERNEL_H_ + +#ifndef offsetof +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#endif + +#ifndef container_of +/** + * container_of - cast a member of a structure out to the containing structure + * @ptr: the pointer to the member. + * @type: the type of the container struct this is embedded in. + * @member: the name of the member within the struct. + * + */ +#define container_of(ptr, type, member) ({ \ + const typeof(((type *)0)->member) * __mptr = (ptr); \ + (type *)((char *)__mptr - offsetof(type, member)); }) +#endif + +#endif diff --git a/tools/perf/util/include/linux/module.h b/tools/perf/util/include/linux/module.h new file mode 100644 index 0000000..b43e2dc --- /dev/null +++ b/tools/perf/util/include/linux/module.h @@ -0,0 +1,6 @@ +#ifndef PERF_LINUX_MODULE_H +#define PERF_LINUX_MODULE_H + +#define EXPORT_SYMBOL(name) + +#endif diff --git a/tools/perf/util/include/linux/rbtree.h b/tools/perf/util/include/linux/rbtree.h new file mode 100644 index 0000000..7a243a1 --- /dev/null +++ b/tools/perf/util/include/linux/rbtree.h @@ -0,0 +1 @@ +#include "../../../../include/linux/rbtree.h" diff --git a/tools/perf/util/rbtree.c b/tools/perf/util/rbtree.c deleted file mode 100644 index b15ba9c..0000000 --- a/tools/perf/util/rbtree.c +++ /dev/null @@ -1,383 +0,0 @@ -/* - Red Black Trees - (C) 1999 Andrea Arcangeli - (C) 2002 David Woodhouse - - This program is free software; you can redistribute it and/or modify - it under the terms of the GNU General Public License as published by - the Free Software Foundation; either version 2 of the License, or - (at your option) any later version. - - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. - - You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - - linux/lib/rbtree.c -*/ - -#include "rbtree.h" - -static void __rb_rotate_left(struct rb_node *node, struct rb_root *root) -{ - struct rb_node *right = node->rb_right; - struct rb_node *parent = rb_parent(node); - - if ((node->rb_right = right->rb_left)) - rb_set_parent(right->rb_left, node); - right->rb_left = node; - - rb_set_parent(right, parent); - - if (parent) - { - if (node == parent->rb_left) - parent->rb_left = right; - else - parent->rb_right = right; - } - else - root->rb_node = right; - rb_set_parent(node, right); -} - -static void __rb_rotate_right(struct rb_node *node, struct rb_root *root) -{ - struct rb_node *left = node->rb_left; - struct rb_node *parent = rb_parent(node); - - if ((node->rb_left = left->rb_right)) - rb_set_parent(left->rb_right, node); - left->rb_right = node; - - rb_set_parent(left, parent); - - if (parent) - { - if (node == parent->rb_right) - parent->rb_right = left; - else - parent->rb_left = left; - } - else - root->rb_node = left; - rb_set_parent(node, left); -} - -void rb_insert_color(struct rb_node *node, struct rb_root *root) -{ - struct rb_node *parent, *gparent; - - while ((parent = rb_parent(node)) && rb_is_red(parent)) - { - gparent = rb_parent(parent); - - if (parent == gparent->rb_left) - { - { - register struct rb_node *uncle = gparent->rb_right; - if (uncle && rb_is_red(uncle)) - { - rb_set_black(uncle); - rb_set_black(parent); - rb_set_red(gparent); - node = gparent; - continue; - } - } - - if (parent->rb_right == node) - { - register struct rb_node *tmp; - __rb_rotate_left(parent, root); - tmp = parent; - parent = node; - node = tmp; - } - - rb_set_black(parent); - rb_set_red(gparent); - __rb_rotate_right(gparent, root); - } else { - { - register struct rb_node *uncle = gparent->rb_left; - if (uncle && rb_is_red(uncle)) - { - rb_set_black(uncle); - rb_set_black(parent); - rb_set_red(gparent); - node = gparent; - continue; - } - } - - if (parent->rb_left == node) - { - register struct rb_node *tmp; - __rb_rotate_right(parent, root); - tmp = parent; - parent = node; - node = tmp; - } - - rb_set_black(parent); - rb_set_red(gparent); - __rb_rotate_left(gparent, root); - } - } - - rb_set_black(root->rb_node); -} - -static void __rb_erase_color(struct rb_node *node, struct rb_node *parent, - struct rb_root *root) -{ - struct rb_node *other; - - while ((!node || rb_is_black(node)) && node != root->rb_node) - { - if (parent->rb_left == node) - { - other = parent->rb_right; - if (rb_is_red(other)) - { - rb_set_black(other); - rb_set_red(parent); - __rb_rotate_left(parent, root); - other = parent->rb_right; - } - if ((!other->rb_left || rb_is_black(other->rb_left)) && - (!other->rb_right || rb_is_black(other->rb_right))) - { - rb_set_red(other); - node = parent; - parent = rb_parent(node); - } - else - { - if (!other->rb_right || rb_is_black(other->rb_right)) - { - rb_set_black(other->rb_left); - rb_set_red(other); - __rb_rotate_right(other, root); - other = parent->rb_right; - } - rb_set_color(other, rb_color(parent)); - rb_set_black(parent); - rb_set_black(other->rb_right); - __rb_rotate_left(parent, root); - node = root->rb_node; - break; - } - } - else - { - other = parent->rb_left; - if (rb_is_red(other)) - { - rb_set_black(other); - rb_set_red(parent); - __rb_rotate_right(parent, root); - other = parent->rb_left; - } - if ((!other->rb_left || rb_is_black(other->rb_left)) && - (!other->rb_right || rb_is_black(other->rb_right))) - { - rb_set_red(other); - node = parent; - parent = rb_parent(node); - } - else - { - if (!other->rb_left || rb_is_black(other->rb_left)) - { - rb_set_black(other->rb_right); - rb_set_red(other); - __rb_rotate_left(other, root); - other = parent->rb_left; - } - rb_set_color(other, rb_color(parent)); - rb_set_black(parent); - rb_set_black(other->rb_left); - __rb_rotate_right(parent, root); - node = root->rb_node; - break; - } - } - } - if (node) - rb_set_black(node); -} - -void rb_erase(struct rb_node *node, struct rb_root *root) -{ - struct rb_node *child, *parent; - int color; - - if (!node->rb_left) - child = node->rb_right; - else if (!node->rb_right) - child = node->rb_left; - else - { - struct rb_node *old = node, *left; - - node = node->rb_right; - while ((left = node->rb_left) != NULL) - node = left; - child = node->rb_right; - parent = rb_parent(node); - color = rb_color(node); - - if (child) - rb_set_parent(child, parent); - if (parent == old) { - parent->rb_right = child; - parent = node; - } else - parent->rb_left = child; - - node->rb_parent_color = old->rb_parent_color; - node->rb_right = old->rb_right; - node->rb_left = old->rb_left; - - if (rb_parent(old)) - { - if (rb_parent(old)->rb_left == old) - rb_parent(old)->rb_left = node; - else - rb_parent(old)->rb_right = node; - } else - root->rb_node = node; - - rb_set_parent(old->rb_left, node); - if (old->rb_right) - rb_set_parent(old->rb_right, node); - goto color; - } - - parent = rb_parent(node); - color = rb_color(node); - - if (child) - rb_set_parent(child, parent); - if (parent) - { - if (parent->rb_left == node) - parent->rb_left = child; - else - parent->rb_right = child; - } - else - root->rb_node = child; - - color: - if (color == RB_BLACK) - __rb_erase_color(child, parent, root); -} - -/* - * This function returns the first node (in sort order) of the tree. - */ -struct rb_node *rb_first(const struct rb_root *root) -{ - struct rb_node *n; - - n = root->rb_node; - if (!n) - return NULL; - while (n->rb_left) - n = n->rb_left; - return n; -} - -struct rb_node *rb_last(const struct rb_root *root) -{ - struct rb_node *n; - - n = root->rb_node; - if (!n) - return NULL; - while (n->rb_right) - n = n->rb_right; - return n; -} - -struct rb_node *rb_next(const struct rb_node *node) -{ - struct rb_node *parent; - - if (rb_parent(node) == node) - return NULL; - - /* If we have a right-hand child, go down and then left as far - as we can. */ - if (node->rb_right) { - node = node->rb_right; - while (node->rb_left) - node=node->rb_left; - return (struct rb_node *)node; - } - - /* No right-hand children. Everything down and left is - smaller than us, so any 'next' node must be in the general - direction of our parent. Go up the tree; any time the - ancestor is a right-hand child of its parent, keep going - up. First time it's a left-hand child of its parent, said - parent is our 'next' node. */ - while ((parent = rb_parent(node)) && node == parent->rb_right) - node = parent; - - return parent; -} - -struct rb_node *rb_prev(const struct rb_node *node) -{ - struct rb_node *parent; - - if (rb_parent(node) == node) - return NULL; - - /* If we have a left-hand child, go down and then right as far - as we can. */ - if (node->rb_left) { - node = node->rb_left; - while (node->rb_right) - node=node->rb_right; - return (struct rb_node *)node; - } - - /* No left-hand children. Go up till we find an ancestor which - is a right-hand child of its parent */ - while ((parent = rb_parent(node)) && node == parent->rb_left) - node = parent; - - return parent; -} - -void rb_replace_node(struct rb_node *victim, struct rb_node *new, - struct rb_root *root) -{ - struct rb_node *parent = rb_parent(victim); - - /* Set the surrounding nodes to point to the replacement */ - if (parent) { - if (victim == parent->rb_left) - parent->rb_left = new; - else - parent->rb_right = new; - } else { - root->rb_node = new; - } - if (victim->rb_left) - rb_set_parent(victim->rb_left, new); - if (victim->rb_right) - rb_set_parent(victim->rb_right, new); - - /* Copy the pointers/colour from the victim to the replacement */ - *new = *victim; -} diff --git a/tools/perf/util/rbtree.h b/tools/perf/util/rbtree.h deleted file mode 100644 index 6bdc488..0000000 --- a/tools/perf/util/rbtree.h +++ /dev/null @@ -1,171 +0,0 @@ -/* - Red Black Trees - (C) 1999 Andrea Arcangeli - - This program is free software; you can redistribute it and/or modify - it under the terms of the GNU General Public License as published by - the Free Software Foundation; either version 2 of the License, or - (at your option) any later version. - - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. - - You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - - linux/include/linux/rbtree.h - - To use rbtrees you'll have to implement your own insert and search cores. - This will avoid us to use callbacks and to drop drammatically performances. - I know it's not the cleaner way, but in C (not in C++) to get - performances and genericity... - - Some example of insert and search follows here. The search is a plain - normal search over an ordered tree. The insert instead must be implemented - int two steps: as first thing the code must insert the element in - order as a red leaf in the tree, then the support library function - rb_insert_color() must be called. Such function will do the - not trivial work to rebalance the rbtree if necessary. - ------------------------------------------------------------------------ -static inline struct page * rb_search_page_cache(struct inode * inode, - unsigned long offset) -{ - struct rb_node * n = inode->i_rb_page_cache.rb_node; - struct page * page; - - while (n) - { - page = rb_entry(n, struct page, rb_page_cache); - - if (offset < page->offset) - n = n->rb_left; - else if (offset > page->offset) - n = n->rb_right; - else - return page; - } - return NULL; -} - -static inline struct page * __rb_insert_page_cache(struct inode * inode, - unsigned long offset, - struct rb_node * node) -{ - struct rb_node ** p = &inode->i_rb_page_cache.rb_node; - struct rb_node * parent = NULL; - struct page * page; - - while (*p) - { - parent = *p; - page = rb_entry(parent, struct page, rb_page_cache); - - if (offset < page->offset) - p = &(*p)->rb_left; - else if (offset > page->offset) - p = &(*p)->rb_right; - else - return page; - } - - rb_link_node(node, parent, p); - - return NULL; -} - -static inline struct page * rb_insert_page_cache(struct inode * inode, - unsigned long offset, - struct rb_node * node) -{ - struct page * ret; - if ((ret = __rb_insert_page_cache(inode, offset, node))) - goto out; - rb_insert_color(node, &inode->i_rb_page_cache); - out: - return ret; -} ------------------------------------------------------------------------ -*/ - -#ifndef _LINUX_RBTREE_H -#define _LINUX_RBTREE_H - -#include - -/** - * container_of - cast a member of a structure out to the containing structure - * @ptr: the pointer to the member. - * @type: the type of the container struct this is embedded in. - * @member: the name of the member within the struct. - * - */ -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -struct rb_node -{ - unsigned long rb_parent_color; -#define RB_RED 0 -#define RB_BLACK 1 - struct rb_node *rb_right; - struct rb_node *rb_left; -} __attribute__((aligned(sizeof(long)))); - /* The alignment might seem pointless, but allegedly CRIS needs it */ - -struct rb_root -{ - struct rb_node *rb_node; -}; - - -#define rb_parent(r) ((struct rb_node *)((r)->rb_parent_color & ~3)) -#define rb_color(r) ((r)->rb_parent_color & 1) -#define rb_is_red(r) (!rb_color(r)) -#define rb_is_black(r) rb_color(r) -#define rb_set_red(r) do { (r)->rb_parent_color &= ~1; } while (0) -#define rb_set_black(r) do { (r)->rb_parent_color |= 1; } while (0) - -static inline void rb_set_parent(struct rb_node *rb, struct rb_node *p) -{ - rb->rb_parent_color = (rb->rb_parent_color & 3) | (unsigned long)p; -} -static inline void rb_set_color(struct rb_node *rb, int color) -{ - rb->rb_parent_color = (rb->rb_parent_color & ~1) | color; -} - -#define RB_ROOT (struct rb_root) { NULL, } -#define rb_entry(ptr, type, member) container_of(ptr, type, member) - -#define RB_EMPTY_ROOT(root) ((root)->rb_node == NULL) -#define RB_EMPTY_NODE(node) (rb_parent(node) == node) -#define RB_CLEAR_NODE(node) (rb_set_parent(node, node)) - -extern void rb_insert_color(struct rb_node *, struct rb_root *); -extern void rb_erase(struct rb_node *, struct rb_root *); - -/* Find logical next and previous nodes in a tree */ -extern struct rb_node *rb_next(const struct rb_node *); -extern struct rb_node *rb_prev(const struct rb_node *); -extern struct rb_node *rb_first(const struct rb_root *); -extern struct rb_node *rb_last(const struct rb_root *); - -/* Fast replacement of a single node without remove/rebalance/add/rebalance */ -extern void rb_replace_node(struct rb_node *victim, struct rb_node *new, - struct rb_root *root); - -static inline void rb_link_node(struct rb_node * node, struct rb_node * parent, - struct rb_node ** rb_link) -{ - node->rb_parent_color = (unsigned long )parent; - node->rb_left = node->rb_right = NULL; - - *rb_link = node; -} - -#endif /* _LINUX_RBTREE_H */ diff --git a/tools/perf/util/strlist.h b/tools/perf/util/strlist.h index 2fb117f..2fdcfee 100644 --- a/tools/perf/util/strlist.h +++ b/tools/perf/util/strlist.h @@ -1,7 +1,7 @@ #ifndef STRLIST_H_ #define STRLIST_H_ -#include "rbtree.h" +#include #include struct str_node { diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 2c48ace..3a275db 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -4,7 +4,7 @@ #include #include "types.h" #include "list.h" -#include "rbtree.h" +#include struct symbol { struct rb_node rb_node; -- cgit v0.10.2 From 5da50258584469ddfee6545feb4eb2252a8d7e7b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 1 Jul 2009 14:46:08 -0300 Subject: perf_counter tools: Share list.h with the kernel The copy we were using came from another copy I did for the dwarves (pahole) package, that came from the kernel years ago. The only function that is used by the perf tools and that isn't in the kernel is list_del_range, that I'm leaving in the perf tools only for now. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <20090701174608.GA5823@ghostprotocols.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index c4f4882..8f7fd1b 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -290,9 +290,10 @@ LIB_FILE=libperf.a LIB_H += ../../include/linux/perf_counter.h LIB_H += ../../include/linux/rbtree.h +LIB_H += ../../include/linux/list.h +LIB_H += util/include/linux/list.h LIB_H += perf.h LIB_H += util/types.h -LIB_H += util/list.h LIB_H += util/levenshtein.h LIB_H += util/parse-options.h LIB_H += util/parse-events.h diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 1960c22..132de8b 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -10,7 +10,7 @@ #include "util/util.h" #include "util/color.h" -#include "util/list.h" +#include #include "util/cache.h" #include #include "util/symbol.h" diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 0d35e2b..be1b758 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -10,7 +10,7 @@ #include "util/util.h" #include "util/color.h" -#include "util/list.h" +#include #include "util/cache.h" #include #include "util/symbol.h" diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index 0606b8f..e9bd5e8 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -2,7 +2,7 @@ #define __PERF_CALLCHAIN_H #include "../perf.h" -#include "list.h" +#include #include #include "symbol.h" diff --git a/tools/perf/util/include/asm/system.h b/tools/perf/util/include/asm/system.h new file mode 100644 index 0000000..710cecc --- /dev/null +++ b/tools/perf/util/include/asm/system.h @@ -0,0 +1 @@ +/* Empty */ diff --git a/tools/perf/util/include/linux/list.h b/tools/perf/util/include/linux/list.h new file mode 100644 index 0000000..dbe4b81 --- /dev/null +++ b/tools/perf/util/include/linux/list.h @@ -0,0 +1,18 @@ +#include "../../../../include/linux/list.h" + +#ifndef PERF_LIST_H +#define PERF_LIST_H +/** + * list_del_range - deletes range of entries from list. + * @begin: first element in the range to delete from the list. + * @end: last element in the range to delete from the list. + * Note: list_empty on the range of entries does not return true after this, + * the entries is in an undefined state. + */ +static inline void list_del_range(struct list_head *begin, + struct list_head *end) +{ + begin->prev->next = end->next; + end->next->prev = begin->prev; +} +#endif diff --git a/tools/perf/util/include/linux/poison.h b/tools/perf/util/include/linux/poison.h new file mode 100644 index 0000000..fef6dbc --- /dev/null +++ b/tools/perf/util/include/linux/poison.h @@ -0,0 +1 @@ +#include "../../../../include/linux/poison.h" diff --git a/tools/perf/util/include/linux/prefetch.h b/tools/perf/util/include/linux/prefetch.h new file mode 100644 index 0000000..7841e48 --- /dev/null +++ b/tools/perf/util/include/linux/prefetch.h @@ -0,0 +1,6 @@ +#ifndef PERF_LINUX_PREFETCH_H +#define PERF_LINUX_PREFETCH_H + +static inline void prefetch(void *a __attribute__((unused))) { } + +#endif diff --git a/tools/perf/util/list.h b/tools/perf/util/list.h deleted file mode 100644 index e2548e8..0000000 --- a/tools/perf/util/list.h +++ /dev/null @@ -1,603 +0,0 @@ -#ifndef _LINUX_LIST_H -#define _LINUX_LIST_H -/* - Copyright (C) Cast of dozens, comes from the Linux kernel - - This program is free software; you can redistribute it and/or modify it - under the terms of version 2 of the GNU General Public License as - published by the Free Software Foundation. -*/ - -#include - -/* - * These are non-NULL pointers that will result in page faults - * under normal circumstances, used to verify that nobody uses - * non-initialized list entries. - */ -#define LIST_POISON1 ((void *)0x00100100) -#define LIST_POISON2 ((void *)0x00200200) - -/** - * container_of - cast a member of a structure out to the containing structure - * @ptr: the pointer to the member. - * @type: the type of the container struct this is embedded in. - * @member: the name of the member within the struct. - * - */ -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Simple doubly linked list implementation. - * - * Some of the internal functions ("__xxx") are useful when - * manipulating whole lists rather than single entries, as - * sometimes we already know the next/prev entries and we can - * generate better code by using them directly rather than - * using the generic single-entry routines. - */ - -struct list_head { - struct list_head *next, *prev; -}; - -#define LIST_HEAD_INIT(name) { &(name), &(name) } - -#define LIST_HEAD(name) \ - struct list_head name = LIST_HEAD_INIT(name) - -static inline void INIT_LIST_HEAD(struct list_head *list) -{ - list->next = list; - list->prev = list; -} - -/* - * Insert a new entry between two known consecutive entries. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -static inline void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next) -{ - next->prev = new; - new->next = next; - new->prev = prev; - prev->next = new; -} - -/** - * list_add - add a new entry - * @new: new entry to be added - * @head: list head to add it after - * - * Insert a new entry after the specified head. - * This is good for implementing stacks. - */ -static inline void list_add(struct list_head *new, struct list_head *head) -{ - __list_add(new, head, head->next); -} - -/** - * list_add_tail - add a new entry - * @new: new entry to be added - * @head: list head to add it before - * - * Insert a new entry before the specified head. - * This is useful for implementing queues. - */ -static inline void list_add_tail(struct list_head *new, struct list_head *head) -{ - __list_add(new, head->prev, head); -} - -/* - * Delete a list entry by making the prev/next entries - * point to each other. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -static inline void __list_del(struct list_head * prev, struct list_head * next) -{ - next->prev = prev; - prev->next = next; -} - -/** - * list_del - deletes entry from list. - * @entry: the element to delete from the list. - * Note: list_empty on entry does not return true after this, the entry is - * in an undefined state. - */ -static inline void list_del(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - entry->next = LIST_POISON1; - entry->prev = LIST_POISON2; -} - -/** - * list_del_range - deletes range of entries from list. - * @beging: first element in the range to delete from the list. - * @beging: first element in the range to delete from the list. - * Note: list_empty on the range of entries does not return true after this, - * the entries is in an undefined state. - */ -static inline void list_del_range(struct list_head *begin, - struct list_head *end) -{ - begin->prev->next = end->next; - end->next->prev = begin->prev; -} - -/** - * list_replace - replace old entry by new one - * @old : the element to be replaced - * @new : the new element to insert - * Note: if 'old' was empty, it will be overwritten. - */ -static inline void list_replace(struct list_head *old, - struct list_head *new) -{ - new->next = old->next; - new->next->prev = new; - new->prev = old->prev; - new->prev->next = new; -} - -static inline void list_replace_init(struct list_head *old, - struct list_head *new) -{ - list_replace(old, new); - INIT_LIST_HEAD(old); -} - -/** - * list_del_init - deletes entry from list and reinitialize it. - * @entry: the element to delete from the list. - */ -static inline void list_del_init(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - INIT_LIST_HEAD(entry); -} - -/** - * list_move - delete from one list and add as another's head - * @list: the entry to move - * @head: the head that will precede our entry - */ -static inline void list_move(struct list_head *list, struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add(list, head); -} - -/** - * list_move_tail - delete from one list and add as another's tail - * @list: the entry to move - * @head: the head that will follow our entry - */ -static inline void list_move_tail(struct list_head *list, - struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add_tail(list, head); -} - -/** - * list_is_last - tests whether @list is the last entry in list @head - * @list: the entry to test - * @head: the head of the list - */ -static inline int list_is_last(const struct list_head *list, - const struct list_head *head) -{ - return list->next == head; -} - -/** - * list_empty - tests whether a list is empty - * @head: the list to test. - */ -static inline int list_empty(const struct list_head *head) -{ - return head->next == head; -} - -/** - * list_empty_careful - tests whether a list is empty and not being modified - * @head: the list to test - * - * Description: - * tests whether a list is empty _and_ checks that no other CPU might be - * in the process of modifying either member (next or prev) - * - * NOTE: using list_empty_careful() without synchronization - * can only be safe if the only activity that can happen - * to the list entry is list_del_init(). Eg. it cannot be used - * if another CPU could re-list_add() it. - */ -static inline int list_empty_careful(const struct list_head *head) -{ - struct list_head *next = head->next; - return (next == head) && (next == head->prev); -} - -static inline void __list_splice(struct list_head *list, - struct list_head *head) -{ - struct list_head *first = list->next; - struct list_head *last = list->prev; - struct list_head *at = head->next; - - first->prev = head; - head->next = first; - - last->next = at; - at->prev = last; -} - -/** - * list_splice - join two lists - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice(struct list_head *list, struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head); -} - -/** - * list_splice_init - join two lists and reinitialise the emptied list. - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * The list at @list is reinitialised - */ -static inline void list_splice_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head); - INIT_LIST_HEAD(list); - } -} - -/** - * list_entry - get the struct for this entry - * @ptr: the &struct list_head pointer. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - */ -#define list_entry(ptr, type, member) \ - container_of(ptr, type, member) - -/** - * list_first_entry - get the first element from a list - * @ptr: the list head to take the element from. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - * - * Note, that list is expected to be not empty. - */ -#define list_first_entry(ptr, type, member) \ - list_entry((ptr)->next, type, member) - -/** - * list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each(pos, head) \ - for (pos = (head)->next; pos != (head); \ - pos = pos->next) - -/** - * __list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - * - * This variant differs from list_for_each() in that it's the - * simplest possible list iteration code, no prefetching is done. - * Use this for code that knows the list to be very short (empty - * or 1 entry) most of the time. - */ -#define __list_for_each(pos, head) \ - for (pos = (head)->next; pos != (head); pos = pos->next) - -/** - * list_for_each_prev - iterate over a list backwards - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each_prev(pos, head) \ - for (pos = (head)->prev; pos != (head); \ - pos = pos->prev) - -/** - * list_for_each_safe - iterate over a list safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_safe(pos, n, head) \ - for (pos = (head)->next, n = pos->next; pos != (head); \ - pos = n, n = pos->next) - -/** - * list_for_each_entry - iterate over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry(pos, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_reverse - iterate backwards over list of given type. - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_reverse(pos, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member); \ - &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue - * @pos: the type * to use as a start point - * @head: the head of the list - * @member: the name of the list_struct within the struct. - * - * Prepares a pos entry for use as a start point in list_for_each_entry_continue. - */ -#define list_prepare_entry(pos, head, member) \ - ((pos) ? : list_entry(head, typeof(*pos), member)) - -/** - * list_for_each_entry_continue - continue iteration over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Continue to iterate over list of given type, continuing after - * the current position. - */ -#define list_for_each_entry_continue(pos, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_from - iterate over list of given type from the current point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing from current position. - */ -#define list_for_each_entry_from(pos, head, member) \ - for (; &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_safe(pos, n, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_continue - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing after current point, - * safe against removal of list entry. - */ -#define list_for_each_entry_safe_continue(pos, n, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_from - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type from current point, safe against - * removal of list entry. - */ -#define list_for_each_entry_safe_from(pos, n, head, member) \ - for (n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_reverse - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate backwards over list of given type, safe against removal - * of list entry. - */ -#define list_for_each_entry_safe_reverse(pos, n, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member), \ - n = list_entry(pos->member.prev, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.prev, typeof(*n), member)) - -/* - * Double linked lists with a single pointer list head. - * Mostly useful for hash tables where the two pointer list head is - * too wasteful. - * You lose the ability to access the tail in O(1). - */ - -struct hlist_head { - struct hlist_node *first; -}; - -struct hlist_node { - struct hlist_node *next, **pprev; -}; - -#define HLIST_HEAD_INIT { .first = NULL } -#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } -#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) -static inline void INIT_HLIST_NODE(struct hlist_node *h) -{ - h->next = NULL; - h->pprev = NULL; -} - -static inline int hlist_unhashed(const struct hlist_node *h) -{ - return !h->pprev; -} - -static inline int hlist_empty(const struct hlist_head *h) -{ - return !h->first; -} - -static inline void __hlist_del(struct hlist_node *n) -{ - struct hlist_node *next = n->next; - struct hlist_node **pprev = n->pprev; - *pprev = next; - if (next) - next->pprev = pprev; -} - -static inline void hlist_del(struct hlist_node *n) -{ - __hlist_del(n); - n->next = LIST_POISON1; - n->pprev = LIST_POISON2; -} - -static inline void hlist_del_init(struct hlist_node *n) -{ - if (!hlist_unhashed(n)) { - __hlist_del(n); - INIT_HLIST_NODE(n); - } -} - -static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) -{ - struct hlist_node *first = h->first; - n->next = first; - if (first) - first->pprev = &n->next; - h->first = n; - n->pprev = &h->first; -} - -/* next must be != NULL */ -static inline void hlist_add_before(struct hlist_node *n, - struct hlist_node *next) -{ - n->pprev = next->pprev; - n->next = next; - next->pprev = &n->next; - *(n->pprev) = n; -} - -static inline void hlist_add_after(struct hlist_node *n, - struct hlist_node *next) -{ - next->next = n->next; - n->next = next; - next->pprev = &n->next; - - if(next->next) - next->next->pprev = &next->next; -} - -#define hlist_entry(ptr, type, member) container_of(ptr,type,member) - -#define hlist_for_each(pos, head) \ - for (pos = (head)->first; pos; \ - pos = pos->next) - -#define hlist_for_each_safe(pos, n, head) \ - for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ - pos = n) - -/** - * hlist_for_each_entry - iterate over list of given type - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry(tpos, pos, head, member) \ - for (pos = (head)->first; \ - pos && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_continue - iterate over a hlist continuing after current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_continue(tpos, pos, member) \ - for (pos = (pos)->next; \ - pos && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_from - iterate over a hlist continuing from current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_from(tpos, pos, member) \ - for (; pos && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @n: another &struct hlist_node to use as temporary storage - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ - for (pos = (head)->first; \ - pos && ({ n = pos->next; 1; }) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = n) - -#endif diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 3a275db..65a8449 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -3,7 +3,7 @@ #include #include "types.h" -#include "list.h" +#include #include struct symbol { -- cgit v0.10.2 From 0406ca6d8e849d9dd027c8cb6791448e81411aef Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 1 Jul 2009 21:02:09 +0200 Subject: perf_counter: Ignore the nmi call frames in the x86-64 backtraces About every callchains recorded with perf record are filled up including the internal perfcounter nmi frame: perf_callchain perf_counter_overflow intel_pmu_handle_irq perf_counter_nmi_handler notifier_call_chain atomic_notifier_call_chain notify_die do_nmi nmi We want ignore this frame as it's not interesting for instrumentation. To solve this, we simply ignore every frames from nmi context. New example of "perf report -s sym -c" after this patch: 9.59% [k] search_by_key 4.88% search_by_key reiserfs_read_locked_inode reiserfs_iget reiserfs_lookup do_lookup __link_path_walk path_walk do_path_lookup user_path_at vfs_fstatat vfs_lstat sys_newlstat system_call_fastpath __lxstat 0x406fb1 3.19% search_by_key search_by_entry_key reiserfs_find_entry reiserfs_lookup do_lookup __link_path_walk path_walk do_path_lookup user_path_at vfs_fstatat vfs_lstat sys_newlstat system_call_fastpath __lxstat 0x406fb1 [...] For now this patch only solves the problem in x86-64. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246474930-6088-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index f517944..cf86a5e 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -3,6 +3,8 @@ extern int kstack_depth_to_print; +int x86_is_stack_id(int id, char *name); + /* Generic stack tracer with callbacks */ struct stacktrace_ops { diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index d4cf4ce..36c3dc7 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -1561,6 +1561,7 @@ void callchain_store(struct perf_callchain_entry *entry, u64 ip) static DEFINE_PER_CPU(struct perf_callchain_entry, irq_entry); static DEFINE_PER_CPU(struct perf_callchain_entry, nmi_entry); +static DEFINE_PER_CPU(int, in_nmi_frame); static void @@ -1576,7 +1577,9 @@ static void backtrace_warning(void *data, char *msg) static int backtrace_stack(void *data, char *name) { - /* Process all stacks: */ + per_cpu(in_nmi_frame, smp_processor_id()) = + x86_is_stack_id(NMI_STACK, name); + return 0; } @@ -1584,6 +1587,9 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) { struct perf_callchain_entry *entry = data; + if (per_cpu(in_nmi_frame, smp_processor_id())) + return; + if (reliable) callchain_store(entry, addr); } diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index d593cd1..bca5fba 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -19,6 +19,12 @@ #include "dumpstack.h" +/* Just a stub for now */ +int x86_is_stack_id(int id, char *name) +{ + return 0; +} + void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data) diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index d35db59..54b0a32 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -19,10 +19,8 @@ #include "dumpstack.h" -static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, - unsigned *usedp, char **idp) -{ - static char ids[][8] = { + +static char x86_stack_ids[][8] = { [DEBUG_STACK - 1] = "#DB", [NMI_STACK - 1] = "NMI", [DOUBLEFAULT_STACK - 1] = "#DF", @@ -33,6 +31,15 @@ static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, N_EXCEPTION_STACKS + DEBUG_STKSZ / EXCEPTION_STKSZ - 2] = "#DB[?]" #endif }; + +int x86_is_stack_id(int id, char *name) +{ + return x86_stack_ids[id - 1] == name; +} + +static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, + unsigned *usedp, char **idp) +{ unsigned k; /* @@ -61,7 +68,7 @@ static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, if (*usedp & (1U << k)) break; *usedp |= 1U << k; - *idp = ids[k]; + *idp = x86_stack_ids[k]; return (unsigned long *)end; } /* @@ -81,12 +88,13 @@ static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, do { ++j; end -= EXCEPTION_STKSZ; - ids[j][4] = '1' + (j - N_EXCEPTION_STACKS); + x86_stack_ids[j][4] = '1' + + (j - N_EXCEPTION_STACKS); } while (stack < end - EXCEPTION_STKSZ); if (*usedp & (1U << j)) break; *usedp |= 1U << j; - *idp = ids[j]; + *idp = x86_stack_ids[j]; return (unsigned long *)end; } #endif -- cgit v0.10.2 From a92bef0f216bbf3d05c0c0709ea02e267f2b920e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 1 Jul 2009 21:02:10 +0200 Subject: perf stat: Handle pipe read failures in perf stat MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Building builtin-stat.c reports the following errors: cc1: warnings being treated as errors builtin-stat.c: In function ‘run_perf_stat’: builtin-stat.c:242: erreur: ignoring return value of ‘read’, declared with attribute warn_unused_result builtin-stat.c:255: erreur: ignoring return value of ‘read’, declared with attribute warn_unused_result make: *** [builtin-stat.o] Erreur 1 This patch handles the possible pipe read failures. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246474930-6088-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 01cc07e..27921a8 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -239,7 +239,8 @@ static int run_perf_stat(int argc __used, const char **argv) /* * Wait until the parent tells us to go. */ - read(go_pipe[0], &buf, 1); + if (read(go_pipe[0], &buf, 1) == -1) + perror("unable to read pipe"); execvp(argv[0], (char **)argv); @@ -252,7 +253,8 @@ static int run_perf_stat(int argc __used, const char **argv) */ close(child_ready_pipe[1]); close(go_pipe[0]); - read(child_ready_pipe[0], &buf, 1); + if (read(child_ready_pipe[0], &buf, 1) == -1) + perror("unable to read pipe"); close(child_ready_pipe[0]); for (counter = 0; counter < nr_counters; counter++) -- cgit v0.10.2 From 9974f496782b7612e36a143bedda858f1cb953d4 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 2 Jul 2009 08:05:58 +0200 Subject: perf_counter tools: Make symbol loading consistently return number of loaded symbols perf_counter tools: Make symbol loading consistently return number of loaded symbols. Signed-off-by: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246514758.13293.42.camel@marge.simson.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 132de8b..3becc8a 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -172,7 +172,7 @@ static int load_kernel(void) return -1; err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose); - if (err) { + if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; } else diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index be1b758..58d1612 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -189,7 +189,7 @@ static int load_kernel(void) return -1; err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose); - if (err) { + if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; } else diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index cdc74cf..97fde1d 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -364,7 +364,7 @@ static int parse_symbols(void) if (kernel_dso == NULL) return -1; - if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1) != 0) + if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1) <= 0) goto out_delete_dso; node = rb_first(&kernel_dso->syms); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 78c2efd..c077b6a 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -146,6 +146,7 @@ static int dso__load_kallsyms(struct dso *self, symbol_filter_t filter, int verb char *line = NULL; size_t n; FILE *file = fopen("/proc/kallsyms", "r"); + int count = 0; if (file == NULL) goto out_failure; @@ -188,8 +189,10 @@ static int dso__load_kallsyms(struct dso *self, symbol_filter_t filter, int verb if (filter && filter(self, sym)) symbol__delete(sym, self->sym_priv_size); - else + else { dso__insert_symbol(self, sym); + count++; + } } /* @@ -212,7 +215,7 @@ static int dso__load_kallsyms(struct dso *self, symbol_filter_t filter, int verb free(line); fclose(file); - return 0; + return count; out_delete_line: free(line); @@ -639,7 +642,7 @@ int dso__load_kernel(struct dso *self, const char *vmlinux, if (vmlinux) err = dso__load_vmlinux(self, vmlinux, filter, verbose); - if (err < 0) + if (err <= 0) err = dso__load_kallsyms(self, filter, verbose); return err; -- cgit v0.10.2 From 208b4b4a59351011b7f212e273f2b7bc47a9c482 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 2 Jul 2009 08:07:10 +0200 Subject: perf_counter tools: Add infrastructure to support loading of kernel module symbols Add infrastructure for module path discovery and section load addresses. Signed-off-by: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246514830.13293.44.camel@marge.simson.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 8f7fd1b..7822b3d 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -306,6 +306,7 @@ LIB_H += util/strlist.h LIB_H += util/run-command.h LIB_H += util/sigchain.h LIB_H += util/symbol.h +LIB_H += util/module.h LIB_H += util/color.h LIB_OBJS += util/abspath.o @@ -329,6 +330,7 @@ LIB_OBJS += util/usage.o LIB_OBJS += util/wrapper.o LIB_OBJS += util/sigchain.o LIB_OBJS += util/symbol.o +LIB_OBJS += util/module.o LIB_OBJS += util/color.o LIB_OBJS += util/pager.o LIB_OBJS += util/header.o diff --git a/tools/perf/util/module.c b/tools/perf/util/module.c new file mode 100644 index 0000000..ddabe92 --- /dev/null +++ b/tools/perf/util/module.c @@ -0,0 +1,509 @@ +#include "util.h" +#include "../perf.h" +#include "string.h" +#include "module.h" + +#include +#include +#include +#include +#include + +static unsigned int crc32(const char *p, unsigned int len) +{ + int i; + unsigned int crc = 0; + + while (len--) { + crc ^= *p++; + for (i = 0; i < 8; i++) + crc = (crc >> 1) ^ ((crc & 1) ? 0xedb88320 : 0); + } + return crc; +} + +/* module section methods */ + +struct sec_dso *sec_dso__new_dso(const char *name) +{ + struct sec_dso *self = malloc(sizeof(*self) + strlen(name) + 1); + + if (self != NULL) { + strcpy(self->name, name); + self->secs = RB_ROOT; + self->find_section = sec_dso__find_section; + } + + return self; +} + +static void sec_dso__delete_section(struct section *self) +{ + free(((void *)self)); +} + +void sec_dso__delete_sections(struct sec_dso *self) +{ + struct section *pos; + struct rb_node *next = rb_first(&self->secs); + + while (next) { + pos = rb_entry(next, struct section, rb_node); + next = rb_next(&pos->rb_node); + rb_erase(&pos->rb_node, &self->secs); + sec_dso__delete_section(pos); + } +} + +void sec_dso__delete_self(struct sec_dso *self) +{ + sec_dso__delete_sections(self); + free(self); +} + +static void sec_dso__insert_section(struct sec_dso *self, struct section *sec) +{ + struct rb_node **p = &self->secs.rb_node; + struct rb_node *parent = NULL; + const u64 hash = sec->hash; + struct section *s; + + while (*p != NULL) { + parent = *p; + s = rb_entry(parent, struct section, rb_node); + if (hash < s->hash) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + } + rb_link_node(&sec->rb_node, parent, p); + rb_insert_color(&sec->rb_node, &self->secs); +} + +struct section *sec_dso__find_section(struct sec_dso *self, const char *name) +{ + struct rb_node *n; + u64 hash; + int len; + + if (self == NULL) + return NULL; + + len = strlen(name); + hash = crc32(name, len); + + n = self->secs.rb_node; + + while (n) { + struct section *s = rb_entry(n, struct section, rb_node); + + if (hash < s->hash) + n = n->rb_left; + else if (hash > s->hash) + n = n->rb_right; + else { + if (!strcmp(name, s->name)) + return s; + else + n = rb_next(&s->rb_node); + } + } + + return NULL; +} + +static size_t sec_dso__fprintf_section(struct section *self, FILE *fp) +{ + return fprintf(fp, "name:%s vma:%llx path:%s\n", + self->name, self->vma, self->path); +} + +size_t sec_dso__fprintf(struct sec_dso *self, FILE *fp) +{ + size_t ret = fprintf(fp, "dso: %s\n", self->name); + + struct rb_node *nd; + for (nd = rb_first(&self->secs); nd; nd = rb_next(nd)) { + struct section *pos = rb_entry(nd, struct section, rb_node); + ret += sec_dso__fprintf_section(pos, fp); + } + + return ret; +} + +static struct section *section__new(const char *name, const char *path) +{ + struct section *self = calloc(1, sizeof(*self)); + + if (!self) + goto out_failure; + + self->name = calloc(1, strlen(name) + 1); + if (!self->name) + goto out_failure; + + self->path = calloc(1, strlen(path) + 1); + if (!self->path) + goto out_failure; + + strcpy(self->name, name); + strcpy(self->path, path); + self->hash = crc32(self->name, strlen(name)); + + return self; + +out_failure: + if (self) { + if (self->name) + free(self->name); + if (self->path) + free(self->path); + free(self); + } + + return NULL; +} + +/* module methods */ + +struct mod_dso *mod_dso__new_dso(const char *name) +{ + struct mod_dso *self = malloc(sizeof(*self) + strlen(name) + 1); + + if (self != NULL) { + strcpy(self->name, name); + self->mods = RB_ROOT; + self->find_module = mod_dso__find_module; + } + + return self; +} + +static void mod_dso__delete_module(struct module *self) +{ + free(((void *)self)); +} + +void mod_dso__delete_modules(struct mod_dso *self) +{ + struct module *pos; + struct rb_node *next = rb_first(&self->mods); + + while (next) { + pos = rb_entry(next, struct module, rb_node); + next = rb_next(&pos->rb_node); + rb_erase(&pos->rb_node, &self->mods); + mod_dso__delete_module(pos); + } +} + +void mod_dso__delete_self(struct mod_dso *self) +{ + mod_dso__delete_modules(self); + free(self); +} + +static void mod_dso__insert_module(struct mod_dso *self, struct module *mod) +{ + struct rb_node **p = &self->mods.rb_node; + struct rb_node *parent = NULL; + const u64 hash = mod->hash; + struct module *m; + + while (*p != NULL) { + parent = *p; + m = rb_entry(parent, struct module, rb_node); + if (hash < m->hash) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + } + rb_link_node(&mod->rb_node, parent, p); + rb_insert_color(&mod->rb_node, &self->mods); +} + +struct module *mod_dso__find_module(struct mod_dso *self, const char *name) +{ + struct rb_node *n; + u64 hash; + int len; + + if (self == NULL) + return NULL; + + len = strlen(name); + hash = crc32(name, len); + + n = self->mods.rb_node; + + while (n) { + struct module *m = rb_entry(n, struct module, rb_node); + + if (hash < m->hash) + n = n->rb_left; + else if (hash > m->hash) + n = n->rb_right; + else { + if (!strcmp(name, m->name)) + return m; + else + n = rb_next(&m->rb_node); + } + } + + return NULL; +} + +static size_t mod_dso__fprintf_module(struct module *self, FILE *fp) +{ + return fprintf(fp, "name:%s path:%s\n", self->name, self->path); +} + +size_t mod_dso__fprintf(struct mod_dso *self, FILE *fp) +{ + struct rb_node *nd; + size_t ret; + + ret = fprintf(fp, "dso: %s\n", self->name); + + for (nd = rb_first(&self->mods); nd; nd = rb_next(nd)) { + struct module *pos = rb_entry(nd, struct module, rb_node); + + ret += mod_dso__fprintf_module(pos, fp); + } + + return ret; +} + +static struct module *module__new(const char *name, const char *path) +{ + struct module *self = calloc(1, sizeof(*self)); + + if (!self) + goto out_failure; + + self->name = calloc(1, strlen(name) + 1); + if (!self->name) + goto out_failure; + + self->path = calloc(1, strlen(path) + 1); + if (!self->path) + goto out_failure; + + strcpy(self->name, name); + strcpy(self->path, path); + self->hash = crc32(self->name, strlen(name)); + + return self; + +out_failure: + if (self) { + if (self->name) + free(self->name); + if (self->path) + free(self->path); + free(self); + } + + return NULL; +} + +static int mod_dso__load_sections(struct module *mod) +{ + int count = 0, path_len; + struct dirent *entry; + char *line = NULL; + char *dir_path; + DIR *dir; + size_t n; + + path_len = strlen("/sys/module/"); + path_len += strlen(mod->name); + path_len += strlen("/sections/"); + + dir_path = calloc(1, path_len + 1); + if (dir_path == NULL) + goto out_failure; + + strcat(dir_path, "/sys/module/"); + strcat(dir_path, mod->name); + strcat(dir_path, "/sections/"); + + dir = opendir(dir_path); + if (dir == NULL) + goto out_free; + + while ((entry = readdir(dir))) { + struct section *section; + char *path, *vma; + int line_len; + FILE *file; + + if (!strcmp(".", entry->d_name) || !strcmp("..", entry->d_name)) + continue; + + path = calloc(1, path_len + strlen(entry->d_name) + 1); + if (path == NULL) + break; + strcat(path, dir_path); + strcat(path, entry->d_name); + + file = fopen(path, "r"); + if (file == NULL) { + free(path); + break; + } + + line_len = getline(&line, &n, file); + if (line_len < 0) { + free(path); + fclose(file); + break; + } + + if (!line) { + free(path); + fclose(file); + break; + } + + line[--line_len] = '\0'; /* \n */ + + vma = strstr(line, "0x"); + if (!vma) { + free(path); + fclose(file); + break; + } + vma += 2; + + section = section__new(entry->d_name, path); + if (!section) { + fprintf(stderr, "load_sections: allocation error\n"); + free(path); + fclose(file); + break; + } + + hex2u64(vma, §ion->vma); + sec_dso__insert_section(mod->sections, section); + + free(path); + fclose(file); + count++; + } + + closedir(dir); + free(line); + free(dir_path); + + return count; + +out_free: + free(dir_path); + +out_failure: + return count; +} + +static int mod_dso__load_module_paths(struct mod_dso *self) +{ + struct utsname uts; + int count = 0, len; + char *line = NULL; + FILE *file; + char *path; + size_t n; + + if (uname(&uts) < 0) + goto out_failure; + + len = strlen("/lib/modules/"); + len += strlen(uts.release); + len += strlen("/modules.dep"); + + path = calloc(1, len); + if (path == NULL) + goto out_failure; + + strcat(path, "/lib/modules/"); + strcat(path, uts.release); + strcat(path, "/modules.dep"); + + file = fopen(path, "r"); + free(path); + if (file == NULL) + goto out_failure; + + while (!feof(file)) { + char *path, *name, *tmp; + struct module *module; + int line_len, len; + + line_len = getline(&line, &n, file); + if (line_len < 0) + break; + + if (!line) + goto out_failure; + + line[--line_len] = '\0'; /* \n */ + + path = strtok(line, ":"); + if (!path) + goto out_failure; + + name = strdup(path); + name = strtok(name, "/"); + + tmp = name; + + while (tmp) { + tmp = strtok(NULL, "/"); + if (tmp) + name = tmp; + } + name = strsep(&name, "."); + + /* Quirk: replace '-' with '_' in sound modules */ + for (len = strlen(name); len; len--) { + if (*(name+len) == '-') + *(name+len) = '_'; + } + + module = module__new(name, path); + if (!module) { + fprintf(stderr, "load_module_paths: allocation error\n"); + goto out_failure; + } + mod_dso__insert_module(self, module); + + module->sections = sec_dso__new_dso("sections"); + if (!module->sections) { + fprintf(stderr, "load_module_paths: allocation error\n"); + goto out_failure; + } + + module->active = mod_dso__load_sections(module); + + if (module->active > 0) + count++; + } + + free(line); + fclose(file); + + return count; + +out_failure: + return -1; +} + +int mod_dso__load_modules(struct mod_dso *dso) +{ + int err; + + err = mod_dso__load_module_paths(dso); + + return err; +} diff --git a/tools/perf/util/module.h b/tools/perf/util/module.h new file mode 100644 index 0000000..8a592ef --- /dev/null +++ b/tools/perf/util/module.h @@ -0,0 +1,53 @@ +#ifndef _PERF_MODULE_ +#define _PERF_MODULE_ 1 + +#include +#include "../types.h" +#include +#include + +struct section { + struct rb_node rb_node; + u64 hash; + u64 vma; + char *name; + char *path; +}; + +struct sec_dso { + struct list_head node; + struct rb_root secs; + struct section *(*find_section)(struct sec_dso *, const char *name); + char name[0]; +}; + +struct module { + struct rb_node rb_node; + u64 hash; + char *name; + char *path; + struct sec_dso *sections; + int active; +}; + +struct mod_dso { + struct list_head node; + struct rb_root mods; + struct module *(*find_module)(struct mod_dso *, const char *name); + char name[0]; +}; + +struct sec_dso *sec_dso__new_dso(const char *name); +void sec_dso__delete_sections(struct sec_dso *self); +void sec_dso__delete_self(struct sec_dso *self); +size_t sec_dso__fprintf(struct sec_dso *self, FILE *fp); +struct section *sec_dso__find_section(struct sec_dso *self, const char *name); + +struct mod_dso *mod_dso__new_dso(const char *name); +void mod_dso__delete_modules(struct mod_dso *self); +void mod_dso__delete_self(struct mod_dso *self); +size_t mod_dso__fprintf(struct mod_dso *self, FILE *fp); +struct module *mod_dso__find_module(struct mod_dso *self, const char *name); +int mod_dso__load_modules(struct mod_dso *dso); + +#endif /* _PERF_MODULE_ */ -- cgit v0.10.2 From 6cfcc53ed4f3ecb9319e73a03f34f1eddcb644dd Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 2 Jul 2009 08:08:36 +0200 Subject: perf_counter tools: Connect module support infrastructure to symbol loading infrastructure Signed-off-by: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246514916.13293.46.camel@marge.simson.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 3becc8a..8820568 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -171,7 +171,7 @@ static int load_kernel(void) if (!kernel_dso) return -1; - err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose); + err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, 0); if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 58d1612..38d136f 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -188,7 +188,7 @@ static int load_kernel(void) if (!kernel_dso) return -1; - err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose); + err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, 0); if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 97fde1d..9bb25fc 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -364,7 +364,7 @@ static int parse_symbols(void) if (kernel_dso == NULL) return -1; - if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1) <= 0) + if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1, 0) <= 0) goto out_delete_dso; node = rb_first(&kernel_dso->syms); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index c077b6a..98a1311 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -35,7 +35,7 @@ static struct symbol *symbol__new(u64 start, u64 len, self = ((void *)self) + priv_size; } self->start = start; - self->end = start + len - 1; + self->end = len ? start + len - 1 : start; memcpy(self->name, name, namelen); return self; @@ -48,8 +48,12 @@ static void symbol__delete(struct symbol *self, unsigned int priv_size) static size_t symbol__fprintf(struct symbol *self, FILE *fp) { - return fprintf(fp, " %llx-%llx %s\n", + if (!self->module) + return fprintf(fp, " %llx-%llx %s\n", self->start, self->end, self->name); + else + return fprintf(fp, " %llx-%llx %s \t[%s]\n", + self->start, self->end, self->name, self->module->name); } struct dso *dso__new(const char *name, unsigned int sym_priv_size) @@ -310,6 +314,26 @@ static inline int elf_sym__is_function(const GElf_Sym *sym) sym->st_size != 0; } +static inline int elf_sym__is_label(const GElf_Sym *sym) +{ + return elf_sym__type(sym) == STT_NOTYPE && + sym->st_name != 0 && + sym->st_shndx != SHN_UNDEF && + sym->st_shndx != SHN_ABS; +} + +static inline const char *elf_sec__name(const GElf_Shdr *shdr, + const Elf_Data *secstrs) +{ + return secstrs->d_buf + shdr->sh_name; +} + +static inline int elf_sec__is_text(const GElf_Shdr *shdr, + const Elf_Data *secstrs) +{ + return strstr(elf_sec__name(shdr, secstrs), "text") != NULL; +} + static inline const char *elf_sym__name(const GElf_Sym *sym, const Elf_Data *symstrs) { @@ -451,9 +475,9 @@ static int dso__synthesize_plt_symbols(struct dso *self, Elf *elf, } static int dso__load_sym(struct dso *self, int fd, const char *name, - symbol_filter_t filter, int verbose) + symbol_filter_t filter, int verbose, struct module *mod) { - Elf_Data *symstrs; + Elf_Data *symstrs, *secstrs; uint32_t nr_syms; int err = -1; uint32_t index; @@ -461,7 +485,7 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, GElf_Shdr shdr; Elf_Data *syms; GElf_Sym sym; - Elf_Scn *sec, *sec_dynsym; + Elf_Scn *sec, *sec_dynsym, *sec_strndx; Elf *elf; size_t dynsym_idx; int nr = 0; @@ -520,6 +544,14 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, if (symstrs == NULL) goto out_elf_end; + sec_strndx = elf_getscn(elf, ehdr.e_shstrndx); + if (sec_strndx == NULL) + goto out_elf_end; + + secstrs = elf_getdata(sec_strndx, NULL); + if (symstrs == NULL) + goto out_elf_end; + nr_syms = shdr.sh_size / shdr.sh_entsize; memset(&sym, 0, sizeof(sym)); @@ -529,8 +561,11 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, elf_symtab__for_each_symbol(syms, nr_syms, index, sym) { struct symbol *f; u64 obj_start; + struct section *section = NULL; + int is_label = elf_sym__is_label(&sym); + const char *section_name; - if (!elf_sym__is_function(&sym)) + if (!is_label && !elf_sym__is_function(&sym)) continue; sec = elf_getscn(elf, sym.st_shndx); @@ -538,6 +573,11 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, goto out_elf_end; gelf_getshdr(sec, &shdr); + + if (is_label && !elf_sec__is_text(&shdr, secstrs)) + continue; + + section_name = elf_sec__name(&shdr, secstrs); obj_start = sym.st_value; if (self->prelinked) { @@ -548,6 +588,17 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, sym.st_value -= shdr.sh_addr - shdr.sh_offset; } + if (mod) { + section = mod->sections->find_section(mod->sections, section_name); + if (section) + sym.st_value += section->vma; + else { + fprintf(stderr, "dso__load_sym() module %s lookup of %s failed\n", + mod->name, section_name); + goto out_elf_end; + } + } + f = symbol__new(sym.st_value, sym.st_size, elf_sym__name(&sym, symstrs), self->sym_priv_size, obj_start, verbose); @@ -557,6 +608,7 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, if (filter && filter(self, f)) symbol__delete(f, self->sym_priv_size); else { + f->module = mod; dso__insert_symbol(self, f); nr++; } @@ -606,7 +658,7 @@ more: fd = open(name, O_RDONLY); } while (fd < 0); - ret = dso__load_sym(self, fd, name, filter, verbose); + ret = dso__load_sym(self, fd, name, filter, verbose, NULL); close(fd); /* @@ -620,6 +672,86 @@ out: return ret; } +static int dso__load_module(struct dso *self, struct mod_dso *mods, const char *name, + symbol_filter_t filter, int verbose) +{ + struct module *mod = mod_dso__find_module(mods, name); + int err = 0, fd; + + if (mod == NULL || !mod->active) + return err; + + fd = open(mod->path, O_RDONLY); + + if (fd < 0) + return err; + + err = dso__load_sym(self, fd, name, filter, verbose, mod); + close(fd); + + return err; +} + +int dso__load_modules(struct dso *self, symbol_filter_t filter, int verbose) +{ + struct mod_dso *mods = mod_dso__new_dso("modules"); + struct module *pos; + struct rb_node *next; + int err; + + err = mod_dso__load_modules(mods); + + if (err <= 0) + return err; + + /* + * Iterate over modules, and load active symbols. + */ + next = rb_first(&mods->mods); + while (next) { + pos = rb_entry(next, struct module, rb_node); + err = dso__load_module(self, mods, pos->name, filter, verbose); + + if (err < 0) + break; + + next = rb_next(&pos->rb_node); + } + + if (err < 0) { + mod_dso__delete_modules(mods); + mod_dso__delete_self(mods); + } + + return err; +} + +static inline void dso__fill_symbol_holes(struct dso *self) +{ + struct symbol *prev = NULL; + struct rb_node *nd; + + for (nd = rb_last(&self->syms); nd; nd = rb_prev(nd)) { + struct symbol *pos = rb_entry(nd, struct symbol, rb_node); + + if (prev) { + u64 hole = 0; + int alias = pos->start == prev->start; + + if (!alias) + hole = prev->start - pos->end - 1; + + if (hole || alias) { + if (alias) + pos->end = prev->end; + else if (hole) + pos->end = prev->start - 1; + } + } + prev = pos; + } +} + static int dso__load_vmlinux(struct dso *self, const char *vmlinux, symbol_filter_t filter, int verbose) { @@ -628,19 +760,26 @@ static int dso__load_vmlinux(struct dso *self, const char *vmlinux, if (fd < 0) return -1; - err = dso__load_sym(self, fd, vmlinux, filter, verbose); + err = dso__load_sym(self, fd, vmlinux, filter, verbose, NULL); + + if (err > 0) + dso__fill_symbol_holes(self); + close(fd); return err; } int dso__load_kernel(struct dso *self, const char *vmlinux, - symbol_filter_t filter, int verbose) + symbol_filter_t filter, int verbose, int modules) { int err = -1; - if (vmlinux) + if (vmlinux) { err = dso__load_vmlinux(self, vmlinux, filter, verbose); + if (err > 0 && modules) + err = dso__load_modules(self, filter, verbose); + } if (err <= 0) err = dso__load_kallsyms(self, filter, verbose); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 65a8449..4e141a3 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -5,6 +5,7 @@ #include "types.h" #include #include +#include "module.h" struct symbol { struct rb_node rb_node; @@ -13,6 +14,7 @@ struct symbol { u64 obj_start; u64 hist_sum; u64 *hist; + struct module *module; void *priv; char name[0]; }; @@ -41,7 +43,8 @@ static inline void *dso__sym_priv(struct dso *self, struct symbol *sym) struct symbol *dso__find_symbol(struct dso *self, u64 ip); int dso__load_kernel(struct dso *self, const char *vmlinux, - symbol_filter_t filter, int verbose); + symbol_filter_t filter, int verbose, int modules); +int dso__load_modules(struct dso *self, symbol_filter_t filter, int verbose); int dso__load(struct dso *self, symbol_filter_t filter, int verbose); size_t dso__fprintf(struct dso *self, FILE *fp); -- cgit v0.10.2 From 429764873cf3fc3e73142872a674bb27cda589c1 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 2 Jul 2009 08:09:46 +0200 Subject: perf_counter tools: Enable kernel module symbol loading in tools Add the -m/--modules option to perf report and perf annotate, which enables live module symbol/image loading. To be used with -k/--vmlinux. (Also give perf annotate a -P/--full-paths option.) Signed-off-by: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246514986.13293.48.camel@marge.simson.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 8820568..08ea6c5 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -43,6 +43,10 @@ static int dump_trace = 0; static int verbose; +static int modules; + +static int full_paths; + static int print_line; static unsigned long page_size; @@ -171,7 +175,7 @@ static int load_kernel(void) if (!kernel_dso) return -1; - err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, 0); + err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, modules); if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; @@ -1268,19 +1272,25 @@ static void print_summary(char *filename) static void annotate_sym(struct dso *dso, struct symbol *sym) { - char *filename = dso->name; + char *filename = dso->name, *d_filename; u64 start, end, len; char command[PATH_MAX*2]; FILE *file; if (!filename) return; - if (dso == kernel_dso) + if (sym->module) + filename = sym->module->path; + else if (dso == kernel_dso) filename = vmlinux; start = sym->obj_start; if (!start) start = sym->start; + if (full_paths) + d_filename = filename; + else + d_filename = basename(filename); end = start + sym->end - sym->start + 1; len = sym->end - sym->start; @@ -1291,13 +1301,14 @@ static void annotate_sym(struct dso *dso, struct symbol *sym) } printf("\n\n------------------------------------------------\n"); - printf(" Percent | Source code & Disassembly of %s\n", filename); + printf(" Percent | Source code & Disassembly of %s\n", d_filename); printf("------------------------------------------------\n"); if (verbose >= 2) printf("annotating [%p] %30s : [%p] %30s\n", dso, dso->name, sym, sym->name); - sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s", (u64)start, (u64)end, filename); + sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s|grep -v %s", + (u64)start, (u64)end, filename, filename); if (verbose >= 3) printf("doing: %s\n", command); @@ -1472,8 +1483,12 @@ static const struct option options[] = { OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), OPT_STRING('k', "vmlinux", &vmlinux, "file", "vmlinux pathname"), + OPT_BOOLEAN('m', "modules", &modules, + "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_BOOLEAN('l', "print-line", &print_line, "print matching source lines (may be slow)"), + OPT_BOOLEAN('P', "full-paths", &full_paths, + "Don't shorten the displayed pathnames"), OPT_END() }; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 38d136f..b44476c 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -46,6 +46,8 @@ static int dump_trace = 0; static int verbose; #define eprintf(x...) do { if (verbose) fprintf(stderr, x); } while (0) +static int modules; + static int full_paths; static unsigned long page_size; @@ -188,7 +190,7 @@ static int load_kernel(void) if (!kernel_dso) return -1; - err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, 0); + err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, modules); if (err <= 0) { dso__delete(kernel_dso); kernel_dso = NULL; @@ -648,6 +650,9 @@ sort__sym_print(FILE *fp, struct hist_entry *self) ret += fprintf(fp, "[%c] %s", self->dso == kernel_dso ? 'k' : self->dso == hypervisor_dso ? 'h' : '.', self->sym->name); + + if (self->sym->module) + ret += fprintf(fp, "\t[%s]", self->sym->module->name); } else { ret += fprintf(fp, "%#016llx", (u64)self->ip); } @@ -1710,6 +1715,8 @@ static const struct option options[] = { OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), OPT_STRING('k', "vmlinux", &vmlinux, "file", "vmlinux pathname"), + OPT_BOOLEAN('m', "modules", &modules, + "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_STRING('s', "sort", &sort_order, "key[,key2...]", "sort by key(s): pid, comm, dso, symbol, parent"), OPT_BOOLEAN('P', "full-paths", &full_paths, diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 9bb25fc..aa044ea 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -66,6 +66,7 @@ static unsigned int page_size; static unsigned int mmap_pages = 16; static int freq = 0; static int verbose = 0; +static char *vmlinux = NULL; static char *sym_filter; static unsigned long filter_start; @@ -265,7 +266,10 @@ static void print_sym_table(void) printf("%9.1f %10ld - ", syme->weight, syme->snap_count); color_fprintf(stdout, color, "%4.1f%%", pcnt); - printf(" - %016llx : %s\n", sym->start, sym->name); + printf(" - %016llx : %s", sym->start, sym->name); + if (sym->module) + printf("\t[%s]", sym->module->name); + printf("\n"); } } @@ -359,12 +363,13 @@ static int parse_symbols(void) { struct rb_node *node; struct symbol *sym; + int modules = vmlinux ? 1 : 0; kernel_dso = dso__new("[kernel]", sizeof(struct sym_entry)); if (kernel_dso == NULL) return -1; - if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1, 0) <= 0) + if (dso__load_kernel(kernel_dso, vmlinux, symbol_filter, verbose, modules) <= 0) goto out_delete_dso; node = rb_first(&kernel_dso->syms); @@ -680,6 +685,7 @@ static const struct option options[] = { "system-wide collection from all CPUs"), OPT_INTEGER('C', "CPU", &profile_cpu, "CPU to profile on"), + OPT_STRING('k', "vmlinux", &vmlinux, "file", "vmlinux pathname"), OPT_INTEGER('m', "mmap-pages", &mmap_pages, "number of mmap data pages"), OPT_INTEGER('r', "realtime", &realtime_prio, @@ -709,6 +715,8 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) { int counter; + symbol__init(); + page_size = sysconf(_SC_PAGE_SIZE); argc = parse_options(argc, argv, options, top_usage, 0); -- cgit v0.10.2 From 14f4654cbd531d48651e005cf05907c14bddb193 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 17:58:19 +0200 Subject: perf_counter tools: Create new chain_for_each_child() iterator Iterating through children of a node in the callchain tree shows something that may be quite confusing at a first glance. The head is the children field of the parent and the list nodes are in the brothers field of the children. This is because the childs are linked to the parent as a list of "brothers" using the "children" list of the parent as a head: --------------- | Parent (head) |------------------------------------- --------------- | | | children | | | ----------- ----------- | | 1st child |---brother---| 2nd child |---brother----- ----------- ----------- This makes the following strange pattern often occuring: list_for_each_entry(child, &parent->children, brothers) { // do something with children } Abstract it to chain_for_each_child() to factorize and simplify this pattern. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246550301-8954-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 3dceabd..3c4a91f 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -16,6 +16,9 @@ #include "callchain.h" +#define chain_for_each_child(child, parent) \ + list_for_each_entry(child, &parent->children, brothers) + static void rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) @@ -46,7 +49,7 @@ void sort_chain_to_rbtree(struct rb_root *rb_root, struct callchain_node *node) { struct callchain_node *child; - list_for_each_entry(child, &node->children, brothers) + chain_for_each_child(child, node) sort_chain_to_rbtree(rb_root, child); if (node->hit) @@ -77,7 +80,7 @@ create_child(struct callchain_node *parent, bool inherit_children) list_splice(&parent->children, &new->children); INIT_LIST_HEAD(&parent->children); - list_for_each_entry(next, &new->children, brothers) + chain_for_each_child(next, new) next->parent = new; } list_add_tail(&new->brothers, &parent->children); @@ -173,7 +176,7 @@ __append_chain_children(struct callchain_node *root, struct ip_callchain *chain, struct callchain_node *rnode; /* lookup in childrens */ - list_for_each_entry(rnode, &root->children, brothers) { + chain_for_each_child(rnode, root) { unsigned int ret = __append_chain(rnode, chain, start, syms); if (!ret) -- cgit v0.10.2 From 5a4b181721375700124513cdd9f97056e1c66675 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 17:58:20 +0200 Subject: perf_counter tools: Add new OPT_CALLBACK_DEFAULT option There is no predefined macro to create an option that can have a custom value or a default one if none is given. This patch provides a new helper OPT_CALLBACK_DEFAULT() which defines such kind of option. For example, considering an option -c, we want to get the default value in the following cases: perf command -c -d perf command -d -c And the foo value when it's given: perf command -c foo -d perf command -d -c foo That's also why PARSE_OPT_LASTARG_DEFAULT is extended here to support default values whatever the position of the option, not only in the end. Should it now be renamed to PARSE_OPT_ARG_DEFAULT ? Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo Cc: git@vger.kernel.org LKML-Reference: <1246550301-8954-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/parse-options.c b/tools/perf/util/parse-options.c index 9a897b7..1bf6719 100644 --- a/tools/perf/util/parse-options.c +++ b/tools/perf/util/parse-options.c @@ -20,7 +20,8 @@ static int get_arg(struct parse_opt_ctx_t *p, const struct option *opt, if (p->opt) { *arg = p->opt; p->opt = NULL; - } else if (p->argc == 1 && (opt->flags & PARSE_OPT_LASTARG_DEFAULT)) { + } else if ((opt->flags & PARSE_OPT_LASTARG_DEFAULT) && (p->argc == 1 || + **(p->argv + 1) == '-')) { *arg = (const char *)opt->defval; } else if (p->argc > 1) { p->argc--; diff --git a/tools/perf/util/parse-options.h b/tools/perf/util/parse-options.h index 15c8aba..8aa3464 100644 --- a/tools/perf/util/parse-options.h +++ b/tools/perf/util/parse-options.h @@ -104,6 +104,8 @@ struct option { { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), .value = (v), .argh = "time", .help = (h), .callback = parse_opt_approxidate_cb } #define OPT_CALLBACK(s, l, v, a, h, f) \ { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), .value = (v), (a), .help = (h), .callback = (f) } +#define OPT_CALLBACK_DEFAULT(s, l, v, a, h, f, d) \ + { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), .value = (v), (a), .help = (h), .callback = (f), .defval = (intptr_t)d, .flags = PARSE_OPT_LASTARG_DEFAULT } /* parse_options() will filter out the processed options and leave the * non-option argments in argv[]. -- cgit v0.10.2 From 4eb3e4788b8a5e220a0aeb590f88c28850726ebe Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 17:58:21 +0200 Subject: perf report: Add support for callchain graph output Currently, the printing of callchains is done in a single vertical level, this is the "flat" mode: 8.25% [k] copy_user_generic_string 4.19% copy_user_generic_string generic_file_aio_read do_sync_read vfs_read sys_pread64 system_call_fastpath pread64 This patch introduces a new "graph" mode which provides a hierarchical output of factorized paths recursively sorted: 8.25% [k] copy_user_generic_string | |--4.31%-- generic_file_aio_read | do_sync_read | vfs_read | | | |--4.19%-- sys_pread64 | | system_call_fastpath | | pread64 | | | --0.12%-- sys_read | system_call_fastpath | __read | |--3.24%-- generic_file_buffered_write | __generic_file_aio_write_nolock | generic_file_aio_write | do_sync_write | reiserfs_file_write | vfs_write | | | |--3.14%-- sys_pwrite64 | | system_call_fastpath | | __pwrite64 | | | --0.10%-- sys_write [...] The command line has then changed. By providing the -c option, the callchain will output in the flat mode by default. But you can override it: perf report -c graph or perf report -c flat You can also pass the abreviated mode: perf report -c g or perf report -c gra will both make use of the graph mode. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246550301-8954-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index b44476c..0ca4638 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -59,6 +59,7 @@ static regex_t parent_regex; static int exclude_other = 1; static int callchain; +static enum chain_mode callchain_mode; static u64 sample_type; @@ -787,8 +788,103 @@ hist_entry__collapse(struct hist_entry *left, struct hist_entry *right) return cmp; } +static size_t ipchain__fprintf_graph_line(FILE *fp, int depth, int depth_mask) +{ + int i; + size_t ret = 0; + + ret += fprintf(fp, "%s", " "); + + for (i = 0; i < depth; i++) + if (depth_mask & (1 << i)) + ret += fprintf(fp, "| "); + else + ret += fprintf(fp, " "); + + ret += fprintf(fp, "\n"); + + return ret; +} static size_t -callchain__fprintf(FILE *fp, struct callchain_node *self, u64 total_samples) +ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain, int depth, + int depth_mask, int count, u64 total_samples, + int hits) +{ + int i; + size_t ret = 0; + + ret += fprintf(fp, "%s", " "); + for (i = 0; i < depth; i++) { + if (depth_mask & (1 << i)) + ret += fprintf(fp, "|"); + else + ret += fprintf(fp, " "); + if (!count && i == depth - 1) { + double percent; + + percent = hits * 100.0 / total_samples; + ret += fprintf(fp, "--%2.2f%%-- ", percent); + } else + ret += fprintf(fp, "%s", " "); + } + if (chain->sym) + ret += fprintf(fp, "%s\n", chain->sym->name); + else + ret += fprintf(fp, "%p\n", (void *)(long)chain->ip); + + return ret; +} + +static size_t +callchain__fprintf_graph(FILE *fp, struct callchain_node *self, + u64 total_samples, int depth, int depth_mask) +{ + struct rb_node *node, *next; + struct callchain_node *child; + struct callchain_list *chain; + int new_depth_mask = depth_mask; + size_t ret = 0; + int i; + + node = rb_first(&self->rb_root); + while (node) { + child = rb_entry(node, struct callchain_node, rb_node); + + /* + * The depth mask manages the output of pipes that show + * the depth. We don't want to keep the pipes of the current + * level for the last child of this depth + */ + next = rb_next(node); + if (!next) + new_depth_mask &= ~(1 << (depth - 1)); + + /* + * But we keep the older depth mask for the line seperator + * to keep the level link until we reach the last child + */ + ret += ipchain__fprintf_graph_line(fp, depth, depth_mask); + i = 0; + list_for_each_entry(chain, &child->val, list) { + if (chain->ip >= PERF_CONTEXT_MAX) + continue; + ret += ipchain__fprintf_graph(fp, chain, depth, + new_depth_mask, i++, + total_samples, + child->cumul_hit); + } + ret += callchain__fprintf_graph(fp, child, total_samples, + depth + 1, + new_depth_mask | (1 << depth)); + node = next; + } + + return ret; +} + +static size_t +callchain__fprintf_flat(FILE *fp, struct callchain_node *self, + u64 total_samples) { struct callchain_list *chain; size_t ret = 0; @@ -796,7 +892,7 @@ callchain__fprintf(FILE *fp, struct callchain_node *self, u64 total_samples) if (!self) return 0; - ret += callchain__fprintf(fp, self->parent, total_samples); + ret += callchain__fprintf_flat(fp, self->parent, total_samples); list_for_each_entry(chain, &self->val, list) { @@ -826,8 +922,13 @@ hist_entry_callchain__fprintf(FILE *fp, struct hist_entry *self, chain = rb_entry(rb_node, struct callchain_node, rb_node); percent = chain->hit * 100.0 / total_samples; - ret += fprintf(fp, " %6.2f%%\n", percent); - ret += callchain__fprintf(fp, chain, total_samples); + if (callchain_mode == FLAT) { + ret += fprintf(fp, " %6.2f%%\n", percent); + ret += callchain__fprintf_flat(fp, chain, total_samples); + } else if (callchain_mode == GRAPH) { + ret += callchain__fprintf_graph(fp, chain, + total_samples, 1, 1); + } ret += fprintf(fp, "\n"); rb_node = rb_next(rb_node); } @@ -1129,8 +1230,12 @@ static void output__insert_entry(struct hist_entry *he) struct rb_node *parent = NULL; struct hist_entry *iter; - if (callchain) - sort_chain_to_rbtree(&he->sorted_chain, &he->callchain); + if (callchain) { + if (callchain_mode == FLAT) + sort_chain_flat(&he->sorted_chain, &he->callchain); + else if (callchain_mode == GRAPH) + sort_chain_graph(&he->sorted_chain, &he->callchain); + } while (*p != NULL) { parent = *p; @@ -1702,6 +1807,26 @@ done: return rc; } +static int +parse_callchain_opt(const struct option *opt __used, const char *arg, + int unset __used) +{ + callchain = 1; + + if (!arg) + return 0; + + if (!strncmp(arg, "graph", strlen(arg))) + callchain_mode = GRAPH; + + else if (!strncmp(arg, "flat", strlen(arg))) + callchain_mode = FLAT; + else + return -1; + + return 0; +} + static const char * const report_usage[] = { "perf report [] ", NULL @@ -1725,7 +1850,9 @@ static const struct option options[] = { "regex filter to identify parent, see: '--sort parent'"), OPT_BOOLEAN('x', "exclude-other", &exclude_other, "Only display entries with parent-match"), - OPT_BOOLEAN('c', "callchain", &callchain, "Display callchains"), + OPT_CALLBACK_DEFAULT('c', "callchain", NULL, "output_type", + "Display callchains with output_type: flat, graph. " + "Default to flat", &parse_callchain_opt, "flat"), OPT_STRING('d', "dsos", &dso_list_str, "dso[,dso...]", "only consider symbols in these dsos"), OPT_STRING('C', "comms", &comm_list_str, "comm[,comm...]", diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 3c4a91f..a9873aa 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -19,9 +19,9 @@ #define chain_for_each_child(child, parent) \ list_for_each_entry(child, &parent->children, brothers) - static void -rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) +rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, + enum chain_mode mode) { struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; @@ -31,10 +31,22 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) parent = *p; rnode = rb_entry(parent, struct callchain_node, rb_node); - if (rnode->hit < chain->hit) - p = &(*p)->rb_left; - else - p = &(*p)->rb_right; + switch (mode) { + case FLAT: + if (rnode->hit < chain->hit) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + break; + case GRAPH: + if (rnode->cumul_hit < chain->cumul_hit) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + break; + default: + break; + } } rb_link_node(&chain->rb_node, parent, p); @@ -45,15 +57,36 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain) * Once we get every callchains from the stream, we can now * sort them by hit */ -void sort_chain_to_rbtree(struct rb_root *rb_root, struct callchain_node *node) +void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node) { struct callchain_node *child; chain_for_each_child(child, node) - sort_chain_to_rbtree(rb_root, child); + sort_chain_flat(rb_root, child); if (node->hit) - rb_insert_callchain(rb_root, node); + rb_insert_callchain(rb_root, node, FLAT); +} + +static void __sort_chain_graph(struct callchain_node *node) +{ + struct callchain_node *child; + + node->rb_root = RB_ROOT; + node->cumul_hit = node->hit; + + chain_for_each_child(child, node) { + __sort_chain_graph(child); + rb_insert_callchain(&node->rb_root, child, GRAPH); + node->cumul_hit += child->cumul_hit; + } +} + +void +sort_chain_graph(struct rb_root *rb_root, struct callchain_node *chain_root) +{ + __sort_chain_graph(chain_root); + rb_root->rb_node = chain_root->rb_root.rb_node; } /* diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index e9bd5e8..dfa5600 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -6,15 +6,21 @@ #include #include "symbol.h" +enum chain_mode { + FLAT, + GRAPH +}; struct callchain_node { struct callchain_node *parent; struct list_head brothers; struct list_head children; struct list_head val; - struct rb_node rb_node; + struct rb_node rb_node; /* to sort nodes in an rbtree */ + struct rb_root rb_root; /* sorted tree of children */ unsigned int val_nr; u64 hit; + u64 cumul_hit; /* hit + hits of children */ }; struct callchain_list { @@ -32,5 +38,6 @@ static inline void callchain_init(struct callchain_node *node) void append_chain(struct callchain_node *root, struct ip_callchain *chain, struct symbol **syms); -void sort_chain_to_rbtree(struct rb_root *rb_root, struct callchain_node *node); +void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node); +void sort_chain_graph(struct rb_root *rb_root, struct callchain_node *node); #endif -- cgit v0.10.2 From c20ab37ef30f4a874cf27e84c12c73e584e6f5cc Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 20:14:33 +0200 Subject: perf_counter tools: Set the minimum percent for callchains to be displayed Callchains output may become a burden on a trace because even rarely hit site are exposed. This can be too much information. Let the user set a threshold as a minimum percent of hits using the new pattern for the -c option: -c mode,min_percent Example: $ perf report -s sym -c flat,4 8.25% [k] copy_user_generic_string 4.19% copy_user_generic_string generic_file_aio_read do_sync_read vfs_read sys_pread64 system_call_fastpath pread64 5.39% [k] search_by_key 4.63% 0x00000000009e0a 2.36% [k] memcpy_c [...] $ perf report -s sym -c graph,2 8.25% [k] copy_user_generic_string | |--4.31%-- generic_file_aio_read | do_sync_read | vfs_read | | | --4.19%-- sys_pread64 | system_call_fastpath | pread64 | --3.24%-- generic_file_buffered_write __generic_file_aio_write_nolock generic_file_aio_write do_sync_write reiserfs_file_write vfs_write | --3.14%-- sys_pwrite64 system_call_fastpath __pwrite64 5.39% [k] search_by_key | --2.23%-- reiserfs_update_sd_size 4.63% 0x00000000009e0a 2.36% [k] memcpy_c [...] You can also omit it and it will default to 0. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246558475-10624-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 0ca4638..e8c9817 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -60,6 +60,7 @@ static regex_t parent_regex; static int exclude_other = 1; static int callchain; static enum chain_mode callchain_mode; +static double callchain_min_percent = 0.0; static u64 sample_type; @@ -1224,7 +1225,7 @@ static void collapse__resort(void) static struct rb_root output_hists; -static void output__insert_entry(struct hist_entry *he) +static void output__insert_entry(struct hist_entry *he, u64 min_callchain_hits) { struct rb_node **p = &output_hists.rb_node; struct rb_node *parent = NULL; @@ -1232,9 +1233,11 @@ static void output__insert_entry(struct hist_entry *he) if (callchain) { if (callchain_mode == FLAT) - sort_chain_flat(&he->sorted_chain, &he->callchain); + sort_chain_flat(&he->sorted_chain, &he->callchain, + min_callchain_hits); else if (callchain_mode == GRAPH) - sort_chain_graph(&he->sorted_chain, &he->callchain); + sort_chain_graph(&he->sorted_chain, &he->callchain, + min_callchain_hits); } while (*p != NULL) { @@ -1251,11 +1254,14 @@ static void output__insert_entry(struct hist_entry *he) rb_insert_color(&he->rb_node, &output_hists); } -static void output__resort(void) +static void output__resort(u64 total_samples) { struct rb_node *next; struct hist_entry *n; struct rb_root *tree = &hist; + u64 min_callchain_hits; + + min_callchain_hits = total_samples * (callchain_min_percent / 100); if (sort__need_collapse) tree = &collapse_hists; @@ -1267,7 +1273,7 @@ static void output__resort(void) next = rb_next(&n->rb_node); rb_erase(&n->rb_node, tree); - output__insert_entry(n); + output__insert_entry(n, min_callchain_hits); } } @@ -1801,7 +1807,7 @@ done: dsos__fprintf(stdout); collapse__resort(); - output__resort(); + output__resort(total); output__fprintf(stdout, total); return rc; @@ -1811,19 +1817,36 @@ static int parse_callchain_opt(const struct option *opt __used, const char *arg, int unset __used) { + char *tok; + char *endptr; + callchain = 1; if (!arg) return 0; - if (!strncmp(arg, "graph", strlen(arg))) + tok = strtok((char *)arg, ","); + if (!tok) + return -1; + + /* get the output mode */ + if (!strncmp(tok, "graph", strlen(arg))) callchain_mode = GRAPH; - else if (!strncmp(arg, "flat", strlen(arg))) + else if (!strncmp(tok, "flat", strlen(arg))) callchain_mode = FLAT; else return -1; + /* get the min percentage */ + tok = strtok(NULL, ","); + if (!tok) + return 0; + + callchain_min_percent = strtod(tok, &endptr); + if (tok == endptr) + return -1; + return 0; } @@ -1850,9 +1873,9 @@ static const struct option options[] = { "regex filter to identify parent, see: '--sort parent'"), OPT_BOOLEAN('x', "exclude-other", &exclude_other, "Only display entries with parent-match"), - OPT_CALLBACK_DEFAULT('c', "callchain", NULL, "output_type", - "Display callchains with output_type: flat, graph. " - "Default to flat", &parse_callchain_opt, "flat"), + OPT_CALLBACK_DEFAULT('c', "callchain", NULL, "output_type,min_percent", + "Display callchains using output_type and min percent threshold. " + "Default: flat,0", &parse_callchain_opt, "flat,100"), OPT_STRING('d', "dsos", &dso_list_str, "dso[,dso...]", "only consider symbols in these dsos"), OPT_STRING('C', "comms", &comm_list_str, "comm[,comm...]", diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index a9873aa..c9900fe 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -57,18 +57,19 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, * Once we get every callchains from the stream, we can now * sort them by hit */ -void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node) +void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, + u64 min_hit) { struct callchain_node *child; chain_for_each_child(child, node) - sort_chain_flat(rb_root, child); + sort_chain_flat(rb_root, child, min_hit); - if (node->hit) + if (node->hit && node->hit >= min_hit) rb_insert_callchain(rb_root, node, FLAT); } -static void __sort_chain_graph(struct callchain_node *node) +static void __sort_chain_graph(struct callchain_node *node, u64 min_hit) { struct callchain_node *child; @@ -76,16 +77,18 @@ static void __sort_chain_graph(struct callchain_node *node) node->cumul_hit = node->hit; chain_for_each_child(child, node) { - __sort_chain_graph(child); - rb_insert_callchain(&node->rb_root, child, GRAPH); + __sort_chain_graph(child, min_hit); + if (child->cumul_hit >= min_hit) + rb_insert_callchain(&node->rb_root, child, GRAPH); node->cumul_hit += child->cumul_hit; } } void -sort_chain_graph(struct rb_root *rb_root, struct callchain_node *chain_root) +sort_chain_graph(struct rb_root *rb_root, struct callchain_node *chain_root, + u64 min_hit) { - __sort_chain_graph(chain_root); + __sort_chain_graph(chain_root, min_hit); rb_root->rb_node = chain_root->rb_root.rb_node; } diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index dfa5600..f3e4776 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -38,6 +38,8 @@ static inline void callchain_init(struct callchain_node *node) void append_chain(struct callchain_node *root, struct ip_callchain *chain, struct symbol **syms); -void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node); -void sort_chain_graph(struct rb_root *rb_root, struct callchain_node *node); +void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, + u64 min_hit); +void sort_chain_graph(struct rb_root *rb_root, struct callchain_node *node, + u64 min_hit); #endif -- cgit v0.10.2 From 1e11fd82d247e4e48a1d6c49402214434538d3fd Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 20:14:34 +0200 Subject: perf_counter tools: Provide helper to print percents color Among perf annotate, perf report and perf top, we can find the common colored printing of percents according to the following rules: High overhead = > 5%, colored in red Mid overhead = > 0.5%, colored in green Low overhead = < 0.5%, default color Factorize these multiple checks in a single function named percent_color_fprintf() and also provide a get_percent_color() for sites which print percentages and other things at the same time. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246558475-10624-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 08ea6c5..5f9eefe 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -25,10 +25,6 @@ #define SHOW_USER 2 #define SHOW_HV 4 -#define MIN_GREEN 0.5 -#define MIN_RED 5.0 - - static char const *input_name = "perf.data"; static char *vmlinux = "vmlinux"; @@ -1047,24 +1043,6 @@ process_event(event_t *event, unsigned long offset, unsigned long head) return 0; } -static char *get_color(double percent) -{ - char *color = PERF_COLOR_NORMAL; - - /* - * We color high-overhead entries in red, mid-overhead - * entries in green - and keep the low overhead places - * normal: - */ - if (percent >= MIN_RED) - color = PERF_COLOR_RED; - else { - if (percent > MIN_GREEN) - color = PERF_COLOR_GREEN; - } - return color; -} - static int parse_line(FILE *file, struct symbol *sym, u64 start, u64 len) { @@ -1126,7 +1104,7 @@ parse_line(FILE *file, struct symbol *sym, u64 start, u64 len) } else if (sym->hist_sum) percent = 100.0 * hits / sym->hist_sum; - color = get_color(percent); + color = get_percent_color(percent); /* * Also color the filename and line if needed, with @@ -1262,7 +1240,7 @@ static void print_summary(char *filename) sym_ext = rb_entry(node, struct sym_ext, node); percent = sym_ext->percent; - color = get_color(percent); + color = get_percent_color(percent); path = sym_ext->path; color_fprintf(stdout, color, " %7.2f %s", percent, path); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index e8c9817..c9dbe33 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -947,25 +947,10 @@ hist_entry__fprintf(FILE *fp, struct hist_entry *self, u64 total_samples) if (exclude_other && !self->parent) return 0; - if (total_samples) { - double percent = self->count * 100.0 / total_samples; - char *color = PERF_COLOR_NORMAL; - - /* - * We color high-overhead entries in red, mid-overhead - * entries in green - and keep the low overhead places - * normal: - */ - if (percent >= 5.0) { - color = PERF_COLOR_RED; - } else { - if (percent >= 0.5) - color = PERF_COLOR_GREEN; - } - - ret = color_fprintf(fp, color, " %6.2f%%", + if (total_samples) + ret = percent_color_fprintf(fp, " %6.2f%%", (self->count * 100.0) / total_samples); - } else + else ret = fprintf(fp, "%12Ld ", self->count); list_for_each_entry(se, &hist_entry__sort_list, list) { diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index aa044ea..95d5c0a 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -239,7 +239,6 @@ static void print_sym_table(void) for (nd = rb_first(&tmp); nd; nd = rb_next(nd)) { struct sym_entry *syme = rb_entry(nd, struct sym_entry, rb_node); struct symbol *sym = (struct symbol *)(syme + 1); - char *color = PERF_COLOR_NORMAL; double pcnt; if (++printed > print_entries || syme->snap_count < count_filter) @@ -248,24 +247,12 @@ static void print_sym_table(void) pcnt = 100.0 - (100.0 * ((sum_ksamples - syme->snap_count) / sum_ksamples)); - /* - * We color high-overhead entries in red, mid-overhead - * entries in green - and keep the low overhead places - * normal: - */ - if (pcnt >= 5.0) { - color = PERF_COLOR_RED; - } else { - if (pcnt >= 0.5) - color = PERF_COLOR_GREEN; - } - if (nr_counters == 1) printf("%20.2f - ", syme->weight); else printf("%9.1f %10ld - ", syme->weight, syme->snap_count); - color_fprintf(stdout, color, "%4.1f%%", pcnt); + percent_color_fprintf(stdout, "%4.1f%%", pcnt); printf(" - %016llx : %s", sym->start, sym->name); if (sym->module) printf("\t[%s]", sym->module->name); diff --git a/tools/perf/util/color.c b/tools/perf/util/color.c index 26f8231..90a044d 100644 --- a/tools/perf/util/color.c +++ b/tools/perf/util/color.c @@ -242,4 +242,31 @@ int color_fwrite_lines(FILE *fp, const char *color, return 0; } +char *get_percent_color(double percent) +{ + char *color = PERF_COLOR_NORMAL; + + /* + * We color high-overhead entries in red, mid-overhead + * entries in green - and keep the low overhead places + * normal: + */ + if (percent >= MIN_RED) + color = PERF_COLOR_RED; + else { + if (percent > MIN_GREEN) + color = PERF_COLOR_GREEN; + } + return color; +} +int percent_color_fprintf(FILE *fp, const char *fmt, double percent) +{ + int r; + char *color; + + color = get_percent_color(percent); + r = color_fprintf(fp, color, fmt, percent); + + return r; +} diff --git a/tools/perf/util/color.h b/tools/perf/util/color.h index 5abfd37..706cec5 100644 --- a/tools/perf/util/color.h +++ b/tools/perf/util/color.h @@ -15,6 +15,9 @@ #define PERF_COLOR_CYAN "\033[36m" #define PERF_COLOR_BG_RED "\033[41m" +#define MIN_GREEN 0.5 +#define MIN_RED 5.0 + /* * This variable stores the value of color.ui */ @@ -32,5 +35,7 @@ void color_parse_mem(const char *value, int len, const char *var, char *dst); int color_fprintf(FILE *fp, const char *color, const char *fmt, ...); int color_fprintf_ln(FILE *fp, const char *color, const char *fmt, ...); int color_fwrite_lines(FILE *fp, const char *color, size_t count, const char *buf); +int percent_color_fprintf(FILE *fp, const char *fmt, double percent); +char *get_percent_color(double percent); #endif /* COLOR_H */ -- cgit v0.10.2 From 24b57c6988c5791628c89a8838910991abc9cc1e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 2 Jul 2009 20:14:35 +0200 Subject: perf_counter tools: Display percents of hits in callchain with overhead colors This adds the use of colors to signal at a glance the important overhead thresholds in callchains hit rates. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246558475-10624-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index c9dbe33..283773d 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -824,7 +824,7 @@ ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain, int depth, double percent; percent = hits * 100.0 / total_samples; - ret += fprintf(fp, "--%2.2f%%-- ", percent); + ret += percent_color_fprintf(fp, "--%2.2f%%-- ", percent); } else ret += fprintf(fp, "%s", " "); } @@ -924,7 +924,8 @@ hist_entry_callchain__fprintf(FILE *fp, struct hist_entry *self, chain = rb_entry(rb_node, struct callchain_node, rb_node); percent = chain->hit * 100.0 / total_samples; if (callchain_mode == FLAT) { - ret += fprintf(fp, " %6.2f%%\n", percent); + ret += percent_color_fprintf(fp, " %6.2f%%\n", + percent); ret += callchain__fprintf_flat(fp, chain, total_samples); } else if (callchain_mode == GRAPH) { ret += callchain__fprintf_graph(fp, chain, -- cgit v0.10.2 From 30d7a77dd5a9720430af72f6f62f5156fe073e55 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 2 Jul 2009 21:24:14 -0300 Subject: perf_counter tools: Adjust symbols in ET_EXEC files too Ingo Molnar wrote: > i just bisected a 'perf report' bug that would cause us to not > resolve all user-space symbols in a 'git gc' run to: > > f5812a7a336fb952d819e4427b9a2dce02368e82 is first bad commit > commit f5812a7a336fb952d819e4427b9a2dce02368e82 > Author: Arnaldo Carvalho de Melo > Date: Tue Jun 30 11:43:17 2009 -0300 > > perf_counter tools: Adjust only prelinked symbol's addresses Rename ->prelinked to ->adjust_symbols and making what was done only for prelinked libraries also to ET_EXEC binaries, such as /usr/bin/git: [acme@doppio pahole]$ readelf -h /usr/bin/git | grep Type Type: EXEC (Executable file) [acme@doppio pahole]$ And after installing the 'git-debuginfo' package, I get correct results: [acme@doppio linux-2.6-tip]$ perf report --sort comm,dso,symbol -d /usr/bin/git | head -20 # # (1139614 samples) # # Overhead Command Shared Object Symbol # ........ ................ ......................... ...... # 34.98% git /usr/bin/git [.] send_sideband 33.39% git /usr/bin/git [.] enter_repo 6.81% git /usr/bin/git [.] diff_opt_parse 4.95% git /usr/bin/git [.] is_repository_shallow 3.24% git /usr/bin/git [.] odb_mkstemp 1.39% git /usr/bin/git [.] output 1.34% git /usr/bin/git [.] xmmap 1.25% git /usr/bin/git [.] receive_pack_config 1.16% git /usr/bin/git [.] git_pathdup 0.90% git /usr/bin/git [.] read_object_with_reference 0.86% git /usr/bin/git [.] show_patch_diff 0.85% git /usr/bin/git 0x00000000095e2e 0.69% git /usr/bin/git [.] display [acme@doppio linux-2.6-tip]$ I'll check what are the last cases where we can't resolve symbols, like this 0x00000000095e2e later. And I guess this will fix the problems Mike were seeing too: [acme@doppio linux-2.6-tip]$ readelf -h ../build/perf/vmlinux | grep Type Type: EXEC (Executable file) [acme@doppio linux-2.6-tip]$ Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 98a1311..4683b67 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -555,9 +555,10 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, nr_syms = shdr.sh_size / shdr.sh_entsize; memset(&sym, 0, sizeof(sym)); - self->prelinked = elf_section_by_name(elf, &ehdr, &shdr, - ".gnu.prelink_undo", - NULL) != NULL; + self->adjust_symbols = (ehdr.e_type == ET_EXEC || + elf_section_by_name(elf, &ehdr, &shdr, + ".gnu.prelink_undo", + NULL) != NULL); elf_symtab__for_each_symbol(syms, nr_syms, index, sym) { struct symbol *f; u64 obj_start; @@ -580,7 +581,7 @@ static int dso__load_sym(struct dso *self, int fd, const char *name, section_name = elf_sec__name(&shdr, secstrs); obj_start = sym.st_value; - if (self->prelinked) { + if (self->adjust_symbols) { if (verbose >= 2) printf("adjusting symbol: st_value: %Lx sh_addr: %Lx sh_offset: %Lx\n", (u64)sym.st_value, (u64)shdr.sh_addr, (u64)shdr.sh_offset); @@ -632,7 +633,7 @@ int dso__load(struct dso *self, symbol_filter_t filter, int verbose) if (!name) return -1; - self->prelinked = 0; + self->adjust_symbols = 0; if (strncmp(self->name, "/tmp/perf-", 10) == 0) return dso__load_perf_map(self, filter, verbose); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 4e141a3..7918cff 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -24,7 +24,7 @@ struct dso { struct rb_root syms; struct symbol *(*find_symbol)(struct dso *, u64 ip); unsigned int sym_priv_size; - unsigned char prelinked; + unsigned char adjust_symbols; char name[0]; }; -- cgit v0.10.2 From 029e5b1636d0511ef143af3a20c83c48e44c03f3 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 13:17:28 +0200 Subject: perf report: Annotate variable initialization MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Certain versions of GCC dont see the initialization that is done here: builtin-report.c: In function ‘__cmd_report’: builtin-report.c:1038: warning: ‘syms’ may be used uninitialized in this function So annotate it with a NULL initialization. Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 283773d..fa937f5 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -1035,7 +1035,7 @@ resolve_callchain(struct thread *thread, struct map *map __used, struct ip_callchain *chain, struct hist_entry *entry) { u64 context = PERF_CONTEXT_MAX; - struct symbol **syms; + struct symbol **syms = NULL; unsigned int i; if (callchain) { -- cgit v0.10.2 From bbf2a330d92c5afccfd17592ba9ccd50f41cf748 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 3 Jul 2009 00:08:26 +0200 Subject: x86: atomic64: The atomic64_t data type should be 8 bytes aligned on 32-bit too Locked instructions on two cache lines at once are painful. If atomic64_t uses two cache lines, my test program is 10x slower. The chance for that is significant: 4/32 or 12.5%. Make sure an atomic64_t is 8 bytes aligned. Signed-off-by: Eric Dumazet Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: [ changed it to __aligned(8) as per Andrew's suggestion ] Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index 2503d4e..ae0fbb5 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -250,7 +250,7 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) /* An 64bit atomic type */ typedef struct { - unsigned long long counter; + unsigned long long __aligned(8) counter; } atomic64_t; #define ATOMIC64_INIT(val) { (val) } -- cgit v0.10.2 From b7882b7c65abb00194bdb3d4a22d27d70fcc59ba Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 13:26:39 +0200 Subject: x86: atomic64: Move the 32-bit atomic64_t implementation to a .c file Linus noted that the atomic64_t primitives are all inlines currently which is crazy because these functions have a large register footprint anyway. Move them to a separate file: arch/x86/lib/atomic64_32.c Also, while at it, rename all uses of 'unsigned long long' to the much shorter u64. This makes the appearance of the prototypes a lot nicer - and it also uncovered a few bugs where (yet unused) API variants had 'long' as their return type instead of u64. [ More intrusive changes are not yet done in this patch. ] Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index ae0fbb5..311a43e 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -250,7 +250,7 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) /* An 64bit atomic type */ typedef struct { - unsigned long long __aligned(8) counter; + u64 __aligned(8) counter; } atomic64_t; #define ATOMIC64_INIT(val) { (val) } @@ -264,31 +264,7 @@ typedef struct { */ #define __atomic64_read(ptr) ((ptr)->counter) -static inline unsigned long long -cmpxchg8b(unsigned long long *ptr, unsigned long long old, unsigned long long new) -{ - asm volatile( - - LOCK_PREFIX "cmpxchg8b (%[ptr])\n" - - : "=A" (old) - - : [ptr] "D" (ptr), - "A" (old), - "b" (ll_low(new)), - "c" (ll_high(new)) - - : "memory"); - - return old; -} - -static inline unsigned long long -atomic64_cmpxchg(atomic64_t *ptr, unsigned long long old_val, - unsigned long long new_val) -{ - return cmpxchg8b(&ptr->counter, old_val, new_val); -} +extern u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val); /** * atomic64_xchg - xchg atomic64 variable @@ -298,18 +274,7 @@ atomic64_cmpxchg(atomic64_t *ptr, unsigned long long old_val, * Atomically xchgs the value of @ptr to @new_val and returns * the old value. */ - -static inline unsigned long long -atomic64_xchg(atomic64_t *ptr, unsigned long long new_val) -{ - unsigned long long old_val; - - do { - old_val = atomic_read(ptr); - } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); - - return old_val; -} +extern u64 atomic64_xchg(atomic64_t *ptr, u64 new_val); /** * atomic64_set - set atomic64 variable @@ -318,10 +283,7 @@ atomic64_xchg(atomic64_t *ptr, unsigned long long new_val) * * Atomically sets the value of @ptr to @new_val. */ -static inline void atomic64_set(atomic64_t *ptr, unsigned long long new_val) -{ - atomic64_xchg(ptr, new_val); -} +extern void atomic64_set(atomic64_t *ptr, u64 new_val); /** * atomic64_read - read atomic64 variable @@ -329,16 +291,7 @@ static inline void atomic64_set(atomic64_t *ptr, unsigned long long new_val) * * Atomically reads the value of @ptr and returns it. */ -static inline unsigned long long atomic64_read(atomic64_t *ptr) -{ - unsigned long long curr_val; - - do { - curr_val = __atomic64_read(ptr); - } while (atomic64_cmpxchg(ptr, curr_val, curr_val) != curr_val); - - return curr_val; -} +extern u64 atomic64_read(atomic64_t *ptr); /** * atomic64_add_return - add and return @@ -347,34 +300,14 @@ static inline unsigned long long atomic64_read(atomic64_t *ptr) * * Atomically adds @delta to @ptr and returns @delta + *@ptr */ -static inline unsigned long long -atomic64_add_return(unsigned long long delta, atomic64_t *ptr) -{ - unsigned long long old_val, new_val; - - do { - old_val = atomic_read(ptr); - new_val = old_val + delta; - - } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); - - return new_val; -} - -static inline long atomic64_sub_return(unsigned long long delta, atomic64_t *ptr) -{ - return atomic64_add_return(-delta, ptr); -} +extern u64 atomic64_add_return(u64 delta, atomic64_t *ptr); -static inline long atomic64_inc_return(atomic64_t *ptr) -{ - return atomic64_add_return(1, ptr); -} - -static inline long atomic64_dec_return(atomic64_t *ptr) -{ - return atomic64_sub_return(1, ptr); -} +/* + * Other variants with different arithmetic operators: + */ +extern u64 atomic64_sub_return(u64 delta, atomic64_t *ptr); +extern u64 atomic64_inc_return(atomic64_t *ptr); +extern u64 atomic64_dec_return(atomic64_t *ptr); /** * atomic64_add - add integer to atomic64 variable @@ -383,10 +316,7 @@ static inline long atomic64_dec_return(atomic64_t *ptr) * * Atomically adds @delta to @ptr. */ -static inline void atomic64_add(unsigned long long delta, atomic64_t *ptr) -{ - atomic64_add_return(delta, ptr); -} +extern void atomic64_add(u64 delta, atomic64_t *ptr); /** * atomic64_sub - subtract the atomic64 variable @@ -395,10 +325,7 @@ static inline void atomic64_add(unsigned long long delta, atomic64_t *ptr) * * Atomically subtracts @delta from @ptr. */ -static inline void atomic64_sub(unsigned long long delta, atomic64_t *ptr) -{ - atomic64_add(-delta, ptr); -} +extern void atomic64_sub(u64 delta, atomic64_t *ptr); /** * atomic64_sub_and_test - subtract value from variable and test result @@ -409,13 +336,7 @@ static inline void atomic64_sub(unsigned long long delta, atomic64_t *ptr) * true if the result is zero, or false for all * other cases. */ -static inline int -atomic64_sub_and_test(unsigned long long delta, atomic64_t *ptr) -{ - unsigned long long old_val = atomic64_sub_return(delta, ptr); - - return old_val == 0; -} +extern int atomic64_sub_and_test(u64 delta, atomic64_t *ptr); /** * atomic64_inc - increment atomic64 variable @@ -423,10 +344,7 @@ atomic64_sub_and_test(unsigned long long delta, atomic64_t *ptr) * * Atomically increments @ptr by 1. */ -static inline void atomic64_inc(atomic64_t *ptr) -{ - atomic64_add(1, ptr); -} +extern void atomic64_inc(atomic64_t *ptr); /** * atomic64_dec - decrement atomic64 variable @@ -434,10 +352,7 @@ static inline void atomic64_inc(atomic64_t *ptr) * * Atomically decrements @ptr by 1. */ -static inline void atomic64_dec(atomic64_t *ptr) -{ - atomic64_sub(1, ptr); -} +extern void atomic64_dec(atomic64_t *ptr); /** * atomic64_dec_and_test - decrement and test @@ -447,10 +362,7 @@ static inline void atomic64_dec(atomic64_t *ptr) * returns true if the result is 0, or false for all other * cases. */ -static inline int atomic64_dec_and_test(atomic64_t *ptr) -{ - return atomic64_sub_and_test(1, ptr); -} +extern int atomic64_dec_and_test(atomic64_t *ptr); /** * atomic64_inc_and_test - increment and test @@ -460,10 +372,7 @@ static inline int atomic64_dec_and_test(atomic64_t *ptr) * and returns true if the result is zero, or false for all * other cases. */ -static inline int atomic64_inc_and_test(atomic64_t *ptr) -{ - return atomic64_sub_and_test(-1, ptr); -} +extern int atomic64_inc_and_test(atomic64_t *ptr); /** * atomic64_add_negative - add and test if negative @@ -474,13 +383,7 @@ static inline int atomic64_inc_and_test(atomic64_t *ptr) * if the result is negative, or false when * result is greater than or equal to zero. */ -static inline int -atomic64_add_negative(unsigned long long delta, atomic64_t *ptr) -{ - long long old_val = atomic64_add_return(delta, ptr); - - return old_val < 0; -} +extern int atomic64_add_negative(u64 delta, atomic64_t *ptr); #include #endif /* _ASM_X86_ATOMIC_32_H */ diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index f9d3563..c3c657c 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -10,6 +10,7 @@ lib-y += usercopy_$(BITS).o getuser.o putuser.o lib-y += memcpy_$(BITS).o ifeq ($(CONFIG_X86_32),y) + lib-y += atomic64_32.o lib-y += checksum_32.o lib-y += strstr_32.o lib-y += semaphore_32.o string_32.o diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c new file mode 100644 index 0000000..d21e725 --- /dev/null +++ b/arch/x86/lib/atomic64_32.c @@ -0,0 +1,216 @@ +#include +#include +#include +#include +#include + +static inline u64 cmpxchg8b(u64 *ptr, u64 old, u64 new) +{ + asm volatile( + + LOCK_PREFIX "cmpxchg8b (%[ptr])\n" + + : "=A" (old) + + : [ptr] "D" (ptr), + "A" (old), + "b" (ll_low(new)), + "c" (ll_high(new)) + + : "memory"); + + return old; +} + +u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val) +{ + return cmpxchg8b(&ptr->counter, old_val, new_val); +} + +/** + * atomic64_xchg - xchg atomic64 variable + * @ptr: pointer to type atomic64_t + * @new_val: value to assign + * + * Atomically xchgs the value of @ptr to @new_val and returns + * the old value. + */ + +u64 atomic64_xchg(atomic64_t *ptr, u64 new_val) +{ + u64 old_val; + + do { + old_val = atomic_read(ptr); + } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); + + return old_val; +} + +/** + * atomic64_set - set atomic64 variable + * @ptr: pointer to type atomic64_t + * @new_val: value to assign + * + * Atomically sets the value of @ptr to @new_val. + */ +void atomic64_set(atomic64_t *ptr, u64 new_val) +{ + atomic64_xchg(ptr, new_val); +} + +/** + * atomic64_read - read atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically reads the value of @ptr and returns it. + */ +u64 atomic64_read(atomic64_t *ptr) +{ + u64 curr_val; + + do { + curr_val = __atomic64_read(ptr); + } while (atomic64_cmpxchg(ptr, curr_val, curr_val) != curr_val); + + return curr_val; +} + +/** + * atomic64_add_return - add and return + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr and returns @delta + *@ptr + */ +u64 atomic64_add_return(u64 delta, atomic64_t *ptr) +{ + u64 old_val, new_val; + + do { + old_val = atomic_read(ptr); + new_val = old_val + delta; + + } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); + + return new_val; +} + +u64 atomic64_sub_return(u64 delta, atomic64_t *ptr) +{ + return atomic64_add_return(-delta, ptr); +} + +u64 atomic64_inc_return(atomic64_t *ptr) +{ + return atomic64_add_return(1, ptr); +} + +u64 atomic64_dec_return(atomic64_t *ptr) +{ + return atomic64_sub_return(1, ptr); +} + +/** + * atomic64_add - add integer to atomic64 variable + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr. + */ +void atomic64_add(u64 delta, atomic64_t *ptr) +{ + atomic64_add_return(delta, ptr); +} + +/** + * atomic64_sub - subtract the atomic64 variable + * @delta: integer value to subtract + * @ptr: pointer to type atomic64_t + * + * Atomically subtracts @delta from @ptr. + */ +void atomic64_sub(u64 delta, atomic64_t *ptr) +{ + atomic64_add(-delta, ptr); +} + +/** + * atomic64_sub_and_test - subtract value from variable and test result + * @delta: integer value to subtract + * @ptr: pointer to type atomic64_t + * + * Atomically subtracts @delta from @ptr and returns + * true if the result is zero, or false for all + * other cases. + */ +int atomic64_sub_and_test(u64 delta, atomic64_t *ptr) +{ + u64 old_val = atomic64_sub_return(delta, ptr); + + return old_val == 0; +} + +/** + * atomic64_inc - increment atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically increments @ptr by 1. + */ +void atomic64_inc(atomic64_t *ptr) +{ + atomic64_add(1, ptr); +} + +/** + * atomic64_dec - decrement atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically decrements @ptr by 1. + */ +void atomic64_dec(atomic64_t *ptr) +{ + atomic64_sub(1, ptr); +} + +/** + * atomic64_dec_and_test - decrement and test + * @ptr: pointer to type atomic64_t + * + * Atomically decrements @ptr by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +int atomic64_dec_and_test(atomic64_t *ptr) +{ + return atomic64_sub_and_test(1, ptr); +} + +/** + * atomic64_inc_and_test - increment and test + * @ptr: pointer to type atomic64_t + * + * Atomically increments @ptr by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +int atomic64_inc_and_test(atomic64_t *ptr) +{ + return atomic64_sub_and_test(-1, ptr); +} + +/** + * atomic64_add_negative - add and test if negative + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +int atomic64_add_negative(u64 delta, atomic64_t *ptr) +{ + long long old_val = atomic64_add_return(delta, ptr); + + return old_val < 0; +} -- cgit v0.10.2 From aacf682fd8c66b57383c407eecd9d4a28264ee91 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 3 Jul 2009 12:14:27 +0200 Subject: x86: atomic64: Improve atomic64_read() Linus noticed that the 32-bit version of atomic64_read() was being overly complex with re-reading the value and doing a retry loop over that. Instead we can just rely on cmpxchg8b returning either the new value or returning the current value. We can use any 'old' value, which will be faster as it can be loaded via immediates. Using some value that is not equal to the real value in memory the instruction gets faster. This also has the advantage that the CPU could avoid dirtying the cacheline. Reported-by: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index d21e725..afa5d44 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -67,13 +67,9 @@ void atomic64_set(atomic64_t *ptr, u64 new_val) */ u64 atomic64_read(atomic64_t *ptr) { - u64 curr_val; + u64 old = 1LL << 32; - do { - curr_val = __atomic64_read(ptr); - } while (atomic64_cmpxchg(ptr, curr_val, curr_val) != curr_val); - - return curr_val; + return cmpxchg8b(&ptr->counter, old, old); } /** -- cgit v0.10.2 From 69237f94e65d3d7f539f1adb98ef68685c595004 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 3 Jul 2009 13:26:41 +0200 Subject: x86: atomic64: Improve cmpxchg8b() Rewrite cmpxchg8b() to not use %edi register but a generic "+m" constraint, to increase compiler freedom in code generation and possibly better code. Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index afa5d44..5fc1e2c 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -6,19 +6,14 @@ static inline u64 cmpxchg8b(u64 *ptr, u64 old, u64 new) { - asm volatile( - - LOCK_PREFIX "cmpxchg8b (%[ptr])\n" - - : "=A" (old) - - : [ptr] "D" (ptr), - "A" (old), - "b" (ll_low(new)), - "c" (ll_high(new)) - - : "memory"); + u32 low = new; + u32 high = new >> 32; + asm volatile( + LOCK_PREFIX "cmpxchg8b %1\n" + : "+A" (old), "+m" (*ptr) + : "b" (low), "c" (high) + ); return old; } -- cgit v0.10.2 From 824975ef190e7dcb77718d1cc2cb53769b16d918 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 12:39:07 +0200 Subject: x86: atomic64: Improve atomic64_add_return() Linus noted (based on Eric Dumazet's numbers) that we would probably be better off not trying an atomic_read() in atomic64_add_return() but intead intentionally let the first cmpxchg8b fail - to get a cache-friendly 'give me ownership of this cacheline' transaction. That can then be followed by the real cmpxchg8b which sets the value local to the CPU. Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index 5fc1e2c..6195962 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -76,13 +76,22 @@ u64 atomic64_read(atomic64_t *ptr) */ u64 atomic64_add_return(u64 delta, atomic64_t *ptr) { - u64 old_val, new_val; + /* + * Try first with a (probably incorrect) assumption about + * what we have there. We'll do two loops most likely, + * but we'll get an ownership MESI transaction straight away + * instead of a read transaction followed by a + * flush-for-ownership transaction: + */ + u64 old_val, new_val, real_val = 1ULL << 32; do { - old_val = atomic_read(ptr); + old_val = real_val; new_val = old_val + delta; - } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); + real_val = atomic64_cmpxchg(ptr, old_val, new_val); + + } while (real_val != old_val); return new_val; } -- cgit v0.10.2 From 3ac805d2afd3fa4a07cb5bcf352fd7fa83f28935 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 12:51:19 +0200 Subject: x86: atomic64: Reduce size of functions cmpxchg8b is a huge instruction in terms of register footprint, we almost never want to inline it, not even within the same code module. GCC 4.3 still messes up for two functions, under-judging the true cost of this instruction - so annotate two key functions to reduce the bloat: arch/x86/lib/atomic64_32.o: text data bss dec hex filename 1763 0 0 1763 6e3 atomic64_32.o.before 435 0 0 435 1b3 atomic64_32.o.after Cc: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index 6195962..a910238 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -4,7 +4,7 @@ #include #include -static inline u64 cmpxchg8b(u64 *ptr, u64 old, u64 new) +static noinline u64 cmpxchg8b(u64 *ptr, u64 old, u64 new) { u32 low = new; u32 high = new >> 32; @@ -74,7 +74,7 @@ u64 atomic64_read(atomic64_t *ptr) * * Atomically adds @delta to @ptr and returns @delta + *@ptr */ -u64 atomic64_add_return(u64 delta, atomic64_t *ptr) +noinline u64 atomic64_add_return(u64 delta, atomic64_t *ptr) { /* * Try first with a (probably incorrect) assumption about -- cgit v0.10.2 From 3217120873598533234b6dedda9c371ce30001d0 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 13:06:01 +0200 Subject: x86: atomic64: Make atomic_read() type-safe MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Linus noticed that atomic64_xchg() uses atomic_read(), which happens to work because atomic_read() is a macro so the .counter value gets u64-read on 32-bit too - but this is really bogus and serious bugs are waiting to happen. Change atomic_read() to be a type-safe inline, and this exposes the atomic64 bogosity as well: arch/x86/lib/atomic64_32.c: In function ‘atomic64_xchg’: arch/x86/lib/atomic64_32.c:39: warning: passing argument 1 of ‘atomic_read’ from incompatible pointer type Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index 311a43e..b551bb1 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -19,7 +19,10 @@ * * Atomically reads the value of @v. */ -#define atomic_read(v) ((v)->counter) +static inline int atomic_read(const atomic_t *v) +{ + return v->counter; +} /** * atomic_set - set atomic variable -- cgit v0.10.2 From 199e23780a7e75c63a9e3d1108804e3af450ea3e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 13:02:39 +0200 Subject: x86: atomic64: Fix unclean type use in atomic64_xchg() Linus noticed that atomic64_xchg() uses atomic_read(), which happens to work because atomic_read() is a macro so the .counter value gets u64-read on 32-bit too - but this is really bogus and serious bugs are waiting to happen. Fix atomic64_xchg() to use __atomic64_read() instead. No code changed: arch/x86/lib/atomic64_32.o: text data bss dec hex filename 435 0 0 435 1b3 atomic64_32.o.before 435 0 0 435 1b3 atomic64_32.o.after md5: bd8ab95e69c93518578bfaf0ea3be4d9 atomic64_32.o.before.asm bd8ab95e69c93518578bfaf0ea3be4d9 atomic64_32.o.after.asm Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index a910238..fd28fd3 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -36,7 +36,7 @@ u64 atomic64_xchg(atomic64_t *ptr, u64 new_val) u64 old_val; do { - old_val = atomic_read(ptr); + old_val = __atomic64_read(ptr); } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); return old_val; -- cgit v0.10.2 From 8e049ef054f1cc765f05f13e1396bb9a17c19e66 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Thu, 2 Jul 2009 08:57:12 +1000 Subject: x86: atomic64: Code atomic(64)_read and atomic(64)_set in C not CPP Occasionally we get bugs where atomic_read or atomic_set are used on atomic64_t variables or vice versa. These bugs don't generate warnings on x86 because atomic_read and atomic_set are coded as macros rather than C functions, so we don't get any type-checking on their arguments; similarly for atomic64_read and atomic64_set in 64-bit kernels. This converts them to C functions so that the arguments are type-checked and bugs like this will get caught more easily. It also converts atomic_cmpxchg and atomic_xchg, and atomic64_cmpxchg and atomic64_xchg on 64-bit, so we get type-checking on their arguments too. Compiling a typical 64-bit x86 config, this generates no new warnings, and the vmlinux text is 86 bytes smaller. Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index b551bb1..aa045de 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -31,7 +31,10 @@ static inline int atomic_read(const atomic_t *v) * * Atomically sets the value of @v to @i. */ -#define atomic_set(v, i) (((v)->counter) = (i)) +static inline void atomic_set(atomic_t *v, int i) +{ + v->counter = i; +} /** * atomic_add - add integer to atomic variable @@ -203,8 +206,15 @@ static inline int atomic_sub_return(int i, atomic_t *v) return atomic_add_return(-i, v); } -#define atomic_cmpxchg(v, old, new) (cmpxchg(&((v)->counter), (old), (new))) -#define atomic_xchg(v, new) (xchg(&((v)->counter), (new))) +static inline int atomic_cmpxchg(atomic_t *v, int old, int new) +{ + return cmpxchg(&v->counter, old, new); +} + +static inline int atomic_xchg(atomic_t *v, int new) +{ + return xchg(&v->counter, new); +} /** * atomic_add_unless - add unless the number is already a given value diff --git a/arch/x86/include/asm/atomic_64.h b/arch/x86/include/asm/atomic_64.h index 0d63602..d605dc2 100644 --- a/arch/x86/include/asm/atomic_64.h +++ b/arch/x86/include/asm/atomic_64.h @@ -18,7 +18,10 @@ * * Atomically reads the value of @v. */ -#define atomic_read(v) ((v)->counter) +static inline int atomic_read(const atomic_t *v) +{ + return v->counter; +} /** * atomic_set - set atomic variable @@ -27,7 +30,10 @@ * * Atomically sets the value of @v to @i. */ -#define atomic_set(v, i) (((v)->counter) = (i)) +static inline void atomic_set(atomic_t *v, int i) +{ + v->counter = i; +} /** * atomic_add - add integer to atomic variable @@ -192,7 +198,10 @@ static inline int atomic_sub_return(int i, atomic_t *v) * Atomically reads the value of @v. * Doesn't imply a read memory barrier. */ -#define atomic64_read(v) ((v)->counter) +static inline long atomic64_read(const atomic64_t *v) +{ + return v->counter; +} /** * atomic64_set - set atomic64 variable @@ -201,7 +210,10 @@ static inline int atomic_sub_return(int i, atomic_t *v) * * Atomically sets the value of @v to @i. */ -#define atomic64_set(v, i) (((v)->counter) = (i)) +static inline void atomic64_set(atomic64_t *v, long i) +{ + v->counter = i; +} /** * atomic64_add - add integer to atomic64 variable @@ -355,11 +367,25 @@ static inline long atomic64_sub_return(long i, atomic64_t *v) #define atomic64_inc_return(v) (atomic64_add_return(1, (v))) #define atomic64_dec_return(v) (atomic64_sub_return(1, (v))) -#define atomic64_cmpxchg(v, old, new) (cmpxchg(&((v)->counter), (old), (new))) -#define atomic64_xchg(v, new) (xchg(&((v)->counter), new)) +static inline long atomic64_cmpxchg(atomic64_t *v, long old, long new) +{ + return cmpxchg(&v->counter, old, new); +} + +static inline long atomic64_xchg(atomic64_t *v, long new) +{ + return xchg(&v->counter, new); +} -#define atomic_cmpxchg(v, old, new) (cmpxchg(&((v)->counter), (old), (new))) -#define atomic_xchg(v, new) (xchg(&((v)->counter), (new))) +static inline long atomic_cmpxchg(atomic_t *v, int old, int new) +{ + return cmpxchg(&v->counter, old, new); +} + +static inline long atomic_xchg(atomic_t *v, int new) +{ + return xchg(&v->counter, new); +} /** * atomic_add_unless - add unless the number is a given value -- cgit v0.10.2 From 67d7178f8fc64b7f68d7dd8a1b21dfa0d42c220c Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 3 Jul 2009 13:23:02 +0200 Subject: x86: atomic64: Improve atomic64_read() Optimize atomic64_read() as a special open-coded cmpxchg8b variant. This generates nicer code: arch/x86/lib/atomic64_32.o: text data bss dec hex filename 435 0 0 435 1b3 atomic64_32.o.before 431 0 0 431 1af atomic64_32.o.after md5: bd8ab95e69c93518578bfaf0ea3be4d9 atomic64_32.o.before.asm 2bdfd4bd1f6b7b61b7fc127aef90ce3b atomic64_32.o.after.asm Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index fd28fd3..cd11803 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -62,9 +62,17 @@ void atomic64_set(atomic64_t *ptr, u64 new_val) */ u64 atomic64_read(atomic64_t *ptr) { - u64 old = 1LL << 32; + u64 res; - return cmpxchg8b(&ptr->counter, old, old); + asm volatile( + "mov %%ebx, %%eax\n\t" + "mov %%ecx, %%edx\n\t" + LOCK_PREFIX "cmpxchg8b %1\n" + : "+A" (res) + : "m" (*ptr) + ); + + return res; } /** -- cgit v0.10.2 From 1fde902d52ee13ab9fab155bbae757fdf7daf0c1 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 17:28:57 +0200 Subject: x86: atomic64: Export APIs to modules atomic64_t primitives are used by a handful of drivers, so export the APIs consistently. These were inlined before. Also mark atomic64_32.o a core object, so that the symbols are available even if not linked to core kernel pieces. Cc: Eric Dumazet Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index c3c657c..07c3189 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -10,7 +10,7 @@ lib-y += usercopy_$(BITS).o getuser.o putuser.o lib-y += memcpy_$(BITS).o ifeq ($(CONFIG_X86_32),y) - lib-y += atomic64_32.o + obj-y += atomic64_32.o lib-y += checksum_32.o lib-y += strstr_32.o lib-y += semaphore_32.o string_32.o diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index cd11803..6722a09 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -1,5 +1,7 @@ #include +#include #include + #include #include #include @@ -21,6 +23,7 @@ u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val) { return cmpxchg8b(&ptr->counter, old_val, new_val); } +EXPORT_SYMBOL(atomic64_cmpxchg); /** * atomic64_xchg - xchg atomic64 variable @@ -41,6 +44,7 @@ u64 atomic64_xchg(atomic64_t *ptr, u64 new_val) return old_val; } +EXPORT_SYMBOL(atomic64_xchg); /** * atomic64_set - set atomic64 variable @@ -53,6 +57,7 @@ void atomic64_set(atomic64_t *ptr, u64 new_val) { atomic64_xchg(ptr, new_val); } +EXPORT_SYMBOL(atomic64_read); /** * atomic64_read - read atomic64 variable @@ -74,6 +79,7 @@ u64 atomic64_read(atomic64_t *ptr) return res; } +EXPORT_SYMBOL(atomic64_read); /** * atomic64_add_return - add and return @@ -103,21 +109,25 @@ noinline u64 atomic64_add_return(u64 delta, atomic64_t *ptr) return new_val; } +EXPORT_SYMBOL(atomic64_add_return); u64 atomic64_sub_return(u64 delta, atomic64_t *ptr) { return atomic64_add_return(-delta, ptr); } +EXPORT_SYMBOL(atomic64_sub_return); u64 atomic64_inc_return(atomic64_t *ptr) { return atomic64_add_return(1, ptr); } +EXPORT_SYMBOL(atomic64_inc_return); u64 atomic64_dec_return(atomic64_t *ptr) { return atomic64_sub_return(1, ptr); } +EXPORT_SYMBOL(atomic64_dec_return); /** * atomic64_add - add integer to atomic64 variable @@ -130,6 +140,7 @@ void atomic64_add(u64 delta, atomic64_t *ptr) { atomic64_add_return(delta, ptr); } +EXPORT_SYMBOL(atomic64_add); /** * atomic64_sub - subtract the atomic64 variable @@ -142,6 +153,7 @@ void atomic64_sub(u64 delta, atomic64_t *ptr) { atomic64_add(-delta, ptr); } +EXPORT_SYMBOL(atomic64_sub); /** * atomic64_sub_and_test - subtract value from variable and test result @@ -158,6 +170,7 @@ int atomic64_sub_and_test(u64 delta, atomic64_t *ptr) return old_val == 0; } +EXPORT_SYMBOL(atomic64_sub_and_test); /** * atomic64_inc - increment atomic64 variable @@ -169,6 +182,7 @@ void atomic64_inc(atomic64_t *ptr) { atomic64_add(1, ptr); } +EXPORT_SYMBOL(atomic64_inc); /** * atomic64_dec - decrement atomic64 variable @@ -180,6 +194,7 @@ void atomic64_dec(atomic64_t *ptr) { atomic64_sub(1, ptr); } +EXPORT_SYMBOL(atomic64_dec); /** * atomic64_dec_and_test - decrement and test @@ -193,6 +208,7 @@ int atomic64_dec_and_test(atomic64_t *ptr) { return atomic64_sub_and_test(1, ptr); } +EXPORT_SYMBOL(atomic64_dec_and_test); /** * atomic64_inc_and_test - increment and test @@ -206,6 +222,7 @@ int atomic64_inc_and_test(atomic64_t *ptr) { return atomic64_sub_and_test(-1, ptr); } +EXPORT_SYMBOL(atomic64_inc_and_test); /** * atomic64_add_negative - add and test if negative @@ -222,3 +239,4 @@ int atomic64_add_negative(u64 delta, atomic64_t *ptr) return old_val < 0; } +EXPORT_SYMBOL(atomic64_add_negative); -- cgit v0.10.2 From 3a8d1788b37435baf6c296f4ea8beb4fa4955f44 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 19:56:36 +0200 Subject: x86: atomic64: Improve atomic64_xchg() Remove the read-first logic from atomic64_xchg() and simplify the loop. This function was the last user of __atomic64_read() - remove it. Also, change the 'real_val' assumption from the somewhat quirky 1ULL << 32 value to the (just as arbitrary, but simpler) value of 0. Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index aa045de..d7c8849 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -268,15 +268,6 @@ typedef struct { #define ATOMIC64_INIT(val) { (val) } -/** - * atomic64_read - read atomic64 variable - * @ptr: pointer of type atomic64_t - * - * Atomically reads the value of @v. - * Doesn't imply a read memory barrier. - */ -#define __atomic64_read(ptr) ((ptr)->counter) - extern u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val); /** diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index 6722a09..a804f96 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -33,14 +33,23 @@ EXPORT_SYMBOL(atomic64_cmpxchg); * Atomically xchgs the value of @ptr to @new_val and returns * the old value. */ - u64 atomic64_xchg(atomic64_t *ptr, u64 new_val) { - u64 old_val; + /* + * Try first with a (possibly incorrect) assumption about + * what we have there. We'll do two loops most likely, + * but we'll get an ownership MESI transaction straight away + * instead of a read transaction followed by a + * flush-for-ownership transaction: + */ + u64 old_val, real_val = 0; do { - old_val = __atomic64_read(ptr); - } while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val); + old_val = real_val; + + real_val = atomic64_cmpxchg(ptr, old_val, new_val); + + } while (real_val != old_val); return old_val; } @@ -91,13 +100,13 @@ EXPORT_SYMBOL(atomic64_read); noinline u64 atomic64_add_return(u64 delta, atomic64_t *ptr) { /* - * Try first with a (probably incorrect) assumption about + * Try first with a (possibly incorrect) assumption about * what we have there. We'll do two loops most likely, * but we'll get an ownership MESI transaction straight away * instead of a read transaction followed by a * flush-for-ownership transaction: */ - u64 old_val, new_val, real_val = 1ULL << 32; + u64 old_val, new_val, real_val = 0; do { old_val = real_val; -- cgit v0.10.2 From ddf9a003d32f720805ac30bcc15755e9289073de Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 3 Jul 2009 20:11:30 +0200 Subject: x86: atomic64: Clean up atomic64_sub_and_test() and atomic64_add_negative() Linus noticed that the variable name 'old_val' is confusingly named in these functions - the correct naming is 'new_val'. Reported-by: Linus Torvalds Cc: Eric Dumazet Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index a804f96..1d98c9e 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -175,9 +175,9 @@ EXPORT_SYMBOL(atomic64_sub); */ int atomic64_sub_and_test(u64 delta, atomic64_t *ptr) { - u64 old_val = atomic64_sub_return(delta, ptr); + u64 new_val = atomic64_sub_return(delta, ptr); - return old_val == 0; + return new_val == 0; } EXPORT_SYMBOL(atomic64_sub_and_test); @@ -244,8 +244,8 @@ EXPORT_SYMBOL(atomic64_inc_and_test); */ int atomic64_add_negative(u64 delta, atomic64_t *ptr) { - long long old_val = atomic64_add_return(delta, ptr); + s64 new_val = atomic64_add_return(delta, ptr); - return old_val < 0; + return new_val < 0; } EXPORT_SYMBOL(atomic64_add_negative); -- cgit v0.10.2 From a79f0da80a508448434476b77f9d3d1a469eab67 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 3 Jul 2009 16:50:10 +0200 Subject: x86: atomic64: Inline atomic64_read() again Now atomic64_read() is light weight (no register pressure and small icache), we can inline it again. Also use "=&A" constraint instead of "+A" to avoid warning about unitialized 'res' variable. (gcc had to force 0 in eax/edx) $ size vmlinux.prev vmlinux.after text data bss dec hex filename 4908667 451676 1684868 7045211 6b805b vmlinux.prev 4908651 451676 1684868 7045195 6b804b vmlinux.after Signed-off-by: Eric Dumazet Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: David Howells Cc: Andrew Morton Cc: Arnd Bergmann LKML-Reference: <4A4E1AA2.30002@gmail.com> [ Also fix typo in atomic64_set() export ] Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index d7c8849..dc5a667 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -295,6 +295,28 @@ extern void atomic64_set(atomic64_t *ptr, u64 new_val); * * Atomically reads the value of @ptr and returns it. */ +static inline u64 atomic64_read(atomic64_t *ptr) +{ + u64 res; + + /* + * Note, we inline this atomic64_t primitive because + * it only clobbers EAX/EDX and leaves the others + * untouched. We also (somewhat subtly) rely on the + * fact that cmpxchg8b returns the current 64-bit value + * of the memory location we are touching: + */ + asm volatile( + "mov %%ebx, %%eax\n\t" + "mov %%ecx, %%edx\n\t" + LOCK_PREFIX "cmpxchg8b %1\n" + : "=&A" (res) + : "m" (*ptr) + ); + + return res; +} + extern u64 atomic64_read(atomic64_t *ptr); /** diff --git a/arch/x86/lib/atomic64_32.c b/arch/x86/lib/atomic64_32.c index 1d98c9e..824fa0b 100644 --- a/arch/x86/lib/atomic64_32.c +++ b/arch/x86/lib/atomic64_32.c @@ -66,31 +66,10 @@ void atomic64_set(atomic64_t *ptr, u64 new_val) { atomic64_xchg(ptr, new_val); } -EXPORT_SYMBOL(atomic64_read); +EXPORT_SYMBOL(atomic64_set); /** - * atomic64_read - read atomic64 variable - * @ptr: pointer to type atomic64_t - * - * Atomically reads the value of @ptr and returns it. - */ -u64 atomic64_read(atomic64_t *ptr) -{ - u64 res; - - asm volatile( - "mov %%ebx, %%eax\n\t" - "mov %%ecx, %%edx\n\t" - LOCK_PREFIX "cmpxchg8b %1\n" - : "+A" (res) - : "m" (*ptr) - ); - - return res; -} EXPORT_SYMBOL(atomic64_read); - -/** * atomic64_add_return - add and return * @delta: integer value to add * @ptr: pointer to type atomic64_t -- cgit v0.10.2 From 91b4eaea93f5be95f4477554399680a53aff2343 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 5 Jul 2009 07:39:17 +0200 Subject: perf report: Warn on callchain output request from non-callchain file perf report segfaults while trying to handle callchains from a non callchain data file. Instead of a segfault, print a useful message to the user. Reported-by: Jens Axboe Reported-by: Arnaldo Carvalho de Melo Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Jens Axboe Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246772361-9960-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index fa937f5..9f9575a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -1684,9 +1684,19 @@ static int __cmd_report(void) sample_type = perf_header__sample_type(); - if (sort__has_parent && !(sample_type & PERF_SAMPLE_CALLCHAIN)) { - fprintf(stderr, "selected --sort parent, but no callchain data\n"); - exit(-1); + if (!(sample_type & PERF_SAMPLE_CALLCHAIN)) { + if (sort__has_parent) { + fprintf(stderr, "selected --sort parent, but no" + " callchain data. Did you call" + " perf record without -g?\n"); + exit(-1); + } + if (callchain) { + fprintf(stderr, "selected -c but no callchain data." + " Did you call perf record without" + " -g?\n"); + exit(-1); + } } if (load_kernel() < 0) { -- cgit v0.10.2 From be9038859e56f729cc9d3b070a35fb8829a73696 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 5 Jul 2009 07:39:18 +0200 Subject: perf report: Use a modifiable string for default callchain options If the user doesn't provide options to tune his callchain output (ie: if he uses -c without arguments) then the default value passed in the OPT_CALLBACK_DEFAULT() macro is used. But it's parsed later by strtok() which will replace comma separators to a zero. This may segfault as we are using a read-only string. Use a modifiable one instead, and also fix the "100%" default minimum threshold value by turning it into a 0 (output every callchains) as it was intended in the origin. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Jens Axboe Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246772361-9960-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 9f9575a..3db99fd 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -58,6 +58,8 @@ static char *parent_pattern = default_parent_pattern; static regex_t parent_regex; static int exclude_other = 1; + +static char callchain_default_opt[] = "flat,0"; static int callchain; static enum chain_mode callchain_mode; static double callchain_min_percent = 0.0; @@ -1871,7 +1873,7 @@ static const struct option options[] = { "Only display entries with parent-match"), OPT_CALLBACK_DEFAULT('c', "callchain", NULL, "output_type,min_percent", "Display callchains using output_type and min percent threshold. " - "Default: flat,0", &parse_callchain_opt, "flat,100"), + "Default: flat,0", &parse_callchain_opt, callchain_default_opt), OPT_STRING('d', "dsos", &dso_list_str, "dso[,dso...]", "only consider symbols in these dsos"), OPT_STRING('C', "comms", &comm_list_str, "comm[,comm...]", -- cgit v0.10.2 From 94a8eb028a57854157a936c7e66b09e2559f115a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 5 Jul 2009 07:39:19 +0200 Subject: perf report: Change default callchain parameters The default callchain parameters are set to use the flat mode and never filter any overhead threshold of backtrace. But flat mode is boring compared to graph mode. Also the number of callchains may be very high if none is filtered. Let's change this to set the graph view and a minimum overhead of 0.5% as default parameters. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Jens Axboe Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246772361-9960-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 3db99fd..8bd5865 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -59,10 +59,10 @@ static regex_t parent_regex; static int exclude_other = 1; -static char callchain_default_opt[] = "flat,0"; +static char callchain_default_opt[] = "graph,0.5"; static int callchain; static enum chain_mode callchain_mode; -static double callchain_min_percent = 0.0; +static double callchain_min_percent = 0.5; static u64 sample_type; -- cgit v0.10.2 From e05b876c222178bc6abcfa9f23d8311731691046 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 5 Jul 2009 07:39:20 +0200 Subject: perf_counter tools: callchains: Manage the cumul hits on the fly The cumul hits are the number of hits of every childs of a node plus the hits of the current nodes, required for percentage computing of a branch. Theses numbers are calculated during the sorting of the branches of the callchain tree using a depth first postfix traversal, so that cumulative hits are propagated in the right order. But if we plan to implement percentages relative to the parent and not absolute percentages (relative to the whole overhead), we need to know the cumulative hits of the parent before computing the children because the relative minimum acceptable number of entries (ie: minimum rate against the cumulative hits from the parent) is the basis to filter the children against a given rate. Then we need to handle the cumul hits on the fly to prepare the implementation of relative overhead rates. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Jens Axboe Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246772361-9960-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index c9900fe..5d244af 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -74,13 +74,11 @@ static void __sort_chain_graph(struct callchain_node *node, u64 min_hit) struct callchain_node *child; node->rb_root = RB_ROOT; - node->cumul_hit = node->hit; chain_for_each_child(child, node) { __sort_chain_graph(child, min_hit); if (child->cumul_hit >= min_hit) rb_insert_callchain(&node->rb_root, child, GRAPH); - node->cumul_hit += child->cumul_hit; } } @@ -159,7 +157,7 @@ add_child(struct callchain_node *parent, struct ip_callchain *chain, new = create_child(parent, false); fill_node(new, chain, start, syms); - new->hit = 1; + new->cumul_hit = new->hit = 1; } /* @@ -189,6 +187,7 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, /* split the hits */ new->hit = parent->hit; + new->cumul_hit = parent->cumul_hit; new->val_nr = parent->val_nr - idx_local; parent->val_nr = idx_local; @@ -216,10 +215,13 @@ __append_chain_children(struct callchain_node *root, struct ip_callchain *chain, unsigned int ret = __append_chain(rnode, chain, start, syms); if (!ret) - return; + goto cumul; } /* nothing in children, add to the current node */ add_child(root, chain, start, syms); + +cumul: + root->cumul_hit++; } static int @@ -261,6 +263,8 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, /* we match 100% of the path, increment the hit */ if (i - start == root->val_nr && i == chain->nr) { root->hit++; + root->cumul_hit++; + return 0; } -- cgit v0.10.2 From 805d127d62472f17c7d79baa001a7651afe2fa47 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 5 Jul 2009 07:39:21 +0200 Subject: perf report: Add "Fractal" mode output - support callchains with relative overhead rate The current callchain displays the overhead rates as absolute: relative to the total overhead. This patch provides relative overhead percentage, in which each branch of the callchain tree is a independant instrumentated object. This provides a 'fractal' view of the call-chain profile: each sub-graph looks like a profile in itself - relative to its parent. You can produce such output by using the "fractal" mode that you can abbreviate via f, fr, fra, frac, etc... ./perf report -s sym -c fractal Example: 8.46% [k] copy_user_generic_string | |--52.01%-- generic_file_aio_read | do_sync_read | vfs_read | | | |--97.20%-- sys_pread64 | | system_call_fastpath | | pread64 | | | --2.81%-- sys_read | system_call_fastpath | __read | |--39.85%-- generic_file_buffered_write | __generic_file_aio_write_nolock | generic_file_aio_write | do_sync_write | reiserfs_file_write | vfs_write | | | |--97.05%-- sys_pwrite64 | | system_call_fastpath | | __pwrite64 | | | --2.95%-- sys_write | system_call_fastpath | __write_nocancel [...] Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Anton Blanchard Cc: Jens Axboe Cc: Arnaldo Carvalho de Melo LKML-Reference: <1246772361-9960-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 8bd5865..4e5cc26 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -59,10 +59,15 @@ static regex_t parent_regex; static int exclude_other = 1; -static char callchain_default_opt[] = "graph,0.5"; +static char callchain_default_opt[] = "fractal,0.5"; + static int callchain; -static enum chain_mode callchain_mode; -static double callchain_min_percent = 0.5; + +static +struct callchain_param callchain_param = { + .mode = CHAIN_GRAPH_ABS, + .min_percent = 0.5 +}; static u64 sample_type; @@ -846,9 +851,15 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, struct callchain_node *child; struct callchain_list *chain; int new_depth_mask = depth_mask; + u64 new_total; size_t ret = 0; int i; + if (callchain_param.mode == CHAIN_GRAPH_REL) + new_total = self->cumul_hit; + else + new_total = total_samples; + node = rb_first(&self->rb_root); while (node) { child = rb_entry(node, struct callchain_node, rb_node); @@ -873,10 +884,10 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, continue; ret += ipchain__fprintf_graph(fp, chain, depth, new_depth_mask, i++, - total_samples, + new_total, child->cumul_hit); } - ret += callchain__fprintf_graph(fp, child, total_samples, + ret += callchain__fprintf_graph(fp, child, new_total, depth + 1, new_depth_mask | (1 << depth)); node = next; @@ -925,13 +936,18 @@ hist_entry_callchain__fprintf(FILE *fp, struct hist_entry *self, chain = rb_entry(rb_node, struct callchain_node, rb_node); percent = chain->hit * 100.0 / total_samples; - if (callchain_mode == FLAT) { + switch (callchain_param.mode) { + case CHAIN_FLAT: ret += percent_color_fprintf(fp, " %6.2f%%\n", percent); ret += callchain__fprintf_flat(fp, chain, total_samples); - } else if (callchain_mode == GRAPH) { + break; + case CHAIN_GRAPH_ABS: /* Falldown */ + case CHAIN_GRAPH_REL: ret += callchain__fprintf_graph(fp, chain, total_samples, 1, 1); + default: + break; } ret += fprintf(fp, "\n"); rb_node = rb_next(rb_node); @@ -1219,14 +1235,9 @@ static void output__insert_entry(struct hist_entry *he, u64 min_callchain_hits) struct rb_node *parent = NULL; struct hist_entry *iter; - if (callchain) { - if (callchain_mode == FLAT) - sort_chain_flat(&he->sorted_chain, &he->callchain, - min_callchain_hits); - else if (callchain_mode == GRAPH) - sort_chain_graph(&he->sorted_chain, &he->callchain, - min_callchain_hits); - } + if (callchain) + callchain_param.sort(&he->sorted_chain, &he->callchain, + min_callchain_hits, &callchain_param); while (*p != NULL) { parent = *p; @@ -1249,7 +1260,7 @@ static void output__resort(u64 total_samples) struct rb_root *tree = &hist; u64 min_callchain_hits; - min_callchain_hits = total_samples * (callchain_min_percent / 100); + min_callchain_hits = total_samples * (callchain_param.min_percent / 100); if (sort__need_collapse) tree = &collapse_hists; @@ -1829,22 +1840,31 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, /* get the output mode */ if (!strncmp(tok, "graph", strlen(arg))) - callchain_mode = GRAPH; + callchain_param.mode = CHAIN_GRAPH_ABS; else if (!strncmp(tok, "flat", strlen(arg))) - callchain_mode = FLAT; + callchain_param.mode = CHAIN_FLAT; + + else if (!strncmp(tok, "fractal", strlen(arg))) + callchain_param.mode = CHAIN_GRAPH_REL; + else return -1; /* get the min percentage */ tok = strtok(NULL, ","); if (!tok) - return 0; + goto setup; - callchain_min_percent = strtod(tok, &endptr); + callchain_param.min_percent = strtod(tok, &endptr); if (tok == endptr) return -1; +setup: + if (register_callchain_param(&callchain_param) < 0) { + fprintf(stderr, "Can't register callchain params\n"); + return -1; + } return 0; } diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 5d244af..9d3c814 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -32,13 +32,14 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, rnode = rb_entry(parent, struct callchain_node, rb_node); switch (mode) { - case FLAT: + case CHAIN_FLAT: if (rnode->hit < chain->hit) p = &(*p)->rb_left; else p = &(*p)->rb_right; break; - case GRAPH: + case CHAIN_GRAPH_ABS: /* Falldown */ + case CHAIN_GRAPH_REL: if (rnode->cumul_hit < chain->cumul_hit) p = &(*p)->rb_left; else @@ -53,43 +54,96 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, rb_insert_color(&chain->rb_node, root); } +static void +__sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, + u64 min_hit) +{ + struct callchain_node *child; + + chain_for_each_child(child, node) + __sort_chain_flat(rb_root, child, min_hit); + + if (node->hit && node->hit >= min_hit) + rb_insert_callchain(rb_root, node, CHAIN_FLAT); +} + /* * Once we get every callchains from the stream, we can now * sort them by hit */ -void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, - u64 min_hit) +static void +sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, + u64 min_hit, struct callchain_param *param __used) +{ + __sort_chain_flat(rb_root, node, min_hit); +} + +static void __sort_chain_graph_abs(struct callchain_node *node, + u64 min_hit) { struct callchain_node *child; - chain_for_each_child(child, node) - sort_chain_flat(rb_root, child, min_hit); + node->rb_root = RB_ROOT; - if (node->hit && node->hit >= min_hit) - rb_insert_callchain(rb_root, node, FLAT); + chain_for_each_child(child, node) { + __sort_chain_graph_abs(child, min_hit); + if (child->cumul_hit >= min_hit) + rb_insert_callchain(&node->rb_root, child, + CHAIN_GRAPH_ABS); + } +} + +static void +sort_chain_graph_abs(struct rb_root *rb_root, struct callchain_node *chain_root, + u64 min_hit, struct callchain_param *param __used) +{ + __sort_chain_graph_abs(chain_root, min_hit); + rb_root->rb_node = chain_root->rb_root.rb_node; } -static void __sort_chain_graph(struct callchain_node *node, u64 min_hit) +static void __sort_chain_graph_rel(struct callchain_node *node, + double min_percent) { struct callchain_node *child; + u64 min_hit; node->rb_root = RB_ROOT; + min_hit = node->cumul_hit * min_percent / 100.0; chain_for_each_child(child, node) { - __sort_chain_graph(child, min_hit); + __sort_chain_graph_rel(child, min_percent); if (child->cumul_hit >= min_hit) - rb_insert_callchain(&node->rb_root, child, GRAPH); + rb_insert_callchain(&node->rb_root, child, + CHAIN_GRAPH_REL); } } -void -sort_chain_graph(struct rb_root *rb_root, struct callchain_node *chain_root, - u64 min_hit) +static void +sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_node *chain_root, + u64 min_hit __used, struct callchain_param *param) { - __sort_chain_graph(chain_root, min_hit); + __sort_chain_graph_rel(chain_root, param->min_percent); rb_root->rb_node = chain_root->rb_root.rb_node; } +int register_callchain_param(struct callchain_param *param) +{ + switch (param->mode) { + case CHAIN_GRAPH_ABS: + param->sort = sort_chain_graph_abs; + break; + case CHAIN_GRAPH_REL: + param->sort = sort_chain_graph_rel; + break; + case CHAIN_FLAT: + param->sort = sort_chain_flat; + break; + default: + return -1; + } + return 0; +} + /* * Create a child for a parent. If inherit_children, then the new child * will become the new parent of it's parent children diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index f3e4776..7812122 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -7,8 +7,9 @@ #include "symbol.h" enum chain_mode { - FLAT, - GRAPH + CHAIN_FLAT, + CHAIN_GRAPH_ABS, + CHAIN_GRAPH_REL }; struct callchain_node { @@ -23,6 +24,17 @@ struct callchain_node { u64 cumul_hit; /* hit + hits of children */ }; +struct callchain_param; + +typedef void (*sort_chain_func_t)(struct rb_root *, struct callchain_node *, + u64, struct callchain_param *); + +struct callchain_param { + enum chain_mode mode; + double min_percent; + sort_chain_func_t sort; +}; + struct callchain_list { u64 ip; struct symbol *sym; @@ -36,10 +48,7 @@ static inline void callchain_init(struct callchain_node *node) INIT_LIST_HEAD(&node->val); } +int register_callchain_param(struct callchain_param *param); void append_chain(struct callchain_node *root, struct ip_callchain *chain, struct symbol **syms); -void sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, - u64 min_hit); -void sort_chain_graph(struct rb_root *rb_root, struct callchain_node *node, - u64 min_hit); #endif -- cgit v0.10.2