From 119a5d573622ae90ba730d18acfae9bb75d77b9a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 30 Jun 2025 18:04:40 -0400 Subject: ring-buffer: Remove ring_buffer_read_prepare_sync() When the ring buffer was first introduced, reading the non-consuming "trace" file required disabling the writing of the ring buffer. To make sure the writing was fully disabled before iterating the buffer with a non-consuming read, it would set the disable flag of the buffer and then call an RCU synchronization to make sure all the buffers were synchronized. The function ring_buffer_read_start() originally would initialize the iterator and call an RCU synchronization, but this was for each individual per CPU buffer where this would get called many times on a machine with many CPUs before the trace file could be read. The commit 72c9ddfd4c5bf ("ring-buffer: Make non-consuming read less expensive with lots of cpus.") separated ring_buffer_read_start into ring_buffer_read_prepare(), ring_buffer_read_sync() and then ring_buffer_read_start() to allow each of the per CPU buffers to be prepared, call the read_buffer_read_sync() once, and then the ring_buffer_read_start() for each of the CPUs which made things much faster. The commit 1039221cc278 ("ring-buffer: Do not disable recording when there is an iterator") removed the requirement of disabling the recording of the ring buffer in order to iterate it, but it did not remove the synchronization that was happening that was required to wait for all the buffers to have no more writers. It's now OK for the buffers to have writers and no synchronization is needed. Remove the synchronization and put back the interface for the ring buffer iterator back before commit 72c9ddfd4c5bf was applied. Cc: Mathieu Desnoyers Link: https://lore.kernel.org/20250630180440.3eabb514@batman.local.home Reported-by: David Howells Fixes: 1039221cc278 ("ring-buffer: Do not disable recording when there is an iterator") Tested-by: David Howells Reviewed-by: Masami Hiramatsu (Google) Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) (limited to 'kernel/trace/trace.c') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 95ae7c4e5835..7996f26c3f46 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4735,21 +4735,15 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot) if (iter->cpu_file == RING_BUFFER_ALL_CPUS) { for_each_tracing_cpu(cpu) { iter->buffer_iter[cpu] = - ring_buffer_read_prepare(iter->array_buffer->buffer, - cpu, GFP_KERNEL); - } - ring_buffer_read_prepare_sync(); - for_each_tracing_cpu(cpu) { - ring_buffer_read_start(iter->buffer_iter[cpu]); + ring_buffer_read_start(iter->array_buffer->buffer, + cpu, GFP_KERNEL); tracing_iter_reset(iter, cpu); } } else { cpu = iter->cpu_file; iter->buffer_iter[cpu] = - ring_buffer_read_prepare(iter->array_buffer->buffer, - cpu, GFP_KERNEL); - ring_buffer_read_prepare_sync(); - ring_buffer_read_start(iter->buffer_iter[cpu]); + ring_buffer_read_start(iter->array_buffer->buffer, + cpu, GFP_KERNEL); tracing_iter_reset(iter, cpu); } -- cgit v1.2.3 From c897c1e5b19dd4fc32e84fa1ab2065c2507be3a7 Mon Sep 17 00:00:00 2001 From: Nam Cao Date: Thu, 26 Jun 2025 17:19:40 +0200 Subject: tracing: Remove pointless memory barriers Memory barriers are useful to ensure memory accesses from one CPU appear in the original order as seen by other CPUs. Some smp_rmb() and smp_wmb() are used, but they are not ordering multiple memory accesses. Remove them. Cc: Masami Hiramatsu Cc: Mathieu Desnoyers Cc: Gabriele Monaco Link: https://lore.kernel.org/20250626151940.1756398-1-namcao@linutronix.de Signed-off-by: Nam Cao Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'kernel/trace/trace.c') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 95ae7c4e5835..0dff4298fc0e 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -936,7 +936,6 @@ int tracing_is_enabled(void) * return the mirror variable of the state of the ring buffer. * It's a little racy, but we don't really care. */ - smp_rmb(); return !global_trace.buffer_disabled; } @@ -1107,8 +1106,6 @@ void tracer_tracing_on(struct trace_array *tr) * important to be fast than accurate. */ tr->buffer_disabled = 0; - /* Make the flag seen by readers */ - smp_wmb(); } /** @@ -1640,8 +1637,6 @@ void tracer_tracing_off(struct trace_array *tr) * important to be fast than accurate. */ tr->buffer_disabled = 1; - /* Make the flag seen by readers */ - smp_wmb(); } /** @@ -2710,8 +2705,6 @@ void trace_buffered_event_enable(void) static void enable_trace_buffered_event(void *data) { - /* Probably not needed, but do it anyway */ - smp_rmb(); this_cpu_dec(trace_buffered_event_cnt); } -- cgit v1.2.3 From 9ba817fb7c6afd3c86a6d4c3b822924b87ef0348 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 22 Jul 2025 17:08:06 -0400 Subject: tracing: Deprecate auto-mounting tracefs in debugfs In January 2015, tracefs was created to allow access to the tracing infrastructure without needing to compile in debugfs. When tracefs is configured, the directory /sys/kernel/tracing will exist and tooling is expected to use that path to access the tracing infrastructure. To allow backward compatibility, when debugfs is mounted, it would automount tracefs in its "tracing" directory so that tooling that had hard coded /sys/kernel/debug/tracing would still work. It has been over 10 years since the new interface was introduced, and all tooling should now be using it. Start the process of deprecating the old path so that it doesn't need to be maintained anymore. A new config is added to allow distributions to disable automounting of tracefs on debugfs. If /sys/kernel/debug/tracing is accessed, a pr_warn() will trigger stating: "NOTICE: Automounting of tracing to debugfs is deprecated and will be removed in 2030" Expect to remove this feature in 5 years (2030). Cc: Cc: Masami Hiramatsu Cc: Mathieu Desnoyers Cc: Mark Rutland Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Sebastian Andrzej Siewior Cc: Namhyung Kim Cc: Linus Torvalds Cc: Andrew Morton Cc: Greg Kroah-Hartman Cc: Al Viro Cc: Christian Brauner Cc: Jan Kara Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Ian Rogers Link: https://lore.kernel.org/20250722170806.40c068c6@gandalf.local.home Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'kernel/trace/trace.c') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0dff4298fc0e..06ab5b7a8711 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -6296,7 +6296,7 @@ static bool tracer_options_updated; static void add_tracer_options(struct trace_array *tr, struct tracer *t) { /* Only enable if the directory has been created already. */ - if (!tr->dir) + if (!tr->dir && !(tr->flags & TRACE_ARRAY_FL_GLOBAL)) return; /* Only create trace option files after update_tracer_options finish */ @@ -8977,13 +8977,13 @@ static inline __init int register_snapshot_cmd(void) { return 0; } static struct dentry *tracing_get_dentry(struct trace_array *tr) { - if (WARN_ON(!tr->dir)) - return ERR_PTR(-ENODEV); - /* Top directory uses NULL as the parent */ if (tr->flags & TRACE_ARRAY_FL_GLOBAL) return NULL; + if (WARN_ON(!tr->dir)) + return ERR_PTR(-ENODEV); + /* All sub buffers have a descriptor */ return tr->dir; } @@ -10249,6 +10249,7 @@ init_tracer_tracefs(struct trace_array *tr, struct dentry *d_tracer) ftrace_init_tracefs(tr, d_tracer); } +#ifdef CONFIG_TRACEFS_AUTOMOUNT_DEPRECATED static struct vfsmount *trace_automount(struct dentry *mntpt, void *ingore) { struct vfsmount *mnt; @@ -10270,6 +10271,8 @@ static struct vfsmount *trace_automount(struct dentry *mntpt, void *ingore) if (IS_ERR(fc)) return ERR_CAST(fc); + pr_warn("NOTICE: Automounting of tracing to debugfs is deprecated and will be removed in 2030\n"); + ret = vfs_parse_fs_string(fc, "source", "tracefs", strlen("tracefs")); if (!ret) @@ -10280,6 +10283,7 @@ static struct vfsmount *trace_automount(struct dentry *mntpt, void *ingore) put_fs_context(fc); return mnt; } +#endif /** * tracing_init_dentry - initialize top level trace array @@ -10304,6 +10308,7 @@ int tracing_init_dentry(void) if (WARN_ON(!tracefs_initialized())) return -ENODEV; +#ifdef CONFIG_TRACEFS_AUTOMOUNT_DEPRECATED /* * As there may still be users that expect the tracing * files to exist in debugfs/tracing, we must automount @@ -10312,6 +10317,7 @@ int tracing_init_dentry(void) */ tr->dir = debugfs_create_automount("tracing", NULL, trace_automount, NULL); +#endif return 0; } -- cgit v1.2.3 From 1a967e92bf47cf5170336b88d748117c700edc47 Mon Sep 17 00:00:00 2001 From: "Masami Hiramatsu (Google)" Date: Tue, 29 Jul 2025 14:10:35 +0900 Subject: tracing: Remove "__attribute__()" from the type field of event format With CONFIG_DEBUG_INFO_BTF=y and PAHOLE_HAS_BTF_TAG=y, `__user` is converted to `__attribute__((btf_type_tag("user")))`. In this case, some syscall events have it for __user data, like below; /sys/kernel/tracing # cat events/syscalls/sys_enter_openat/format name: sys_enter_openat ID: 720 format: field:unsigned short common_type; offset:0; size:2; signed:0; field:unsigned char common_flags; offset:2; size:1; signed:0; field:unsigned char common_preempt_count; offset:3; size:1; signed:0; field:int common_pid; offset:4; size:4; signed:1; field:int __syscall_nr; offset:8; size:4; signed:1; field:int dfd; offset:16; size:8; signed:0; field:const char __attribute__((btf_type_tag("user"))) * filename; offset:24; size:8; signed:0; field:int flags; offset:32; size:8; signed:0; field:umode_t mode; offset:40; size:8; signed:0; Then the trace event filter fails to set the string acceptable flag (FILTER_PTR_STRING) to the field and rejects setting string filter; # echo 'filename.ustring ~ "*ftracetest-dir.wbx24v*"' \ >> events/syscalls/sys_enter_openat/filter sh: write error: Invalid argument # cat error_log [ 723.743637] event filter parse error: error: Expecting numeric field Command: filename.ustring ~ "*ftracetest-dir.wbx24v*" Since this __attribute__ makes format parsing complicated and not needed, remove the __attribute__(.*) from the type string. Cc: Mathieu Desnoyers Link: https://lore.kernel.org/175376583493.1688759.12333973498014733551.stgit@mhiramat.tok.corp.google.com Signed-off-by: Masami Hiramatsu (Google) Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 28 ++++++++++++++++++---------- 1 file changed, 18 insertions(+), 10 deletions(-) (limited to 'kernel/trace/trace.c') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 06ab5b7a8711..945a8ecf2c62 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5930,17 +5930,27 @@ static inline void trace_insert_eval_map_file(struct module *mod, struct trace_eval_map **start, int len) { } #endif /* !CONFIG_TRACE_EVAL_MAP_FILE */ -static void trace_insert_eval_map(struct module *mod, - struct trace_eval_map **start, int len) +static void +trace_event_update_with_eval_map(struct module *mod, + struct trace_eval_map **start, + int len) { struct trace_eval_map **map; - if (len <= 0) - return; + /* Always run sanitizer only if btf_type_tag attr exists. */ + if (len <= 0) { + if (!(IS_ENABLED(CONFIG_DEBUG_INFO_BTF) && + IS_ENABLED(CONFIG_PAHOLE_HAS_BTF_TAG) && + __has_attribute(btf_type_tag))) + return; + } map = start; - trace_event_eval_update(map, len); + trace_event_update_all(map, len); + + if (len <= 0) + return; trace_insert_eval_map_file(mod, start, len); } @@ -10334,7 +10344,7 @@ static void __init eval_map_work_func(struct work_struct *work) int len; len = __stop_ftrace_eval_maps - __start_ftrace_eval_maps; - trace_insert_eval_map(NULL, __start_ftrace_eval_maps, len); + trace_event_update_with_eval_map(NULL, __start_ftrace_eval_maps, len); } static int __init trace_eval_init(void) @@ -10387,9 +10397,6 @@ bool module_exists(const char *module) static void trace_module_add_evals(struct module *mod) { - if (!mod->num_trace_evals) - return; - /* * Modules with bad taint do not have events created, do * not bother with enums either. @@ -10397,7 +10404,8 @@ static void trace_module_add_evals(struct module *mod) if (trace_module_has_bad_taint(mod)) return; - trace_insert_eval_map(mod, mod->trace_evals, mod->num_trace_evals); + /* Even if no trace_evals, this need to sanitize field types. */ + trace_event_update_with_eval_map(mod, mod->trace_evals, mod->num_trace_evals); } #ifdef CONFIG_TRACE_EVAL_MAP_FILE -- cgit v1.2.3