Refactoring: Privatize ring buffer config header
[lttng-ust.git] / libringbuffer / ringbuffer-config.h
diff --git a/libringbuffer/ringbuffer-config.h b/libringbuffer/ringbuffer-config.h
new file mode 100644 (file)
index 0000000..1052ed6
--- /dev/null
@@ -0,0 +1,241 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright (C) 2010-2021 Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
+ *
+ * Ring buffer configuration header. Note: after declaring the standard inline
+ * functions, clients should also include linux/ringbuffer/api.h.
+ */
+
+#ifndef _LTTNG_RING_BUFFER_CONFIG_H
+#define _LTTNG_RING_BUFFER_CONFIG_H
+
+#include <errno.h>
+#include <lttng/ust-tracer.h>
+#include <stdint.h>
+#include <stddef.h>
+#include <urcu/arch.h>
+#include <string.h>
+#include <lttng/align.h>
+#include <lttng/ust-compiler.h>
+
+struct lttng_ust_lib_ring_buffer;
+struct lttng_ust_lib_ring_buffer_channel;
+struct lttng_ust_lib_ring_buffer_config;
+struct lttng_ust_lib_ring_buffer_ctx;
+struct lttng_ust_shm_handle;
+
+/*
+ * Ring buffer client callbacks. Only used by slow path, never on fast path.
+ * For the fast path, record_header_size(), ring_buffer_clock_read() should be
+ * provided as inline functions too.  These may simply return 0 if not used by
+ * the client.
+ */
+struct lttng_ust_lib_ring_buffer_client_cb {
+       /* Mandatory callbacks */
+
+       /* A static inline version is also required for fast path */
+       uint64_t (*ring_buffer_clock_read) (struct lttng_ust_lib_ring_buffer_channel *chan);
+       size_t (*record_header_size) (const struct lttng_ust_lib_ring_buffer_config *config,
+                                     struct lttng_ust_lib_ring_buffer_channel *chan,
+                                     size_t offset,
+                                     size_t *pre_header_padding,
+                                     struct lttng_ust_lib_ring_buffer_ctx *ctx,
+                                     void *client_ctx);
+
+       /* Slow path only, at subbuffer switch */
+       size_t (*subbuffer_header_size) (void);
+       void (*buffer_begin) (struct lttng_ust_lib_ring_buffer *buf, uint64_t tsc,
+                             unsigned int subbuf_idx,
+                             struct lttng_ust_shm_handle *handle);
+       void (*buffer_end) (struct lttng_ust_lib_ring_buffer *buf, uint64_t tsc,
+                           unsigned int subbuf_idx, unsigned long data_size,
+                           struct lttng_ust_shm_handle *handle);
+
+       /* Optional callbacks (can be set to NULL) */
+
+       /* Called at buffer creation/finalize */
+       int (*buffer_create) (struct lttng_ust_lib_ring_buffer *buf, void *priv,
+                             int cpu, const char *name,
+                             struct lttng_ust_shm_handle *handle);
+       /*
+        * Clients should guarantee that no new reader handle can be opened
+        * after finalize.
+        */
+       void (*buffer_finalize) (struct lttng_ust_lib_ring_buffer *buf,
+                                void *priv, int cpu,
+                                struct lttng_ust_shm_handle *handle);
+
+       /*
+        * Extract header length, payload length and timestamp from event
+        * record. Used by buffer iterators. Timestamp is only used by channel
+        * iterator.
+        */
+       void (*record_get) (const struct lttng_ust_lib_ring_buffer_config *config,
+                           struct lttng_ust_lib_ring_buffer_channel *chan,
+                           struct lttng_ust_lib_ring_buffer *buf,
+                           size_t offset, size_t *header_len,
+                           size_t *payload_len, uint64_t *timestamp,
+                           struct lttng_ust_shm_handle *handle);
+       /*
+        * Offset and size of content size field in client.
+        */
+       void (*content_size_field) (const struct lttng_ust_lib_ring_buffer_config *config,
+                               size_t *offset, size_t *length);
+       void (*packet_size_field) (const struct lttng_ust_lib_ring_buffer_config *config,
+                               size_t *offset, size_t *length);
+};
+
+/*
+ * Ring buffer instance configuration.
+ *
+ * Declare as "static const" within the client object to ensure the inline fast
+ * paths can be optimized.
+ *
+ * alloc/sync pairs:
+ *
+ * RING_BUFFER_ALLOC_PER_CPU and RING_BUFFER_SYNC_PER_CPU :
+ *   Per-cpu buffers with per-cpu synchronization. Tracing must be performed
+ *   with preemption disabled (lib_ring_buffer_get_cpu() and
+ *   lib_ring_buffer_put_cpu()).
+ *
+ * RING_BUFFER_ALLOC_PER_CPU and RING_BUFFER_SYNC_GLOBAL :
+ *   Per-cpu buffer with global synchronization. Tracing can be performed with
+ *   preemption enabled, statistically stays on the local buffers.
+ *
+ * RING_BUFFER_ALLOC_GLOBAL and RING_BUFFER_SYNC_PER_CPU :
+ *   Should only be used for buffers belonging to a single thread or protected
+ *   by mutual exclusion by the client. Note that periodical sub-buffer switch
+ *   should be disabled in this kind of configuration.
+ *
+ * RING_BUFFER_ALLOC_GLOBAL and RING_BUFFER_SYNC_GLOBAL :
+ *   Global shared buffer with global synchronization.
+ *
+ * wakeup:
+ *
+ * RING_BUFFER_WAKEUP_BY_TIMER uses per-cpu deferrable timers to poll the
+ * buffers and wake up readers if data is ready. Mainly useful for tracers which
+ * don't want to call into the wakeup code on the tracing path. Use in
+ * combination with "read_timer_interval" channel_create() argument.
+ *
+ * RING_BUFFER_WAKEUP_BY_WRITER directly wakes up readers when a subbuffer is
+ * ready to read. Lower latencies before the reader is woken up. Mainly suitable
+ * for drivers.
+ *
+ * RING_BUFFER_WAKEUP_NONE does not perform any wakeup whatsoever. The client
+ * has the responsibility to perform wakeups.
+ */
+#define LTTNG_UST_RING_BUFFER_CONFIG_PADDING   20
+
+enum lttng_ust_lib_ring_buffer_alloc_types {
+       RING_BUFFER_ALLOC_PER_CPU,
+       RING_BUFFER_ALLOC_GLOBAL,
+};
+
+enum lttng_ust_lib_ring_buffer_sync_types {
+       RING_BUFFER_SYNC_PER_CPU,       /* Wait-free */
+       RING_BUFFER_SYNC_GLOBAL,        /* Lock-free */
+};
+
+enum lttng_ust_lib_ring_buffer_mode_types {
+       RING_BUFFER_OVERWRITE = 0,      /* Overwrite when buffer full */
+       RING_BUFFER_DISCARD = 1,        /* Discard when buffer full */
+};
+
+enum lttng_ust_lib_ring_buffer_output_types {
+       RING_BUFFER_SPLICE,
+       RING_BUFFER_MMAP,
+       RING_BUFFER_READ,               /* TODO */
+       RING_BUFFER_ITERATOR,
+       RING_BUFFER_NONE,
+};
+
+enum lttng_ust_lib_ring_buffer_backend_types {
+       RING_BUFFER_PAGE,
+       RING_BUFFER_VMAP,               /* TODO */
+       RING_BUFFER_STATIC,             /* TODO */
+};
+
+enum lttng_ust_lib_ring_buffer_oops_types {
+       RING_BUFFER_NO_OOPS_CONSISTENCY,
+       RING_BUFFER_OOPS_CONSISTENCY,
+};
+
+enum lttng_ust_lib_ring_buffer_ipi_types {
+       RING_BUFFER_IPI_BARRIER,
+       RING_BUFFER_NO_IPI_BARRIER,
+};
+
+enum lttng_ust_lib_ring_buffer_wakeup_types {
+       RING_BUFFER_WAKEUP_BY_TIMER,    /* wake up performed by timer */
+       RING_BUFFER_WAKEUP_BY_WRITER,   /*
+                                        * writer wakes up reader,
+                                        * not lock-free
+                                        * (takes spinlock).
+                                        */
+};
+
+struct lttng_ust_lib_ring_buffer_config {
+       enum lttng_ust_lib_ring_buffer_alloc_types alloc;
+       enum lttng_ust_lib_ring_buffer_sync_types sync;
+       enum lttng_ust_lib_ring_buffer_mode_types mode;
+       enum lttng_ust_lib_ring_buffer_output_types output;
+       enum lttng_ust_lib_ring_buffer_backend_types backend;
+       enum lttng_ust_lib_ring_buffer_oops_types oops;
+       enum lttng_ust_lib_ring_buffer_ipi_types ipi;
+       enum lttng_ust_lib_ring_buffer_wakeup_types wakeup;
+       /*
+        * tsc_bits: timestamp bits saved at each record.
+        *   0 and 64 disable the timestamp compression scheme.
+        */
+       unsigned int tsc_bits;
+       struct lttng_ust_lib_ring_buffer_client_cb cb;
+       /*
+        * client_type is used by the consumer process (which is in a
+        * different address space) to lookup the appropriate client
+        * callbacks and update the cb pointers.
+        */
+       int client_type;
+       int _unused1;
+       const struct lttng_ust_lib_ring_buffer_client_cb *cb_ptr;
+       char padding[LTTNG_UST_RING_BUFFER_CONFIG_PADDING];
+};
+
+/*
+ * Reservation flags.
+ *
+ * RING_BUFFER_RFLAG_FULL_TSC
+ *
+ * This flag is passed to record_header_size() and to the primitive used to
+ * write the record header. It indicates that the full 64-bit time value is
+ * needed in the record header. If this flag is not set, the record header needs
+ * only to contain "tsc_bits" bit of time value.
+ *
+ * Reservation flags can be added by the client, starting from
+ * "(RING_BUFFER_FLAGS_END << 0)". It can be used to pass information from
+ * record_header_size() to lib_ring_buffer_write_record_header().
+ */
+#define        RING_BUFFER_RFLAG_FULL_TSC              (1U << 0)
+#define RING_BUFFER_RFLAG_END                  (1U << 1)
+
+/*
+ * lib_ring_buffer_check_config() returns 0 on success.
+ * Used internally to check for valid configurations at channel creation.
+ */
+static inline lttng_ust_notrace
+int lib_ring_buffer_check_config(const struct lttng_ust_lib_ring_buffer_config *config,
+                            unsigned int switch_timer_interval,
+                            unsigned int read_timer_interval);
+static inline
+int lib_ring_buffer_check_config(const struct lttng_ust_lib_ring_buffer_config *config,
+                            unsigned int switch_timer_interval,
+                            unsigned int read_timer_interval)
+{
+       if (config->alloc == RING_BUFFER_ALLOC_GLOBAL
+           && config->sync == RING_BUFFER_SYNC_PER_CPU
+           && switch_timer_interval)
+               return -EINVAL;
+       return 0;
+}
+
+#endif /* _LTTNG_RING_BUFFER_CONFIG_H */
This page took 0.026074 seconds and 4 git commands to generate.