X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=fio.h;h=2c8d9040e732a351d78ea9cfbf586045561941a5;hp=b0307df9b72f33663c1517efb25c82d735c0cfb5;hb=9a3f110001f68968def40797bed5632c2406cdfa;hpb=7b9f733afb91a5c92f44bb6e68860f17ba14f585 diff --git a/fio.h b/fio.h index b0307df9..2c8d9040 100644 --- a/fio.h +++ b/fio.h @@ -17,9 +17,9 @@ struct thread_data; #include "compiler/compiler.h" +#include "thread_options.h" #include "flist.h" #include "fifo.h" -#include "rbtree.h" #include "arch/arch.h" #include "os/os.h" #include "mutex.h" @@ -33,36 +33,27 @@ struct thread_data; #include "options.h" #include "profile.h" #include "time.h" +#include "gettime.h" #include "lib/getopt.h" #include "lib/rand.h" +#include "lib/rbtree.h" #include "server.h" +#include "stat.h" +#include "flow.h" -#ifdef FIO_HAVE_GUASI -#include -#endif - -#ifdef FIO_HAVE_SOLARISAIO +#ifdef CONFIG_SOLARISAIO #include #endif -struct group_run_stats { - uint64_t max_run[2], min_run[2]; - uint64_t max_bw[2], min_bw[2]; - uint64_t io_kb[2]; - uint64_t agg[2]; - uint32_t kb_base; -}; +#ifdef CONFIG_LIBNUMA +#include +#include /* - * What type of allocation to use for io buffers + * "local" is pseudo-policy */ -enum fio_memtype { - MEM_MALLOC = 0, /* ordinary malloc */ - MEM_SHM, /* use shared memory segments */ - MEM_SHMHUGE, /* use shared memory segments with huge pages */ - MEM_MMAP, /* use anonynomous mmap */ - MEM_MMAPHUGE, /* memory mapped huge file */ -}; +#define MPOL_LOCAL MPOL_MAX +#endif /* * offset generator types @@ -72,355 +63,35 @@ enum { RW_SEQ_IDENT, }; -/* - * How many depth levels to log - */ -#define FIO_IO_U_MAP_NR 7 -#define FIO_IO_U_LAT_U_NR 10 -#define FIO_IO_U_LAT_M_NR 12 - -/* - * Aggregate clat samples to report percentile(s) of them. - * - * EXECUTIVE SUMMARY - * - * FIO_IO_U_PLAT_BITS determines the maximum statistical error on the - * value of resulting percentiles. The error will be approximately - * 1/2^(FIO_IO_U_PLAT_BITS+1) of the value. - * - * FIO_IO_U_PLAT_GROUP_NR and FIO_IO_U_PLAT_BITS determine the maximum - * range being tracked for latency samples. The maximum value tracked - * accurately will be 2^(GROUP_NR + PLAT_BITS -1) microseconds. - * - * FIO_IO_U_PLAT_GROUP_NR and FIO_IO_U_PLAT_BITS determine the memory - * requirement of storing those aggregate counts. The memory used will - * be (FIO_IO_U_PLAT_GROUP_NR * 2^FIO_IO_U_PLAT_BITS) * sizeof(int) - * bytes. - * - * FIO_IO_U_PLAT_NR is the total number of buckets. - * - * DETAILS - * - * Suppose the clat varies from 0 to 999 (usec), the straightforward - * method is to keep an array of (999 + 1) buckets, in which a counter - * keeps the count of samples which fall in the bucket, e.g., - * {[0],[1],...,[999]}. However this consumes a huge amount of space, - * and can be avoided if an approximation is acceptable. - * - * One such method is to let the range of the bucket to be greater - * than one. This method has low accuracy when the value is small. For - * example, let the buckets be {[0,99],[100,199],...,[900,999]}, and - * the represented value of each bucket be the mean of the range. Then - * a value 0 has an round-off error of 49.5. To improve on this, we - * use buckets with non-uniform ranges, while bounding the error of - * each bucket within a ratio of the sample value. A simple example - * would be when error_bound = 0.005, buckets are { - * {[0],[1],...,[99]}, {[100,101],[102,103],...,[198,199]},.., - * {[900,909],[910,919]...} }. The total range is partitioned into - * groups with different ranges, then buckets with uniform ranges. An - * upper bound of the error is (range_of_bucket/2)/value_of_bucket - * - * For better efficiency, we implement this using base two. We group - * samples by their Most Significant Bit (MSB), extract the next M bit - * of them as an index within the group, and discard the rest of the - * bits. - * - * E.g., assume a sample 'x' whose MSB is bit n (starting from bit 0), - * and use M bit for indexing - * - * | n | M bits | bit (n-M-1) ... bit 0 | - * - * Because x is at least 2^n, and bit 0 to bit (n-M-1) is at most - * (2^(n-M) - 1), discarding bit 0 to (n-M-1) makes the round-off - * error - * - * 2^(n-M)-1 2^(n-M) 1 - * e <= --------- <= ------- = --- - * 2^n 2^n 2^M - * - * Furthermore, we use "mean" of the range to represent the bucket, - * the error e can be lowered by half to 1 / 2^(M+1). By using M bits - * as the index, each group must contains 2^M buckets. - * - * E.g. Let M (FIO_IO_U_PLAT_BITS) be 6 - * Error bound is 1/2^(6+1) = 0.0078125 (< 1%) - * - * Group MSB #discarded range of #buckets - * error_bits value - * ---------------------------------------------------------------- - * 0* 0~5 0 [0,63] 64 - * 1* 6 0 [64,127] 64 - * 2 7 1 [128,255] 64 - * 3 8 2 [256,511] 64 - * 4 9 3 [512,1023] 64 - * ... ... ... [...,...] ... - * 18 23 17 [8838608,+inf]** 64 - * - * * Special cases: when n < (M-1) or when n == (M-1), in both cases, - * the value cannot be rounded off. Use all bits of the sample as - * index. - * - * ** If a sample's MSB is greater than 23, it will be counted as 23. - */ - -#define FIO_IO_U_PLAT_BITS 6 -#define FIO_IO_U_PLAT_VAL (1 << FIO_IO_U_PLAT_BITS) -#define FIO_IO_U_PLAT_GROUP_NR 19 -#define FIO_IO_U_PLAT_NR (FIO_IO_U_PLAT_GROUP_NR * FIO_IO_U_PLAT_VAL) -#define FIO_IO_U_LIST_MAX_LEN 20 /* The size of the default and user-specified - list of percentiles */ - -#define MAX_PATTERN_SIZE 512 - -struct thread_stat { - char *name; - char *verror; - int32_t error; - int32_t groupid; - uint32_t pid; - char *description; - uint32_t members; - - /* - * bandwidth and latency stats - */ - struct io_stat clat_stat[2]; /* completion latency */ - struct io_stat slat_stat[2]; /* submission latency */ - struct io_stat lat_stat[2]; /* total latency */ - struct io_stat bw_stat[2]; /* bandwidth stats */ - - uint64_t stat_io_bytes[2]; - struct timeval stat_sample_time[2]; - - /* - * fio system usage accounting - */ - struct rusage ru_start; - struct rusage ru_end; - uint64_t usr_time; - uint64_t sys_time; - uint64_t ctx; - uint64_t minf, majf; - - /* - * IO depth and latency stats - */ - uint64_t clat_percentiles; - double *percentile_list; - - uint32_t io_u_map[FIO_IO_U_MAP_NR]; - uint32_t io_u_submit[FIO_IO_U_MAP_NR]; - uint32_t io_u_complete[FIO_IO_U_MAP_NR]; - uint32_t io_u_lat_u[FIO_IO_U_LAT_U_NR]; - uint32_t io_u_lat_m[FIO_IO_U_LAT_M_NR]; - uint32_t io_u_plat[2][FIO_IO_U_PLAT_NR]; - uint64_t total_io_u[3]; - uint64_t short_io_u[3]; - uint64_t total_submit; - uint64_t total_complete; - - uint64_t io_bytes[2]; - uint64_t runtime[2]; - uint64_t total_run_time; - - /* - * IO Error related stats - */ - uint16_t continue_on_error; - uint64_t total_err_count; - int32_t first_error; - - uint32_t kb_base; -}; - -struct bssplit { - unsigned int bs; - unsigned char perc; +enum { + TD_F_VER_BACKLOG = 1, + TD_F_TRIM_BACKLOG = 2, + TD_F_READ_IOLOG = 4, + TD_F_REFILL_BUFFERS = 8, + TD_F_SCRAMBLE_BUFFERS = 16, + TD_F_VER_NONE = 32, + TD_F_PROFILE_OPS = 64, }; -struct thread_options { - int pad; - char *description; - char *name; - char *directory; - char *filename; - char *opendir; - char *ioengine; - enum td_ddir td_ddir; - unsigned int rw_seq; - unsigned int kb_base; - unsigned int ddir_seq_nr; - long ddir_seq_add; - unsigned int iodepth; - unsigned int iodepth_low; - unsigned int iodepth_batch; - unsigned int iodepth_batch_complete; - - unsigned long long size; - unsigned int size_percent; - unsigned int fill_device; - unsigned long long file_size_low; - unsigned long long file_size_high; - unsigned long long start_offset; - - unsigned int bs[2]; - unsigned int ba[2]; - unsigned int min_bs[2]; - unsigned int max_bs[2]; - struct bssplit *bssplit[2]; - unsigned int bssplit_nr[2]; - - unsigned int nr_files; - unsigned int open_files; - enum file_lock_mode file_lock_mode; - unsigned int lockfile_batch; - - unsigned int odirect; - unsigned int invalidate_cache; - unsigned int create_serialize; - unsigned int create_fsync; - unsigned int create_on_open; - unsigned int end_fsync; - unsigned int pre_read; - unsigned int sync_io; - unsigned int verify; - unsigned int do_verify; - unsigned int verifysort; - unsigned int verify_interval; - unsigned int verify_offset; - char verify_pattern[MAX_PATTERN_SIZE]; - unsigned int verify_pattern_bytes; - unsigned int verify_fatal; - unsigned int verify_dump; - unsigned int verify_async; - unsigned long long verify_backlog; - unsigned int verify_batch; - unsigned int use_thread; - unsigned int unlink; - unsigned int do_disk_util; - unsigned int override_sync; - unsigned int rand_repeatable; - unsigned int use_os_rand; - unsigned int write_lat_log; - unsigned int write_bw_log; - unsigned int norandommap; - unsigned int softrandommap; - unsigned int bs_unaligned; - unsigned int fsync_on_close; - - unsigned int hugepage_size; - unsigned int rw_min_bs; - unsigned int thinktime; - unsigned int thinktime_spin; - unsigned int thinktime_blocks; - unsigned int fsync_blocks; - unsigned int fdatasync_blocks; - unsigned int barrier_blocks; - unsigned long long start_delay; - unsigned long long timeout; - unsigned long long ramp_time; - unsigned int overwrite; - unsigned int bw_avg_time; - unsigned int loops; - unsigned long long zone_size; - unsigned long long zone_skip; - enum fio_memtype mem_type; - unsigned int mem_align; - - unsigned int stonewall; - unsigned int new_group; - unsigned int numjobs; - os_cpu_mask_t cpumask; - unsigned int cpumask_set; - os_cpu_mask_t verify_cpumask; - unsigned int verify_cpumask_set; - unsigned int iolog; - unsigned int rwmixcycle; - unsigned int rwmix[2]; - unsigned int nice; - unsigned int file_service_type; - unsigned int group_reporting; - unsigned int fadvise_hint; - enum fio_fallocate_mode fallocate_mode; - unsigned int zero_buffers; - unsigned int refill_buffers; - unsigned int scramble_buffers; - unsigned int time_based; - unsigned int disable_lat; - unsigned int disable_clat; - unsigned int disable_slat; - unsigned int disable_bw; - unsigned int gtod_reduce; - unsigned int gtod_cpu; - unsigned int gtod_offload; - enum fio_cs clocksource; - unsigned int no_stall; - unsigned int trim_percentage; - unsigned int trim_batch; - unsigned int trim_zero; - unsigned long long trim_backlog; - unsigned int clat_percentiles; - unsigned int overwrite_plist; - double percentile_list[FIO_IO_U_LIST_MAX_LEN]; - - char *read_iolog_file; - char *write_iolog_file; - char *bw_log_file; - char *lat_log_file; - char *replay_redirect; - - /* - * Pre-run and post-run shell - */ - char *exec_prerun; - char *exec_postrun; - - unsigned int rate[2]; - unsigned int ratemin[2]; - unsigned int ratecycle; - unsigned int rate_iops[2]; - unsigned int rate_iops_min[2]; - - char *ioscheduler; - - /* - * CPU "io" cycle burner - */ - unsigned int cpuload; - unsigned int cpucycle; - - /* - * I/O Error handling - */ - unsigned int continue_on_error; - - /* - * Benchmark profile type - */ - char *profile; - - /* - * blkio cgroup support - */ - char *cgroup; - unsigned int cgroup_weight; - unsigned int cgroup_nodelete; - - unsigned int uid; - unsigned int gid; - - unsigned int sync_file_range; - - unsigned int userspace_libaio_reap; +enum { + FIO_RAND_BS_OFF = 0, + FIO_RAND_VER_OFF, + FIO_RAND_MIX_OFF, + FIO_RAND_FILE_OFF, + FIO_RAND_BLOCK_OFF, + FIO_RAND_FILE_SIZE_OFF, + FIO_RAND_TRIM_OFF, + FIO_RAND_BUF_OFF, + FIO_RAND_NR_OFFS, }; -#define FIO_VERROR_SIZE 128 - /* * This describes a single thread/process executing a fio job. */ struct thread_data { struct thread_options o; + unsigned long flags; + void *eo; char verror[FIO_VERROR_SIZE]; pthread_t thread; int thread_number; @@ -431,8 +102,21 @@ struct thread_data { struct io_log *clat_log; struct io_log *lat_log; struct io_log *bw_log; + struct io_log *iops_log; + + uint64_t stat_io_bytes[DDIR_RWDIR_CNT]; + struct timeval bw_sample_time; + + uint64_t stat_io_blocks[DDIR_RWDIR_CNT]; + struct timeval iops_sample_time; + + volatile int update_rusage; + struct fio_mutex *rusage_sem; + struct rusage ru_start; + struct rusage ru_end; struct fio_file **files; + unsigned char *file_locks; unsigned int files_size; unsigned int files_index; unsigned int nr_open_files; @@ -444,6 +128,7 @@ struct thread_data { struct frand_state __next_file_state; }; int error; + int sig; int done; pid_t pid; char *orig_buffer; @@ -463,7 +148,7 @@ struct thread_data { char *sysfs_root; - unsigned long rand_seeds[8]; + unsigned long rand_seeds[FIO_RAND_NR_OFFS]; union { os_random_state_t bsrange_state; @@ -492,10 +177,23 @@ struct thread_data { struct ioengine_ops *io_ops; /* - * Current IO depth and list of free and busy io_u's. + * Queue depth of io_u's that fio MIGHT do */ unsigned int cur_depth; + + /* + * io_u's about to be committed + */ unsigned int io_u_queued; + + /* + * io_u's submitted but not completed yet + */ + unsigned int io_u_in_flight; + + /* + * List of free and busy io_u's + */ struct flist_head io_u_freelist; struct flist_head io_u_busylist; struct flist_head io_u_requeues; @@ -514,21 +212,22 @@ struct thread_data { /* * Rate state */ - unsigned long rate_nsec_cycle[2]; - long rate_pending_usleep[2]; - unsigned long rate_bytes[2]; - unsigned long rate_blocks[2]; - struct timeval lastrate[2]; - - unsigned long long total_io_size; - unsigned long long fill_device_size; - - unsigned long io_issues[2]; - unsigned long long io_blocks[2]; - unsigned long long io_bytes[2]; - unsigned long long io_skip_bytes; - unsigned long long this_io_bytes[2]; - unsigned long long zone_bytes; + uint64_t rate_bps[DDIR_RWDIR_CNT]; + long rate_pending_usleep[DDIR_RWDIR_CNT]; + unsigned long rate_bytes[DDIR_RWDIR_CNT]; + unsigned long rate_blocks[DDIR_RWDIR_CNT]; + struct timeval lastrate[DDIR_RWDIR_CNT]; + + uint64_t total_io_size; + uint64_t fill_device_size; + + unsigned long io_issues[DDIR_RWDIR_CNT]; + uint64_t io_blocks[DDIR_RWDIR_CNT]; + uint64_t this_io_blocks[DDIR_RWDIR_CNT]; + uint64_t io_bytes[DDIR_RWDIR_CNT]; + uint64_t io_skip_bytes; + uint64_t this_io_bytes[DDIR_RWDIR_CNT]; + uint64_t zone_bytes; struct fio_mutex *mutex; /* @@ -577,6 +276,8 @@ struct thread_data { struct flist_head trim_list; unsigned long trim_entries; + struct flist_head next_rand_list; + /* * for fileservice, how often to switch to a new file */ @@ -600,11 +301,15 @@ struct thread_data { unsigned int total_err_count; int first_error; + struct fio_flow *flow; + /* * Can be overloaded by profiles */ struct prof_io_ops prof_io_ops; void *prof_data; + + void *pinned_mem; }; /* @@ -623,7 +328,7 @@ enum { int e = (err); \ (td)->error = e; \ if (!(td)->first_error) \ - snprintf(td->verror, sizeof(td->verror) - 1, "file:%s:%d, func=%s, error=%s", __FILE__, __LINE__, (func), (msg)); \ + snprintf(td->verror, sizeof(td->verror), "file:%s:%d, func=%s, error=%s", __FILE__, __LINE__, (func), (msg)); \ } while (0) @@ -634,27 +339,33 @@ enum { #define td_vmsg(td, err, msg, func) \ __td_verror((td), (err), (msg), (func)) +#define __fio_stringify_1(x) #x +#define __fio_stringify(x) __fio_stringify_1(x) + extern int exitall_on_terminate; -extern int thread_number; -extern int nr_process, nr_thread; +extern unsigned int thread_number; +extern unsigned int stat_number; +extern unsigned int nr_process, nr_thread; extern int shm_id; extern int groupid; -extern int terse_output; +extern int output_format; extern int temp_stall_ts; -extern unsigned long long mlock_size; -extern unsigned long page_mask, page_size; +extern uintptr_t page_mask, page_size; extern int read_only; extern int eta_print; +extern int eta_new_line; extern unsigned long done_secs; extern char *job_section; extern int fio_gtod_offload; extern int fio_gtod_cpu; extern enum fio_cs fio_clock_source; +extern int fio_clock_source_set; extern int warnings_fatal; extern int terse_version; extern int is_backend; extern int nr_clients; extern int log_syslog; +extern const char fio_version_string[]; extern struct thread_data *threads; @@ -663,14 +374,35 @@ static inline void fio_ro_check(struct thread_data *td, struct io_u *io_u) assert(!(io_u->ddir == DDIR_WRITE && !td_write(td))); } -#define BLOCKS_PER_MAP (8 * sizeof(unsigned long)) -#define TO_MAP_BLOCK(f, b) (b) -#define RAND_MAP_IDX(f, b) (TO_MAP_BLOCK(f, b) / BLOCKS_PER_MAP) -#define RAND_MAP_BIT(f, b) (TO_MAP_BLOCK(f, b) & (BLOCKS_PER_MAP - 1)) - #define REAL_MAX_JOBS 2048 -#define td_non_fatal_error(e) ((e) == EIO || (e) == EILSEQ) +static inline enum error_type_bit td_error_type(enum fio_ddir ddir, int err) +{ + if (err == EILSEQ) + return ERROR_TYPE_VERIFY_BIT; + if (ddir == DDIR_READ) + return ERROR_TYPE_READ_BIT; + return ERROR_TYPE_WRITE_BIT; +} + +static int __NON_FATAL_ERR[] = {EIO, EILSEQ}; +static inline int td_non_fatal_error(struct thread_data *td, + enum error_type_bit etype, int err) +{ + int i; + if (!td->o.ignore_error[etype]) { + td->o.ignore_error[etype] = __NON_FATAL_ERR; + td->o.ignore_error_nr[etype] = sizeof(__NON_FATAL_ERR) + / sizeof(int); + } + + if (!(td->o.continue_on_error & (1 << etype))) + return 0; + for (i = 0; i < td->o.ignore_error_nr[etype]; i++) + if (td->o.ignore_error[etype][i] == err) + return 1; + return 0; +} static inline void update_error_count(struct thread_data *td, int err) { @@ -683,8 +415,6 @@ static inline int should_fsync(struct thread_data *td) { if (td->last_was_sync) return 0; - if (td->o.odirect) - return 0; if (td_write(td) || td_rw(td) || td->o.override_sync) return 1; @@ -696,21 +426,27 @@ static inline int should_fsync(struct thread_data *td) */ extern int __must_check parse_options(int, char **); extern int parse_jobs_ini(char *, int, int); -extern int exec_run(void); +extern int parse_cmd_line(int, char **); +extern int fio_backend(void); extern void reset_fio_state(void); +extern void clear_io_state(struct thread_data *); extern int fio_options_parse(struct thread_data *, char **, int); extern void fio_keywords_init(void); extern int fio_cmd_option_parse(struct thread_data *, const char *, char *); +extern int fio_cmd_ioengine_option_parse(struct thread_data *, const char *, char *); extern void fio_fill_default_options(struct thread_data *); extern int fio_show_option_help(const char *); +extern void fio_options_set_ioengine_opts(struct option *long_options, struct thread_data *td); extern void fio_options_dup_and_init(struct option *); -extern void options_mem_dupe(struct thread_data *); -extern void options_mem_free(struct thread_data *); +extern void fio_options_mem_dupe(struct thread_data *); +extern void options_mem_dupe(void *data, struct fio_option *options); extern void td_fill_rand_seeds(struct thread_data *); extern void add_job_opts(const char **); -extern char *num2str(unsigned long, int, int, int); +extern char *num2str(unsigned long, int, int, int, int); +extern int ioengine_load(struct thread_data *); -#define FIO_GETOPT_JOB 0x89988998 +#define FIO_GETOPT_JOB 0x89000000 +#define FIO_GETOPT_IOENGINE 0x98000000 #define FIO_NR_OPTIONS (FIO_MAX_OPTS + 128) /* @@ -730,6 +466,7 @@ enum { TD_CREATED, TD_INITIALIZED, TD_RAMP, + TD_SETTING_UP, TD_RUNNING, TD_PRE_READING, TD_VERIFYING, @@ -739,12 +476,14 @@ enum { }; extern void td_set_runstate(struct thread_data *, int); +#define TERMINATE_ALL (-1) +extern void fio_terminate_threads(int); /* * Memory helpers */ -extern int __must_check fio_pin_memory(void); -extern void fio_unpin_memory(void); +extern int __must_check fio_pin_memory(struct thread_data *); +extern void fio_unpin_memory(struct thread_data *); extern int __must_check allocate_io_mem(struct thread_data *); extern void free_io_mem(struct thread_data *); @@ -811,18 +550,28 @@ static inline int __should_check_rate(struct thread_data *td, } static inline int should_check_rate(struct thread_data *td, - unsigned long *bytes_done) + uint64_t *bytes_done) { int ret = 0; - if (bytes_done[0]) - ret |= __should_check_rate(td, 0); - if (bytes_done[1]) - ret |= __should_check_rate(td, 1); + if (bytes_done[DDIR_READ]) + ret |= __should_check_rate(td, DDIR_READ); + if (bytes_done[DDIR_WRITE]) + ret |= __should_check_rate(td, DDIR_WRITE); + if (bytes_done[DDIR_TRIM]) + ret |= __should_check_rate(td, DDIR_TRIM); return ret; } +static inline unsigned int td_max_bs(struct thread_data *td) +{ + unsigned int max_bs; + + max_bs = max(td->o.max_bs[DDIR_READ], td->o.max_bs[DDIR_WRITE]); + return max(td->o.max_bs[DDIR_TRIM], max_bs); +} + static inline int is_power_of_2(unsigned int val) { return (val != 0 && ((val & (val - 1)) == 0)); @@ -850,4 +599,24 @@ static inline void td_io_u_free_notify(struct thread_data *td) pthread_cond_signal(&td->free_cond); } +extern const char *fio_get_arch_string(int); +extern const char *fio_get_os_string(int); + +enum { + FIO_OUTPUT_TERSE = 0, + FIO_OUTPUT_JSON, + FIO_OUTPUT_NORMAL, +}; + +enum { + FIO_RAND_DIST_RANDOM = 0, + FIO_RAND_DIST_ZIPF, + FIO_RAND_DIST_PARETO, +}; + +enum { + FIO_RAND_GEN_TAUSWORTHE = 0, + FIO_RAND_GEN_LFSR, +}; + #endif