*/
#define _GNU_SOURCE
+#include <xseg/xseg.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/syscall.h>
#include <sys/types.h>
#include <pthread.h>
-#include <xseg/xseg.h>
#include <peer.h>
#include <time.h>
#include <sys/util.h>
#include <signal.h>
#include <bench-xseg.h>
+#include <bench-lfsr.h>
+#include <limits.h>
+#include <math.h>
-struct timespec delay = {0, 4000000};
+/*
+ * This macro checks two things:
+ * a) If in-flight requests are less than given iodepth
+ * b) If we have submitted all of the requests
+ * c) If we are not in ping mode
+ * d) If we have been asked to terminate
+ */
+#define CAN_SEND_REQUEST(__p) \
+ ((__p->status->submitted - __p->status->received < __p->iodepth) && \
+ (__p->status->submitted < __p->status->max) && \
+ (GET_FLAG(PING, __p->flags) == PING_MODE_OFF) && \
+ !isTerminate())
+
+#define CAN_VERIFY(__p) \
+ ((GET_FLAG(VERIFY, __p->flags) != VERIFY_NO) && __p->op == X_READ)
+
+#define CAN_PRINT_PROGRESS(__p, __nr) \
+ ((GET_FLAG(PROGRESS, __p->flags) != PROGRESS_NO) && \
+ (GET_FLAG(PING, __p->flags) == PING_MODE_OFF) && \
+ (__p->status->received == __nr))
void custom_peer_usage()
{
fprintf(stderr, "Custom peer options: \n"
" --------------------------------------------\n"
+ "\n"
+ "a) Benchmark options: \n"
+ " --------------------------------------------\n"
+ " -op | None | XSEG operation:\n"
+ " | | [read|write|info|delete]\n"
+ " --pattern | None | I/O pattern [seq|rand]\n"
+ " -rc | None | Request cap\n"
+ " -to | None | Total objects\n"
" -ts | None | Total I/O size\n"
" -os | 4M | Object size\n"
" -bs | 4k | Block size\n"
- " -dp | None | Destination port\n"
+ " -tp | None | Target port\n"
" --iodepth | 1 | Number of in-flight I/O requests\n"
+ " --seed | None | Initialize LFSR and target names\n"
+ "\n"
+ "b) Object naming options: \n"
+ " --------------------------------------------\n"
+ " --prefix | bench | Add a common prefix to all object names\n"
+ " --objname | None | Use only one object with this name\n"
+ "\n"
+ "c) Data verification options: \n"
+ " --------------------------------------------\n"
+ " --verify | no | Verify written requests:\n"
+ " | | [no|meta|full]\n"
+ "\n"
+ "d) Progress report options: \n"
+ " --------------------------------------------\n"
+ " --insanity | sane | Adjust insanity level of benchmark:\n"
+ " | | [sane|eccentric|manic|paranoid]\n"
+ " --progress | both | Show progress of benchmark:\n"
+ " | | [req|io|both|no]\n"
+ " --interval | 5%% | Intervals at which progress is shown\n"
+ "\n"
+ "e) Misc options: \n"
+ " --------------------------------------------\n"
+ " --ping | no | Ping target before starting:\n"
+ " | | [yes|no]\n"
+ "\n"
+ "Additional information:\n"
+ " --------------------------------------------\n"
+ " * The -to and -ts options are mutually exclusive\n"
+ "\n"
+ " * The object name is always not null-terminated and\n"
+ " defaults to the following structure:\n"
+ " <prefix>-<seed>-<object number>\n"
+ "\n"
+ " where:\n"
+ " a. <prefix> is given by user or defaults to 'bench'\n"
+ " b. <seed> is given by user or defaults to a random value.\n"
+ " Its length will be 9 digits, with trailing zeros where\n"
+ " necessary\n"
+ " c. <object number> is out of the user's control. It is\n"
+ " calculated during the benchmark and is a 15-digit\n"
+ " number, allowing a maximum of 1 quadrillion objects\n"
+ "\n"
+ " So, if bench is called with the arguments:\n"
+ " --prefix obj --seed 999\n"
+ "\n"
+ " and <object number> is 9,the resulting object name will\n"
+ " be:\n"
+ " obj-000000999-000000000000009\n"
+ "\n"
+ " * The above object name structure can by bypassed with the\n"
+ " --objname <object name> argument. This implies the\n"
+ " following:\n"
+ "\n"
+ " a. -to option is strictly restricted to 1\n"
+ " b. -ts option defaults to (and can't be larger than)\n"
+ " the object size (-os argument)\n"
+ " c. --prefix is must be unused. If used, it produces an\n"
+ " error to alert the user\n"
+ "\n"
+ " * The progress report is printed by default at intervals of\n"
+ " 5%%.\n"
+ " There are three progress types:\n"
+ "\n"
+ " a. req: it prints the request status so far i.e. how many\n"
+ " requests have been subitted, received, failed etc.\n"
+ " b. io: it prints the bandwidth and IOPS status of the\n"
+ " elapsed 5%% of the benchmark\n"
+ " c. both: it combines the output of <req> and <io>.\n"
+ "\n"
+ " * Interval is commonly a percentage of max requests. This\n"
+ " means that when a user gives:\n"
+ " --interval 33%%\n"
+ "\n"
+ " the progress report will be printed 3 times during the\n"
+ " benchmark. Else, if the user wants to, he/she can give:\n"
+ " --interval 1234\n"
+ "\n"
+ " and the progress report will be printed every 1234\n"
+ " requests.\n"
"\n");
}
-/*
- * Convert string to size in bytes.
- * If syntax is invalid, return 0. Values such as zero and non-integer
- * multiples of segment's page size should not be accepted.
- */
-static uint64_t str2num(char *str)
-{
- char *unit;
- uint64_t num;
-
- num = strtoll(str, &unit, 10);
- if (strlen(unit) > 1) //Invalid syntax
- return 0;
- else if (strlen(unit) < 1) //Plain number in bytes
- return num;
-
- switch (*unit) {
- case 'g':
- case 'G':
- num *= 1024;
- case 'm':
- case 'M':
- num *= 1024;
- case 'k':
- case 'K':
- num *= 1024;
- break;
- default:
- num = 0;
- }
- return num;
-}
-
int custom_peer_init(struct peerd *peer, int argc, char *argv[])
{
struct bench *prefs;
+ char request_cap[MAX_ARG_LEN + 1];
+ char total_objects[MAX_ARG_LEN + 1];
char total_size[MAX_ARG_LEN + 1];
char object_size[MAX_ARG_LEN + 1];
char block_size[MAX_ARG_LEN + 1];
+ char op[MAX_ARG_LEN + 1];
+ char pattern[MAX_ARG_LEN + 1];
+ char insanity[MAX_ARG_LEN + 1];
+ char verify[MAX_ARG_LEN + 1];
+ char progress[MAX_ARG_LEN + 1];
+ char interval[MAX_ARG_LEN + 1];
+ char ping[MAX_ARG_LEN + 1];
+ char prefix[XSEG_MAX_TARGETLEN + 1];
+ char objname[XSEG_MAX_TARGETLEN + 1];
struct xseg *xseg = peer->xseg;
+ struct object_vars *obv;
unsigned int xseg_page_size = 1 << xseg->config.page_shift;
+ long iodepth = -1;
long dst_port = -1;
-
+ unsigned long seed = -1;
+ unsigned long seed_max;
+ uint64_t rc;
+ struct timespec *ts;
+ int set_by_hand = 1;
+ int j, r;
+
+ op[0] = 0;
+ pattern[0] = 0;
+ total_objects[0] = 0;
total_size[0] = 0;
block_size[0] = 0;
object_size[0] = 0;
-
+ insanity[0] = 0;
+ verify[0] = 0;
+ request_cap[0] = 0;
+ progress[0] = 0;
+ interval[0] = 0;
+ ping[0] = 0;
+ prefix[0] = 0;
+ objname[0] = 0;
+
+ /* allocate struct bench */
prefs = malloc(sizeof(struct bench));
if (!prefs) {
perror("malloc");
- return -1;
+ goto prefs_fail;
+ }
+ memset(prefs, 0, sizeof(struct bench));
+
+ /* allocate struct req_status */
+ prefs->status = malloc(sizeof(struct req_status));
+ if (!prefs->status) {
+ perror("malloc");
+ goto status_fail;
+ }
+ memset(prefs->status, 0, sizeof(struct req_status));
+
+ /* allocate struct object_name */
+ prefs->objvars = malloc(sizeof(struct object_vars));
+ if (!prefs->objvars) {
+ perror("malloc");
+ goto object_name_fail;
+ }
+ memset(prefs->objvars, 0, sizeof(struct object_vars));
+
+ /* allocate struct object_name */
+ prefs->rep = malloc(sizeof(struct progress_report));
+ if (!prefs->rep) {
+ perror("malloc");
+ goto progress_report_fail;
+ }
+ memset(prefs->rep, 0, sizeof(struct progress_report));
+
+ /* allocate a struct timespec for each peer request */
+ for (j = 0; j < peer->nr_ops; j++) {
+ ts = malloc(sizeof(struct timespec));
+ if (!ts) {
+ perror("malloc");
+ goto priv_fail;
+ }
+ peer->peer_reqs[j].priv = ts;
}
//Begin reading the benchmark-specific arguments
BEGIN_READ_ARGS(argc, argv);
+ READ_ARG_STRING("-rc", request_cap, MAX_ARG_LEN);
+ READ_ARG_STRING("-op", op, MAX_ARG_LEN);
+ READ_ARG_STRING("--pattern", pattern, MAX_ARG_LEN);
+ READ_ARG_STRING("-to", total_objects, MAX_ARG_LEN);
READ_ARG_STRING("-ts", total_size, MAX_ARG_LEN);
READ_ARG_STRING("-os", object_size, MAX_ARG_LEN);
READ_ARG_STRING("-bs", block_size, MAX_ARG_LEN);
- READ_ARG_ULONG("--iodepth", prefs->iodepth);
- READ_ARG_ULONG("-dp", dst_port);
+ READ_ARG_ULONG("--iodepth", iodepth);
+ READ_ARG_ULONG("-tp", dst_port);
+ READ_ARG_ULONG("--seed", seed);
+ READ_ARG_STRING("--insanity", insanity, MAX_ARG_LEN);
+ READ_ARG_STRING("--verify", verify, MAX_ARG_LEN);
+ READ_ARG_STRING("--progress", progress, MAX_ARG_LEN);
+ READ_ARG_STRING("--interval", interval, MAX_ARG_LEN);
+ READ_ARG_STRING("--ping", ping, MAX_ARG_LEN);
+ READ_ARG_STRING("--prefix", prefix, XSEG_MAX_TARGETLEN);
+ READ_ARG_STRING("--objname", objname, XSEG_MAX_TARGETLEN);
END_READ_ARGS();
- /*************************
- * Check size parameters *
- *************************/
+ /********************************\
+ * Check object name parameters *
+ \********************************/
+ if (objname[0] && prefix[0]) {
+ XSEGLOG2(&lc, E, "--objname and --prefix options cannot be"
+ "used together.");
+ goto arg_fail;
+ }
- //Block size (bs): Defaults to 4K.
- //It must be a number followed by one of these characters: [k|K|m|M|g|G].
- //If not, it will be considered as size in bytes.
- //Must be integer multiple of segment's page size (typically 4k).
- if (!block_size[0])
- strcpy(block_size,"4k");
+ obv = prefs->objvars;
+ obv->seedlen = SEEDLEN;
+ obv->objnumlen = OBJNUMLEN;
+ if (objname[0]) {
+ strncpy(obv->name, objname, XSEG_MAX_TARGETLEN);
+ obv->prefixlen = 0;
+ obv->namelen = strlen(objname);
+ } else {
+ if (!prefix[0]) /* In this case we use a default value */
+ strcpy(prefix, "bench");
+ strncpy(obv->prefix, prefix, XSEG_MAX_TARGETLEN);
+ obv->prefixlen = strlen(prefix);
+ /* We add 2 for the extra dashes */
+ obv->namelen = obv->prefixlen + obv->seedlen +
+ obv->objnumlen + 2;
+ }
- if (!prefs->iodepth)
- prefs->iodepth = 1;
+ /* Only --prefix can exceed bounds since --objname is bounded */
+ if (obv->namelen > XSEG_MAX_TARGETLEN) {
+ XSEGLOG2(&lc, E, "--prefix %s: Prefix is too long.", prefix);
+ goto arg_fail;
+ }
- prefs->bs = str2num(block_size);
- if (!prefs->bs) {
- XSEGLOG2(&lc, E, "Invalid syntax: %s\n", block_size);
+ /*****************************\
+ * Check I/O type parameters *
+ \*****************************/
+
+ //We support 4 xseg operations: X_READ, X_WRITE, X_DELETE, X_INFO
+ //The I/O pattern of these operations can be either sequential (seq) or
+ //random (rand)
+ if (!op[0]) {
+ XSEGLOG2(&lc, E, "xseg operation needs to be supplied\n");
goto arg_fail;
- } else if (prefs->bs % xseg_page_size) {
- XSEGLOG2(&lc, E, "Misaligned block size: %s\n", block_size);
+ }
+ r = read_op(op);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: -op %s\n", op);
goto arg_fail;
}
+ prefs->op = r;
- //Total I/O size (ts): Must be supplied by user.
- //Must have the same format as "total size"
- //Must be integer multiple of "block size"
- if (!total_size[0]) {
- XSEGLOG2(&lc, E, "Total I/O size needs to be supplied\n");
+ if (!pattern[0]) {
+ XSEGLOG2(&lc, E, "I/O pattern needs to be supplied\n");
+ goto arg_fail;
+ }
+ r = read_pattern(pattern);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --pattern %s\n", pattern);
goto arg_fail;
}
+ SET_FLAG(PATTERN, prefs->flags, r);
+
+ if (!verify[0])
+ strcpy(verify, "no");
+ r = read_verify(verify);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --verify %s\n", verify);
+ goto arg_fail;
+ }
+ SET_FLAG(VERIFY, prefs->flags, r);
+
+ //Default iodepth value is 1
+ if (iodepth < 0)
+ prefs->iodepth = 1;
+ else
+ prefs->iodepth = iodepth;
- prefs->ts = str2num(total_size);
- if (!prefs->ts) {
- XSEGLOG2(&lc, E, "Invalid syntax: %s\n", total_size);
+ /**************************\
+ * Check timer parameters *
+ \**************************/
+
+ //Most of the times, not all timers need to be used.
+ //We can choose which timers will be used by adjusting the "insanity"
+ //level of the benchmark i.e. the obscurity of code paths (get request,
+ //submit request) that will be timed.
+ if (!insanity[0])
+ strcpy(insanity, "sane");
+
+ r = read_insanity(insanity);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --insanity %s\n", insanity);
goto arg_fail;
- } else if (prefs->ts % prefs->bs) {
- XSEGLOG2(&lc, E, "Misaligned total I/O size: %s\n", total_size);
+ }
+ SET_FLAG(INSANITY, prefs->flags, r);
+
+ /*****************************\
+ * Check I/O size parameters *
+ \*****************************/
+
+ //Block size (bs): Defaults to 4K.
+ //It must be a number followed by one of these characters:
+ // [k|K|m|M|g|G]
+ //If not, it will be considered as size in bytes.
+ //Must be integer multiple of segment's page size (typically 4k).
+ if (!block_size[0])
+ strcpy(block_size,"4k");
+
+ prefs->bs = str2num(block_size);
+ if (!prefs->bs) {
+ XSEGLOG2(&lc, E, "Invalid syntax: -bs %s\n", block_size);
goto arg_fail;
- } else if (prefs->ts > xseg->segment_size) {
- XSEGLOG2(&lc, E, "Total I/O size exceeds segment size\n", total_size);
+ } else if (prefs->bs % xseg_page_size) {
+ XSEGLOG2(&lc, E, "Misaligned block size: %s\n", block_size);
goto arg_fail;
}
//Object size (os): Defaults to 4M.
- //Must have the same format as "total size"
+ //Must have the same format as "block size"
//Must be integer multiple of "block size"
if (!object_size[0])
strcpy(object_size,"4M");
prefs->os = str2num(object_size);
if (!prefs->os) {
- XSEGLOG2(&lc, E, "Invalid syntax: %s\n", object_size);
+ XSEGLOG2(&lc, E, "Invalid syntax: -os %s\n", object_size);
goto arg_fail;
} else if (prefs->os % prefs->bs) {
XSEGLOG2(&lc, E, "Misaligned object size: %s\n", object_size);
goto arg_fail;
}
- /*************************
+ //Total objects (to) or total I/O size (ts).
+ //Must have the same format as "block size"
+ //They are mutually exclusive
+ if (total_objects[0] && total_size[0]) {
+ XSEGLOG2(&lc, E, "Total objects and total size are "
+ "mutually exclusive\n");
+ goto arg_fail;
+ } else if (total_objects[0]) {
+ prefs->to = str2num(total_objects);
+ if (!prefs->to) {
+ XSEGLOG2(&lc, E, "Invalid syntax: -to %s\n",
+ total_objects);
+ goto arg_fail;
+ }
+ //In this case, the maximum number of requests is the total
+ //number of objects we will handle
+ prefs->status->max = prefs->to;
+ } else if (total_size[0]) {
+ if (prefs->op != X_READ && prefs->op != X_WRITE) {
+ XSEGLOG2(&lc, E, "Total objects must be supplied "
+ "(required by -op %s)\n", op);
+ goto arg_fail;
+ }
+ prefs->ts = str2num(total_size);
+ if (!prefs->ts) {
+ XSEGLOG2(&lc, E, "Invalid syntax: -ts %s\n", total_size);
+ goto arg_fail;
+ } else if (prefs->ts % prefs->bs) {
+ XSEGLOG2(&lc, E, "Misaligned total I/O size: %s\n", total_size);
+ goto arg_fail;
+ }
+ //In this case, the maximum number of requests is the number of
+ //blocks we need to cover the total I/O size
+ prefs->status->max = prefs->ts / prefs->bs;
+ } else if (!objname[0]) {
+ XSEGLOG2(&lc, E, "Total objects or total size must be supplied\n");
+ goto arg_fail;
+ }
+
+ /*
+ * Enforce --objname restrictions here.
+ */
+ if (obv->name[0]) {
+ if (prefs->to > 1) {
+ XSEGLOG2(&lc, E, "-to %s: Total objects are restricted "
+ "to 1 due to --objname %s\n",
+ total_objects, objname);
+ goto arg_fail;
+ } else if (prefs->ts > prefs->os) {
+ XSEGLOG2(&lc, E, "-ts %s: Total size can't be larger "
+ "than object size (%s) due to "
+ "--objname %s\n",
+ total_size, object_size, objname);
+ goto arg_fail;
+ } else if (prefs->op == X_READ || prefs->op == X_WRITE) {
+ prefs->ts = prefs->os;
+ prefs->status->max = prefs->ts / prefs->bs;
+ } else {
+ prefs->to = 1;
+ prefs->status->max = 1;
+ }
+ }
+
+ if (prefs->status->max == 1)
+ SET_FLAG(PATTERN, prefs->flags, PATTERN_SEQ);
+
+ /*************************\
* Check port parameters *
- *************************/
+ \*************************/
if (dst_port < 0){
- XSEGLOG2(&lc, E, "Destination port needs to be supplied\n");
+ XSEGLOG2(&lc, E, "Target port must be supplied\n");
goto arg_fail;
}
prefs->src_port = peer->portno_start; //TODO: allow user to change this
prefs->dst_port = (xport) dst_port;
- /*********************************
+ /*********************************\
* Create timers for all metrics *
- *********************************/
-
- prefs->total_tm = malloc(sizeof(struct timer));
- prefs->get_tm = malloc(sizeof(struct timer));
- prefs->sub_tm = malloc(sizeof(struct timer));
- prefs->rec_tm = malloc(sizeof(struct timer));
- if (!prefs->total_tm || !prefs->get_tm || !prefs->sub_tm ||
- !prefs->rec_tm) {
- perror("malloc");
- return -1;
+ \*********************************/
+
+ if (init_timer(&prefs->total_tm, INSANITY_SANE))
+ goto tm_fail;
+ if (init_timer(&prefs->sub_tm, INSANITY_MANIC))
+ goto tm_fail;
+ if (init_timer(&prefs->get_tm, INSANITY_PARANOID))
+ goto tm_fail;
+ if (init_timer(&prefs->rec_tm, INSANITY_ECCENTRIC))
+ goto tm_fail;
+
+ /***********************\
+ * Initialize the LFSR *
+ \***********************/
+
+ seed_max = pow(10, obv->seedlen + 1) - 1;
+ if (seed == -1) {
+ srand(time(NULL));
+ set_by_hand = 0;
+ } else if (validate_seed(prefs, seed)) {
+ XSEGLOG2(&lc, E, "--seed %lu: Seed larger than %lu. Only its "
+ "first %d digits will be used",
+ seed, seed_max, obv->seedlen);
+ goto arg_fail;
}
- memset(prefs->total_tm, 0, sizeof(struct timer));
- memset(prefs->get_tm, 0, sizeof(struct timer));
- memset(prefs->sub_tm, 0, sizeof(struct timer));
- memset(prefs->rec_tm, 0, sizeof(struct timer));
- /**************************
- * Customize struct peerd *
- **************************/
+reseed:
+ if (!set_by_hand)
+ seed = rand() % seed_max + 1;
- peer->peerd_loop = custom_peerd_loop;
+ if (GET_FLAG(PATTERN, prefs->flags) == PATTERN_RAND) {
+ prefs->lfsr = malloc(sizeof(struct bench_lfsr));
+ if (!prefs->lfsr) {
+ perror("malloc");
+ goto lfsr_fail;
+ }
+
+ r = lfsr_init(prefs->lfsr, prefs->status->max,
+ seed, seed & 0xF);
+ if (r) {
+ if (!set_by_hand) {
+ free(prefs->lfsr);
+ goto reseed;
+ }
+ XSEGLOG2(&lc, E, "LFSR could not be initialized.\n");
+ goto lfsr_fail;
+ }
+ }
+ obv->seed = seed;
+
+ /*********************************\
+ * Miscellaneous initializations *
+ \*********************************/
+
+ /* The request cap must be enforced only after the LFSR is initialized */
+ if (request_cap[0]) {
+ rc = str2num(request_cap);
+ if (!rc) {
+ XSEGLOG2(&lc, E, "Invalid syntax: -rc %s\n", request_cap);
+ goto arg_fail;
+ } else if (rc > prefs->status->max) {
+ XSEGLOG2(&lc, E, "Request cap exceeds current request total.\n");
+ goto arg_fail;
+ }
+ prefs->status->max = rc;
+ }
+
+ /* Benchmarking progress printing is on by default */
+ if (!progress[0])
+ strcpy(progress, "both");
+ r = read_progress(progress);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --progress %s\n", progress);
+ goto arg_fail;
+ }
+ SET_FLAG(PROGRESS, prefs->flags, r);
+ prefs->rep->lines = calculate_report_lines(prefs);
+
+ /*
+ * Progress report interval definition makes no sense with disabled
+ * progress reports.
+ */
+ if ((GET_FLAG(PROGRESS, prefs->flags) == PROGRESS_NO) &&
+ interval[0]) {
+ XSEGLOG2(&lc, E, "Cannot define progress interval without "
+ "enabling progress report\n");
+ goto arg_fail;
+ }
+
+ if (GET_FLAG(PROGRESS, prefs->flags) != PROGRESS_NO) {
+ /* By default, we print every 5% */
+ if (!interval[0])
+ strcpy(interval, "5%");
+ prefs->rep->interval = read_interval(prefs, interval);
+ if (prefs->rep->interval == 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --interval %s\n",
+ interval);
+ goto arg_fail;
+ }
+ }
+
+ /* Pinging the target peer is on by default */
+ if (!ping[0])
+ strcpy(ping, "no");
+ r = read_ping(ping);
+ if (r < 0) {
+ XSEGLOG2(&lc, E, "Invalid syntax: --ping %s\n", ping);
+ goto arg_fail;
+ }
+ SET_FLAG(PING, prefs->flags, r);
+
+ prefs->peer = peer;
+ peer->peerd_loop = bench_peerd_loop;
peer->priv = (void *) prefs;
+
+ if (obv->prefixlen)
+ XSEGLOG2(&lc, I, "Seed is %u, prefix is %s",
+ obv->seed, obv->prefix);
+ else
+ XSEGLOG2(&lc, I, "Seed is %u, object name is %s",
+ obv->seed, obv->name);
+
return 0;
arg_fail:
- free(prefs);
custom_peer_usage();
+lfsr_fail:
+ free(prefs->lfsr);
+tm_fail:
+ free(prefs->total_tm);
+ free(prefs->sub_tm);
+ free(prefs->get_tm);
+ free(prefs->rec_tm);
+priv_fail:
+ j--;
+ for (; j >= 0; j--) {
+ free(peer->peer_reqs[j].priv);
+ }
+progress_report_fail:
+ free(prefs->rep);
+object_name_fail:
+ free(prefs->objvars);
+status_fail:
+ free(prefs->status);
+prefs_fail:
+ free(prefs);
return -1;
}
struct xseg_request *req;
struct xseg *xseg = peer->xseg;
struct peer_req *pr;
+ struct object_vars *obv = prefs->objvars;
xport srcport = prefs->src_port;
xport dstport = prefs->dst_port;
xport p;
int r;
- uint32_t targetlen = 10; //FIXME: handle it better
+ uint64_t new;
uint64_t size = prefs->bs;
+ struct timespec *ts;
//srcport and dstport must already be provided by the user.
//returns struct xseg_request with basic initializations
- XSEGLOG2(&lc, D, "Get request %lu\n", prefs->get_tm->completed);
- timer_start(prefs->get_tm);
+ XSEGLOG2(&lc, D, "Get new request\n");
+ timer_start(prefs, prefs->get_tm);
req = xseg_get_request(xseg, srcport, dstport, X_ALLOC);
if (!req) {
XSEGLOG2(&lc, W, "Cannot get request\n");
return -1;
}
- timer_stop(prefs->get_tm, NULL);
+ timer_stop(prefs, prefs->get_tm, NULL);
- //Allocate enough space for the data and the target's name
- XSEGLOG2(&lc, D, "Prepare request %lu\n", prefs->sub_tm->completed);
- r = xseg_prep_request(xseg, req, targetlen, size);
+ /*
+ * Allocate enough space for the data and the target's name.
+ * Also, allocate one extra byte to prevent buffer overflow due to the
+ * obligatory null termination of snprint(). This extra byte will not be
+ * counted as part of the target's name.
+ */
+ XSEGLOG2(&lc, D, "Prepare new request\n");
+ r = xseg_prep_request(xseg, req, obv->namelen + 1, size);
if (r < 0) {
XSEGLOG2(&lc, W, "Cannot prepare request! (%lu, %llu)\n",
- (unsigned long)targetlen, (unsigned long long)size);
+ obv->namelen + 1, (unsigned long long)size);
goto put_xseg_request;
}
+ req->targetlen--;
+
+ //Determine what the next target/chunk will be, based on I/O pattern
+ new = determine_next(prefs);
+ req->op = prefs->op;
+ XSEGLOG2(&lc, I, "Our new request is %lu\n", new);
+ obv->objnum = __get_object(prefs, new);
+ create_target(prefs, req);
+
+ if (prefs->op == X_WRITE || prefs->op == X_READ) {
+ req->size = size;
+ //Calculate the chunk's offset inside the object
+ req->offset = calculate_offset(prefs, new);
+ XSEGLOG2(&lc, D, "Offset of request %lu is %lu\n", new, req->offset);
+
+ if (prefs->op == X_WRITE)
+ create_chunk(prefs, req, new);
+ }
-#if 0
- //TODO: allow strcpy, memcpy
- //Copy target's name to the newly allocated space
- req_target = xseg_get_target(xseg, req);
- strncpy(req_target, target, targetlen);
-
- //Copy data buffer to the newly allocated space
- req_data = xseg_get_data(xseg, req);
- memcpy(req_data, buf, size);
- req->offset = offset;
- req->size = size;
- req->op = X_WRITE;
-#endif
-
- //Measure this?
XSEGLOG2(&lc, D, "Allocate peer request\n");
pr = alloc_peer_req(peer);
if (!pr) {
pr->peer = peer;
pr->portno = srcport;
pr->req = req;
- pr->priv = malloc(sizeof(struct timespec));
- XSEGLOG2(&lc, D, "Set request data\n");
+ //XSEGLOG2(&lc, D, "Set request data\n");
r = xseg_set_req_data(xseg, req, pr);
- if (r<0) {
+ if (r < 0) {
XSEGLOG2(&lc, W, "Cannot set request data\n");
goto put_peer_request;
}
/*
* Start measuring receive time.
* When we receive a request, we need to have its submission time to
- * measure elapsed time. Thus, we memcpy its submission time to pr->priv.
+ * measure elapsed time. Thus, we copy its submission time to pr->priv.
* QUESTION: Is this the fastest way?
*/
- timer_start(prefs->rec_tm);
- memcpy(pr->priv, &prefs->rec_tm->start_time, sizeof(struct timespec));
+ timer_start(prefs, prefs->rec_tm);
+ if (prefs->rec_tm->insanity <= GET_FLAG(INSANITY, prefs->flags)) {
+ ts = (struct timespec *)pr->priv;
+ ts->tv_sec = prefs->rec_tm->start_time.tv_sec;
+ ts->tv_nsec = prefs->rec_tm->start_time.tv_nsec;
+ }
//Submit the request from the source port to the target port
- XSEGLOG2(&lc, D, "Submit request %lu\n", prefs->sub_tm->completed);
- //QUESTION: Can't we just use the submision time calculated previously?
- timer_start(prefs->sub_tm);
+ XSEGLOG2(&lc, D, "Submit request %lu\n", new);
+ timer_start(prefs, prefs->sub_tm);
p = xseg_submit(xseg, req, srcport, X_ALLOC);
if (p == NoPort) {
XSEGLOG2(&lc, W, "Cannot submit request\n");
goto put_peer_request;
}
- timer_stop(prefs->sub_tm, NULL);
+ prefs->status->submitted++;
+ timer_stop(prefs, prefs->sub_tm, NULL);
- //Send SIGIO to the process that has binded this port to inform that
+ //Send SIGIO to the process that has bound this port to inform that
//IO is possible
- xseg_signal(xseg, p);
+ r = xseg_signal(xseg, p);
+ //if (r < 0)
+ // XSEGLOG2(&lc, W, "Cannot signal destination peer (reason %d)\n", r);
+
+ return 0;
+
+put_peer_request:
+ free_peer_req(peer, pr);
+put_xseg_request:
+ if (xseg_put_request(xseg, req, srcport))
+ XSEGLOG2(&lc, W, "Cannot put request\n");
+ return -1;
+}
+
+static int send_ping_request(struct peerd *peer, struct bench *prefs)
+{
+ struct xseg_request *req;
+ struct xseg *xseg = peer->xseg;
+ struct peer_req *pr;
+ xport srcport = prefs->src_port;
+ xport dstport = prefs->dst_port;
+ xport p;
+ int r;
+
+ XSEGLOG2(&lc, I, "Sending ping request...");
+ //srcport and dstport must already be provided by the user.
+ //returns struct xseg_request with basic initializations
+ XSEGLOG2(&lc, D, "Get new request\n");
+ req = xseg_get_request(xseg, srcport, dstport, X_ALLOC);
+ if (!req) {
+ XSEGLOG2(&lc, W, "Cannot get request\n");
+ return -1;
+ }
+ req->op = X_PING;
+
+ XSEGLOG2(&lc, D, "Allocate peer request\n");
+ pr = alloc_peer_req(peer);
+ if (!pr) {
+ XSEGLOG2(&lc, W, "Cannot allocate peer request (%ld remaining)\n",
+ peer->nr_ops - xq_count(&peer->free_reqs));
+ goto put_xseg_request;
+ }
+ pr->peer = peer;
+ pr->portno = srcport;
+ pr->req = req;
+
+ r = xseg_set_req_data(xseg, req, pr);
+ if (r < 0) {
+ XSEGLOG2(&lc, W, "Cannot set request data\n");
+ goto put_peer_request;
+ }
+
+ //Submit the request from the source port to the target port
+ XSEGLOG2(&lc, D, "Submit ping request");
+ p = xseg_submit(xseg, req, srcport, X_ALLOC);
+ if (p == NoPort) {
+ XSEGLOG2(&lc, W, "Cannot submit request\n");
+ goto put_peer_request;
+ }
+ timer_stop(prefs, prefs->sub_tm, NULL);
+
+ //Send SIGIO to the process that has bound this port to inform that
+ //IO is possible
+ r = xseg_signal(xseg, p);
+ //if (r < 0)
+ // XSEGLOG2(&lc, W, "Cannot signal destination peer (reason %d)\n", r);
return 0;
put_peer_request:
- free(pr->priv);
free_peer_req(peer, pr);
put_xseg_request:
if (xseg_put_request(xseg, req, srcport))
* This function substitutes the default generic_peerd_loop of peer.c.
* It's plugged to struct peerd at custom peer's initialisation
*/
-int custom_peerd_loop(void *arg)
+int bench_peerd_loop(void *arg)
{
#ifdef MT
struct thread *t = (struct thread *) arg;
struct bench *prefs = peer->priv;
xport portno_start = peer->portno_start;
xport portno_end = peer->portno_end;
+ pid_t pid = syscall(SYS_gettid);
uint64_t threshold=1000/(1 + portno_end - portno_start);
- pid_t pid =syscall(SYS_gettid);
+ uint64_t next_report = prefs->rep->interval;
+ uint64_t loops;
int r;
XSEGLOG2(&lc, I, "%s has tid %u.\n",id, pid);
xseg_init_local_signal(xseg, peer->portno_start);
- uint64_t loops;
- unsigned long max_completed = prefs->ts / prefs->bs;
- timer_start(prefs->total_tm);
+ if (GET_FLAG(PROGRESS, prefs->flags) != PROGRESS_NO)
+ print_dummy_progress(prefs);
+
+ /* If no ping is going to be sent, we can begin the benchmark now. */
+ if (GET_FLAG(PING, prefs->flags) == PING_MODE_ON)
+ send_ping_request(peer, prefs);
+ else
+ timer_start(prefs, prefs->total_tm);
- //while (!isTerminate() && xq_count(&peer->free_reqs) == peer->nr_ops) {
- while (!isTerminate()) {
-#ifdef MT
- if (t->func) {
- XSEGLOG2(&lc, D, "%s executes function\n", id);
- xseg_cancel_wait(xseg, peer->portno_start);
- t->func(t->arg);
- t->func = NULL;
- t->arg = NULL;
- continue;
- }
-#endif
send_request:
- while (prefs->sub_tm->completed - prefs->rec_tm->completed <
- prefs->iodepth){
+ while (!(isTerminate() && all_peer_reqs_free(peer))) {
+ while (CAN_SEND_REQUEST(prefs)) {
+ xseg_cancel_wait(xseg, peer->portno_start);
+ XSEGLOG2(&lc, D, "...because %lu < %lu && %lu < %lu\n",
+ prefs->status->submitted - prefs->status->received,
+ prefs->iodepth, prefs->status->received,
+ prefs->status->max);
XSEGLOG2(&lc, D, "Start sending new request\n");
r = send_request(peer, prefs);
- if (r<0)
+ if (r < 0)
break;
}
-
//Heart of peerd_loop. This loop is common for everyone.
for (loops = threshold; loops > 0; loops--) {
+ if (loops == 1)
+ xseg_prepare_wait(xseg, peer->portno_start);
+
+ if (CAN_PRINT_PROGRESS(prefs, next_report)) {
+ print_progress(prefs);
+ next_report += prefs->rep->interval;
+ }
+
if (check_ports(peer)) {
- if (max_completed == prefs->rec_tm->completed)
- return 0;
- else
- //If an old request has just been acked, the most sensible
- //thing to do is to immediately send a new one
+ //If an old request has just been acked, the
+ //most sensible thing to do is to immediately
+ //send a new one
+ if (prefs->status->received < prefs->status->max)
goto send_request;
+ else
+ return 0;
}
}
- xseg_prepare_wait(xseg, peer->portno_start);
-#ifdef ST_THREADS
- if (ta){
- st_sleep(0);
- continue;
- }
-#endif
- XSEGLOG2(&lc, I, "%s goes to sleep\n",id);
+ XSEGLOG2(&lc, I, "%s goes to sleep\n", id);
xseg_wait_signal(xseg, 10000000UL);
xseg_cancel_wait(xseg, peer->portno_start);
XSEGLOG2(&lc, I, "%s woke up\n", id);
void custom_peer_finalize(struct peerd *peer)
{
struct bench *prefs = peer->priv;
- unsigned int s, ms, us, ns;
-
- if (!prefs->total_tm->completed)
- timer_stop(prefs->total_tm, NULL);
-
- struct timespec tm = prefs->total_tm->sum;
- ns = tm.tv_nsec % 1000;
- tm.tv_nsec /= 1000;
- us = tm.tv_nsec % 1000;
- ms = tm.tv_nsec / 1000;
- s = tm.tv_sec;
-
- printf("\n");
- printf(" Total time spent\n");
- printf("================================\n");
- printf(" |-s-||-ms-|-us-|-ns-|\n");
- printf("Time: %03u, %03u %03u %03u\n", s, ms, us, ns);
+ //TODO: Measure mean time, standard variation
+
+ timer_stop(prefs, prefs->total_tm, NULL);
+
+ if (GET_FLAG(PROGRESS, prefs->flags) != PROGRESS_NO)
+ clear_report_lines(prefs->rep->lines);
+
+ print_req_stats(prefs);
+ print_remaining(prefs);
+ print_total_res(prefs);
+
+ if (GET_FLAG(INSANITY, prefs->flags) >= prefs->rec_tm->insanity)
+ print_rec_res(prefs);
+
+ print_divider();
+ /*
+ * This is kinda hacky but is reasonable.
+ * During the benchmark, we need to calculate the bandwidth within an
+ * interval.
+ * After the benchmark, we need to calculate the total bandwidth. To do
+ * so, we treat the benchmark as one single interval, and that's what
+ * the two lines below do. This way, the same code can aplly to both
+ * cases
+ */
+ prefs->total_tm->elapsed_time = prefs->total_tm->sum;
+ prefs->rep->interval = prefs->status->received;
+ print_io_stats(prefs);
+ fflush(stdout);
return;
}
-
+/*
+ * handle_received: +1 to our received requests.
+ * Do some sanity checks and then check if request is failed.
+ * If not try to verify the request if asked.
+ */
static void handle_received(struct peerd *peer, struct peer_req *pr)
{
//FIXME: handle null pointer
struct bench *prefs = peer->priv;
struct timer *rec = prefs->rec_tm;
+ int start_timer = 0;
if (!pr->req) {
//This is a serious error, so we must stop
return;
}
- if (!pr->priv) {
+ /*
+ * If we were in ping mode, we can now switch off and start the
+ * benchmark.
+ */
+ if (GET_FLAG(PING, prefs->flags) == PING_MODE_ON) {
+ XSEGLOG2(&lc, I, "Ping received. Benchmark can start now.");
+ SET_FLAG(PING, prefs->flags, PING_MODE_OFF);
+ start_timer = 1;
+ goto out;
+ }
+
+ prefs->status->received++;
+
+ if ((GET_FLAG(INSANITY, prefs->flags) < rec->insanity) && !pr->priv) {
XSEGLOG2(&lc, W, "Cannot find submission time of request");
return;
}
- timer_stop(rec, pr->priv);
+ timer_stop(prefs, rec, (struct timespec *)pr->priv);
+
+ if (!(pr->req->state & XS_SERVED))
+ prefs->status->failed++;
+ else if (CAN_VERIFY(prefs) && read_chunk(prefs, pr->req))
+ prefs->status->corrupted++;
+out:
if (xseg_put_request(peer->xseg, pr->req, pr->portno))
XSEGLOG2(&lc, W, "Cannot put xseg request\n");
- //QUESTION, can't we just keep the malloced memory for future use?
- free(pr->priv);
free_peer_req(peer, pr);
+
+ if (start_timer)
+ timer_start(prefs, prefs->total_tm);
}
int dispatch(struct peerd *peer, struct peer_req *pr, struct xseg_request *req,