OpenCL GPU miner
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 

9263 lines
262 KiB

/*
* Copyright 2013-2014 sgminer developers (see AUTHORS.md)
* Copyright 2011-2013 Con Kolivas
* Copyright 2011-2012 Luke Dashjr
* Copyright 2010 Jeff Garzik
*
* 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 3 of the License, or (at your option)
* any later version. See COPYING for more details.
*/
#include "config.h"
#ifdef HAVE_CURSES
#include <curses.h>
#endif
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdbool.h>
#include <stdint.h>
#include <unistd.h>
#include <sys/time.h>
#include <time.h>
#include <math.h>
#include <stdarg.h>
#include <assert.h>
#include <signal.h>
#include <limits.h>
#include <sys/stat.h>
#include <sys/types.h>
#ifndef WIN32
#include <sys/resource.h>
#else
#include <winsock2.h>
#include <windows.h>
#endif
#include <ccan/opt/opt.h>
#include <jansson.h>
#ifdef HAVE_LIBCURL
#include <curl/curl.h>
#else
char *curly = ":D";
#endif
#include <libgen.h>
#include "sph/sph_sha2.h"
#include "sph/sph_blake.h"
#include "compat.h"
#include "miner.h"
#include "findnonce.h"
#include "adl.h"
#include "driver-opencl.h"
#include "bench_block.h"
#include "algorithm.h"
#include "pool.h"
#include "config_parser.h"
#include "events.h"
#if defined(unix) || defined(__APPLE__)
#include <errno.h>
#include <fcntl.h>
#include <sys/wait.h>
#endif
static char packagename[256];
static bool startup = true; //sgminer is starting up
static bool gpu_initialized = false; //gpu initialized
static int init_pool; //pool used to initialize gpus
static bool on_backup_pool = false; //for simple connect strategy... flag if we're on a backup pool
bool opt_work_update;
bool opt_protocol;
bool have_longpoll;
bool want_per_device_stats;
bool use_syslog;
bool opt_quiet;
bool opt_realquiet;
bool opt_loginput;
bool opt_compact;
bool opt_incognito;
// remote config options...
int opt_remoteconf_retry = 3; // number of retries
int opt_remoteconf_wait = 10; // wait in secs between retries
bool opt_remoteconf_usecache = false; // use last downloaded copy of the config file when download fails
const int opt_cutofftemp = 95;
int opt_log_interval = 5;
int opt_queue = 1;
int opt_scantime = 7;
int opt_expiry = 28;
unsigned long long global_hashrate;
unsigned long global_quota_gcd = 1;
bool opt_show_coindiff = false;
time_t last_getwork;
int nDevs;
int opt_dynamic_interval = 7;
int opt_g_threads = -1;
bool opt_restart = true;
int opt_vote = 0;
/*****************************************
* Xn Algorithm options
*****************************************/
int opt_hamsi_expand_big = 4;
int opt_keccak_unroll = 0;
bool opt_hamsi_short = false;
bool opt_blake_compact = false;
bool opt_luffa_parallel = false;
struct list_head scan_devices;
bool devices_enabled[MAX_DEVICES];
int opt_devs_enabled;
static bool opt_display_devs;
bool opt_removedisabled;
int total_devices;
static int most_devices;
struct cgpu_info **devices;
int mining_threads;
#ifdef HAVE_CURSES
bool use_curses = true;
#else
bool use_curses;
#endif
static bool opt_submit_stale = true;
int opt_shares;
bool opt_fail_only;
int opt_fail_switch_delay = 60;
int opt_watchpool_refresh = 30;
static bool opt_fix_protocol;
static bool opt_lowmem;
static bool opt_morenotices;
bool opt_autofan;
bool opt_autoengine;
bool opt_noadl;
char *opt_api_allow = NULL;
char *opt_api_groups;
char *opt_api_description = PACKAGE_STRING;
int opt_api_port = 4028;
bool opt_api_listen;
bool opt_api_mcast;
char *opt_api_mcast_addr = API_MCAST_ADDR;
char *opt_api_mcast_code = API_MCAST_CODE;
char *opt_api_mcast_des = "";
int opt_api_mcast_port = 4028;
bool opt_api_network;
bool opt_delaynet;
bool opt_disable_pool;
bool opt_disable_client_reconnect = false;
static bool no_work;
bool opt_worktime;
#if defined(HAVE_LIBCURL) && defined(CURL_HAS_KEEPALIVE)
int opt_tcp_keepalive = 30;
#else
int opt_tcp_keepalive;
#endif
double opt_diff_mult = 0.0;
char *opt_kernel_path;
char *sgminer_path;
#define QUIET (opt_quiet || opt_realquiet)
struct thr_info *control_thr;
struct thr_info **mining_thr = NULL;
static int gwsched_thr_id;
static int watchpool_thr_id;
static int watchdog_thr_id;
#ifdef HAVE_CURSES
static int input_thr_id;
#endif
int gpur_thr_id;
static int api_thr_id;
static int total_control_threads;
#if LOCK_TRACKING
pthread_mutex_t lockstat_lock;
#endif
pthread_mutex_t hash_lock;
static pthread_mutex_t *stgd_lock;
pthread_mutex_t console_lock;
cglock_t ch_lock;
static pthread_rwlock_t blk_lock;
static pthread_mutex_t sshare_lock;
pthread_rwlock_t netacc_lock;
pthread_rwlock_t mining_thr_lock;
pthread_rwlock_t devices_lock;
static pthread_mutex_t lp_lock;
static pthread_cond_t lp_cond;
static pthread_mutex_t algo_switch_lock;
static int algo_switch_n = 0;
static unsigned long pool_switch_options = 0;
static pthread_mutex_t algo_switch_wait_lock;
static pthread_cond_t algo_switch_wait_cond;
pthread_mutex_t restart_lock;
pthread_cond_t restart_cond;
pthread_cond_t gws_cond;
double total_rolling;
double total_mhashes_done;
static struct timeval total_tv_start, total_tv_end, launch_time;
cglock_t control_lock;
pthread_mutex_t stats_lock;
static void *restart_mining_threads_thread(void *userdata);
static void apply_initial_gpu_settings(struct pool *pool);
static unsigned long compare_pool_settings(struct pool *oldpool, struct pool *newpool);
static void apply_switcher_options(unsigned long options, struct pool *pool);
static void restart_mining_threads(unsigned int new_n_threads);
static void probe_pools(void);
static bool test_pool(struct pool *pool);
int hw_errors;
int total_accepted, total_rejected;
double total_diff1;
int total_getworks, total_stale, total_discarded;
double total_diff_accepted, total_diff_rejected, total_diff_stale;
static int staged_rollable;
unsigned int new_blocks;
static unsigned int work_block;
unsigned int found_blocks;
unsigned int local_work;
unsigned int total_go, total_ro;
struct pool **pools;
static struct pool *currentpool = NULL;
struct strategies strategies[] = {
{ "Failover" },
{ "Round Robin" },
{ "Rotate" },
{ "Load Balance" },
{ "Balance" },
};
int total_pools, enabled_pools;
enum pool_strategy pool_strategy = POOL_FAILOVER;
int opt_rotate_period;
static int total_urls;
//default mode apply algorithm/gpu settings when pool changes
int opt_switchmode = SWITCH_POOL;
static
#ifndef HAVE_CURSES
const
#endif
bool curses_active;
/* Protected by ch_lock */
char current_hash[68];
static char prev_block[12];
static char current_block[32];
static char datestamp[40];
static char blocktime[32];
struct timeval block_timeval;
static char best_share[8] = "0";
double current_diff = 0xFFFFFFFFFFFFFFFFULL;
static char block_diff[8];
double best_diff = 0;
struct block {
char hash[68];
UT_hash_handle hh;
int block_no;
};
static struct block *blocks = NULL;
int swork_id;
/* For creating a hash database of stratum shares submitted that have not had
* a response yet */
struct stratum_share {
UT_hash_handle hh;
bool block;
struct work *work;
int id;
time_t sshare_time;
time_t sshare_sent;
};
static struct stratum_share *stratum_shares = NULL;
char *opt_socks_proxy = NULL;
#if defined(unix) || defined(__APPLE__)
char *opt_stderr_cmd = NULL;
static int forkpid;
#endif // defined(unix)
#ifndef _MSC_VER
struct sigaction termhandler, inthandler;
#endif
struct thread_q *getq;
static int total_work;
struct work *staged_work = NULL;
struct schedtime schedstart;
struct schedtime schedstop;
bool sched_paused;
static bool time_before(struct tm *tm1, struct tm *tm2)
{
if (tm1->tm_hour < tm2->tm_hour)
return true;
if (tm1->tm_hour == tm2->tm_hour && tm1->tm_min < tm2->tm_min)
return true;
return false;
}
static bool should_run(void)
{
struct timeval tv;
struct tm *tm;
if (!schedstart.enable && !schedstop.enable)
return true;
cgtime(&tv);
const time_t tmp_time = tv.tv_sec;
tm = localtime(&tmp_time);
if (schedstart.enable) {
if (!schedstop.enable) {
if (time_before(tm, &schedstart.tm))
return false;
/* This is a once off event with no stop time set */
schedstart.enable = false;
return true;
}
if (time_before(&schedstart.tm, &schedstop.tm)) {
if (time_before(tm, &schedstop.tm) && !time_before(tm, &schedstart.tm))
return true;
return false;
} /* Times are reversed */
if (time_before(tm, &schedstart.tm)) {
if (time_before(tm, &schedstop.tm))
return true;
return false;
}
return true;
}
/* only schedstop.enable == true */
if (!time_before(tm, &schedstop.tm))
return false;
return true;
}
void get_datestamp(char *f, size_t fsiz, struct timeval *tv)
{
struct tm *tm;
const time_t tmp_time = tv->tv_sec;
tm = localtime(&tmp_time);
snprintf(f, fsiz, "[%d-%02d-%02d %02d:%02d:%02d]",
tm->tm_year + 1900,
tm->tm_mon + 1,
tm->tm_mday,
tm->tm_hour,
tm->tm_min,
tm->tm_sec);
}
static void get_timestamp(char *f, size_t fsiz, struct timeval *tv)
{
struct tm *tm;
const time_t tmp_time = tv->tv_sec;
tm = localtime(&tmp_time);
snprintf(f, fsiz, "[%02d:%02d:%02d]",
tm->tm_hour,
tm->tm_min,
tm->tm_sec);
}
static char exit_buf[512];
static void applog_and_exit(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
vsnprintf(exit_buf, sizeof(exit_buf), fmt, ap);
va_end(ap);
_applog(LOG_ERR, exit_buf, true);
exit(1);
}
static pthread_mutex_t sharelog_lock;
static FILE *sharelog_file = NULL;
static struct cgpu_info *get_thr_cgpu(int thr_id)
{
struct cgpu_info *cgpu = NULL;
rd_lock(&mining_thr_lock);
if (thr_id < mining_threads && mining_thr[thr_id])
cgpu = mining_thr[thr_id]->cgpu;
rd_unlock(&mining_thr_lock);
return cgpu;
}
struct cgpu_info *get_devices(int id)
{
struct cgpu_info *cgpu = NULL;
rd_lock(&devices_lock);
if (id < total_devices)
cgpu = devices[id];
rd_unlock(&devices_lock);
return cgpu;
}
void enable_device(int i);
static void sharelog(const char*disposition, const struct work*work)
{
char *target, *hash, *data;
struct cgpu_info *cgpu;
unsigned long int t;
struct pool *pool;
int thr_id, rv;
char s[1024];
size_t ret;
if (!sharelog_file)
return;
thr_id = work->thr_id;
cgpu = get_thr_cgpu(thr_id);
pool = work->pool;
t = (unsigned long int)(work->tv_work_found.tv_sec);
target = bin2hex(work->target, sizeof(work->target));
hash = bin2hex(work->hash, sizeof(work->hash));
data = bin2hex(work->data, sizeof(work->data));
// timestamp,disposition,target,pool,dev,thr,sharehash,sharedata
rv = snprintf(s, sizeof(s), "%lu,%s,%s,%s,%s%u,%u,%s,%s\n", t, disposition, target, pool->rpc_url, cgpu->drv->name, cgpu->device_id, thr_id, hash, data);
free(target);
free(hash);
free(data);
if (rv >= (int)(sizeof(s)))
s[sizeof(s) - 1] = '\0';
else if (rv < 0) {
applog(LOG_ERR, "sharelog printf error");
return;
}
mutex_lock(&sharelog_lock);
ret = fwrite(s, rv, 1, sharelog_file);
fflush(sharelog_file);
mutex_unlock(&sharelog_lock);
if (ret != 1)
applog(LOG_ERR, "sharelog fwrite error");
}
static char *getwork_req = "{\"method\": \"getwork\", \"params\": [], \"id\":0}\n";
static char *gbt_req = "{\"id\": 0, \"method\": \"getblocktemplate\", \"params\": [{\"capabilities\": [\"coinbasetxn\", \"workid\", \"coinbase/append\"]}]}\n";
/* Adjust all the pools' quota to the greatest common denominator after a pool
* has been added or the quotas changed. */
void adjust_quota_gcd(void)
{
unsigned long gcd, lowest_quota = ~0UL, quota;
struct pool *pool;
int i;
for (i = 0; i < total_pools; i++) {
pool = pools[i];
quota = pool->quota;
if (!quota)
continue;
if (quota < lowest_quota)
lowest_quota = quota;
}
if (likely(lowest_quota < ~0UL)) {
gcd = lowest_quota;
for (i = 0; i < total_pools; i++) {
pool = pools[i];
quota = pool->quota;
if (!quota)
continue;
while (quota % gcd)
gcd--;
}
} else
gcd = 1;
for (i = 0; i < total_pools; i++) {
pool = pools[i];
pool->quota_used *= global_quota_gcd;
pool->quota_used /= gcd;
pool->quota_gcd = pool->quota / gcd;
}
global_quota_gcd = gcd;
applog(LOG_DEBUG, "Global quota greatest common denominator set to %lu", gcd);
}
/* Return value is ignored if not called from add_pool_details */
struct pool *add_pool(void)
{
struct pool *pool;
pool = (struct pool *)calloc(sizeof(struct pool), 1);
if (!pool)
quit(1, "Failed to calloc pool in add_pool");
pool->pool_no = pool->prio = total_pools;
/* Default pool name is "" (empty string) */
char buf[32];
buf[0] = '\0';
pool->name = strdup(buf);
pool->profile = strdup(buf); //profile blank by default
pool->algorithm.name[0] = '\0'; //blank algorithm name
/* intensities default to blank */
pool->intensity = strdup(buf);
pool->xintensity = strdup(buf);
pool->rawintensity = strdup(buf);
pools = (struct pool **)realloc(pools, sizeof(struct pool *) * (total_pools + 2));
pools[total_pools++] = pool;
mutex_init(&pool->pool_lock);
if (unlikely(pthread_cond_init(&pool->cr_cond, NULL)))
quit(1, "Failed to pthread_cond_init in add_pool");
cglock_init(&pool->data_lock);
mutex_init(&pool->stratum_lock);
cglock_init(&pool->gbt_lock);
INIT_LIST_HEAD(&pool->curlring);
/* Make sure the pool doesn't think we've been idle since time 0 */
pool->tv_idle.tv_sec = ~0UL;
pool->rpc_req = getwork_req;
pool->rpc_proxy = NULL;
pool->quota = 1;
adjust_quota_gcd();
pool->extranonce_subscribe = true;
pool->description = "";
return pool;
}
/* Used in configuration parsing. */
static struct pool* get_current_pool()
{
while ((json_array_index + 1) > total_pools)
add_pool();
if (json_array_index < 0) {
if (!total_pools)
add_pool();
return pools[total_pools - 1];
}
return pools[json_array_index];
}
/* Used everywhere else (to get pool currently mined on). */
struct pool *current_pool(void)
{
struct pool *pool;
cg_rlock(&control_lock);
pool = currentpool;
cg_runlock(&control_lock);
return pool;
}
/* Pool variant of test and set */
static bool pool_tset(struct pool *pool, bool *var)
{
bool ret;
mutex_lock(&pool->pool_lock);
ret = *var;
*var = true;
mutex_unlock(&pool->pool_lock);
return ret;
}
bool pool_tclear(struct pool *pool, bool *var)
{
bool ret;
mutex_lock(&pool->pool_lock);
ret = *var;
*var = false;
mutex_unlock(&pool->pool_lock);
return ret;
}
char *set_int_range(const char *arg, int *i, int min, int max)
{
char *err = opt_set_intval(arg, i);
if (err)
return err;
if (*i < min || *i > max)
return "Value out of range";
return NULL;
}
char *set_int_0_to_9999(const char *arg, int *i)
{
return set_int_range(arg, i, 0, 9999);
}
char *set_int_1_to_65535(const char *arg, int *i)
{
return set_int_range(arg, i, 1, 65535);
}
char *set_int_0_to_10(const char *arg, int *i)
{
return set_int_range(arg, i, 0, 10);
}
char *set_int_1_to_10(const char *arg, int *i)
{
return set_int_range(arg, i, 1, 10);
}
void get_intrange(char *arg, int *val1, int *val2)
{
if (sscanf(arg, "%d-%d", val1, val2) == 1)
*val2 = *val1;
}
char *set_devices(char *arg)
{
int i, val1 = 0, val2 = 0;
char *p, *nextptr;
if(arg[0] != '\0')
{
if(!strcasecmp(arg, "?"))
{
opt_display_devs = true;
return NULL;
}
//all devices enabled
else if(!strcasecmp(arg, "*") || !strcasecmp(arg, "all"))
{
applog(LOG_DEBUG, "set_devices(all)");
opt_devs_enabled = 0;
return NULL;
}
}
else
return "Invalid device parameters";
applog(LOG_DEBUG, "set_devices(%s)", arg);
p = strdup(arg);
nextptr = strtok(p, ",");
do {
if (nextptr == NULL)
{
free(p);
return "Invalid parameters for set devices";
}
get_intrange(nextptr, &val1, &val2);
if (val1 < 0 || val1 > MAX_DEVICES || val2 < 0 || val2 > MAX_DEVICES || val1 > val2)
{
free(p);
return "Invalid value passed to set devices";
}
for (i = val1; i <= val2; i++)
{
devices_enabled[i] = true;
opt_devs_enabled++;
}
} while ((nextptr = strtok(NULL, ",")) != NULL);
applog(LOG_DEBUG, "set_devices(%s) done.", arg);
free(p);
return NULL;
}
static char *set_balance(enum pool_strategy *strategy)
{
*strategy = POOL_BALANCE;
return NULL;
}
static char *set_loadbalance(enum pool_strategy *strategy)
{
*strategy = POOL_LOADBALANCE;
return NULL;
}
static char *set_rotate(const char *arg, int *i)
{
pool_strategy = POOL_ROTATE;
return set_int_range(arg, i, 0, 9999);
}
static char *set_rr(enum pool_strategy *strategy)
{
*strategy = POOL_ROUNDROBIN;
return NULL;
}
/* Detect that url is for a stratum protocol either via the presence of
* stratum+tcp or by detecting a stratum server response */
bool detect_stratum(struct pool *pool, char *url)
{
if (!extract_sockaddr(url, &pool->sockaddr_url, &pool->stratum_port))
return false;
if (!strncasecmp(url, "stratum+tcp://", 14)) {
pool->rpc_url = strdup(url);
pool->has_stratum = true;
pool->stratum_url = pool->sockaddr_url;
return true;
}
return false;
}
static struct pool *add_url(void)
{
total_urls++;
if (total_urls > total_pools)
add_pool();
return pools[total_urls - 1];
}
static void setup_url(struct pool *pool, char *arg)
{
arg = get_proxy(arg, pool);
if (detect_stratum(pool, arg))
return;
opt_set_charp(arg, &pool->rpc_url);
if (strncmp(arg, "http://", 7) && strncmp(arg, "https://", 8)) {
char *httpinput;
httpinput = (char *)malloc(255);
if (!httpinput)
quit(1, "Failed to malloc httpinput");
strcpy(httpinput, "http://");
strncat(httpinput, arg, 248);
pool->rpc_url = httpinput;
}
}
static char *set_url(char *arg)
{
struct pool *pool = add_url();
setup_url(pool, arg);
return NULL;
}
static char *set_pool_algorithm(const char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i algorithm to %s", pool->pool_no, arg);
set_algorithm(&pool->algorithm, arg);
return NULL;
}
static char *set_pool_backup(const char *arg)
{
struct pool *pool = get_current_pool();
pool->backup = TRUE;
return NULL;
}
static char *set_pool_devices(const char *arg)
{
struct pool *pool = get_current_pool();
pool->devices = arg;
return NULL;
}
static char *set_pool_kernelfile(const char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i algorithm kernel file to %s", pool->pool_no, arg);
pool->algorithm.kernelfile = arg;
return NULL;
}
static char *set_pool_lookup_gap(const char *arg)
{
struct pool *pool = get_current_pool();
pool->lookup_gap = arg;
return NULL;
}
static char *set_pool_intensity(const char *arg)
{
struct pool *pool = get_current_pool();
opt_set_charp(arg, &pool->intensity);
return NULL;
}
static char *set_pool_xintensity(const char *arg)
{
struct pool *pool = get_current_pool();
opt_set_charp(arg, &pool->xintensity);
return NULL;
}
static char *set_pool_rawintensity(const char *arg)
{
struct pool *pool = get_current_pool();
opt_set_charp(arg, &pool->rawintensity);
return NULL;
}
static char *set_pool_thread_concurrency(const char *arg)
{
struct pool *pool = get_current_pool();
pool->thread_concurrency = arg;
return NULL;
}
#ifdef HAVE_ADL
static char *set_pool_gpu_engine(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_engine = arg;
return NULL;
}
static char *set_pool_gpu_memclock(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_memclock = arg;
return NULL;
}
static char *set_pool_gpu_threads(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_threads = arg;
return NULL;
}
static char *set_pool_gpu_fan(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_fan = arg;
return NULL;
}
static char *set_pool_gpu_powertune(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_powertune = arg;
return NULL;
}
static char *set_pool_gpu_vddc(const char *arg)
{
struct pool *pool = get_current_pool();
pool->gpu_vddc = arg;
return NULL;
}
#endif
static char *set_pool_nfactor(const char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i N-factor to %s", pool->pool_no, arg);
set_algorithm_nfactor(&pool->algorithm, (const uint8_t) atoi(arg));
return NULL;
}
static char *set_pool_name(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i name to %s", pool->pool_no, arg);
opt_set_charp(arg, &pool->name);
return NULL;
}
static char *set_pool_profile(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i profile to %s", pool->pool_no, arg);
opt_set_charp(arg, &pool->profile);
return NULL;
}
static char *set_poolname_deprecated(char *arg)
{
applog(LOG_ERR, "Specifying pool name by --poolname is deprecated. Use --name instead.");
set_pool_name(arg);
return NULL;
}
static char *set_pool_shaders(const char *arg)
{
struct pool *pool = get_current_pool();
pool->shaders = arg;
return NULL;
}
static char *set_pool_worksize(const char *arg)
{
struct pool *pool = get_current_pool();
pool->worksize = arg;
return NULL;
}
static void enable_pool(struct pool *pool)
{
if (pool->state != POOL_ENABLED)
enabled_pools++;
pool->state = POOL_ENABLED;
}
static void disable_pool(struct pool *pool)
{
if (pool->state == POOL_ENABLED)
enabled_pools--;
pool->state = POOL_DISABLED;
}
static void reject_pool(struct pool *pool)
{
if (pool->state == POOL_ENABLED)
enabled_pools--;
pool->state = POOL_REJECTING;
}
/* We can't remove the memory used for this struct pool because there may
* still be work referencing it. We just remove it from the pools list */
void remove_pool(struct pool *pool)
{
int i, last_pool = total_pools - 1;
struct pool *other;
/* Boost priority of any lower prio than this one */
for (i = 0; i < total_pools; i++) {
other = pools[i];
if (other->prio > pool->prio)
other->prio--;
}
if (pool->pool_no < last_pool) {
/* Swap the last pool for this one */
(pools[last_pool])->pool_no = pool->pool_no;
pools[pool->pool_no] = pools[last_pool];
}
/* Give it an invalid number */
pool->pool_no = total_pools;
pool->removed = true;
total_pools--;
}
static char *set_pool_state(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_INFO, "Setting pool %s state to %s", get_pool_name(pool), arg);
if (strcmp(arg, "disabled") == 0) {
pool->state = POOL_DISABLED;
} else if (strcmp(arg, "enabled") == 0) {
pool->state = POOL_ENABLED;
} else if (strcmp(arg, "hidden") == 0) {
pool->state = POOL_HIDDEN;
} else if (strcmp(arg, "rejecting") == 0) {
pool->state = POOL_REJECTING;
} else {
pool->state = POOL_ENABLED;
}
return NULL;
}
static char *set_switcher_mode(char *arg)
{
if(!strcasecmp(arg, "off"))
opt_switchmode = SWITCH_OFF;
else if(!strcasecmp(arg, "algorithm"))
opt_switchmode = SWITCH_ALGO;
else if(!strcasecmp(arg, "pool"))
opt_switchmode = SWITCH_POOL;
else
return NULL;
applog(LOG_INFO, "Setting switcher mode to %s", arg);
return NULL;
}
static char *set_quota(char *arg)
{
char *semicolon = strchr(arg, ';'), *url;
size_t len, qlen;
int quota;
struct pool *pool;
if (!semicolon)
return "No semicolon separated quota;URL pair found";
len = strlen(arg);
*semicolon = '\0';
qlen = strlen(arg);
if (!qlen)
return "No parameter for quota found";
len -= qlen + 1;
if (len < 1)
return "No parameter for URL found";
quota = atoi(arg);
if (quota < 0)
return "Invalid negative parameter for quota set";
url = arg + qlen + 1;
pool = add_url();
setup_url(pool, url);
pool->quota = quota;
applog(LOG_INFO, "Setting %s to quota %d", get_pool_name(pool), pool->quota);
adjust_quota_gcd();
return NULL;
}
static char *set_user(const char *arg)
{
struct pool *pool = get_current_pool();
opt_set_charp(arg, &pool->rpc_user);
return NULL;
}
static char *set_pass(const char *arg)
{
struct pool *pool = get_current_pool();
opt_set_charp(arg, &pool->rpc_pass);
return NULL;
}
static char *set_userpass(const char *arg)
{
struct pool *pool = get_current_pool();
char *updup;
updup = strdup(arg);
opt_set_charp(arg, &pool->rpc_userpass);
pool->rpc_user = strtok(updup, ":");
if (!pool->rpc_user)
return "Failed to find : delimited user info";
pool->rpc_pass = strtok(NULL, ":");
if (!pool->rpc_pass)
pool->rpc_pass = "";
return NULL;
}
static char *set_no_extranonce_subscribe(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Disable extranonce subscribe on %d", pool->pool_no);
opt_set_invbool(&pool->extranonce_subscribe);
return NULL;
}
static char *set_pool_priority(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i priority to %s", pool->pool_no, arg);
opt_set_intval(arg, &pool->prio);
return NULL;
}
static char *set_pool_description(char *arg)
{
struct pool *pool = get_current_pool();
applog(LOG_DEBUG, "Setting pool %i description to %s", pool->pool_no, arg);
opt_set_charp(arg, &pool->description);
return NULL;
}
static char *enable_debug(bool *flag)
{
*flag = true;
opt_debug_console = true;
/* Turn on verbose output, too. */
opt_verbose = true;
return NULL;
}
static char *set_schedtime(const char *arg, struct schedtime *st)
{
if (sscanf(arg, "%d:%d", &st->tm.tm_hour, &st->tm.tm_min) != 2)
return "Invalid time set, should be HH:MM";
if (st->tm.tm_hour > 23 || st->tm.tm_min > 59 || st->tm.tm_hour < 0 || st->tm.tm_min < 0)
return "Invalid time set.";
st->enable = true;
return NULL;
}
static char *set_log_file(char *arg)
{
char *r = "";
long int i = strtol(arg, &r, 10);
int fd, stderr_fd = fileno(stderr);
if ((!*r) && i >= 0 && i <= INT_MAX)
fd = i;
else
if (!strcmp(arg, "-"))
{
fd = fileno(stdout);
if (unlikely(fd == -1))
return "Standard output missing for log-file";
}
else
{
fd = open(arg, O_WRONLY | O_APPEND | O_CREAT, S_IRUSR | S_IWUSR);
if (unlikely(fd == -1))
return "Failed to open %s for log-file";
}
close(stderr_fd);
if (unlikely(-1 == dup2(fd, stderr_fd)))
return "Failed to dup2 for log-file";
close(fd);
return NULL;
}
static char* set_sharelog(char *arg)
{
char *r = "";
long int i = strtol(arg, &r, 10);
if ((!*r) && i >= 0 && i <= INT_MAX) {
sharelog_file = fdopen((int)i, "a");
if (!sharelog_file)
applog(LOG_ERR, "Failed to open fd %u for share log", (unsigned int)i);
} else if (!strcmp(arg, "-")) {
sharelog_file = stdout;
if (!sharelog_file)
applog(LOG_ERR, "Standard output missing for share log");
} else {
sharelog_file = fopen(arg, "a");
if (!sharelog_file)
applog(LOG_ERR, "Failed to open %s for share log", arg);
}
return NULL;
}
static char *temp_cutoff_str = NULL;
char *set_temp_cutoff(char *arg)
{
int val;
if (!(arg && arg[0]))
return "Invalid parameters for set temp cutoff";
val = atoi(arg);
if (val < 0 || val > 200)
return "Invalid value passed to set temp cutoff";
temp_cutoff_str = arg;
return NULL;
}
static void load_temp_cutoffs()
{
int i, val = 0, device = 0;
char *nextptr;
if (temp_cutoff_str) {
for (device = 0, nextptr = strtok(temp_cutoff_str, ","); nextptr; ++device, nextptr = strtok(NULL, ",")) {
if (device >= total_devices)
quit(1, "Too many values passed to set temp cutoff");
val = atoi(nextptr);
if (val < 0 || val > 200)
quit(1, "Invalid value passed to set temp cutoff");
rd_lock(&devices_lock);
devices[device]->cutofftemp = val;
rd_unlock(&devices_lock);
}
} else {
rd_lock(&devices_lock);
for (i = device; i < total_devices; ++i) {
if (!devices[i]->cutofftemp)
devices[i]->cutofftemp = opt_cutofftemp;
}
rd_unlock(&devices_lock);
return;
}
if (device <= 1) {
rd_lock(&devices_lock);
for (i = device; i < total_devices; ++i)
devices[i]->cutofftemp = val;
rd_unlock(&devices_lock);
}
}
static char *set_api_allow(const char *arg)
{
opt_set_charp(arg, &opt_api_allow);
return NULL;
}
static char *set_api_groups(const char *arg)
{
opt_set_charp(arg, &opt_api_groups);
return NULL;
}
static char *set_api_description(const char *arg)
{
opt_set_charp(arg, &opt_api_description);
return NULL;
}
static char *set_api_mcast_addr(const char *arg)
{
opt_set_charp(arg, &opt_api_mcast_addr);
return NULL;
}
static char *set_api_mcast_code(const char *arg)
{
opt_set_charp(arg, &opt_api_mcast_code);
return NULL;
}
static char *set_api_mcast_des(const char *arg)
{
opt_set_charp(arg, &opt_api_mcast_des);
return NULL;
}
static char *set_null(const char __maybe_unused *arg)
{
return NULL;
}
char *set_difficulty_multiplier(char *arg)
{
applog(LOG_WARNING, "Option difficulty-multiplier is deprecated and will be removed in v5.1. Pools that need this option should fix their setup.");
if (!(arg && arg[0]))
return "Invalid parameters for set difficulty multiplier";
opt_diff_mult = strtod(arg, NULL);
if (opt_diff_mult == 0.0)
return "Invalid value passed to set difficulty multiplier";
return NULL;
}
/* These options are available from config file or commandline */
struct opt_table opt_config_table[] = {
OPT_WITH_ARG("--algorithm|--kernel|-k",
set_default_algorithm, NULL, NULL,
"Set mining algorithm and most common defaults, default: scrypt"),
OPT_WITH_ARG("--api-allow",
set_api_allow, NULL, NULL,
"Allow API access only to the given list of [G:]IP[/Prefix] addresses[/subnets]"),
OPT_WITH_ARG("--api-description",
set_api_description, NULL, NULL,
"Description placed in the API status header, default: sgminer version"),
OPT_WITH_ARG("--api-groups",
set_api_groups, NULL, NULL,
"API one letter groups G:cmd:cmd[,P:cmd:*...] defining the cmds a groups can use"),
OPT_WITHOUT_ARG("--api-listen",
opt_set_bool, &opt_api_listen,
"Enable API, default: disabled"),
OPT_WITHOUT_ARG("--api-mcast",
opt_set_bool, &opt_api_mcast,
"Enable API Multicast listener, default: disabled"),
OPT_WITH_ARG("--api-mcast-addr",
set_api_mcast_addr, NULL, NULL,
"API Multicast listen address"),
OPT_WITH_ARG("--api-mcast-code",
set_api_mcast_code, NULL, NULL,
"Code expected in the API Multicast message, don't use '-'"),
OPT_WITH_ARG("--api-mcast-des",
set_api_mcast_des, NULL, NULL,
"Description appended to the API Multicast reply, default: ''"),
OPT_WITH_ARG("--api-mcast-port",
set_int_1_to_65535, opt_show_intval, &opt_api_mcast_port,
"API Multicast listen port"),
OPT_WITHOUT_ARG("--api-network",
opt_set_bool, &opt_api_network,
"Allow API (if enabled) to listen on/for any address, default: only 127.0.0.1"),
OPT_WITH_ARG("--api-port",
set_int_1_to_65535, opt_show_intval, &opt_api_port,
"Port number of miner API"),
#ifdef HAVE_ADL
OPT_WITHOUT_ARG("--auto-fan",
opt_set_bool, &opt_autofan,
"Automatically adjust all GPU fan speeds to maintain a target temperature"),
OPT_WITHOUT_ARG("--auto-gpu",
opt_set_bool, &opt_autoengine,
"Automatically adjust all GPU engine clock speeds to maintain a target temperature"),
#endif
OPT_WITHOUT_ARG("--balance",
set_balance, &pool_strategy,
"Change multipool strategy from failover to even share balance"),
OPT_WITHOUT_ARG("--blake-compact",
opt_set_bool, &opt_blake_compact,
"Set SPH_COMPACT_BLAKE64 for Xn derived algorithms (Can give better hashrate for some GPUs)"),
#ifdef HAVE_CURSES
OPT_WITHOUT_ARG("--compact",
opt_set_bool, &opt_compact,
"Use compact display without per device statistics"),
#endif
OPT_WITHOUT_ARG("--debug|-D",
enable_debug, &opt_debug,
"Enable debug output"),
OPT_WITHOUT_ARG("--debug-log",
opt_set_bool, &opt_debug,
"Enable debug logging when stderr is redirected to file"),
OPT_WITH_ARG("--default-profile",
set_default_profile, NULL, NULL,
"Set Default Profile"),
OPT_WITH_ARG("--description",
set_pool_description, NULL, NULL,
"Pool description"),
OPT_WITH_ARG("--device|-d",
set_default_devices, NULL, NULL,
"Select device to use, one value, range and/or comma separated (e.g. 0-2,4) default: all"),
OPT_WITHOUT_ARG("--disable-rejecting",
opt_set_bool, &opt_disable_pool,
"Automatically disable pools that continually reject shares"),
OPT_WITH_ARG("--expiry|-E",
set_int_0_to_9999, opt_show_intval, &opt_expiry,
"Upper bound on how many seconds after getting work we consider a share from it stale"),
// event options
OPT_WITH_ARG("--event-on",
set_event_type, NULL, NULL,
"Select event type to perform task on"),
OPT_WITH_ARG("--event-runcmd",
set_event_runcmd, NULL, NULL,
"Command to perform on event"),
OPT_WITH_ARG("--event-reboot",
set_event_reboot, NULL, NULL,
"Reboot the system on event"),
OPT_WITH_ARG("--event-reboot-delay",
set_event_reboot_delay, NULL, NULL,
"Delay in seconds to wait before rebooting"),
OPT_WITH_ARG("--event-quit",
set_event_quit, NULL, NULL,
"Quit sgminer on event"),
OPT_WITH_ARG("--event-quit-message",
set_event_quit_message, NULL, NULL,
"Quit message when quitting sgminer on event"),
OPT_WITHOUT_ARG("--failover-only",
opt_set_bool, &opt_fail_only,
"Don't leak work to backup pools when primary pool is lagging"),
OPT_WITH_ARG("--failover-switch-delay",
set_int_1_to_65535, opt_show_intval, &opt_fail_switch_delay,
"Delay in seconds before switching back to a failed pool"),
OPT_WITHOUT_ARG("--fix-protocol",
opt_set_bool, &opt_fix_protocol,
"Do not redirect to a different getwork protocol (eg. stratum)"),
OPT_WITH_ARG("--gpu-dyninterval",
set_int_1_to_65535, opt_show_intval, &opt_dynamic_interval,
"Set the refresh interval in ms for GPUs using dynamic intensity"),
OPT_WITH_ARG("--gpu-platform",
set_int_0_to_9999, opt_show_intval, &opt_platform_id,
"Select OpenCL platform ID to use for GPU mining"),
#ifndef HAVE_ADL
// gpu-threads can only be set per-card if ADL is available
OPT_WITH_ARG("--gpu-threads|-g",
set_int_1_to_10, opt_show_intval, &opt_g_threads,
"Number of threads per GPU (1 - 10)"),
#else
OPT_WITH_ARG("--gpu-threads|-g",
set_default_gpu_threads, NULL, NULL,
"Number of threads per GPU - one value or comma separated list (e.g. 1,2,1)"),
OPT_WITH_ARG("--gpu-engine",
set_default_gpu_engine, NULL, NULL,
"GPU engine (over)clock range in Mhz - one value, range and/or comma separated list (e.g. 850-900,900,750-850)"),
OPT_WITH_ARG("--gpu-fan",
set_default_gpu_fan, NULL, NULL,
"GPU fan percentage range - one value, range and/or comma separated list (e.g. 0-85,85,65)"),
OPT_WITH_ARG("--gpu-map",
set_gpu_map, NULL, NULL,
"Map OpenCL to ADL device order manually, paired CSV (e.g. 1:0,2:1 maps OpenCL 1 to ADL 0, 2 to 1)"),
OPT_WITH_ARG("--gpu-memclock",
set_default_gpu_memclock, NULL, NULL,
"Set the GPU memory (over)clock in Mhz - one value for all or separate by commas for per card"),
OPT_WITH_ARG("--gpu-memdiff",
set_gpu_memdiff, NULL, NULL,
"Set a fixed difference in clock speed between the GPU and memory in auto-gpu mode"),
OPT_WITH_ARG("--gpu-powertune",
set_default_gpu_powertune, NULL, NULL,
"Set the GPU powertune percentage - one value for all or separate by commas for per card"),
OPT_WITHOUT_ARG("--gpu-reorder",
opt_set_bool, &opt_reorder,
"Attempt to reorder GPU devices according to PCI Bus ID"),
OPT_WITH_ARG("--gpu-vddc",
set_default_gpu_vddc, NULL, NULL,
"Set the GPU voltage in Volts - one value for all or separate by commas for per card"),
#endif
OPT_WITH_ARG("--hamsi-expand-big",
set_int_1_to_10, opt_show_intval, &opt_hamsi_expand_big,
"Set SPH_HAMSI_EXPAND_BIG for X13 derived algorithms (1 or 4 are common)"),
OPT_WITHOUT_ARG("--hamsi-short",
opt_set_bool, &opt_hamsi_short,
"Set SPH_HAMSI_SHORT for X13 derived algorithms (Can give better hashrate for some GPUs)"),
OPT_WITH_ARG("--keccak-unroll",
set_int_0_to_9999, opt_show_intval, &opt_keccak_unroll,
"Set SPH_KECCAK_UNROLL for Xn derived algorithms (Default: 0)"),
OPT_WITH_ARG("--kernelfile",
set_default_kernelfile, NULL, NULL,
"Set the algorithm kernel source file (without file extension)."),
OPT_WITH_ARG("--lookup-gap",
set_default_lookup_gap, NULL, NULL,
"Set GPU lookup gap for scrypt mining, comma separated"),
OPT_WITHOUT_ARG("--luffa-parallel",
opt_set_bool, &opt_luffa_parallel,
"Set SPH_LUFFA_PARALLEL for Xn derived algorithms (Can give better hashrate for some GPUs)"),
#ifdef HAVE_CURSES
OPT_WITHOUT_ARG("--incognito",
opt_set_bool, &opt_incognito,
"Do not display user name in status window"),
#endif
OPT_WITHOUT_ARG("--more-notices",
opt_set_bool, &opt_morenotices,
"Shows work restart and new block notices, hidden by default"),
OPT_WITH_ARG("--intensity|-I",
set_default_intensity, NULL, NULL,
"Intensity of GPU scanning (d or " MIN_INTENSITY_STR
" -> " MAX_INTENSITY_STR
",default: d to maintain desktop interactivity), overridden by --xintensity or --rawintensity."),
OPT_WITH_ARG("--xintensity|-X",
set_default_xintensity, NULL, NULL,
"Shader based intensity of GPU scanning (" MIN_XINTENSITY_STR " to "
MAX_XINTENSITY_STR "), overridden --xintensity|-X and --rawintensity."),
OPT_WITH_ARG("--rawintensity",
set_default_rawintensity, NULL, NULL,
"Raw intensity of GPU scanning (" MIN_RAWINTENSITY_STR " to "
MAX_RAWINTENSITY_STR "), overrides --intensity|-I and --xintensity|-X."),
OPT_WITH_ARG("--kernel-path|-K",
opt_set_charp, opt_show_charp, &opt_kernel_path,
"Specify a path to where kernel files are"),
OPT_WITHOUT_ARG("--load-balance",
set_loadbalance, &pool_strategy,
"Change multipool strategy from failover to quota based balance"),
OPT_WITH_ARG("--log|-l",
set_int_0_to_9999, opt_show_intval, &opt_log_interval,
"Interval in seconds between log output"),
OPT_WITH_ARG("--log-file|-L",
set_log_file, NULL, NULL,
"Log stderr to file"),
OPT_WITHOUT_ARG("--log-show-date|-L",
opt_set_bool, &opt_log_show_date,
"Show date on every log line"),
OPT_WITHOUT_ARG("--lowmem",
opt_set_bool, &opt_lowmem,
"Minimise caching of shares for low memory applications"),
#if defined(unix) || defined(__APPLE__)
OPT_WITH_ARG("--monitor|-m",
opt_set_charp, NULL, &opt_stderr_cmd,
"Use custom pipe cmd for output messages"),
#endif // defined(unix)
OPT_WITH_ARG("--name|--pool-name",
set_pool_name, NULL, NULL,
"Name of pool"),
OPT_WITHOUT_ARG("--net-delay",
opt_set_bool, &opt_delaynet,
"Impose small delays in networking to not overload slow routers"),
OPT_WITH_ARG("--nfactor",
set_default_nfactor, NULL, NULL,
"Override default scrypt N-factor parameter."),
#ifdef HAVE_ADL
OPT_WITHOUT_ARG("--no-adl",
opt_set_bool, &opt_noadl,
"Disable the ATI display library used for monitoring and setting GPU parameters"),
#else
OPT_WITHOUT_ARG("--no-adl",
opt_set_bool, &opt_noadl, opt_hidden),
#endif
OPT_WITHOUT_ARG("--no-pool-disable",
opt_set_invbool, &opt_disable_pool,
opt_hidden),
OPT_WITHOUT_ARG("--no-client-reconnect",
opt_set_invbool, &opt_disable_client_reconnect,
"Disable 'client.reconnect' stratum functionality"),
OPT_WITHOUT_ARG("--no-restart",
opt_set_invbool, &opt_restart,
"Do not attempt to restart GPUs that hang"),
OPT_WITHOUT_ARG("--no-submit-stale",
opt_set_invbool, &opt_submit_stale,
"Don't submit shares if they are detected as stale"),
OPT_WITHOUT_ARG("--no-extranonce|--pool-no-extranonce",
set_no_extranonce_subscribe, NULL,
"Disable 'extranonce' stratum subscribe for pool"),
OPT_WITH_ARG("--pass|--pool-pass|-p",
set_pass, NULL, NULL,
"Password for bitcoin JSON-RPC server"),
OPT_WITHOUT_ARG("--per-device-stats",
opt_set_bool, &want_per_device_stats,
"Force verbose mode and output per-device statistics"),
OPT_WITH_ARG("--poolname", /* TODO: Backward compatibility, to be removed. */
set_poolname_deprecated, NULL, NULL,
opt_hidden),
OPT_WITH_ARG("--pool-algorithm|--pool-kernel",
set_pool_algorithm, NULL, NULL,
"Set algorithm for pool"),
OPT_WITHOUT_ARG("--pool-backup",
set_pool_backup, NULL,
"Mark this pool as a backup for simple connect strategy"),
OPT_WITH_ARG("--pool-device",
set_pool_devices, NULL, NULL,
"Select devices to use with pool, one value, range and/or comma separated (e.g. 0-2,4) default: all"),
OPT_WITH_ARG("--pool-kernelfile",
set_pool_kernelfile, NULL, NULL,
"Set the pool's algorithm kernel source file (without file extension)."),
OPT_WITH_ARG("--pool-lookup-gap",
set_pool_lookup_gap, NULL, NULL,
"Set Pool GPU lookup gap for scrypt mining, comma separated"),
#ifdef HAVE_ADL
OPT_WITH_ARG("--pool-gpu-engine",
set_pool_gpu_engine, NULL, NULL,
"Pool GPU engine (over)clock range in Mhz - one value, range and/or comma separated list (e.g. 850-900,900,750-850)"),
OPT_WITH_ARG("--pool-gpu-fan",
set_pool_gpu_fan, NULL, NULL,
"GPU fan for pool"),
OPT_WITH_ARG("--pool-gpu-memclock",
set_pool_gpu_memclock, NULL, NULL,
"Set the Pool GPU memory (over)clock in Mhz - one value for all or separate by commas for per card"),
OPT_WITH_ARG("--pool-gpu-powertune",
set_pool_gpu_powertune, NULL, NULL,
"Set the Pool GPU powertune percentage - one value for all or separate by commas for per card"),
OPT_WITH_ARG("--pool-gpu-threads",
set_pool_gpu_threads, NULL, NULL,
"Number of threads per GPU for pool"),
OPT_WITH_ARG("--pool-gpu-vddc",
set_pool_gpu_vddc, NULL, NULL,
"Set the Pool GPU voltage in Volts - one value for all or separate by commas for per card"),
#endif
OPT_WITH_ARG("--pool-intensity",
set_pool_intensity, NULL, NULL,
"Intensity of GPU scanning (pool-specific)"),
OPT_WITH_ARG("--pool-nfactor",
set_pool_nfactor, NULL, NULL,
"Set N-factor for pool"),
OPT_WITH_ARG("--pool-profile",
set_pool_profile, NULL, NULL,
"Profile to use with the pool"),
OPT_WITH_ARG("--pool-rawintensity",
set_pool_rawintensity, NULL, NULL,
"Raw intensity of GPU scanning (pool-specific)"),
OPT_WITH_ARG("--pool-shaders",
set_pool_shaders, NULL, NULL,
"Pool GPU shaders per card for tuning scrypt, comma separated"),
OPT_WITH_ARG("--pool-thread-concurrency",
set_pool_thread_concurrency, NULL, NULL,
"Set thread concurrency for pool"),
OPT_WITH_ARG("--pool-worksize",
set_pool_worksize, NULL, NULL,
"Override detected optimal worksize for pool - one value or comma separated list"),
OPT_WITH_ARG("--pool-xintensity",
set_pool_xintensity, NULL, NULL,
"Shader based intensity of GPU scanning (pool-specific)"),
OPT_WITH_ARG("--priority|--pool-priority",
set_pool_priority, NULL, NULL,
"Pool priority"),
OPT_WITH_ARG("--profile-algorithm|--profile-kernel",
set_profile_algorithm, NULL, NULL,
"Set algorithm for profile"),
OPT_WITH_ARG("--profile-device",
set_profile_devices, NULL, NULL,
"Select devices to use with profile, one value, range and/or comma separated (e.g. 0-2,4) default: all"),
OPT_WITH_ARG("--profile-kernelfile",
set_profile_kernelfile, NULL, NULL,
"Set the profile's algorithm kernel source file (without file extension)."),
OPT_WITH_ARG("--profile-lookup-gap",
set_profile_lookup_gap, NULL, NULL,
"Set Profile GPU lookup gap for scrypt mining, comma separated"),
#ifdef HAVE_ADL
OPT_WITH_ARG("--profile-gpu-engine",
set_profile_gpu_engine, NULL, NULL,
"Profile GPU engine (over)clock range in Mhz - one value, range and/or comma separated list (e.g. 850-900,900,750-850)"),
OPT_WITH_ARG("--profile-gpu-fan",
set_profile_gpu_fan, NULL, NULL,
"GPU fan for profile"),
OPT_WITH_ARG("--profile-gpu-memclock",
set_profile_gpu_memclock, NULL, NULL,
"Set the Profile GPU memory (over)clock in Mhz - one value for all or separate by commas for per card"),
OPT_WITH_ARG("--profile-gpu-powertune",
set_profile_gpu_powertune, NULL, NULL,
"Set the Profile GPU powertune percentage - one value for all or separate by commas for per card"),
OPT_WITH_ARG("--profile-gpu-threads",
set_profile_gpu_threads, NULL, NULL,
"Number of threads per GPU for profile"),
OPT_WITH_ARG("--profile-gpu-vddc",
set_profile_gpu_vddc, NULL, NULL,
"Set the Profile GPU voltage in Volts - one value for all or separate by commas for per card"),
#endif
OPT_WITH_ARG("--profile-intensity",
set_profile_intensity, NULL, NULL,
"Intensity of GPU scanning (profile-specific)"),
OPT_WITH_ARG("--profile-name",
set_profile_name, NULL, NULL,
"Profile Name"),
OPT_WITH_ARG("--profile-nfactor",
set_profile_nfactor, NULL, NULL,
"Set N-factor for profile"),
OPT_WITH_ARG("--profile-rawintensity",
set_profile_rawintensity, NULL, NULL,
"Raw intensity of GPU scanning (profile-specific)"),
OPT_WITH_ARG("--profile-shaders",
set_profile_shaders, NULL, NULL,
"Profile GPU shaders per card for tuning scrypt, comma separated"),
OPT_WITH_ARG("--profile-thread-concurrency",
set_profile_thread_concurrency, NULL, NULL,
"Set thread concurrency for profile"),
OPT_WITH_ARG("--profile-worksize",
set_profile_worksize, NULL, NULL,
"Override detected optimal worksize for profile - one value or comma separated list"),
OPT_WITH_ARG("--profile-xintensity",
set_profile_xintensity, NULL, NULL,
"Shader based intensity of GPU scanning (profile-specific)"),
OPT_WITHOUT_ARG("--protocol-dump|-P",
opt_set_bool, &opt_protocol,
"Verbose dump of protocol-level activities"),
OPT_WITH_ARG("--queue|-Q",
set_int_0_to_9999, opt_show_intval, &opt_queue,
"Minimum number of work items to have queued (0+)"),
OPT_WITHOUT_ARG("--quiet|-q",
opt_set_bool, &opt_quiet,
"Disable logging output, display status and errors"),
OPT_WITH_ARG("--quota|--pool-quota|-U",
set_quota, NULL, NULL,
"quota;URL combination for server with load-balance strategy quotas"),
OPT_WITHOUT_ARG("--real-quiet",
opt_set_bool, &opt_realquiet,
"Disable all output"),
OPT_WITHOUT_ARG("--remove-disabled",
opt_set_bool, &opt_removedisabled,
"Remove disabled devices entirely, as if they didn't exist"),
OPT_WITH_ARG("--retries",
set_null, NULL, NULL,
opt_hidden),
OPT_WITH_ARG("--retry-pause",
set_null, NULL, NULL,
opt_hidden),
OPT_WITH_ARG("--rotate",
set_rotate, opt_show_intval, &opt_rotate_period,
"Change multipool strategy from failover to regularly rotate at N minutes"),
OPT_WITHOUT_ARG("--round-robin",
set_rr, &pool_strategy,
"Change multipool strategy from failover to round robin on failure"),
OPT_WITH_ARG("--scan-time|-s",
set_int_0_to_9999, opt_show_intval, &opt_scantime,
"Upper bound on time spent scanning current work, in seconds"),
OPT_WITH_ARG("--sched-start",
set_schedtime, NULL, &schedstart,
"Set a time of day in HH:MM to start mining (a once off without a stop time)"),
OPT_WITH_ARG("--sched-stop",
set_schedtime, NULL, &schedstop,
"Set a time of day in HH:MM to stop mining (will quit without a start time)"),
OPT_WITH_ARG("--shaders",
set_default_shaders, NULL, NULL,
"GPU shaders per card for tuning scrypt, comma separated"),
OPT_WITH_ARG("--sharelog",
set_sharelog, NULL, NULL,
"Append share log to file"),
OPT_WITH_ARG("--shares",
opt_set_intval, NULL, &opt_shares,
"Quit after mining N shares (default: unlimited)"),
OPT_WITH_ARG("--socks-proxy",
opt_set_charp, NULL, &opt_socks_proxy,
"Set socks4 proxy (host:port)"),
OPT_WITHOUT_ARG("--show-coindiff",
opt_set_bool, &opt_show_coindiff,
"Show coin difficulty rather than hash value of a share"),
OPT_WITH_ARG("--state|--pool-state",
set_pool_state, NULL, NULL,
"Specify pool state at startup (default: enabled)"),
OPT_WITH_ARG("--switcher-mode",
set_switcher_mode, NULL, NULL,
"Algorithm/gpu settings switcher mode."),
#ifdef HAVE_SYSLOG_H
OPT_WITHOUT_ARG("--syslog",
opt_set_bool, &use_syslog,
"Use system log for output messages (default: standard error)"),
#endif
#if defined(HAVE_LIBCURL) && defined(CURL_HAS_KEEPALIVE)
OPT_WITH_ARG("--tcp-keepalive",
set_int_0_to_9999, opt_show_intval, &opt_tcp_keepalive,
"TCP keepalive packet idle time"),
#else
OPT_WITH_ARG("--tcp-keepalive",
set_int_0_to_9999, opt_show_intval, &opt_tcp_keepalive,
opt_hidden),
#endif
#ifdef HAVE_ADL
OPT_WITH_ARG("--temp-cutoff",
set_temp_cutoff, opt_show_intval, &opt_cutofftemp,
"Temperature which a device will be automatically disabled at, one value or comma separated list"),
OPT_WITH_ARG("--temp-hysteresis",
set_int_1_to_10, opt_show_intval, &opt_hysteresis,
"Set how much the temperature can fluctuate outside limits when automanaging speeds"),
OPT_WITH_ARG("--temp-overheat",
set_temp_overheat, opt_show_intval, &opt_overheattemp,
"Temperature which a device will be throttled at while automanaging fan and/or GPU, one value or comma separated list"),
OPT_WITH_ARG("--temp-target",
set_temp_target, opt_show_intval, &opt_targettemp,
"Temperature which a device should stay at while automanaging fan and/or GPU, one value or comma separated list"),
#endif
#ifdef HAVE_CURSES
OPT_WITHOUT_ARG("--text-only|-T",
opt_set_invbool, &use_curses,
"Disable ncurses formatted screen output"),
#else
OPT_WITHOUT_ARG("--text-only|-T",
opt_set_invbool, &use_curses,
opt_hidden),
#endif
OPT_WITH_ARG("--thread-concurrency",
set_default_thread_concurrency, NULL, NULL,
"Set GPU thread concurrency for scrypt mining, comma separated"),
OPT_WITH_ARG("--url|--pool-url|-o",
set_url, NULL, NULL,
"URL for bitcoin JSON-RPC server"),
OPT_WITH_ARG("--user|--pool-user|-u",
set_user, NULL, NULL,
"Username for bitcoin JSON-RPC server"),
OPT_WITH_ARG("--vectors",
set_vector, NULL, NULL,
opt_hidden),
/* All current kernels only support vectors=1 */
/* "Override detected optimal vector (1, 2 or 4) - one value or comma separated list"), */
OPT_WITHOUT_ARG("--verbose|-v",
opt_set_bool, &opt_verbose,
"Log verbose output to stderr as well as status output"),
OPT_WITH_ARG("--vote",
set_int_1_to_65535, opt_show_intval, &opt_vote,
"Optional vote value for decred blocks"),
OPT_WITH_ARG("--watchpool-refresh",
set_int_1_to_65535, opt_show_intval, &opt_watchpool_refresh,
"Interval in seconds to refresh pool status"),
OPT_WITH_ARG("--worksize|-w",
set_default_worksize, NULL, NULL,
"Override detected optimal worksize - one value or comma separated list"),
OPT_WITH_ARG("--userpass|--pool-userpass|-O",
set_userpass, NULL, NULL,
"Username:Password pair for bitcoin JSON-RPC server"),
OPT_WITHOUT_ARG("--worktime",
opt_set_bool, &opt_worktime,
"Display extra work time debug information"),
OPT_WITH_ARG("--pools",
opt_set_bool, NULL, NULL, opt_hidden),
OPT_WITH_ARG("--profiles",
opt_set_bool, NULL, NULL, opt_hidden),
OPT_WITH_ARG("--includes",
opt_set_bool, NULL, NULL, opt_hidden),
OPT_WITH_ARG("--events",
opt_set_bool, NULL, NULL, opt_hidden),
OPT_WITH_ARG("--difficulty-multiplier",
set_difficulty_multiplier, NULL, NULL,
"(deprecated) Difficulty multiplier for jobs received from stratum pools"),
OPT_ENDTABLE
};
void default_save_file(char *filename);
extern const char *opt_argv0;
static char *opt_verusage_and_exit(const char *extra)
{
printf("%s\n", packagename);
printf("%s", opt_usage(opt_argv0, extra));
fflush(stdout);
exit(0);
}
char *display_devs(int *ndevs)
{
*ndevs = 0;
print_ndevs(ndevs);
exit(*ndevs);
}
/* These options are available from commandline only */
static struct opt_table opt_cmdline_table[] = {
OPT_WITH_ARG("--config|-c",
load_config, NULL, NULL,
"Load a JSON-format configuration file\n"
"See example.conf for an example configuration."),
OPT_WITH_ARG("--default-config",
set_default_config, NULL, NULL,
"Specify the filename of the default config file\n"
"Loaded at start and used when saving without a name."),
OPT_WITH_ARG("--remote-config-retry",
set_int_0_to_9999, opt_show_intval, &opt_remoteconf_retry,
"Number of times to retry downloading remote config file. Default: 3"),
OPT_WITH_ARG("--remote-config-wait",
set_int_0_to_9999, opt_show_intval, &opt_remoteconf_wait,
"Time in seconds to wait between download retries of remote config files. Default: 10secs"),
OPT_WITHOUT_ARG("--remote-config-usecache",
opt_set_bool, &opt_remoteconf_usecache,
"Use cached copy of the remote config file when download fails. Default: No"),
OPT_WITHOUT_ARG("--help|-h",
opt_verusage_and_exit, NULL,
"Print this message"),
OPT_WITHOUT_ARG("--ndevs|-n",
display_devs, &nDevs,
"Display number of detected GPUs, OpenCL platform "
"information, and exit"),
OPT_WITHOUT_ARG("--version|-V",
opt_version_and_exit, packagename,
"Display version and exit"),
OPT_ENDTABLE
};
#ifdef HAVE_LIBCURL
static bool jobj_binary(const json_t *obj, const char *key,
void *buf, size_t buflen, bool required)
{
const char *hexstr;
json_t *tmp;
tmp = json_object_get(obj, key);
if (unlikely(!tmp)) {
if (unlikely(required))
if (opt_morenotices)
applog(LOG_ERR, "JSON key '%s' not found", key);
return false;
}
hexstr = json_string_value(tmp);
if (unlikely(!hexstr)) {
applog(LOG_ERR, "JSON key '%s' is not a string", key);
return false;
}
if (!hex2bin((unsigned char *)buf, hexstr, buflen))
return false;
return true;
}
#endif
static struct work *make_work(void)
{
struct work *w = (struct work *)calloc(1, sizeof(struct work));
if (unlikely(!w))
quit(1, "Failed to calloc work in make_work");
cg_wlock(&control_lock);
w->id = total_work++;
cg_wunlock(&control_lock);
return w;
}
/* This is the central place all work that is about to be retired should be
* cleaned to remove any dynamically allocated arrays within the struct */
void clean_work(struct work *w)
{
free(w->job_id);
free(w->ntime);
free(w->coinbase);
free(w->nonce1);
memset(w, 0, sizeof(struct work));
}
/* All dynamically allocated work structs should be freed here to not leak any
* ram from arrays allocated within the work struct */
void free_work(struct work *w)
{
clean_work(w);
free(w);
}
static void calc_diff(struct work *work, double known);
char *workpadding = "000000800000000000000000000000000000000000000000000000000000000000000000000000000000000080020000";
#ifdef HAVE_LIBCURL
/* Process transactions with GBT by storing the binary value of the first
* transaction, and the hashes of the remaining transactions since these
* remain constant with an altered coinbase when generating work. Must be
* entered under gbt_lock */
static bool __build_gbt_txns(struct pool *pool, json_t *res_val)
{
json_t *txn_array;
bool ret = false;
size_t cal_len;
int i;
free(pool->txn_hashes);
pool->txn_hashes = NULL;
pool->gbt_txns = 0;
txn_array = json_object_get(res_val, "transactions");
if (!json_is_array(txn_array))
goto out;
ret = true;
pool->gbt_txns = json_array_size(txn_array);
if (!pool->gbt_txns)
goto out;
pool->txn_hashes = (unsigned char *)calloc(32 * (pool->gbt_txns + 1), 1);
if (unlikely(!pool->txn_hashes))
quit(1, "Failed to calloc txn_hashes in __build_gbt_txns");
for (i = 0; i < pool->gbt_txns; i++) {
json_t *txn_val = json_object_get(json_array_get(txn_array, i), "data");
const char *txn = json_string_value(txn_val);
size_t txn_len = strlen(txn);
unsigned char *txn_bin;
cal_len = txn_len;
align_len(&cal_len);
txn_bin = (unsigned char *)calloc(cal_len, 1);
if (unlikely(!txn_bin))
quit(1, "Failed to calloc txn_bin in __build_gbt_txns");
if (unlikely(!hex2bin(txn_bin, txn, txn_len / 2)))
quit(1, "Failed to hex2bin txn_bin");
gen_hash(txn_bin, txn_len / 2, pool->txn_hashes + (32 * i));
free(txn_bin);
}
out:
return ret;
}
static unsigned char *__gbt_merkleroot(struct pool *pool)
{
unsigned char *merkle_hash;
int i, txns;
merkle_hash = (unsigned char *)calloc(32 * (pool->gbt_txns + 2), 1);
if (unlikely(!merkle_hash))
quit(1, "Failed to calloc merkle_hash in __gbt_merkleroot");
gen_hash(pool->coinbase, pool->coinbase_len, merkle_hash);
if (pool->gbt_txns)
memcpy(merkle_hash + 32, pool->txn_hashes, pool->gbt_txns * 32);
txns = pool->gbt_txns + 1;
while (txns > 1) {
if (txns % 2) {
memcpy(&merkle_hash[txns * 32], &merkle_hash[(txns - 1) * 32], 32);
txns++;
}
for (i = 0; i < txns; i += 2){
unsigned char hashout[32];
gen_hash(merkle_hash + (i * 32), 64, hashout);
memcpy(merkle_hash + (i / 2 * 32), hashout, 32);
}
txns /= 2;
}
return merkle_hash;
}
static bool work_decode(struct pool *pool, struct work *work, json_t *val);
static void update_gbt(struct pool *pool)
{
int rolltime;
json_t *val;
CURL *curl;
char curl_err_str[CURL_ERROR_SIZE];
curl = curl_easy_init();
if (unlikely(!curl))
quit (1, "CURL initialisation failed in update_gbt");
val = json_rpc_call(curl, curl_err_str, pool->rpc_url, pool->rpc_userpass,
pool->rpc_req, true, false, &rolltime, pool, false);
if (val) {
struct work *work = make_work();
bool rc = work_decode(pool, work, val);
total_getworks++;
pool->getwork_requested++;
if (rc) {
applog(LOG_DEBUG, "Successfully retrieved and updated GBT from %s", get_pool_name(pool));
cgtime(&pool->tv_idle);
if (pool == current_pool())
opt_work_update = true;
} else {
applog(LOG_DEBUG, "Successfully retrieved but FAILED to decipher GBT from %s", get_pool_name(pool));
}
json_decref(val);
free_work(work);
} else {
applog(LOG_DEBUG, "FAILED to update GBT from %s", get_pool_name(pool));
}
curl_easy_cleanup(curl);
}
/* Return the work coin/network difficulty */
static double get_work_blockdiff(const struct work *work)
{
uint32_t* data = (uint32_t*) work->data;
uint64_t diff64;
double numerator;
int powdiff;
uint8_t shift;
// Neoscrypt has the data reversed
if (work->pool->algorithm.type == ALGO_NEOSCRYPT) {
diff64 = bswap_64(((uint64_t)(be32toh(*((uint32_t *)(work->data + 72))) & 0xFFFFFF00)) << 8);
numerator = (double)work->pool->algorithm.diff_numerator;
}
else if (work->pool->algorithm.type == ALGO_DECRED) {
shift = work->data[116+3];
powdiff = (8 * (0x1d - 3)) - (8 * (shift - 3));
diff64 = data[29] & 0xFFFFFF;
if (!diff64) diff64 = 1;
double d = (double)work->pool->algorithm.diff_numerator / (double)diff64;
for (int m = shift; m < 29; m++) d *= 256.0;
for (int m = 29; m < shift; m++) d /= 256.0;
if (shift == 28) d *= 256.0; // testnet
return d;
}
else {
shift = work->data[72];
powdiff = (8 * (0x1d - 3)) - (8 * (shift - 3));;
diff64 = be32toh(*((uint32_t *)(work->data + 72))) & 0x0000000000FFFFFF;
numerator = work->pool->algorithm.diff_numerator << powdiff;
}
if (unlikely(!diff64)) {
diff64 = 1;
}
return numerator / (double)diff64;
}
static void gen_gbt_work(struct pool *pool, struct work *work)
{
unsigned char *merkleroot;
struct timeval now;
uint64_t nonce2le;
cgtime(&now);
if (now.tv_sec - pool->tv_lastwork.tv_sec > 60)
update_gbt(pool);
cg_wlock(&pool->gbt_lock);
nonce2le = htole64(pool->nonce2);
memcpy(pool->coinbase + pool->nonce2_offset, &nonce2le, pool->n2size);
pool->nonce2++;
cg_dwlock(&pool->gbt_lock);
merkleroot = __gbt_merkleroot(pool);
memcpy(work->data, &pool->gbt_version, 4);
memcpy(work->data + 4, pool->previousblockhash, 32);
memcpy(work->data + 4 + 32 + 32, &pool->curtime, 4);
memcpy(work->data + 4 + 32 + 32 + 4, &pool->gbt_bits, 4);
memcpy(work->target, pool->gbt_target, 32);
work->coinbase = bin2hex(pool->coinbase, pool->coinbase_len);
/* For encoding the block data on submission */
work->gbt_txns = pool->gbt_txns + 1;
if (pool->gbt_workid)
work->job_id = strdup(pool->gbt_workid);
cg_runlock(&pool->gbt_lock);
flip32(work->data + 4 + 32, merkleroot);
free(merkleroot);
memset(work->data + 4 + 32 + 32 + 4 + 4, 0, 4); /* nonce */
hex2bin(work->data + 4 + 32 + 32 + 4 + 4 + 4, workpadding, 48);
if (opt_debug) {
char *header = bin2hex(work->data, 128);
applog(LOG_DEBUG, "Generated GBT header %s", header);
applog(LOG_DEBUG, "Work coinbase %s", work->coinbase);
free(header);
}
if (pool->algorithm.calc_midstate) pool->algorithm.calc_midstate(work);
local_work++;
work->pool = pool;
work->gbt = true;
work->id = total_work++;
work->longpoll = false;
work->getwork_mode = GETWORK_MODE_GBT;
work->work_block = work_block;
/* Nominally allow a driver to ntime roll 60 seconds */
work->drv_rolllimit = 60;
calc_diff(work, 0);
cgtime(&work->tv_staged);
}
static bool gbt_decode(struct pool *pool, json_t *res_val)
{
const char *previousblockhash;
const char *target;
const char *coinbasetxn;
const char *longpollid;
unsigned char hash_swap[32];
int expires;
int version;
int curtime;
bool submitold;
const char *bits;
const char *workid;
size_t cbt_len, orig_len;
uint8_t *extra_len;
size_t cal_len;
previousblockhash = json_string_value(json_object_get(res_val, "previousblockhash"));
target = json_string_value(json_object_get(res_val, "target"));
coinbasetxn = json_string_value(json_object_get(json_object_get(res_val, "coinbasetxn"), "data"));
longpollid = json_string_value(json_object_get(res_val, "longpollid"));
expires = json_integer_value(json_object_get(res_val, "expires"));
version = json_integer_value(json_object_get(res_val, "version"));
curtime = json_integer_value(json_object_get(res_val, "curtime"));
submitold = json_is_true(json_object_get(res_val, "submitold"));
bits = json_string_value(json_object_get(res_val, "bits"));
workid = json_string_value(json_object_get(res_val, "workid"));
if (!previousblockhash || !target || !coinbasetxn || !longpollid ||
!expires || !version || !curtime || !bits) {
applog(LOG_ERR, "JSON failed to decode GBT");
return false;
}
applog(LOG_DEBUG, "previousblockhash: %s", previousblockhash);
applog(LOG_DEBUG, "target: %s", target);
applog(LOG_DEBUG, "coinbasetxn: %s", coinbasetxn);
applog(LOG_DEBUG, "longpollid: %s", longpollid);
applog(LOG_DEBUG, "expires: %d", expires);
applog(LOG_DEBUG, "version: %d", version);
applog(LOG_DEBUG, "curtime: %d", curtime);
applog(LOG_DEBUG, "submitold: %s", submitold ? "true" : "false");
applog(LOG_DEBUG, "bits: %s", bits);
if (workid)
applog(LOG_DEBUG, "workid: %s", workid);
cg_wlock(&pool->gbt_lock);
free(pool->coinbasetxn);
pool->coinbasetxn = strdup(coinbasetxn);
cbt_len = strlen(pool->coinbasetxn) / 2;
/* We add 8 bytes of extra data corresponding to nonce2 */
pool->n2size = 8;
pool->coinbase_len = cbt_len + pool->n2size;
cal_len = pool->coinbase_len + 1;
align_len(&cal_len);
free(pool->coinbase);
pool->coinbase = (unsigned char *)calloc(cal_len, 1);
if (unlikely(!pool->coinbase))
quit(1, "Failed to calloc pool coinbase in gbt_decode");
hex2bin(pool->coinbase, pool->coinbasetxn, 42);
extra_len = (uint8_t *)(pool->coinbase + 41);
orig_len = *extra_len;
hex2bin(pool->coinbase + 42, pool->coinbasetxn + 84, orig_len);
*extra_len += pool->n2size;
hex2bin(pool->coinbase + 42 + *extra_len, pool->coinbasetxn + 84 + (orig_len * 2),
cbt_len - orig_len - 42);
pool->nonce2_offset = orig_len + 42;
free(pool->longpollid);
pool->longpollid = strdup(longpollid);
free(pool->gbt_workid);
if (workid)
pool->gbt_workid = strdup(workid);
else
pool->gbt_workid = NULL;
hex2bin(hash_swap, previousblockhash, 32);
swap256(pool->previousblockhash, hash_swap);
hex2bin(hash_swap, target, 32);
swab256(pool->gbt_target, hash_swap);
pool->gbt_expires = expires;
pool->gbt_version = htobe32(version);
pool->curtime = htobe32(curtime);
pool->submit_old = submitold;
hex2bin((unsigned char *)&pool->gbt_bits, bits, 4);
__build_gbt_txns(pool, res_val);
cg_wunlock(&pool->gbt_lock);
return true;
}
static bool getwork_decode(json_t *res_val, struct work *work)
{
size_t worklen = 128;
if (work->pool->algorithm.type == ALGO_CRE) worklen = 168;
else if (work->pool->algorithm.type == ALGO_DECRED) worklen = 192;
if (unlikely(!jobj_binary(res_val, "data", work->data, worklen, true))) {
if (opt_morenotices)
applog(LOG_ERR, "%s: JSON inval data", isnull(get_pool_name(work->pool), ""));
return false;
}
if (work->pool->algorithm.type == ALGO_CRE || work->pool->algorithm.type == ALGO_SCRYPT) {
if (!jobj_binary(res_val, "midstate", work->midstate, sizeof(work->midstate), false)) {
// Calculate it ourselves
if (opt_morenotices) {
applog(LOG_DEBUG, "%s: Calculating midstate locally", isnull(get_pool_name(work->pool), ""));
}
if (work->pool->algorithm.calc_midstate) work->pool->algorithm.calc_midstate(work);
}
}
if (unlikely(!jobj_binary(res_val, "target", work->target, sizeof(work->target), true))) {
if (opt_morenotices)
applog(LOG_ERR, "%s: JSON inval target", isnull(get_pool_name(work->pool), ""));
return false;
}
if (work->pool->algorithm.type == ALGO_DECRED) {
uint16_t vote = (uint16_t) (opt_vote << 1) | 1;
memcpy(&work->data[100], &vote, 2);
// some random extradata to make it unique
((uint32_t*)work->data)[36] = (rand()*4);
((uint32_t*)work->data)[37] = (rand()*4) << 8 | work->thr_id;
}
return true;
}
/* Returns whether the pool supports local work generation or not. */
static bool pool_localgen(struct pool *pool)
{
return (pool->has_stratum || pool->has_gbt);
}
static bool work_decode(struct pool *pool, struct work *work, json_t *val)
{
json_t *res_val = json_object_get(val, "result");
bool ret = false;
cgtime(&pool->tv_lastwork);
if (!res_val || json_is_null(res_val)) {
applog(LOG_ERR, "JSON Failed to decode result");
goto out;
}
work->pool = pool;
if (pool->has_gbt) {
if (unlikely(!gbt_decode(pool, res_val)))
goto out;
work->gbt = true;
ret = true;
goto out;
} else if (unlikely(!getwork_decode(res_val, work)))
goto out;
memset(work->hash, 0, sizeof(work->hash));
cgtime(&work->tv_staged);
ret = true;
out:
return ret;
}
#else /* HAVE_LIBCURL */
/* Always true with stratum */
#define pool_localgen(pool) (true)
#define json_rpc_call(curl, curl_err_str, url, userpass, rpc_req, probe, longpoll, rolltime, pool, share) (NULL)
#define work_decode(pool, work, val) (false)
#define gen_gbt_work(pool, work) {}
#endif /* HAVE_LIBCURL */
int dev_from_id(int thr_id)
{
struct cgpu_info *cgpu = get_thr_cgpu(thr_id);
return cgpu->device_id;
}
/* Create an exponentially decaying average over the opt_log_interval */
void decay_time(double *f, double fadd, double fsecs)
{
double ftotal, fprop;
fprop = 1.0 - 1 / (exp(fsecs / (double)opt_log_interval));
ftotal = 1.0 + fprop;
*f += (fadd * fprop);
*f /= ftotal;
}
static int __total_staged(void)
{
return HASH_COUNT(staged_work);
}
static int total_staged(void)
{
int ret;
mutex_lock(stgd_lock);
ret = __total_staged();
mutex_unlock(stgd_lock);
return ret;
}
#ifdef HAVE_CURSES
WINDOW *mainwin, *statuswin, *logwin;
#endif
double total_secs = 1.0;
static char statusline[256];
/* logstart is where the log window should start */
static int devcursor, logstart, logcursor;
#ifdef HAVE_CURSES
/* statusy is where the status window goes up to in cases where it won't fit at startup */
static int statusy;
#endif
extern struct cgpu_info gpus[MAX_GPUDEVICES]; /* Maximum number apparently possible */
#ifdef HAVE_CURSES
static inline void unlock_curses(void)
{
mutex_unlock(&console_lock);
}
static inline void lock_curses(void)
{
mutex_lock(&console_lock);
}
static bool curses_active_locked(void)
{
bool ret;
lock_curses();
ret = curses_active;
if (!ret)
unlock_curses();
return ret;
}
#endif
/* Convert a uint64_t value into a truncated string for displaying with its
* associated suitable for Mega, Giga etc. Buf array needs to be long enough */
static void suffix_string(uint64_t val, char *buf, size_t bufsiz, int sigdigits)
{
const double dkilo = 1000.0;
const uint64_t kilo = 1000ull;
const uint64_t mega = 1000000ull;
const uint64_t giga = 1000000000ull;
const uint64_t tera = 1000000000000ull;
const uint64_t peta = 1000000000000000ull;
const uint64_t exa = 1000000000000000000ull;
char suffix[2] = "";
bool decimal = true;
double dval;
if (val >= exa) {
val /= peta;
dval = (double)val / dkilo;
strcpy(suffix, "E");
} else if (val >= peta) {
val /= tera;
dval = (double)val / dkilo;
strcpy(suffix, "P");
} else if (val >= tera) {
val /= giga;
dval = (double)val / dkilo;
strcpy(suffix, "T");
} else if (val >= giga) {
val /= mega;
dval = (double)val / dkilo;
strcpy(suffix, "G");
} else if (val >= mega) {
val /= kilo;
dval = (double)val / dkilo;
strcpy(suffix, "M");
} else if (val >= kilo) {
dval = (double)val / dkilo;
strcpy(suffix, "K");
} else {
dval = val;
decimal = false;
}
if (!sigdigits) {
if (decimal)
snprintf(buf, bufsiz, "%.3g%s", dval, suffix);
else
snprintf(buf, bufsiz, "%d%s", (unsigned int)dval, suffix);
} else {
/* Always show sigdigits + 1, padded on right with zeroes
* followed by suffix */
int ndigits = sigdigits - 1 - (dval > 0.0 ? (int)floor(log10(dval)) : 0);
snprintf(buf, bufsiz, "%*.*f%s", sigdigits + 1, ndigits, dval, suffix);
}
}
/* Convert a double value into a truncated string for displaying with its
* associated suitable for Mega, Giga etc. Buf array needs to be long enough */
static void suffix_string_double(double val, char *buf, size_t bufsiz, int sigdigits)
{
if (val < 10) {
snprintf(buf, bufsiz, "%.3f", val);
} else {
return suffix_string(val, buf, bufsiz, sigdigits);
}
}
double cgpu_runtime(struct cgpu_info *cgpu)
{
struct timeval now;
double dev_runtime;
if (cgpu->dev_start_tv.tv_sec == 0)
dev_runtime = total_secs;
else {
cgtime(&now);
dev_runtime = tdiff(&now, &(cgpu->dev_start_tv));
}
if (dev_runtime < 1.0)
dev_runtime = 1.0;
return dev_runtime;
}
static void get_statline(char *buf, size_t bufsiz, struct cgpu_info *cgpu)
{
char displayed_hashes[16], displayed_rolling[16];
double dev_runtime, wu;
uint64_t dh64, dr64;
dev_runtime = cgpu_runtime(cgpu);
wu = cgpu->diff1 / dev_runtime * 60.0;
dh64 = (double)cgpu->total_mhashes / dev_runtime * 1000000ull;
dr64 = (double)cgpu->rolling * 1000000ull;
suffix_string(dh64, displayed_hashes, sizeof(displayed_hashes), 4);
suffix_string(dr64, displayed_rolling, sizeof(displayed_rolling), 4);
snprintf(buf, bufsiz, "%s%d ", cgpu->drv->name, cgpu->device_id);
cgpu->drv->get_statline_before(buf, bufsiz, cgpu);
tailsprintf(buf, bufsiz, "(%ds):%s (avg):%sh/s | A:%.0f R:%.0f HW:%d WU:%.3f/m",
opt_log_interval,
displayed_rolling,
displayed_hashes,
cgpu->diff_accepted,
cgpu->diff_rejected,
cgpu->hw_errors,
wu);
cgpu->drv->get_statline(buf, bufsiz, cgpu);
}
static bool shared_strategy(void)
{
return (pool_strategy == POOL_LOADBALANCE || pool_strategy == POOL_BALANCE);
}
#ifdef HAVE_CURSES
#define CURBUFSIZ 256
#define cg_mvwprintw(win, y, x, fmt, ...) do { \
char tmp42[CURBUFSIZ]; \
snprintf(tmp42, sizeof(tmp42), fmt, ##__VA_ARGS__); \
mvwprintw(win, y, x, "%s", tmp42); \
} while (0)
#define cg_wprintw(win, fmt, ...) do { \
char tmp42[CURBUFSIZ]; \
snprintf(tmp42, sizeof(tmp42), fmt, ##__VA_ARGS__); \
wprintw(win, "%s", tmp42); \
} while (0)
/* Must be called with curses mutex lock held and curses_active */
static void curses_print_uptime(struct timeval *start_time)
{
struct timeval now, tv;
unsigned int days, hours;
div_t d;
cgtime(&now);
timersub(&now, start_time, &tv);
d = div((int)tv.tv_sec, 86400);
days = d.quot;
d = div(d.rem, 3600);
hours = d.quot;
d = div(d.rem, 60);
cg_wprintw(statuswin, " - [%u day%c %02d:%02d:%02d]"
, days
, (days == 1) ? ' ' : 's'
, hours
, d.quot
, d.rem
);
}
/* Must be called with curses mutex lock held and curses_active */
static void curses_print_status(void)
{
struct pool *pool = current_pool();
unsigned short int line = 0;
wattron(statuswin, A_BOLD);
cg_mvwprintw(statuswin, line, 0, PACKAGE " " CGMINER_VERSION " - Started: %s", datestamp);
curses_print_uptime(&launch_time);
wattroff(statuswin, A_BOLD);
mvwhline(statuswin, ++line, 0, '-', 80);
cg_mvwprintw(statuswin, ++line, 0, "%s", statusline);
wclrtoeol(statuswin);
cg_mvwprintw(statuswin, ++line, 0, "ST: %d SS: %d NB: %d LW: %d GF: %d RF: %d",
total_staged(), total_stale, new_blocks,
local_work, total_go, total_ro);
wclrtoeol(statuswin);
if (shared_strategy() && total_pools > 1) {
cg_mvwprintw(statuswin, ++line, 0, "Connected to multiple pools %s block change notify",
have_longpoll ? "with": "without");
} else {
cg_mvwprintw(statuswin, ++line, 0, "Connected to %s (%s) diff %s as user %s",
get_pool_name(pool),
pool->has_stratum ? "stratum" : (pool->has_gbt ? "GBT" : "longpoll"),
pool->diff,
get_pool_user(pool));
}
wclrtoeol(statuswin);
cg_mvwprintw(statuswin, ++line, 0, "Block: %s... Diff:%s Started: %s Best share: %s ",
prev_block, block_diff, blocktime, best_share);
mvwhline(statuswin, ++line, 0, '-', 80);
mvwhline(statuswin, statusy - 1, 0, '-', 80);
cg_mvwprintw(statuswin, devcursor - 1, 0, "[P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit");
}
static void adj_width(int var, int *length)
{
if ((int)(log10((double)var) + 1) > *length)
(*length)++;
}
static int dev_width;
static void curses_print_devstatus(struct cgpu_info *cgpu, int count)
{
static int drwidth = 5, hwwidth = 1, wuwidth = 1;
char logline[256];
char displayed_hashes[16], displayed_rolling[16];
float reject_pct = 0.0;
uint64_t dh64, dr64;
struct timeval now;
double dev_runtime, wu;
/* Do not print if window vertical size too small. */
if (devcursor + count > LINES - 2)
return;
if (count >= (opt_removedisabled ? most_devices : total_devices))
return;
if (cgpu->dev_start_tv.tv_sec == 0)
dev_runtime = total_secs;
else {
cgtime(&now);
dev_runtime = tdiff(&now, &(cgpu->dev_start_tv));
}
if (dev_runtime < 1.0)
dev_runtime = 1.0;
cgpu->utility = cgpu->accepted / dev_runtime * 60;
wu = cgpu->diff1 / dev_runtime * 60;
wmove(statuswin, devcursor + count, 0);
cg_wprintw(statuswin, "%s %*d: ", cgpu->drv->name, dev_width, cgpu->device_id);
logline[0] = '\0';
cgpu->drv->get_statline_before(logline, sizeof(logline), cgpu);
cg_wprintw(statuswin, "%s", logline);
dh64 = (double)cgpu->total_mhashes / dev_runtime * 1000000ull;
dr64 = (double)cgpu->rolling * 1000000ull;
suffix_string(dh64, displayed_hashes, sizeof(displayed_hashes), 4);
suffix_string(dr64, displayed_rolling, sizeof(displayed_rolling), 4);
if (cgpu->status == LIFE_DEAD)
cg_wprintw(statuswin, "DEAD ");
else if (cgpu->status == LIFE_SICK)
cg_wprintw(statuswin, "SICK ");
else if (cgpu->deven == DEV_DISABLED)
cg_wprintw(statuswin, "OFF ");
else if (cgpu->deven == DEV_RECOVER)
cg_wprintw(statuswin, "REST ");
else
cg_wprintw(statuswin, "%6s", displayed_rolling);
if ((cgpu->diff_accepted + cgpu->diff_rejected) > 0)
reject_pct = (cgpu->diff_rejected / (cgpu->diff_accepted + cgpu->diff_rejected)) * 100;
adj_width(cgpu->hw_errors, &hwwidth);
adj_width(wu, &wuwidth);
cg_wprintw(statuswin, "/%6sh/s | R:%*.1f%% HW:%*d WU:%*.3f/m",
displayed_hashes,
drwidth, reject_pct,
hwwidth, cgpu->hw_errors,
wuwidth + 2, wu);
logline[0] = '\0';
cgpu->drv->get_statline(logline, sizeof(logline), cgpu);
cg_wprintw(statuswin, "%s", logline);
wclrtoeol(statuswin);
}
#endif
#ifdef HAVE_CURSES
/* Check for window resize. Called with curses mutex locked */
static inline void change_logwinsize(void)
{
int x, y, logx, logy;
getmaxyx(mainwin, y, x);
if (x < 80 || y < 25)
return;
if (y > statusy + 2 && statusy < logstart) {
if (y - 2 < logstart)
statusy = y - 2;
else
statusy = logstart;
logcursor = statusy + 1;
mvwin(logwin, logcursor, 0);
wresize(statuswin, statusy, x);
}
y -= logcursor;
getmaxyx(logwin, logy, logx);
/* Detect screen size change */
if (x != logx || y != logy)
wresize(logwin, y, x);
}
static void check_winsizes(void)
{
if (!use_curses)
return;
if (curses_active_locked()) {
int y, x;
erase();
x = getmaxx(statuswin);
if (logstart > LINES - 2)
statusy = LINES - 2;
else
statusy = logstart;
logcursor = statusy;
wresize(statuswin, statusy, x);
getmaxyx(mainwin, y, x);
y -= logcursor;
wresize(logwin, y, x);
mvwin(logwin, logcursor, 0);
unlock_curses();
}
}
static void disable_curses_windows(void);
static void enable_curses_windows(void);
static void switch_logsize(bool __maybe_unused newdevs)
{
if (curses_active_locked()) {
#ifdef WIN32
if (newdevs)
disable_curses_windows();
#endif
if (opt_compact) {
logstart = devcursor + 1;
} else {
logstart = devcursor + (opt_removedisabled ? most_devices : total_devices) + 1;
}
logcursor = logstart + 1;
#ifdef WIN32
if (newdevs)
enable_curses_windows();
#endif
unlock_curses();
check_winsizes();
}
}
/* For mandatory printing when mutex is already locked */
void _wlog(const char *str)
{
wprintw(logwin, "%s", str);
}
/* Mandatory printing */
void _wlogprint(const char *str)
{
if (curses_active_locked()) {
wprintw(logwin, "%s", str);
unlock_curses();
}
}
#else
static void switch_logsize(bool __maybe_unused newdevs)
{
}
#endif
#ifdef HAVE_CURSES
bool _log_curses_only(int prio, const char *datetime, const char *str)
{
bool high_prio;
high_prio = (prio == LOG_WARNING || prio == LOG_ERR);
if (curses_active) {
if (!opt_loginput || high_prio) {
wprintw(logwin, "%s%s\n", datetime, str);
if (high_prio) {
touchwin(logwin);
wrefresh(logwin);
}
}
return true;
}
return false;
}
void clear_logwin(void)
{
if (curses_active_locked()) {
erase();
wclear(logwin);
unlock_curses();
}
}
void logwin_update(void)
{
if (curses_active_locked()) {
touchwin(logwin);
wrefresh(logwin);
unlock_curses();
}
}
#endif
static void restart_threads(void);
/* Theoretically threads could race when modifying accepted and
* rejected values but the chance of two submits completing at the
* same time is zero so there is no point adding extra locking */
static void
share_result(json_t *val, json_t *res, json_t *err, const struct work *work,
char *hashshow, bool resubmit, char *worktime)
{
struct pool *pool = work->pool;
struct cgpu_info *cgpu;
cgpu = get_thr_cgpu(work->thr_id);
if (json_is_true(res) || (work->gbt && json_is_null(res))) {
mutex_lock(&stats_lock);
cgpu->accepted++;
total_accepted++;
pool->accepted++;
cgpu->diff_accepted += work->work_difficulty;
total_diff_accepted += work->work_difficulty;
pool->diff_accepted += work->work_difficulty;
mutex_unlock(&stats_lock);
pool->seq_rejects = 0;
cgpu->last_share_pool = pool->pool_no;
cgpu->last_share_pool_time = time(NULL);
cgpu->last_share_diff = work->work_difficulty;
pool->last_share_time = cgpu->last_share_pool_time;
pool->last_share_diff = work->work_difficulty;
applog(LOG_DEBUG, "[THR%d] PROOF OF WORK RESULT: true (yay!!!)", work->thr_id);
if (!QUIET) {
if (total_pools > 1) {
applog(LOG_NOTICE, "Accepted %s %s %d at %s %s%s",
hashshow,
cgpu->drv->name,
cgpu->device_id,
get_pool_name(pool),
resubmit ? "(resubmit)" : "", worktime);
} else {
applog(LOG_DEBUG, "[THR%d] Accepted %s %s %d %s%s",
work->thr_id, hashshow, cgpu->drv->name, cgpu->device_id, resubmit ? "(resubmit)" : "", worktime);
applog(LOG_NOTICE, "Accepted %s %s %d %s%s",
hashshow, cgpu->drv->name, cgpu->device_id, resubmit ? "(resubmit)" : "", worktime);
}
}
sharelog("accept", work);
if (opt_shares && total_diff_accepted >= opt_shares) {
applog(LOG_WARNING, "Successfully mined %d accepted shares as requested and exiting.", opt_shares);
kill_work();
return;
}
/* Detect if a pool that has been temporarily disabled for
* continually rejecting shares has started accepting shares.
* This will only happen with the work returned from a
* longpoll */
if (unlikely(pool->state == POOL_REJECTING)) {
applog(LOG_WARNING, "Rejecting %s now accepting shares, re-enabling!",
get_pool_name(pool));
enable_pool(pool);
switch_pools(NULL);
}
/* If we know we found the block we know better than anyone
* that new work is needed. */
if (unlikely(work->block))
restart_threads();
} else {
mutex_lock(&stats_lock);
cgpu->rejected++;
total_rejected++;
pool->rejected++;
cgpu->diff_rejected += work->work_difficulty;
total_diff_rejected += work->work_difficulty;
pool->diff_rejected += work->work_difficulty;
pool->seq_rejects++;
mutex_unlock(&stats_lock);
applog(LOG_DEBUG, "[THR%d] PROOF OF WORK RESULT: false (booooo)", work->thr_id);
if (!QUIET) {
char disposition[36] = "reject";
char reason[32];
strcpy(reason, "");
if (!work->gbt)
res = json_object_get(val, "reject-reason");
if (res) {
const char *reasontmp = json_string_value(res);
size_t reasonLen = strlen(reasontmp);
if (reasonLen > 28)
reasonLen = 28;
reason[0] = ' '; reason[1] = '(';
memcpy(2 + reason, reasontmp, reasonLen);
reason[reasonLen + 2] = ')'; reason[reasonLen + 3] = '\0';
memcpy(disposition + 7, reasontmp, reasonLen);
disposition[6] = ':'; disposition[reasonLen + 7] = '\0';
} else if (work->stratum && err && json_is_array(err)) {
json_t *reason_val = json_array_get(err, 1);
char *reason_str;
if (reason_val && json_is_string(reason_val)) {
reason_str = (char *)json_string_value(reason_val);
snprintf(reason, 31, " (%s)", reason_str);
}
}
applog(LOG_NOTICE, "Rejected %s %s %d %s%s %s%s",
hashshow,
cgpu->drv->name,
cgpu->device_id,
(total_pools > 1) ? get_pool_name(pool) : "",
reason, resubmit ? "(resubmit)" : "",
worktime);
sharelog(disposition, work);
}
/* Once we have more than a nominal amount of sequential rejects,
* at least 10 and more than 3 mins at the current utility,
* disable the pool because some pool error is likely to have
* ensued. Do not do this if we know the share just happened to
* be stale due to networking delays.
*/
if (pool->seq_rejects > 10 && !work->stale && opt_disable_pool && enabled_pools > 1) {
double utility = total_accepted / total_secs * 60;
if (pool->seq_rejects > utility * 3) {
applog(LOG_WARNING, "%s rejected %d sequential shares, disabling!",
get_pool_name(pool),
pool->seq_rejects);
reject_pool(pool);
if (pool == current_pool())
switch_pools(NULL);
pool->seq_rejects = 0;
}
}
}
}
static void show_hash(struct work *work, char *hashshow)
{
unsigned char rhash[32];
char diffdisp[16], wdiffdisp[16];
unsigned long h32;
uint32_t *hash32;
int ofs;
suffix_string_double(work->share_diff, diffdisp, sizeof (diffdisp), 0);
suffix_string_double(work->work_difficulty, wdiffdisp, sizeof (wdiffdisp), 0);
if (opt_show_coindiff) {
snprintf(hashshow, 64, "Coin %.0f Diff %s/%s%s", get_work_blockdiff(work), diffdisp, wdiffdisp,
work->block? " BLOCK!" : "");
} else {
swab256(rhash, work->hash);
for (ofs = 0; ofs <= 28; ++ofs) {
if (rhash[ofs]) {
break;
}
}
hash32 = (uint32_t *)(rhash + ofs);
h32 = be32toh(*hash32);
snprintf(hashshow, 64, "%08lx Diff %s/%s%s", h32, diffdisp, wdiffdisp,
work->block ? " BLOCK!" : "");
}
}
#ifdef HAVE_LIBCURL
static void text_print_status(int thr_id)
{
struct cgpu_info *cgpu;
char logline[256];
cgpu = get_thr_cgpu(thr_id);
if (cgpu) {
get_statline(logline, sizeof(logline), cgpu);
printf("%s\n", logline);
}
}
static void print_status(int thr_id)
{
if (!curses_active)
text_print_status(thr_id);
}
static bool submit_upstream_work(struct work *work, CURL *curl, char *curl_err_str, bool resubmit)
{
char *hexstr = NULL;
json_t *val, *res, *err;
char *s;
bool rc = false;
int thr_id = work->thr_id;
struct cgpu_info *cgpu;
struct pool *pool = work->pool;
int rolltime;
struct timeval tv_submit, tv_submit_reply;
char hashshow[64 + 4] = "";
char worktime[200] = "";
struct timeval now;
double dev_runtime;
cgpu = get_thr_cgpu(thr_id);
if (work->pool->algorithm.type == ALGO_DECRED) {
endian_flip180(work->data, work->data);
} else if (work->pool->algorithm.type == ALGO_CRE) {
endian_flip168(work->data, work->data);
} else {
endian_flip128(work->data, work->data);
}
/* build hex string - Make sure to restrict to 80 bytes for Neoscrypt */
int datasize = 128;
if (work->pool->algorithm.type == ALGO_NEOSCRYPT) datasize = 80;
else if (work->pool->algorithm.type == ALGO_CRE) datasize = 168;
else if (work->pool->algorithm.type == ALGO_DECRED) {
datasize = 192;
((uint32_t*)work->data)[45] = 0x80000001UL;
((uint32_t*)work->data)[46] = 0;
((uint32_t*)work->data)[47] = 0x000005a0UL;
}
hexstr = bin2hex(work->data, datasize);
/* build JSON-RPC request */
if (work->gbt) {
char *gbt_block, *varint;
unsigned char data[80];
flip80(data, work->data);
gbt_block = bin2hex(data, 80);
if (work->gbt_txns < 0xfd) {
uint8_t val = work->gbt_txns;
varint = bin2hex((const unsigned char *)&val, 1);
} else if (work->gbt_txns <= 0xffff) {
uint16_t val = htole16(work->gbt_txns);
gbt_block = (char *)realloc_strcat(gbt_block, "fd");
varint = bin2hex((const unsigned char *)&val, 2);
} else {
uint32_t val = htole32(work->gbt_txns);
gbt_block = (char *)realloc_strcat(gbt_block, "fe");
varint = bin2hex((const unsigned char *)&val, 4);
}
gbt_block = (char *)realloc_strcat(gbt_block, varint);
free(varint);
gbt_block = (char *)realloc_strcat(gbt_block, work->coinbase);
s = strdup("{\"id\": 0, \"method\": \"submitblock\", \"params\": [\"");
s = (char *)realloc_strcat(s, gbt_block);
if (work->job_id) {
s = (char *)realloc_strcat(s, "\", {\"workid\": \"");
s = (char *)realloc_strcat(s, work->job_id);
s = (char *)realloc_strcat(s, "\"}]}");
} else
s = (char *)realloc_strcat(s, "\", {}]}");
free(gbt_block);
} else {
s = strdup("{\"method\": \"getwork\", \"params\": [ \"");
s = (char *)realloc_strcat(s, hexstr);
s = (char *)realloc_strcat(s, "\" ], \"id\":1}");
}
applog(LOG_DEBUG, "DBG: sending %s submit RPC call: %s", pool->rpc_url, s);
s = (char *)realloc_strcat(s, "\n");
cgtime(&tv_submit);
/* issue JSON-RPC request */
val = json_rpc_call(curl, curl_err_str, pool->rpc_url, pool->rpc_userpass, s, false, false, &rolltime, pool, true);
cgtime(&tv_submit_reply);
free(s);
if (unlikely(!val)) {
applog(LOG_INFO, "submit_upstream_work json_rpc_call failed");
if (!pool_tset(pool, &pool->submit_fail)) {
total_ro++;
pool->remotefail_occasions++;
if (opt_lowmem) {
applog(LOG_WARNING, "%s communication failure, discarding shares", get_pool_name(pool));
goto out;
}
applog(LOG_WARNING, "%s communication failure, caching submissions", get_pool_name(pool));
}
cgsleep_ms(5000);
goto out;
} else if (pool_tclear(pool, &pool->submit_fail))
applog(LOG_WARNING, "%s communication resumed, submitting work", get_pool_name(pool));
res = json_object_get(val, "result");
err = json_object_get(val, "error");
if (!QUIET) {
show_hash(work, hashshow);
if (opt_worktime) {
char workclone[20];
struct tm *tm, tm_getwork, tm_submit_reply;
double getwork_time = tdiff((struct timeval *)&(work->tv_getwork_reply),
(struct timeval *)&(work->tv_getwork));
double getwork_to_work = tdiff((struct timeval *)&(work->tv_work_start),
(struct timeval *)&(work->tv_getwork_reply));
double work_time = tdiff((struct timeval *)&(work->tv_work_found),
(struct timeval *)&(work->tv_work_start));
double work_to_submit = tdiff(&tv_submit,
(struct timeval *)&(work->tv_work_found));
double submit_time = tdiff(&tv_submit_reply, &tv_submit);
int diffplaces = 3;
time_t tmp_time = work->tv_getwork.tv_sec;
tm = localtime(&tmp_time);
memcpy(&tm_getwork, tm, sizeof(struct tm));
tmp_time = tv_submit_reply.tv_sec;
tm = localtime(&tmp_time);
memcpy(&tm_submit_reply, tm, sizeof(struct tm));
if (work->clone) {
snprintf(workclone, sizeof(workclone), "C:%1.3f",
tdiff((struct timeval *)&(work->tv_cloned),
(struct timeval *)&(work->tv_getwork_reply)));
}
else
strcpy(workclone, "O");
if (work->work_difficulty < 1)
diffplaces = 6;
snprintf(worktime, sizeof(worktime),
" <-%08lx.%08lx M:%c D:%1.*f G:%02d:%02d:%02d:%1.3f %s (%1.3f) W:%1.3f (%1.3f) S:%1.3f R:%02d:%02d:%02d",
(unsigned long)be32toh(*(uint32_t *)&(work->data[32])),
(unsigned long)be32toh(*(uint32_t *)&(work->data[28])),
work->getwork_mode, diffplaces, work->work_difficulty,
tm_getwork.tm_hour, tm_getwork.tm_min,
tm_getwork.tm_sec, getwork_time, workclone,
getwork_to_work, work_time, work_to_submit, submit_time,
tm_submit_reply.tm_hour, tm_submit_reply.tm_min,
tm_submit_reply.tm_sec);
}
}
share_result(val, res, err, work, hashshow, resubmit, worktime);
if (cgpu->dev_start_tv.tv_sec == 0)
dev_runtime = total_secs;
else {
cgtime(&now);
dev_runtime = tdiff(&now, &(cgpu->dev_start_tv));
}
if (dev_runtime < 1.0)
dev_runtime = 1.0;
cgpu->utility = cgpu->accepted / dev_runtime * 60;
if (!opt_realquiet)
print_status(thr_id);
if (!want_per_device_stats) {
char logline[256];
get_statline(logline, sizeof(logline), cgpu);
applog(LOG_INFO, "%s", logline);
}
json_decref(val);
rc = true;
out:
free(hexstr);
return rc;
}
static bool get_upstream_work(struct work *work, CURL *curl, char *curl_err_str)
{
struct pool *pool = work->pool;
struct sgminer_pool_stats *pool_stats = &(pool->sgminer_pool_stats);
struct timeval tv_elapsed;
json_t *val = NULL;
bool rc = false;
char *url;
applog(LOG_DEBUG, "DBG: sending %s get RPC call: %s", pool->rpc_url, pool->rpc_req);
url = pool->rpc_url;
cgtime(&work->tv_getwork);
val = json_rpc_call(curl, curl_err_str, url, pool->rpc_userpass, pool->rpc_req, false,
false, &work->rolltime, pool, false);
pool_stats->getwork_attempts++;
if (likely(val)) {
rc = work_decode(pool, work, val);
if (unlikely(!rc))
applog(LOG_DEBUG, "Failed to decode work in get_upstream_work");
} else
applog(LOG_DEBUG, "Failed json_rpc_call in get_upstream_work");
cgtime(&work->tv_getwork_reply);
timersub(&(work->tv_getwork_reply), &(work->tv_getwork), &tv_elapsed);
pool_stats->getwork_wait_rolling += ((double)tv_elapsed.tv_sec + ((double)tv_elapsed.tv_usec / 1000000)) * 0.63;
pool_stats->getwork_wait_rolling /= 1.63;
timeradd(&tv_elapsed, &(pool_stats->getwork_wait), &(pool_stats->getwork_wait));
if (timercmp(&tv_elapsed, &(pool_stats->getwork_wait_max), >)) {
pool_stats->getwork_wait_max.tv_sec = tv_elapsed.tv_sec;
pool_stats->getwork_wait_max.tv_usec = tv_elapsed.tv_usec;
}
if (timercmp(&tv_elapsed, &(pool_stats->getwork_wait_min), <)) {
pool_stats->getwork_wait_min.tv_sec = tv_elapsed.tv_sec;
pool_stats->getwork_wait_min.tv_usec = tv_elapsed.tv_usec;
}
pool_stats->getwork_calls++;
work->pool = pool;
work->longpoll = false;
work->getwork_mode = GETWORK_MODE_POOL;
calc_diff(work, 0);
total_getworks++;
pool->getwork_requested++;
if (likely(val))
json_decref(val);
return rc;
}
#endif /* HAVE_LIBCURL */
/* Specifies whether we can use this pool for work or not. */
static bool pool_unworkable(struct pool *pool)
{
if (pool->idle)
return true;
if (pool->state != POOL_ENABLED)
return true;
if (pool->has_stratum && !pool->stratum_active)
return true;
return false;
}
/* In balanced mode, the amount of diff1 solutions per pool is monitored as a
* rolling average per 10 minutes and if pools start getting more, it biases
* away from them to distribute work evenly. The share count is reset to the
* rolling average every 10 minutes to not send all work to one pool after it
* has been disabled/out for an extended period. */
static struct pool *select_balanced(struct pool *cp)
{
int i, lowest = cp->shares;
struct pool *ret = cp;
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
if (pool_unworkable(pool))
continue;
if (pool->shares < lowest) {
lowest = pool->shares;
ret = pool;
}
}
ret->shares++;
return ret;
}
static struct pool *priority_pool(int choice);
static bool pool_unusable(struct pool *pool);
/* Select any active pool in a rotating fashion when loadbalance is chosen if
* it has any quota left. */
static inline struct pool *select_pool(bool lagging)
{
static int rotating_pool = 0;
struct pool *pool, *cp;
bool avail = false;
int tested, i;
cp = current_pool();
if (pool_strategy == POOL_BALANCE) {
pool = select_balanced(cp);
goto out;
}
if (pool_strategy != POOL_LOADBALANCE && (!lagging || opt_fail_only)) {
pool = cp;
goto out;
} else
pool = NULL;
for (i = 0; i < total_pools; i++) {
struct pool *tp = pools[i];
if (tp->quota_used < tp->quota_gcd) {
avail = true;
break;
}
}
/* There are no pools with quota, so reset them. */
if (!avail) {
for (i = 0; i < total_pools; i++)
pools[i]->quota_used = 0;
if (++rotating_pool >= total_pools)
rotating_pool = 0;
}
/* Try to find the first pool in the rotation that is usable */
tested = 0;
while (!pool && tested++ < total_pools) {
pool = pools[rotating_pool];
if (pool->quota_used++ < pool->quota_gcd) {
if (!pool_unworkable(pool))
break;
/* Failover-only flag for load-balance means distribute
* unused quota to priority pool 0. */
if (opt_fail_only)
priority_pool(0)->quota_used--;
}
pool = NULL;
if (++rotating_pool >= total_pools)
rotating_pool = 0;
}
/* If there are no alive pools with quota, choose according to
* priority. */
if (!pool) {
for (i = 0; i < total_pools; i++) {
struct pool *tp = priority_pool(i);
if (!pool_unusable(tp)) {
pool = tp;
break;
}
}
}
/* If still nothing is usable, use the current pool */
if (!pool)
pool = cp;
out:
applog(LOG_DEBUG, "Selecting %s for work", get_pool_name(pool));
return pool;
}
/* truediffone == 0x00000000FFFF0000000000000000000000000000000000000000000000000000
* Generate a 256 bit binary LE target by cutting up diff into 64 bit sized
* portions or vice versa. */
static const double truediffone = 26959535291011309493156476344723991336010898738574164086137773096960.0;
static const double bits192 = 6277101735386680763835789423207666416102355444464034512896.0;
static const double bits128 = 340282366920938463463374607431768211456.0;
static const double bits64 = 18446744073709551616.0;
/* Converts a little endian 256 bit value to a double */
static double le256todouble(const void *target)
{
uint64_t *data64;
double dcut64;
data64 = (uint64_t *)((unsigned char *)target + 24);
dcut64 = le64toh(*data64) * bits192;
data64 = (uint64_t *)((unsigned char *)target + 16);
dcut64 += le64toh(*data64) * bits128;
data64 = (uint64_t *)((unsigned char *)target + 8);
dcut64 += le64toh(*data64) * bits64;
data64 = (uint64_t *)target;
dcut64 += le64toh(*data64);
return dcut64;
}
/*
* Calculate the work->work_difficulty based on the work->target
*/
static void calc_diff(struct work *work, double known)
{
struct sgminer_pool_stats *pool_stats = &(work->pool->sgminer_pool_stats);
double difficulty;
if (known) {
work->work_difficulty = known;
} else {
double d64, dcut64;
d64 = work->pool->algorithm.diff_multiplier2 * truediffone;
applog(LOG_DEBUG, "calc_diff() algorithm = %s", work->pool->algorithm.name);
// Neoscrypt
if (work->pool->algorithm.type == ALGO_NEOSCRYPT) {
dcut64 = (double)*((uint64_t *)(work->target + 22));
}
else {
dcut64 = le256todouble(work->target);
}
if (unlikely(!dcut64))
dcut64 = 1;
work->work_difficulty = d64 / dcut64;
}
difficulty = work->work_difficulty;
pool_stats->last_diff = difficulty;
suffix_string_double(difficulty, work->pool->diff, sizeof(work->pool->diff), 0);
if (difficulty == pool_stats->min_diff)
pool_stats->min_diff_count++;
else if (difficulty < pool_stats->min_diff || pool_stats->min_diff == 0) {
pool_stats->min_diff = difficulty;
pool_stats->min_diff_count = 1;
}
if (difficulty == pool_stats->max_diff)
pool_stats->max_diff_count++;
else if (difficulty > pool_stats->max_diff) {
pool_stats->max_diff = difficulty;
pool_stats->max_diff_count = 1;
}
}
#ifdef HAVE_CURSES
static void disable_curses_windows(void)
{
leaveok(logwin, false);
leaveok(statuswin, false);
leaveok(mainwin, false);
nocbreak();
echo();
delwin(logwin);
delwin(statuswin);
}
/* Force locking of curses console_lock on shutdown since a dead thread might
* have grabbed the lock. */
static bool curses_active_forcelocked(void)
{
bool ret;
mutex_trylock(&console_lock);
ret = curses_active;
if (!ret)
unlock_curses();
return ret;
}
static void disable_curses(void)
{
if (curses_active_forcelocked()) {
use_curses = false;
curses_active = false;
disable_curses_windows();
delwin(mainwin);
endwin();
#ifdef WIN32
// Move the cursor to after curses output.
HANDLE hout = GetStdHandle(STD_OUTPUT_HANDLE);
CONSOLE_SCREEN_BUFFER_INFO csbi;
COORD coord;
if (GetConsoleScreenBufferInfo(hout, &csbi)) {
coord.X = 0;
coord.Y = csbi.dwSize.Y - 1;
SetConsoleCursorPosition(hout, coord);
}
#endif
unlock_curses();
}
}
#endif
static void kill_timeout(struct thr_info *thr)
{
cg_completion_timeout(&thr_info_cancel_join, thr, 1000);
}
static void kill_mining(void)
{
struct thr_info *thr;
int i;
forcelog(LOG_DEBUG, "Killing off mining threads");
/* Kill the mining threads*/
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++) {
thr = mining_thr[i];
forcelog(LOG_DEBUG, "Waiting for thread %d to finish...", thr->id);
thr_info_cancel_join(thr);
}
rd_unlock(&mining_thr_lock);
}
static void __kill_work(void)
{
struct thr_info *thr;
int i;
if (!successful_connect)
return;
forcelog(LOG_INFO, "Received kill message");
forcelog(LOG_DEBUG, "Killing off watchpool thread");
/* Kill the watchpool thread */
thr = &control_thr[watchpool_thr_id];
kill_timeout(thr);
forcelog(LOG_DEBUG, "Killing off watchdog thread");
/* Kill the watchdog thread */
thr = &control_thr[watchdog_thr_id];
kill_timeout(thr);
forcelog(LOG_DEBUG, "Shutting down mining threads");
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++) {
struct cgpu_info *cgpu;
thr = mining_thr[i];
if (!thr)
continue;
cgpu = thr->cgpu;
if (!cgpu)
continue;
cgpu->shutdown = true;
}
rd_unlock(&mining_thr_lock);
sleep(1);
cg_completion_timeout(&kill_mining, NULL, 3000);
/* Stop the others */
forcelog(LOG_DEBUG, "Killing off API thread");
thr = &control_thr[api_thr_id];
kill_timeout(thr);
}
/* This should be the common exit path */
void kill_work(void)
{
__kill_work();
quit(0, "Shutdown signal received.");
}
static
#ifdef WIN32
const
#endif
const char **initial_args;
static void clean_up(bool restarting);
void app_restart(void)
{
applog(LOG_WARNING, "Attempting to restart %s", packagename);
__kill_work();
clean_up(true);
#if defined(unix) || defined(__APPLE__)
if (forkpid > 0) {
kill(forkpid, SIGTERM);
forkpid = 0;
}
#endif
execv(initial_args[0], (EXECV_2ND_ARG_TYPE)initial_args);
applog(LOG_WARNING, "Failed to restart application");
}
static void sighandler(int __maybe_unused sig)
{
#ifndef _MSC_VER
/* Restore signal handlers so we can still quit if kill_work fails */
sigaction(SIGTERM, &termhandler, NULL);
sigaction(SIGINT, &inthandler, NULL);
#endif
kill_work();
}
#ifdef HAVE_LIBCURL
/* Called with pool_lock held. Recruit an extra curl if none are available for
* this pool. */
static void recruit_curl(struct pool *pool)
{
struct curl_ent *ce = (struct curl_ent *)calloc(sizeof(struct curl_ent), 1);
if (unlikely(!ce))
quit(1, "Failed to calloc in recruit_curl");
ce->curl = curl_easy_init();
if (unlikely(!ce->curl))
quit(1, "Failed to init in recruit_curl");
list_add(&ce->node, &pool->curlring);
pool->curls++;
}
/* Grab an available curl if there is one. If not, then recruit extra curls
* unless we are in a submit_fail situation, or we have opt_delaynet enabled
* and there are already 5 curls in circulation. Limit total number to the
* number of mining threads per pool as well to prevent blasting a pool during
* network delays/outages. */
static struct curl_ent *pop_curl_entry(struct pool *pool)
{
int curl_limit = opt_delaynet ? 5 : (mining_threads + opt_queue) * 2;
bool recruited = false;
struct curl_ent *ce;
mutex_lock(&pool->pool_lock);
retry:
if (!pool->curls) {
recruit_curl(pool);
recruited = true;
} else if (list_empty(&pool->curlring)) {
if (pool->curls >= curl_limit) {
pthread_cond_wait(&pool->cr_cond, &pool->pool_lock);
goto retry;
} else {
recruit_curl(pool);
recruited = true;
}
}
ce = list_entry(pool->curlring.next, struct curl_ent*, node);
list_del(&ce->node);
mutex_unlock(&pool->pool_lock);
if (recruited)
applog(LOG_DEBUG, "Recruited curl for %s", get_pool_name(pool));
return ce;
}
static void push_curl_entry(struct curl_ent *ce, struct pool *pool)
{
mutex_lock(&pool->pool_lock);
list_add_tail(&ce->node, &pool->curlring);
cgtime(&ce->tv);
pthread_cond_broadcast(&pool->cr_cond);
mutex_unlock(&pool->pool_lock);
}
static bool stale_work(struct work *work, bool share);
static inline bool should_roll(struct work *work)
{
struct timeval now;
time_t expiry;
if (work->pool != current_pool() && pool_strategy != POOL_LOADBALANCE && pool_strategy != POOL_BALANCE)
return false;
if (work->rolltime > opt_scantime)
expiry = work->rolltime;
else
expiry = opt_scantime;
expiry = expiry * 2 / 3;
/* We shouldn't roll if we're unlikely to get one shares' duration
* work out of doing so */
cgtime(&now);
if (now.tv_sec - work->tv_staged.tv_sec > expiry)
return false;
return true;
}
/* Limit rolls to 7000 to not beyond 2 hours in the future where bitcoind will
* reject blocks as invalid. */
static inline bool can_roll(struct work *work)
{
return (!work->stratum && work->pool && work->rolltime && !work->clone &&
work->rolls < 7000 && !stale_work(work, false));
}
static uint32_t _get_work_time(struct work *work)
{
uint32_t *data = (uint32_t*) work->data;
uint32_t work_ntime = data[17];
if (work->pool && work->pool->algorithm.type == ALGO_DECRED) {
work_ntime = data[34];
}
return work_ntime;
}
static void _set_work_time(struct work *work, uint32_t ntime)
{
uint32_t *data = (uint32_t*) work->data;
uint32_t *work_ntime = &data[17];
if (work->pool && work->pool->algorithm.type == ALGO_DECRED) {
work_ntime = &data[34];
}
(*work_ntime) = ntime;
}
static void roll_work(struct work *work)
{
uint32_t work_ntime;
uint32_t ntime;
work_ntime = _get_work_time(work);
ntime = be32toh(work_ntime);
ntime++;
if (work->pool->algorithm.type == ALGO_DECRED) {
uint32_t* data = (uint32_t*) work->data;
// dont mess with ntime, use extranonce
data[36]++;
data[37] = ((rand()*4) << 8) | work->thr_id;
} else {
_set_work_time(work, htobe32(ntime));
}
local_work++;
work->rolls++;
work->blk.nonce = 0;
applog(LOG_DEBUG, "Successfully rolled work");
/* This is now a different work item so it needs a different ID for the
* hashtable */
work->id = total_work++;
}
static void *submit_work_thread(void *userdata)
{
struct work *work = (struct work *)userdata;
struct pool *pool = work->pool;
bool resubmit = false;
struct curl_ent *ce;
pthread_detach(pthread_self());
RenameThread("SubmitWork");
applog(LOG_DEBUG, "Creating extra submit work thread");
ce = pop_curl_entry(pool);
/* submit solution to bitcoin via JSON-RPC */
while (!submit_upstream_work(work, ce->curl, ce->curl_err_str, resubmit)) {
if (opt_lowmem) {
applog(LOG_NOTICE, "%s share being discarded to minimise memory cache", get_pool_name(pool));
break;
}
resubmit = true;
if (stale_work(work, true)) {
applog(LOG_NOTICE, "%s share became stale while retrying submit, discarding", get_pool_name(pool));
mutex_lock(&stats_lock);
total_stale++;
pool->stale_shares++;
total_diff_stale += work->work_difficulty;
pool->diff_stale += work->work_difficulty;
mutex_unlock(&stats_lock);
free_work(work);
break;
}
/* pause, then restart work-request loop */
applog(LOG_INFO, "json_rpc_call failed on submit_work, retrying");
}
push_curl_entry(ce, pool);
return NULL;
}
static struct work *make_clone(struct work *work)
{
struct work *work_clone = copy_work(work);
if (work->pool->algorithm.type == ALGO_DECRED) {
// maybe not useful here
((uint32_t*)work->data)[36] = (rand()*4);
((uint32_t*)work->data)[37] = (rand()*4) << 8;
}
work_clone->clone = true;
cgtime((struct timeval *)&(work_clone->tv_cloned));
work_clone->longpoll = false;
work_clone->mandatory = false;
/* Make cloned work appear slightly older to bias towards keeping the
* master work item which can be further rolled */
work_clone->tv_staged.tv_sec -= 1;
return work_clone;
}
static void stage_work(struct work *work);
static bool clone_available(void)
{
struct work *work_clone = NULL, *work, *tmp;
bool cloned = false;
mutex_lock(stgd_lock);
if (!staged_rollable)
goto out_unlock;
HASH_ITER(hh, staged_work, work, tmp) {
if (can_roll(work) && should_roll(work)) {
roll_work(work);
work_clone = make_clone(work);
roll_work(work);
cloned = true;
break;
}
}
out_unlock:
mutex_unlock(stgd_lock);
if (cloned) {
applog(LOG_DEBUG, "Pushing cloned available work to stage thread");
stage_work(work_clone);
}
return cloned;
}
/* Clones work by rolling it if possible, and returning a clone instead of the
* original work item which gets staged again to possibly be rolled again in
* the future */
static struct work *clone_work(struct work *work)
{
int mrs = mining_threads + opt_queue - total_staged();
struct work *work_clone;
bool cloned;
if (mrs < 1)
return work;
cloned = false;
work_clone = make_clone(work);
while (mrs-- > 0 && can_roll(work) && should_roll(work)) {
applog(LOG_DEBUG, "Pushing rolled converted work to stage thread");
stage_work(work_clone);
roll_work(work);
work_clone = make_clone(work);
/* Roll it again to prevent duplicates should this be used
* directly later on */
roll_work(work);
cloned = true;
}
if (cloned) {
stage_work(work);
return work_clone;
}
free_work(work_clone);
return work;
}
#else /* HAVE_LIBCURL */
static void *submit_work_thread(void __maybe_unused *userdata)
{
pthread_detach(pthread_self());
return NULL;
}
#endif /* HAVE_LIBCURL */
/* Return an adjusted ntime if we're submitting work that a device has
* internally offset the ntime. */
static char *offset_ntime(const char *ntime, int noffset)
{
unsigned char bin[4];
uint32_t h32, *be32 = (uint32_t *)bin;
hex2bin(bin, ntime, 4);
h32 = be32toh(*be32) + noffset;
*be32 = htobe32(h32);
return bin2hex(bin, 4);
}
/* Duplicates any dynamically allocated arrays within the work struct to
* prevent a copied work struct from freeing ram belonging to another struct */
static void _copy_work(struct work *work, const struct work *base_work, int noffset)
{
int id = work->id;
clean_work(work);
memcpy(work, base_work, sizeof(struct work));
/* Keep the unique new id assigned during make_work to prevent copied
* work from having the same id. */
work->id = id;
if (base_work->job_id)
work->job_id = strdup(base_work->job_id);
if (base_work->nonce1)
work->nonce1 = strdup(base_work->nonce1);
if (base_work->ntime) {
/* If we are passed an noffset the binary work->data ntime and
* the work->ntime hex string need to be adjusted. */
if (noffset) {
uint32_t work_ntime = _get_work_time(work);
uint32_t ntime = be32toh(work_ntime);
ntime += noffset;
_set_work_time(work, htobe32(ntime));
work->ntime = offset_ntime(base_work->ntime, noffset);
} else
work->ntime = strdup(base_work->ntime);
} else if (noffset) {
uint32_t work_ntime = _get_work_time(work);
uint32_t ntime = be32toh(work_ntime);
ntime += noffset;
_set_work_time(work, htobe32(ntime));
}
if (base_work->coinbase)
work->coinbase = strdup(base_work->coinbase);
}
/* Generates a copy of an existing work struct, creating fresh heap allocations
* for all dynamically allocated arrays within the struct. noffset is used for
* when a driver has internally rolled the ntime, noffset is a relative value.
* The macro copy_work() calls this function with an noffset of 0. */
struct work *copy_work_noffset(struct work *base_work, int noffset)
{
struct work *work = make_work();
_copy_work(work, base_work, noffset);
return work;
}
void pool_failed(struct pool *pool)
{
if (!pool_tset(pool, &pool->idle)) {
cgtime(&pool->tv_idle);
if (pool == current_pool()) {
switch_pools(NULL);
}
}
}
static void pool_died(struct pool *pool)
{
if (!pool_tset(pool, &pool->idle)) {
cgtime(&pool->tv_idle);
if (pool == current_pool()) {
applog(LOG_WARNING, "%s not responding!", get_pool_name(pool));
switch_pools(NULL);
} else {
applog(LOG_INFO, "%s failed to return work", get_pool_name(pool));
}
}
}
static bool stale_work(struct work *work, bool share)
{
struct timeval now;
time_t work_expiry;
struct pool *pool;
int getwork_delay;
if (work->work_block != work_block) {
applog(LOG_DEBUG, "Work stale due to block mismatch");
return true;
}
/* Technically the rolltime should be correct but some pools
* advertise a broken expire= that is lower than a meaningful
* scantime */
if (work->rolltime > opt_scantime)
work_expiry = work->rolltime;
else
work_expiry = opt_expiry;
pool = work->pool;
if (!share && pool->has_stratum) {
bool same_job;
if (!pool->stratum_active || !pool->stratum_notify) {
applog(LOG_DEBUG, "Work stale due to stratum inactive");
return true;
}
same_job = true;
cg_rlock(&pool->data_lock);
if (strcmp(work->job_id, pool->swork.job_id))
same_job = false;
cg_runlock(&pool->data_lock);
if (!same_job) {
applog(LOG_DEBUG, "Work stale due to stratum job_id mismatch");
return true;
}
}
/* Factor in the average getwork delay of this pool, rounding it up to
* the nearest second */
getwork_delay = pool->sgminer_pool_stats.getwork_wait_rolling * 5 + 1;
work_expiry -= getwork_delay;
if (unlikely(work_expiry < 5))
work_expiry = 5;
cgtime(&now);
if ((now.tv_sec - work->tv_staged.tv_sec) >= work_expiry) {
applog(LOG_DEBUG, "Work stale due to expiry");
return true;
}
if (opt_fail_only && !share && pool != current_pool() && !work->mandatory &&
pool_strategy != POOL_LOADBALANCE && pool_strategy != POOL_BALANCE) {
applog(LOG_DEBUG, "Work stale due to fail only pool mismatch");
return true;
}
return false;
}
static double share_diff(const struct work *work)
{
bool new_best = false;
double d64, s64;
double ret;
d64 = work->pool->algorithm.share_diff_multiplier * truediffone;
s64 = le256todouble(work->hash);
if (unlikely(!s64))
s64 = 0;
ret = d64 / s64;
applog(LOG_DEBUG, "Found share with difficulty %.3f", ret);
cg_wlock(&control_lock);
if (unlikely(ret > best_diff)) {
new_best = true;
best_diff = ret;
suffix_string_double(best_diff, best_share, sizeof(best_share), 0);
}
if (unlikely(ret > work->pool->best_diff))
work->pool->best_diff = ret;
cg_wunlock(&control_lock);
if (unlikely(new_best))
applog(LOG_INFO, "New best share: %s", best_share);
return ret;
}
static bool cnx_needed(struct pool *pool);
/* Find the pool that currently has the highest priority */
static struct pool *priority_pool(int choice)
{
struct pool *ret = NULL;
int i;
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
if (pool->prio == choice) {
ret = pool;
break;
}
}
if (unlikely(!ret)) {
applog(LOG_ERR, "WTF No pool %d found!", choice);
return pools[choice];
}
return ret;
}
/* Specifies whether we can switch to this pool or not. */
static bool pool_unusable(struct pool *pool)
{
if (pool->idle)
return true;
if (pool->state != POOL_ENABLED)
return true;
return false;
}
void __switch_pools(struct pool *selected, bool saveprio)
{
struct pool *pool, *last_pool;
int i, pool_no, next_pool;
cg_wlock(&control_lock);
last_pool = currentpool;
pool_no = currentpool->pool_no;
/* If a specific pool was selected, prioritise it over others */
if (selected && saveprio)
{
if (selected->prio != 0)
{
for (i = 0; i < total_pools; i++)
{
pool = pools[i];
if (pool->prio < selected->prio)
pool->prio++;
}
selected->prio = 0;
}
}
switch (pool_strategy)
{
/* All of these set to the master pool */
case POOL_BALANCE:
case POOL_FAILOVER:
case POOL_LOADBALANCE:
for (i = 0; i < total_pools; i++)
{
pool = priority_pool(i);
if (pool_unusable(pool))
continue;
pool_no = pool->pool_no;
break;
}
break;
/* Both of these simply increment and cycle */
case POOL_ROUNDROBIN:
case POOL_ROTATE:
if (selected && !selected->idle)
{
pool_no = selected->pool_no;
break;
}
next_pool = pool_no;
/* Select the next alive pool */
for (i = 1; i < total_pools; i++)
{
next_pool++;
if (next_pool >= total_pools)
next_pool = 0;
pool = pools[next_pool];
if (pool_unusable(pool))
continue;
pool_no = next_pool;
break;
}
break;
default:
break;
}
currentpool = pools[pool_no];
pool = currentpool;
on_backup_pool = pool->backup;
cg_wunlock(&control_lock);
/* Set the lagging flag to avoid pool not providing work fast enough
* messages in failover only mode since we have to get all fresh work
* as in restart_threads */
if (opt_fail_only) {
pool_tset(pool, &pool->lagging);
}
if (pool != last_pool && pool_strategy != POOL_LOADBALANCE && pool_strategy != POOL_BALANCE) {
//if the gpus have been initialized or first pool during startup, it's ok to switch...
if(gpu_initialized || startup) {
applog(LOG_WARNING, "Switching to %s", get_pool_name(pool));
if (pool_localgen(pool) || opt_fail_only) {
clear_pool_work(last_pool);
}
}
}
//if startup, initialize gpus and start mining threads
if(startup) {
startup = false; //remove startup flag so we don't enter this block again
applog(LOG_NOTICE, "Startup GPU initialization... Using settings from pool %s.", get_pool_name(pool));
//set initial pool number for restart_mining_threads to prevent mismatched GPU settings
init_pool = pool->pool_no;
//apply gpu settings based on first alive pool
apply_initial_gpu_settings(pool);
gpu_initialized = true; //gpus initialized
}
mutex_lock(&lp_lock);
pthread_cond_broadcast(&lp_cond);
mutex_unlock(&lp_lock);
}
void discard_work(struct work *work)
{
if (!work->clone && !work->rolls && !work->mined) {
if (work->pool) {
work->pool->discarded_work++;
work->pool->quota_used--;
work->pool->works--;
}
total_discarded++;
applog(LOG_DEBUG, "[THR%d] Discarded work", work->thr_id);
} else
applog(LOG_DEBUG, "[THR%d] Discarded cloned or rolled work", work->thr_id);
free_work(work);
}
static void wake_gws(void)
{
mutex_lock(stgd_lock);
pthread_cond_signal(&gws_cond);
mutex_unlock(stgd_lock);
}
static void discard_stale(void)
{
struct work *work, *tmp;
int stale = 0;
mutex_lock(stgd_lock);
HASH_ITER(hh, staged_work, work, tmp) {
if (stale_work(work, false)) {
HASH_DEL(staged_work, work);
discard_work(work);
stale++;
}
}
pthread_cond_signal(&gws_cond);
mutex_unlock(stgd_lock);
if (stale)
applog(LOG_DEBUG, "Discarded %d stales that didn't match current hash", stale);
}
static void *restart_thread(void __maybe_unused *arg)
{
struct pool *cp = current_pool();
struct cgpu_info *cgpu;
int i;
pthread_detach(pthread_self());
/* Artificially set the lagging flag to avoid pool not providing work
* fast enough messages after every long poll */
pool_tset(cp, &cp->lagging);
/* Discard staged work that is now stale */
discard_stale();
if (mining_thr) {
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++) {
cgpu = mining_thr[i]->cgpu;
if (unlikely(!cgpu))
continue;
if (cgpu->deven != DEV_ENABLED)
continue;
mining_thr[i]->work_restart = true;
cgpu->drv->flush_work(cgpu);
}
rd_unlock(&mining_thr_lock);
}
mutex_lock(&restart_lock);
pthread_cond_broadcast(&restart_cond);
mutex_unlock(&restart_lock);
return NULL;
}
/* In order to prevent a deadlock via the various drv->flush_work
* implementations we send the restart messages via a separate thread. */
static void restart_threads(void)
{
pthread_t rthread;
if (unlikely(pthread_create(&rthread, NULL, restart_thread, NULL)))
quit(1, "Failed to create restart thread");
}
static void signal_work_update(void)
{
int i;
applog(LOG_INFO, "Work update message received");
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++)
mining_thr[i]->work_update = true;
rd_unlock(&mining_thr_lock);
}
static void set_curblock(char *hexstr, unsigned char *bedata)
{
int ofs;
cg_wlock(&ch_lock);
cgtime(&block_timeval);
strcpy(current_hash, hexstr);
memcpy(current_block, bedata, 32);
get_timestamp(blocktime, sizeof(blocktime), &block_timeval);
cg_wunlock(&ch_lock);
for (ofs = 0; ofs <= 56; ofs++) {
if (memcmp(&current_hash[ofs], "0", 1))
break;
}
strncpy(prev_block, &current_hash[ofs], 8);
prev_block[8] = '\0';
applog(LOG_INFO, "New block: %s... diff %s", current_hash, block_diff);
}
/* Search to see if this string is from a block that has been seen before */
static bool block_exists(char *hexstr)
{
struct block *s;
rd_lock(&blk_lock);
HASH_FIND_STR(blocks, hexstr, s);
rd_unlock(&blk_lock);
if (s)
return true;
return false;
}
static int block_sort(struct block *blocka, struct block *blockb)
{
return blocka->block_no - blockb->block_no;
}
/* Decode the current block difficulty which is in packed form */
static void set_blockdiff(const struct work *work)
{
double ddiff = get_work_blockdiff(work);
if (unlikely(current_diff != ddiff)) {
suffix_string(ddiff, block_diff, sizeof(block_diff), 0);
current_diff = ddiff;
if (opt_morenotices)
applog(LOG_NOTICE, "Network diff set to %s", block_diff);
}
}
static bool test_work_current(struct work *work)
{
struct pool *pool = work->pool;
unsigned char bedata[32];
char hexstr[68];
bool ret = true;
if (work->mandatory)
return ret;
swap256(bedata, work->data + 4);
__bin2hex(hexstr, bedata, 32);
/* Search to see if this block exists yet and if not, consider it a
* new block and set the current block details to this one */
if (!block_exists(hexstr)) {
struct block *s = (struct block *)calloc(sizeof(struct block), 1);
int deleted_block = 0;
if (unlikely(!s))
quit (1, "test_work_current OOM");
strcpy(s->hash, hexstr);
s->block_no = new_blocks++;
wr_lock(&blk_lock);
/* Only keep the last hour's worth of blocks in memory since
* work from blocks before this is virtually impossible and we
* want to prevent memory usage from continually rising */
if (HASH_COUNT(blocks) > 6) {
struct block *oldblock;
HASH_SORT(blocks, block_sort);
oldblock = blocks;
deleted_block = oldblock->block_no;
HASH_DEL(blocks, oldblock);
free(oldblock);
}
HASH_ADD_STR(blocks, hash, s);
set_blockdiff(work);
wr_unlock(&blk_lock);
if (deleted_block)
applog(LOG_DEBUG, "Deleted block %d from database", deleted_block);
set_curblock(hexstr, bedata);
/* Copy the information to this pool's prev_block since it
* knows the new block exists. */
memcpy(pool->prev_block, bedata, 32);
if (unlikely(new_blocks == 1)) {
ret = false;
goto out;
}
work->work_block = ++work_block;
if (opt_morenotices)
{
if (work->longpoll) {
if (work->stratum) {
applog(LOG_NOTICE, "Stratum from %s detected new block", get_pool_name(pool));
} else {
applog(LOG_NOTICE, "%sLONGPOLL from %s detected new block",
work->gbt ? "GBT " : "", get_pool_name(pool));
}
} else if (have_longpoll)
applog(LOG_NOTICE, "New block detected on network before pool notification");
else
applog(LOG_NOTICE, "New block detected on network");
}
restart_threads();
} else {
if (memcmp(pool->prev_block, bedata, 32)) {
/* Work doesn't match what this pool has stored as
* prev_block. Let's see if the work is from an old
* block or the pool is just learning about a new
* block. */
if (memcmp(bedata, current_block, 32)) {
/* Doesn't match current block. It's stale */
applog(LOG_DEBUG, "Stale data from %s", get_pool_name(pool));
ret = false;
} else {
/* Work is from new block and pool is up now
* current. */
applog(LOG_INFO, "%s now up to date", get_pool_name(pool));
memcpy(pool->prev_block, bedata, 32);
}
}
#if 0
/* This isn't ideal, this pool is still on an old block but
* accepting shares from it. To maintain fair work distribution
* we work on it anyway. */
if (memcmp(bedata, current_block, 32))
applog(LOG_DEBUG, "%s still on old block", get_pool_name(pool));
#endif
if (work->longpoll) {
work->work_block = ++work_block;
if (shared_strategy() || work->pool == current_pool()) {
if(opt_morenotices) {
if (work->stratum)
applog(LOG_NOTICE, "Stratum from %s requested work restart", get_pool_name(pool));
else
applog(LOG_NOTICE, "%sLONGPOLL from %s requested work restart", work->gbt ? "GBT " : "", get_pool_name(pool));
}
restart_threads();
}
}
}
out:
work->longpoll = false;
return ret;
}
static int tv_sort(struct work *worka, struct work *workb)
{
return worka->tv_staged.tv_sec - workb->tv_staged.tv_sec;
}
static bool work_rollable(struct work *work)
{
return (!work->clone && work->rolltime);
}
static bool hash_push(struct work *work)
{
bool rc = true;
mutex_lock(stgd_lock);
if (work_rollable(work))
staged_rollable++;
if (likely(!getq->frozen)) {
HASH_ADD_INT(staged_work, id, work);
HASH_SORT(staged_work, tv_sort);
} else
rc = false;
pthread_cond_broadcast(&getq->cond);
mutex_unlock(stgd_lock);
return rc;
}
static void stage_work(struct work *work)
{
applog(LOG_DEBUG, "[THR%d] Pushing work from %s to hash queue", work->thr_id, get_pool_name(work->pool));
work->work_block = work_block;
test_work_current(work);
work->pool->works++;
hash_push(work);
}
#ifdef HAVE_CURSES
int curses_int(const char *query)
{
int ret;
char *cvar;
cvar = curses_input(query);
ret = atoi(cvar);
free(cvar);
return ret;
}
#endif
#ifdef HAVE_CURSES
static bool input_pool(bool live);
#endif
#ifdef HAVE_CURSES
static void display_pool_summary(struct pool *pool)
{
double efficiency = 0.0;
if (curses_active_locked()) {
wlog("Pool: %s\n", pool->rpc_url);
if (pool->solved)
wlog("SOLVED %d BLOCK%s!\n", pool->solved, pool->solved > 1 ? "S" : "");
if (!pool->has_stratum)
wlog("%s own long-poll support\n", pool->hdr_path ? "Has" : "Does not have");
wlog(" Queued work requests: %d\n", pool->getwork_requested);
wlog(" Share submissions: %d\n", pool->accepted + pool->rejected);
wlog(" Accepted shares: %d\n", pool->accepted);
wlog(" Rejected shares: %d\n", pool->rejected);
wlog(" Accepted difficulty shares: %1.f\n", pool->diff_accepted);
wlog(" Rejected difficulty shares: %1.f\n", pool->diff_rejected);
if (pool->accepted || pool->rejected)
wlog(" Reject ratio: %.1f%%\n", (double)(pool->rejected * 100) / (double)(pool->accepted + pool->rejected));
efficiency = pool->getwork_requested ? pool->accepted * 100.0 / pool->getwork_requested : 0.0;
if (!pool_localgen(pool))
wlog(" Efficiency (accepted / queued): %.0f%%\n", efficiency);
wlog(" Items worked on: %d\n", pool->works);
wlog(" Discarded work due to new blocks: %d\n", pool->discarded_work);
wlog(" Stale submissions discarded due to new blocks: %d\n", pool->stale_shares);
wlog(" Unable to get work from server occasions: %d\n", pool->getfail_occasions);
wlog(" Submitting work remotely delay occasions: %d\n\n", pool->remotefail_occasions);
unlock_curses();
}
}
#endif
void zero_bestshare(void)
{
int i;
best_diff = 0;
memset(best_share, 0, 8);
suffix_string_double(best_diff, best_share, sizeof(best_share), 0);
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
pool->best_diff = 0;
}
}
void zero_stats(void)
{
int i;
cgtime(&total_tv_start);
total_rolling = 0;
total_mhashes_done = 0;
total_getworks = 0;
total_accepted = 0;
total_rejected = 0;
hw_errors = 0;
total_stale = 0;
total_discarded = 0;
local_work = 0;
total_go = 0;
total_ro = 0;
total_secs = 1.0;
total_diff1 = 0;
found_blocks = 0;
total_diff_accepted = 0;
total_diff_rejected = 0;
total_diff_stale = 0;
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
pool->getwork_requested = 0;
pool->accepted = 0;
pool->rejected = 0;
pool->stale_shares = 0;
pool->discarded_work = 0;
pool->getfail_occasions = 0;
pool->remotefail_occasions = 0;
pool->last_share_time = 0;
pool->diff1 = 0;
pool->diff_accepted = 0;
pool->diff_rejected = 0;
pool->diff_stale = 0;
pool->last_share_diff = 0;
}
zero_bestshare();
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = get_devices(i);
mutex_lock(&hash_lock);
cgpu->total_mhashes = 0;
cgpu->accepted = 0;
cgpu->rejected = 0;
cgpu->hw_errors = 0;
cgpu->utility = 0.0;
cgpu->last_share_pool_time = 0;
cgpu->diff1 = 0;
cgpu->diff_accepted = 0;
cgpu->diff_rejected = 0;
cgpu->last_share_diff = 0;
mutex_unlock(&hash_lock);
/* Don't take any locks in the driver zero stats function, as
* it's called async from everything else and we don't want to
* deadlock. */
cgpu->drv->zero_stats(cgpu);
}
}
static void set_highprio(void)
{
#ifndef WIN32
int ret = nice(-10);
if (!ret)
applog(LOG_DEBUG, "Unable to set thread to high priority");
#else
SetThreadPriority(GetCurrentThread(), THREAD_PRIORITY_HIGHEST);
#endif
}
static void set_lowprio(void)
{
#ifndef WIN32
int ret = nice(10);
if (!ret)
applog(LOG_INFO, "Unable to set thread to low priority");
#else
SetThreadPriority(GetCurrentThread(), THREAD_PRIORITY_LOWEST);
#endif
}
#ifdef HAVE_CURSES
static void display_pools(void)
{
struct pool *pool;
int selected, i;
char input;
opt_loginput = true;
immedok(logwin, true);
clear_logwin();
updated:
for (i = 0; i < total_pools; i++) {
pool = pools[i];
if (pool == current_pool())
wattron(logwin, A_BOLD);
if (pool->state != POOL_ENABLED)
wattron(logwin, A_DIM);
wlogprint("%d: ", pool->pool_no);
switch (pool->state) {
case POOL_ENABLED:
wlogprint("Enabled ");
break;
case POOL_DISABLED:
wlogprint("Disabled ");
break;
case POOL_REJECTING:
wlogprint("Rejecting ");
break;
case POOL_HIDDEN:
default:
break;
}
wlogprint("%s Quota %d Prio %d '%s' User:%s\n",
pool->idle ? "Dead" : "Alive",
pool->quota,
pool->prio,
get_pool_name(pool),
get_pool_user(pool));
wattroff(logwin, A_BOLD | A_DIM);
}
retry:
wlogprint("\nCurrent pool management strategy: %s\n",
strategies[pool_strategy].s);
if (pool_strategy == POOL_ROTATE)
wlogprint("Set to rotate every %d minutes\n", opt_rotate_period);
wlogprint("[F]ailover only %s\n", opt_fail_only ? "enabled" : "disabled");
wlogprint("Pool [A]dd [R]emove [D]isable [E]nable [Q]uota change\n");
wlogprint("[C]hange management strategy [S]witch pool [I]nformation\n");
wlogprint("Or press any other key to continue\n");
logwin_update();
input = getch();
if (!strncasecmp(&input, "a", 1)) {
input_pool(true);
goto updated;
} else if (!strncasecmp(&input, "r", 1)) {
if (total_pools <= 1) {
wlogprint("Cannot remove last pool");
goto retry;
}
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
if (pool == current_pool())
switch_pools(NULL);
if (pool == current_pool()) {
wlogprint("Unable to remove pool due to activity\n");
goto retry;
}
disable_pool(pool);
remove_pool(pool);
goto updated;
} else if (!strncasecmp(&input, "s", 1)) {
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
enable_pool(pool);
switch_pools(pool);
goto updated;
} else if (!strncasecmp(&input, "d", 1)) {
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
disable_pool(pool);
if (pool == current_pool())
switch_pools(NULL);
goto updated;
} else if (!strncasecmp(&input, "e", 1)) {
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
enable_pool(pool);
if (pool->prio < current_pool()->prio)
switch_pools(pool);
goto updated;
} else if (!strncasecmp(&input, "c", 1)) {
for (i = 0; i <= TOP_STRATEGY; i++)
wlogprint("%d: %s\n", i, strategies[i].s);
selected = curses_int("Select strategy number type");
if (selected < 0 || selected > TOP_STRATEGY) {
wlogprint("Invalid selection\n");
goto retry;
}
if (selected == POOL_ROTATE) {
opt_rotate_period = curses_int("Select interval in minutes");
if (opt_rotate_period < 0 || opt_rotate_period > 9999) {
opt_rotate_period = 0;
wlogprint("Invalid selection\n");
goto retry;
}
}
pool_strategy = (enum pool_strategy)selected;
switch_pools(NULL);
goto updated;
} else if (!strncasecmp(&input, "i", 1)) {
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
display_pool_summary(pool);
goto retry;
} else if (!strncasecmp(&input, "q", 1)) {
selected = curses_int("Select pool number");
if (selected < 0 || selected >= total_pools) {
wlogprint("Invalid selection\n");
goto retry;
}
pool = pools[selected];
selected = curses_int("Set quota");
if (selected < 0) {
wlogprint("Invalid negative quota\n");
goto retry;
}
pool->quota = selected;
adjust_quota_gcd();
goto updated;
} else if (!strncasecmp(&input, "f", 1)) {
opt_fail_only ^= true;
goto updated;
} else
clear_logwin();
immedok(logwin, false);
opt_loginput = false;
}
static void display_options(void)
{
int selected;
char input;
opt_loginput = true;
immedok(logwin, true);
retry:
clear_logwin();
wlogprint("[N]ormal [C]lear [S]ilent mode (disable all output)\n");
wlogprint("[D]ebug: %s\n[P]er-device: %s\n[Q]uiet: %s\n[V]erbose: %s\n"
"[R]PC debug: %s\n[W]orkTime details: %s\n[I]ncognito: %s\n"
"co[M]pact: %s\n[L]og interval: %d\n[Z]ero statistics\n",
opt_debug_console ? "on" : "off",
want_per_device_stats? "on" : "off",
opt_quiet ? "on" : "off",
opt_verbose ? "on" : "off",
opt_protocol ? "on" : "off",
opt_worktime ? "on" : "off",
opt_incognito ? "on" : "off",
opt_compact ? "on" : "off",
opt_log_interval);
wlogprint("Select an option or any other key to return\n");
logwin_update();
input = getch();
if (!strncasecmp(&input, "q", 1)) {
opt_quiet ^= true;
wlogprint("Quiet mode %s\n", opt_quiet ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "v", 1)) {
opt_verbose ^= true;
if (opt_verbose)
opt_quiet = false;
wlogprint("Verbose mode %s\n", opt_verbose ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "n", 1)) {
opt_verbose = false;
opt_debug_console = false;
opt_quiet = false;
opt_protocol = false;
opt_compact = false;
want_per_device_stats = false;
wlogprint("Output mode reset to normal\n");
switch_logsize(false);
goto retry;
} else if (!strncasecmp(&input, "d", 1)) {
opt_debug = true;
opt_debug_console ^= true;
opt_verbose = opt_debug_console;
if (opt_debug_console)
opt_quiet = false;
wlogprint("Debug mode %s\n", opt_debug_console ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "i", 1)) {
opt_incognito ^= true;
wlogprint("Incognito mode %s\n", opt_incognito ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "m", 1)) {
opt_compact ^= true;
wlogprint("Compact mode %s\n", opt_compact ? "enabled" : "disabled");
switch_logsize(false);
goto retry;
} else if (!strncasecmp(&input, "p", 1)) {
want_per_device_stats ^= true;
opt_verbose = want_per_device_stats;
wlogprint("Per-device stats %s\n", want_per_device_stats ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "r", 1)) {
opt_protocol ^= true;
if (opt_protocol)
opt_quiet = false;
wlogprint("RPC protocol debugging %s\n", opt_protocol ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "c", 1))
clear_logwin();
else if (!strncasecmp(&input, "l", 1)) {
selected = curses_int("Interval in seconds");
if (selected < 0 || selected > 9999) {
wlogprint("Invalid selection\n");
goto retry;
}
opt_log_interval = selected;
wlogprint("Log interval set to %d seconds\n", opt_log_interval);
goto retry;
} else if (!strncasecmp(&input, "s", 1)) {
opt_realquiet = true;
} else if (!strncasecmp(&input, "w", 1)) {
opt_worktime ^= true;
wlogprint("WorkTime details %s\n", opt_worktime ? "enabled" : "disabled");
goto retry;
} else if (!strncasecmp(&input, "z", 1)) {
zero_stats();
goto retry;
} else
clear_logwin();
immedok(logwin, false);
opt_loginput = false;
}
#endif /* HAVE_CURSES */
void default_save_file(char *filename)
{
if (default_config && *default_config) {
strcpy(filename, default_config);
return;
}
#if defined(unix) || defined(__APPLE__)
if (getenv("HOME") && *getenv("HOME")) {
strcpy(filename, getenv("HOME"));
strcat(filename, "/");
}
else
strcpy(filename, "");
strcat(filename, ".sgminer/");
mkdir(filename, 0777);
#else
strcpy(filename, "");
#endif
strcat(filename, def_conf);
}
#ifdef HAVE_CURSES
static void set_options(void)
{
int selected;
char input;
opt_loginput = true;
immedok(logwin, true);
clear_logwin();
retry:
wlogprint("[Q]ueue: %d\n[S]cantime: %d\n[E]xpiry: %d\n"
"[W]rite config file\n[R]estart\n",
opt_queue, opt_scantime, opt_expiry);
wlogprint("Select an option or any other key to return\n");
logwin_update();
input = getch();
if (!strncasecmp(&input, "q", 1)) {
selected = curses_int("Extra work items to queue");
if (selected < 0 || selected > 9999) {
wlogprint("Invalid selection\n");
goto retry;
}
opt_queue = selected;
goto retry;
} else if (!strncasecmp(&input, "s", 1)) {
selected = curses_int("Set scantime in seconds");
if (selected < 0 || selected > 9999) {
wlogprint("Invalid selection\n");
goto retry;
}
opt_scantime = selected;
goto retry;
} else if (!strncasecmp(&input, "e", 1)) {
selected = curses_int("Set expiry time in seconds");
if (selected < 0 || selected > 9999) {
wlogprint("Invalid selection\n");
goto retry;
}
opt_expiry = selected;
goto retry;
} else if (!strncasecmp(&input, "w", 1)) {
char *str, filename[PATH_MAX], prompt[PATH_MAX + 50];
default_save_file(filename);
snprintf(prompt, sizeof(prompt), "Config filename to write (Enter for default) [%s]", filename);
str = curses_input(prompt);
if (strcmp(str, "-1")) {
struct stat statbuf;
strcpy(filename, str);
free(str);
if (!stat(filename, &statbuf)) {
wlogprint("File exists, overwrite?\n");
input = getch();
if (strncasecmp(&input, "y", 1))
goto retry;
}
}
else
free(str);
write_config(filename);
goto retry;
} else if (!strncasecmp(&input, "r", 1)) {
wlogprint("Are you sure?\n");
input = getch();
if (!strncasecmp(&input, "y", 1))
app_restart();
else
clear_logwin();
} else
clear_logwin();
immedok(logwin, false);
opt_loginput = false;
}
static void *input_thread(void __maybe_unused *userdata)
{
pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
RenameThread("Input");
if (!curses_active)
return NULL;
while (1) {
char input;
input = getch();
if (!strncasecmp(&input, "q", 1)) {
kill_work();
return NULL;
} else if (!strncasecmp(&input, "d", 1))
display_options();
else if (!strncasecmp(&input, "p", 1))
display_pools();
else if (!strncasecmp(&input, "s", 1))
set_options();
else if (!strncasecmp(&input, "g", 1))
manage_gpu();
if (opt_realquiet) {
disable_curses();
break;
}
}
return NULL;
}
#endif /* HAVE_CURSES */
static void *api_thread(void *userdata)
{
struct thr_info *mythr = (struct thr_info *)userdata;
pthread_detach(pthread_self());
pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
RenameThread("API");
set_lowprio();
api(api_thr_id);
PTH(mythr) = 0L;
return NULL;
}
/* Sole work devices are serialised wrt calling get_work so they report in on
* each pass through their scanhash function as well as in get_work whereas
* queued work devices work asynchronously so get them to report in and out
* only across get_work. */
static void thread_reportin(struct thr_info *thr)
{
cgtime(&thr->last);
thr->getwork = false;
thr->cgpu->status = LIFE_WELL;
thr->cgpu->device_last_well = time(NULL);
}
/* Tell the watchdog thread this thread is waiting on get work and should not
* be restarted */
static void thread_reportout(struct thr_info *thr)
{
cgtime(&thr->last);
thr->getwork = true;
thr->cgpu->status = LIFE_WELL;
thr->cgpu->device_last_well = time(NULL);
}
static void hashmeter(int thr_id, struct timeval *diff,
uint64_t hashes_done)
{
struct timeval temp_tv_end, total_diff;
double secs;
double local_secs;
static double local_mhashes_done = 0;
double local_mhashes;
bool showlog = false;
char displayed_hashes[16], displayed_rolling[16];
uint64_t dh64, dr64;
struct thr_info *thr = NULL;
local_mhashes = (double)hashes_done / 1000000.0;
/* Update the last time this thread reported in */
rd_lock(&mining_thr_lock);
if (thr_id >= 0 && thr_id < mining_threads) {
thr = mining_thr[thr_id];
cgtime(&thr->last);
thr->cgpu->device_last_well = time(NULL);
}
rd_unlock(&mining_thr_lock);
secs = (double)diff->tv_sec + ((double)diff->tv_usec / 1000000.0);
/* So we can call hashmeter from a non worker thread */
if (thr) {
struct cgpu_info *cgpu = thr->cgpu;
double thread_rolling = 0.0;
int i;
applog(LOG_DEBUG, "[thread %d: %"PRIu64" hashes, %.1f khash/sec]",
thr_id, hashes_done, hashes_done / 1000 / secs);
/* Rolling average for each thread and each device */
decay_time(&thr->rolling, local_mhashes / secs, secs);
for (i = 0; i < cgpu->threads; i++)
thread_rolling += cgpu->thr[i]->rolling;
mutex_lock(&hash_lock);
decay_time(&cgpu->rolling, thread_rolling, secs);
cgpu->total_mhashes += local_mhashes;
mutex_unlock(&hash_lock);
// If needed, output detailed, per-device stats
if (want_per_device_stats) {
struct timeval now;
struct timeval elapsed;
cgtime(&now);
timersub(&now, &thr->cgpu->last_message_tv, &elapsed);
if (opt_log_interval <= elapsed.tv_sec) {
struct cgpu_info *cgpu = thr->cgpu;
char logline[255];
cgpu->last_message_tv = now;
get_statline(logline, sizeof(logline), cgpu);
if (!curses_active) {
printf("%s \r", logline);
fflush(stdout);
} else
applog(LOG_INFO, "%s", logline);
}
}
}
/* Totals are updated by all threads so can race without locking */
mutex_lock(&hash_lock);
cgtime(&temp_tv_end);
timersub(&temp_tv_end, &total_tv_end, &total_diff);
total_mhashes_done += local_mhashes;
local_mhashes_done += local_mhashes;
/* Only update with opt_log_interval */
if (total_diff.tv_sec < opt_log_interval)
goto out_unlock;
showlog = true;
cgtime(&total_tv_end);
local_secs = (double)total_diff.tv_sec + ((double)total_diff.tv_usec / 1000000.0);
decay_time(&total_rolling, local_mhashes_done / local_secs, local_secs);
global_hashrate = ((unsigned long long)lround(total_rolling)) * 1000000;
timersub(&total_tv_end, &total_tv_start, &total_diff);
total_secs = (double)total_diff.tv_sec +
((double)total_diff.tv_usec / 1000000.0);
dh64 = (double)total_mhashes_done / total_secs * 1000000ull;
dr64 = (double)total_rolling * 1000000ull;
suffix_string(dh64, displayed_hashes, sizeof(displayed_hashes), 4);
suffix_string(dr64, displayed_rolling, sizeof(displayed_rolling), 4);
snprintf(statusline, sizeof(statusline),
"%s(%ds):%s (avg):%sh/s | A:%.0f R:%.0f HW:%d WU:%.3f/m",
want_per_device_stats ? "ALL " : "",
opt_log_interval, displayed_rolling, displayed_hashes,
total_diff_accepted, total_diff_rejected, hw_errors,
total_diff1 / total_secs * 60);
local_mhashes_done = 0;
out_unlock:
mutex_unlock(&hash_lock);
if (showlog) {
if (!curses_active) {
printf("%s \r", statusline);
fflush(stdout);
} else
applog(LOG_INFO, "%s", statusline);
}
}
static void stratum_share_result(json_t *val, json_t *res_val, json_t *err_val,
struct stratum_share *sshare)
{
struct work *work = sshare->work;
time_t now_t = time(NULL);
char hashshow[64];
int srdiff;
srdiff = now_t - sshare->sshare_sent;
if (opt_debug || srdiff > 0) {
applog(LOG_INFO, "Pool %d stratum share result lag time %d seconds",
work->pool->pool_no, srdiff);
}
show_hash(work, hashshow);
share_result(val, res_val, err_val, work, hashshow, false, "");
}
/* Parses stratum json responses and tries to find the id that the request
* matched to and treat it accordingly. */
static bool parse_stratum_response(struct pool *pool, char *s)
{
json_t *val = NULL, *err_val, *res_val, *id_val;
struct stratum_share *sshare;
json_error_t err;
bool ret = false;
int id;
val = JSON_LOADS(s, &err);
if (!val) {
applog(LOG_INFO, "JSON decode failed(%d): %s", err.line, err.text);
goto out;
}
res_val = json_object_get(val, "result");
err_val = json_object_get(val, "error");
id_val = json_object_get(val, "id");
if (json_is_null(id_val) || !id_val) {
char *ss;
if (err_val)
ss = json_dumps(err_val, JSON_INDENT(3));
else
ss = strdup("(unknown reason)");
applog(LOG_INFO, "JSON-RPC non method decode failed: %s", ss);
free(ss);
goto out;
}
id = json_integer_value(id_val);
mutex_lock(&sshare_lock);
HASH_FIND_INT(stratum_shares, &id, sshare);
if (sshare) {
HASH_DEL(stratum_shares, sshare);
pool->sshares--;
}
mutex_unlock(&sshare_lock);
if (!sshare) {
double pool_diff;
/* Since the share is untracked, we can only guess at what the
* work difficulty is based on the current pool diff. */
cg_rlock(&pool->data_lock);
pool_diff = pool->swork.diff;
cg_runlock(&pool->data_lock);
if (json_is_true(res_val)) {
applog(LOG_NOTICE, "Accepted untracked stratum share from %s", get_pool_name(pool));
/* We don't know what device this came from so we can't
* attribute the work to the relevant cgpu */
mutex_lock(&stats_lock);
total_accepted++;
pool->accepted++;
total_diff_accepted += pool_diff;
pool->diff_accepted += pool_diff;
mutex_unlock(&stats_lock);
} else {
applog(LOG_NOTICE, "Rejected untracked stratum share from %s", get_pool_name(pool));
mutex_lock(&stats_lock);
total_rejected++;
pool->rejected++;
total_diff_rejected += pool_diff;
pool->diff_rejected += pool_diff;
mutex_unlock(&stats_lock);
}
goto out;
}
stratum_share_result(val, res_val, err_val, sshare);
free_work(sshare->work);
free(sshare);
ret = true;
out:
if (val)
json_decref(val);
return ret;
}
void clear_stratum_shares(struct pool *pool)
{
struct stratum_share *sshare, *tmpshare;
double diff_cleared = 0;
int cleared = 0;
mutex_lock(&sshare_lock);
HASH_ITER(hh, stratum_shares, sshare, tmpshare) {
if (sshare->work->pool == pool) {
HASH_DEL(stratum_shares, sshare);
diff_cleared += sshare->work->work_difficulty;
free_work(sshare->work);
pool->sshares--;
free(sshare);
cleared++;
}
}
mutex_unlock(&sshare_lock);
if (cleared) {
applog(LOG_WARNING, "Lost %d shares due to stratum disconnect on %s", cleared, get_pool_name(pool));
pool->stale_shares += cleared;
total_stale += cleared;
pool->diff_stale += diff_cleared;
total_diff_stale += diff_cleared;
}
}
void clear_pool_work(struct pool *pool)
{
struct work *work, *tmp;
int cleared = 0;
mutex_lock(stgd_lock);
HASH_ITER(hh, staged_work, work, tmp) {
if (work->pool == pool) {
HASH_DEL(staged_work, work);
free_work(work);
cleared++;
}
}
mutex_unlock(stgd_lock);
if (cleared)
applog(LOG_INFO, "Cleared %d work items due to stratum disconnect on pool %d", cleared, pool->pool_no);
}
static int cp_prio(void)
{
int prio;
cg_rlock(&control_lock);
prio = currentpool->prio;
cg_runlock(&control_lock);
return prio;
}
/* We only need to maintain a secondary pool connection when we need the
* capacity to get work from the backup pools while still on the primary */
static bool cnx_needed(struct pool *pool)
{
struct pool *cp;
if (pool->state != POOL_ENABLED)
return false;
/* Balance strategies need all pools online */
if (pool_strategy == POOL_BALANCE)
return true;
if (pool_strategy == POOL_LOADBALANCE)
return true;
/* Idle stratum pool needs something to kick it alive again */
if (pool->has_stratum && pool->idle)
return true;
/* Getwork pools without opt_fail_only need backup pools up to be able
* to leak shares */
cp = current_pool();
if (cp == pool)
return true;
if (!pool_localgen(cp) && (!opt_fail_only || !cp->hdr_path))
return true;
/* If we're waiting for a response from shares submitted, keep the
* connection open. */
if (pool->sshares)
return true;
/* If the pool has only just come to life and is higher priority than
* the current pool keep the connection open so we can fail back to
* it. */
if (pool_strategy == POOL_FAILOVER && pool->prio < cp_prio())
return true;
if (pool_unworkable(cp))
return true;
/* We've run out of work, bring anything back to life. */
if (no_work)
return true;
return false;
}
static void wait_lpcurrent(struct pool *pool);
static void pool_resus(struct pool *pool);
static void gen_stratum_work(struct pool *pool, struct work *work);
static void stratum_resumed(struct pool *pool)
{
if (!pool->stratum_notify)
return;
if (pool_tclear(pool, &pool->idle)) {
applog(LOG_INFO, "Stratum connection to %s resumed", get_pool_name(pool));
pool_resus(pool);
}
}
static bool supports_resume(struct pool *pool)
{
bool ret;
cg_rlock(&pool->data_lock);
ret = (pool->sessionid != NULL);
cg_runlock(&pool->data_lock);
return ret;
}
/* One stratum receive thread per pool that has stratum waits on the socket
* checking for new messages and for the integrity of the socket connection. We
* reset the connection based on the integrity of the receive side only as the
* send side will eventually expire data it fails to send. */
static void *stratum_rthread(void *userdata)
{
struct pool *pool = (struct pool *)userdata;
char threadname[16];
pthread_detach(pthread_self());
snprintf(threadname, sizeof(threadname), "%d/RStratum", pool->pool_no);
RenameThread(threadname);
while (42) {
struct timeval timeout;
int sel_ret;
fd_set rd;
char *s;
if (unlikely(pool->removed))
break;
/* Check to see whether we need to maintain this connection
* indefinitely or just bring it up when we switch to this
* pool */
if (!sock_full(pool) && !cnx_needed(pool)) {
applog(LOG_INFO, "Suspending stratum on %s",
get_pool_name(pool));
suspend_stratum(pool);
clear_stratum_shares(pool);
clear_pool_work(pool);
wait_lpcurrent(pool);
if (!restart_stratum(pool)) {
pool_died(pool);
while (!restart_stratum(pool)) {
pool_failed(pool);
if (pool->removed)
goto out;
cgsleep_ms(30000);
}
}
}
FD_ZERO(&rd);
FD_SET(pool->sock, &rd);
timeout.tv_sec = 90;
timeout.tv_usec = 0;
/* The protocol specifies that notify messages should be sent
* every minute so if we fail to receive any for 90 seconds we
* assume the connection has been dropped and treat this pool
* as dead */
if (!sock_full(pool) && (sel_ret = select(pool->sock + 1, &rd, NULL, NULL, &timeout)) < 1) {
applog(LOG_DEBUG, "Stratum select failed on %s with value %d", get_pool_name(pool), sel_ret);
s = NULL;
} else
s = recv_line(pool);
if (!s) {
applog(LOG_NOTICE, "Stratum connection to %s interrupted", get_pool_name(pool));
pool->getfail_occasions++;
total_go++;
/* If the socket to our stratum pool disconnects, all
* tracked submitted shares are lost and we will leak
* the memory if we don't discard their records. */
if (!supports_resume(pool) || opt_lowmem)
clear_stratum_shares(pool);
clear_pool_work(pool);
if (pool == current_pool())
restart_threads();
if (restart_stratum(pool))
continue;
pool_died(pool);
while (!restart_stratum(pool)) {
pool_failed(pool);
if (pool->removed)
goto out;
cgsleep_ms(30000);
}
stratum_resumed(pool);
continue;
}
/* Check this pool hasn't died while being a backup pool and
* has not had its idle flag cleared */
stratum_resumed(pool);
if (!parse_method(pool, s) && !parse_stratum_response(pool, s))
applog(LOG_INFO, "Unknown stratum msg: %s", s);
else if (pool->swork.clean) {
struct work *work = make_work();
/* Generate a single work item to update the current
* block database */
pool->swork.clean = false;
gen_stratum_work(pool, work);
work->longpoll = true;
/* Return value doesn't matter. We're just informing
* that we may need to restart. */
test_work_current(work);
free_work(work);
}
free(s);
}
out:
return NULL;
}
/* Each pool has one stratum send thread for sending shares to avoid many
* threads being created for submission since all sends need to be serialised
* anyway. */
static void *stratum_sthread(void *userdata)
{
struct pool *pool = (struct pool *)userdata;
char threadname[16];
pthread_detach(pthread_self());
snprintf(threadname, sizeof(threadname), "%d/SStratum", pool->pool_no);
RenameThread(threadname);
pool->stratum_q = tq_new();
if (!pool->stratum_q)
quit(1, "Failed to create stratum_q in stratum_sthread");
while (42) {
char noncehex[12], nonce2hex[33], s[1024] = { 0 };
struct stratum_share *sshare;
uint32_t *hash32, nonce;
unsigned char nonce2[16];
struct work *work;
bool submitted;
if (unlikely(pool->removed)) {
break;
}
work = (struct work *)tq_pop(pool->stratum_q, NULL);
if (unlikely(!work))
quit(1, "Stratum q returned empty work");
if (unlikely(work->nonce2_len > 8)) {
applog(LOG_ERR, "%s asking for inappropriately long nonce2 length %d", get_pool_name(pool), (int)work->nonce2_len);
applog(LOG_ERR, "Not attempting to submit shares");
free_work(work);
continue;
}
// TODO: check for memory leaks
sshare = (struct stratum_share *)calloc(sizeof(struct stratum_share), 1);
hash32 = (uint32_t *)work->hash;
submitted = false;
sshare->sshare_time = time(NULL);
/* This work item is freed in parse_stratum_response */
sshare->work = work;
applog(LOG_DEBUG, "stratum_sthread() algorithm = %s", pool->algorithm.name);
// Neoscrypt is little endian
if (pool->algorithm.type == ALGO_NEOSCRYPT) {
nonce = htobe32(*((uint32_t *)(work->data + 76)));
//*((uint32_t *)nonce2) = htole32(work->nonce2);
}
else if (pool->algorithm.type == ALGO_DECRED) {
nonce = *((uint32_t *)(work->data + 140));
}
else {
nonce = *((uint32_t *)(work->data + 76));
}
__bin2hex(noncehex, (const unsigned char *)&nonce, 4);
*((uint64_t *)nonce2) = htole64(work->nonce2);
__bin2hex(nonce2hex, nonce2, work->nonce2_len);
memset(s, 0, 1024);
mutex_lock(&sshare_lock);
/* Give the stratum share a unique id */
sshare->id = swork_id++;
mutex_unlock(&sshare_lock);
if (pool->algorithm.type == ALGO_DECRED && opt_vote) {
snprintf(s, sizeof(s),
"{\"params\": [\"%s\", \"%s\", \"%s\", \"%s\", \"%s\", \"%04x\"], \"id\": %d, \"method\": \"mining.submit\"}",
pool->rpc_user, work->job_id, nonce2hex, work->ntime, noncehex, (opt_vote << 1) | 1, sshare->id);
} else {
snprintf(s, sizeof(s),
"{\"params\": [\"%s\", \"%s\", \"%s\", \"%s\", \"%s\"], \"id\": %d, \"method\": \"mining.submit\"}",
pool->rpc_user, work->job_id, nonce2hex, work->ntime, noncehex, sshare->id);
}
applog(LOG_INFO, "Submitting share %08lx to %s", (long unsigned int)htole32(hash32[6]), get_pool_name(pool));
/* Try resubmitting for up to 2 minutes if we fail to submit
* once and the stratum pool nonce1 still matches suggesting
* we may be able to resume. */
while (time(NULL) < sshare->sshare_time + 120) {
bool sessionid_match;
mutex_lock(&sshare_lock);
if (likely(stratum_send(pool, s, strlen(s)))) {
int ssdiff;
if (pool_tclear(pool, &pool->submit_fail))
applog(LOG_WARNING, "%s communication resumed, submitting work", get_pool_name(pool));
sshare->sshare_sent = time(NULL);
ssdiff = sshare->sshare_sent - sshare->sshare_time;
if (opt_debug || ssdiff > 0) {
applog(LOG_INFO, "Pool %d stratum share submission lag time %d seconds",
pool->pool_no, ssdiff);
}
HASH_ADD_INT(stratum_shares, id, sshare);
pool->sshares++;
mutex_unlock(&sshare_lock);
applog(LOG_DEBUG, "Successfully submitted, adding to stratum_shares db");
submitted = true;
break;
}
else {
mutex_unlock(&sshare_lock);
}
if (!pool_tset(pool, &pool->submit_fail) && cnx_needed(pool)) {
applog(LOG_WARNING, "%s stratum share submission failure", get_pool_name(pool));
total_ro++;
pool->remotefail_occasions++;
}
if (opt_lowmem) {
applog(LOG_DEBUG, "Lowmem option prevents resubmitting stratum share");
break;
}
cg_rlock(&pool->data_lock);
sessionid_match = (pool->nonce1 && !strcmp(work->nonce1, pool->nonce1));
cg_runlock(&pool->data_lock);
if (!sessionid_match) {
applog(LOG_DEBUG, "No matching session id for resubmitting stratum share");
break;
}
/* Retry every 5 seconds */
sleep(5);
}
if (unlikely(!submitted)) {
applog(LOG_DEBUG, "Failed to submit stratum share, discarding");
free_work(work);
free(sshare);
pool->stale_shares++;
total_stale++;
}
}
/* Freeze the work queue but don't free up its memory in case there is
* work still trying to be submitted to the removed pool. */
tq_freeze(pool->stratum_q);
return NULL;
}
static void init_stratum_threads(struct pool *pool)
{
have_longpoll = true;
if (unlikely(pthread_create(&pool->stratum_sthread, NULL, stratum_sthread, (void *)pool)))
quit(1, "Failed to create stratum sthread");
if (unlikely(pthread_create(&pool->stratum_rthread, NULL, stratum_rthread, (void *)pool)))
quit(1, "Failed to create stratum rthread");
}
static void *longpoll_thread(void *userdata);
static bool stratum_works(struct pool *pool)
{
applog(LOG_INFO, "Testing %s stratum %s", get_pool_name(pool), pool->stratum_url);
if (!extract_sockaddr(pool->stratum_url, &pool->sockaddr_url, &pool->stratum_port))
return false;
if (!initiate_stratum(pool))
return false;
return true;
}
static bool pool_active(struct pool *pool, bool pinging)
{
struct timeval tv_getwork, tv_getwork_reply;
bool ret = false;
json_t *val;
CURL *curl;
char curl_err_str[CURL_ERROR_SIZE];
int rolltime = 0;
if (pool->has_gbt)
applog(LOG_DEBUG, "Retrieving block template from %s", get_pool_name(pool));
else
applog(LOG_INFO, "Testing %s", get_pool_name(pool));
/* This is the central point we activate stratum when we can */
retry_stratum:
if (pool->has_stratum) {
/* We create the stratum thread for each pool just after
* successful authorisation. Once the init flag has been set
* we never unset it and the stratum thread is responsible for
* setting/unsetting the active flag */
bool init = pool_tset(pool, &pool->stratum_init);
if (!init) {
bool ret = initiate_stratum(pool) && (!pool->extranonce_subscribe || subscribe_extranonce(pool)) && auth_stratum(pool);
if (ret)
init_stratum_threads(pool);
else
pool_tclear(pool, &pool->stratum_init);
return ret;
}
return pool->stratum_active;
}
curl = curl_easy_init();
if (unlikely(!curl)) {
applog(LOG_ERR, "CURL initialisation failed");
return false;
}
/* Probe for GBT support on first pass */
if (!pool->probed) {
applog(LOG_DEBUG, "Probing for GBT support");
val = json_rpc_call(curl, curl_err_str, pool->rpc_url, pool->rpc_userpass,
gbt_req, true, false, &rolltime, pool, false);
if (val) {
bool append = false, submit = false;
json_t *res_val, *mutables;
int i, mutsize = 0;
res_val = json_object_get(val, "result");
if (res_val) {
mutables = json_object_get(res_val, "mutable");
mutsize = json_array_size(mutables);
}
for (i = 0; i < mutsize; i++) {
json_t *arrval = json_array_get(mutables, i);
if (json_is_string(arrval)) {
const char *mut = json_string_value(arrval);
if (!strncasecmp(mut, "coinbase/append", 15))
append = true;
else if (!strncasecmp(mut, "submit/coinbase", 15))
submit = true;
}
}
json_decref(val);
/* Only use GBT if it supports coinbase append and
* submit coinbase */
if (append && submit) {
pool->has_gbt = true;
pool->rpc_req = gbt_req;
}
}
/* Reset this so we can probe fully just after this. It will be
* set to true that time.*/
pool->probed = false;
if (pool->has_gbt)
applog(LOG_DEBUG, "GBT coinbase + append support found, switching to GBT protocol");
else
applog(LOG_DEBUG, "No GBT coinbase + append support found, using getwork protocol");
}
cgtime(&tv_getwork);
val = json_rpc_call(curl, curl_err_str, pool->rpc_url, pool->rpc_userpass,
pool->rpc_req, true, false, &rolltime, pool, false);
cgtime(&tv_getwork_reply);
/* Detect if a http getwork pool has an X-Stratum header at startup,
* and if so, switch to that in preference to getwork if it works */
if (pool->stratum_url && !opt_fix_protocol && stratum_works(pool)) {
applog(LOG_NOTICE, "Switching %s to %s", get_pool_name(pool), pool->stratum_url);
if (!pool->rpc_url)
pool->rpc_url = strdup(pool->stratum_url);
pool->has_stratum = true;
curl_easy_cleanup(curl);
goto retry_stratum;
}
/* json_rpc_call() above succeeded */
if (val) {
struct work *work = make_work();
bool rc;
rc = work_decode(pool, work, val);
if (rc) {
applog(LOG_DEBUG, "Successfully retrieved and deciphered work from %s", get_pool_name(pool));
work->pool = pool;
work->rolltime = rolltime;
copy_time(&work->tv_getwork, &tv_getwork);
copy_time(&work->tv_getwork_reply, &tv_getwork_reply);
work->getwork_mode = GETWORK_MODE_TESTPOOL;
calc_diff(work, 0);
applog(LOG_DEBUG, "Pushing pooltest work to base pool");
stage_work(work);
total_getworks++;
pool->getwork_requested++;
ret = true;
cgtime(&pool->tv_idle);
} else {
applog(LOG_DEBUG, "Successfully retrieved but FAILED to decipher work from %s", get_pool_name(pool));
free_work(work);
}
json_decref(val);
if (pool->lp_url)
goto out;
/* Decipher the longpoll URL, if any, and store it in ->lp_url */
if (pool->hdr_path) {
char *copy_start, *hdr_path;
bool need_slash = false;
size_t siz;
hdr_path = pool->hdr_path;
if (strstr(hdr_path, "://")) {
pool->lp_url = hdr_path;
hdr_path = NULL;
} else {
/* absolute path, on current server */
copy_start = (*hdr_path == '/') ? (hdr_path + 1) : hdr_path;
if (pool->rpc_url[strlen(pool->rpc_url) - 1] != '/')
need_slash = true;
siz = strlen(pool->rpc_url) + strlen(copy_start) + 2;
pool->lp_url = (char *)malloc(siz);
if (!pool->lp_url) {
applog(LOG_ERR, "Malloc failure in pool_active");
goto out;
}
snprintf(pool->lp_url, siz, "%s%s%s", pool->rpc_url, need_slash ? "/" : "", copy_start);
}
} else
pool->lp_url = NULL;
if (!pool->lp_started) {
pool->lp_started = true;
if (unlikely(pthread_create(&pool->longpoll_thread, NULL, longpoll_thread, (void *)pool)))
quit(1, "Failed to create pool longpoll thread");
}
} else {
/* If we failed to parse a getwork, this could be a stratum
* url without the prefix stratum+tcp:// so let's check it */
if (initiate_stratum(pool)) {
pool->has_stratum = true;
goto retry_stratum;
}
applog(LOG_DEBUG, "FAILED to retrieve work from %s", get_pool_name(pool));
if (!pinging)
applog(LOG_WARNING, "%s slow/down or URL or credentials invalid", get_pool_name(pool));
}
out:
curl_easy_cleanup(curl);
return ret;
}
static void pool_resus(struct pool *pool)
{
if (pool_strategy == POOL_FAILOVER && pool->prio < cp_prio())
applog(LOG_WARNING, "%s alive, testing stability", get_pool_name(pool));
else
applog(LOG_INFO, "%s alive", get_pool_name(pool));
}
/* If this is called non_blocking, it will return NULL for work so that must
* be handled. */
static struct work *hash_pop(bool blocking)
{
struct work *work = NULL, *tmp;
int hc;
mutex_lock(stgd_lock);
if (!HASH_COUNT(staged_work)) {
if (!blocking)
goto out_unlock;
do {
struct timespec then;
struct timeval now;
int rc;
cgtime(&now);
then.tv_sec = now.tv_sec + 10;
then.tv_nsec = now.tv_usec * 1000;
pthread_cond_signal(&gws_cond);
rc = pthread_cond_timedwait(&getq->cond, stgd_lock, &then);
/* Check again for !no_work as multiple threads may be
* waiting on this condition and another may set the
* bool separately. */
if (rc && !no_work) {
no_work = true;
applog(LOG_WARNING, "Waiting for work to be available from pools.");
event_notify("idle");
}
} while (!HASH_COUNT(staged_work));
}
if (no_work) {
applog(LOG_WARNING, "Work available from pools, resuming.");
no_work = false;
}
hc = HASH_COUNT(staged_work);
/* Find clone work if possible, to allow masters to be reused */
if (hc > staged_rollable) {
HASH_ITER(hh, staged_work, work, tmp) {
if (!work_rollable(work))
break;
}
} else
work = staged_work;
HASH_DEL(staged_work, work);
if (work_rollable(work))
staged_rollable--;
/* Signal the getwork scheduler to look for more work */
pthread_cond_signal(&gws_cond);
/* Signal hash_pop again in case there are mutliple hash_pop waiters */
pthread_cond_signal(&getq->cond);
/* Keep track of last getwork grabbed */
last_getwork = time(NULL);
out_unlock:
mutex_unlock(stgd_lock);
return work;
}
void set_target(unsigned char *dest_target, double diff, double diff_multiplier2, const int thr_id)
{
unsigned char target[32];
uint64_t *data64, h64;
double d64, dcut64;
if (unlikely(diff == 0.0)) {
/* This shouldn't happen but best we check to prevent a crash */
applog(LOG_ERR, "[THR%d] Diff zero passed to set_target", thr_id);
diff = 1.0;
}
d64 = diff_multiplier2 * truediffone;
d64 /= diff;
dcut64 = d64 / bits192;
h64 = dcut64;
data64 = (uint64_t *)(target + 24);
*data64 = htole64(h64);
dcut64 = h64;
dcut64 *= bits192;
d64 -= dcut64;
dcut64 = d64 / bits128;
h64 = dcut64;
data64 = (uint64_t *)(target + 16);
*data64 = htole64(h64);
dcut64 = h64;
dcut64 *= bits128;
d64 -= dcut64;
dcut64 = d64 / bits64;
h64 = dcut64;
data64 = (uint64_t *)(target + 8);
*data64 = htole64(h64);
dcut64 = h64;
dcut64 *= bits64;
d64 -= dcut64;
h64 = d64;
data64 = (uint64_t *)(target);
*data64 = htole64(h64);
if (opt_debug) {
char *htarget = bin2hex(target, 32);
applog(LOG_DEBUG, "[THR%d] Generated target %s", thr_id, htarget);
free(htarget);
}
memcpy(dest_target, target, 32);
}
/*****************************************************
* Special set_target() function for Neoscrypt
****************************************************/
void set_target_neoscrypt(unsigned char *target, double diff, const int thr_id)
{
uint64_t m;
int k;
diff /= 65536.0;
for (k = 6; k > 0 && diff > 1.0; --k) {
diff /= 4294967296.0;
}
m = 4294901760.0 / diff;
if (m == 0 && k == 6) {
memset(target, 0xff, 32);
}
else {
memset(target, 0, 32);
((uint32_t *)target)[k] = (uint32_t)m;
((uint32_t *)target)[k + 1] = (uint32_t)(m >> 32);
}
if (opt_debug) {
/* The target is computed in this systems endianess and stored
* in its endianess on a uint32-level. But because the target are
* eight uint32s, they are stored in mixed mode, i.e., each uint32
* is stored in the local endianess, but the least significant bit
* is stored in target[0] bit 0.
*
* To print this large number in a native human readable form the
* order of the array entries is swapped, i.e., target[7] <-> target[0]
* and each array entry is byte swapped to have the least significant
* bit to the right. */
uint32_t swaped[8];
swab256(swaped, target);
char *htarget = bin2hex((unsigned char *)swaped, 32);
applog(LOG_DEBUG, "[THR%d] Generated neoscrypt target 0x%s", thr_id, htarget);
free(htarget);
}
}
/* Generates stratum based work based on the most recent notify information
* from the pool. This will keep generating work while a pool is down so we use
* other means to detect when the pool has died in stratum_thread */
static void gen_stratum_work(struct pool *pool, struct work *work)
{
unsigned char merkle_root[32], merkle_sha[64];
uint32_t *data32, *swap32;
uint64_t nonce2le;
int i, j;
cg_wlock(&pool->data_lock);
nonce2le = htole64(pool->nonce2);
if (pool->algorithm.type != ALGO_DECRED) {
/* Update coinbase. Always use an LE encoded nonce2 to fill in values
* from left to right and prevent overflow errors with small n2sizes */
memcpy(pool->coinbase + pool->nonce2_offset, &nonce2le, pool->n2size);
}
work->nonce2 = pool->nonce2++;
work->nonce2_len = pool->n2size;
/* Downgrade to a read lock to read off the pool variables */
cg_dwlock(&pool->data_lock);
if (pool->algorithm.type != ALGO_DECRED) {
/* Generate merkle root */
pool->algorithm.gen_hash(pool->coinbase, pool->swork.cb_len, merkle_root);
memcpy(merkle_sha, merkle_root, 32);
for (i = 0; i < pool->swork.merkles; i++) {
memcpy(merkle_sha + 32, pool->swork.merkle_bin[i], 32);
gen_hash(merkle_sha, 64, merkle_root);
memcpy(merkle_sha, merkle_root, 32);
}
}
applog(LOG_DEBUG, "[THR%d] gen_stratum_work() - algorithm = %s", work->thr_id, pool->algorithm.name);
// Different for Neoscrypt because of Little Endian
if (pool->algorithm.type == ALGO_NEOSCRYPT) {
/* Incoming data is in little endian. */
memcpy(merkle_root, merkle_sha, 32);
uint32_t temp = pool->merkle_offset / sizeof(uint32_t), i;
/* Put version (4 byte) + prev_hash (4 byte* 8) but big endian encoded
* into work. */
for (i = 0; i < temp; ++i) {
((uint32_t *)work->data)[i] = be32toh(((uint32_t *)pool->header_bin)[i]);
}
/* Now add the merkle_root (4 byte* 8), but it is encoded in little endian. */
temp += 8;
for (j = 0; i < temp; ++i, ++j) {
((uint32_t *)work->data)[i] = le32toh(((uint32_t *)merkle_root)[j]);
}
/* Add the time encoded in big endianess. */
hex2bin((unsigned char *)&temp, pool->swork.ntime, 4);
/* Add the nbits (big endianess). */
((uint32_t *)work->data)[17] = be32toh(temp);
hex2bin((unsigned char *)&temp, pool->swork.nbit, 4);
((uint32_t *)work->data)[18] = be32toh(temp);
((uint32_t *)work->data)[20] = 0x80000000;
((uint32_t *)work->data)[31] = 0x00000280;
}
else if (pool->algorithm.type == ALGO_DECRED) {
uint16_t vote = (uint16_t) (opt_vote << 1) | 1;
size_t nonce2_offset = MIN(pool->n1_len, 36);
memcpy(work->data, pool->header_bin, 4); // version
flip32(work->data + 4, pool->header_bin + 4); // prevhash
memcpy(work->data + 4 + 32, pool->coinbase, MIN((int)pool->swork.cb_len, 108));
memcpy(work->data + 100, &vote, 2);
for (i = 36; i < 45; i++)
((uint32_t *)work->data)[i] = 0;
memcpy(work->data + 144, pool->nonce1bin, nonce2_offset);
memcpy(work->data + 144 + nonce2_offset, &nonce2le, pool->n2size);
}
else {
data32 = (uint32_t *)merkle_sha;
swap32 = (uint32_t *)merkle_root;
flip32(swap32, data32);
/* Copy the data template from header_bin */
memcpy(work->data, pool->header_bin, 128);
memcpy(work->data + pool->merkle_offset, merkle_root, 32);
}
/* Store the stratum work diff to check it still matches the pool's
* stratum diff when submitting shares */
work->sdiff = pool->swork.diff;
/* Copy parameters required for share submission */
work->job_id = strdup(pool->swork.job_id);
work->nonce1 = strdup(pool->nonce1);
work->ntime = strdup(pool->swork.ntime);
cg_runlock(&pool->data_lock);
if (opt_debug) {
char *header, *merkle_hash;
int datasize = 128;
if (pool->algorithm.type == ALGO_DECRED) datasize = 180;
header = bin2hex(work->data, datasize);
if (pool->algorithm.type != ALGO_DECRED) {
merkle_hash = bin2hex((const unsigned char *)merkle_root, 32);
applog(LOG_DEBUG, "[THR%d] Generated stratum merkle %s", work->thr_id, merkle_hash);
free(merkle_hash);
}
applog(LOG_DEBUG, "[THR%d] Generated stratum header %s", work->thr_id, header);
applog(LOG_DEBUG, "[THR%d] Work job_id %s nonce2 %"PRIu64" ntime %s", work->thr_id, work->job_id,
work->nonce2, work->ntime);
free(header);
}
// For Neoscrypt use set_target_neoscrypt() function
if (pool->algorithm.type == ALGO_NEOSCRYPT) {
set_target_neoscrypt(work->target, work->sdiff, work->thr_id);
} else {
if (pool->algorithm.calc_midstate) pool->algorithm.calc_midstate(work);
set_target(work->target, work->sdiff, pool->algorithm.diff_multiplier2, work->thr_id);
}
local_work++;
work->pool = pool;
work->stratum = true;
work->blk.nonce = 0;
work->id = total_work++;
work->longpoll = false;
work->getwork_mode = GETWORK_MODE_STRATUM;
work->work_block = work_block;
/* Nominally allow a driver to ntime roll 60 seconds */
work->drv_rolllimit = 60;
calc_diff(work, work->sdiff);
cgtime(&work->tv_staged);
}
static void enable_devices(void)
{
int i;
//enable/disable devices as needed
if(opt_devs_enabled)
{
for (i = 0; i < total_devices; i++)
{
//device should be enabled
if(devices_enabled[i])
enable_device(i);
else
{
//mark as disabled
devices[i]->deven = DEV_DISABLED;
}
}
}
//enable all devices
else
{
for (i = 0; i < total_devices; ++i)
enable_device(i);
}
}
static void apply_initial_gpu_settings(struct pool *pool)
{
int i;
const char *opt;
unsigned long options; //gpu adl options to apply
unsigned int needed_threads = 0; //number of mining threads needed after we change devices
applog(LOG_NOTICE, "Startup Pool No = %d", pool->pool_no);
//get compare options
options = compare_pool_settings(NULL, pool);
//apply gpu settings
rd_lock(&mining_thr_lock);
apply_switcher_options(options, pool);
/*
//reset devices
opt_devs_enabled = 0;
for (i = 0; i < MAX_DEVICES; i++)
devices_enabled[i] = false;
//assign pool devices if any
if(!empty_string((opt = get_pool_setting(pool->devices, ((!empty_string(default_profile.devices))?default_profile.devices:"all"))))) {
set_devices((char *)opt);
}
//lookup gap
if(!empty_string((opt = get_pool_setting(pool->lookup_gap, default_profile.lookup_gap))))
set_lookup_gap((char *)opt);
//set intensity
if(!empty_string((opt = get_pool_setting(pool->rawintensity, default_profile.rawintensity)))) {
set_rawintensity((char *)opt);
}
else if(!empty_string((opt = get_pool_setting(pool->xintensity, default_profile.xintensity)))) {
set_xintensity((char *)opt);
}
else if(!empty_string((opt = get_pool_setting(pool->intensity, ((!empty_string(default_profile.intensity))?default_profile.intensity:"8"))))) {
set_intensity((char *)opt);
}
//shaders
if(!empty_string((opt = get_pool_setting(pool->shaders, default_profile.shaders))))
set_shaders((char *)opt);
//thread-concurrency
// neoscrypt - if not specified set TC to 0 so that TC will be calculated by intensity settings
if (pool->algorithm.type == ALGO_NEOSCRYPT) {
opt = ((empty_string(pool->thread_concurrency))?"0":get_pool_setting(pool->thread_concurrency, default_profile.thread_concurrency));
}
// otherwise use pool/profile setting or default to default profile setting
else {
opt = get_pool_setting(pool->thread_concurrency, default_profile.thread_concurrency);
}
if (!empty_string(opt)) {
set_thread_concurrency(opt);
}
//worksize
if(!empty_string((opt = get_pool_setting(pool->worksize, default_profile.worksize))))
set_worksize(opt);
*/
//manually apply algorithm
for (i = 0; i < nDevs; i++)
{
applog(LOG_DEBUG, "Set GPU %d to %s", i, isnull(pool->algorithm.name, ""));
gpus[i].algorithm = pool->algorithm;
}
/*
#ifdef HAVE_ADL
options = APPLY_ENGINE | APPLY_MEMCLOCK | APPLY_FANSPEED | APPLY_POWERTUNE | APPLY_VDDC;
//GPU clock
if(!empty_string((opt = get_pool_setting(pool->gpu_engine, default_profile.gpu_engine))))
set_gpu_engine((char *)opt);
else
options ^= APPLY_ENGINE;
//GPU memory clock
if(!empty_string((opt = get_pool_setting(pool->gpu_memclock, default_profile.gpu_memclock))))
set_gpu_memclock((char *)opt);
else
options ^= APPLY_MEMCLOCK;
//GPU fans
if(!empty_string((opt = get_pool_setting(pool->gpu_fan, default_profile.gpu_fan))))
set_gpu_fan((char *)opt);
else
options ^= APPLY_FANSPEED;
//GPU powertune
if(!empty_string((opt = get_pool_setting(pool->gpu_powertune, default_profile.gpu_powertune))))
set_gpu_powertune((char *)opt);
else
options ^= APPLY_POWERTUNE;
//GPU vddc
if(!empty_string((opt = get_pool_setting(pool->gpu_vddc, default_profile.gpu_vddc))))
set_gpu_vddc((char *)opt);
else
options ^= APPLY_VDDC;
//apply gpu settings
for (i = 0; i < nDevs; i++)
{
if(opt_isset(options, APPLY_ENGINE))
set_engineclock(i, gpus[i].min_engine);
if(opt_isset(options, APPLY_MEMCLOCK))
set_memoryclock(i, gpus[i].gpu_memclock);
if(opt_isset(options, APPLY_FANSPEED))
set_fanspeed(i, gpus[i].min_fan);
if(opt_isset(options, APPLY_POWERTUNE))
set_powertune(i, gpus[i].gpu_powertune);
if(opt_isset(options, APPLY_VDDC))
set_vddc(i, gpus[i].gpu_vddc);
}
#endif*/
rd_unlock(&mining_thr_lock);
//enable/disable devices as needed
enable_devices();
//recount the number of needed mining threads
#ifdef HAVE_ADL
if(!empty_string((opt = get_pool_setting(pool->gpu_threads, default_profile.gpu_threads))))
set_gpu_threads((char *)opt);
rd_lock(&devices_lock);
for (i = 0; i < total_devices; i++)
if (!opt_removedisabled || !opt_devs_enabled || devices_enabled[i])
needed_threads += devices[i]->threads;
rd_unlock(&devices_lock);
#else
needed_threads = mining_threads;
#endif
//bad thread count?
if(needed_threads == 0)
quit(1, "No GPUs Initialized.");
restart_mining_threads(needed_threads);
}
static unsigned long compare_pool_settings(struct pool *oldpool, struct pool *newpool)
{
unsigned int options = 0;
const char *opt1, *opt2;
applog(LOG_DEBUG, "compare_pool_settings()");
if (!newpool) {
return 0;
}
//compare pool devices
opt1 = ((oldpool)?get_pool_setting(oldpool->devices, ((!empty_string(default_profile.devices))?default_profile.devices:"all")):"");
opt2 = get_pool_setting(newpool->devices, ((!empty_string(default_profile.devices))?default_profile.devices:"all"));
//changing devices means a hard reset of mining threads
if (strcasecmp(opt1, opt2) != 0) {
options |= (SWITCHER_APPLY_DEVICE | SWITCHER_HARD_RESET);
}
//compare gpu threads
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_threads, default_profile.gpu_threads):"");
opt2 = get_pool_setting(newpool->gpu_threads, default_profile.gpu_threads);
//changing gpu threads means a hard reset of mining threads
if (strcasecmp(opt1, opt2) != 0) {
options |= (SWITCHER_APPLY_GT | SWITCHER_HARD_RESET);
}
//compare algorithm
if ((oldpool && !cmp_algorithm(&oldpool->algorithm, &newpool->algorithm)) || (!oldpool)) {
options |= (SWITCHER_APPLY_ALGO | SWITCHER_SOFT_RESET);
}
//lookup gap
opt1 = ((oldpool)?get_pool_setting(oldpool->lookup_gap, default_profile.lookup_gap):"");
opt2 = get_pool_setting(newpool->lookup_gap, default_profile.lookup_gap);
//lookup gap means soft reset but only if hard reset isnt set
if (strcasecmp(opt1, opt2) != 0) {
options |= (SWITCHER_APPLY_LG | SWITCHER_SOFT_RESET);
}
// Intensity - First determine the intensity type that we are going to use with the new pool
unsigned int intoptions = 0;
if (!empty_string(newpool->rawintensity)) {
intoptions = SWITCHER_APPLY_RAWINT | SWITCHER_SOFT_RESET;
opt1 = ((oldpool && !empty_string(oldpool->rawintensity))?oldpool->rawintensity:"");
opt2 = get_pool_setting(newpool->rawintensity, default_profile.rawintensity);
}
else if (!empty_string(newpool->xintensity)) {
intoptions = SWITCHER_APPLY_XINT | SWITCHER_SOFT_RESET;
opt1 = ((oldpool && !empty_string(oldpool->xintensity))?oldpool->xintensity:"");
opt2 = get_pool_setting(newpool->xintensity, default_profile.xintensity);
}
else {
intoptions = SWITCHER_APPLY_INT | SWITCHER_SOFT_RESET;
opt1 = ((oldpool && !empty_string(oldpool->intensity))?oldpool->intensity:"");
opt2 = get_pool_setting(newpool->intensity, default_profile.intensity);
}
// if old intensity and new intensity different, set flags to update
if(strcasecmp(opt1, opt2) != 0) {
// in case we compared 2 empty strings make sure new intensity is not empty
if (!empty_string(opt2)) {
options |= intoptions;
}
}
//shaders
opt1 = ((oldpool)?get_pool_setting(oldpool->shaders, default_profile.shaders):"");
opt2 = get_pool_setting(newpool->shaders, default_profile.shaders);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= (SWITCHER_APPLY_SHADER | SWITCHER_SOFT_RESET);
}
//thread-concurrency
// neoscrypt - if not specified set TC to 0 so that TC will be calculated by intensity settings
if (newpool->algorithm.type == ALGO_NEOSCRYPT) {
opt2 = ((empty_string(newpool->thread_concurrency))?"0":get_pool_setting(newpool->thread_concurrency, default_profile.thread_concurrency));
}
// otherwise use pool/profile setting or default to default profile setting
else {
opt2 = get_pool_setting(newpool->thread_concurrency, default_profile.thread_concurrency);
}
opt1 = ((oldpool)?get_pool_setting(oldpool->thread_concurrency, default_profile.thread_concurrency):"");
//thread-concurrency is soft reset
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= (SWITCHER_APPLY_TC | SWITCHER_SOFT_RESET);
}
//worksize
opt1 = ((oldpool)?get_pool_setting(oldpool->worksize, default_profile.worksize):"");
opt2 = get_pool_setting(newpool->worksize, default_profile.worksize);
//worksize is soft reset
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= (SWITCHER_APPLY_WORKSIZE | SWITCHER_SOFT_RESET);
}
#ifdef HAVE_ADL
//gpu-engine
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_engine, default_profile.gpu_engine):"");
opt2 = get_pool_setting(newpool->gpu_engine, default_profile.gpu_engine);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= SWITCHER_APPLY_GPU_ENGINE;
}
//gpu-memclock
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_memclock, default_profile.gpu_memclock):"");
opt2 = get_pool_setting(newpool->gpu_memclock, default_profile.gpu_memclock);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= SWITCHER_APPLY_GPU_MEMCLOCK;
}
//GPU fans
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_fan, default_profile.gpu_fan):"");
opt2 = get_pool_setting(newpool->gpu_fan, default_profile.gpu_fan);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= SWITCHER_APPLY_GPU_FAN;
}
//GPU powertune
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_powertune, default_profile.gpu_powertune):"");
opt2 = get_pool_setting(newpool->gpu_powertune, default_profile.gpu_powertune);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= SWITCHER_APPLY_GPU_POWERTUNE;
}
//GPU vddc
opt1 = ((oldpool)?get_pool_setting(oldpool->gpu_vddc, default_profile.gpu_vddc):"");
opt2 = get_pool_setting(newpool->gpu_vddc, default_profile.gpu_vddc);
if (strcasecmp(opt1, opt2) != 0 && !empty_string(opt2)) {
options |= SWITCHER_APPLY_GPU_VDDC;
}
#endif
// Remove soft reset if hard reset is set
if (opt_isset(options, SWITCHER_HARD_RESET) && opt_isset(options, SWITCHER_SOFT_RESET)) {
options &= ~SWITCHER_SOFT_RESET;
}
return options;
}
static void apply_switcher_options(unsigned long options, struct pool *pool)
{
int i;
const char *opt;
//nothing to change, abort
if (!options) {
return;
}
if(opt_isset(options, SWITCHER_APPLY_DEVICE))
{
//reset devices flags
opt_devs_enabled = 0;
for (i = 0; i < MAX_DEVICES; i++)
devices_enabled[i] = false;
//assign pool devices if any
if(!empty_string((opt = get_pool_setting(pool->devices, ((!empty_string(default_profile.devices))?default_profile.devices:"all"))))) {
set_devices((char *)opt);
}
}
//lookup gap
if(opt_isset(options, SWITCHER_APPLY_LG))
{
if(!empty_string((opt = get_pool_setting(pool->lookup_gap, default_profile.lookup_gap))))
set_lookup_gap((char *)opt);
}
//raw intensity from pool
if(opt_isset(options, SWITCHER_APPLY_RAWINT))
{
applog(LOG_DEBUG, "Switching to rawintensity: pool = %s, default = %s", pool->rawintensity, default_profile.rawintensity);
opt = get_pool_setting(pool->rawintensity, default_profile.rawintensity);
applog(LOG_DEBUG, "rawintensity -> %s", opt);
set_rawintensity(opt);
}
//xintensity
else if(opt_isset(options, SWITCHER_APPLY_XINT))
{
applog(LOG_DEBUG, "Switching to xintensity: pool = %s, default = %s", pool->xintensity, default_profile.xintensity);
opt = get_pool_setting(pool->xintensity, default_profile.xintensity);
applog(LOG_DEBUG, "xintensity -> %s", opt);
set_xintensity(opt);
}
//intensity
else if(opt_isset(options, SWITCHER_APPLY_INT))
{
applog(LOG_DEBUG, "Switching to intensity: pool = %s, default = %s", pool->intensity, default_profile.intensity);
opt = get_pool_setting(pool->intensity, default_profile.intensity);
applog(LOG_DEBUG, "intensity -> %s", opt);
set_intensity(opt);
}
//default basic intensity
else if(opt_isset(options, SWITCHER_APPLY_INT8))
{
default_profile.intensity = strdup("8");
set_intensity(default_profile.intensity);
}
//shaders
if(opt_isset(options, SWITCHER_APPLY_SHADER))
{
if(!empty_string((opt = get_pool_setting(pool->shaders, default_profile.shaders))))
set_shaders((char *)opt);
}
//thread-concurrency
if(opt_isset(options, SWITCHER_APPLY_TC))
{
// neoscrypt - if not specified set TC to 0 so that TC will be calculated by intensity settings
if (pool->algorithm.type == ALGO_NEOSCRYPT) {
opt = ((empty_string(pool->thread_concurrency))?"0":get_pool_setting(pool->thread_concurrency, default_profile.thread_concurrency));
}
// otherwise use pool/profile setting or default to default profile setting
else {
opt = get_pool_setting(pool->thread_concurrency, default_profile.thread_concurrency);
}
if(!empty_string(opt)) {
set_thread_concurrency((char *)opt);
}
}
//worksize
if(opt_isset(options, SWITCHER_APPLY_WORKSIZE))
{
if(!empty_string((opt = get_pool_setting(pool->worksize, default_profile.worksize))))
set_worksize(opt);
}
#ifdef HAVE_ADL
//GPU clock
if(opt_isset(options, SWITCHER_APPLY_GPU_ENGINE))
{
if(!empty_string((opt = get_pool_setting(pool->gpu_engine, default_profile.gpu_engine))))
set_gpu_engine((char *)opt);
}
//GPU memory clock
if(opt_isset(options, SWITCHER_APPLY_GPU_MEMCLOCK))
{
if(!empty_string((opt = get_pool_setting(pool->gpu_memclock, default_profile.gpu_memclock))))
set_gpu_memclock((char *)opt);
}
//GPU fans
if(opt_isset(options, SWITCHER_APPLY_GPU_FAN))
{
if(!empty_string((opt = get_pool_setting(pool->gpu_fan, default_profile.gpu_fan))))
set_gpu_fan((char *)opt);
}
//GPU powertune
if(opt_isset(options, SWITCHER_APPLY_GPU_POWERTUNE))
{
if(!empty_string((opt = get_pool_setting(pool->gpu_powertune, default_profile.gpu_powertune))))
set_gpu_powertune((char *)opt);
}
//GPU vddc
if(opt_isset(options, SWITCHER_APPLY_GPU_VDDC))
{
if(!empty_string((opt = get_pool_setting(pool->gpu_vddc, default_profile.gpu_vddc))))
set_gpu_vddc((char *)opt);
}
//apply gpu settings
for (i = 0; i < nDevs; ++i) {
if(opt_isset(options, SWITCHER_APPLY_GPU_ENGINE))
set_engineclock(i, gpus[i].min_engine);
if(opt_isset(options, SWITCHER_APPLY_GPU_MEMCLOCK))
set_memoryclock(i, gpus[i].gpu_memclock);
if(opt_isset(options, SWITCHER_APPLY_GPU_FAN))
set_fanspeed(i, gpus[i].min_fan);
if(opt_isset(options, SWITCHER_APPLY_GPU_POWERTUNE))
set_powertune(i, gpus[i].gpu_powertune);
if(opt_isset(options, SWITCHER_APPLY_GPU_VDDC))
set_vddc(i, gpus[i].gpu_vddc);
}
#endif
}
static void mutex_unlock_cleanup_handler(void *mutex)
{
mutex_unlock((pthread_mutex_t *) mutex);
}
static void get_work_prepare_thread(struct thr_info *mythr, struct work *work)
{
int i;
applog(LOG_DEBUG, "[THR%d] get_work_prepare_thread", mythr->id);
//if switcher is disabled
if(opt_switchmode == SWITCH_OFF)
return;
pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, NULL);
mutex_lock(&algo_switch_lock);
if(algo_switch_n == 0)
{
//get pool options to apply on switch
pool_switch_options = 0;
//switcher mode - switch on algorithm change
if(opt_switchmode == SWITCH_ALGO)
{
//if algorithms are different
if(!cmp_algorithm(&work->pool->algorithm, &mythr->cgpu->algorithm))
pool_switch_options = compare_pool_settings(pools[mythr->pool_no], work->pool);
}
//switcher mode - switch on pool change
else if(opt_switchmode == SWITCH_POOL)
{
if(work->pool->pool_no != mythr->pool_no)
{
if((pool_switch_options = compare_pool_settings(pools[mythr->pool_no], work->pool)) == 0)
{
applog(LOG_DEBUG, "No settings change from pool %s...", isnull(get_pool_name(work->pool), ""));
rd_lock(&mining_thr_lock);
//apply new pool_no to all mining threads
for (i = 0; i < mining_threads; i++)
{
struct thr_info *thr = mining_thr[i];
thr->pool_no = work->pool->pool_no;
}
rd_unlock(&mining_thr_lock);
}
}
}
if(pool_switch_options == 0)
{
mutex_unlock(&algo_switch_lock);
pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL);
return;
}
}
mutex_lock(&algo_switch_wait_lock);
algo_switch_n++;
mutex_unlock(&algo_switch_wait_lock);
//get the number of active threads to know when to switch... if we only check total threads, we may wait for ever on a disabled GPU
int active_threads = 0;
rd_lock(&mining_thr_lock);
for(i = 0; i < mining_threads; i++)
{
struct cgpu_info *cgpu = mining_thr[i]->cgpu;
//dont count dead/sick GPU threads or we may wait for ever also...
if(cgpu->deven != DEV_DISABLED && cgpu->status != LIFE_SICK && cgpu->status != LIFE_DEAD)
active_threads++;
}
rd_unlock(&mining_thr_lock);
// If all threads are waiting now
if(algo_switch_n >= active_threads)
{
const char *opt;
applog(LOG_DEBUG, "Applying pool settings for %s...", isnull(get_pool_name(work->pool), ""));
rd_lock(&mining_thr_lock);
// Shutdown all threads first (necessary)
if(opt_isset(pool_switch_options, SWITCHER_SOFT_RESET))
{
applog(LOG_DEBUG, "Soft Reset... Shutdown threads...");
for (i = 0; i < mining_threads; i++)
{
struct thr_info *thr = mining_thr[i];
thr->cgpu->drv->thread_shutdown(thr);
}
}
// Reset stats (e.g. for working_diff to be set properly in hash_sole_work)
zero_stats();
//apply switcher options
apply_switcher_options(pool_switch_options, work->pool);
//devices
/*if(opt_isset(pool_switch_options, SWITCHER_APPLY_DEVICE))
{
//reset devices flags
opt_devs_enabled = 0;
for (i = 0; i < MAX_DEVICES; i++)
devices_enabled[i] = false;
//assign pool devices if any
if(!empty_string((opt = get_pool_setting(work->pool->devices, ((!empty_string(default_profile.devices))?default_profile.devices:"all"))))) {
set_devices((char *)opt);
}
}
//lookup gap
if(opt_isset(pool_switch_options, SWITCHER_APPLY_LG))
{
if(!empty_string((opt = get_pool_setting(work->pool->lookup_gap, default_profile.lookup_gap))))
set_lookup_gap((char *)opt);
}
//raw intensity from pool
if(opt_isset(pool_switch_options, SWITCHER_APPLY_RAWINT))
{
applog(LOG_DEBUG, "Switching to rawintensity: pool = %s, default = %s", work->pool->rawintensity, default_profile.rawintensity);
opt = get_pool_setting(work->pool->rawintensity, default_profile.rawintensity);
applog(LOG_DEBUG, "rawintensity -> %s", opt);
set_rawintensity(opt);
}
//xintensity
else if(opt_isset(pool_switch_options, SWITCHER_APPLY_XINT))
{
applog(LOG_DEBUG, "Switching to xintensity: pool = %s, default = %s", work->pool->xintensity, default_profile.xintensity);
opt = get_pool_setting(work->pool->xintensity, default_profile.xintensity);
applog(LOG_DEBUG, "xintensity -> %s", opt);
set_xintensity(opt);
}
//intensity
else if(opt_isset(pool_switch_options, SWITCHER_APPLY_INT))
{
applog(LOG_DEBUG, "Switching to intensity: pool = %s, default = %s", work->pool->intensity, default_profile.intensity);
opt = get_pool_setting(work->pool->intensity, default_profile.intensity);
applog(LOG_DEBUG, "intensity -> %s", opt);
set_intensity(opt);
}
//default basic intensity
else if(opt_isset(pool_switch_options, SWITCHER_APPLY_INT8))
{
default_profile.intensity = strdup("8");
set_intensity(default_profile.intensity);
}
//shaders
if(opt_isset(pool_switch_options, SWITCHER_APPLY_SHADER))
{
if(!empty_string((opt = get_pool_setting(work->pool->shaders, default_profile.shaders))))
set_shaders((char *)opt);
}
//thread-concurrency
if(opt_isset(pool_switch_options, SWITCHER_APPLY_TC))
{
// neoscrypt - if not specified set TC to 0 so that TC will be calculated by intensity settings
if (work->pool->algorithm.type == ALGO_NEOSCRYPT) {
opt = ((empty_string(work->pool->thread_concurrency))?"0":get_pool_setting(work->pool->thread_concurrency, default_profile.thread_concurrency));
}
// otherwise use pool/profile setting or default to default profile setting
else {
opt = get_pool_setting(work->pool->thread_concurrency, default_profile.thread_concurrency);
}
if(!empty_string(opt)) {
set_thread_concurrency((char *)opt);
}
}
//worksize
if(opt_isset(pool_switch_options, SWITCHER_APPLY_WORKSIZE))
{
if(!empty_string((opt = get_pool_setting(work->pool->worksize, default_profile.worksize))))
set_worksize(opt);
}
#ifdef HAVE_ADL
//GPU clock
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_ENGINE))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_engine, default_profile.gpu_engine))))
set_gpu_engine((char *)opt);
}
//GPU memory clock
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_MEMCLOCK))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_memclock, default_profile.gpu_memclock))))
set_gpu_memclock((char *)opt);
}
//GPU fans
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_FAN))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_fan, default_profile.gpu_fan))))
set_gpu_fan((char *)opt);
}
//GPU powertune
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_POWERTUNE))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_powertune, default_profile.gpu_powertune))))
set_gpu_powertune((char *)opt);
}
//GPU vddc
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_VDDC))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_vddc, default_profile.gpu_vddc))))
set_gpu_vddc((char *)opt);
}
//apply gpu settings
for (i = 0; i < nDevs; i++)
{
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_ENGINE))
set_engineclock(i, gpus[i].min_engine);
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_MEMCLOCK))
set_memoryclock(i, gpus[i].gpu_memclock);
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_FAN))
set_fanspeed(i, gpus[i].min_fan);
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_POWERTUNE))
set_powertune(i, gpus[i].gpu_powertune);
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GPU_VDDC))
set_vddc(i, gpus[i].gpu_vddc);
}
#endif
*/
// Change algorithm for each thread (thread_prepare calls initCl)
if(opt_isset(pool_switch_options, SWITCHER_SOFT_RESET))
applog(LOG_DEBUG, "Soft Reset... Restarting threads...");
struct thr_info *thr;
for (i = 0; i < mining_threads; i++)
{
thr = mining_thr[i];
thr->pool_no = work->pool->pool_no; //set thread on new pool
//apply new algorithm if set
if(opt_isset(pool_switch_options, SWITCHER_APPLY_ALGO))
thr->cgpu->algorithm = work->pool->algorithm;
if(opt_isset(pool_switch_options, SWITCHER_SOFT_RESET))
{
thr->cgpu->drv->thread_prepare(thr);
thr->cgpu->drv->thread_init(thr);
}
// Necessary because algorithms can have dramatically different diffs
thr->cgpu->drv->working_diff = 1;
}
rd_unlock(&mining_thr_lock);
mutex_unlock(&algo_switch_lock);
// Hard restart if needed
if(opt_isset(pool_switch_options, SWITCHER_HARD_RESET))
{
applog(LOG_DEBUG, "Hard Reset Mining Threads...");
//if devices changed... enable/disable as needed
if(opt_isset(pool_switch_options, SWITCHER_APPLY_DEVICE))
enable_devices();
//figure out how many mining threads we'll need
unsigned int n_threads = 0;
pthread_t restart_thr;
#ifdef HAVE_ADL
//change gpu threads if needed
if(opt_isset(pool_switch_options, SWITCHER_APPLY_GT))
{
if(!empty_string((opt = get_pool_setting(work->pool->gpu_threads, default_profile.gpu_threads))))
set_gpu_threads(opt);
}
rd_lock(&devices_lock);
for (i = 0; i < total_devices; i++)
if (!opt_removedisabled || !opt_devs_enabled || devices_enabled[i])
n_threads += devices[i]->threads;
rd_unlock(&devices_lock);
#else
n_threads = mining_threads;
#endif
if (unlikely(pthread_create(&restart_thr, NULL, restart_mining_threads_thread, (void *) (intptr_t) n_threads)))
quit(1, "restart_mining_threads create thread failed");
applog(LOG_DEBUG, "Hard reset: Exiting mining thread %d", mythr->id);
pthread_exit(NULL);
}
else
{
// Signal other threads to start working now
mutex_lock(&algo_switch_wait_lock);
algo_switch_n = 0;
pthread_cond_broadcast(&algo_switch_wait_cond);
mutex_unlock(&algo_switch_wait_lock);
pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL);
// no need to wait, exit
return;
}
}
else {
bool hard_reset = opt_isset(pool_switch_options, SWITCHER_HARD_RESET);
mutex_unlock(&algo_switch_lock);
if (hard_reset) {
applog(LOG_DEBUG, "Hard reset: Exiting mining thread %d", mythr->id);
pthread_exit(NULL);
}
}
pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL);
// Set cleanup instructions in the event that the thread is cancelled
pthread_cleanup_push(mutex_unlock_cleanup_handler, (void *)&algo_switch_wait_lock);
// Wait for signal to start working again
mutex_lock(&algo_switch_wait_lock);
while(algo_switch_n > 0)
pthread_cond_wait(&algo_switch_wait_cond, &algo_switch_wait_lock);
// Non-zero argument will execute the cleanup handler after popping it
pthread_cleanup_pop(1);
}
struct work *get_work(struct thr_info *thr, const int thr_id)
{
struct work *work = NULL;
time_t diff_t;
thread_reportout(thr);
applog(LOG_DEBUG, "[THR%d] Popping work from get queue to get work", thr_id);
diff_t = time(NULL);
while (!work) {
work = hash_pop(true);
if (stale_work(work, false)) {
applog(LOG_DEBUG, "[THR%d] Work is stale, discarding", thr_id);
discard_work(work);
work = NULL;
wake_gws();
}
}
applog(LOG_DEBUG, "[THR%d] preparing thread...", thr_id);
get_work_prepare_thread(thr, work);
diff_t = time(NULL) - diff_t;
/* Since this is a blocking function, we need to add grace time to
* the device's last valid work to not make outages appear to be
* device failures. */
if (diff_t > 0) {
applog(LOG_DEBUG, "[THR%d] Get work blocked for %d seconds", thr_id, (int)diff_t);
thr->cgpu->last_device_valid_work += diff_t;
}
applog(LOG_DEBUG, "[THR%d] Got work from get queue", thr_id);
work->thr_id = thr_id;
thread_reportin(thr);
work->mined = true;
work->device_diff = MIN(thr->cgpu->drv->max_diff, work->work_difficulty);
return work;
}
/* Submit a copy of the tested, statistic recorded work item asynchronously */
static void submit_work_async(struct work *work)
{
struct pool *pool = work->pool;
pthread_t submit_thread;
cgtime(&work->tv_work_found);
if (stale_work(work, true)) {
if (opt_submit_stale)
applog(LOG_NOTICE, "%s stale share detected, submitting (user)", get_pool_name(pool));
else if (pool->submit_old)
applog(LOG_NOTICE, "%s stale share detected, submitting (pool)", get_pool_name(pool));
else {
applog(LOG_NOTICE, "%s stale share detected, discarding", get_pool_name(pool));
sharelog("discard", work);
mutex_lock(&stats_lock);
total_stale++;
pool->stale_shares++;
total_diff_stale += work->work_difficulty;
pool->diff_stale += work->work_difficulty;
mutex_unlock(&stats_lock);
free_work(work);
return;
}
work->stale = true;
}
if (work->stratum) {
applog(LOG_DEBUG, "Pushing %s work to stratum queue", get_pool_name(pool));
if (unlikely(!tq_push(pool->stratum_q, work))) {
applog(LOG_DEBUG, "Discarding work from removed pool");
free_work(work);
}
} else {
applog(LOG_DEBUG, "Pushing submit work to work thread");
if (unlikely(pthread_create(&submit_thread, NULL, submit_work_thread, (void *)work)))
quit(1, "Failed to create submit_work_thread");
}
}
void inc_hw_errors(struct thr_info *thr)
{
applog(LOG_INFO, "[THR%d] %s%d: invalid nonce - HW error", thr->id, thr->cgpu->drv->name,
thr->cgpu->device_id);
mutex_lock(&stats_lock);
hw_errors++;
thr->cgpu->hw_errors++;
mutex_unlock(&stats_lock);
thr->cgpu->drv->hw_error(thr);
}
/* Fills in the work nonce and builds the output data in work->hash */
static void rebuild_nonce(struct work *work, uint32_t nonce)
{
uint32_t nonce_pos = 76;
if (work->pool->algorithm.type == ALGO_CRE) nonce_pos = 140;
else if (work->pool->algorithm.type == ALGO_DECRED) nonce_pos = 140;
uint32_t *work_nonce = (uint32_t *)(work->data + nonce_pos);
*work_nonce = htole32(nonce);
work->pool->algorithm.regenhash(work);
}
/* For testing a nonce against diff 1 */
bool test_nonce(struct work *work, uint32_t nonce)
{
uint32_t *hash_32 = (uint32_t *)(work->hash + 28);
uint32_t diff1targ;
rebuild_nonce(work, nonce);
// for Neoscrypt, the diff1targ value is in work->target
if (work->pool->algorithm.type == ALGO_NEOSCRYPT || work->pool->algorithm.type == ALGO_PLUCK
|| work->pool->algorithm.type == ALGO_YESCRYPT || work->pool->algorithm.type == ALGO_YESCRYPT_MULTI) {
diff1targ = ((uint32_t *)work->target)[7];
}
else {
diff1targ = work->pool->algorithm.diff1targ;
}
return (le32toh(*hash_32) <= diff1targ);
}
static void update_work_stats(struct thr_info *thr, struct work *work)
{
double test_diff = current_diff;
test_diff *= work->pool->algorithm.share_diff_multiplier;
work->share_diff = share_diff(work);
test_diff *= work->pool->algorithm.share_diff_multiplier;
if (unlikely(work->share_diff >= test_diff)) {
work->block = true;
work->pool->solved++;
found_blocks++;
work->mandatory = true;
applog(LOG_NOTICE, "Found block for %s!", get_pool_name(work->pool));
}
mutex_lock(&stats_lock);
total_diff1 += work->device_diff;
thr->cgpu->diff1 += work->device_diff;
work->pool->diff1 += work->device_diff;
thr->cgpu->last_device_valid_work = time(NULL);
mutex_unlock(&stats_lock);
}
/* To be used once the work has been tested to be meet diff1 and has had its
* nonce adjusted. Returns true if the work target is met. */
bool submit_tested_work(struct thr_info *thr, struct work *work)
{
struct work *work_out;
update_work_stats(thr, work);
if (!fulltest(work->hash, work->target)) {
applog(LOG_INFO, "%s %d: Share above target", thr->cgpu->drv->name,
thr->cgpu->device_id);
return false;
}
work_out = copy_work(work);
submit_work_async(work_out);
return true;
}
/* Returns true if nonce for work was a valid share */
bool submit_nonce(struct thr_info *thr, struct work *work, uint32_t nonce)
{
if (test_nonce(work, nonce)) {
submit_tested_work(thr, work);
return true;
}
inc_hw_errors(thr);
return false;
}
static inline bool abandon_work(struct work *work, struct timeval *wdiff, uint64_t hashes)
{
if (wdiff->tv_sec > opt_scantime ||
work->blk.nonce >= MAXTHREADS - hashes ||
hashes >= 0xfffffffe ||
stale_work(work, false))
return true;
return false;
}
static void mt_disable(struct thr_info *mythr, const int thr_id,
struct device_drv *drv)
{
applog(LOG_WARNING, "Thread %d being disabled", thr_id);
mythr->rolling = mythr->cgpu->rolling = 0;
applog(LOG_DEBUG, "Waiting on sem in miner thread");
mythr->paused = true;
cgsem_wait(&mythr->sem);
applog(LOG_WARNING, "Thread %d being re-enabled", thr_id);
mythr->paused = false;
drv->thread_enable(mythr);
}
/* The main hashing loop for devices that are slow enough to work on one work
* item at a time, without a queue, aborting work before the entire nonce
* range has been hashed if needed. */
static void hash_sole_work(struct thr_info *mythr)
{
const int thr_id = mythr->id;
struct cgpu_info *cgpu = mythr->cgpu;
struct device_drv *drv = cgpu->drv;
struct timeval getwork_start, tv_start, *tv_end, tv_workstart, tv_lastupdate;
struct sgminer_stats *dev_stats = &(cgpu->sgminer_stats);
struct sgminer_stats *pool_stats;
/* Try to cycle approximately 5 times before each log update */
const long cycle = opt_log_interval / 5 ? 5 : 1;
const bool primary = mythr->device_thread == 0;
struct timeval diff, sdiff, wdiff = {0, 0};
uint32_t max_nonce = drv->can_limit_work(mythr);
int64_t hashes_done = 0;
tv_end = &getwork_start;
cgtime(&getwork_start);
sdiff.tv_sec = sdiff.tv_usec = 0;
cgtime(&tv_lastupdate);
while (likely(!cgpu->shutdown)) {
struct work *work = get_work(mythr, thr_id);
int64_t hashes;
mythr->work_restart = false;
cgpu->new_work = true;
cgtime(&tv_workstart);
work->blk.nonce = 0;
cgpu->max_hashes = 0;
if (!drv->prepare_work(mythr, work)) {
applog(LOG_ERR, "work prepare failed, exiting "
"mining thread %d", thr_id);
break;
}
work->device_diff = MIN(drv->working_diff, work->work_difficulty);
/* Dynamically adjust the working diff even if the target
* diff is very high to ensure we can still validate scrypt is
* returning shares. */
double wu;
wu = total_diff1 / total_secs * 60;
if (wu > 30 && drv->working_diff < drv->max_diff &&
drv->working_diff < work->work_difficulty) {
drv->working_diff++;
applog(LOG_DEBUG, "Driver %s working diff changed to %.0f",
drv->dname, drv->working_diff);
work->device_diff = MIN(drv->working_diff, work->work_difficulty);
} else if (drv->working_diff > work->work_difficulty)
drv->working_diff = work->work_difficulty;
if (work->pool->algorithm.type == ALGO_NEOSCRYPT) {
set_target_neoscrypt(work->device_target, work->device_diff, work->thr_id);
} else {
set_target(work->device_target, work->device_diff, work->pool->algorithm.diff_multiplier2, work->thr_id);
}
do {
cgtime(&tv_start);
subtime(&tv_start, &getwork_start);
addtime(&getwork_start, &dev_stats->getwork_wait);
if (time_more(&getwork_start, &dev_stats->getwork_wait_max))
copy_time(&dev_stats->getwork_wait_max, &getwork_start);
if (time_less(&getwork_start, &dev_stats->getwork_wait_min))
copy_time(&dev_stats->getwork_wait_min, &getwork_start);
dev_stats->getwork_calls++;
pool_stats = &(work->pool->sgminer_stats);
addtime(&getwork_start, &pool_stats->getwork_wait);
if (time_more(&getwork_start, &pool_stats->getwork_wait_max))
copy_time(&pool_stats->getwork_wait_max, &getwork_start);
if (time_less(&getwork_start, &pool_stats->getwork_wait_min))
copy_time(&pool_stats->getwork_wait_min, &getwork_start);
pool_stats->getwork_calls++;
cgtime(&(work->tv_work_start));
/* Only allow the mining thread to be cancelled when
* it is not in the driver code. */
pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, NULL);
thread_reportin(mythr);
hashes = drv->scanhash(mythr, work, work->blk.nonce + max_nonce);
thread_reportout(mythr);
pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL);
pthread_testcancel();
/* tv_end is == &getwork_start */
cgtime(&getwork_start);
if (unlikely(hashes == -1)) {
applog(LOG_ERR, "%s %d failure, disabling!", drv->name, cgpu->device_id);
cgpu->deven = DEV_DISABLED;
dev_error(cgpu, REASON_THREAD_ZERO_HASH);
event_notify("idle");
cgpu->shutdown = true;
break;
}
hashes_done += hashes;
if (hashes > cgpu->max_hashes)
cgpu->max_hashes = hashes;
timersub(tv_end, &tv_start, &diff);
sdiff.tv_sec += diff.tv_sec;
sdiff.tv_usec += diff.tv_usec;
if (sdiff.tv_usec > 1000000) {
++sdiff.tv_sec;
sdiff.tv_usec -= 1000000;
}
timersub(tv_end, &tv_workstart, &wdiff);
if (unlikely((long)sdiff.tv_sec < cycle)) {
int mult;
if (likely(max_nonce == 0xffffffff))
continue;
mult = 1000000 / ((sdiff.tv_usec + 0x400) / 0x400) + 0x10;
mult *= cycle;
if (max_nonce > (0xffffffff * 0x400) / mult)
max_nonce = 0xffffffff;
else
max_nonce = (max_nonce * mult) / 0x400;
} else if (unlikely(sdiff.tv_sec > cycle))
max_nonce = max_nonce * cycle / sdiff.tv_sec;
else if (unlikely(sdiff.tv_usec > 100000))
max_nonce = max_nonce * 0x400 / (((cycle * 1000000) + sdiff.tv_usec) / (cycle * 1000000 / 0x400));
timersub(tv_end, &tv_lastupdate, &diff);
/* Update the hashmeter at most 5 times per second */
if ((hashes_done && (diff.tv_sec > 0 || diff.tv_usec > 200000)) ||
diff.tv_sec >= opt_log_interval) {
hashmeter(thr_id, &diff, hashes_done);
hashes_done = 0;
copy_time(&tv_lastupdate, tv_end);
}
if (unlikely(mythr->work_restart)) {
/* Apart from device_thread 0, we stagger the
* starting of every next thread to try and get
* all devices busy before worrying about
* getting work for their extra threads */
if (!primary) {
struct timespec rgtp;
rgtp.tv_sec = 0;
rgtp.tv_nsec = 250 * mythr->device_thread * 1000000;
nanosleep(&rgtp, NULL);
}
break;
}
if (unlikely(mythr->pause || cgpu->deven != DEV_ENABLED))
mt_disable(mythr, thr_id, drv);
sdiff.tv_sec = sdiff.tv_usec = 0;
} while (!abandon_work(work, &wdiff, cgpu->max_hashes));
free_work(work);
}
cgpu->deven = DEV_DISABLED;
}
void *miner_thread(void *userdata)
{
struct thr_info *mythr = (struct thr_info *)userdata;
const int thr_id = mythr->id;
struct cgpu_info *cgpu = mythr->cgpu;
struct device_drv *drv = cgpu->drv;
char threadname[16];
snprintf(threadname, sizeof(threadname), "%d/Miner", thr_id);
RenameThread(threadname);
thread_reportout(mythr);
if (!drv->thread_init(mythr)) {
dev_error(cgpu, REASON_THREAD_FAIL_INIT);
goto out;
}
applog(LOG_DEBUG, "Waiting on sem in miner thread");
cgsem_wait(&mythr->sem);
set_highprio();
drv->hash_work(mythr);
out:
drv->thread_shutdown(mythr);
return NULL;
}
enum {
STAT_SLEEP_INTERVAL = 1,
STAT_CTR_INTERVAL = 10000000,
FAILURE_INTERVAL = 30,
};
#ifdef HAVE_LIBCURL
/* Stage another work item from the work returned in a longpoll */
static void convert_to_work(json_t *val, int rolltime, struct pool *pool, struct timeval *tv_lp, struct timeval *tv_lp_reply)
{
struct work *work;
bool rc;
work = make_work();
rc = work_decode(pool, work, val);
if (unlikely(!rc)) {
applog(LOG_ERR, "Could not convert longpoll data to work");
free_work(work);
return;
}
total_getworks++;
pool->getwork_requested++;
work->pool = pool;
work->rolltime = rolltime;
copy_time(&work->tv_getwork, tv_lp);
copy_time(&work->tv_getwork_reply, tv_lp_reply);
calc_diff(work, 0);
if (pool->state == POOL_REJECTING)
work->mandatory = true;
if (pool->has_gbt)
gen_gbt_work(pool, work);
work->longpoll = true;
work->getwork_mode = GETWORK_MODE_LP;
/* We'll be checking this work item twice, but we already know it's
* from a new block so explicitly force the new block detection now
* rather than waiting for it to hit the stage thread. This also
* allows testwork to know whether LP discovered the block or not. */
test_work_current(work);
/* Don't use backup LPs as work if we have failover-only enabled. Use
* the longpoll work from a pool that has been rejecting shares as a
* way to detect when the pool has recovered.
*/
if (pool != current_pool() && opt_fail_only && pool->state != POOL_REJECTING) {
free_work(work);
return;
}
work = clone_work(work);
applog(LOG_DEBUG, "Pushing converted work to stage thread");
stage_work(work);
applog(LOG_DEBUG, "Converted longpoll data to work");
}
/* If we want longpoll, enable it for the chosen default pool, or, if
* the pool does not support longpoll, find the first one that does
* and use its longpoll support */
static struct pool *select_longpoll_pool(struct pool *cp)
{
int i;
if (cp->hdr_path || cp->has_gbt)
return cp;
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
if (pool->has_stratum || pool->hdr_path)
return pool;
}
return NULL;
}
#endif /* HAVE_LIBCURL */
/* This will make the longpoll thread wait till it's the current pool, or it
* has been flagged as rejecting, before attempting to open any connections.
*/
static void wait_lpcurrent(struct pool *pool)
{
while (!cnx_needed(pool) && (pool->state == POOL_DISABLED ||
(pool != current_pool() && pool_strategy != POOL_LOADBALANCE &&
pool_strategy != POOL_BALANCE))) {
mutex_lock(&lp_lock);
pthread_cond_wait(&lp_cond, &lp_lock);
mutex_unlock(&lp_lock);
}
}
#ifdef HAVE_LIBCURL
static void *longpoll_thread(void *userdata)
{
struct pool *cp = (struct pool *)userdata;
/* This *pool is the source of the actual longpoll, not the pool we've
* tied it to */
struct timeval start, reply, end;
struct pool *pool = NULL;
char threadname[16];
CURL *curl = NULL;
char curl_err_str[CURL_ERROR_SIZE];
int failures = 0;
char lpreq[1024];
char *lp_url;
int rolltime;
snprintf(threadname, sizeof(threadname), "%d/Longpoll", cp->pool_no);
RenameThread(threadname);
curl = curl_easy_init();
if (unlikely(!curl)) {
applog(LOG_ERR, "CURL initialisation failed");
return NULL;
}
retry_pool:
pool = select_longpoll_pool(cp);
if (!pool) {
applog(LOG_WARNING, "No suitable long-poll found for %s", cp->rpc_url);
while (!pool) {
cgsleep_ms(60000);
pool = select_longpoll_pool(cp);
}
}
if (pool->has_stratum) {
applog(LOG_WARNING, "Block change for %s detection via %s stratum",
cp->rpc_url, pool->rpc_url);
goto out;
}
/* Any longpoll from any pool is enough for this to be true */
have_longpoll = true;
wait_lpcurrent(cp);
if (pool->has_gbt) {
lp_url = pool->rpc_url;
applog(LOG_WARNING, "GBT longpoll ID activated for %s", lp_url);
} else {
strcpy(lpreq, getwork_req);
lp_url = pool->lp_url;
if (cp == pool)
applog(LOG_WARNING, "Long-polling activated for %s", lp_url);
else
applog(LOG_WARNING, "Long-polling activated for %s via %s", cp->rpc_url, lp_url);
}
while (42) {
json_t *val, *soval;
wait_lpcurrent(cp);
cgtime(&start);
/* Update the longpollid every time, but do it under lock to
* avoid races */
if (pool->has_gbt) {
cg_rlock(&pool->gbt_lock);
snprintf(lpreq, sizeof(lpreq),
"{\"id\": 0, \"method\": \"getblocktemplate\", \"params\": "
"[{\"capabilities\": [\"coinbasetxn\", \"workid\", \"coinbase/append\"], "
"\"longpollid\": \"%s\"}]}\n", pool->longpollid);
cg_runlock(&pool->gbt_lock);
}
/* Longpoll connections can be persistent for a very long time
* and any number of issues could have come up in the meantime
* so always establish a fresh connection instead of relying on
* a persistent one. */
curl_easy_setopt(curl, CURLOPT_FRESH_CONNECT, 1);
val = json_rpc_call(curl, curl_err_str, lp_url, pool->rpc_userpass,
lpreq, false, true, &rolltime, pool, false);
cgtime(&reply);
if (likely(val)) {
soval = json_object_get(json_object_get(val, "result"), "submitold");
if (soval)
pool->submit_old = json_is_true(soval);
else
pool->submit_old = false;
convert_to_work(val, rolltime, pool, &start, &reply);
failures = 0;
json_decref(val);
} else {
/* Some pools regularly drop the longpoll request so
* only see this as longpoll failure if it happens
* immediately and just restart it the rest of the
* time. */
cgtime(&end);
if (end.tv_sec - start.tv_sec > 30)
continue;
if (failures == 1)
applog(LOG_WARNING, "longpoll failed for %s, retrying every 30s", lp_url);
cgsleep_ms(30000);
}
if (pool != cp) {
pool = select_longpoll_pool(cp);
if (pool->has_stratum) {
applog(LOG_WARNING, "Block change for %s detection via %s stratum",
cp->rpc_url, pool->rpc_url);
break;
}
if (unlikely(!pool))
goto retry_pool;
}
if (unlikely(pool->removed))
break;
}
out:
curl_easy_cleanup(curl);
return NULL;
}
#else /* HAVE_LIBCURL */
static void *longpoll_thread(void __maybe_unused *userdata)
{
pthread_detach(pthread_self());
return NULL;
}
#endif /* HAVE_LIBCURL */
void reinit_device(struct cgpu_info *cgpu)
{
mutex_lock(&algo_switch_lock);
cgpu->drv->reinit_device(cgpu);
mutex_unlock(&algo_switch_lock);
}
static struct timeval rotate_tv;
/* We reap curls if they are unused for over a minute */
static void reap_curl(struct pool *pool)
{
struct curl_ent *ent, *iter;
struct timeval now;
int reaped = 0;
cgtime(&now);
mutex_lock(&pool->pool_lock);
list_for_each_entry_safe(ent, iter, &pool->curlring, node) {
if (pool->curls < 2)
break;
if (now.tv_sec - ent->tv.tv_sec > 300) {
reaped++;
pool->curls--;
list_del(&ent->node);
curl_easy_cleanup(ent->curl);
free(ent);
}
}
mutex_unlock(&pool->pool_lock);
if (reaped)
applog(LOG_DEBUG, "Reaped %d curl%s from %s", reaped, reaped > 1 ? "s" : "", get_pool_name(pool));
}
static void *watchpool_thread(void __maybe_unused *userdata)
{
int intervals = 0;
pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
RenameThread("Watchpool");
set_lowprio();
while (42) {
struct timeval now;
int sleeptimeout/* = 1000*/;
int i;
// get current time
cgtime(&now);
// sleep timeout is 30 secs for most cases
sleeptimeout = opt_watchpool_refresh * 1000;
//limit to 5 secs minimum...
if (sleeptimeout < 5000) {
sleeptimeout = 5000;
}
// check the status of each pool
for (i = 0; i < total_pools; ++i) {
struct pool *pool = pools[i];
reap_curl(pool);
/* Get a rolling utility per pool over 10 mins */
if (intervals >= 600) {
int shares = pool->diff1 - pool->last_shares;
pool->last_shares = pool->diff1;
pool->utility = (pool->utility + (double)shares * 0.63) / 1.63;
pool->shares = pool->utility;
intervals = 0;
}
// if this pool is disabled, skip it
if (pool->state == POOL_DISABLED) {
continue;
}
/* Don't start testing any pools if the test threads
* from startup are still doing their first attempt. */
if (unlikely(pool->testing)) {
pthread_join(pool->test_thread, NULL);
pool->testing = false;
}
/* Test pool is idle once every minute */
if (pool->idle && now.tv_sec - pool->tv_idle.tv_sec > 30) {
cgtime(&pool->tv_idle);
if (pool_active(pool, true) && pool_tclear(pool, &pool->idle)) {
pool_resus(pool);
}
}
// if this pool is alive and the priority is greater (lower) than currently connected pool
if (!pool->idle && pool->prio < cp_prio()) {
// failover strategy - switch when failover delay is met
if (pool_strategy == POOL_FAILOVER && (now.tv_sec - pool->tv_idle.tv_sec > opt_fail_switch_delay)) {
applog(LOG_WARNING, "%s stable for %d seconds", get_pool_name(pool), opt_fail_switch_delay);
switch_pools(NULL);
}
}
} //end pool loop
// if the pool stategy is rotation and we have been over the rotate delay, switch pool
if (pool_strategy == POOL_ROTATE && now.tv_sec - rotate_tv.tv_sec > 60 * opt_rotate_period) {
cgtime(&rotate_tv);
switch_pools(NULL);
}
// if the current pool is dead/idle switch pool
if (current_pool()->idle) {
switch_pools(NULL);
}
cgsleep_ms(sleeptimeout);
intervals += (sleeptimeout / 1000);
} //end main loop
return NULL;
}
/* Makes sure the hashmeter keeps going even if mining threads stall, updates
* the screen at regular intervals, and restarts threads if they appear to have
* died. */
#define WATCHDOG_INTERVAL 2
#define WATCHDOG_SICK_TIME 120
#define WATCHDOG_DEAD_TIME 600
#define WATCHDOG_SICK_COUNT (WATCHDOG_SICK_TIME/WATCHDOG_INTERVAL)
#define WATCHDOG_DEAD_COUNT (WATCHDOG_DEAD_TIME/WATCHDOG_INTERVAL)
static void *watchdog_thread(void __maybe_unused *userdata)
{
const unsigned int interval = WATCHDOG_INTERVAL;
struct timeval zero_tv;
pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
RenameThread("Watchdog");
set_lowprio();
memset(&zero_tv, 0, sizeof(struct timeval));
cgtime(&rotate_tv);
while (1) {
int i;
struct timeval now;
sleep(interval);
discard_stale();
hashmeter(-1, &zero_tv, 0);
rd_lock(&mining_thr_lock);
#ifdef HAVE_CURSES
if (curses_active_locked()) {
struct cgpu_info *cgpu;
int count;
change_logwinsize();
curses_print_status();
if (!opt_compact) {
count = 0;
for (i = 0; i < total_devices; i++) {
cgpu = get_devices(i);
if (cgpu && (!opt_removedisabled || cgpu->deven != DEV_DISABLED || devices_enabled[i]))
curses_print_devstatus(cgpu, count++);
}
}
touchwin(statuswin);
wrefresh(statuswin);
touchwin(logwin);
wrefresh(logwin);
unlock_curses();
}
#endif
cgtime(&now);
// check last getwork time if greater than 10 mins, declare idle...
if ((time(NULL) - last_getwork) >= 600) {
event_notify("idle");
}
if (!sched_paused && !should_run()) {
applog(LOG_WARNING, "Pausing execution as per stop time %02d:%02d scheduled",
schedstop.tm.tm_hour, schedstop.tm.tm_min);
if (!schedstart.enable) {
quit(0, "Terminating execution as planned");
break;
}
applog(LOG_WARNING, "Will restart execution as scheduled at %02d:%02d",
schedstart.tm.tm_hour, schedstart.tm.tm_min);
sched_paused = true;
for (i = 0; i < mining_threads; i++)
mining_thr[i]->pause = true;
} else if (sched_paused && should_run()) {
applog(LOG_WARNING, "Restarting execution as per start time %02d:%02d scheduled",
schedstart.tm.tm_hour, schedstart.tm.tm_min);
if (schedstop.enable)
applog(LOG_WARNING, "Will pause execution as scheduled at %02d:%02d",
schedstop.tm.tm_hour, schedstop.tm.tm_min);
sched_paused = false;
for (i = 0; i < mining_threads; i++) {
struct thr_info *thr;
thr = mining_thr[i];
/* Don't touch disabled devices */
if (thr->cgpu->deven == DEV_DISABLED)
continue;
thr->pause = false;
applog(LOG_DEBUG, "Pushing sem post to thread %d", thr->id);
cgsem_post(&thr->sem);
}
}
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = get_devices(i);
struct thr_info *thr = cgpu->thr[0];
enum dev_enable *denable;
char dev_str[8];
int gpu;
cgpu->drv->get_stats(cgpu);
gpu = cgpu->device_id;
denable = &cgpu->deven;
snprintf(dev_str, sizeof(dev_str), "%s%d", cgpu->drv->name, gpu);
#ifdef HAVE_ADL
if (adl_active && cgpu->has_adl)
gpu_autotune(gpu, denable);
if (opt_debug && cgpu->has_adl) {
int engineclock = 0, memclock = 0, activity = 0, fanspeed = 0, fanpercent = 0, powertune = 0;
float temp = 0, vddc = 0;
if (gpu_stats(gpu, &temp, &engineclock, &memclock, &vddc, &activity, &fanspeed, &fanpercent, &powertune))
applog(LOG_DEBUG, "%.1f C F: %d%%(%dRPM) E: %dMHz M: %dMhz V: %.3fV A: %d%% P: %d%%",
temp, fanpercent, fanspeed, engineclock, memclock, vddc, activity, powertune);
}
#endif
/* Thread is disabled or waiting on getwork */
if (*denable == DEV_DISABLED || thr->getwork)
continue;
if (cgpu->status != LIFE_WELL && (now.tv_sec - thr->last.tv_sec < WATCHDOG_SICK_TIME)) {
if (cgpu->status != LIFE_INIT)
applog(LOG_ERR, "%s: Recovered, declaring WELL!", dev_str);
cgpu->status = LIFE_WELL;
cgpu->device_last_well = time(NULL);
} else if (cgpu->status == LIFE_WELL && (now.tv_sec - thr->last.tv_sec > WATCHDOG_SICK_TIME)) {
thr->rolling = cgpu->rolling = 0;
cgpu->status = LIFE_SICK;
applog(LOG_ERR, "%s: Idle for more than 2 minutes, declaring SICK!", dev_str);
cgtime(&thr->sick);
dev_error(cgpu, REASON_DEV_SICK_IDLE_60);
event_notify("gpu_sick");
#ifdef HAVE_ADL
if (adl_active && cgpu->has_adl && gpu_activity(gpu) > 50) {
applog(LOG_ERR, "GPU still showing activity suggesting a hard hang.");
applog(LOG_ERR, "Will not attempt to auto-restart it.");
} else
#endif
if (opt_restart) {
applog(LOG_ERR, "%s: Attempting to restart", dev_str);
reinit_device(cgpu);
}
} else if (cgpu->status == LIFE_SICK && (now.tv_sec - thr->last.tv_sec > WATCHDOG_DEAD_TIME)) {
cgpu->status = LIFE_DEAD;
applog(LOG_ERR, "%s: Not responded for more than 10 minutes, declaring DEAD!", dev_str);
cgtime(&thr->sick);
dev_error(cgpu, REASON_DEV_DEAD_IDLE_600);
event_notify("gpu_dead");
} else if (now.tv_sec - thr->sick.tv_sec > 60 &&
(cgpu->status == LIFE_SICK || cgpu->status == LIFE_DEAD)) {
/* Attempt to restart a GPU that's sick or dead once every minute */
cgtime(&thr->sick);
#ifdef HAVE_ADL
if (adl_active && cgpu->has_adl && gpu_activity(gpu) > 50) {
/* Again do not attempt to restart a device that may have hard hung */
} else
#endif
if (opt_restart)
reinit_device(cgpu);
}
}
rd_unlock(&mining_thr_lock);
}
return NULL;
}
static void log_print_status(struct cgpu_info *cgpu)
{
char logline[255];
get_statline(logline, sizeof(logline), cgpu);
applog(LOG_WARNING, "%s", logline);
}
static void noop_get_statline(char __maybe_unused *buf, size_t __maybe_unused bufsiz, struct cgpu_info __maybe_unused *cgpu);
void blank_get_statline_before(char *buf, size_t bufsiz, struct cgpu_info __maybe_unused *cgpu);
void print_summary(void)
{
struct timeval diff;
int hours, mins, secs, i;
double utility, displayed_hashes, work_util;
bool mhash_base = true;
timersub(&total_tv_end, &total_tv_start, &diff);
hours = diff.tv_sec / 3600;
mins = (diff.tv_sec % 3600) / 60;
secs = diff.tv_sec % 60;
utility = total_accepted / total_secs * 60;
work_util = total_diff1 / total_secs * 60;
applog(LOG_WARNING, "\nSummary of runtime statistics:\n");
applog(LOG_WARNING, "Started at %s", datestamp);
if (total_pools == 1)
applog(LOG_WARNING, "Pool: %s", pools[0]->rpc_url);
applog(LOG_WARNING, "Runtime: %d hrs : %d mins : %d secs", hours, mins, secs);
displayed_hashes = total_mhashes_done / total_secs;
if (displayed_hashes < 1) {
displayed_hashes *= 1000;
mhash_base = false;
}
applog(LOG_WARNING, "Average hashrate: %.1f %shash/s", displayed_hashes, mhash_base? "Mega" : "Kilo");
applog(LOG_WARNING, "Solved blocks: %d", found_blocks);
applog(LOG_WARNING, "Best share difficulty: %s", best_share);
applog(LOG_WARNING, "Share submissions: %d", total_accepted + total_rejected);
applog(LOG_WARNING, "Accepted shares: %d", total_accepted);
applog(LOG_WARNING, "Rejected shares: %d", total_rejected);
applog(LOG_WARNING, "Accepted difficulty shares: %1.f", total_diff_accepted);
applog(LOG_WARNING, "Rejected difficulty shares: %1.f", total_diff_rejected);
if (total_accepted || total_rejected)
applog(LOG_WARNING, "Reject ratio: %.1f%%", (double)(total_rejected * 100) / (double)(total_accepted + total_rejected));
applog(LOG_WARNING, "Hardware errors: %d", hw_errors);
applog(LOG_WARNING, "Utility (accepted shares / min): %.2f/min", utility);
applog(LOG_WARNING, "Work Utility (diff1 shares solved / min): %.2f/min\n", work_util);
applog(LOG_WARNING, "Stale submissions discarded due to new blocks: %d", total_stale);
applog(LOG_WARNING, "Unable to get work from server occasions: %d", total_go);
applog(LOG_WARNING, "Work items generated locally: %d", local_work);
applog(LOG_WARNING, "Submitting work remotely delay occasions: %d", total_ro);
applog(LOG_WARNING, "New blocks detected on network: %d\n", new_blocks);
if (total_pools > 1) {
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
applog(LOG_WARNING, "Pool: %s", pool->rpc_url);
if (pool->solved)
applog(LOG_WARNING, "SOLVED %d BLOCK%s!", pool->solved, pool->solved > 1 ? "S" : "");
applog(LOG_WARNING, " Share submissions: %d", pool->accepted + pool->rejected);
applog(LOG_WARNING, " Accepted shares: %d", pool->accepted);
applog(LOG_WARNING, " Rejected shares: %d", pool->rejected);
applog(LOG_WARNING, " Accepted difficulty shares: %1.f", pool->diff_accepted);
applog(LOG_WARNING, " Rejected difficulty shares: %1.f", pool->diff_rejected);
if (pool->accepted || pool->rejected)
applog(LOG_WARNING, " Reject ratio: %.1f%%", (double)(pool->rejected * 100) / (double)(pool->accepted + pool->rejected));
applog(LOG_WARNING, " Items worked on: %d", pool->works);
applog(LOG_WARNING, " Stale submissions discarded due to new blocks: %d", pool->stale_shares);
applog(LOG_WARNING, " Unable to get work from server occasions: %d", pool->getfail_occasions);
applog(LOG_WARNING, " Submitting work remotely delay occasions: %d\n", pool->remotefail_occasions);
}
}
applog(LOG_WARNING, "Summary of per device statistics:\n");
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = get_devices(i);
cgpu->drv->get_statline_before = &blank_get_statline_before;
cgpu->drv->get_statline = &noop_get_statline;
log_print_status(cgpu);
}
if (opt_shares) {
applog(LOG_WARNING, "Mined %.0f accepted shares of %d requested\n", total_diff_accepted, opt_shares);
if (opt_shares > total_diff_accepted)
applog(LOG_WARNING, "WARNING - Mined only %.0f shares of %d requested.", total_diff_accepted, opt_shares);
}
applog(LOG_WARNING, " ");
fflush(stderr);
fflush(stdout);
}
static void clean_up(bool restarting)
{
#ifdef HAVE_ADL
clear_adl(nDevs);
#endif
cgtime(&total_tv_end);
#ifdef WIN32
timeEndPeriod(1);
#endif
#ifdef HAVE_CURSES
disable_curses();
#endif
if (!restarting && !opt_realquiet && successful_connect)
print_summary();
curl_global_cleanup();
}
void _quit(int status)
{
clean_up(false);
#if defined(unix) || defined(__APPLE__)
if (forkpid > 0) {
kill(forkpid, SIGTERM);
forkpid = 0;
}
#endif
exit(status);
}
#ifdef HAVE_CURSES
char *curses_input(const char *query)
{
char *input;
echo();
input = (char *)malloc(255);
if (!input)
quit(1, "Failed to malloc input");
leaveok(logwin, false);
wlogprint("%s:\n", query);
wgetnstr(logwin, input, 255);
if (!strlen(input))
strcpy(input, "-1");
leaveok(logwin, true);
noecho();
return input;
}
#endif
static bool pools_active = false;
static void *test_pool_thread(void *arg)
{
struct pool *pool = (struct pool *)arg;
if (pool_active(pool, false)) {
pool_tset(pool, &pool->lagging);
pool_tclear(pool, &pool->idle);
bool first_pool = false;
cg_wlock(&control_lock);
if (!pools_active) {
currentpool = pool;
if (pool->pool_no != 0)
first_pool = true;
pools_active = true;
}
cg_wunlock(&control_lock);
if (unlikely(first_pool))
applog(LOG_NOTICE, "Switching to %s - first alive pool", get_pool_name(pool));
pool_resus(pool);
switch_pools(NULL);
} else {
pool_died(pool);
}
return NULL;
}
/* Always returns true that the pool details were added unless we are not
* live, implying this is the only pool being added, so if no pools are
* active it returns false. */
bool add_pool_details(struct pool *pool, bool live, char *url, char *user, char *pass, char *name, char *desc, char *profile, char *algo)
{
size_t siz;
url = get_proxy(url, pool);
pool->rpc_url = url;
pool->rpc_user = user;
pool->rpc_pass = pass;
pool->name = name;
pool->description = desc;
pool->profile = profile;
//if a profile was supplied, apply pool properties from profile
if(!empty_string(profile))
apply_pool_profile(pool); //remove profile if was invalid
//if profile is empty, assign algorithm or default algorithm
if(empty_string(pool->profile))
{
if(!empty_string(algo))
set_algorithm(&pool->algorithm, algo);
else
set_algorithm(&pool->algorithm, default_profile.algorithm.name);
}
siz = strlen(pool->rpc_user) + strlen(pool->rpc_pass) + 2;
pool->rpc_userpass = (char *)malloc(siz);
if (!pool->rpc_userpass)
quit(1, "Failed to malloc userpass");
snprintf(pool->rpc_userpass, siz, "%s:%s", pool->rpc_user, pool->rpc_pass);
pool->testing = true;
pool->idle = true;
enable_pool(pool);
pthread_create(&pool->test_thread, NULL, test_pool_thread, (void *)pool);
if (!live) {
pthread_join(pool->test_thread, NULL);
pool->testing = false;
return pools_active;
}
return true;
}
#ifdef HAVE_CURSES
static bool input_pool(bool live)
{
char *url = NULL, *user = NULL, *pass = NULL;
char *name = NULL, *desc = NULL, *profile = NULL, *algo = NULL;
struct pool *pool;
bool ret = false;
immedok(logwin, true);
wlogprint("Input server details.\n");
/* Get user input */
url = curses_input("URL");
if (!url) goto out;
user = curses_input("User name");
if (!user) goto out;
pass = curses_input("Password");
if (!pass) goto out;
name = curses_input("Pool name (optional)");
if (strcmp(name, "-1") == 0) strcpy(name, "");
desc = curses_input("Description (optional)");
if (strcmp(desc, "-1") == 0) strcpy(desc, "");
profile = curses_input("Profile (optional)");
if (strcmp(profile, "-1") == 0) profile[0] = '\0';
algo = curses_input("Algorithm (optional)");
if (strcmp(algo, "-1") == 0) algo[0] = '\0';
pool = add_pool();
if (!detect_stratum(pool, url) && strncmp(url, "http://", 7) &&
strncmp(url, "https://", 8)) {
char *httpinput;
httpinput = (char *)malloc(256);
if (!httpinput)
quit(1, "Failed to malloc httpinput");
strcpy(httpinput, "http://");
strncat(httpinput, url, 248);
free(url);
url = httpinput;
}
ret = add_pool_details(pool, live, url, user, pass,
name, desc, profile, algo);
out:
immedok(logwin, false);
if (!ret) {
if (url)
free(url);
if (user)
free(user);
if (pass)
free(pass);
if (name)
free(name);
if (desc)
free(desc);
if (algo)
free(algo);
}
return ret;
}
#endif
#if defined(unix) || defined(__APPLE__)
static void fork_monitor()
{
// Make a pipe: [readFD, writeFD]
int pfd[2];
int r = pipe(pfd);
if (r < 0) {
perror("pipe - failed to create pipe for --monitor");
exit(1);
}
// Make stderr write end of pipe
fflush(stderr);
r = dup2(pfd[1], 2);
if (r < 0) {
perror("dup2 - failed to alias stderr to write end of pipe for --monitor");
exit(1);
}
r = close(pfd[1]);
if (r < 0) {
perror("close - failed to close write end of pipe for --monitor");
exit(1);
}
// Don't allow a dying monitor to kill the main process
sighandler_t sr0 = signal(SIGPIPE, SIG_IGN);
sighandler_t sr1 = signal(SIGPIPE, SIG_IGN);
if (SIG_ERR == sr0 || SIG_ERR == sr1) {
perror("signal - failed to edit signal mask for --monitor");
exit(1);
}
// Fork a child process
forkpid = fork();
if (forkpid < 0) {
perror("fork - failed to fork child process for --monitor");
exit(1);
}
// Child: launch monitor command
if (0 == forkpid) {
// Make stdin read end of pipe
r = dup2(pfd[0], 0);
if (r < 0) {
perror("dup2 - in child, failed to alias read end of pipe to stdin for --monitor");
exit(1);
}
close(pfd[0]);
if (r < 0) {
perror("close - in child, failed to close read end of pipe for --monitor");
exit(1);
}
// Launch user specified command
execl("/bin/bash", "/bin/bash", "-c", opt_stderr_cmd, (char*)NULL);
perror("execl - in child failed to exec user specified command for --monitor");
exit(1);
}
// Parent: clean up unused fds and bail
r = close(pfd[0]);
if (r < 0) {
perror("close - failed to close read end of pipe for --monitor");
exit(1);
}
}
#endif // defined(unix)
#ifdef HAVE_CURSES
static void enable_curses_windows(void)
{
int x,y;
getmaxyx(mainwin, y, x);
statuswin = newwin(logstart, x, 0, 0);
leaveok(statuswin, true);
logwin = newwin(y - logcursor, 0, logcursor, 0);
idlok(logwin, true);
scrollok(logwin, true);
leaveok(logwin, true);
cbreak();
noecho();
}
void enable_curses(void) {
lock_curses();
if (curses_active) {
unlock_curses();
return;
}
mainwin = initscr();
enable_curses_windows();
curses_active = true;
statusy = logstart;
unlock_curses();
}
#endif
/* Various noop functions for drivers that don't support or need their
* variants. */
static void noop_reinit_device(struct cgpu_info __maybe_unused *cgpu)
{
}
void blank_get_statline_before(char *buf, size_t bufsiz, struct cgpu_info __maybe_unused *cgpu)
{
tailsprintf(buf, bufsiz, " | ");
}
static void noop_get_statline(char __maybe_unused *buf, size_t __maybe_unused bufsiz, struct cgpu_info __maybe_unused *cgpu)
{
}
static bool noop_get_stats(struct cgpu_info __maybe_unused *cgpu)
{
return true;
}
static bool noop_thread_prepare(struct thr_info __maybe_unused *thr)
{
return true;
}
static uint64_t noop_can_limit_work(struct thr_info __maybe_unused *thr)
{
return 0xffffffff;
}
static bool noop_thread_init(struct thr_info __maybe_unused *thr)
{
return true;
}
static bool noop_prepare_work(struct thr_info __maybe_unused *thr, struct work __maybe_unused *work)
{
return true;
}
static void noop_hw_error(struct thr_info __maybe_unused *thr)
{
}
static void noop_thread_shutdown(struct thr_info __maybe_unused *thr)
{
}
static void noop_thread_enable(struct thr_info __maybe_unused *thr)
{
}
static void noop_detect(void)
{
}
#define noop_flush_work noop_reinit_device
#define noop_update_work noop_reinit_device
#define noop_queue_full noop_get_stats
#define noop_zero_stats noop_reinit_device
/* Fill missing driver drv functions with noops */
void fill_device_drv(struct device_drv *drv)
{
if (!drv->drv_detect)
drv->drv_detect = &noop_detect;
if (!drv->reinit_device)
drv->reinit_device = &noop_reinit_device;
if (!drv->get_statline_before)
drv->get_statline_before = &blank_get_statline_before;
if (!drv->get_statline)
drv->get_statline = &noop_get_statline;
if (!drv->get_stats)
drv->get_stats = &noop_get_stats;
if (!drv->thread_prepare)
drv->thread_prepare = &noop_thread_prepare;
if (!drv->can_limit_work)
drv->can_limit_work = &noop_can_limit_work;
if (!drv->thread_init)
drv->thread_init = &noop_thread_init;
if (!drv->prepare_work)
drv->prepare_work = &noop_prepare_work;
if (!drv->hw_error)
drv->hw_error = &noop_hw_error;
if (!drv->thread_shutdown)
drv->thread_shutdown = &noop_thread_shutdown;
if (!drv->thread_enable)
drv->thread_enable = &noop_thread_enable;
if (!drv->hash_work)
drv->hash_work = &hash_sole_work;
if (!drv->flush_work)
drv->flush_work = &noop_flush_work;
if (!drv->update_work)
drv->update_work = &noop_update_work;
if (!drv->queue_full)
drv->queue_full = &noop_queue_full;
if (!drv->zero_stats)
drv->zero_stats = &noop_zero_stats;
if (!drv->max_diff)
drv->max_diff = 1;
if (!drv->working_diff)
drv->working_diff = 1;
}
void enable_device(int i)
{
rd_lock(&devices_lock);
devices[i]->deven = DEV_ENABLED;
rd_unlock(&devices_lock);
}
struct _cgpu_devid_counter {
char name[4];
int lastid;
UT_hash_handle hh;
};
static void adjust_mostdevs(void)
{
int i;
most_devices = 0;
for (i = 0; i < total_devices; i++) {
if (devices_enabled[i]) {
most_devices++;
}
}
}
bool add_cgpu(struct cgpu_info *cgpu)
{
static struct _cgpu_devid_counter *devids = NULL;
struct _cgpu_devid_counter *d;
HASH_FIND_STR(devids, cgpu->drv->name, d);
if (d)
cgpu->device_id = ++d->lastid;
else {
d = (struct _cgpu_devid_counter *)malloc(sizeof(*d));
memcpy(d->name, cgpu->drv->name, sizeof(d->name));
cgpu->device_id = d->lastid = 0;
HASH_ADD_STR(devids, name, d);
}
wr_lock(&devices_lock);
devices = (struct cgpu_info **)realloc(devices, sizeof(struct cgpu_info *) * (total_devices + 2));
wr_unlock(&devices_lock);
mutex_lock(&stats_lock);
cgpu->last_device_valid_work = time(NULL);
mutex_unlock(&stats_lock);
wr_lock(&devices_lock);
devices[total_devices++] = cgpu;
wr_unlock(&devices_lock);
adjust_mostdevs();
return true;
}
static void probe_pools(void)
{
int i;
for (i = 0; i < total_pools; i++)
{
struct pool *pool = pools[i];
pool->testing = true;
pthread_create(&pool->test_thread, NULL, test_pool_thread, (void *)pool);
}
}
static void restart_mining_threads(unsigned int new_n_threads)
{
struct thr_info *thr;
unsigned int i, j, k;
// Stop and free threads
if (mining_thr)
{
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++)
{
applog(LOG_DEBUG, "Shutting down thread %d", i);
mining_thr[i]->cgpu->shutdown = true;
}
rd_unlock(&mining_thr_lock);
// kill_mining will rd lock mining_thr_lock
kill_mining();
applog(LOG_DEBUG, "Finish switching pools");
// Finish switching pools
mutex_lock(&algo_switch_wait_lock);
algo_switch_n = 0;
mutex_unlock(&algo_switch_wait_lock);
applog(LOG_DEBUG, "Shutdown OpenCL contexts...");
rd_lock(&mining_thr_lock);
for (i = 0; i < mining_threads; i++)
{
thr = mining_thr[i];
thr->cgpu->drv->thread_shutdown(thr);
thr->cgpu->shutdown = false;
}
rd_unlock(&mining_thr_lock);
}
wr_lock(&mining_thr_lock);
if (mining_thr)
{
applog(LOG_DEBUG, "Free old mining and device thread memory...");
rd_lock(&devices_lock);
for (i = 0; i < total_devices; i++) {
if (devices[i]->thr) free(devices[i]->thr);
}
rd_unlock(&devices_lock);
for (i = 0; i < mining_threads; i++) {
free(mining_thr[i]);
}
free(mining_thr);
}
// Alloc
applog(LOG_DEBUG, "Allocate new threads...");
mining_threads = (int) new_n_threads;
#ifdef HAVE_CURSES
adj_width(mining_threads, &dev_width);
#endif
mining_thr = (struct thr_info **)calloc(mining_threads, sizeof(thr));
if (!mining_thr)
quit(1, "Failed to calloc mining_thr");
for (i = 0; i < mining_threads; i++) {
mining_thr[i] = (struct thr_info *)calloc(1, sizeof(struct thr_info));
if (!mining_thr[i])
quit(1, "Failed to calloc mining_thr[%d]", i);
}
rd_lock(&devices_lock);
// Start threads
struct pool *pool;
if(gpu_initialized)
pool = current_pool();
else
pool = pools[init_pool];
k = 0;
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = devices[i];
cgpu->thr = (struct thr_info **)malloc(sizeof(struct thr_info *) * (cgpu->threads+1));
cgpu->thr[cgpu->threads] = NULL;
cgpu->status = LIFE_INIT;
if (cgpu->deven == DEV_DISABLED && opt_removedisabled) {
cgpu->threads = 0;
continue;
}
applog(LOG_DEBUG, "Assign threads for device %d", i);
for (j = 0; j < cgpu->threads; ++j, ++k)
{
thr = mining_thr[k];
thr->id = k;
thr->pool_no = pool->pool_no;
applog(LOG_DEBUG, "Thread %d set pool = %d (%s)", k, thr->pool_no, isnull(get_pool_name(pools[thr->pool_no]), ""));
thr->cgpu = cgpu;
thr->device_thread = j;
cgtime(&thr->last);
cgpu->thr[j] = thr;
if (!cgpu->drv->thread_prepare(thr)) {
applog(LOG_ERR, "thread_prepare failed for thread %d", thr->id);
continue;
}
}
}
rd_unlock(&devices_lock);
wr_unlock(&mining_thr_lock);
rd_lock(&devices_lock);
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = devices[i];
for (j = 0; j < cgpu->threads; ++j) {
thr = cgpu->thr[j];
applog(LOG_DEBUG, "Starting device %d mining thread %d...", i, j);
if (unlikely(thr_info_create(thr, NULL, miner_thread, thr)))
quit(1, "thread %d create failed", thr->id);
/* Enable threads for devices set not to mine but disable
* their queue in case we wish to enable them later */
if (cgpu->deven != DEV_DISABLED) {
applog(LOG_DEBUG, "Pushing sem post to thread %d", thr->id);
cgsem_post(&thr->sem);
}
}
}
rd_unlock(&devices_lock);
}
static void *restart_mining_threads_thread(void *userdata)
{
//get thread id
pthread_t t = pthread_self();
//detach
pthread_detach(t);
//restart mining threads
restart_mining_threads((unsigned int) (intptr_t) userdata);
return NULL;
}
#define DRIVER_FILL_DEVICE_DRV(X) fill_device_drv(&X##_drv);
int main(int argc, char *argv[])
{
#ifndef _MSC_VER
struct sigaction handler;
#endif
struct thr_info *thr;
struct block *block;
int i;
char *s;
/* This dangerous function tramples random dynamically allocated
* variables so do it before anything at all */
if (unlikely(curl_global_init(CURL_GLOBAL_ALL)))
quit(1, "Failed to curl_global_init");
#if LOCK_TRACKING
// Must be first
if (unlikely(pthread_mutex_init(&lockstat_lock, NULL)))
quithere(1, "Failed to pthread_mutex_init lockstat_lock errno=%d", errno);
#endif
// initialize default profile (globals) before reading config options
init_default_profile();
initial_args = (const char **)malloc(sizeof(char *)* (argc + 1));
for (i = 0; i < argc; i++)
initial_args[i] = (const char *)strdup(argv[i]);
initial_args[argc] = NULL;
mutex_init(&hash_lock);
mutex_init(&console_lock);
cglock_init(&control_lock);
mutex_init(&stats_lock);
mutex_init(&sharelog_lock);
cglock_init(&ch_lock);
mutex_init(&sshare_lock);
rwlock_init(&blk_lock);
rwlock_init(&netacc_lock);
rwlock_init(&mining_thr_lock);
rwlock_init(&devices_lock);
mutex_init(&algo_switch_lock);
mutex_init(&lp_lock);
if (unlikely(pthread_cond_init(&lp_cond, NULL)))
quit(1, "Failed to pthread_cond_init lp_cond");
mutex_init(&restart_lock);
if (unlikely(pthread_cond_init(&restart_cond, NULL)))
quit(1, "Failed to pthread_cond_init restart_cond");
mutex_init(&algo_switch_wait_lock);
if (unlikely(pthread_cond_init(&algo_switch_wait_cond, NULL)))
quit(1, "Failed to pthread_cond_init algo_switch_wait_cond");
if (unlikely(pthread_cond_init(&gws_cond, NULL)))
quit(1, "Failed to pthread_cond_init gws_cond");
/* Create a unique get work queue */
getq = tq_new();
if (!getq)
quit(1, "Failed to create getq");
/* We use the getq mutex as the staged lock */
stgd_lock = &getq->mutex;
snprintf(packagename, sizeof(packagename), "%s %s", PACKAGE, CGMINER_VERSION);
#ifndef WIN32
signal(SIGPIPE, SIG_IGN);
#else
timeBeginPeriod(1);
#endif
#ifndef _MSC_VER
handler.sa_handler = &sighandler;
handler.sa_flags = 0;
sigemptyset(&handler.sa_mask);
sigaction(SIGTERM, &handler, &termhandler);
sigaction(SIGINT, &handler, &inthandler);
#endif
/* enable debug output to file by default is stderr is redirected to a file */
opt_debug = !isatty(fileno((FILE *)stderr));
/* opt_kernel_path defaults to SGMINER_PREFIX */
opt_kernel_path = (char *)alloca(PATH_MAX);
strcpy(opt_kernel_path, SGMINER_PREFIX);
/* sgminer_path is current dir */
sgminer_path = (char *)alloca(PATH_MAX);
#ifndef _MSC_VER
s = strdup(argv[0]);
strcpy(sgminer_path, dirname(s));
free(s);
#else
GetCurrentDirectory(PATH_MAX - 1, sgminer_path);
#endif
/* Default algorithm specified in algorithm.c ATM */
set_algorithm(&default_profile.algorithm, "scrypt");
devcursor = 8;
logstart = devcursor + 1;
logcursor = logstart + 1;
block = (struct block *)calloc(sizeof(struct block), 1);
if (unlikely(!block))
quit (1, "main OOM");
for (i = 0; i < 36; i++)
strcat(block->hash, "0");
HASH_ADD_STR(blocks, hash, block);
strcpy(current_hash, block->hash);
INIT_LIST_HEAD(&scan_devices);
memset(gpus, 0, sizeof(gpus));
for (i = 0; i < MAX_GPUDEVICES; i++)
gpus[i].dynamic = true;
/* parse config and command line */
opt_register_table(opt_config_table,
"Options for both config file and command line");
opt_register_table(opt_cmdline_table,
"Options for command line only");
opt_parse(&argc, argv, applog_and_exit);
if (argc != 1)
quit(1, "Unexpected extra commandline arguments");
if (!config_loaded)
load_default_config();
//load default profile if specified in config
load_default_profile();
#ifdef HAVE_CURSES
if (opt_realquiet || opt_display_devs)
use_curses = false;
if (use_curses)
enable_curses();
#endif
applog(LOG_WARNING, "Started %s", packagename);
applog(LOG_WARNING, "* using Jansson %s", JANSSON_VERSION);
if (cnfbuf) {
applog(LOG_NOTICE, "Loaded configuration file %s", cnfbuf);
switch (fileconf_load) {
case 0:
applog(LOG_WARNING, "Fatal JSON error in configuration file.");
applog(LOG_WARNING, "Configuration file could not be used.");
break;
case -1:
applog(LOG_WARNING, "Error in configuration file, partially loaded.");
if (use_curses)
applog(LOG_WARNING, "Start sgminer with -T to see what failed to load.");
break;
default:
break;
}
free(cnfbuf);
cnfbuf = NULL;
}
if (want_per_device_stats)
opt_verbose = true;
total_control_threads = 8;
control_thr = (struct thr_info *)calloc(total_control_threads, sizeof(*thr));
if (!control_thr)
quit(1, "Failed to calloc control_thr");
gwsched_thr_id = 0;
//Detect GPUs
/* Use the DRIVER_PARSE_COMMANDS macro to fill all the device_drvs */
DRIVER_PARSE_COMMANDS(DRIVER_FILL_DEVICE_DRV)
// this will set total_devices
opencl_drv.drv_detect();
if (opt_display_devs) {
applog(LOG_ERR, "Devices detected:");
for (i = 0; i < total_devices; ++i) {
struct cgpu_info *cgpu = devices[i];
if (cgpu->name)
applog(LOG_ERR, " %2d. %s %d: %s (driver: %s)", i, cgpu->drv->name, cgpu->device_id, cgpu->name, cgpu->drv->dname);
else
applog(LOG_ERR, " %2d. %s %d (driver: %s)", i, cgpu->drv->name, cgpu->device_id, cgpu->drv->dname);
}
quit(0, "%d devices listed", total_devices);
}
//apply default settings to GPUs
apply_defaults();
//apply pool-specific config from profiles
apply_pool_profiles();
most_devices = 0;
mining_threads = 0;
if (opt_devs_enabled) {
for (i = 0; i < MAX_DEVICES; i++) {
if (devices_enabled[i]) {
if (i >= total_devices)
quit (1, "Command line options set a device that doesn't exist");
enable_device(i);
mining_threads += devices[i]->threads;
most_devices++;
} else if (i < total_devices) {
devices[i]->deven = DEV_DISABLED;
if (!opt_removedisabled)
mining_threads += devices[i]->threads;
}
}
} else {
for (i = 0; i < total_devices; ++i) {
enable_device(i);
mining_threads += devices[i]->threads;
}
most_devices = total_devices;
}
#ifdef HAVE_CURSES
adj_width(mining_threads, &dev_width);
#endif
if (mining_threads == 0)
quit(1, "All devices disabled, cannot mine!");
load_temp_cutoffs();
rd_lock(&devices_lock);
for (i = 0; i < total_devices; ++i)
devices[i]->sgminer_stats.getwork_wait_min.tv_sec = MIN_SEC_UNSET;
rd_unlock(&devices_lock);
if (!opt_compact) {
logstart += (opt_removedisabled ? most_devices : total_devices);
logcursor = logstart + 1;
#ifdef HAVE_CURSES
check_winsizes();
#endif
}
if (!getenv("GPU_MAX_ALLOC_PERCENT"))
applog(LOG_WARNING, "WARNING: GPU_MAX_ALLOC_PERCENT is not specified!");
if (!getenv("GPU_USE_SYNC_OBJECTS"))
applog(LOG_WARNING, "WARNING: GPU_USE_SYNC_OBJECTS is not specified!");
if (!total_pools) {
applog(LOG_WARNING, "Need to specify at least one pool server.");
#ifdef HAVE_CURSES
if (!use_curses || !input_pool(false))
#endif
quit(1, "Pool setup failed");
}
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
size_t siz;
pool->sgminer_stats.getwork_wait_min.tv_sec = MIN_SEC_UNSET;
pool->sgminer_pool_stats.getwork_wait_min.tv_sec = MIN_SEC_UNSET;
if (!pool->rpc_userpass) {
if (!pool->rpc_user || !pool->rpc_pass)
quit(1, "No login credentials supplied for %s", get_pool_name(pool));
siz = strlen(pool->rpc_user) + strlen(pool->rpc_pass) + 2;
pool->rpc_userpass = (char *)malloc(siz);
if (!pool->rpc_userpass)
quit(1, "Failed to malloc userpass");
snprintf(pool->rpc_userpass, siz, "%s:%s", pool->rpc_user, pool->rpc_pass);
}
}
/* Set the currentpool to pool 0 */
currentpool = pools[0];
#ifdef HAVE_SYSLOG_H
if (use_syslog)
openlog(PACKAGE, LOG_PID, LOG_USER);
#endif
#if defined(unix) || defined(__APPLE__)
if (opt_stderr_cmd)
fork_monitor();
#endif // defined(unix)
/* Set pool state */
for (i = 0; i < total_pools; i++) {
struct pool *pool = pools[i];
switch (pool->state) {
case POOL_DISABLED:
disable_pool(pool);
break;
case POOL_ENABLED:
enable_pool(pool);
break;
case POOL_HIDDEN:
i--; /* Reiterate over this index. */
remove_pool(pool);
break;
case POOL_REJECTING:
reject_pool(pool);
break;
default:
enable_pool(pool);
break;
}
pool->idle = true;
}
applog(LOG_NOTICE, "Probing for an alive pool");
int slept = 0;
do {
/* Look for at least one active pool before starting */
probe_pools();
do {
sleep(1);
slept++;
} while (!pools_active && slept < 60);
if (!pools_active) {
applog(LOG_ERR, "No servers were found that could be used to get work from.");
applog(LOG_ERR, "Please check the details from the list below of the servers you have input");
applog(LOG_ERR, "Most likely you have input the wrong URL, forgotten to add a port, or have not set up workers");
for (i = 0; i < total_pools; i++) {
struct pool *pool;
pool = pools[i];
applog(LOG_WARNING, "Pool: %d URL: %s User: %s Password: %s",
i, pool->rpc_url, pool->rpc_user, pool->rpc_pass);
}
#ifdef HAVE_CURSES
if (use_curses) {
halfdelay(150);
applog(LOG_ERR, "Press any key to exit, or sgminer will try again in 15s.");
if (getch() != ERR)
quit(0, "No servers could be used! Exiting.");
cbreak();
} else
#endif
quit(0, "No servers could be used! Exiting.");
}
} while (!pools_active);
//wait for GPUs to be initialized after first alive pool is found
slept = 0;
do {
sleep(1);
slept++;
} while (!gpu_initialized && slept < 60);
if(slept >= 60)
applog(LOG_WARNING, "GPUs did not become initialized in 60 seconds...");
rd_lock(&devices_lock);
total_mhashes_done = 0;
for (i = 0; i < total_devices; i++) {
struct cgpu_info *cgpu = devices[i];
cgpu->rolling = cgpu->total_mhashes = 0;
}
rd_unlock(&devices_lock);
cgtime(&total_tv_start);
cgtime(&total_tv_end);
get_datestamp(datestamp, sizeof(datestamp), &total_tv_start);
launch_time = total_tv_start;
watchpool_thr_id = 2;
thr = &control_thr[watchpool_thr_id];
/* start watchpool thread */
if (thr_info_create(thr, NULL, watchpool_thread, NULL))
quit(1, "watchpool thread create failed");
pthread_detach(thr->pth);
watchdog_thr_id = 3;
thr = &control_thr[watchdog_thr_id];
/* start watchdog thread */
if (thr_info_create(thr, NULL, watchdog_thread, NULL))
quit(1, "watchdog thread create failed");
pthread_detach(thr->pth);
/* Create reinit gpu thread */
gpur_thr_id = 4;
thr = &control_thr[gpur_thr_id];
thr->q = tq_new();
if (!thr->q)
quit(1, "tq_new failed for gpur_thr_id");
if (thr_info_create(thr, NULL, reinit_gpu, thr))
quit(1, "reinit_gpu thread create failed");
/* Create API socket thread */
api_thr_id = 5;
thr = &control_thr[api_thr_id];
if (thr_info_create(thr, NULL, api_thread, thr))
quit(1, "API thread create failed");
#ifdef HAVE_CURSES
/* Create curses input thread for keyboard input. Create this last so
* that we know all threads are created since this can call kill_work
* to try and shut down all previous threads. */
input_thr_id = 7;
thr = &control_thr[input_thr_id];
if (thr_info_create(thr, NULL, input_thread, thr))
quit(1, "input thread create failed");
pthread_detach(thr->pth);
#endif
/* Just to be sure */
if (total_control_threads != 8)
quit(1, "incorrect total_control_threads (%d) should be 8", total_control_threads);
/* Once everything is set up, main() becomes the getwork scheduler */
while (42) {
int ts, max_staged = opt_queue;
struct pool *pool, *cp;
bool lagging = false;
struct timespec then;
struct timeval now;
struct work *work;
if (opt_work_update)
signal_work_update();
opt_work_update = false;
cp = current_pool();
/* If the primary pool is a getwork pool and cannot roll work,
* try to stage one extra work per mining thread */
if (!pool_localgen(cp) && !staged_rollable)
max_staged += mining_threads;
cgtime(&now);
then.tv_sec = now.tv_sec + 2;
then.tv_nsec = now.tv_usec * 1000;
mutex_lock(stgd_lock);
ts = __total_staged();
if (!pool_localgen(cp) && !ts && !opt_fail_only)
lagging = true;
/* Wait until hash_pop tells us we need to create more work */
if (ts > max_staged) {
pthread_cond_timedwait(&gws_cond, stgd_lock, &then);
ts = __total_staged();
}
mutex_unlock(stgd_lock);
if (ts > max_staged) {
/* Keeps slowly generating work even if it's not being
* used to keep last_getwork incrementing and to see
* if pools are still alive. */
work = hash_pop(false);
if (work) {
applog(LOG_DEBUG,
"[THR%d] Staged work: total (%d) > max (%d), discarding",
work->thr_id, ts, max_staged);
discard_work(work);
}
continue;
}
work = make_work();
if (lagging && !pool_tset(cp, &cp->lagging)) {
applog(LOG_WARNING, "%s not providing work fast enough", cp->name);
cp->getfail_occasions++;
total_go++;
if (!pool_localgen(cp))
applog(LOG_INFO, "Increasing queue to %d", ++opt_queue);
}
pool = select_pool(lagging);
retry:
if (pool->has_stratum) {
while (!pool->stratum_active || !pool->stratum_notify) {
struct pool *altpool = select_pool(true);
cgsleep_ms(5000);
if (altpool != pool) {
pool = altpool;
goto retry;
}
}
gen_stratum_work(pool, work);
applog(LOG_DEBUG, "Generated stratum work");
stage_work(work);
continue;
}
#ifdef HAVE_LIBCURL
struct curl_ent *ce;
if (pool->has_gbt) {
while (pool->idle) {
struct pool *altpool = select_pool(true);
cgsleep_ms(5000);
if (altpool != pool) {
pool = altpool;
goto retry;
}
}
gen_gbt_work(pool, work);
applog(LOG_DEBUG, "Generated GBT work");
stage_work(work);
continue;
}
if (clone_available()) {
applog(LOG_DEBUG, "Cloned getwork work");
free_work(work);
continue;
}
work->pool = pool;
ce = pop_curl_entry(pool);
/* obtain new work from bitcoin via JSON-RPC */
if (!get_upstream_work(work, ce->curl, ce->curl_err_str)) {
applog(LOG_DEBUG, "%s json_rpc_call failed on get work, retrying in 5s", get_pool_name(pool));
/* Make sure the pool just hasn't stopped serving
* requests but is up as we'll keep hammering it */
if (++pool->seq_getfails > mining_threads + opt_queue)
pool_died(pool);
cgsleep_ms(5000);
push_curl_entry(ce, pool);
pool = select_pool(!opt_fail_only);
goto retry;
}
if (ts >= max_staged)
pool_tclear(pool, &pool->lagging);
if (pool_tclear(pool, &pool->idle))
pool_resus(pool);
applog(LOG_DEBUG, "Generated getwork work");
stage_work(work);
push_curl_entry(ce, pool);
#endif /* HAVE_LIBCURL */
}
return 0;
}