X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=cmd%2Fztest%2Fztest.c;h=b38d7b1d0ee492808853f2f48446214433b6ede1;hb=c99c90015ece64746e20b74245caca41d1dbefe1;hp=b5cee5b3bd8dfeb387227aece1622b43cf55149e;hpb=090ff0929ec7a2e2e65efaaddb9981d15964f7d9;p=zfs.git diff --git a/cmd/ztest/ztest.c b/cmd/ztest/ztest.c index b5cee5b..b38d7b1 100644 --- a/cmd/ztest/ztest.c +++ b/cmd/ztest/ztest.c @@ -20,6 +20,8 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. + * Copyright (c) 2012 by Delphix. All rights reserved. + * Copyright 2011 Nexenta Systems, Inc. All rights reserved. */ /* @@ -49,7 +51,9 @@ * At random times, the child self-immolates with a SIGKILL. * This is the software equivalent of pulling the power cord. * The parent then runs the test again, using the existing - * storage pool, as many times as desired. + * storage pool, as many times as desired. If backwards compatability + * testing is enabled ztest will sometimes run the "older" version + * of ztest after a SIGKILL. * * (6) To verify that we don't have future leaks or temporal incursions, * many of the functional tests record the transaction group number @@ -57,6 +61,9 @@ * the transaction group number is less than the current, open txg. * If you add a new test, please do this if applicable. * + * (7) Threads are created with a reduced stack size, for sanity checking. + * Therefore, it's important not to allocate huge buffers on the stack. + * * When run with no arguments, ztest runs for about five minutes and * produces no output if successful. To get a little bit of information, * specify -V. To get more information, specify -VV, and so on. @@ -66,9 +73,15 @@ * You can ask more more vdevs [-v], datasets [-d], or threads [-t] * to increase the pool capacity, fanout, and overall stress level. * - * The -N(okill) option will suppress kills, so each child runs to completion. - * This can be useful when you're trying to distinguish temporal incursions - * from plain old race conditions. + * Use the -k option to set the desired frequency of kills. + * + * When ztest invokes itself it passes all relevant information through a + * temporary file which is mmap-ed in the child process. This allows shared + * memory to survive the exec syscall. The ztest_shared_hdr_t struct is always + * stored at offset 0 of this file and contains information on the size and + * number of shared structures in the file. The information stored in this file + * must remain backwards compatible with older versions of ztest so that + * ztest can invoke them during backwards compatibility testing (-B). */ #include @@ -96,6 +109,7 @@ #include #include #include +#include #include #include #include @@ -108,28 +122,82 @@ #include #include -static char cmdname[] = "ztest"; -static char *zopt_pool = cmdname; - -static uint64_t zopt_vdevs = 5; -static uint64_t zopt_vdevtime; -static int zopt_ashift = SPA_MINBLOCKSHIFT; -static int zopt_mirrors = 2; -static int zopt_raidz = 4; -static int zopt_raidz_parity = 1; -static size_t zopt_vdev_size = SPA_MINDEVSIZE; -static int zopt_datasets = 7; -static int zopt_threads = 23; -static uint64_t zopt_passtime = 60; /* 60 seconds */ -static uint64_t zopt_killrate = 70; /* 70% kill rate */ -static int zopt_verbose = 0; -static int zopt_init = 1; -static char *zopt_dir = "/tmp"; -static uint64_t zopt_time = 300; /* 5 minutes */ -static uint64_t zopt_maxloops = 50; /* max loops during spa_freeze() */ +static int ztest_fd_data = -1; +static int ztest_fd_rand = -1; + +typedef struct ztest_shared_hdr { + uint64_t zh_hdr_size; + uint64_t zh_opts_size; + uint64_t zh_size; + uint64_t zh_stats_size; + uint64_t zh_stats_count; + uint64_t zh_ds_size; + uint64_t zh_ds_count; +} ztest_shared_hdr_t; + +static ztest_shared_hdr_t *ztest_shared_hdr; + +typedef struct ztest_shared_opts { + char zo_pool[MAXNAMELEN]; + char zo_dir[MAXNAMELEN]; + char zo_alt_ztest[MAXNAMELEN]; + char zo_alt_libpath[MAXNAMELEN]; + uint64_t zo_vdevs; + uint64_t zo_vdevtime; + size_t zo_vdev_size; + int zo_ashift; + int zo_mirrors; + int zo_raidz; + int zo_raidz_parity; + int zo_datasets; + int zo_threads; + uint64_t zo_passtime; + uint64_t zo_killrate; + int zo_verbose; + int zo_init; + uint64_t zo_time; + uint64_t zo_maxloops; + uint64_t zo_metaslab_gang_bang; +} ztest_shared_opts_t; + +static const ztest_shared_opts_t ztest_opts_defaults = { + .zo_pool = { 'z', 't', 'e', 's', 't', '\0' }, + .zo_dir = { '/', 't', 'm', 'p', '\0' }, + .zo_alt_ztest = { '\0' }, + .zo_alt_libpath = { '\0' }, + .zo_vdevs = 5, + .zo_ashift = SPA_MINBLOCKSHIFT, + .zo_mirrors = 2, + .zo_raidz = 4, + .zo_raidz_parity = 1, + .zo_vdev_size = SPA_MINDEVSIZE, + .zo_datasets = 7, + .zo_threads = 23, + .zo_passtime = 60, /* 60 seconds */ + .zo_killrate = 70, /* 70% kill rate */ + .zo_verbose = 0, + .zo_init = 1, + .zo_time = 300, /* 5 minutes */ + .zo_maxloops = 50, /* max loops during spa_freeze() */ + .zo_metaslab_gang_bang = 32 << 10 +}; + +extern uint64_t metaslab_gang_bang; +extern uint64_t metaslab_df_alloc_threshold; + +static ztest_shared_opts_t *ztest_shared_opts; +static ztest_shared_opts_t ztest_opts; + +typedef struct ztest_shared_ds { + uint64_t zd_seq; +} ztest_shared_ds_t; + +static ztest_shared_ds_t *ztest_shared_ds; +#define ZTEST_GET_SHARED_DS(d) (&ztest_shared_ds[d]) #define BT_MAGIC 0x123456789abcdefULL -#define MAXFAULTS() (MAX(zs->zs_mirrors, 1) * (zopt_raidz_parity + 1) - 1) +#define MAXFAULTS() \ + (MAX(zs->zs_mirrors, 1) * (ztest_opts.zo_raidz_parity + 1) - 1) enum ztest_io_type { ZTEST_IO_WRITE_TAG, @@ -168,8 +236,8 @@ typedef enum { typedef struct rll { void *rll_writer; int rll_readers; - mutex_t rll_lock; - cond_t rll_cv; + kmutex_t rll_lock; + kcondvar_t rll_cv; } rll_t; typedef struct rl { @@ -201,12 +269,13 @@ typedef struct ztest_od { * Per-dataset state. */ typedef struct ztest_ds { + ztest_shared_ds_t *zd_shared; objset_t *zd_os; + krwlock_t zd_zilog_lock; zilog_t *zd_zilog; - uint64_t zd_seq; ztest_od_t *zd_od; /* debugging aid */ char zd_name[MAXNAMELEN]; - mutex_t zd_dirobj_lock; + kmutex_t zd_dirobj_lock; rll_t zd_object_lock[ZTEST_OBJECT_LOCKS]; rll_t zd_range_lock[ZTEST_RANGE_LOCKS]; } ztest_ds_t; @@ -220,11 +289,17 @@ typedef struct ztest_info { ztest_func_t *zi_func; /* test function */ uint64_t zi_iters; /* iterations per execution */ uint64_t *zi_interval; /* execute every seconds */ - uint64_t zi_call_count; /* per-pass count */ - uint64_t zi_call_time; /* per-pass time */ - uint64_t zi_call_next; /* next time to call this function */ } ztest_info_t; +typedef struct ztest_shared_callstate { + uint64_t zc_count; /* per-pass count */ + uint64_t zc_time; /* per-pass time */ + uint64_t zc_next; /* next time to call this function */ +} ztest_shared_callstate_t; + +static ztest_shared_callstate_t *ztest_shared_callstate; +#define ZTEST_GET_SHARED_CALLSTATE(c) (&ztest_shared_callstate[c]) + /* * Note: these aren't static because we want dladdr() to work. */ @@ -235,6 +310,7 @@ ztest_func_t ztest_dmu_commit_callbacks; ztest_func_t ztest_zap; ztest_func_t ztest_zap_parallel; ztest_func_t ztest_zil_commit; +ztest_func_t ztest_zil_remount; ztest_func_t ztest_dmu_read_write_zcopy; ztest_func_t ztest_dmu_objset_create_destroy; ztest_func_t ztest_dmu_prealloc; @@ -254,6 +330,8 @@ ztest_func_t ztest_vdev_LUN_growth; ztest_func_t ztest_vdev_add_remove; ztest_func_t ztest_vdev_aux_add_remove; ztest_func_t ztest_split_pool; +ztest_func_t ztest_reguid; +ztest_func_t ztest_spa_upgrade; uint64_t zopt_always = 0ULL * NANOSEC; /* all the time */ uint64_t zopt_incessant = 1ULL * NANOSEC / 10; /* every 1/10 second */ @@ -270,6 +348,7 @@ ztest_info_t ztest_info[] = { { ztest_zap_parallel, 100, &zopt_always }, { ztest_split_pool, 1, &zopt_always }, { ztest_zil_commit, 1, &zopt_incessant }, + { ztest_zil_remount, 1, &zopt_sometimes }, { ztest_dmu_read_write_zcopy, 1, &zopt_often }, { ztest_dmu_objset_create_destroy, 1, &zopt_often }, { ztest_dsl_prop_get_set, 1, &zopt_often }, @@ -283,13 +362,17 @@ ztest_info_t ztest_info[] = { { ztest_fault_inject, 1, &zopt_sometimes }, { ztest_ddt_repair, 1, &zopt_sometimes }, { ztest_dmu_snapshot_hold, 1, &zopt_sometimes }, + { ztest_reguid, 1, &zopt_sometimes }, { ztest_spa_rename, 1, &zopt_rarely }, { ztest_scrub, 1, &zopt_rarely }, + { ztest_spa_upgrade, 1, &zopt_rarely }, { ztest_dsl_dataset_promote_busy, 1, &zopt_rarely }, - { ztest_vdev_attach_detach, 1, &zopt_rarely }, + { ztest_vdev_attach_detach, 1, &zopt_rarely }, { ztest_vdev_LUN_growth, 1, &zopt_rarely }, - { ztest_vdev_add_remove, 1, &zopt_vdevtime }, - { ztest_vdev_aux_add_remove, 1, &zopt_vdevtime }, + { ztest_vdev_add_remove, 1, + &ztest_opts.zo_vdevtime }, + { ztest_vdev_aux_add_remove, 1, + &ztest_opts.zo_vdevtime }, }; #define ZTEST_FUNCS (sizeof (ztest_info) / sizeof (ztest_info_t)) @@ -299,16 +382,15 @@ ztest_info_t ztest_info[] = { * The callbacks are ordered by txg number. */ typedef struct ztest_cb_list { - mutex_t zcl_callbacks_lock; - list_t zcl_callbacks; + kmutex_t zcl_callbacks_lock; + list_t zcl_callbacks; } ztest_cb_list_t; /* * Stuff we need to share writably between parent and child. */ typedef struct ztest_shared { - char *zs_pool; - spa_t *zs_spa; + boolean_t zs_do_init; hrtime_t zs_proc_start; hrtime_t zs_proc_stop; hrtime_t zs_thread_start; @@ -319,12 +401,11 @@ typedef struct ztest_shared { uint64_t zs_vdev_aux; uint64_t zs_alloc; uint64_t zs_space; - mutex_t zs_vdev_lock; - rwlock_t zs_name_lock; - ztest_info_t zs_info[ZTEST_FUNCS]; uint64_t zs_splits; uint64_t zs_mirrors; - ztest_ds_t zs_zd[]; + uint64_t zs_metaslab_sz; + uint64_t zs_metaslab_df_alloc_threshold; + uint64_t zs_guid; } ztest_shared_t; #define ID_PARALLEL -1ULL @@ -332,11 +413,21 @@ typedef struct ztest_shared { static char ztest_dev_template[] = "%s/%s.%llua"; static char ztest_aux_template[] = "%s/%s.%s.%llu"; ztest_shared_t *ztest_shared; -uint64_t *ztest_seq; -static int ztest_random_fd; -static int ztest_dump_core = 1; +static spa_t *ztest_spa = NULL; +static ztest_ds_t *ztest_ds; + +static kmutex_t ztest_vdev_lock; + +/* + * The ztest_name_lock protects the pool and dataset namespace used by + * the individual tests. To modify the namespace, consumers must grab + * this lock as writer. Grabbing the lock as reader will ensure that the + * namespace does not change while the lock is held. + */ +static krwlock_t ztest_name_lock; +static boolean_t ztest_dump_core = B_TRUE; static boolean_t ztest_exiting; /* Global commit callback list */ @@ -360,7 +451,6 @@ static int zc_cb_counter = 0; extern uint64_t metaslab_gang_bang; extern uint64_t metaslab_df_alloc_threshold; -static uint64_t metaslab_sz; enum ztest_object { ZTEST_META_DNODE = 0, @@ -395,9 +485,10 @@ fatal(int do_perror, char *message, ...) { va_list args; int save_errno = errno; - char buf[FATAL_MSG_SZ]; + char *buf; (void) fflush(stdout); + buf = umem_alloc(FATAL_MSG_SZ, UMEM_NOFAIL); va_start(args, message); (void) sprintf(buf, "ztest: "); @@ -474,12 +565,14 @@ nicenumtoull(const char *buf) static void usage(boolean_t requested) { + const ztest_shared_opts_t *zo = &ztest_opts_defaults; + char nice_vdev_size[10]; char nice_gang_bang[10]; FILE *fp = requested ? stdout : stderr; - nicenum(zopt_vdev_size, nice_vdev_size); - nicenum(metaslab_gang_bang, nice_gang_bang); + nicenum(zo->zo_vdev_size, nice_vdev_size); + nicenum(zo->zo_metaslab_gang_bang, nice_gang_bang); (void) fprintf(fp, "Usage: %s\n" "\t[-v vdevs (default: %llu)]\n" @@ -500,39 +593,43 @@ usage(boolean_t requested) "\t[-T time (default: %llu sec)] total run time\n" "\t[-F freezeloops (default: %llu)] max loops in spa_freeze()\n" "\t[-P passtime (default: %llu sec)] time per pass\n" + "\t[-B alt_ztest (default: )] alternate ztest path\n" "\t[-h] (print help)\n" "", - cmdname, - (u_longlong_t)zopt_vdevs, /* -v */ + zo->zo_pool, + (u_longlong_t)zo->zo_vdevs, /* -v */ nice_vdev_size, /* -s */ - zopt_ashift, /* -a */ - zopt_mirrors, /* -m */ - zopt_raidz, /* -r */ - zopt_raidz_parity, /* -R */ - zopt_datasets, /* -d */ - zopt_threads, /* -t */ + zo->zo_ashift, /* -a */ + zo->zo_mirrors, /* -m */ + zo->zo_raidz, /* -r */ + zo->zo_raidz_parity, /* -R */ + zo->zo_datasets, /* -d */ + zo->zo_threads, /* -t */ nice_gang_bang, /* -g */ - zopt_init, /* -i */ - (u_longlong_t)zopt_killrate, /* -k */ - zopt_pool, /* -p */ - zopt_dir, /* -f */ - (u_longlong_t)zopt_time, /* -T */ - (u_longlong_t)zopt_maxloops, /* -F */ - (u_longlong_t)zopt_passtime); /* -P */ + zo->zo_init, /* -i */ + (u_longlong_t)zo->zo_killrate, /* -k */ + zo->zo_pool, /* -p */ + zo->zo_dir, /* -f */ + (u_longlong_t)zo->zo_time, /* -T */ + (u_longlong_t)zo->zo_maxloops, /* -F */ + (u_longlong_t)zo->zo_passtime); exit(requested ? 0 : 1); } static void process_options(int argc, char **argv) { + char *path; + ztest_shared_opts_t *zo = &ztest_opts; + int opt; uint64_t value; + char altdir[MAXNAMELEN] = { 0 }; - /* By default, test gang blocks for blocks 32K and greater */ - metaslab_gang_bang = 32 << 10; + bcopy(&ztest_opts_defaults, zo, sizeof (*zo)); while ((opt = getopt(argc, argv, - "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:hF:")) != EOF) { + "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:hF:B:")) != EOF) { value = 0; switch (opt) { case 'v': @@ -553,58 +650,71 @@ process_options(int argc, char **argv) } switch (opt) { case 'v': - zopt_vdevs = value; + zo->zo_vdevs = value; break; case 's': - zopt_vdev_size = MAX(SPA_MINDEVSIZE, value); + zo->zo_vdev_size = MAX(SPA_MINDEVSIZE, value); break; case 'a': - zopt_ashift = value; + zo->zo_ashift = value; break; case 'm': - zopt_mirrors = value; + zo->zo_mirrors = value; break; case 'r': - zopt_raidz = MAX(1, value); + zo->zo_raidz = MAX(1, value); break; case 'R': - zopt_raidz_parity = MIN(MAX(value, 1), 3); + zo->zo_raidz_parity = MIN(MAX(value, 1), 3); break; case 'd': - zopt_datasets = MAX(1, value); + zo->zo_datasets = MAX(1, value); break; case 't': - zopt_threads = MAX(1, value); + zo->zo_threads = MAX(1, value); break; case 'g': - metaslab_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, value); + zo->zo_metaslab_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, + value); break; case 'i': - zopt_init = value; + zo->zo_init = value; break; case 'k': - zopt_killrate = value; + zo->zo_killrate = value; break; case 'p': - zopt_pool = strdup(optarg); + (void) strlcpy(zo->zo_pool, optarg, + sizeof (zo->zo_pool)); break; case 'f': - zopt_dir = strdup(optarg); + path = realpath(optarg, NULL); + if (path == NULL) { + (void) fprintf(stderr, "error: %s: %s\n", + optarg, strerror(errno)); + usage(B_FALSE); + } else { + (void) strlcpy(zo->zo_dir, path, + sizeof (zo->zo_dir)); + } break; case 'V': - zopt_verbose++; + zo->zo_verbose++; break; case 'E': - zopt_init = 0; + zo->zo_init = 0; break; case 'T': - zopt_time = value; + zo->zo_time = value; break; case 'P': - zopt_passtime = MAX(1, value); + zo->zo_passtime = MAX(1, value); break; case 'F': - zopt_maxloops = MAX(1, value); + zo->zo_maxloops = MAX(1, value); + break; + case 'B': + (void) strlcpy(altdir, optarg, sizeof (altdir)); break; case 'h': usage(B_TRUE); @@ -616,17 +726,65 @@ process_options(int argc, char **argv) } } - zopt_raidz_parity = MIN(zopt_raidz_parity, zopt_raidz - 1); + zo->zo_raidz_parity = MIN(zo->zo_raidz_parity, zo->zo_raidz - 1); - zopt_vdevtime = (zopt_vdevs > 0 ? zopt_time * NANOSEC / zopt_vdevs : + zo->zo_vdevtime = + (zo->zo_vdevs > 0 ? zo->zo_time * NANOSEC / zo->zo_vdevs : UINT64_MAX >> 2); + + if (strlen(altdir) > 0) { + char *cmd; + char *realaltdir; + char *bin; + char *ztest; + char *isa; + int isalen; + + cmd = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + realaltdir = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + + VERIFY(NULL != realpath(getexecname(), cmd)); + if (0 != access(altdir, F_OK)) { + ztest_dump_core = B_FALSE; + fatal(B_TRUE, "invalid alternate ztest path: %s", + altdir); + } + VERIFY(NULL != realpath(altdir, realaltdir)); + + /* + * 'cmd' should be of the form "/usr/bin//ztest". + * We want to extract to determine if we should use + * 32 or 64 bit binaries. + */ + bin = strstr(cmd, "/usr/bin/"); + ztest = strstr(bin, "/ztest"); + isa = bin + 9; + isalen = ztest - isa; + (void) snprintf(zo->zo_alt_ztest, sizeof (zo->zo_alt_ztest), + "%s/usr/bin/%.*s/ztest", realaltdir, isalen, isa); + (void) snprintf(zo->zo_alt_libpath, sizeof (zo->zo_alt_libpath), + "%s/usr/lib/%.*s", realaltdir, isalen, isa); + + if (0 != access(zo->zo_alt_ztest, X_OK)) { + ztest_dump_core = B_FALSE; + fatal(B_TRUE, "invalid alternate ztest: %s", + zo->zo_alt_ztest); + } else if (0 != access(zo->zo_alt_libpath, X_OK)) { + ztest_dump_core = B_FALSE; + fatal(B_TRUE, "invalid alternate lib directory %s", + zo->zo_alt_libpath); + } + + umem_free(cmd, MAXPATHLEN); + umem_free(realaltdir, MAXPATHLEN); + } } static void ztest_kill(ztest_shared_t *zs) { - zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(zs->zs_spa)); - zs->zs_space = metaslab_class_get_space(spa_normal_class(zs->zs_spa)); + zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(ztest_spa)); + zs->zs_space = metaslab_class_get_space(spa_normal_class(ztest_spa)); (void) kill(getpid(), SIGKILL); } @@ -635,10 +793,12 @@ ztest_random(uint64_t range) { uint64_t r; + ASSERT3S(ztest_fd_rand, >=, 0); + if (range == 0) return (0); - if (read(ztest_random_fd, &r, sizeof (r)) != sizeof (r)) + if (read(ztest_fd_rand, &r, sizeof (r)) != sizeof (r)) fatal(1, "short read from /dev/urandom"); return (r % range); @@ -654,18 +814,20 @@ ztest_record_enospc(const char *s) static uint64_t ztest_get_ashift(void) { - if (zopt_ashift == 0) + if (ztest_opts.zo_ashift == 0) return (SPA_MINBLOCKSHIFT + ztest_random(3)); - return (zopt_ashift); + return (ztest_opts.zo_ashift); } static nvlist_t * -make_vdev_file(char *path, char *aux, size_t size, uint64_t ashift) +make_vdev_file(char *path, char *aux, char *pool, size_t size, uint64_t ashift) { - char pathbuf[MAXPATHLEN]; + char *pathbuf; uint64_t vdev; nvlist_t *file; + pathbuf = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + if (ashift == 0) ashift = ztest_get_ashift(); @@ -674,12 +836,15 @@ make_vdev_file(char *path, char *aux, size_t size, uint64_t ashift) if (aux != NULL) { vdev = ztest_shared->zs_vdev_aux; - (void) sprintf(path, ztest_aux_template, - zopt_dir, zopt_pool, aux, vdev); + (void) snprintf(path, MAXPATHLEN, + ztest_aux_template, ztest_opts.zo_dir, + pool == NULL ? ztest_opts.zo_pool : pool, + aux, vdev); } else { vdev = ztest_shared->zs_vdev_next_leaf++; - (void) sprintf(path, ztest_dev_template, - zopt_dir, zopt_pool, vdev); + (void) snprintf(path, MAXPATHLEN, + ztest_dev_template, ztest_opts.zo_dir, + pool == NULL ? ztest_opts.zo_pool : pool, vdev); } } @@ -696,28 +861,30 @@ make_vdev_file(char *path, char *aux, size_t size, uint64_t ashift) VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, path) == 0); VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); + umem_free(pathbuf, MAXPATHLEN); return (file); } static nvlist_t * -make_vdev_raidz(char *path, char *aux, size_t size, uint64_t ashift, int r) +make_vdev_raidz(char *path, char *aux, char *pool, size_t size, + uint64_t ashift, int r) { nvlist_t *raidz, **child; int c; if (r < 2) - return (make_vdev_file(path, aux, size, ashift)); + return (make_vdev_file(path, aux, pool, size, ashift)); child = umem_alloc(r * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < r; c++) - child[c] = make_vdev_file(path, aux, size, ashift); + child[c] = make_vdev_file(path, aux, pool, size, ashift); VERIFY(nvlist_alloc(&raidz, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(raidz, ZPOOL_CONFIG_TYPE, VDEV_TYPE_RAIDZ) == 0); VERIFY(nvlist_add_uint64(raidz, ZPOOL_CONFIG_NPARITY, - zopt_raidz_parity) == 0); + ztest_opts.zo_raidz_parity) == 0); VERIFY(nvlist_add_nvlist_array(raidz, ZPOOL_CONFIG_CHILDREN, child, r) == 0); @@ -730,19 +897,19 @@ make_vdev_raidz(char *path, char *aux, size_t size, uint64_t ashift, int r) } static nvlist_t * -make_vdev_mirror(char *path, char *aux, size_t size, uint64_t ashift, - int r, int m) +make_vdev_mirror(char *path, char *aux, char *pool, size_t size, + uint64_t ashift, int r, int m) { nvlist_t *mirror, **child; int c; if (m < 1) - return (make_vdev_raidz(path, aux, size, ashift, r)); + return (make_vdev_raidz(path, aux, pool, size, ashift, r)); child = umem_alloc(m * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < m; c++) - child[c] = make_vdev_raidz(path, aux, size, ashift, r); + child[c] = make_vdev_raidz(path, aux, pool, size, ashift, r); VERIFY(nvlist_alloc(&mirror, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(mirror, ZPOOL_CONFIG_TYPE, @@ -759,8 +926,8 @@ make_vdev_mirror(char *path, char *aux, size_t size, uint64_t ashift, } static nvlist_t * -make_vdev_root(char *path, char *aux, size_t size, uint64_t ashift, - int log, int r, int m, int t) +make_vdev_root(char *path, char *aux, char *pool, size_t size, uint64_t ashift, + int log, int r, int m, int t) { nvlist_t *root, **child; int c; @@ -770,7 +937,8 @@ make_vdev_root(char *path, char *aux, size_t size, uint64_t ashift, child = umem_alloc(t * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < t; c++) { - child[c] = make_vdev_mirror(path, aux, size, ashift, r, m); + child[c] = make_vdev_mirror(path, aux, pool, size, ashift, + r, m); VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_IS_LOG, log) == 0); } @@ -788,6 +956,27 @@ make_vdev_root(char *path, char *aux, size_t size, uint64_t ashift, return (root); } +/* + * Find a random spa version. Returns back a random spa version in the + * range [initial_version, SPA_VERSION_FEATURES]. + */ +static uint64_t +ztest_random_spa_version(uint64_t initial_version) +{ + uint64_t version = initial_version; + + if (version <= SPA_VERSION_BEFORE_FEATURES) { + version = version + + ztest_random(SPA_VERSION_BEFORE_FEATURES - version + 1); + } + + if (version > SPA_VERSION_BEFORE_FEATURES) + version = SPA_VERSION_FEATURES; + + ASSERT(SPA_VERSION_IS_SUPPORTED(version)); + return (version); +} + static int ztest_random_blocksize(void) { @@ -838,7 +1027,7 @@ ztest_dsl_prop_set_uint64(char *osname, zfs_prop_t prop, uint64_t value, { const char *propname = zfs_prop_to_name(prop); const char *valname; - char setpoint[MAXPATHLEN]; + char *setpoint; uint64_t curval; int error; @@ -850,24 +1039,26 @@ ztest_dsl_prop_set_uint64(char *osname, zfs_prop_t prop, uint64_t value, ztest_record_enospc(FTAG); return (error); } - ASSERT3U(error, ==, 0); + ASSERT0(error); + setpoint = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); VERIFY3U(dsl_prop_get(osname, propname, sizeof (curval), 1, &curval, setpoint), ==, 0); - if (zopt_verbose >= 6) { + if (ztest_opts.zo_verbose >= 6) { VERIFY(zfs_prop_index_to_string(prop, curval, &valname) == 0); (void) printf("%s %s = %s at '%s'\n", osname, propname, valname, setpoint); } + umem_free(setpoint, MAXPATHLEN); return (error); } static int -ztest_spa_prop_set_uint64(ztest_shared_t *zs, zpool_prop_t prop, uint64_t value) +ztest_spa_prop_set_uint64(zpool_prop_t prop, uint64_t value) { - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; nvlist_t *props = NULL; int error; @@ -882,7 +1073,7 @@ ztest_spa_prop_set_uint64(ztest_shared_t *zs, zpool_prop_t prop, uint64_t value) ztest_record_enospc(FTAG); return (error); } - ASSERT3U(error, ==, 0); + ASSERT0(error); return (error); } @@ -892,8 +1083,8 @@ ztest_rll_init(rll_t *rll) { rll->rll_writer = NULL; rll->rll_readers = 0; - VERIFY(_mutex_init(&rll->rll_lock, USYNC_THREAD, NULL) == 0); - VERIFY(cond_init(&rll->rll_cv, USYNC_THREAD, NULL) == 0); + mutex_init(&rll->rll_lock, NULL, MUTEX_DEFAULT, NULL); + cv_init(&rll->rll_cv, NULL, CV_DEFAULT, NULL); } static void @@ -901,32 +1092,32 @@ ztest_rll_destroy(rll_t *rll) { ASSERT(rll->rll_writer == NULL); ASSERT(rll->rll_readers == 0); - VERIFY(_mutex_destroy(&rll->rll_lock) == 0); - VERIFY(cond_destroy(&rll->rll_cv) == 0); + mutex_destroy(&rll->rll_lock); + cv_destroy(&rll->rll_cv); } static void ztest_rll_lock(rll_t *rll, rl_type_t type) { - VERIFY(mutex_lock(&rll->rll_lock) == 0); + mutex_enter(&rll->rll_lock); if (type == RL_READER) { while (rll->rll_writer != NULL) - (void) cond_wait(&rll->rll_cv, &rll->rll_lock); + (void) cv_wait(&rll->rll_cv, &rll->rll_lock); rll->rll_readers++; } else { while (rll->rll_writer != NULL || rll->rll_readers) - (void) cond_wait(&rll->rll_cv, &rll->rll_lock); + (void) cv_wait(&rll->rll_cv, &rll->rll_lock); rll->rll_writer = curthread; } - VERIFY(mutex_unlock(&rll->rll_lock) == 0); + mutex_exit(&rll->rll_lock); } static void ztest_rll_unlock(rll_t *rll) { - VERIFY(mutex_lock(&rll->rll_lock) == 0); + mutex_enter(&rll->rll_lock); if (rll->rll_writer) { ASSERT(rll->rll_readers == 0); @@ -938,9 +1129,9 @@ ztest_rll_unlock(rll_t *rll) } if (rll->rll_writer == NULL && rll->rll_readers == 0) - VERIFY(cond_broadcast(&rll->rll_cv) == 0); + cv_broadcast(&rll->rll_cv); - VERIFY(mutex_unlock(&rll->rll_lock) == 0); + mutex_exit(&rll->rll_lock); } static void @@ -989,15 +1180,19 @@ ztest_range_unlock(rl_t *rl) } static void -ztest_zd_init(ztest_ds_t *zd, objset_t *os) +ztest_zd_init(ztest_ds_t *zd, ztest_shared_ds_t *szd, objset_t *os) { zd->zd_os = os; zd->zd_zilog = dmu_objset_zil(os); - zd->zd_seq = 0; + zd->zd_shared = szd; dmu_objset_name(os, zd->zd_name); int l; - VERIFY(_mutex_init(&zd->zd_dirobj_lock, USYNC_THREAD, NULL) == 0); + if (zd->zd_shared != NULL) + zd->zd_shared->zd_seq = 0; + + rw_init(&zd->zd_zilog_lock, NULL, RW_DEFAULT, NULL); + mutex_init(&zd->zd_dirobj_lock, NULL, MUTEX_DEFAULT, NULL); for (l = 0; l < ZTEST_OBJECT_LOCKS; l++) ztest_rll_init(&zd->zd_object_lock[l]); @@ -1011,7 +1206,8 @@ ztest_zd_fini(ztest_ds_t *zd) { int l; - VERIFY(_mutex_destroy(&zd->zd_dirobj_lock) == 0); + mutex_destroy(&zd->zd_dirobj_lock); + rw_destroy(&zd->zd_zilog_lock); for (l = 0; l < ZTEST_OBJECT_LOCKS; l++) ztest_rll_destroy(&zd->zd_object_lock[l]); @@ -1580,7 +1776,7 @@ ztest_replay_setattr(ztest_ds_t *zd, lr_setattr_t *lr, boolean_t byteswap) ASSERT3U(lr->lr_size, >=, sizeof (*bbt)); ASSERT3U(lr->lr_size, <=, db->db_size); - VERIFY3U(dmu_set_bonus(db, lr->lr_size, tx), ==, 0); + VERIFY0(dmu_set_bonus(db, lr->lr_size, tx)); bbt = ztest_bt_bonus(db); ztest_bt_generate(bbt, os, lr->lr_foid, -1ULL, lr->lr_mode, txg, crtxg); @@ -1596,19 +1792,19 @@ ztest_replay_setattr(ztest_ds_t *zd, lr_setattr_t *lr, boolean_t byteswap) return (0); } -zil_replay_func_t *ztest_replay_vector[TX_MAX_TYPE] = { +zil_replay_func_t ztest_replay_vector[TX_MAX_TYPE] = { NULL, /* 0 no such transaction type */ - (zil_replay_func_t *)ztest_replay_create, /* TX_CREATE */ + (zil_replay_func_t)ztest_replay_create, /* TX_CREATE */ NULL, /* TX_MKDIR */ NULL, /* TX_MKXATTR */ NULL, /* TX_SYMLINK */ - (zil_replay_func_t *)ztest_replay_remove, /* TX_REMOVE */ + (zil_replay_func_t)ztest_replay_remove, /* TX_REMOVE */ NULL, /* TX_RMDIR */ NULL, /* TX_LINK */ NULL, /* TX_RENAME */ - (zil_replay_func_t *)ztest_replay_write, /* TX_WRITE */ - (zil_replay_func_t *)ztest_replay_truncate, /* TX_TRUNCATE */ - (zil_replay_func_t *)ztest_replay_setattr, /* TX_SETATTR */ + (zil_replay_func_t)ztest_replay_write, /* TX_WRITE */ + (zil_replay_func_t)ztest_replay_truncate, /* TX_TRUNCATE */ + (zil_replay_func_t)ztest_replay_setattr, /* TX_SETATTR */ NULL, /* TX_ACL */ NULL, /* TX_CREATE_ACL */ NULL, /* TX_CREATE_ATTR */ @@ -1754,7 +1950,7 @@ ztest_lookup(ztest_ds_t *zd, ztest_od_t *od, int count) int error; int i; - ASSERT(_mutex_held(&zd->zd_dirobj_lock)); + ASSERT(mutex_held(&zd->zd_dirobj_lock)); for (i = 0; i < count; i++, od++) { od->od_object = 0; @@ -1795,7 +1991,7 @@ ztest_create(ztest_ds_t *zd, ztest_od_t *od, int count) int missing = 0; int i; - ASSERT(_mutex_held(&zd->zd_dirobj_lock)); + ASSERT(mutex_held(&zd->zd_dirobj_lock)); for (i = 0; i < count; i++, od++) { if (missing) { @@ -1841,7 +2037,7 @@ ztest_remove(ztest_ds_t *zd, ztest_od_t *od, int count) int error; int i; - ASSERT(_mutex_held(&zd->zd_dirobj_lock)); + ASSERT(mutex_held(&zd->zd_dirobj_lock)); od += count - 1; @@ -1983,6 +2179,8 @@ ztest_io(ztest_ds_t *zd, uint64_t object, uint64_t offset) if (ztest_random(2) == 0) io_type = ZTEST_IO_WRITE_TAG; + (void) rw_enter(&zd->zd_zilog_lock, RW_READER); + switch (io_type) { case ZTEST_IO_WRITE_TAG: @@ -2020,6 +2218,8 @@ ztest_io(ztest_ds_t *zd, uint64_t object, uint64_t offset) break; } + (void) rw_exit(&zd->zd_zilog_lock); + umem_free(data, blocksize); } @@ -2057,13 +2257,13 @@ ztest_object_init(ztest_ds_t *zd, ztest_od_t *od, size_t size, boolean_t remove) int count = size / sizeof (*od); int rv = 0; - VERIFY(mutex_lock(&zd->zd_dirobj_lock) == 0); + mutex_enter(&zd->zd_dirobj_lock); if ((ztest_lookup(zd, od, count) != 0 || remove) && (ztest_remove(zd, od, count) != 0 || ztest_create(zd, od, count) != 0)) rv = -1; zd->zd_od = od; - VERIFY(mutex_unlock(&zd->zd_dirobj_lock) == 0); + mutex_exit(&zd->zd_dirobj_lock); return (rv); } @@ -2074,6 +2274,8 @@ ztest_zil_commit(ztest_ds_t *zd, uint64_t id) { zilog_t *zilog = zd->zd_zilog; + (void) rw_enter(&zd->zd_zilog_lock, RW_READER); + zil_commit(zilog, ztest_random(ZTEST_OBJECTS)); /* @@ -2082,9 +2284,37 @@ ztest_zil_commit(ztest_ds_t *zd, uint64_t id) * will verify that the log really does contain this record. */ mutex_enter(&zilog->zl_lock); - ASSERT(zd->zd_seq <= zilog->zl_commit_lr_seq); - zd->zd_seq = zilog->zl_commit_lr_seq; + ASSERT(zd->zd_shared != NULL); + ASSERT3U(zd->zd_shared->zd_seq, <=, zilog->zl_commit_lr_seq); + zd->zd_shared->zd_seq = zilog->zl_commit_lr_seq; mutex_exit(&zilog->zl_lock); + + (void) rw_exit(&zd->zd_zilog_lock); +} + +/* + * This function is designed to simulate the operations that occur during a + * mount/unmount operation. We hold the dataset across these operations in an + * attempt to expose any implicit assumptions about ZIL management. + */ +/* ARGSUSED */ +void +ztest_zil_remount(ztest_ds_t *zd, uint64_t id) +{ + objset_t *os = zd->zd_os; + + mutex_enter(&zd->zd_dirobj_lock); + (void) rw_enter(&zd->zd_zilog_lock, RW_WRITER); + + /* zfs_sb_teardown() */ + zil_close(zd->zd_zilog); + + /* zfsvfs_setup() */ + VERIFY(zil_open(os, ztest_get_data) == zd->zd_zilog); + zil_replay(os, zd, ztest_replay_vector); + + (void) rw_exit(&zd->zd_zilog_lock); + mutex_exit(&zd->zd_dirobj_lock); } /* @@ -2095,14 +2325,14 @@ ztest_zil_commit(ztest_ds_t *zd, uint64_t id) void ztest_spa_create_destroy(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; + ztest_shared_opts_t *zo = &ztest_opts; spa_t *spa; nvlist_t *nvroot; /* * Attempt to create using a bad file. */ - nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); + nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, 0, 0, 0, 1); VERIFY3U(ENOENT, ==, spa_create("ztest_bad_file", nvroot, NULL, NULL, NULL)); nvlist_free(nvroot); @@ -2110,7 +2340,7 @@ ztest_spa_create_destroy(ztest_ds_t *zd, uint64_t id) /* * Attempt to create using a bad mirror. */ - nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 2, 1); + nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, 0, 0, 2, 1); VERIFY3U(ENOENT, ==, spa_create("ztest_bad_mirror", nvroot, NULL, NULL, NULL)); nvlist_free(nvroot); @@ -2119,15 +2349,87 @@ ztest_spa_create_destroy(ztest_ds_t *zd, uint64_t id) * Attempt to create an existing pool. It shouldn't matter * what's in the nvroot; we should fail with EEXIST. */ - (void) rw_rdlock(&zs->zs_name_lock); - nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); - VERIFY3U(EEXIST, ==, spa_create(zs->zs_pool, nvroot, NULL, NULL, NULL)); + (void) rw_enter(&ztest_name_lock, RW_READER); + nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, 0, 0, 0, 1); + VERIFY3U(EEXIST, ==, spa_create(zo->zo_pool, nvroot, NULL, NULL, NULL)); nvlist_free(nvroot); - VERIFY3U(0, ==, spa_open(zs->zs_pool, &spa, FTAG)); - VERIFY3U(EBUSY, ==, spa_destroy(zs->zs_pool)); + VERIFY3U(0, ==, spa_open(zo->zo_pool, &spa, FTAG)); + VERIFY3U(EBUSY, ==, spa_destroy(zo->zo_pool)); spa_close(spa, FTAG); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); +} + +/* ARGSUSED */ +void +ztest_spa_upgrade(ztest_ds_t *zd, uint64_t id) +{ + spa_t *spa; + uint64_t initial_version = SPA_VERSION_INITIAL; + uint64_t version, newversion; + nvlist_t *nvroot, *props; + char *name; + + mutex_enter(&ztest_vdev_lock); + name = kmem_asprintf("%s_upgrade", ztest_opts.zo_pool); + + /* + * Clean up from previous runs. + */ + (void) spa_destroy(name); + + nvroot = make_vdev_root(NULL, NULL, name, ztest_opts.zo_vdev_size, 0, + 0, ztest_opts.zo_raidz, ztest_opts.zo_mirrors, 1); + + /* + * If we're configuring a RAIDZ device then make sure that the + * the initial version is capable of supporting that feature. + */ + switch (ztest_opts.zo_raidz_parity) { + case 0: + case 1: + initial_version = SPA_VERSION_INITIAL; + break; + case 2: + initial_version = SPA_VERSION_RAIDZ2; + break; + case 3: + initial_version = SPA_VERSION_RAIDZ3; + break; + } + + /* + * Create a pool with a spa version that can be upgraded. Pick + * a value between initial_version and SPA_VERSION_BEFORE_FEATURES. + */ + do { + version = ztest_random_spa_version(initial_version); + } while (version > SPA_VERSION_BEFORE_FEATURES); + + props = fnvlist_alloc(); + fnvlist_add_uint64(props, + zpool_prop_to_name(ZPOOL_PROP_VERSION), version); + VERIFY3S(spa_create(name, nvroot, props, NULL, NULL), ==, 0); + fnvlist_free(nvroot); + fnvlist_free(props); + + VERIFY3S(spa_open(name, &spa, FTAG), ==, 0); + VERIFY3U(spa_version(spa), ==, version); + newversion = ztest_random_spa_version(version + 1); + + if (ztest_opts.zo_verbose >= 4) { + (void) printf("upgrading spa version from %llu to %llu\n", + (u_longlong_t)version, (u_longlong_t)newversion); + } + + spa_upgrade(spa, newversion); + VERIFY3U(spa_version(spa), >, version); + VERIFY3U(spa_version(spa), ==, fnvlist_lookup_uint64(spa->spa_config, + zpool_prop_to_name(ZPOOL_PROP_VERSION))); + spa_close(spa, FTAG); + + strfree(name); + mutex_exit(&ztest_vdev_lock); } static vdev_t * @@ -2175,14 +2477,15 @@ void ztest_vdev_add_remove(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; uint64_t leaves; uint64_t guid; nvlist_t *nvroot; int error; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); - leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) * zopt_raidz; + mutex_enter(&ztest_vdev_lock); + leaves = + MAX(zs->zs_mirrors + zs->zs_splits, 1) * ztest_opts.zo_raidz; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); @@ -2207,9 +2510,9 @@ ztest_vdev_add_remove(ztest_ds_t *zd, uint64_t id) * dmu_objset_destroy() to fail with EBUSY thus * leaving the dataset in an inconsistent state. */ - VERIFY(rw_wrlock(&ztest_shared->zs_name_lock) == 0); + rw_enter(&ztest_name_lock, RW_WRITER); error = spa_vdev_remove(spa, guid, B_FALSE); - VERIFY(rw_unlock(&ztest_shared->zs_name_lock) == 0); + rw_exit(&ztest_name_lock); if (error && error != EEXIST) fatal(0, "spa_vdev_remove() = %d", error); @@ -2219,8 +2522,10 @@ ztest_vdev_add_remove(ztest_ds_t *zd, uint64_t id) /* * Make 1/4 of the devices be log devices. */ - nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, - ztest_random(4) == 0, zopt_raidz, zs->zs_mirrors, 1); + nvroot = make_vdev_root(NULL, NULL, NULL, + ztest_opts.zo_vdev_size, 0, + ztest_random(4) == 0, ztest_opts.zo_raidz, + zs->zs_mirrors, 1); error = spa_vdev_add(spa, nvroot); nvlist_free(nvroot); @@ -2231,7 +2536,7 @@ ztest_vdev_add_remove(ztest_ds_t *zd, uint64_t id) fatal(0, "spa_vdev_add() = %d", error); } - VERIFY(mutex_unlock(&ztest_shared->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); } /* @@ -2242,13 +2547,16 @@ void ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; vdev_t *rvd = spa->spa_root_vdev; spa_aux_vdev_t *sav; char *aux; + char *path; uint64_t guid = 0; int error; + path = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + if (ztest_random(2) == 0) { sav = &spa->spa_spares; aux = ZPOOL_CONFIG_SPARES; @@ -2257,7 +2565,7 @@ ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id) aux = ZPOOL_CONFIG_L2CACHE; } - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); + mutex_enter(&ztest_vdev_lock); spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); @@ -2272,10 +2580,10 @@ ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id) */ zs->zs_vdev_aux = 0; for (;;) { - char path[MAXPATHLEN]; int c; - (void) sprintf(path, ztest_aux_template, zopt_dir, - zopt_pool, aux, zs->zs_vdev_aux); + (void) snprintf(path, sizeof (path), ztest_aux_template, + ztest_opts.zo_dir, ztest_opts.zo_pool, aux, + zs->zs_vdev_aux); for (c = 0; c < sav->sav_count; c++) if (strcmp(sav->sav_vdevs[c]->vdev_path, path) == 0) @@ -2293,8 +2601,8 @@ ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id) /* * Add a new device. */ - nvlist_t *nvroot = make_vdev_root(NULL, aux, - (zopt_vdev_size * 5) / 4, 0, 0, 0, 0, 1); + nvlist_t *nvroot = make_vdev_root(NULL, aux, NULL, + (ztest_opts.zo_vdev_size * 5) / 4, 0, 0, 0, 0, 1); error = spa_vdev_add(spa, nvroot); if (error != 0) fatal(0, "spa_vdev_add(%p) = %d", nvroot, error); @@ -2313,7 +2621,9 @@ ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id) fatal(0, "spa_vdev_remove(%llu) = %d", guid, error); } - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); + + umem_free(path, MAXPATHLEN); } /* @@ -2324,17 +2634,17 @@ void ztest_split_pool(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; vdev_t *rvd = spa->spa_root_vdev; nvlist_t *tree, **child, *config, *split, **schild; uint_t c, children, schildren = 0, lastlogid = 0; int error = 0; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); + mutex_enter(&ztest_vdev_lock); /* ensure we have a useable config; mirrors of raidz aren't supported */ - if (zs->zs_mirrors < 3 || zopt_raidz > 1) { - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + if (zs->zs_mirrors < 3 || ztest_opts.zo_raidz > 1) { + mutex_exit(&ztest_vdev_lock); return; } @@ -2393,9 +2703,9 @@ ztest_split_pool(ztest_ds_t *zd, uint64_t id) spa_config_exit(spa, SCL_VDEV, FTAG); - (void) rw_wrlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_WRITER); error = spa_vdev_split_mirror(spa, "splitp", config, NULL, B_FALSE); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); nvlist_free(config); @@ -2408,7 +2718,7 @@ ztest_split_pool(ztest_ds_t *zd, uint64_t id) ++zs->zs_splits; --zs->zs_mirrors; } - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); } @@ -2420,7 +2730,7 @@ void ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; spa_aux_vdev_t *sav = &spa->spa_spares; vdev_t *rvd = spa->spa_root_vdev; vdev_t *oldvd, *newvd, *pvd; @@ -2430,15 +2740,18 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) uint64_t ashift = ztest_get_ashift(); uint64_t oldguid, pguid; size_t oldsize, newsize; - char oldpath[MAXPATHLEN], newpath[MAXPATHLEN]; + char *oldpath, *newpath; int replacing; int oldvd_has_siblings = B_FALSE; int newvd_is_spare = B_FALSE; int oldvd_is_log; int error, expected_error; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); - leaves = MAX(zs->zs_mirrors, 1) * zopt_raidz; + oldpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + newpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + + mutex_enter(&ztest_vdev_lock); + leaves = MAX(zs->zs_mirrors, 1) * ztest_opts.zo_raidz; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); @@ -2464,12 +2777,12 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) if (zs->zs_mirrors >= 1) { ASSERT(oldvd->vdev_ops == &vdev_mirror_ops); ASSERT(oldvd->vdev_children >= zs->zs_mirrors); - oldvd = oldvd->vdev_child[leaf / zopt_raidz]; + oldvd = oldvd->vdev_child[leaf / ztest_opts.zo_raidz]; } - if (zopt_raidz > 1) { + if (ztest_opts.zo_raidz > 1) { ASSERT(oldvd->vdev_ops == &vdev_raidz_ops); - ASSERT(oldvd->vdev_children == zopt_raidz); - oldvd = oldvd->vdev_child[leaf % zopt_raidz]; + ASSERT(oldvd->vdev_children == ztest_opts.zo_raidz); + oldvd = oldvd->vdev_child[leaf % ztest_opts.zo_raidz]; } /* @@ -2498,8 +2811,7 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) if (error != 0 && error != ENODEV && error != EBUSY && error != ENOTSUP) fatal(0, "detach (%s) returned %d", oldpath, error); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); - return; + goto out; } /* @@ -2511,8 +2823,9 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) newvd_is_spare = B_TRUE; (void) strcpy(newpath, newvd->vdev_path); } else { - (void) snprintf(newpath, sizeof (newpath), ztest_dev_template, - zopt_dir, zopt_pool, top * leaves + leaf); + (void) snprintf(newpath, MAXPATHLEN, ztest_dev_template, + ztest_opts.zo_dir, ztest_opts.zo_pool, + top * leaves + leaf); if (ztest_random(2) == 0) newpath[strlen(newpath) - 1] = 'b'; newvd = vdev_lookup_by_path(rvd, newpath); @@ -2561,7 +2874,7 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) /* * Build the nvlist describing newpath. */ - root = make_vdev_root(newpath, NULL, newvd == NULL ? newsize : 0, + root = make_vdev_root(newpath, NULL, NULL, newvd == NULL ? newsize : 0, ashift, 0, 0, 0, 1); error = spa_vdev_attach(spa, oldguid, root, replacing); @@ -2590,8 +2903,11 @@ ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id) oldpath, (longlong_t)oldsize, newpath, (longlong_t)newsize, replacing, error, expected_error); } +out: + mutex_exit(&ztest_vdev_lock); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + umem_free(oldpath, MAXPATHLEN); + umem_free(newpath, MAXPATHLEN); } /* @@ -2614,7 +2930,7 @@ grow_vdev(vdev_t *vd, void *arg) fsize = lseek(fd, 0, SEEK_END); VERIFY(ftruncate(fd, *newsize) == 0); - if (zopt_verbose >= 6) { + if (ztest_opts.zo_verbose >= 6) { (void) printf("%s grew from %lu to %lu bytes\n", vd->vdev_path, (ulong_t)fsize, (ulong_t)*newsize); } @@ -2650,7 +2966,7 @@ online_vdev(vdev_t *vd, void *arg) * vdev_open fails is by checking the returned newstate. */ if (error || newstate != VDEV_STATE_HEALTHY) { - if (zopt_verbose >= 5) { + if (ztest_opts.zo_verbose >= 5) { (void) printf("Unable to expand vdev, state %llu, " "error %d\n", (u_longlong_t)newstate, error); } @@ -2665,7 +2981,7 @@ online_vdev(vdev_t *vd, void *arg) * trying to online it. */ if (generation != spa->spa_config_generation) { - if (zopt_verbose >= 5) { + if (ztest_opts.zo_verbose >= 5) { (void) printf("vdev configuration has changed, " "guid %llu, state %llu, expected gen %llu, " "got gen %llu\n", @@ -2713,8 +3029,7 @@ vdev_walk_tree(vdev_t *vd, vdev_t *(*func)(vdev_t *, void *), void *arg) void ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; vdev_t *vd, *tvd; metaslab_class_t *mc; metaslab_group_t *mg; @@ -2722,7 +3037,7 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) uint64_t top; uint64_t old_class_space, new_class_space, old_ms_count, new_ms_count; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); + mutex_enter(&ztest_vdev_lock); spa_config_enter(spa, SCL_STATE, spa, RW_READER); top = ztest_random_vdev_top(spa, B_TRUE); @@ -2748,16 +3063,16 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) * original size, and it has a valid psize. */ if (tvd->vdev_state != VDEV_STATE_HEALTHY || - psize == 0 || psize >= 4 * zopt_vdev_size) { + psize == 0 || psize >= 4 * ztest_opts.zo_vdev_size) { spa_config_exit(spa, SCL_STATE, spa); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); return; } ASSERT(psize > 0); newsize = psize + psize / 8; ASSERT3U(newsize, >, psize); - if (zopt_verbose >= 6) { + if (ztest_opts.zo_verbose >= 6) { (void) printf("Expanding LUN %s from %lu to %lu\n", vd->vdev_path, (ulong_t)psize, (ulong_t)newsize); } @@ -2770,12 +3085,12 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) if (vdev_walk_tree(tvd, grow_vdev, &newsize) != NULL || vdev_walk_tree(tvd, online_vdev, NULL) != NULL || tvd->vdev_state != VDEV_STATE_HEALTHY) { - if (zopt_verbose >= 5) { + if (ztest_opts.zo_verbose >= 5) { (void) printf("Could not expand LUN because " "the vdev configuration changed.\n"); } spa_config_exit(spa, SCL_STATE, spa); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); return; } @@ -2804,12 +3119,12 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) new_class_space = metaslab_class_get_space(mc); if (tvd->vdev_mg != mg || mg->mg_class != mc) { - if (zopt_verbose >= 5) { + if (ztest_opts.zo_verbose >= 5) { (void) printf("Could not verify LUN expansion due to " "intervening vdev offline or remove.\n"); } spa_config_exit(spa, SCL_STATE, spa); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); return; } @@ -2827,7 +3142,7 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) fatal(0, "LUN expansion failed: class_space %llu <= %llu\n", old_class_space, new_class_space); - if (zopt_verbose >= 5) { + if (ztest_opts.zo_verbose >= 5) { char oldnumbuf[6], newnumbuf[6]; nicenum(old_class_space, oldnumbuf); @@ -2837,7 +3152,7 @@ ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id) } spa_config_exit(spa, SCL_STATE, spa); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); } /* @@ -2864,7 +3179,8 @@ ztest_dataset_create(char *dsname) if (err || zilset < 80) return (err); - (void) printf("Setting dataset %s to sync always\n", dsname); + if (ztest_opts.zo_verbose >= 5) + (void) printf("Setting dataset %s to sync always\n", dsname); return (ztest_dsl_prop_set_uint64(dsname, ZFS_PROP_SYNC, ZFS_SYNC_ALWAYS, B_FALSE)); } @@ -2884,7 +3200,7 @@ ztest_objset_destroy_cb(const char *name, void *arg) error = dmu_object_info(os, ZTEST_DIROBJ, &doi); if (error != ENOENT) { /* We could have crashed in the middle of destroying it */ - ASSERT3U(error, ==, 0); + ASSERT0(error); ASSERT3U(doi.doi_type, ==, DMU_OT_ZAP_OTHER); ASSERT3S(doi.doi_physical_blocks_512, >=, 0); } @@ -2936,19 +3252,21 @@ ztest_snapshot_destroy(char *osname, uint64_t id) void ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; - ztest_ds_t zdtmp; + ztest_ds_t *zdtmp; int iters; int error; objset_t *os, *os2; - char name[MAXNAMELEN]; + char *name; zilog_t *zilog; int i; - (void) rw_rdlock(&zs->zs_name_lock); + zdtmp = umem_alloc(sizeof (ztest_ds_t), UMEM_NOFAIL); + name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + + (void) rw_enter(&ztest_name_lock, RW_READER); (void) snprintf(name, MAXNAMELEN, "%s/temp_%llu", - zs->zs_pool, (u_longlong_t)id); + ztest_opts.zo_pool, (u_longlong_t)id); /* * If this dataset exists from a previous run, process its replay log @@ -2957,9 +3275,9 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) */ if (ztest_random(2) == 0 && dmu_objset_own(name, DMU_OST_OTHER, B_FALSE, FTAG, &os) == 0) { - ztest_zd_init(&zdtmp, os); - zil_replay(os, &zdtmp, ztest_replay_vector); - ztest_zd_fini(&zdtmp); + ztest_zd_init(zdtmp, NULL, os); + zil_replay(os, zdtmp, ztest_replay_vector); + ztest_zd_fini(zdtmp); dmu_objset_disown(os, FTAG); } @@ -2983,8 +3301,7 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) if (error) { if (error == ENOSPC) { ztest_record_enospc(FTAG); - (void) rw_unlock(&zs->zs_name_lock); - return; + goto out; } fatal(0, "dmu_objset_create(%s) = %d", name, error); } @@ -2992,7 +3309,7 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) VERIFY3U(0, ==, dmu_objset_own(name, DMU_OST_OTHER, B_FALSE, FTAG, &os)); - ztest_zd_init(&zdtmp, os); + ztest_zd_init(zdtmp, NULL, os); /* * Open the intent log for it. @@ -3005,7 +3322,7 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) */ iters = ztest_random(5); for (i = 0; i < iters; i++) { - ztest_dmu_object_alloc_free(&zdtmp, id); + ztest_dmu_object_alloc_free(zdtmp, id); if (ztest_random(iters) == 0) (void) ztest_snapshot_create(name, i); } @@ -3030,9 +3347,12 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) zil_close(zilog); dmu_objset_disown(os, FTAG); - ztest_zd_fini(&zdtmp); + ztest_zd_fini(zdtmp); +out: + (void) rw_exit(&ztest_name_lock); - (void) rw_unlock(&zs->zs_name_lock); + umem_free(name, MAXNAMELEN); + umem_free(zdtmp, sizeof (ztest_ds_t)); } /* @@ -3041,12 +3361,10 @@ ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id) void ztest_dmu_snapshot_create_destroy(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; - - (void) rw_rdlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_READER); (void) ztest_snapshot_destroy(zd->zd_name, id); (void) ztest_snapshot_create(zd->zd_name, id); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } /* @@ -3055,13 +3373,19 @@ ztest_dmu_snapshot_create_destroy(ztest_ds_t *zd, uint64_t id) void ztest_dsl_dataset_cleanup(char *osname, uint64_t id) { - char snap1name[MAXNAMELEN]; - char clone1name[MAXNAMELEN]; - char snap2name[MAXNAMELEN]; - char clone2name[MAXNAMELEN]; - char snap3name[MAXNAMELEN]; + char *snap1name; + char *clone1name; + char *snap2name; + char *clone2name; + char *snap3name; int error; + snap1name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + clone1name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + snap2name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + clone2name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + snap3name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + (void) snprintf(snap1name, MAXNAMELEN, "%s@s1_%llu", osname, (u_longlong_t)id); (void) snprintf(clone1name, MAXNAMELEN, "%s/c1_%llu", @@ -3088,6 +3412,12 @@ ztest_dsl_dataset_cleanup(char *osname, uint64_t id) error = dmu_objset_destroy(snap1name, B_FALSE); if (error && error != ENOENT) fatal(0, "dmu_objset_destroy(%s) = %d", snap1name, error); + + umem_free(snap1name, MAXNAMELEN); + umem_free(clone1name, MAXNAMELEN); + umem_free(snap2name, MAXNAMELEN); + umem_free(clone2name, MAXNAMELEN); + umem_free(snap3name, MAXNAMELEN); } /* @@ -3096,18 +3426,23 @@ ztest_dsl_dataset_cleanup(char *osname, uint64_t id) void ztest_dsl_dataset_promote_busy(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; objset_t *clone; dsl_dataset_t *ds; - char snap1name[MAXNAMELEN]; - char clone1name[MAXNAMELEN]; - char snap2name[MAXNAMELEN]; - char clone2name[MAXNAMELEN]; - char snap3name[MAXNAMELEN]; + char *snap1name; + char *clone1name; + char *snap2name; + char *clone2name; + char *snap3name; char *osname = zd->zd_name; int error; - (void) rw_rdlock(&zs->zs_name_lock); + snap1name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + clone1name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + snap2name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + clone2name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + snap3name = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + + (void) rw_enter(&ztest_name_lock, RW_READER); ztest_dsl_dataset_cleanup(osname, id); @@ -3192,42 +3527,65 @@ ztest_dsl_dataset_promote_busy(ztest_ds_t *zd, uint64_t id) out: ztest_dsl_dataset_cleanup(osname, id); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); + + umem_free(snap1name, MAXNAMELEN); + umem_free(clone1name, MAXNAMELEN); + umem_free(snap2name, MAXNAMELEN); + umem_free(clone2name, MAXNAMELEN); + umem_free(snap3name, MAXNAMELEN); } +#undef OD_ARRAY_SIZE +#define OD_ARRAY_SIZE 4 + /* * Verify that dmu_object_{alloc,free} work as expected. */ void ztest_dmu_object_alloc_free(ztest_ds_t *zd, uint64_t id) { - ztest_od_t od[4]; - int batchsize = sizeof (od) / sizeof (od[0]); + ztest_od_t *od; + int batchsize; + int size; int b; + size = sizeof(ztest_od_t) * OD_ARRAY_SIZE; + od = umem_alloc(size, UMEM_NOFAIL); + batchsize = OD_ARRAY_SIZE; + for (b = 0; b < batchsize; b++) - ztest_od_init(&od[b], id, FTAG, b, DMU_OT_UINT64_OTHER, 0, 0); + ztest_od_init(od + b, id, FTAG, b, DMU_OT_UINT64_OTHER, 0, 0); /* * Destroy the previous batch of objects, create a new batch, * and do some I/O on the new objects. */ - if (ztest_object_init(zd, od, sizeof (od), B_TRUE) != 0) + if (ztest_object_init(zd, od, size, B_TRUE) != 0) return; while (ztest_random(4 * batchsize) != 0) ztest_io(zd, od[ztest_random(batchsize)].od_object, ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT); + + umem_free(od, size); } +#undef OD_ARRAY_SIZE +#define OD_ARRAY_SIZE 2 + /* * Verify that dmu_{read,write} work as expected. */ void ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) { + int size; + ztest_od_t *od; + objset_t *os = zd->zd_os; - ztest_od_t od[2]; + size = sizeof(ztest_od_t) * OD_ARRAY_SIZE; + od = umem_alloc(size, UMEM_NOFAIL); dmu_tx_t *tx; int i, freeit, error; uint64_t n, s, txg; @@ -3266,11 +3624,13 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) /* * Read the directory info. If it's the first time, set things up. */ - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, chunksize); - ztest_od_init(&od[1], id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, chunksize); + ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, chunksize); + ztest_od_init(od + 1, id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, chunksize); - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, size, B_FALSE) != 0) { + umem_free(od, size); return; + } bigobj = od[0].od_object; packobj = od[1].od_object; @@ -3313,10 +3673,10 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) */ error = dmu_read(os, packobj, packoff, packsize, packbuf, DMU_READ_PREFETCH); - ASSERT3U(error, ==, 0); + ASSERT0(error); error = dmu_read(os, bigobj, bigoff, bigsize, bigbuf, DMU_READ_PREFETCH); - ASSERT3U(error, ==, 0); + ASSERT0(error); /* * Get a tx for the mods to both packobj and bigobj. @@ -3334,6 +3694,7 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) if (txg == 0) { umem_free(packbuf, packsize); umem_free(bigbuf, bigsize); + umem_free(od, size); return; } @@ -3392,7 +3753,7 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) dmu_write(os, packobj, packoff, packsize, packbuf, tx); if (freeit) { - if (zopt_verbose >= 7) { + if (ztest_opts.zo_verbose >= 7) { (void) printf("freeing offset %llx size %llx" " txg %llx\n", (u_longlong_t)bigoff, @@ -3401,7 +3762,7 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) } VERIFY(0 == dmu_free_range(os, bigobj, bigoff, bigsize, tx)); } else { - if (zopt_verbose >= 7) { + if (ztest_opts.zo_verbose >= 7) { (void) printf("writing offset %llx size %llx" " txg %llx\n", (u_longlong_t)bigoff, @@ -3434,6 +3795,7 @@ ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id) umem_free(packbuf, packsize); umem_free(bigbuf, bigsize); + umem_free(od, size); } void @@ -3485,14 +3847,18 @@ compare_and_update_pbbufs(uint64_t s, bufwad_t *packbuf, bufwad_t *bigbuf, } } +#undef OD_ARRAY_SIZE +#define OD_ARRAY_SIZE 2 + void ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) { objset_t *os = zd->zd_os; - ztest_od_t od[2]; + ztest_od_t *od; dmu_tx_t *tx; uint64_t i; int error; + int size; uint64_t n, s, txg; bufwad_t *packbuf, *bigbuf; uint64_t packobj, packoff, packsize, bigobj, bigoff, bigsize; @@ -3505,6 +3871,9 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) arc_buf_t **bigbuf_arcbufs; dmu_object_info_t doi; + size = sizeof(ztest_od_t) * OD_ARRAY_SIZE; + od = umem_alloc(size, UMEM_NOFAIL); + /* * This test uses two objects, packobj and bigobj, that are always * updated together (i.e. in the same tx) so that their contents are @@ -3524,11 +3893,14 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) /* * Read the directory info. If it's the first time, set things up. */ - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); - ztest_od_init(&od[1], id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, chunksize); + ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); + ztest_od_init(od + 1, id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, chunksize); + - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, size, B_FALSE) != 0) { + umem_free(od, size); return; + } bigobj = od[0].od_object; packobj = od[1].od_object; @@ -3614,6 +3986,7 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) } } umem_free(bigbuf_arcbufs, 2 * s * sizeof (arc_buf_t *)); + umem_free(od, size); dmu_buf_rele(bonus_db, FTAG); return; } @@ -3626,10 +3999,10 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) if (i != 0 || ztest_random(2) != 0) { error = dmu_read(os, packobj, packoff, packsize, packbuf, DMU_READ_PREFETCH); - ASSERT3U(error, ==, 0); + ASSERT0(error); error = dmu_read(os, bigobj, bigoff, bigsize, bigbuf, DMU_READ_PREFETCH); - ASSERT3U(error, ==, 0); + ASSERT0(error); } compare_and_update_pbbufs(s, packbuf, bigbuf, bigsize, n, chunksize, txg); @@ -3639,7 +4012,7 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) * Now write them out. */ dmu_write(os, packobj, packoff, packsize, packbuf, tx); - if (zopt_verbose >= 7) { + if (ztest_opts.zo_verbose >= 7) { (void) printf("writing offset %llx size %llx" " txg %llx\n", (u_longlong_t)bigoff, @@ -3710,13 +4083,16 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id) umem_free(packbuf, packsize); umem_free(bigbuf, bigsize); umem_free(bigbuf_arcbufs, 2 * s * sizeof (arc_buf_t *)); + umem_free(od, size); } /* ARGSUSED */ void ztest_dmu_write_parallel(ztest_ds_t *zd, uint64_t id) { - ztest_od_t od[1]; + ztest_od_t *od; + + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); uint64_t offset = (1ULL << (ztest_random(20) + 43)) + (ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT); @@ -3725,47 +4101,56 @@ ztest_dmu_write_parallel(ztest_ds_t *zd, uint64_t id) * to verify that parallel writes to an object -- even to the * same blocks within the object -- doesn't cause any trouble. */ - ztest_od_init(&od[0], ID_PARALLEL, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0); + ztest_od_init(od, ID_PARALLEL, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0); - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) return; while (ztest_random(10) != 0) - ztest_io(zd, od[0].od_object, offset); + ztest_io(zd, od->od_object, offset); + + umem_free(od, sizeof(ztest_od_t)); } void ztest_dmu_prealloc(ztest_ds_t *zd, uint64_t id) { - ztest_od_t od[1]; + ztest_od_t *od; uint64_t offset = (1ULL << (ztest_random(4) + SPA_MAXBLOCKSHIFT)) + (ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT); uint64_t count = ztest_random(20) + 1; uint64_t blocksize = ztest_random_blocksize(); void *data; - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); - if (ztest_object_init(zd, od, sizeof (od), !ztest_random(2)) != 0) + ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); + + if (ztest_object_init(zd, od, sizeof (ztest_od_t), !ztest_random(2)) != 0) { + umem_free(od, sizeof(ztest_od_t)); return; + } - if (ztest_truncate(zd, od[0].od_object, offset, count * blocksize) != 0) + if (ztest_truncate(zd, od->od_object, offset, count * blocksize) != 0) { + umem_free(od, sizeof(ztest_od_t)); return; + } - ztest_prealloc(zd, od[0].od_object, offset, count * blocksize); + ztest_prealloc(zd, od->od_object, offset, count * blocksize); data = umem_zalloc(blocksize, UMEM_NOFAIL); while (ztest_random(count) != 0) { uint64_t randoff = offset + (ztest_random(count) * blocksize); - if (ztest_write(zd, od[0].od_object, randoff, blocksize, + if (ztest_write(zd, od->od_object, randoff, blocksize, data) != 0) break; while (ztest_random(4) != 0) - ztest_io(zd, od[0].od_object, randoff); + ztest_io(zd, od->od_object, randoff); } umem_free(data, blocksize); + umem_free(od, sizeof(ztest_od_t)); } /* @@ -3779,7 +4164,7 @@ void ztest_zap(ztest_ds_t *zd, uint64_t id) { objset_t *os = zd->zd_os; - ztest_od_t od[1]; + ztest_od_t *od; uint64_t object; uint64_t txg, last_txg; uint64_t value[ZTEST_ZAP_MAX_INTS]; @@ -3790,12 +4175,14 @@ ztest_zap(ztest_ds_t *zd, uint64_t id) int error; char *hc[2] = { "s.acl.h", ".s.open.h.hyLZlg" }; - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); + ztest_od_init(od, id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0); - if (ztest_object_init(zd, od, sizeof (od), !ztest_random(2)) != 0) - return; + if (ztest_object_init(zd, od, sizeof (ztest_od_t), + !ztest_random(2)) != 0) + goto out; - object = od[0].od_object; + object = od->od_object; /* * Generate a known hash collision, and verify that @@ -3805,7 +4192,7 @@ ztest_zap(ztest_ds_t *zd, uint64_t id) dmu_tx_hold_zap(tx, object, B_TRUE, NULL); txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG); if (txg == 0) - return; + goto out; for (i = 0; i < 2; i++) { value[i] = i; VERIFY3U(0, ==, zap_add(os, object, hc[i], sizeof (uint64_t), @@ -3873,7 +4260,7 @@ ztest_zap(ztest_ds_t *zd, uint64_t id) dmu_tx_hold_zap(tx, object, B_TRUE, NULL); txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG); if (txg == 0) - return; + goto out; if (last_txg > txg) fatal(0, "zap future leak: old %llu new %llu", last_txg, txg); @@ -3898,18 +4285,20 @@ ztest_zap(ztest_ds_t *zd, uint64_t id) error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); if (error == ENOENT) - return; + goto out; - ASSERT3U(error, ==, 0); + ASSERT0(error); tx = dmu_tx_create(os); dmu_tx_hold_zap(tx, object, B_TRUE, NULL); txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG); if (txg == 0) - return; + goto out; VERIFY3U(0, ==, zap_remove(os, object, txgname, tx)); VERIFY3U(0, ==, zap_remove(os, object, propname, tx)); dmu_tx_commit(tx); +out: + umem_free(od, sizeof(ztest_od_t)); } /* @@ -3919,16 +4308,17 @@ void ztest_fzap(ztest_ds_t *zd, uint64_t id) { objset_t *os = zd->zd_os; - ztest_od_t od[1]; + ztest_od_t *od; uint64_t object, txg; int i; - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); + ztest_od_init(od, id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0); - if (ztest_object_init(zd, od, sizeof (od), !ztest_random(2)) != 0) - return; - - object = od[0].od_object; + if (ztest_object_init(zd, od, sizeof (ztest_od_t), + !ztest_random(2)) != 0) + goto out; + object = od->od_object; /* * Add entries to this ZAP and make sure it spills over @@ -3948,12 +4338,14 @@ ztest_fzap(ztest_ds_t *zd, uint64_t id) dmu_tx_hold_zap(tx, object, B_TRUE, name); txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG); if (txg == 0) - return; + goto out; error = zap_add(os, object, name, sizeof (uint64_t), 1, &value, tx); ASSERT(error == 0 || error == EEXIST); dmu_tx_commit(tx); } +out: + umem_free(od, sizeof(ztest_od_t)); } /* ARGSUSED */ @@ -3961,7 +4353,7 @@ void ztest_zap_parallel(ztest_ds_t *zd, uint64_t id) { objset_t *os = zd->zd_os; - ztest_od_t od[1]; + ztest_od_t *od; uint64_t txg, object, count, wsize, wc, zl_wsize, zl_wc; dmu_tx_t *tx; int i, namelen, error; @@ -3969,12 +4361,15 @@ ztest_zap_parallel(ztest_ds_t *zd, uint64_t id) char name[20], string_value[20]; void *data; - ztest_od_init(&od[0], ID_PARALLEL, FTAG, micro, DMU_OT_ZAP_OTHER, 0, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); + ztest_od_init(od, ID_PARALLEL, FTAG, micro, DMU_OT_ZAP_OTHER, 0, 0); - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) { + umem_free(od, sizeof(ztest_od_t)); return; + } - object = od[0].od_object; + object = od->od_object; /* * Generate a random name of the form 'xxx.....' where each @@ -4063,6 +4458,8 @@ ztest_zap_parallel(ztest_ds_t *zd, uint64_t id) if (tx != NULL) dmu_tx_commit(tx); + + umem_free(od, sizeof(ztest_od_t)); } /* @@ -4097,7 +4494,7 @@ ztest_commit_callback(void *arg, int error) data->zcd_called = B_TRUE; if (error == ECANCELED) { - ASSERT3U(data->zcd_txg, ==, 0); + ASSERT0(data->zcd_txg); ASSERT(!data->zcd_added); /* @@ -4111,7 +4508,7 @@ ztest_commit_callback(void *arg, int error) ASSERT(data->zcd_added); ASSERT3U(data->zcd_txg, !=, 0); - (void) mutex_lock(&zcl.zcl_callbacks_lock); + (void) mutex_enter(&zcl.zcl_callbacks_lock); /* See if this cb was called more quickly */ if ((synced_txg - data->zcd_txg) < zc_min_txg_delay) @@ -4120,7 +4517,7 @@ ztest_commit_callback(void *arg, int error) /* Remove our callback from the list */ list_remove(&zcl.zcl_callbacks, data); - (void) mutex_unlock(&zcl.zcl_callbacks_lock); + (void) mutex_exit(&zcl.zcl_callbacks_lock); umem_free(data, sizeof (ztest_cb_data_t)); } @@ -4135,6 +4532,7 @@ ztest_create_cb_data(objset_t *os, uint64_t txg) cb_data->zcd_txg = txg; cb_data->zcd_spa = dmu_objset_spa(os); + list_link_init(&cb_data->zcd_node); return (cb_data); } @@ -4146,23 +4544,26 @@ void ztest_dmu_commit_callbacks(ztest_ds_t *zd, uint64_t id) { objset_t *os = zd->zd_os; - ztest_od_t od[1]; + ztest_od_t *od; dmu_tx_t *tx; ztest_cb_data_t *cb_data[3], *tmp_cb; uint64_t old_txg, txg; int i, error = 0; - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); + ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0); - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) { + umem_free(od, sizeof(ztest_od_t)); return; + } tx = dmu_tx_create(os); cb_data[0] = ztest_create_cb_data(os, 0); dmu_tx_callback_register(tx, ztest_commit_callback, cb_data[0]); - dmu_tx_hold_write(tx, od[0].od_object, 0, sizeof (uint64_t)); + dmu_tx_hold_write(tx, od->od_object, 0, sizeof (uint64_t)); /* Every once in a while, abort the transaction on purpose */ if (ztest_random(100) == 0) @@ -4196,6 +4597,7 @@ ztest_dmu_commit_callbacks(ztest_ds_t *zd, uint64_t id) umem_free(cb_data[i], sizeof (ztest_cb_data_t)); } + umem_free(od, sizeof(ztest_od_t)); return; } @@ -4205,16 +4607,16 @@ ztest_dmu_commit_callbacks(ztest_ds_t *zd, uint64_t id) /* * Read existing data to make sure there isn't a future leak. */ - VERIFY(0 == dmu_read(os, od[0].od_object, 0, sizeof (uint64_t), + VERIFY(0 == dmu_read(os, od->od_object, 0, sizeof (uint64_t), &old_txg, DMU_READ_PREFETCH)); if (old_txg > txg) fatal(0, "future leak: got %" PRIu64 ", open txg is %" PRIu64, old_txg, txg); - dmu_write(os, od[0].od_object, 0, sizeof (uint64_t), &txg, tx); + dmu_write(os, od->od_object, 0, sizeof (uint64_t), &txg, tx); - (void) mutex_lock(&zcl.zcl_callbacks_lock); + (void) mutex_enter(&zcl.zcl_callbacks_lock); /* * Since commit callbacks don't have any ordering requirement and since @@ -4263,9 +4665,11 @@ ztest_dmu_commit_callbacks(ztest_ds_t *zd, uint64_t id) zc_cb_counter += 3; - (void) mutex_unlock(&zcl.zcl_callbacks_lock); + (void) mutex_exit(&zcl.zcl_callbacks_lock); dmu_tx_commit(tx); + + umem_free(od, sizeof(ztest_od_t)); } /* ARGSUSED */ @@ -4278,38 +4682,36 @@ ztest_dsl_prop_get_set(ztest_ds_t *zd, uint64_t id) ZFS_PROP_COPIES, ZFS_PROP_DEDUP }; - ztest_shared_t *zs = ztest_shared; int p; - (void) rw_rdlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_READER); for (p = 0; p < sizeof (proplist) / sizeof (proplist[0]); p++) (void) ztest_dsl_prop_set_uint64(zd->zd_name, proplist[p], ztest_random_dsl_prop(proplist[p]), (int)ztest_random(2)); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } /* ARGSUSED */ void ztest_spa_prop_get_set(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; nvlist_t *props = NULL; - (void) rw_rdlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_READER); - (void) ztest_spa_prop_set_uint64(zs, ZPOOL_PROP_DEDUPDITTO, + (void) ztest_spa_prop_set_uint64(ZPOOL_PROP_DEDUPDITTO, ZIO_DEDUPDITTO_MIN + ztest_random(ZIO_DEDUPDITTO_MIN)); - VERIFY3U(spa_prop_get(zs->zs_spa, &props), ==, 0); + VERIFY0(spa_prop_get(ztest_spa, &props)); - if (zopt_verbose >= 6) + if (ztest_opts.zo_verbose >= 6) dump_nvlist(props, 4); nvlist_free(props); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } /* @@ -4327,14 +4729,14 @@ ztest_dmu_snapshot_hold(ztest_ds_t *zd, uint64_t id) char tag[100]; char osname[MAXNAMELEN]; - (void) rw_rdlock(&ztest_shared->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_READER); dmu_objset_name(os, osname); - (void) snprintf(snapname, 100, "sh1_%llu", id); + (void) snprintf(snapname, 100, "sh1_%llu", (u_longlong_t)id); (void) snprintf(fullname, 100, "%s@%s", osname, snapname); - (void) snprintf(clonename, 100, "%s/ch1_%llu", osname, id); - (void) snprintf(tag, 100, "%tag_%llu", id); + (void) snprintf(clonename, 100, "%s/ch1_%llu",osname,(u_longlong_t)id); + (void) snprintf(tag, 100, "tag_%llu", (u_longlong_t)id); /* * Clean up from any previous run. @@ -4424,7 +4826,7 @@ ztest_dmu_snapshot_hold(ztest_ds_t *zd, uint64_t id) VERIFY(dmu_objset_hold(fullname, FTAG, &origin) == ENOENT); out: - (void) rw_unlock(&ztest_shared->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } /* @@ -4435,14 +4837,14 @@ void ztest_fault_inject(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; int fd; uint64_t offset; uint64_t leaves; uint64_t bad = 0x1990c0ffeedecadeull; uint64_t top, leaf; - char path0[MAXPATHLEN]; - char pathrand[MAXPATHLEN]; + char *path0; + char *pathrand; size_t fsize; int bshift = SPA_MAXBLOCKSHIFT + 2; /* don't scrog all labels */ int iters = 1000; @@ -4452,11 +4854,14 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) uint64_t guid0 = 0; boolean_t islog = B_FALSE; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); + path0 = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + pathrand = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + + mutex_enter(&ztest_vdev_lock); maxfaults = MAXFAULTS(); - leaves = MAX(zs->zs_mirrors, 1) * zopt_raidz; + leaves = MAX(zs->zs_mirrors, 1) * ztest_opts.zo_raidz; mirror_save = zs->zs_mirrors; - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); ASSERT(leaves >= 1); @@ -4478,10 +4883,12 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) * write failures and random online/offline activity on leaf 0, * and we'll write random garbage to the randomly chosen leaf. */ - (void) snprintf(path0, sizeof (path0), ztest_dev_template, - zopt_dir, zopt_pool, top * leaves + zs->zs_splits); - (void) snprintf(pathrand, sizeof (pathrand), ztest_dev_template, - zopt_dir, zopt_pool, top * leaves + leaf); + (void) snprintf(path0, MAXPATHLEN, ztest_dev_template, + ztest_opts.zo_dir, ztest_opts.zo_pool, + top * leaves + zs->zs_splits); + (void) snprintf(pathrand, MAXPATHLEN, ztest_dev_template, + ztest_opts.zo_dir, ztest_opts.zo_pool, + top * leaves + leaf); vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0); if (vd0 != NULL && vd0->vdev_top->vdev_islog) @@ -4518,7 +4925,7 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) if (sav->sav_count == 0) { spa_config_exit(spa, SCL_STATE, FTAG); - return; + goto out; } vd0 = sav->sav_vdevs[ztest_random(sav->sav_count)]; guid0 = vd0->vdev_guid; @@ -4550,19 +4957,31 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) * leaving the dataset in an inconsistent state. */ if (islog) - (void) rw_wrlock(&ztest_shared->zs_name_lock); + (void) rw_enter(&ztest_name_lock, + RW_WRITER); VERIFY(vdev_offline(spa, guid0, flags) != EBUSY); if (islog) - (void) rw_unlock(&ztest_shared->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } else { + /* + * Ideally we would like to be able to randomly + * call vdev_[on|off]line without holding locks + * to force unpredictable failures but the side + * effects of vdev_[on|off]line prevent us from + * doing so. We grab the ztest_vdev_lock here to + * prevent a race between injection testing and + * aux_vdev removal. + */ + mutex_enter(&ztest_vdev_lock); (void) vdev_online(spa, guid0, 0, NULL); + mutex_exit(&ztest_vdev_lock); } } if (maxfaults == 0) - return; + goto out; /* * We have at least single-fault tolerance, so inject data corruption. @@ -4570,7 +4989,7 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) fd = open(pathrand, O_RDWR); if (fd == -1) /* we hit a gap in the device namespace */ - return; + goto out; fsize = lseek(fd, 0, SEEK_END); @@ -4582,25 +5001,28 @@ ztest_fault_inject(ztest_ds_t *zd, uint64_t id) if (offset >= fsize) continue; - VERIFY(mutex_lock(&zs->zs_vdev_lock) == 0); + mutex_enter(&ztest_vdev_lock); if (mirror_save != zs->zs_mirrors) { - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); (void) close(fd); - return; + goto out; } if (pwrite(fd, &bad, sizeof (bad), offset) != sizeof (bad)) fatal(1, "can't inject bad word at 0x%llx in %s", offset, pathrand); - VERIFY(mutex_unlock(&zs->zs_vdev_lock) == 0); + mutex_exit(&ztest_vdev_lock); - if (zopt_verbose >= 7) + if (ztest_opts.zo_verbose >= 7) (void) printf("injected bad word into %s," " offset 0x%llx\n", pathrand, (u_longlong_t)offset); } (void) close(fd); +out: + umem_free(path0, MAXPATHLEN); + umem_free(pathrand, MAXPATHLEN); } /* @@ -4610,9 +5032,9 @@ void ztest_ddt_repair(ztest_ds_t *zd, uint64_t id) { ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; objset_t *os = zd->zd_os; - ztest_od_t od[1]; + ztest_od_t *od; uint64_t object, blocksize, txg, pattern, psize; enum zio_checksum checksum = spa_dedup_checksum(spa); dmu_buf_t *db; @@ -4625,28 +5047,32 @@ ztest_ddt_repair(ztest_ds_t *zd, uint64_t id) blocksize = ztest_random_blocksize(); blocksize = MIN(blocksize, 2048); /* because we write so many */ - ztest_od_init(&od[0], id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); + od = umem_alloc(sizeof(ztest_od_t), UMEM_NOFAIL); + ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0); - if (ztest_object_init(zd, od, sizeof (od), B_FALSE) != 0) + if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) { + umem_free(od, sizeof(ztest_od_t)); return; + } /* * Take the name lock as writer to prevent anyone else from changing * the pool and dataset properies we need to maintain during this test. */ - (void) rw_wrlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_WRITER); if (ztest_dsl_prop_set_uint64(zd->zd_name, ZFS_PROP_DEDUP, checksum, B_FALSE) != 0 || ztest_dsl_prop_set_uint64(zd->zd_name, ZFS_PROP_COPIES, 1, B_FALSE) != 0) { - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); + umem_free(od, sizeof(ztest_od_t)); return; } object = od[0].od_object; blocksize = od[0].od_blocksize; - pattern = spa_guid(spa) ^ dmu_objset_fsid_guid(os); + pattern = zs->zs_guid ^ dmu_objset_fsid_guid(os); ASSERT(object != 0); @@ -4654,7 +5080,8 @@ ztest_ddt_repair(ztest_ds_t *zd, uint64_t id) dmu_tx_hold_write(tx, object, 0, copies * blocksize); txg = ztest_tx_assign(tx, TXG_WAIT, FTAG); if (txg == 0) { - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); + umem_free(od, sizeof(ztest_od_t)); return; } @@ -4698,7 +5125,8 @@ ztest_ddt_repair(ztest_ds_t *zd, uint64_t id) zio_buf_free(buf, psize); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); + umem_free(od, sizeof(ztest_od_t)); } /* @@ -4708,8 +5136,7 @@ ztest_ddt_repair(ztest_ds_t *zd, uint64_t id) void ztest_scrub(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; - spa_t *spa = zs->zs_spa; + spa_t *spa = ztest_spa; (void) spa_scan(spa, POOL_SCAN_SCRUB); (void) poll(NULL, 0, 100); /* wait a moment, then force a restart */ @@ -4717,19 +5144,48 @@ ztest_scrub(ztest_ds_t *zd, uint64_t id) } /* + * Change the guid for the pool. + */ +/* ARGSUSED */ +void +ztest_reguid(ztest_ds_t *zd, uint64_t id) +{ + spa_t *spa = ztest_spa; + uint64_t orig, load; + int error; + + orig = spa_guid(spa); + load = spa_load_guid(spa); + + (void) rw_enter(&ztest_name_lock, RW_WRITER); + error = spa_change_guid(spa); + (void) rw_exit(&ztest_name_lock); + + if (error != 0) + return; + + if (ztest_opts.zo_verbose >= 4) { + (void) printf("Changed guid old %llu -> %llu\n", + (u_longlong_t)orig, (u_longlong_t)spa_guid(spa)); + } + + VERIFY3U(orig, !=, spa_guid(spa)); + VERIFY3U(load, ==, spa_load_guid(spa)); +} + +/* * Rename the pool to a different name and then rename it back. */ /* ARGSUSED */ void ztest_spa_rename(ztest_ds_t *zd, uint64_t id) { - ztest_shared_t *zs = ztest_shared; char *oldname, *newname; spa_t *spa; - (void) rw_wrlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_WRITER); - oldname = zs->zs_pool; + oldname = ztest_opts.zo_pool; newname = umem_alloc(strlen(oldname) + 5, UMEM_NOFAIL); (void) strcpy(newname, oldname); (void) strcat(newname, "_tmp"); @@ -4749,7 +5205,7 @@ ztest_spa_rename(ztest_ds_t *zd, uint64_t id) */ VERIFY3U(0, ==, spa_open(newname, &spa, FTAG)); - ASSERT(spa == zs->zs_spa); + ASSERT(spa == ztest_spa); spa_close(spa, FTAG); /* @@ -4762,12 +5218,12 @@ ztest_spa_rename(ztest_ds_t *zd, uint64_t id) */ VERIFY3U(0, ==, spa_open(oldname, &spa, FTAG)); - ASSERT(spa == zs->zs_spa); + ASSERT(spa == ztest_spa); spa_close(spa, FTAG); umem_free(newname, strlen(newname) + 1); - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); } /* @@ -4777,52 +5233,56 @@ static void ztest_run_zdb(char *pool) { int status; - char zdb[MAXPATHLEN + MAXNAMELEN + 20]; - char zbuf[1024]; char *bin; - char *ztest; - char *isa; - int isalen; + char *zdb; + char *zbuf; FILE *fp; - (void) realpath(getexecname(), zdb); + bin = umem_alloc(MAXPATHLEN + MAXNAMELEN + 20, UMEM_NOFAIL); + zdb = umem_alloc(MAXPATHLEN + MAXNAMELEN + 20, UMEM_NOFAIL); + zbuf = umem_alloc(1024, UMEM_NOFAIL); - /* zdb lives in /usr/sbin, while ztest lives in /usr/bin */ - bin = strstr(zdb, "/usr/bin/"); - ztest = strstr(bin, "/ztest"); - isa = bin + 8; - isalen = ztest - isa; - isa = strdup(isa); - /* LINTED */ - (void) sprintf(bin, - "/usr/sbin%.*s/zdb -bcc%s%s -U %s %s", - isalen, - isa, - zopt_verbose >= 3 ? "s" : "", - zopt_verbose >= 4 ? "v" : "", + VERIFY(realpath(getexecname(), bin) != NULL); + if (strncmp(bin, "/usr/sbin/ztest", 15) == 0) { + strcpy(bin, "/usr/sbin/zdb"); /* Installed */ + } else if (strncmp(bin, "/sbin/ztest", 11) == 0) { + strcpy(bin, "/sbin/zdb"); /* Installed */ + } else { + strstr(bin, "/ztest/")[0] = '\0'; /* In-tree */ + strcat(bin, "/zdb/zdb"); + } + + (void) sprintf(zdb, + "%s -bcc%s%s -U %s %s", + bin, + ztest_opts.zo_verbose >= 3 ? "s" : "", + ztest_opts.zo_verbose >= 4 ? "v" : "", spa_config_path, pool); - free(isa); - if (zopt_verbose >= 5) + if (ztest_opts.zo_verbose >= 5) (void) printf("Executing %s\n", strstr(zdb, "zdb ")); fp = popen(zdb, "r"); - while (fgets(zbuf, sizeof (zbuf), fp) != NULL) - if (zopt_verbose >= 3) + while (fgets(zbuf, 1024, fp) != NULL) + if (ztest_opts.zo_verbose >= 3) (void) printf("%s", zbuf); status = pclose(fp); if (status == 0) - return; + goto out; ztest_dump_core = 0; if (WIFEXITED(status)) fatal(0, "'%s' exit code %d", zdb, WEXITSTATUS(status)); else fatal(0, "'%s' died with signal %d", zdb, WTERMSIG(status)); +out: + umem_free(bin, MAXPATHLEN + MAXNAMELEN + 20); + umem_free(zdb, MAXPATHLEN + MAXNAMELEN + 20); + umem_free(zbuf, 1024); } static void @@ -4830,12 +5290,12 @@ ztest_walk_pool_directory(char *header) { spa_t *spa = NULL; - if (zopt_verbose >= 6) + if (ztest_opts.zo_verbose >= 6) (void) printf("%s\n", header); mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) - if (zopt_verbose >= 6) + if (ztest_opts.zo_verbose >= 6) (void) printf("\t%s\n", spa_name(spa)); mutex_exit(&spa_namespace_lock); } @@ -4847,7 +5307,7 @@ ztest_spa_import_export(char *oldname, char *newname) uint64_t pool_guid; spa_t *spa; - if (zopt_verbose >= 4) { + if (ztest_opts.zo_verbose >= 4) { (void) printf("import/export: old = %s, new = %s\n", oldname, newname); } @@ -4922,7 +5382,7 @@ ztest_spa_import_export(char *oldname, char *newname) static void ztest_resume(spa_t *spa) { - if (spa_suspended(spa) && zopt_verbose >= 6) + if (spa_suspended(spa) && ztest_opts.zo_verbose >= 6) (void) printf("resuming from suspended state\n"); spa_vdev_state_enter(spa, SCL_NONE); vdev_clear(spa, NULL); @@ -4940,30 +5400,27 @@ ztest_resume_thread(void *arg) ztest_resume(spa); (void) poll(NULL, 0, 100); } - return (NULL); -} -static void * -ztest_deadman_thread(void *arg) -{ - ztest_shared_t *zs = arg; - int grace = 300; - hrtime_t delta; - - delta = (zs->zs_thread_stop - zs->zs_thread_start) / NANOSEC + grace; + thread_exit(); - (void) poll(NULL, 0, (int)(1000 * delta)); + return (NULL); +} - fatal(0, "failed to complete within %d seconds of deadline", grace); +#define GRACE 300 - return (NULL); +#if 0 +static void +ztest_deadman_alarm(int sig) +{ + fatal(0, "failed to complete within %d seconds of deadline", GRACE); } +#endif static void -ztest_execute(ztest_info_t *zi, uint64_t id) +ztest_execute(int test, ztest_info_t *zi, uint64_t id) { - ztest_shared_t *zs = ztest_shared; - ztest_ds_t *zd = &zs->zs_zd[id % zopt_datasets]; + ztest_ds_t *zd = &ztest_ds[id % ztest_opts.zo_datasets]; + ztest_shared_callstate_t *zc = ZTEST_GET_SHARED_CALLSTATE(test); hrtime_t functime = gethrtime(); int i; @@ -4972,10 +5429,10 @@ ztest_execute(ztest_info_t *zi, uint64_t id) functime = gethrtime() - functime; - atomic_add_64(&zi->zi_call_count, 1); - atomic_add_64(&zi->zi_call_time, functime); + atomic_add_64(&zc->zc_count, 1); + atomic_add_64(&zc->zc_time, functime); - if (zopt_verbose >= 4) { + if (ztest_opts.zo_verbose >= 4) { Dl_info dli; (void) dladdr((void *)zi->zi_func, &dli); (void) printf("%6.2f sec in %s\n", @@ -4986,11 +5443,13 @@ ztest_execute(ztest_info_t *zi, uint64_t id) static void * ztest_thread(void *arg) { + int rand; uint64_t id = (uintptr_t)arg; ztest_shared_t *zs = ztest_shared; uint64_t call_next; hrtime_t now; ztest_info_t *zi; + ztest_shared_callstate_t *zc; while ((now = gethrtime()) < zs->zs_thread_stop) { /* @@ -5008,15 +5467,20 @@ ztest_thread(void *arg) /* * Pick a random function to execute. */ - zi = &zs->zs_info[ztest_random(ZTEST_FUNCS)]; - call_next = zi->zi_call_next; + rand = ztest_random(ZTEST_FUNCS); + zi = &ztest_info[rand]; + zc = ZTEST_GET_SHARED_CALLSTATE(rand); + call_next = zc->zc_next; if (now >= call_next && - atomic_cas_64(&zi->zi_call_next, call_next, call_next + - ztest_random(2 * zi->zi_interval[0] + 1)) == call_next) - ztest_execute(zi, id); + atomic_cas_64(&zc->zc_next, call_next, call_next + + ztest_random(2 * zi->zi_interval[0] + 1)) == call_next) { + ztest_execute(rand, zi, id); + } } + thread_exit(); + return (NULL); } @@ -5027,14 +5491,14 @@ ztest_dataset_name(char *dsname, char *pool, int d) } static void -ztest_dataset_destroy(ztest_shared_t *zs, int d) +ztest_dataset_destroy(int d) { char name[MAXNAMELEN]; int t; - ztest_dataset_name(name, zs->zs_pool, d); + ztest_dataset_name(name, ztest_opts.zo_pool, d); - if (zopt_verbose >= 3) + if (ztest_opts.zo_verbose >= 3) (void) printf("Destroying %s to free up space\n", name); /* @@ -5042,7 +5506,8 @@ ztest_dataset_destroy(ztest_shared_t *zs, int d) * ztest thread t operates on dataset (t % zopt_datasets), * so there may be more than one thing to clean up. */ - for (t = d; t < zopt_threads; t += zopt_datasets) + for (t = d; t < ztest_opts.zo_threads; + t += ztest_opts.zo_datasets) ztest_dsl_dataset_cleanup(name, t); (void) dmu_objset_find(name, ztest_objset_destroy_cb, NULL, @@ -5071,31 +5536,31 @@ ztest_dataset_dirobj_verify(ztest_ds_t *zd) } static int -ztest_dataset_open(ztest_shared_t *zs, int d) +ztest_dataset_open(int d) { - ztest_ds_t *zd = &zs->zs_zd[d]; - uint64_t committed_seq = zd->zd_seq; + ztest_ds_t *zd = &ztest_ds[d]; + uint64_t committed_seq = ZTEST_GET_SHARED_DS(d)->zd_seq; objset_t *os; zilog_t *zilog; char name[MAXNAMELEN]; int error; - ztest_dataset_name(name, zs->zs_pool, d); + ztest_dataset_name(name, ztest_opts.zo_pool, d); - (void) rw_rdlock(&zs->zs_name_lock); + (void) rw_enter(&ztest_name_lock, RW_READER); error = ztest_dataset_create(name); if (error == ENOSPC) { - (void) rw_unlock(&zs->zs_name_lock); + (void) rw_exit(&ztest_name_lock); ztest_record_enospc(FTAG); return (error); } ASSERT(error == 0 || error == EEXIST); - VERIFY3U(dmu_objset_hold(name, zd, &os), ==, 0); - (void) rw_unlock(&zs->zs_name_lock); + VERIFY0(dmu_objset_hold(name, zd, &os)); + (void) rw_exit(&ztest_name_lock); - ztest_zd_init(zd, os); + ztest_zd_init(zd, ZTEST_GET_SHARED_DS(d), os); zilog = zd->zd_zilog; @@ -5110,7 +5575,7 @@ ztest_dataset_open(ztest_shared_t *zs, int d) ztest_dataset_dirobj_verify(zd); - if (zopt_verbose >= 6) + if (ztest_opts.zo_verbose >= 6) (void) printf("%s replay %llu blocks, %llu records, seq %llu\n", zd->zd_name, (u_longlong_t)zilog->zl_parse_blk_count, @@ -5128,9 +5593,9 @@ ztest_dataset_open(ztest_shared_t *zs, int d) } static void -ztest_dataset_close(ztest_shared_t *zs, int d) +ztest_dataset_close(int d) { - ztest_ds_t *zd = &zs->zs_zd[d]; + ztest_ds_t *zd = &ztest_ds[d]; zil_close(zd->zd_zilog); dmu_objset_rele(zd->zd_os, zd); @@ -5144,9 +5609,11 @@ ztest_dataset_close(ztest_shared_t *zs, int d) static void ztest_run(ztest_shared_t *zs) { - thread_t *tid; + kt_did_t *tid; spa_t *spa; - thread_t resume_tid; + objset_t *os; + kthread_t *resume_thread; + uint64_t object; int error; int t, d; @@ -5155,17 +5622,20 @@ ztest_run(ztest_shared_t *zs) /* * Initialize parent/child shared state. */ - VERIFY(_mutex_init(&zs->zs_vdev_lock, USYNC_THREAD, NULL) == 0); - VERIFY(rwlock_init(&zs->zs_name_lock, USYNC_THREAD, NULL) == 0); + mutex_init(&ztest_vdev_lock, NULL, MUTEX_DEFAULT, NULL); + rw_init(&ztest_name_lock, NULL, RW_DEFAULT, NULL); zs->zs_thread_start = gethrtime(); - zs->zs_thread_stop = zs->zs_thread_start + zopt_passtime * NANOSEC; + zs->zs_thread_stop = + zs->zs_thread_start + ztest_opts.zo_passtime * NANOSEC; zs->zs_thread_stop = MIN(zs->zs_thread_stop, zs->zs_proc_stop); zs->zs_thread_kill = zs->zs_thread_stop; - if (ztest_random(100) < zopt_killrate) - zs->zs_thread_kill -= ztest_random(zopt_passtime * NANOSEC); + if (ztest_random(100) < ztest_opts.zo_killrate) { + zs->zs_thread_kill -= + ztest_random(ztest_opts.zo_passtime * NANOSEC); + } - (void) _mutex_init(&zcl.zcl_callbacks_lock, USYNC_THREAD, NULL); + mutex_init(&zcl.zcl_callbacks_lock, NULL, MUTEX_DEFAULT, NULL); list_create(&zcl.zcl_callbacks, sizeof (ztest_cb_data_t), offsetof(ztest_cb_data_t, zcd_node)); @@ -5174,8 +5644,13 @@ ztest_run(ztest_shared_t *zs) * Open our pool. */ kernel_init(FREAD | FWRITE); - VERIFY(spa_open(zs->zs_pool, &spa, FTAG) == 0); - zs->zs_spa = spa; + VERIFY(spa_open(ztest_opts.zo_pool, &spa, FTAG) == 0); + spa->spa_debug = B_TRUE; + ztest_spa = spa; + + VERIFY3U(0, ==, dmu_objset_hold(ztest_opts.zo_pool, FTAG, &os)); + zs->zs_guid = dmu_objset_fsid_guid(os); + dmu_objset_rele(os, FTAG); spa->spa_dedup_ditto = 2 * ZIO_DEDUPDITTO_MIN; @@ -5192,14 +5667,17 @@ ztest_run(ztest_shared_t *zs) /* * Create a thread to periodically resume suspended I/O. */ - VERIFY(thr_create(0, 0, ztest_resume_thread, spa, THR_BOUND, - &resume_tid) == 0); + VERIFY3P((resume_thread = zk_thread_create(NULL, 0, + (thread_func_t)ztest_resume_thread, spa, TS_RUN, NULL, 0, 0, + PTHREAD_CREATE_JOINABLE)), !=, NULL); +#if 0 /* - * Create a deadman thread to abort() if we hang. + * Set a deadman alarm to abort() if we hang. */ - VERIFY(thr_create(0, 0, ztest_deadman_thread, zs, THR_BOUND, - NULL) == 0); + signal(SIGALRM, ztest_deadman_alarm); + alarm((zs->zs_thread_stop - zs->zs_thread_start) / NANOSEC + GRACE); +#endif /* * Verify that we can safely inquire about about any object, @@ -5220,34 +5698,42 @@ ztest_run(ztest_shared_t *zs) * If we got any ENOSPC errors on the previous run, destroy something. */ if (zs->zs_enospc_count != 0) { - int d = ztest_random(zopt_datasets); - ztest_dataset_destroy(zs, d); + int d = ztest_random(ztest_opts.zo_datasets); + ztest_dataset_destroy(d); } zs->zs_enospc_count = 0; - tid = umem_zalloc(zopt_threads * sizeof (thread_t), UMEM_NOFAIL); + tid = umem_zalloc(ztest_opts.zo_threads * sizeof (kt_did_t), + UMEM_NOFAIL); - if (zopt_verbose >= 4) + if (ztest_opts.zo_verbose >= 4) (void) printf("starting main threads...\n"); /* * Kick off all the tests that run in parallel. */ - for (t = 0; t < zopt_threads; t++) { - if (t < zopt_datasets && ztest_dataset_open(zs, t) != 0) + for (t = 0; t < ztest_opts.zo_threads; t++) { + kthread_t *thread; + + if (t < ztest_opts.zo_datasets && + ztest_dataset_open(t) != 0) return; - VERIFY(thr_create(0, 0, ztest_thread, (void *)(uintptr_t)t, - THR_BOUND, &tid[t]) == 0); + + VERIFY3P(thread = zk_thread_create(NULL, 0, + (thread_func_t)ztest_thread, + (void *)(uintptr_t)t, TS_RUN, NULL, 0, 0, + PTHREAD_CREATE_JOINABLE), !=, NULL); + tid[t] = thread->t_tid; } /* * Wait for all of the tests to complete. We go in reverse order * so we don't close datasets while threads are still using them. */ - for (t = zopt_threads - 1; t >= 0; t--) { - VERIFY(thr_join(tid[t], NULL, NULL) == 0); - if (t < zopt_datasets) - ztest_dataset_close(zs, t); + for (t = ztest_opts.zo_threads - 1; t >= 0; t--) { + thread_join(tid[t]); + if (t < ztest_opts.zo_datasets) + ztest_dataset_close(t); } txg_wait_synced(spa_get_dsl(spa), 0); @@ -5255,23 +5741,23 @@ ztest_run(ztest_shared_t *zs) zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(spa)); zs->zs_space = metaslab_class_get_space(spa_normal_class(spa)); - umem_free(tid, zopt_threads * sizeof (thread_t)); + umem_free(tid, ztest_opts.zo_threads * sizeof (kt_did_t)); /* Kill the resume thread */ ztest_exiting = B_TRUE; - VERIFY(thr_join(resume_tid, NULL, NULL) == 0); + thread_join(resume_thread->t_tid); ztest_resume(spa); /* * Right before closing the pool, kick off a bunch of async I/O; * spa_close() should wait for it to complete. */ - for (uint64_t object = 1; object < 50; object++) + for (object = 1; object < 50; object++) dmu_prefetch(spa->spa_meta_objset, object, 0, 1ULL << 20); /* Verify that at least one commit cb was called in a timely fashion */ if (zc_cb_counter >= ZTEST_COMMIT_CB_MIN_REG) - VERIFY3U(zc_min_txg_delay, ==, 0); + VERIFY0(zc_min_txg_delay); spa_close(spa, FTAG); @@ -5280,7 +5766,7 @@ ztest_run(ztest_shared_t *zs) */ mutex_enter(&spa_namespace_lock); for (spa = spa_next(NULL); spa != NULL; spa = spa_next(spa)) - if (zopt_verbose > 3) + if (ztest_opts.zo_verbose > 3) (void) printf("spa_next: found %s\n", spa_name(spa)); mutex_exit(&spa_namespace_lock); @@ -5290,34 +5776,33 @@ ztest_run(ztest_shared_t *zs) */ if (ztest_random(2) == 0) { char name[MAXNAMELEN]; - (void) snprintf(name, MAXNAMELEN, "%s_import", zs->zs_pool); - ztest_spa_import_export(zs->zs_pool, name); - ztest_spa_import_export(name, zs->zs_pool); + (void) snprintf(name, MAXNAMELEN, "%s_import", + ztest_opts.zo_pool); + ztest_spa_import_export(ztest_opts.zo_pool, name); + ztest_spa_import_export(name, ztest_opts.zo_pool); } kernel_fini(); list_destroy(&zcl.zcl_callbacks); - - (void) _mutex_destroy(&zcl.zcl_callbacks_lock); - - (void) rwlock_destroy(&zs->zs_name_lock); - (void) _mutex_destroy(&zs->zs_vdev_lock); + mutex_destroy(&zcl.zcl_callbacks_lock); + rw_destroy(&ztest_name_lock); + mutex_destroy(&ztest_vdev_lock); } static void -ztest_freeze(ztest_shared_t *zs) +ztest_freeze(void) { - ztest_ds_t *zd = &zs->zs_zd[0]; + ztest_ds_t *zd = &ztest_ds[0]; spa_t *spa; int numloops = 0; - if (zopt_verbose >= 3) + if (ztest_opts.zo_verbose >= 3) (void) printf("testing spa_freeze()...\n"); kernel_init(FREAD | FWRITE); - VERIFY3U(0, ==, spa_open(zs->zs_pool, &spa, FTAG)); - VERIFY3U(0, ==, ztest_dataset_open(zs, 0)); + VERIFY3U(0, ==, spa_open(ztest_opts.zo_pool, &spa, FTAG)); + VERIFY3U(0, ==, ztest_dataset_open(0)); /* * Force the first log block to be transactionally allocated. @@ -5344,7 +5829,8 @@ ztest_freeze(ztest_shared_t *zs) * to increase well beyond the last synced value in the uberblock. * The ZIL should be OK with that. */ - while (ztest_random(10) != 0 && numloops++ < zopt_maxloops) { + while (ztest_random(10) != 0 && + numloops++ < ztest_opts.zo_maxloops) { ztest_dmu_write_parallel(zd, 0); ztest_dmu_object_alloc_free(zd, 0); txg_wait_synced(spa_get_dsl(spa), 0); @@ -5359,7 +5845,7 @@ ztest_freeze(ztest_shared_t *zs) /* * Close our dataset and close the pool. */ - ztest_dataset_close(zs, 0); + ztest_dataset_close(0); spa_close(spa, FTAG); kernel_fini(); @@ -5367,9 +5853,16 @@ ztest_freeze(ztest_shared_t *zs) * Open and close the pool and dataset to induce log replay. */ kernel_init(FREAD | FWRITE); - VERIFY3U(0, ==, spa_open(zs->zs_pool, &spa, FTAG)); - VERIFY3U(0, ==, ztest_dataset_open(zs, 0)); - ztest_dataset_close(zs, 0); + VERIFY3U(0, ==, spa_open(ztest_opts.zo_pool, &spa, FTAG)); + ASSERT(spa_freeze_txg(spa) == UINT64_MAX); + VERIFY3U(0, ==, ztest_dataset_open(0)); + ztest_dataset_close(0); + + spa->spa_debug = B_TRUE; + ztest_spa = spa; + txg_wait_synced(spa_get_dsl(spa), 0); + ztest_reguid(NULL, 0); + spa_close(spa, FTAG); kernel_fini(); } @@ -5404,15 +5897,11 @@ make_random_props(void) { nvlist_t *props; - if (ztest_random(2) == 0) - return (NULL); - VERIFY(nvlist_alloc(&props, NV_UNIQUE_NAME, 0) == 0); + if (ztest_random(2) == 0) + return (props); VERIFY(nvlist_add_uint64(props, "autoreplace", 1) == 0); - (void) printf("props:\n"); - dump_nvlist(props, 4); - return (props); } @@ -5425,39 +5914,223 @@ ztest_init(ztest_shared_t *zs) { spa_t *spa; nvlist_t *nvroot, *props; + int i; - VERIFY(_mutex_init(&zs->zs_vdev_lock, USYNC_THREAD, NULL) == 0); - VERIFY(rwlock_init(&zs->zs_name_lock, USYNC_THREAD, NULL) == 0); + mutex_init(&ztest_vdev_lock, NULL, MUTEX_DEFAULT, NULL); + rw_init(&ztest_name_lock, NULL, RW_DEFAULT, NULL); kernel_init(FREAD | FWRITE); /* * Create the storage pool. */ - (void) spa_destroy(zs->zs_pool); + (void) spa_destroy(ztest_opts.zo_pool); ztest_shared->zs_vdev_next_leaf = 0; zs->zs_splits = 0; - zs->zs_mirrors = zopt_mirrors; - nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, - 0, zopt_raidz, zs->zs_mirrors, 1); + zs->zs_mirrors = ztest_opts.zo_mirrors; + nvroot = make_vdev_root(NULL, NULL, NULL, ztest_opts.zo_vdev_size, 0, + 0, ztest_opts.zo_raidz, zs->zs_mirrors, 1); props = make_random_props(); - VERIFY3U(0, ==, spa_create(zs->zs_pool, nvroot, props, NULL, NULL)); + for (i = 0; i < SPA_FEATURES; i++) { + char *buf; + VERIFY3S(-1, !=, asprintf(&buf, "feature@%s", + spa_feature_table[i].fi_uname)); + VERIFY3U(0, ==, nvlist_add_uint64(props, buf, 0)); + free(buf); + } + VERIFY3U(0, ==, spa_create(ztest_opts.zo_pool, nvroot, props, + NULL, NULL)); nvlist_free(nvroot); - VERIFY3U(0, ==, spa_open(zs->zs_pool, &spa, FTAG)); - metaslab_sz = 1ULL << spa->spa_root_vdev->vdev_child[0]->vdev_ms_shift; + VERIFY3U(0, ==, spa_open(ztest_opts.zo_pool, &spa, FTAG)); + zs->zs_metaslab_sz = + 1ULL << spa->spa_root_vdev->vdev_child[0]->vdev_ms_shift; spa_close(spa, FTAG); kernel_fini(); - ztest_run_zdb(zs->zs_pool); + ztest_run_zdb(ztest_opts.zo_pool); + + ztest_freeze(); + + ztest_run_zdb(ztest_opts.zo_pool); + + rw_destroy(&ztest_name_lock); + mutex_destroy(&ztest_vdev_lock); +} + +static void +setup_data_fd(void) +{ + static char ztest_name_data[] = "/tmp/ztest.data.XXXXXX"; + + ztest_fd_data = mkstemp(ztest_name_data); + ASSERT3S(ztest_fd_data, >=, 0); + (void) unlink(ztest_name_data); +} + +static int +shared_data_size(ztest_shared_hdr_t *hdr) +{ + int size; + + size = hdr->zh_hdr_size; + size += hdr->zh_opts_size; + size += hdr->zh_size; + size += hdr->zh_stats_size * hdr->zh_stats_count; + size += hdr->zh_ds_size * hdr->zh_ds_count; + + return (size); +} + +static void +setup_hdr(void) +{ + int size; + ztest_shared_hdr_t *hdr; + + hdr = (void *)mmap(0, P2ROUNDUP(sizeof (*hdr), getpagesize()), + PROT_READ | PROT_WRITE, MAP_SHARED, ztest_fd_data, 0); + ASSERT(hdr != MAP_FAILED); + + VERIFY3U(0, ==, ftruncate(ztest_fd_data, sizeof (ztest_shared_hdr_t))); + + hdr->zh_hdr_size = sizeof (ztest_shared_hdr_t); + hdr->zh_opts_size = sizeof (ztest_shared_opts_t); + hdr->zh_size = sizeof (ztest_shared_t); + hdr->zh_stats_size = sizeof (ztest_shared_callstate_t); + hdr->zh_stats_count = ZTEST_FUNCS; + hdr->zh_ds_size = sizeof (ztest_shared_ds_t); + hdr->zh_ds_count = ztest_opts.zo_datasets; - ztest_freeze(zs); + size = shared_data_size(hdr); + VERIFY3U(0, ==, ftruncate(ztest_fd_data, size)); - ztest_run_zdb(zs->zs_pool); + (void) munmap((caddr_t)hdr, P2ROUNDUP(sizeof (*hdr), getpagesize())); +} + +static void +setup_data(void) +{ + int size, offset; + ztest_shared_hdr_t *hdr; + uint8_t *buf; + + hdr = (void *)mmap(0, P2ROUNDUP(sizeof (*hdr), getpagesize()), + PROT_READ, MAP_SHARED, ztest_fd_data, 0); + ASSERT(hdr != MAP_FAILED); + + size = shared_data_size(hdr); + + (void) munmap((caddr_t)hdr, P2ROUNDUP(sizeof (*hdr), getpagesize())); + hdr = ztest_shared_hdr = (void *)mmap(0, P2ROUNDUP(size, getpagesize()), + PROT_READ | PROT_WRITE, MAP_SHARED, ztest_fd_data, 0); + ASSERT(hdr != MAP_FAILED); + buf = (uint8_t *)hdr; + + offset = hdr->zh_hdr_size; + ztest_shared_opts = (void *)&buf[offset]; + offset += hdr->zh_opts_size; + ztest_shared = (void *)&buf[offset]; + offset += hdr->zh_size; + ztest_shared_callstate = (void *)&buf[offset]; + offset += hdr->zh_stats_size * hdr->zh_stats_count; + ztest_shared_ds = (void *)&buf[offset]; +} + +static boolean_t +exec_child(char *cmd, char *libpath, boolean_t ignorekill, int *statusp) +{ + pid_t pid; + int status; + char *cmdbuf = NULL; + + pid = fork(); + + if (cmd == NULL) { + cmdbuf = umem_alloc(MAXPATHLEN, UMEM_NOFAIL); + (void) strlcpy(cmdbuf, getexecname(), MAXPATHLEN); + cmd = cmdbuf; + } + + if (pid == -1) + fatal(1, "fork failed"); + + if (pid == 0) { /* child */ + char *emptyargv[2] = { cmd, NULL }; + char fd_data_str[12]; + + struct rlimit rl = { 1024, 1024 }; + (void) setrlimit(RLIMIT_NOFILE, &rl); + + (void) close(ztest_fd_rand); + VERIFY(11 >= snprintf(fd_data_str, 12, "%d", ztest_fd_data)); + VERIFY(0 == setenv("ZTEST_FD_DATA", fd_data_str, 1)); + + (void) enable_extended_FILE_stdio(-1, -1); + if (libpath != NULL) + VERIFY(0 == setenv("LD_LIBRARY_PATH", libpath, 1)); + (void) execv(cmd, emptyargv); + ztest_dump_core = B_FALSE; + fatal(B_TRUE, "exec failed: %s", cmd); + } + + if (cmdbuf != NULL) { + umem_free(cmdbuf, MAXPATHLEN); + cmd = NULL; + } - (void) rwlock_destroy(&zs->zs_name_lock); - (void) _mutex_destroy(&zs->zs_vdev_lock); + while (waitpid(pid, &status, 0) != pid) + continue; + if (statusp != NULL) + *statusp = status; + + if (WIFEXITED(status)) { + if (WEXITSTATUS(status) != 0) { + (void) fprintf(stderr, "child exited with code %d\n", + WEXITSTATUS(status)); + exit(2); + } + return (B_FALSE); + } else if (WIFSIGNALED(status)) { + if (!ignorekill || WTERMSIG(status) != SIGKILL) { + (void) fprintf(stderr, "child died with signal %d\n", + WTERMSIG(status)); + exit(3); + } + return (B_TRUE); + } else { + (void) fprintf(stderr, "something strange happened to child\n"); + exit(4); + /* NOTREACHED */ + } +} + +static void +ztest_run_init(void) +{ + int i; + + ztest_shared_t *zs = ztest_shared; + + ASSERT(ztest_opts.zo_init != 0); + + /* + * Blow away any existing copy of zpool.cache + */ + (void) remove(spa_config_path); + + /* + * Create and initialize our storage pool. + */ + for (i = 1; i <= ztest_opts.zo_init; i++) { + bzero(zs, sizeof (ztest_shared_t)); + if (ztest_opts.zo_verbose >= 3 && + ztest_opts.zo_init != 1) { + (void) printf("ztest_init(), pass %d\n", i); + } + ztest_init(zs); + } } int @@ -5465,65 +6138,97 @@ main(int argc, char **argv) { int kills = 0; int iters = 0; + int older = 0; + int newer = 0; ztest_shared_t *zs; - size_t shared_size; ztest_info_t *zi; + ztest_shared_callstate_t *zc; char timebuf[100]; char numbuf[6]; spa_t *spa; - int i, f; + char *cmd; + boolean_t hasalt; + int f; + char *fd_data_str = getenv("ZTEST_FD_DATA"); (void) setvbuf(stdout, NULL, _IOLBF, 0); - ztest_random_fd = open("/dev/urandom", O_RDONLY); + ztest_fd_rand = open("/dev/urandom", O_RDONLY); + ASSERT3S(ztest_fd_rand, >=, 0); + + if (!fd_data_str) { + dprintf_setup(&argc, argv); + process_options(argc, argv); - process_options(argc, argv); + setup_data_fd(); + setup_hdr(); + setup_data(); + bcopy(&ztest_opts, ztest_shared_opts, + sizeof (*ztest_shared_opts)); + } else { + ztest_fd_data = atoi(fd_data_str); + setup_data(); + bcopy(ztest_shared_opts, &ztest_opts, sizeof (ztest_opts)); + } + ASSERT3U(ztest_opts.zo_datasets, ==, ztest_shared_hdr->zh_ds_count); /* Override location of zpool.cache */ VERIFY(asprintf((char **)&spa_config_path, "%s/zpool.cache", - zopt_dir) != -1); + ztest_opts.zo_dir) != -1); - /* - * Blow away any existing copy of zpool.cache - */ - if (zopt_init != 0) - (void) remove(spa_config_path); + ztest_ds = umem_alloc(ztest_opts.zo_datasets * sizeof (ztest_ds_t), + UMEM_NOFAIL); + zs = ztest_shared; - shared_size = sizeof (*zs) + zopt_datasets * sizeof (ztest_ds_t); + if (fd_data_str) { + metaslab_gang_bang = ztest_opts.zo_metaslab_gang_bang; + metaslab_df_alloc_threshold = + zs->zs_metaslab_df_alloc_threshold; + + if (zs->zs_do_init) + ztest_run_init(); + else + ztest_run(zs); + exit(0); + } - zs = ztest_shared = (void *)mmap(0, - P2ROUNDUP(shared_size, getpagesize()), - PROT_READ | PROT_WRITE, MAP_SHARED | MAP_ANON, -1, 0); + hasalt = (strlen(ztest_opts.zo_alt_ztest) != 0); - if (zopt_verbose >= 1) { + if (ztest_opts.zo_verbose >= 1) { (void) printf("%llu vdevs, %d datasets, %d threads," " %llu seconds...\n", - (u_longlong_t)zopt_vdevs, zopt_datasets, zopt_threads, - (u_longlong_t)zopt_time); + (u_longlong_t)ztest_opts.zo_vdevs, + ztest_opts.zo_datasets, + ztest_opts.zo_threads, + (u_longlong_t)ztest_opts.zo_time); } - /* - * Create and initialize our storage pool. - */ - for (i = 1; i <= zopt_init; i++) { - bzero(zs, sizeof (ztest_shared_t)); - if (zopt_verbose >= 3 && zopt_init != 1) - (void) printf("ztest_init(), pass %d\n", i); - zs->zs_pool = zopt_pool; - ztest_init(zs); + cmd = umem_alloc(MAXNAMELEN, UMEM_NOFAIL); + (void) strlcpy(cmd, getexecname(), MAXNAMELEN); + + zs->zs_do_init = B_TRUE; + if (strlen(ztest_opts.zo_alt_ztest) != 0) { + if (ztest_opts.zo_verbose >= 1) { + (void) printf("Executing older ztest for " + "initialization: %s\n", ztest_opts.zo_alt_ztest); + } + VERIFY(!exec_child(ztest_opts.zo_alt_ztest, + ztest_opts.zo_alt_libpath, B_FALSE, NULL)); + } else { + VERIFY(!exec_child(NULL, NULL, B_FALSE, NULL)); } + zs->zs_do_init = B_FALSE; - zs->zs_pool = zopt_pool; zs->zs_proc_start = gethrtime(); - zs->zs_proc_stop = zs->zs_proc_start + zopt_time * NANOSEC; + zs->zs_proc_stop = zs->zs_proc_start + ztest_opts.zo_time * NANOSEC; for (f = 0; f < ZTEST_FUNCS; f++) { - zi = &zs->zs_info[f]; - *zi = ztest_info[f]; + zi = &ztest_info[f]; + zc = ZTEST_GET_SHARED_CALLSTATE(f); if (zs->zs_proc_start + zi->zi_interval[0] > zs->zs_proc_stop) - zi->zi_call_next = UINT64_MAX; + zc->zc_next = UINT64_MAX; else - zi->zi_call_next = zs->zs_proc_start + + zc->zc_next = zs->zs_proc_start + ztest_random(2 * zi->zi_interval[0] + 1); } @@ -5534,60 +6239,43 @@ main(int argc, char **argv) */ while (gethrtime() < zs->zs_proc_stop) { int status; - pid_t pid; + boolean_t killed; /* * Initialize the workload counters for each function. */ for (f = 0; f < ZTEST_FUNCS; f++) { - zi = &zs->zs_info[f]; - zi->zi_call_count = 0; - zi->zi_call_time = 0; + zc = ZTEST_GET_SHARED_CALLSTATE(f); + zc->zc_count = 0; + zc->zc_time = 0; } /* Set the allocation switch size */ - metaslab_df_alloc_threshold = ztest_random(metaslab_sz / 4) + 1; - - pid = fork(); - - if (pid == -1) - fatal(1, "fork failed"); + zs->zs_metaslab_df_alloc_threshold = + ztest_random(zs->zs_metaslab_sz / 4) + 1; - if (pid == 0) { /* child */ - struct rlimit rl = { 1024, 1024 }; - (void) setrlimit(RLIMIT_NOFILE, &rl); - (void) enable_extended_FILE_stdio(-1, -1); - ztest_run(zs); - exit(0); - } - - while (waitpid(pid, &status, 0) != pid) - continue; - - if (WIFEXITED(status)) { - if (WEXITSTATUS(status) != 0) { - (void) fprintf(stderr, - "child exited with code %d\n", - WEXITSTATUS(status)); - exit(2); + if (!hasalt || ztest_random(2) == 0) { + if (hasalt && ztest_opts.zo_verbose >= 1) { + (void) printf("Executing newer ztest: %s\n", + cmd); } - } else if (WIFSIGNALED(status)) { - if (WTERMSIG(status) != SIGKILL) { - (void) fprintf(stderr, - "child died with signal %d\n", - WTERMSIG(status)); - exit(3); - } - kills++; + newer++; + killed = exec_child(cmd, NULL, B_TRUE, &status); } else { - (void) fprintf(stderr, "something strange happened " - "to child\n"); - exit(4); + if (hasalt && ztest_opts.zo_verbose >= 1) { + (void) printf("Executing older ztest: %s\n", + ztest_opts.zo_alt_ztest); + } + older++; + killed = exec_child(ztest_opts.zo_alt_ztest, + ztest_opts.zo_alt_libpath, B_TRUE, &status); } + if (killed) + kills++; iters++; - if (zopt_verbose >= 1) { + if (ztest_opts.zo_verbose >= 1) { hrtime_t now = gethrtime(); now = MIN(now, zs->zs_proc_stop); @@ -5602,10 +6290,10 @@ main(int argc, char **argv) 100.0 * zs->zs_alloc / zs->zs_space, numbuf, 100.0 * (now - zs->zs_proc_start) / - (zopt_time * NANOSEC), timebuf); + (ztest_opts.zo_time * NANOSEC), timebuf); } - if (zopt_verbose >= 2) { + if (ztest_opts.zo_verbose >= 2) { (void) printf("\nWorkload summary:\n\n"); (void) printf("%7s %9s %s\n", "Calls", "Time", "Function"); @@ -5614,11 +6302,12 @@ main(int argc, char **argv) for (f = 0; f < ZTEST_FUNCS; f++) { Dl_info dli; - zi = &zs->zs_info[f]; - print_time(zi->zi_call_time, timebuf); + zi = &ztest_info[f]; + zc = ZTEST_GET_SHARED_CALLSTATE(f); + print_time(zc->zc_time, timebuf); (void) dladdr((void *)zi->zi_func, &dli); (void) printf("%7llu %9s %s\n", - (u_longlong_t)zi->zi_call_count, timebuf, + (u_longlong_t)zc->zc_count, timebuf, dli.dli_sname); } (void) printf("\n"); @@ -5630,25 +6319,33 @@ main(int argc, char **argv) * instead of 'ztest'. Do a blind rename in case this happened. */ kernel_init(FREAD); - if (spa_open(zopt_pool, &spa, FTAG) == 0) { + if (spa_open(ztest_opts.zo_pool, &spa, FTAG) == 0) { spa_close(spa, FTAG); } else { char tmpname[MAXNAMELEN]; kernel_fini(); kernel_init(FREAD | FWRITE); (void) snprintf(tmpname, sizeof (tmpname), "%s_tmp", - zopt_pool); - (void) spa_rename(tmpname, zopt_pool); + ztest_opts.zo_pool); + (void) spa_rename(tmpname, ztest_opts.zo_pool); } kernel_fini(); - ztest_run_zdb(zopt_pool); + ztest_run_zdb(ztest_opts.zo_pool); } - if (zopt_verbose >= 1) { + if (ztest_opts.zo_verbose >= 1) { + if (hasalt) { + (void) printf("%d runs of older ztest: %s\n", older, + ztest_opts.zo_alt_ztest); + (void) printf("%d runs of newer ztest: %s\n", newer, + cmd); + } (void) printf("%d killed, %d completed, %.0f%% kill rate\n", kills, iters - kills, (100.0 * kills) / MAX(1, iters)); } + umem_free(cmd, MAXNAMELEN); + return (0); }