diff options
author | André Fabian Silva Delgado <emulatorman@parabola.nu> | 2015-09-08 01:01:14 -0300 |
---|---|---|
committer | André Fabian Silva Delgado <emulatorman@parabola.nu> | 2015-09-08 01:01:14 -0300 |
commit | e5fd91f1ef340da553f7a79da9540c3db711c937 (patch) | |
tree | b11842027dc6641da63f4bcc524f8678263304a3 /net/ceph | |
parent | 2a9b0348e685a63d97486f6749622b61e9e3292f (diff) |
Linux-libre 4.2-gnu
Diffstat (limited to 'net/ceph')
-rw-r--r-- | net/ceph/ceph_common.c | 66 | ||||
-rw-r--r-- | net/ceph/crush/crush.c | 13 | ||||
-rw-r--r-- | net/ceph/crush/crush_ln_table.h | 32 | ||||
-rw-r--r-- | net/ceph/crush/hash.c | 8 | ||||
-rw-r--r-- | net/ceph/crush/mapper.c | 148 | ||||
-rw-r--r-- | net/ceph/messenger.c | 29 | ||||
-rw-r--r-- | net/ceph/mon_client.c | 13 | ||||
-rw-r--r-- | net/ceph/osd_client.c | 42 | ||||
-rw-r--r-- | net/ceph/pagevec.c | 5 |
9 files changed, 223 insertions, 133 deletions
diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c index 79e8f71ae..f30329f72 100644 --- a/net/ceph/ceph_common.c +++ b/net/ceph/ceph_common.c @@ -9,6 +9,7 @@ #include <keys/ceph-type.h> #include <linux/module.h> #include <linux/mount.h> +#include <linux/nsproxy.h> #include <linux/parser.h> #include <linux/sched.h> #include <linux/seq_file.h> @@ -16,8 +17,6 @@ #include <linux/statfs.h> #include <linux/string.h> #include <linux/vmalloc.h> -#include <linux/nsproxy.h> -#include <net/net_namespace.h> #include <linux/ceph/ceph_features.h> @@ -131,6 +130,13 @@ int ceph_compare_options(struct ceph_options *new_opt, int i; int ret; + /* + * Don't bother comparing options if network namespaces don't + * match. + */ + if (!net_eq(current->nsproxy->net_ns, read_pnet(&client->msgr.net))) + return -1; + ret = memcmp(opt1, opt2, ofs); if (ret) return ret; @@ -335,9 +341,6 @@ ceph_parse_options(char *options, const char *dev_name, int err = -ENOMEM; substring_t argstr[MAX_OPT_ARGS]; - if (current->nsproxy->net_ns != &init_net) - return ERR_PTR(-EINVAL); - opt = kzalloc(sizeof(*opt), GFP_KERNEL); if (!opt) return ERR_PTR(-ENOMEM); @@ -352,8 +355,8 @@ ceph_parse_options(char *options, const char *dev_name, /* start with defaults */ opt->flags = CEPH_OPT_DEFAULT; opt->osd_keepalive_timeout = CEPH_OSD_KEEPALIVE_DEFAULT; - opt->mount_timeout = CEPH_MOUNT_TIMEOUT_DEFAULT; /* seconds */ - opt->osd_idle_ttl = CEPH_OSD_IDLE_TTL_DEFAULT; /* seconds */ + opt->mount_timeout = CEPH_MOUNT_TIMEOUT_DEFAULT; + opt->osd_idle_ttl = CEPH_OSD_IDLE_TTL_DEFAULT; /* get mon ip(s) */ /* ip1[:port1][,ip2[:port2]...] */ @@ -439,13 +442,32 @@ ceph_parse_options(char *options, const char *dev_name, pr_warn("ignoring deprecated osdtimeout option\n"); break; case Opt_osdkeepalivetimeout: - opt->osd_keepalive_timeout = intval; + /* 0 isn't well defined right now, reject it */ + if (intval < 1 || intval > INT_MAX / 1000) { + pr_err("osdkeepalive out of range\n"); + err = -EINVAL; + goto out; + } + opt->osd_keepalive_timeout = + msecs_to_jiffies(intval * 1000); break; case Opt_osd_idle_ttl: - opt->osd_idle_ttl = intval; + /* 0 isn't well defined right now, reject it */ + if (intval < 1 || intval > INT_MAX / 1000) { + pr_err("osd_idle_ttl out of range\n"); + err = -EINVAL; + goto out; + } + opt->osd_idle_ttl = msecs_to_jiffies(intval * 1000); break; case Opt_mount_timeout: - opt->mount_timeout = intval; + /* 0 is "wait forever" (i.e. infinite timeout) */ + if (intval < 0 || intval > INT_MAX / 1000) { + pr_err("mount_timeout out of range\n"); + err = -EINVAL; + goto out; + } + opt->mount_timeout = msecs_to_jiffies(intval * 1000); break; case Opt_share: @@ -512,12 +534,14 @@ int ceph_print_client_options(struct seq_file *m, struct ceph_client *client) seq_puts(m, "notcp_nodelay,"); if (opt->mount_timeout != CEPH_MOUNT_TIMEOUT_DEFAULT) - seq_printf(m, "mount_timeout=%d,", opt->mount_timeout); + seq_printf(m, "mount_timeout=%d,", + jiffies_to_msecs(opt->mount_timeout) / 1000); if (opt->osd_idle_ttl != CEPH_OSD_IDLE_TTL_DEFAULT) - seq_printf(m, "osd_idle_ttl=%d,", opt->osd_idle_ttl); + seq_printf(m, "osd_idle_ttl=%d,", + jiffies_to_msecs(opt->osd_idle_ttl) / 1000); if (opt->osd_keepalive_timeout != CEPH_OSD_KEEPALIVE_DEFAULT) seq_printf(m, "osdkeepalivetimeout=%d,", - opt->osd_keepalive_timeout); + jiffies_to_msecs(opt->osd_keepalive_timeout) / 1000); /* drop redundant comma */ if (m->count != pos) @@ -587,6 +611,7 @@ struct ceph_client *ceph_create_client(struct ceph_options *opt, void *private, fail_monc: ceph_monc_stop(&client->monc); fail: + ceph_messenger_fini(&client->msgr); kfree(client); return ERR_PTR(err); } @@ -600,8 +625,8 @@ void ceph_destroy_client(struct ceph_client *client) /* unmount */ ceph_osdc_stop(&client->osdc); - ceph_monc_stop(&client->monc); + ceph_messenger_fini(&client->msgr); ceph_debugfs_client_cleanup(client); @@ -626,8 +651,8 @@ static int have_mon_and_osd_map(struct ceph_client *client) */ int __ceph_open_session(struct ceph_client *client, unsigned long started) { - int err; - unsigned long timeout = client->options->mount_timeout * HZ; + unsigned long timeout = client->options->mount_timeout; + long err; /* open session, and wait for mon and osd maps */ err = ceph_monc_open_session(&client->monc); @@ -635,16 +660,15 @@ int __ceph_open_session(struct ceph_client *client, unsigned long started) return err; while (!have_mon_and_osd_map(client)) { - err = -EIO; if (timeout && time_after_eq(jiffies, started + timeout)) - return err; + return -ETIMEDOUT; /* wait */ dout("mount waiting for mon_map\n"); err = wait_event_interruptible_timeout(client->auth_wq, have_mon_and_osd_map(client) || (client->auth_err < 0), - timeout); - if (err == -EINTR || err == -ERESTARTSYS) + ceph_timeout_jiffies(timeout)); + if (err < 0) return err; if (client->auth_err < 0) return client->auth_err; @@ -721,5 +745,5 @@ module_exit(exit_ceph_lib); MODULE_AUTHOR("Sage Weil <sage@newdream.net>"); MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>"); MODULE_AUTHOR("Patience Warnick <patience@newdream.net>"); -MODULE_DESCRIPTION("Ceph filesystem for Linux"); +MODULE_DESCRIPTION("Ceph core library"); MODULE_LICENSE("GPL"); diff --git a/net/ceph/crush/crush.c b/net/ceph/crush/crush.c index 9d84ce4ea..80d7c3a97 100644 --- a/net/ceph/crush/crush.c +++ b/net/ceph/crush/crush.c @@ -1,15 +1,11 @@ - #ifdef __KERNEL__ # include <linux/slab.h> +# include <linux/crush/crush.h> #else -# include <stdlib.h> -# include <assert.h> -# define kfree(x) do { if (x) free(x); } while (0) -# define BUG_ON(x) assert(!(x)) +# include "crush_compat.h" +# include "crush.h" #endif -#include <linux/crush/crush.h> - const char *crush_bucket_alg_name(int alg) { switch (alg) { @@ -134,6 +130,9 @@ void crush_destroy(struct crush_map *map) kfree(map->rules); } +#ifndef __KERNEL__ + kfree(map->choose_tries); +#endif kfree(map); } diff --git a/net/ceph/crush/crush_ln_table.h b/net/ceph/crush/crush_ln_table.h index 6192c7fc9..aae534c90 100644 --- a/net/ceph/crush/crush_ln_table.h +++ b/net/ceph/crush/crush_ln_table.h @@ -10,20 +10,20 @@ * */ -#if defined(__linux__) -#include <linux/types.h> -#elif defined(__FreeBSD__) -#include <sys/types.h> -#endif - #ifndef CEPH_CRUSH_LN_H #define CEPH_CRUSH_LN_H +#ifdef __KERNEL__ +# include <linux/types.h> +#else +# include "crush_compat.h" +#endif -// RH_LH_tbl[2*k] = 2^48/(1.0+k/128.0) -// RH_LH_tbl[2*k+1] = 2^48*log2(1.0+k/128.0) - -static int64_t __RH_LH_tbl[128*2+2] = { +/* + * RH_LH_tbl[2*k] = 2^48/(1.0+k/128.0) + * RH_LH_tbl[2*k+1] = 2^48*log2(1.0+k/128.0) + */ +static __s64 __RH_LH_tbl[128*2+2] = { 0x0001000000000000ll, 0x0000000000000000ll, 0x0000fe03f80fe040ll, 0x000002dfca16dde1ll, 0x0000fc0fc0fc0fc1ll, 0x000005b9e5a170b4ll, 0x0000fa232cf25214ll, 0x0000088e68ea899all, 0x0000f83e0f83e0f9ll, 0x00000b5d69bac77ell, 0x0000f6603d980f67ll, 0x00000e26fd5c8555ll, @@ -89,11 +89,12 @@ static int64_t __RH_LH_tbl[128*2+2] = { 0x0000820820820821ll, 0x0000fa2f045e7832ll, 0x000081848da8faf1ll, 0x0000fba577877d7dll, 0x0000810204081021ll, 0x0000fd1a708bbe11ll, 0x0000808080808081ll, 0x0000fe8df263f957ll, 0x0000800000000000ll, 0x0000ffff00000000ll, - }; - +}; - // LL_tbl[k] = 2^48*log2(1.0+k/2^15); -static int64_t __LL_tbl[256] = { +/* + * LL_tbl[k] = 2^48*log2(1.0+k/2^15) + */ +static __s64 __LL_tbl[256] = { 0x0000000000000000ull, 0x00000002e2a60a00ull, 0x000000070cb64ec5ull, 0x00000009ef50ce67ull, 0x0000000cd1e588fdull, 0x0000000fb4747e9cull, 0x0000001296fdaf5eull, 0x0000001579811b58ull, 0x000000185bfec2a1ull, 0x0000001b3e76a552ull, 0x0000001e20e8c380ull, 0x0000002103551d43ull, @@ -160,7 +161,4 @@ static int64_t __LL_tbl[256] = { 0x000002d4562d2ec6ull, 0x000002d73330209dull, 0x000002da102d63b0ull, 0x000002dced24f814ull, }; - - - #endif diff --git a/net/ceph/crush/hash.c b/net/ceph/crush/hash.c index 5bb63e37a..ed123af49 100644 --- a/net/ceph/crush/hash.c +++ b/net/ceph/crush/hash.c @@ -1,6 +1,8 @@ - -#include <linux/types.h> -#include <linux/crush/hash.h> +#ifdef __KERNEL__ +# include <linux/crush/hash.h> +#else +# include "hash.h" +#endif /* * Robert Jenkins' function for mixing 32-bit values diff --git a/net/ceph/crush/mapper.c b/net/ceph/crush/mapper.c index 5b47736d2..393bfb22d 100644 --- a/net/ceph/crush/mapper.c +++ b/net/ceph/crush/mapper.c @@ -1,27 +1,31 @@ +/* + * Ceph - scalable distributed file system + * + * Copyright (C) 2015 Intel Corporation All Rights Reserved + * + * This is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License version 2.1, as published by the Free Software + * Foundation. See file COPYING. + * + */ #ifdef __KERNEL__ # include <linux/string.h> # include <linux/slab.h> # include <linux/bug.h> # include <linux/kernel.h> -# ifndef dprintk -# define dprintk(args...) -# endif +# include <linux/crush/crush.h> +# include <linux/crush/hash.h> #else -# include <string.h> -# include <stdio.h> -# include <stdlib.h> -# include <assert.h> -# define BUG_ON(x) assert(!(x)) -# define dprintk(args...) /* printf(args) */ -# define kmalloc(x, f) malloc(x) -# define kfree(x) free(x) +# include "crush_compat.h" +# include "crush.h" +# include "hash.h" #endif - -#include <linux/crush/crush.h> -#include <linux/crush/hash.h> #include "crush_ln_table.h" +#define dprintk(args...) /* printf(args) */ + /* * Implement the core CRUSH mapping algorithm. */ @@ -139,7 +143,7 @@ static int bucket_list_choose(struct crush_bucket_list *bucket, int i; for (i = bucket->h.size-1; i >= 0; i--) { - __u64 w = crush_hash32_4(bucket->h.hash,x, bucket->h.items[i], + __u64 w = crush_hash32_4(bucket->h.hash, x, bucket->h.items[i], r, bucket->h.id); w &= 0xffff; dprintk("list_choose i=%d x=%d r=%d item %d weight %x " @@ -238,43 +242,46 @@ static int bucket_straw_choose(struct crush_bucket_straw *bucket, return bucket->h.items[high]; } -// compute 2^44*log2(input+1) -uint64_t crush_ln(unsigned xin) +/* compute 2^44*log2(input+1) */ +static __u64 crush_ln(unsigned int xin) { - unsigned x=xin, x1; - int iexpon, index1, index2; - uint64_t RH, LH, LL, xl64, result; + unsigned int x = xin, x1; + int iexpon, index1, index2; + __u64 RH, LH, LL, xl64, result; - x++; + x++; - // normalize input - iexpon = 15; - while(!(x&0x18000)) { x<<=1; iexpon--; } + /* normalize input */ + iexpon = 15; + while (!(x & 0x18000)) { + x <<= 1; + iexpon--; + } - index1 = (x>>8)<<1; - // RH ~ 2^56/index1 - RH = __RH_LH_tbl[index1 - 256]; - // LH ~ 2^48 * log2(index1/256) - LH = __RH_LH_tbl[index1 + 1 - 256]; + index1 = (x >> 8) << 1; + /* RH ~ 2^56/index1 */ + RH = __RH_LH_tbl[index1 - 256]; + /* LH ~ 2^48 * log2(index1/256) */ + LH = __RH_LH_tbl[index1 + 1 - 256]; - // RH*x ~ 2^48 * (2^15 + xf), xf<2^8 - xl64 = (int64_t)x * RH; - xl64 >>= 48; - x1 = xl64; + /* RH*x ~ 2^48 * (2^15 + xf), xf<2^8 */ + xl64 = (__s64)x * RH; + xl64 >>= 48; + x1 = xl64; - result = iexpon; - result <<= (12 + 32); + result = iexpon; + result <<= (12 + 32); - index2 = x1 & 0xff; - // LL ~ 2^48*log2(1.0+index2/2^15) - LL = __LL_tbl[index2]; + index2 = x1 & 0xff; + /* LL ~ 2^48*log2(1.0+index2/2^15) */ + LL = __LL_tbl[index2]; - LH = LH + LL; + LH = LH + LL; - LH >>= (48-12 - 32); - result += LH; + LH >>= (48 - 12 - 32); + result += LH; - return result; + return result; } @@ -290,9 +297,9 @@ uint64_t crush_ln(unsigned xin) static int bucket_straw2_choose(struct crush_bucket_straw2 *bucket, int x, int r) { - unsigned i, high = 0; - unsigned u; - unsigned w; + unsigned int i, high = 0; + unsigned int u; + unsigned int w; __s64 ln, draw, high_draw = 0; for (i = 0; i < bucket->h.size; i++) { @@ -567,6 +574,10 @@ reject: out[outpos] = item; outpos++; count--; +#ifndef __KERNEL__ + if (map->choose_tries && ftotal <= map->choose_total_tries) + map->choose_tries[ftotal]++; +#endif } dprintk("CHOOSE returns %d\n", outpos); @@ -610,6 +621,20 @@ static void crush_choose_indep(const struct crush_map *map, } for (ftotal = 0; left > 0 && ftotal < tries; ftotal++) { +#ifdef DEBUG_INDEP + if (out2 && ftotal) { + dprintk("%u %d a: ", ftotal, left); + for (rep = outpos; rep < endpos; rep++) { + dprintk(" %d", out[rep]); + } + dprintk("\n"); + dprintk("%u %d b: ", ftotal, left); + for (rep = outpos; rep < endpos; rep++) { + dprintk(" %d", out2[rep]); + } + dprintk("\n"); + } +#endif for (rep = outpos; rep < endpos; rep++) { if (out[rep] != CRUSH_ITEM_UNDEF) continue; @@ -726,6 +751,24 @@ static void crush_choose_indep(const struct crush_map *map, out2[rep] = CRUSH_ITEM_NONE; } } +#ifndef __KERNEL__ + if (map->choose_tries && ftotal <= map->choose_total_tries) + map->choose_tries[ftotal]++; +#endif +#ifdef DEBUG_INDEP + if (out2) { + dprintk("%u %d a: ", ftotal, left); + for (rep = outpos; rep < endpos; rep++) { + dprintk(" %d", out[rep]); + } + dprintk("\n"); + dprintk("%u %d b: ", ftotal, left); + for (rep = outpos; rep < endpos; rep++) { + dprintk(" %d", out2[rep]); + } + dprintk("\n"); + } +#endif } /** @@ -790,8 +833,15 @@ int crush_do_rule(const struct crush_map *map, switch (curstep->op) { case CRUSH_RULE_TAKE: - w[0] = curstep->arg1; - wsize = 1; + if ((curstep->arg1 >= 0 && + curstep->arg1 < map->max_devices) || + (-1-curstep->arg1 < map->max_buckets && + map->buckets[-1-curstep->arg1])) { + w[0] = curstep->arg1; + wsize = 1; + } else { + dprintk(" bad take value %d\n", curstep->arg1); + } break; case CRUSH_RULE_SET_CHOOSE_TRIES: @@ -877,7 +927,7 @@ int crush_do_rule(const struct crush_map *map, 0); } else { out_size = ((numrep < (result_max-osize)) ? - numrep : (result_max-osize)); + numrep : (result_max-osize)); crush_choose_indep( map, map->buckets[-1-w[i]], @@ -923,5 +973,3 @@ int crush_do_rule(const struct crush_map *map, } return result_len; } - - diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c index 967080a9f..e3be1d22a 100644 --- a/net/ceph/messenger.c +++ b/net/ceph/messenger.c @@ -6,6 +6,7 @@ #include <linux/inet.h> #include <linux/kthread.h> #include <linux/net.h> +#include <linux/nsproxy.h> #include <linux/slab.h> #include <linux/socket.h> #include <linux/string.h> @@ -278,7 +279,6 @@ static void _ceph_msgr_exit(void) ceph_msgr_slab_exit(); BUG_ON(zero_page == NULL); - kunmap(zero_page); page_cache_release(zero_page); zero_page = NULL; } @@ -480,8 +480,8 @@ static int ceph_tcp_connect(struct ceph_connection *con) int ret; BUG_ON(con->sock); - ret = sock_create_kern(con->peer_addr.in_addr.ss_family, SOCK_STREAM, - IPPROTO_TCP, &sock); + ret = sock_create_kern(read_pnet(&con->msgr->net), paddr->ss_family, + SOCK_STREAM, IPPROTO_TCP, &sock); if (ret) return ret; sock->sk->sk_allocation = GFP_NOFS; @@ -1545,7 +1545,7 @@ static int write_partial_message_data(struct ceph_connection *con) page = ceph_msg_data_next(&msg->cursor, &page_offset, &length, &last_piece); ret = ceph_tcp_sendpage(con->sock, page, page_offset, - length, last_piece); + length, !last_piece); if (ret <= 0) { if (do_datacrc) msg->footer.data_crc = cpu_to_le32(crc); @@ -1732,17 +1732,17 @@ static int verify_hello(struct ceph_connection *con) static bool addr_is_blank(struct sockaddr_storage *ss) { + struct in_addr *addr = &((struct sockaddr_in *)ss)->sin_addr; + struct in6_addr *addr6 = &((struct sockaddr_in6 *)ss)->sin6_addr; + switch (ss->ss_family) { case AF_INET: - return ((struct sockaddr_in *)ss)->sin_addr.s_addr == 0; + return addr->s_addr == htonl(INADDR_ANY); case AF_INET6: - return - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[0] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[1] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[2] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[3] == 0; + return ipv6_addr_any(addr6); + default: + return true; } - return false; } static int addr_port(struct sockaddr_storage *ss) @@ -2945,11 +2945,18 @@ void ceph_messenger_init(struct ceph_messenger *msgr, msgr->tcp_nodelay = tcp_nodelay; atomic_set(&msgr->stopping, 0); + write_pnet(&msgr->net, get_net(current->nsproxy->net_ns)); dout("%s %p\n", __func__, msgr); } EXPORT_SYMBOL(ceph_messenger_init); +void ceph_messenger_fini(struct ceph_messenger *msgr) +{ + put_net(read_pnet(&msgr->net)); +} +EXPORT_SYMBOL(ceph_messenger_fini); + static void clear_standby(struct ceph_connection *con) { /* come back from STANDBY? */ diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c index 2b3cf05e8..9d6ff1215 100644 --- a/net/ceph/mon_client.c +++ b/net/ceph/mon_client.c @@ -298,21 +298,28 @@ void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc) } EXPORT_SYMBOL(ceph_monc_request_next_osdmap); +/* + * Wait for an osdmap with a given epoch. + * + * @epoch: epoch to wait for + * @timeout: in jiffies, 0 means "wait forever" + */ int ceph_monc_wait_osdmap(struct ceph_mon_client *monc, u32 epoch, unsigned long timeout) { unsigned long started = jiffies; - int ret; + long ret; mutex_lock(&monc->mutex); while (monc->have_osdmap < epoch) { mutex_unlock(&monc->mutex); - if (timeout != 0 && time_after_eq(jiffies, started + timeout)) + if (timeout && time_after_eq(jiffies, started + timeout)) return -ETIMEDOUT; ret = wait_event_interruptible_timeout(monc->client->auth_wq, - monc->have_osdmap >= epoch, timeout); + monc->have_osdmap >= epoch, + ceph_timeout_jiffies(timeout)); if (ret < 0) return ret; diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index c4ec92392..50033677c 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -296,6 +296,9 @@ static void osd_req_op_data_release(struct ceph_osd_request *osd_req, case CEPH_OSD_OP_CMPXATTR: ceph_osd_data_release(&op->xattr.osd_data); break; + case CEPH_OSD_OP_STAT: + ceph_osd_data_release(&op->raw_data_in); + break; default: break; } @@ -450,7 +453,7 @@ __CEPH_FORALL_OSD_OPS(GENERATE_CASE) */ static struct ceph_osd_req_op * _osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which, - u16 opcode) + u16 opcode, u32 flags) { struct ceph_osd_req_op *op; @@ -460,14 +463,15 @@ _osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which, op = &osd_req->r_ops[which]; memset(op, 0, sizeof (*op)); op->op = opcode; + op->flags = flags; return op; } void osd_req_op_init(struct ceph_osd_request *osd_req, - unsigned int which, u16 opcode) + unsigned int which, u16 opcode, u32 flags) { - (void)_osd_req_op_init(osd_req, which, opcode); + (void)_osd_req_op_init(osd_req, which, opcode, flags); } EXPORT_SYMBOL(osd_req_op_init); @@ -476,7 +480,8 @@ void osd_req_op_extent_init(struct ceph_osd_request *osd_req, u64 offset, u64 length, u64 truncate_size, u32 truncate_seq) { - struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode); + struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, + opcode, 0); size_t payload_len = 0; BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE && @@ -515,7 +520,8 @@ EXPORT_SYMBOL(osd_req_op_extent_update); void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which, u16 opcode, const char *class, const char *method) { - struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode); + struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, + opcode, 0); struct ceph_pagelist *pagelist; size_t payload_len = 0; size_t size; @@ -552,7 +558,8 @@ int osd_req_op_xattr_init(struct ceph_osd_request *osd_req, unsigned int which, u16 opcode, const char *name, const void *value, size_t size, u8 cmp_op, u8 cmp_mode) { - struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode); + struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, + opcode, 0); struct ceph_pagelist *pagelist; size_t payload_len; @@ -585,7 +592,8 @@ void osd_req_op_watch_init(struct ceph_osd_request *osd_req, unsigned int which, u16 opcode, u64 cookie, u64 version, int flag) { - struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, opcode); + struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, + opcode, 0); BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH); @@ -602,7 +610,8 @@ void osd_req_op_alloc_hint_init(struct ceph_osd_request *osd_req, u64 expected_write_size) { struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which, - CEPH_OSD_OP_SETALLOCHINT); + CEPH_OSD_OP_SETALLOCHINT, + 0); op->alloc_hint.expected_object_size = expected_object_size; op->alloc_hint.expected_write_size = expected_write_size; @@ -786,7 +795,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, } if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) { - osd_req_op_init(req, which, opcode); + osd_req_op_init(req, which, opcode, 0); } else { u32 object_size = le32_to_cpu(layout->fl_object_size); u32 object_base = off - objoff; @@ -1088,7 +1097,7 @@ static void __move_osd_to_lru(struct ceph_osd_client *osdc, BUG_ON(!list_empty(&osd->o_osd_lru)); list_add_tail(&osd->o_osd_lru, &osdc->osd_lru); - osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ; + osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl; } static void maybe_move_osd_to_lru(struct ceph_osd_client *osdc, @@ -1199,7 +1208,7 @@ static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o) static void __schedule_osd_timeout(struct ceph_osd_client *osdc) { schedule_delayed_work(&osdc->timeout_work, - osdc->client->options->osd_keepalive_timeout * HZ); + osdc->client->options->osd_keepalive_timeout); } static void __cancel_osd_timeout(struct ceph_osd_client *osdc) @@ -1567,10 +1576,9 @@ static void handle_timeout(struct work_struct *work) { struct ceph_osd_client *osdc = container_of(work, struct ceph_osd_client, timeout_work.work); + struct ceph_options *opts = osdc->client->options; struct ceph_osd_request *req; struct ceph_osd *osd; - unsigned long keepalive = - osdc->client->options->osd_keepalive_timeout * HZ; struct list_head slow_osds; dout("timeout\n"); down_read(&osdc->map_sem); @@ -1586,7 +1594,8 @@ static void handle_timeout(struct work_struct *work) */ INIT_LIST_HEAD(&slow_osds); list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) { - if (time_before(jiffies, req->r_stamp + keepalive)) + if (time_before(jiffies, + req->r_stamp + opts->osd_keepalive_timeout)) break; osd = req->r_osd; @@ -1613,8 +1622,7 @@ static void handle_osds_timeout(struct work_struct *work) struct ceph_osd_client *osdc = container_of(work, struct ceph_osd_client, osds_timeout_work.work); - unsigned long delay = - osdc->client->options->osd_idle_ttl * HZ >> 2; + unsigned long delay = osdc->client->options->osd_idle_ttl / 4; dout("osds timeout\n"); down_read(&osdc->map_sem); @@ -2619,7 +2627,7 @@ int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client) osdc->event_count = 0; schedule_delayed_work(&osdc->osds_timeout_work, - round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ)); + round_jiffies_relative(osdc->client->options->osd_idle_ttl)); err = -ENOMEM; osdc->req_mempool = mempool_create_kmalloc_pool(10, diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c index 096d91447..d4f5f220a 100644 --- a/net/ceph/pagevec.c +++ b/net/ceph/pagevec.c @@ -51,10 +51,7 @@ void ceph_put_page_vector(struct page **pages, int num_pages, bool dirty) set_page_dirty_lock(pages[i]); put_page(pages[i]); } - if (is_vmalloc_addr(pages)) - vfree(pages); - else - kfree(pages); + kvfree(pages); } EXPORT_SYMBOL(ceph_put_page_vector); |