diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2017-02-22 10:15:09 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2017-02-22 10:15:09 -0800 |
commit | 3051bf36c25d5153051704291782f8d44e744d36 (patch) | |
tree | 72dfc8a1d12675c6f2981d13102df954b678f11b /lib | |
parent | 1e74a2eb1f5cc7f2f2b5aa9c9eeecbcf352220a3 (diff) | |
parent | 005c3490e9db23738d91e02788606c0fe4734723 (diff) |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
Pull networking updates from David Miller:
"Highlights:
1) Support TX_RING in AF_PACKET TPACKET_V3 mode, from Sowmini
Varadhan.
2) Simplify classifier state on sk_buff in order to shrink it a bit.
From Willem de Bruijn.
3) Introduce SIPHASH and it's usage for secure sequence numbers and
syncookies. From Jason A. Donenfeld.
4) Reduce CPU usage for ICMP replies we are going to limit or
suppress, from Jesper Dangaard Brouer.
5) Introduce Shared Memory Communications socket layer, from Ursula
Braun.
6) Add RACK loss detection and allow it to actually trigger fast
recovery instead of just assisting after other algorithms have
triggered it. From Yuchung Cheng.
7) Add xmit_more and BQL support to mvneta driver, from Simon Guinot.
8) skb_cow_data avoidance in esp4 and esp6, from Steffen Klassert.
9) Export MPLS packet stats via netlink, from Robert Shearman.
10) Significantly improve inet port bind conflict handling, especially
when an application is restarted and changes it's setting of
reuseport. From Josef Bacik.
11) Implement TX batching in vhost_net, from Jason Wang.
12) Extend the dummy device so that VF (virtual function) features,
such as configuration, can be more easily tested. From Phil
Sutter.
13) Avoid two atomic ops per page on x86 in bnx2x driver, from Eric
Dumazet.
14) Add new bpf MAP, implementing a longest prefix match trie. From
Daniel Mack.
15) Packet sample offloading support in mlxsw driver, from Yotam Gigi.
16) Add new aquantia driver, from David VomLehn.
17) Add bpf tracepoints, from Daniel Borkmann.
18) Add support for port mirroring to b53 and bcm_sf2 drivers, from
Florian Fainelli.
19) Remove custom busy polling in many drivers, it is done in the core
networking since 4.5 times. From Eric Dumazet.
20) Support XDP adjust_head in virtio_net, from John Fastabend.
21) Fix several major holes in neighbour entry confirmation, from
Julian Anastasov.
22) Add XDP support to bnxt_en driver, from Michael Chan.
23) VXLAN offloads for enic driver, from Govindarajulu Varadarajan.
24) Add IPVTAP driver (IP-VLAN based tap driver) from Sainath Grandhi.
25) Support GRO in IPSEC protocols, from Steffen Klassert"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1764 commits)
Revert "ath10k: Search SMBIOS for OEM board file extension"
net: socket: fix recvmmsg not returning error from sock_error
bnxt_en: use eth_hw_addr_random()
bpf: fix unlocking of jited image when module ronx not set
arch: add ARCH_HAS_SET_MEMORY config
net: napi_watchdog() can use napi_schedule_irqoff()
tcp: Revert "tcp: tcp_probe: use spin_lock_bh()"
net/hsr: use eth_hw_addr_random()
net: mvpp2: enable building on 64-bit platforms
net: mvpp2: switch to build_skb() in the RX path
net: mvpp2: simplify MVPP2_PRS_RI_* definitions
net: mvpp2: fix indentation of MVPP2_EXT_GLOBAL_CTRL_DEFAULT
net: mvpp2: remove unused register definitions
net: mvpp2: simplify mvpp2_bm_bufs_add()
net: mvpp2: drop useless fields in mvpp2_bm_pool and related code
net: mvpp2: remove unused 'tx_skb' field of 'struct mvpp2_tx_queue'
net: mvpp2: release reference to txq_cpu[] entry after unmapping
net: mvpp2: handle too large value in mvpp2_rx_time_coal_set()
net: mvpp2: handle too large value handling in mvpp2_rx_pkts_coal_set()
net: mvpp2: remove useless arguments in mvpp2_rx_{pkts, time}_coal_set
...
Diffstat (limited to 'lib')
-rw-r--r-- | lib/Kconfig | 3 | ||||
-rw-r--r-- | lib/Kconfig.debug | 16 | ||||
-rw-r--r-- | lib/Makefile | 8 | ||||
-rw-r--r-- | lib/parman.c | 376 | ||||
-rw-r--r-- | lib/rhashtable.c | 270 | ||||
-rw-r--r-- | lib/siphash.c | 551 | ||||
-rw-r--r-- | lib/test_parman.c | 395 | ||||
-rw-r--r-- | lib/test_siphash.c | 223 |
8 files changed, 1787 insertions, 55 deletions
diff --git a/lib/Kconfig b/lib/Kconfig index 260a80e313b9..5d644f180fe5 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -550,4 +550,7 @@ config STACKDEPOT config SBITMAP bool +config PARMAN + tristate "parman" + endmenu diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index acedbe626d47..8397e4e3db9c 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1831,13 +1831,23 @@ config TEST_HASH tristate "Perform selftest on hash functions" default n help - Enable this option to test the kernel's integer (<linux/hash,h>) - and string (<linux/stringhash.h>) hash functions on boot - (or module load). + Enable this option to test the kernel's integer (<linux/hash.h>), + string (<linux/stringhash.h>), and siphash (<linux/siphash.h>) + hash functions on boot (or module load). This is intended to help people writing architecture-specific optimized versions. If unsure, say N. +config TEST_PARMAN + tristate "Perform selftest on priority array manager" + default n + depends on PARMAN + help + Enable this option to test priority array manager on boot + (or module load). + + If unsure, say N. + endmenu # runtime tests config PROVIDE_OHCI1394_DMA_INIT diff --git a/lib/Makefile b/lib/Makefile index 19ea76149a37..6b768b58a38d 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -22,7 +22,8 @@ lib-y := ctype.o string.o vsprintf.o cmdline.o \ sha1.o chacha20.o md5.o irq_regs.o argv_split.o \ flex_proportions.o ratelimit.o show_mem.o \ is_single_threaded.o plist.o decompress.o kobject_uevent.o \ - earlycpio.o seq_buf.o nmi_backtrace.o nodemask.o win_minmax.o + earlycpio.o seq_buf.o siphash.o \ + nmi_backtrace.o nodemask.o win_minmax.o lib-$(CONFIG_MMU) += ioremap.o lib-$(CONFIG_SMP) += cpumask.o @@ -44,7 +45,7 @@ obj-$(CONFIG_TEST_HEXDUMP) += test_hexdump.o obj-y += kstrtox.o obj-$(CONFIG_TEST_BPF) += test_bpf.o obj-$(CONFIG_TEST_FIRMWARE) += test_firmware.o -obj-$(CONFIG_TEST_HASH) += test_hash.o +obj-$(CONFIG_TEST_HASH) += test_hash.o test_siphash.o obj-$(CONFIG_TEST_KASAN) += test_kasan.o obj-$(CONFIG_TEST_KSTRTOX) += test-kstrtox.o obj-$(CONFIG_TEST_LKM) += test_module.o @@ -55,6 +56,7 @@ obj-$(CONFIG_TEST_STATIC_KEYS) += test_static_key_base.o obj-$(CONFIG_TEST_PRINTF) += test_printf.o obj-$(CONFIG_TEST_BITMAP) += test_bitmap.o obj-$(CONFIG_TEST_UUID) += test_uuid.o +obj-$(CONFIG_TEST_PARMAN) += test_parman.o ifeq ($(CONFIG_DEBUG_KOBJECT),y) CFLAGS_kobject.o += -DDEBUG @@ -229,3 +231,5 @@ obj-$(CONFIG_UBSAN) += ubsan.o UBSAN_SANITIZE_ubsan.o := n obj-$(CONFIG_SBITMAP) += sbitmap.o + +obj-$(CONFIG_PARMAN) += parman.o diff --git a/lib/parman.c b/lib/parman.c new file mode 100644 index 000000000000..c6e42a8db824 --- /dev/null +++ b/lib/parman.c @@ -0,0 +1,376 @@ +/* + * lib/parman.c - Manager for linear priority array areas + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * Copyright (c) 2017 Jiri Pirko <jiri@mellanox.com> + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the names of the copyright holders nor the names of its + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * Alternatively, this software may be distributed under the terms of the + * GNU General Public License ("GPL") version 2 as published by the Free + * Software Foundation. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGE. + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/export.h> +#include <linux/list.h> +#include <linux/err.h> +#include <linux/parman.h> + +struct parman_algo { + int (*item_add)(struct parman *parman, struct parman_prio *prio, + struct parman_item *item); + void (*item_remove)(struct parman *parman, struct parman_prio *prio, + struct parman_item *item); +}; + +struct parman { + const struct parman_ops *ops; + void *priv; + const struct parman_algo *algo; + unsigned long count; + unsigned long limit_count; + struct list_head prio_list; +}; + +static int parman_enlarge(struct parman *parman) +{ + unsigned long new_count = parman->limit_count + + parman->ops->resize_step; + int err; + + err = parman->ops->resize(parman->priv, new_count); + if (err) + return err; + parman->limit_count = new_count; + return 0; +} + +static int parman_shrink(struct parman *parman) +{ + unsigned long new_count = parman->limit_count - + parman->ops->resize_step; + int err; + + if (new_count < parman->ops->base_count) + return 0; + err = parman->ops->resize(parman->priv, new_count); + if (err) + return err; + parman->limit_count = new_count; + return 0; +} + +static bool parman_prio_used(struct parman_prio *prio) + +{ + return !list_empty(&prio->item_list); +} + +static struct parman_item *parman_prio_first_item(struct parman_prio *prio) +{ + return list_first_entry(&prio->item_list, + typeof(struct parman_item), list); +} + +static unsigned long parman_prio_first_index(struct parman_prio *prio) +{ + return parman_prio_first_item(prio)->index; +} + +static struct parman_item *parman_prio_last_item(struct parman_prio *prio) +{ + return list_last_entry(&prio->item_list, + typeof(struct parman_item), list); +} + +static unsigned long parman_prio_last_index(struct parman_prio *prio) +{ + return parman_prio_last_item(prio)->index; +} + +static unsigned long parman_lsort_new_index_find(struct parman *parman, + struct parman_prio *prio) +{ + list_for_each_entry_from_reverse(prio, &parman->prio_list, list) { + if (!parman_prio_used(prio)) + continue; + return parman_prio_last_index(prio) + 1; + } + return 0; +} + +static void __parman_prio_move(struct parman *parman, struct parman_prio *prio, + struct parman_item *item, unsigned long to_index, + unsigned long count) +{ + parman->ops->move(parman->priv, item->index, to_index, count); +} + +static void parman_prio_shift_down(struct parman *parman, + struct parman_prio *prio) +{ + struct parman_item *item; + unsigned long to_index; + + if (!parman_prio_used(prio)) + return; + item = parman_prio_first_item(prio); + to_index = parman_prio_last_index(prio) + 1; + __parman_prio_move(parman, prio, item, to_index, 1); + list_move_tail(&item->list, &prio->item_list); + item->index = to_index; +} + +static void parman_prio_shift_up(struct parman *parman, + struct parman_prio *prio) +{ + struct parman_item *item; + unsigned long to_index; + + if (!parman_prio_used(prio)) + return; + item = parman_prio_last_item(prio); + to_index = parman_prio_first_index(prio) - 1; + __parman_prio_move(parman, prio, item, to_index, 1); + list_move(&item->list, &prio->item_list); + item->index = to_index; +} + +static void parman_prio_item_remove(struct parman *parman, + struct parman_prio *prio, + struct parman_item *item) +{ + struct parman_item *last_item; + unsigned long to_index; + + last_item = parman_prio_last_item(prio); + if (last_item == item) { + list_del(&item->list); + return; + } + to_index = item->index; + __parman_prio_move(parman, prio, last_item, to_index, 1); + list_del(&last_item->list); + list_replace(&item->list, &last_item->list); + last_item->index = to_index; +} + +static int parman_lsort_item_add(struct parman *parman, + struct parman_prio *prio, + struct parman_item *item) +{ + struct parman_prio *prio2; + unsigned long new_index; + int err; + + if (parman->count + 1 > parman->limit_count) { + err = parman_enlarge(parman); + if (err) + return err; + } + + new_index = parman_lsort_new_index_find(parman, prio); + list_for_each_entry_reverse(prio2, &parman->prio_list, list) { + if (prio2 == prio) + break; + parman_prio_shift_down(parman, prio2); + } + item->index = new_index; + list_add_tail(&item->list, &prio->item_list); + parman->count++; + return 0; +} + +static void parman_lsort_item_remove(struct parman *parman, + struct parman_prio *prio, + struct parman_item *item) +{ + parman_prio_item_remove(parman, prio, item); + list_for_each_entry_continue(prio, &parman->prio_list, list) + parman_prio_shift_up(parman, prio); + parman->count--; + if (parman->limit_count - parman->count >= parman->ops->resize_step) + parman_shrink(parman); +} + +static const struct parman_algo parman_lsort = { + .item_add = parman_lsort_item_add, + .item_remove = parman_lsort_item_remove, +}; + +static const struct parman_algo *parman_algos[] = { + &parman_lsort, +}; + +/** + * parman_create - creates a new parman instance + * @ops: caller-specific callbacks + * @priv: pointer to a private data passed to the ops + * + * Note: all locking must be provided by the caller. + * + * Each parman instance manages an array area with chunks of entries + * with the same priority. Consider following example: + * + * item 1 with prio 10 + * item 2 with prio 10 + * item 3 with prio 10 + * item 4 with prio 20 + * item 5 with prio 20 + * item 6 with prio 30 + * item 7 with prio 30 + * item 8 with prio 30 + * + * In this example, there are 3 priority chunks. The order of the priorities + * matters, however the order of items within a single priority chunk does not + * matter. So the same array could be ordered as follows: + * + * item 2 with prio 10 + * item 3 with prio 10 + * item 1 with prio 10 + * item 5 with prio 20 + * item 4 with prio 20 + * item 7 with prio 30 + * item 8 with prio 30 + * item 6 with prio 30 + * + * The goal of parman is to maintain the priority ordering. The caller + * provides @ops with callbacks parman uses to move the items + * and resize the array area. + * + * Returns a pointer to newly created parman instance in case of success, + * otherwise it returns NULL. + */ +struct parman *parman_create(const struct parman_ops *ops, void *priv) +{ + struct parman *parman; + + parman = kzalloc(sizeof(*parman), GFP_KERNEL); + if (!parman) + return NULL; + INIT_LIST_HEAD(&parman->prio_list); + parman->ops = ops; + parman->priv = priv; + parman->limit_count = ops->base_count; + parman->algo = parman_algos[ops->algo]; + return parman; +} +EXPORT_SYMBOL(parman_create); + +/** + * parman_destroy - destroys existing parman instance + * @parman: parman instance + * + * Note: all locking must be provided by the caller. + */ +void parman_destroy(struct parman *parman) +{ + WARN_ON(!list_empty(&parman->prio_list)); + kfree(parman); +} +EXPORT_SYMBOL(parman_destroy); + +/** + * parman_prio_init - initializes a parman priority chunk + * @parman: parman instance + * @prio: parman prio structure to be initialized + * @prority: desired priority of the chunk + * + * Note: all locking must be provided by the caller. + * + * Before caller could add an item with certain priority, he has to + * initialize a priority chunk for it using this function. + */ +void parman_prio_init(struct parman *parman, struct parman_prio *prio, + unsigned long priority) +{ + struct parman_prio *prio2; + struct list_head *pos; + + INIT_LIST_HEAD(&prio->item_list); + prio->priority = priority; + + /* Position inside the list according to priority */ + list_for_each(pos, &parman->prio_list) { + prio2 = list_entry(pos, typeof(*prio2), list); + if (prio2->priority > prio->priority) + break; + } + list_add_tail(&prio->list, pos); +} +EXPORT_SYMBOL(parman_prio_init); + +/** + * parman_prio_fini - finalizes use of parman priority chunk + * @prio: parman prio structure + * + * Note: all locking must be provided by the caller. + */ +void parman_prio_fini(struct parman_prio *prio) +{ + WARN_ON(parman_prio_used(prio)); + list_del(&prio->list); +} +EXPORT_SYMBOL(parman_prio_fini); + +/** + * parman_item_add - adds a parman item under defined priority + * @parman: parman instance + * @prio: parman prio instance to add the item to + * @item: parman item instance + * + * Note: all locking must be provided by the caller. + * + * Adds item to a array managed by parman instance under the specified priority. + * + * Returns 0 in case of success, negative number to indicate an error. + */ +int parman_item_add(struct parman *parman, struct parman_prio *prio, + struct parman_item *item) +{ + return parman->algo->item_add(parman, prio, item); +} +EXPORT_SYMBOL(parman_item_add); + +/** + * parman_item_del - deletes parman item + * @parman: parman instance + * @prio: parman prio instance to delete the item from + * @item: parman item instance + * + * Note: all locking must be provided by the caller. + */ +void parman_item_remove(struct parman *parman, struct parman_prio *prio, + struct parman_item *item) +{ + parman->algo->item_remove(parman, prio, item); +} +EXPORT_SYMBOL(parman_item_remove); + +MODULE_LICENSE("Dual BSD/GPL"); +MODULE_AUTHOR("Jiri Pirko <jiri@mellanox.com>"); +MODULE_DESCRIPTION("Priority-based array manager"); diff --git a/lib/rhashtable.c b/lib/rhashtable.c index 32d0ad058380..172454e6b979 100644 --- a/lib/rhashtable.c +++ b/lib/rhashtable.c @@ -32,6 +32,11 @@ #define HASH_MIN_SIZE 4U #define BUCKET_LOCKS_PER_CPU 32UL +union nested_table { + union nested_table __rcu *table; + struct rhash_head __rcu *bucket; +}; + static u32 head_hashfn(struct rhashtable *ht, const struct bucket_table *tbl, const struct rhash_head *he) @@ -76,6 +81,9 @@ static int alloc_bucket_locks(struct rhashtable *ht, struct bucket_table *tbl, /* Never allocate more than 0.5 locks per bucket */ size = min_t(unsigned int, size, tbl->size >> 1); + if (tbl->nest) + size = min(size, 1U << tbl->nest); + if (sizeof(spinlock_t) != 0) { tbl->locks = NULL; #ifdef CONFIG_NUMA @@ -99,8 +107,45 @@ static int alloc_bucket_locks(struct rhashtable *ht, struct bucket_table *tbl, return 0; } +static void nested_table_free(union nested_table *ntbl, unsigned int size) +{ + const unsigned int shift = PAGE_SHIFT - ilog2(sizeof(void *)); + const unsigned int len = 1 << shift; + unsigned int i; + + ntbl = rcu_dereference_raw(ntbl->table); + if (!ntbl) + return; + + if (size > len) { + size >>= shift; + for (i = 0; i < len; i++) + nested_table_free(ntbl + i, size); + } + + kfree(ntbl); +} + +static void nested_bucket_table_free(const struct bucket_table *tbl) +{ + unsigned int size = tbl->size >> tbl->nest; + unsigned int len = 1 << tbl->nest; + union nested_table *ntbl; + unsigned int i; + + ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]); + + for (i = 0; i < len; i++) + nested_table_free(ntbl + i, size); + + kfree(ntbl); +} + static void bucket_table_free(const struct bucket_table *tbl) { + if (tbl->nest) + nested_bucket_table_free(tbl); + if (tbl) kvfree(tbl->locks); @@ -112,6 +157,59 @@ static void bucket_table_free_rcu(struct rcu_head *head) bucket_table_free(container_of(head, struct bucket_table, rcu)); } +static union nested_table *nested_table_alloc(struct rhashtable *ht, + union nested_table __rcu **prev, + unsigned int shifted, + unsigned int nhash) +{ + union nested_table *ntbl; + int i; + + ntbl = rcu_dereference(*prev); + if (ntbl) + return ntbl; + + ntbl = kzalloc(PAGE_SIZE, GFP_ATOMIC); + + if (ntbl && shifted) { + for (i = 0; i < PAGE_SIZE / sizeof(ntbl[0].bucket); i++) + INIT_RHT_NULLS_HEAD(ntbl[i].bucket, ht, + (i << shifted) | nhash); + } + + rcu_assign_pointer(*prev, ntbl); + + return ntbl; +} + +static struct bucket_table *nested_bucket_table_alloc(struct rhashtable *ht, + size_t nbuckets, + gfp_t gfp) +{ + const unsigned int shift = PAGE_SHIFT - ilog2(sizeof(void *)); + struct bucket_table *tbl; + size_t size; + + if (nbuckets < (1 << (shift + 1))) + return NULL; + + size = sizeof(*tbl) + sizeof(tbl->buckets[0]); + + tbl = kzalloc(size, gfp); + if (!tbl) + return NULL; + + if (!nested_table_alloc(ht, (union nested_table __rcu **)tbl->buckets, + 0, 0)) { + kfree(tbl); + return NULL; + } + + tbl->nest = (ilog2(nbuckets) - 1) % shift + 1; + + return tbl; +} + static struct bucket_table *bucket_table_alloc(struct rhashtable *ht, size_t nbuckets, gfp_t gfp) @@ -126,10 +224,17 @@ static struct bucket_table *bucket_table_alloc(struct rhashtable *ht, tbl = kzalloc(size, gfp | __GFP_NOWARN | __GFP_NORETRY); if (tbl == NULL && gfp == GFP_KERNEL) tbl = vzalloc(size); + + size = nbuckets; + + if (tbl == NULL && gfp != GFP_KERNEL) { + tbl = nested_bucket_table_alloc(ht, nbuckets, gfp); + nbuckets = 0; + } if (tbl == NULL) return NULL; - tbl->size = nbuckets; + tbl->size = size; if (alloc_bucket_locks(ht, tbl, gfp) < 0) { bucket_table_free(tbl); @@ -164,12 +269,17 @@ static int rhashtable_rehash_one(struct rhashtable *ht, unsigned int old_hash) struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht); struct bucket_table *new_tbl = rhashtable_last_table(ht, rht_dereference_rcu(old_tbl->future_tbl, ht)); - struct rhash_head __rcu **pprev = &old_tbl->buckets[old_hash]; - int err = -ENOENT; + struct rhash_head __rcu **pprev = rht_bucket_var(old_tbl, old_hash); + int err = -EAGAIN; struct rhash_head *head, *next, *entry; spinlock_t *new_bucket_lock; unsigned int new_hash; + if (new_tbl->nest) + goto out; + + err = -ENOENT; + rht_for_each(entry, old_tbl, old_hash) { err = 0; next = rht_dereference_bucket(entry->next, old_tbl, old_hash); @@ -202,19 +312,26 @@ out: return err; } -static void rhashtable_rehash_chain(struct rhashtable *ht, +static int rhashtable_rehash_chain(struct rhashtable *ht, unsigned int old_hash) { struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht); spinlock_t *old_bucket_lock; + int err; old_bucket_lock = rht_bucket_lock(old_tbl, old_hash); spin_lock_bh(old_bucket_lock); - while (!rhashtable_rehash_one(ht, old_hash)) + while (!(err = rhashtable_rehash_one(ht, old_hash))) ; - old_tbl->rehash++; + + if (err == -ENOENT) { + old_tbl->rehash++; + err = 0; + } spin_unlock_bh(old_bucket_lock); + + return err; } static int rhashtable_rehash_attach(struct rhashtable *ht, @@ -246,13 +363,17 @@ static int rhashtable_rehash_table(struct rhashtable *ht) struct bucket_table *new_tbl; struct rhashtable_walker *walker; unsigned int old_hash; + int err; new_tbl = rht_dereference(old_tbl->future_tbl, ht); if (!new_tbl) return 0; - for (old_hash = 0; old_hash < old_tbl->size; old_hash++) - rhashtable_rehash_chain(ht, old_hash); + for (old_hash = 0; old_hash < old_tbl->size; old_hash++) { + err = rhashtable_rehash_chain(ht, old_hash); + if (err) + return err; + } /* Publish the new table pointer. */ rcu_assign_pointer(ht->tbl, new_tbl); @@ -271,31 +392,16 @@ static int rhashtable_rehash_table(struct rhashtable *ht) return rht_dereference(new_tbl->future_tbl, ht) ? -EAGAIN : 0; } -/** - * rhashtable_expand - Expand hash table while allowing concurrent lookups - * @ht: the hash table to expand - * - * A secondary bucket array is allocated and the hash entries are migrated. - * - * This function may only be called in a context where it is safe to call - * synchronize_rcu(), e.g. not within a rcu_read_lock() section. - * - * The caller must ensure that no concurrent resizing occurs by holding - * ht->mutex. - * - * It is valid to have concurrent insertions and deletions protected by per - * bucket locks or concurrent RCU protected lookups and traversals. - */ -static int rhashtable_expand(struct rhashtable *ht) +static int rhashtable_rehash_alloc(struct rhashtable *ht, + struct bucket_table *old_tbl, + unsigned int size) { - struct bucket_table *new_tbl, *old_tbl = rht_dereference(ht->tbl, ht); + struct bucket_table *new_tbl; int err; ASSERT_RHT_MUTEX(ht); - old_tbl = rhashtable_last_table(ht, old_tbl); - - new_tbl = bucket_table_alloc(ht, old_tbl->size * 2, GFP_KERNEL); + new_tbl = bucket_table_alloc(ht, size, GFP_KERNEL); if (new_tbl == NULL) return -ENOMEM; @@ -324,12 +430,9 @@ static int rhashtable_expand(struct rhashtable *ht) */ static int rhashtable_shrink(struct rhashtable *ht) { - struct bucket_table *new_tbl, *old_tbl = rht_dereference(ht->tbl, ht); + struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht); unsigned int nelems = atomic_read(&ht->nelems); unsigned int size = 0; - int err; - - ASSERT_RHT_MUTEX(ht); if (nelems) size = roundup_pow_of_two(nelems * 3 / 2); @@ -342,15 +445,7 @@ static int rhashtable_shrink(struct rhashtable *ht) if (rht_dereference(old_tbl->future_tbl, ht)) return -EEXIST; - new_tbl = bucket_table_alloc(ht, size, GFP_KERNEL); - if (new_tbl == NULL) - return -ENOMEM; - - err = rhashtable_rehash_attach(ht, old_tbl, new_tbl); - if (err) - bucket_table_free(new_tbl); - - return err; + return rhashtable_rehash_alloc(ht, old_tbl, size); } static void rht_deferred_worker(struct work_struct *work) @@ -366,11 +461,14 @@ static void rht_deferred_worker(struct work_struct *work) tbl = rhashtable_last_table(ht, tbl); if (rht_grow_above_75(ht, tbl)) - rhashtable_expand(ht); + err = rhashtable_rehash_alloc(ht, tbl, tbl->size * 2); else if (ht->p.automatic_shrinking && rht_shrink_below_30(ht, tbl)) - rhashtable_shrink(ht); + err = rhashtable_shrink(ht); + else if (tbl->nest) + err = rhashtable_rehash_alloc(ht, tbl, tbl->size); - err = rhashtable_rehash_table(ht); + if (!err) + err = rhashtable_rehash_table(ht); mutex_unlock(&ht->mutex); @@ -439,8 +537,8 @@ static void *rhashtable_lookup_one(struct rhashtable *ht, int elasticity; elasticity = ht->elasticity; - pprev = &tbl->buckets[hash]; - rht_for_each(head, tbl, hash) { + pprev = rht_bucket_var(tbl, hash); + rht_for_each_continue(head, *pprev, tbl, hash) { struct rhlist_head *list; struct rhlist_head *plist; @@ -477,6 +575,7 @@ static struct bucket_table *rhashtable_insert_one(struct rhashtable *ht, struct rhash_head *obj, void *data) { + struct rhash_head __rcu **pprev; struct bucket_table *new_tbl; struct rhash_head *head; @@ -499,7 +598,11 @@ static struct bucket_table *rhashtable_insert_one(struct rhashtable *ht, if (unlikely(rht_grow_above_100(ht, tbl))) return ERR_PTR(-EAGAIN); - head = rht_dereference_bucket(tbl->buckets[hash], tbl, hash); + pprev = rht_bucket_insert(ht, tbl, hash); + if (!pprev) + return ERR_PTR(-ENOMEM); + + head = rht_dereference_bucket(*pprev, tbl, hash); RCU_INIT_POINTER(obj->next, head); if (ht->rhlist) { @@ -509,7 +612,7 @@ static struct bucket_table *rhashtable_insert_one(struct rhashtable *ht, RCU_INIT_POINTER(list->next, NULL); } - rcu_assign_pointer(tbl->buckets[hash], obj); + rcu_assign_pointer(*pprev, obj); atomic_inc(&ht->nelems); if (rht_grow_above_75(ht, tbl)) @@ -975,7 +1078,7 @@ void rhashtable_free_and_destroy(struct rhashtable *ht, void (*free_fn)(void *ptr, void *arg), void *arg) { - const struct bucket_table *tbl; + struct bucket_table *tbl; unsigned int i; cancel_work_sync(&ht->run_work); @@ -986,7 +1089,7 @@ void rhashtable_free_and_destroy(struct rhashtable *ht, for (i = 0; i < tbl->size; i++) { struct rhash_head *pos, *next; - for (pos = rht_dereference(tbl->buckets[i], ht), + for (pos = rht_dereference(*rht_bucket(tbl, i), ht), next = !rht_is_a_nulls(pos) ? rht_dereference(pos->next, ht) : NULL; !rht_is_a_nulls(pos); @@ -1007,3 +1110,70 @@ void rhashtable_destroy(struct rhashtable *ht) return rhashtable_free_and_destroy(ht, NULL, NULL); } EXPORT_SYMBOL_GPL(rhashtable_destroy); + +struct rhash_head __rcu **rht_bucket_nested(const struct bucket_table *tbl, + unsigned int hash) +{ + const unsigned int shift = PAGE_SHIFT - ilog2(sizeof(void *)); + static struct rhash_head __rcu *rhnull = + (struct rhash_head __rcu *)NULLS_MARKER(0); + unsigned int index = hash & ((1 << tbl->nest) - 1); + unsigned int size = tbl->size >> tbl->nest; + unsigned int subhash = hash; + union nested_table *ntbl; + + ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]); + ntbl = rht_dereference_bucket(ntbl[index].table, tbl, hash); + subhash >>= tbl->nest; + + while (ntbl && size > (1 << shift)) { + index = subhash & ((1 << shift) - 1); + ntbl = rht_dereference_bucket(ntbl[index].table, tbl, hash); + size >>= shift; + subhash >>= shift; + } + + if (!ntbl) + return &rhnull; + + return &ntbl[subhash].bucket; + +} +EXPORT_SYMBOL_GPL(rht_bucket_nested); + +struct rhash_head __rcu **rht_bucket_nested_insert(struct rhashtable *ht, + struct bucket_table *tbl, + unsigned int hash) +{ + const unsigned int shift = PAGE_SHIFT - ilog2(sizeof(void *)); + unsigned int index = hash & ((1 << tbl->nest) - 1); + unsigned int size = tbl->size >> tbl->nest; + union nested_table *ntbl; + unsigned int shifted; + unsigned int nhash; + + ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]); + hash >>= tbl->nest; + nhash = index; + shifted = tbl->nest; + ntbl = nested_table_alloc(ht, &ntbl[index].table, + size <= (1 << shift) ? shifted : 0, nhash); + + while (ntbl && size > (1 << shift)) { + index = hash & ((1 << shift) - 1); + size >>= shift; + hash >>= shift; + nhash |= index << shifted; + shifted += shift; + ntbl = nested_table_alloc(ht, &ntbl[index].table, + size <= (1 << shift) ? shifted : 0, + nhash); + } + + if (!ntbl) + return NULL; + + return &ntbl[hash].bucket; + +} +EXPORT_SYMBOL_GPL(rht_bucket_nested_insert); diff --git a/lib/siphash.c b/lib/siphash.c new file mode 100644 index 000000000000..3ae58b4edad6 --- /dev/null +++ b/lib/siphash.c @@ -0,0 +1,551 @@ +/* Copyright (C) 2016 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved. + * + * This file is provided under a dual BSD/GPLv2 license. + * + * SipHash: a fast short-input PRF + * https://131002.net/siphash/ + * + * This implementation is specifically for SipHash2-4 for a secure PRF + * and HalfSipHash1-3/SipHash1-3 for an insecure PRF only suitable for + * hashtables. + */ + +#include <linux/siphash.h> +#include <asm/unaligned.h> + +#if defined(CONFIG_DCACHE_WORD_ACCESS) && BITS_PER_LONG == 64 +#include <linux/dcache.h> +#include <asm/word-at-a-time.h> +#endif + +#define SIPROUND \ + do { \ + v0 += v1; v1 = rol64(v1, 13); v1 ^= v0; v0 = rol64(v0, 32); \ + v2 += v3; v3 = rol64(v3, 16); v3 ^= v2; \ + v0 += v3; v3 = rol64(v3, 21); v3 ^= v0; \ + v2 += v1; v1 = rol64(v1, 17); v1 ^= v2; v2 = rol64(v2, 32); \ + } while (0) + +#define PREAMBLE(len) \ + u64 v0 = 0x736f6d6570736575ULL; \ + u64 v1 = 0x646f72616e646f6dULL; \ + u64 v2 = 0x6c7967656e657261ULL; \ + u64 v3 = 0x7465646279746573ULL; \ + u64 b = ((u64)(len)) << 56; \ + v3 ^= key->key[1]; \ + v2 ^= key->key[0]; \ + v1 ^= key->key[1]; \ + v0 ^= key->key[0]; + +#define POSTAMBLE \ + v3 ^= b; \ + SIPROUND; \ + SIPROUND; \ + v0 ^= b; \ + v2 ^= 0xff; \ + SIPROUND; \ + SIPROUND; \ + SIPROUND; \ + SIPROUND; \ + return (v0 ^ v1) ^ (v2 ^ v3); + +u64 __siphash_aligned(const void *data, size_t len, const siphash_key_t *key) +{ + const u8 *end = data + len - (len % sizeof(u64)); + const u8 left = len & (sizeof(u64) - 1); + u64 m; + PREAMBLE(len) + for (; data != end; data += sizeof(u64)) { + m = le64_to_cpup(data); + v3 ^= m; + SIPROUND; + SIPROUND; + v0 ^= m; + } +#if defined(CONFIG_DCACHE_WORD_ACCESS) && BITS_PER_LONG == 64 + if (left) + b |= le64_to_cpu((__force __le64)(load_unaligned_zeropad(data) & + bytemask_from_count(left))); +#else + switch (left) { + case 7: b |= ((u64)end[6]) << 48; + case 6: b |= ((u64)end[5]) << 40; + case 5: b |= ((u64)end[4]) << 32; + case 4: b |= le32_to_cpup(data); break; + case 3: b |= ((u64)end[2]) << 16; + case 2: b |= le16_to_cpup(data); break; + case 1: b |= end[0]; + } +#endif + POSTAMBLE +} < |