#define _GNU_SOURCE #include #include #include #include #include "locking.h" #include "cpusched.h" #include "xmalloc.h" #include "hash.h" struct map_entry { int fd; unsigned int cpu; struct map_entry *next; }; static struct hash_table mapper; static unsigned int *cpu_work_map = NULL; static unsigned int cpu_len = 0; static struct rwlock map_lock; static unsigned int get_appropriate_cpu(void) { unsigned int i, cpu = 0; unsigned int work = UINT_MAX; for (i = 0; i < cpu_len; ++i) { if (cpu_work_map[i] < work) { work = cpu_work_map[i]; cpu = i; } } return cpu; } unsigned int socket_to_cpu(int fd) { int cpu = 0; struct map_entry *entry; rwlock_rd_lock(&map_lock); entry = lookup_hash(fd, &mapper); while (entry && fd != entry->fd) entry = entry->next; if (entry && fd == entry->fd) cpu = entry->cpu; rwlock_unlock(&map_lock); return cpu; } unsigned int register_socket(int fd) { void **pos; struct map_entry *entry; rwlock_wr_lock(&map_lock); entry = xzmalloc(sizeof(*entry)); entry->fd = fd; entry->cpu = get_appropriate_cpu(); cpu_work_map[entry->cpu]++; pos = insert_hash(entry->fd, entry, &mapper); if (pos) { entry->next = (*pos); (*pos) = entry; } rwlock_unlock(&map_lock); return entry->cpu; } static struct map_entry *socket_to_map_entry(int fd) { struct map_entry *entry, *ret = NULL; rwlock_rd_lock(&map_lock); entry = lookup_hash(fd, &mapper); while (entry && fd != entry->fd) entry = entry->next; if (entry && fd == entry->fd) ret = entry; rwlock_unlock(&map_lock); return ret; } void unregister_socket(int fd) { struct map_entry *pos; struct map_entry *entry = socket_to_map_entry(fd); if (entry == NULL) return; rwlock_wr_lock(&map_lock); cpu_work_map[entry->cpu]--; pos = remove_hash(entry->fd, entry, entry->next, &mapper); while (pos && pos->next && pos->next != entry) pos = pos->next; if (pos && pos->next && pos->next == entry) pos->next = entry->next; entry->next = NULL; xfree(entry); rwlock_unlock(&map_lock); } static int cleanup_cpusched_batch(void *ptr) { struct map_entry *next; struct map_entry *entry = ptr; if (!entry) return 0; while ((next = entry->next)) { entry->next = NULL; xfree(entry); entry = next; } xfree(entry); return 0; } void init_cpusched(unsigned int cpus) { rwlock_init(&map_lock); cpu_len = cpus; cpu_work_map = xcalloc(cpu_len, sizeof(*cpu_work_map)); init_hash(&mapper); } void destroy_cpusched(void) { xfree(cpu_work_map); for_each_hash(&mapper, cleanup_cpusched_batch); free_hash(&mapper); rwlock_destroy(&map_lock); } /option>

path: root/net/netlabel/netlabel_user.c

/td>

Age	Commit message (Expand)	Author	Files	Lines

author	Dmitry Safonov <dsafonov@virtuozzo.com>	2016-09-05 16:33:08 +0300
committer	Thomas Gleixner <tglx@linutronix.de>	2016-09-14 21:28:11 +0200
commit	6846351052e685c2d1428e80ead2d7ca3d7ed913 (patch)
tree	5bc4ba09c325f2cd672e793f3c23f69b206bea19
parent	cc87324b3dbb9bdf6916c7f479230db24c4aa309 (diff)

x86/signal: Add SA_{X32,IA32}_ABI sa_flags

Introduce new flags that defines which ABI to use on creating sigframe. Those flags kernel will set according to sigaction syscall ABI, which set handler for the signal being delivered. So that will drop the dependency on TIF_IA32/TIF_X32 flags on signal deliver. Those flags will be used only under CONFIG_COMPAT. Similar way ARM uses sa_flags to differ in which mode deliver signal for 26-bit applications (look at SA_THIRYTWO). Signed-off-by: Dmitry Safonov <dsafonov@virtuozzo.com> Reviewed-by: Andy Lutomirski <luto@kernel.org> Cc: 0x7f454c46@gmail.com Cc: oleg@redhat.com Cc: linux-mm@kvack.org Cc: gorcunov@openvz.org Cc: xemul@virtuozzo.com Link: http://lkml.kernel.org/r/20160905133308.28234-7-dsafonov@virtuozzo.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

Diffstat