* Copyright(c) 2017-2018 Intel Corporation
*/
+#include <fcntl.h>
#include <inttypes.h>
#include <limits.h>
-#include <sys/mman.h>
#include <stdint.h>
#include <errno.h>
-#include <sys/file.h>
#include <string.h>
+#include <unistd.h>
#include <rte_common.h>
-#include <rte_log.h>
+#include <rte_eal_paging.h>
#include <rte_errno.h>
+#include <rte_log.h>
+#include <rte_memory.h>
#include <rte_spinlock.h>
#include <rte_tailq.h>
#define MASK_LEN_TO_MOD(x) ((x) - RTE_ALIGN_FLOOR(x, MASK_ALIGN))
#define MASK_GET_IDX(idx, mod) ((idx << MASK_SHIFT) + mod)
+/*
+ * We use this to keep track of created/attached memory areas to prevent user
+ * errors in API usage.
+ */
+struct mem_area {
+ TAILQ_ENTRY(mem_area) next;
+ void *addr;
+ size_t len;
+ int fd;
+};
+TAILQ_HEAD(mem_area_head, mem_area);
+/* local per-process tailq */
+static struct mem_area_head mem_area_tailq =
+ TAILQ_HEAD_INITIALIZER(mem_area_tailq);
+static rte_spinlock_t mem_area_lock = RTE_SPINLOCK_INITIALIZER;
+
/*
* This is a mask that is always stored at the end of array, to provide fast
* way of finding free/used spots without looping through each element.
}
static int
-resize_and_map(int fd, void *addr, size_t len)
+resize_and_map(int fd, const char *path, void *addr, size_t len)
{
- char path[PATH_MAX];
void *map_addr;
- if (ftruncate(fd, len)) {
+ if (eal_file_truncate(fd, len)) {
RTE_LOG(ERR, EAL, "Cannot truncate %s\n", path);
- /* pass errno up the chain */
- rte_errno = errno;
return -1;
}
- map_addr = mmap(addr, len, PROT_READ | PROT_WRITE,
- MAP_SHARED | MAP_FIXED, fd, 0);
+ map_addr = rte_mem_map(addr, len, RTE_PROT_READ | RTE_PROT_WRITE,
+ RTE_MAP_SHARED | RTE_MAP_FORCE_ADDRESS, fd, 0);
if (map_addr != addr) {
- RTE_LOG(ERR, EAL, "mmap() failed: %s\n", strerror(errno));
- /* pass errno up the chain */
- rte_errno = errno;
return -1;
}
return 0;
}
+static int
+overlap(const struct mem_area *ma, const void *start, size_t len)
+{
+ const void *end = RTE_PTR_ADD(start, len);
+ const void *ma_start = ma->addr;
+ const void *ma_end = RTE_PTR_ADD(ma->addr, ma->len);
+
+ /* start overlap? */
+ if (start >= ma_start && start < ma_end)
+ return 1;
+ /* end overlap? */
+ if (end > ma_start && end < ma_end)
+ return 1;
+ return 0;
+}
+
static int
find_next_n(const struct rte_fbarray *arr, unsigned int start, unsigned int n,
bool used)
return 0;
}
-int __rte_experimental
+int
rte_fbarray_init(struct rte_fbarray *arr, const char *name, unsigned int len,
unsigned int elt_sz)
{
size_t page_sz, mmap_len;
char path[PATH_MAX];
struct used_mask *msk;
+ struct mem_area *ma = NULL;
void *data = NULL;
int fd = -1;
+ const struct internal_config *internal_conf =
+ eal_get_internal_configuration();
if (arr == NULL) {
rte_errno = EINVAL;
if (fully_validate(name, elt_sz, len))
return -1;
- page_sz = sysconf(_SC_PAGESIZE);
- if (page_sz == (size_t)-1)
- goto fail;
+ /* allocate mem area before doing anything */
+ ma = malloc(sizeof(*ma));
+ if (ma == NULL) {
+ rte_errno = ENOMEM;
+ return -1;
+ }
+
+ page_sz = rte_mem_page_size();
+ if (page_sz == (size_t)-1) {
+ free(ma);
+ return -1;
+ }
/* calculate our memory limits */
mmap_len = calc_data_size(page_sz, elt_sz, len);
data = eal_get_virtual_area(NULL, &mmap_len, page_sz, 0, 0);
- if (data == NULL)
- goto fail;
+ if (data == NULL) {
+ free(ma);
+ return -1;
+ }
- eal_get_fbarray_path(path, sizeof(path), name);
+ rte_spinlock_lock(&mem_area_lock);
- /*
- * Each fbarray is unique to process namespace, i.e. the filename
- * depends on process prefix. Try to take out a lock and see if we
- * succeed. If we don't, someone else is using it already.
- */
- fd = open(path, O_CREAT | O_RDWR, 0600);
- if (fd < 0) {
- RTE_LOG(DEBUG, EAL, "%s(): couldn't open %s: %s\n", __func__,
- path, strerror(errno));
- rte_errno = errno;
- goto fail;
- } else if (flock(fd, LOCK_EX | LOCK_NB)) {
- RTE_LOG(DEBUG, EAL, "%s(): couldn't lock %s: %s\n", __func__,
- path, strerror(errno));
- rte_errno = EBUSY;
- goto fail;
- }
+ fd = -1;
- /* take out a non-exclusive lock, so that other processes could still
- * attach to it, but no other process could reinitialize it.
- */
- if (flock(fd, LOCK_SH | LOCK_NB)) {
- rte_errno = errno;
- goto fail;
- }
+ if (internal_conf->no_shconf) {
+ /* remap virtual area as writable */
+ static const int flags = RTE_MAP_FORCE_ADDRESS |
+ RTE_MAP_PRIVATE | RTE_MAP_ANONYMOUS;
+ void *new_data = rte_mem_map(data, mmap_len,
+ RTE_PROT_READ | RTE_PROT_WRITE, flags, fd, 0);
+ if (new_data == NULL) {
+ RTE_LOG(DEBUG, EAL, "%s(): couldn't remap anonymous memory: %s\n",
+ __func__, rte_strerror(rte_errno));
+ goto fail;
+ }
+ } else {
+ eal_get_fbarray_path(path, sizeof(path), name);
- if (resize_and_map(fd, data, mmap_len))
- goto fail;
+ /*
+ * Each fbarray is unique to process namespace, i.e. the
+ * filename depends on process prefix. Try to take out a lock
+ * and see if we succeed. If we don't, someone else is using it
+ * already.
+ */
+ fd = eal_file_open(path, EAL_OPEN_CREATE | EAL_OPEN_READWRITE);
+ if (fd < 0) {
+ RTE_LOG(DEBUG, EAL, "%s(): couldn't open %s: %s\n",
+ __func__, path, rte_strerror(rte_errno));
+ goto fail;
+ } else if (eal_file_lock(
+ fd, EAL_FLOCK_EXCLUSIVE, EAL_FLOCK_RETURN)) {
+ RTE_LOG(DEBUG, EAL, "%s(): couldn't lock %s: %s\n",
+ __func__, path, rte_strerror(rte_errno));
+ rte_errno = EBUSY;
+ goto fail;
+ }
- /* we've mmap'ed the file, we can now close the fd */
- close(fd);
+ /* take out a non-exclusive lock, so that other processes could
+ * still attach to it, but no other process could reinitialize
+ * it.
+ */
+ if (eal_file_lock(fd, EAL_FLOCK_SHARED, EAL_FLOCK_RETURN))
+ goto fail;
+
+ if (resize_and_map(fd, path, data, mmap_len))
+ goto fail;
+ }
+ ma->addr = data;
+ ma->len = mmap_len;
+ ma->fd = fd;
+
+ /* do not close fd - keep it until detach/destroy */
+ TAILQ_INSERT_TAIL(&mem_area_tailq, ma, next);
/* initialize the data */
memset(data, 0, mmap_len);
rte_rwlock_init(&arr->rwlock);
+ rte_spinlock_unlock(&mem_area_lock);
+
return 0;
fail:
if (data)
- munmap(data, mmap_len);
+ rte_mem_unmap(data, mmap_len);
if (fd >= 0)
close(fd);
+ free(ma);
+
+ rte_spinlock_unlock(&mem_area_lock);
return -1;
}
-int __rte_experimental
+int
rte_fbarray_attach(struct rte_fbarray *arr)
{
+ struct mem_area *ma = NULL, *tmp = NULL;
size_t page_sz, mmap_len;
char path[PATH_MAX];
void *data = NULL;
if (fully_validate(arr->name, arr->elt_sz, arr->len))
return -1;
- page_sz = sysconf(_SC_PAGESIZE);
- if (page_sz == (size_t)-1)
- goto fail;
+ ma = malloc(sizeof(*ma));
+ if (ma == NULL) {
+ rte_errno = ENOMEM;
+ return -1;
+ }
+
+ page_sz = rte_mem_page_size();
+ if (page_sz == (size_t)-1) {
+ free(ma);
+ return -1;
+ }
mmap_len = calc_data_size(page_sz, arr->elt_sz, arr->len);
+ /* check the tailq - maybe user has already mapped this address space */
+ rte_spinlock_lock(&mem_area_lock);
+
+ TAILQ_FOREACH(tmp, &mem_area_tailq, next) {
+ if (overlap(tmp, arr->data, mmap_len)) {
+ rte_errno = EEXIST;
+ goto fail;
+ }
+ }
+
+ /* we know this memory area is unique, so proceed */
+
data = eal_get_virtual_area(arr->data, &mmap_len, page_sz, 0, 0);
if (data == NULL)
goto fail;
eal_get_fbarray_path(path, sizeof(path), arr->name);
- fd = open(path, O_RDWR);
+ fd = eal_file_open(path, EAL_OPEN_READWRITE);
if (fd < 0) {
- rte_errno = errno;
goto fail;
}
/* lock the file, to let others know we're using it */
- if (flock(fd, LOCK_SH | LOCK_NB)) {
- rte_errno = errno;
+ if (eal_file_lock(fd, EAL_FLOCK_SHARED, EAL_FLOCK_RETURN))
goto fail;
- }
- if (resize_and_map(fd, data, mmap_len))
+ if (resize_and_map(fd, path, data, mmap_len))
goto fail;
- close(fd);
+ /* store our new memory area */
+ ma->addr = data;
+ ma->fd = fd; /* keep fd until detach/destroy */
+ ma->len = mmap_len;
+
+ TAILQ_INSERT_TAIL(&mem_area_tailq, ma, next);
/* we're done */
+ rte_spinlock_unlock(&mem_area_lock);
return 0;
fail:
if (data)
- munmap(data, mmap_len);
+ rte_mem_unmap(data, mmap_len);
if (fd >= 0)
close(fd);
+ free(ma);
+ rte_spinlock_unlock(&mem_area_lock);
return -1;
}
-int __rte_experimental
+int
rte_fbarray_detach(struct rte_fbarray *arr)
{
+ struct mem_area *tmp = NULL;
+ size_t mmap_len;
+ int ret = -1;
+
if (arr == NULL) {
rte_errno = EINVAL;
return -1;
* really do anything about it, things will blow up either way.
*/
- size_t page_sz = sysconf(_SC_PAGESIZE);
-
+ size_t page_sz = rte_mem_page_size();
if (page_sz == (size_t)-1)
return -1;
- /* this may already be unmapped (e.g. repeated call from previously
- * failed destroy(), but this is on user, we can't (easily) know if this
- * is still mapped.
- */
- munmap(arr->data, calc_data_size(page_sz, arr->elt_sz, arr->len));
+ mmap_len = calc_data_size(page_sz, arr->elt_sz, arr->len);
- return 0;
+ /* does this area exist? */
+ rte_spinlock_lock(&mem_area_lock);
+
+ TAILQ_FOREACH(tmp, &mem_area_tailq, next) {
+ if (tmp->addr == arr->data && tmp->len == mmap_len)
+ break;
+ }
+ if (tmp == NULL) {
+ rte_errno = ENOENT;
+ ret = -1;
+ goto out;
+ }
+
+ rte_mem_unmap(arr->data, mmap_len);
+
+ /* area is unmapped, close fd and remove the tailq entry */
+ if (tmp->fd >= 0)
+ close(tmp->fd);
+ TAILQ_REMOVE(&mem_area_tailq, tmp, next);
+ free(tmp);
+
+ ret = 0;
+out:
+ rte_spinlock_unlock(&mem_area_lock);
+ return ret;
}
-int __rte_experimental
+int
rte_fbarray_destroy(struct rte_fbarray *arr)
{
+ struct mem_area *tmp = NULL;
+ size_t mmap_len;
int fd, ret;
char path[PATH_MAX];
+ const struct internal_config *internal_conf =
+ eal_get_internal_configuration();
- ret = rte_fbarray_detach(arr);
- if (ret)
- return ret;
+ if (arr == NULL) {
+ rte_errno = EINVAL;
+ return -1;
+ }
- /* try deleting the file */
- eal_get_fbarray_path(path, sizeof(path), arr->name);
+ /*
+ * we don't need to synchronize detach as two values we need (element
+ * size and total capacity) are constant for the duration of life of
+ * the array, so the parts we care about will not race. if the user is
+ * detaching while doing something else in the same process, we can't
+ * really do anything about it, things will blow up either way.
+ */
- fd = open(path, O_RDONLY);
- if (fd < 0) {
- RTE_LOG(ERR, EAL, "Could not open fbarray file: %s\n",
- strerror(errno));
+ size_t page_sz = rte_mem_page_size();
+ if (page_sz == (size_t)-1)
return -1;
+
+ mmap_len = calc_data_size(page_sz, arr->elt_sz, arr->len);
+
+ /* does this area exist? */
+ rte_spinlock_lock(&mem_area_lock);
+
+ TAILQ_FOREACH(tmp, &mem_area_tailq, next) {
+ if (tmp->addr == arr->data && tmp->len == mmap_len)
+ break;
}
- if (flock(fd, LOCK_EX | LOCK_NB)) {
- RTE_LOG(DEBUG, EAL, "Cannot destroy fbarray - another process is using it\n");
- rte_errno = EBUSY;
+ if (tmp == NULL) {
+ rte_errno = ENOENT;
ret = -1;
- } else {
- ret = 0;
- unlink(path);
- memset(arr, 0, sizeof(*arr));
+ goto out;
}
- close(fd);
+ /* with no shconf, there were never any files to begin with */
+ if (!internal_conf->no_shconf) {
+ /*
+ * attempt to get an exclusive lock on the file, to ensure it
+ * has been detached by all other processes
+ */
+ fd = tmp->fd;
+ if (eal_file_lock(fd, EAL_FLOCK_EXCLUSIVE, EAL_FLOCK_RETURN)) {
+ RTE_LOG(DEBUG, EAL, "Cannot destroy fbarray - another process is using it\n");
+ rte_errno = EBUSY;
+ ret = -1;
+ goto out;
+ }
+ /* we're OK to destroy the file */
+ eal_get_fbarray_path(path, sizeof(path), arr->name);
+ if (unlink(path)) {
+ RTE_LOG(DEBUG, EAL, "Cannot unlink fbarray: %s\n",
+ strerror(errno));
+ rte_errno = errno;
+ /*
+ * we're still holding an exclusive lock, so drop it to
+ * shared.
+ */
+ eal_file_lock(fd, EAL_FLOCK_SHARED, EAL_FLOCK_RETURN);
+
+ ret = -1;
+ goto out;
+ }
+ close(fd);
+ }
+ rte_mem_unmap(arr->data, mmap_len);
+
+ /* area is unmapped, remove the tailq entry */
+ TAILQ_REMOVE(&mem_area_tailq, tmp, next);
+ free(tmp);
+ ret = 0;
+
+ /* reset the fbarray structure */
+ memset(arr, 0, sizeof(*arr));
+out:
+ rte_spinlock_unlock(&mem_area_lock);
return ret;
}
-void * __rte_experimental
+void *
rte_fbarray_get(const struct rte_fbarray *arr, unsigned int idx)
{
void *ret = NULL;
return ret;
}
-int __rte_experimental
+int
rte_fbarray_set_used(struct rte_fbarray *arr, unsigned int idx)
{
return set_used(arr, idx, true);
}
-int __rte_experimental
+int
rte_fbarray_set_free(struct rte_fbarray *arr, unsigned int idx)
{
return set_used(arr, idx, false);
}
-int __rte_experimental
+int
rte_fbarray_is_used(struct rte_fbarray *arr, unsigned int idx)
{
struct used_mask *msk;
return ret;
}
-int __rte_experimental
+int
rte_fbarray_find_next_free(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find(arr, start, true, false);
}
-int __rte_experimental
+int
rte_fbarray_find_next_used(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find(arr, start, true, true);
}
-int __rte_experimental
+int
rte_fbarray_find_prev_free(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find(arr, start, false, false);
}
-int __rte_experimental
+int
rte_fbarray_find_prev_used(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find(arr, start, false, true);
return ret;
}
-int __rte_experimental
+int
rte_fbarray_find_next_n_free(struct rte_fbarray *arr, unsigned int start,
unsigned int n)
{
return fbarray_find_n(arr, start, n, true, false);
}
-int __rte_experimental
+int
rte_fbarray_find_next_n_used(struct rte_fbarray *arr, unsigned int start,
unsigned int n)
{
return fbarray_find_n(arr, start, n, true, true);
}
-int __rte_experimental
+int
rte_fbarray_find_prev_n_free(struct rte_fbarray *arr, unsigned int start,
unsigned int n)
{
return fbarray_find_n(arr, start, n, false, false);
}
-int __rte_experimental
+int
rte_fbarray_find_prev_n_used(struct rte_fbarray *arr, unsigned int start,
unsigned int n)
{
return ret;
}
-int __rte_experimental
+static int
+fbarray_find_biggest(struct rte_fbarray *arr, unsigned int start, bool used,
+ bool rev)
+{
+ int cur_idx, next_idx, cur_len, biggest_idx, biggest_len;
+ /* don't stack if conditions, use function pointers instead */
+ int (*find_func)(struct rte_fbarray *, unsigned int);
+ int (*find_contig_func)(struct rte_fbarray *, unsigned int);
+
+ if (arr == NULL || start >= arr->len) {
+ rte_errno = EINVAL;
+ return -1;
+ }
+ /* the other API calls already do their fair share of cheap checks, so
+ * no need to do them here.
+ */
+
+ /* the API's called are thread-safe, but something may still happen
+ * between the API calls, so lock the fbarray. all other API's are
+ * read-locking the fbarray, so read lock here is OK.
+ */
+ rte_rwlock_read_lock(&arr->rwlock);
+
+ /* pick out appropriate functions */
+ if (used) {
+ if (rev) {
+ find_func = rte_fbarray_find_prev_used;
+ find_contig_func = rte_fbarray_find_rev_contig_used;
+ } else {
+ find_func = rte_fbarray_find_next_used;
+ find_contig_func = rte_fbarray_find_contig_used;
+ }
+ } else {
+ if (rev) {
+ find_func = rte_fbarray_find_prev_free;
+ find_contig_func = rte_fbarray_find_rev_contig_free;
+ } else {
+ find_func = rte_fbarray_find_next_free;
+ find_contig_func = rte_fbarray_find_contig_free;
+ }
+ }
+
+ cur_idx = start;
+ biggest_idx = -1; /* default is error */
+ biggest_len = 0;
+ for (;;) {
+ cur_idx = find_func(arr, cur_idx);
+
+ /* block found, check its length */
+ if (cur_idx >= 0) {
+ cur_len = find_contig_func(arr, cur_idx);
+ /* decide where we go next */
+ next_idx = rev ? cur_idx - cur_len : cur_idx + cur_len;
+ /* move current index to start of chunk */
+ cur_idx = rev ? next_idx + 1 : cur_idx;
+
+ if (cur_len > biggest_len) {
+ biggest_idx = cur_idx;
+ biggest_len = cur_len;
+ }
+ cur_idx = next_idx;
+ /* in reverse mode, next_idx may be -1 if chunk started
+ * at array beginning. this means there's no more work
+ * to do.
+ */
+ if (cur_idx < 0)
+ break;
+ } else {
+ /* nothing more to find, stop. however, a failed API
+ * call has set rte_errno, which we want to ignore, as
+ * reaching the end of fbarray is not an error.
+ */
+ rte_errno = 0;
+ break;
+ }
+ }
+ /* if we didn't find anything at all, set rte_errno */
+ if (biggest_idx < 0)
+ rte_errno = used ? ENOENT : ENOSPC;
+
+ rte_rwlock_read_unlock(&arr->rwlock);
+ return biggest_idx;
+}
+
+int
+rte_fbarray_find_biggest_free(struct rte_fbarray *arr, unsigned int start)
+{
+ return fbarray_find_biggest(arr, start, false, false);
+}
+
+int
+rte_fbarray_find_biggest_used(struct rte_fbarray *arr, unsigned int start)
+{
+ return fbarray_find_biggest(arr, start, true, false);
+}
+
+int
+rte_fbarray_find_rev_biggest_free(struct rte_fbarray *arr, unsigned int start)
+{
+ return fbarray_find_biggest(arr, start, false, true);
+}
+
+int
+rte_fbarray_find_rev_biggest_used(struct rte_fbarray *arr, unsigned int start)
+{
+ return fbarray_find_biggest(arr, start, true, true);
+}
+
+
+int
rte_fbarray_find_contig_free(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find_contig(arr, start, true, false);
}
-int __rte_experimental
+int
rte_fbarray_find_contig_used(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find_contig(arr, start, true, true);
}
-int __rte_experimental
+int
rte_fbarray_find_rev_contig_free(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find_contig(arr, start, false, false);
}
-int __rte_experimental
+int
rte_fbarray_find_rev_contig_used(struct rte_fbarray *arr, unsigned int start)
{
return fbarray_find_contig(arr, start, false, true);
}
-int __rte_experimental
+int
rte_fbarray_find_idx(const struct rte_fbarray *arr, const void *elt)
{
void *end;
return ret;
}
-void __rte_experimental
+void
rte_fbarray_dump_metadata(struct rte_fbarray *arr, FILE *f)
{
struct used_mask *msk;