/* Reallocates 'b' so that it has exactly 'new_headroom' and 'new_tailroom'
* bytes of headroom and tailroom, respectively. */
static void
ofpbuf_resize__(struct ofpbuf *b, size_t new_headroom, size_t new_tailroom)
{
void *new_base, *new_data;
size_t new_allocated;
new_allocated = new_headroom + ofpbuf_size(b) + new_tailroom;
switch (b->source) {
case OFPBUF_DPDK:
OVS_NOT_REACHED();
case OFPBUF_MALLOC:
if (new_headroom == ofpbuf_headroom(b)) {
new_base = xrealloc(ofpbuf_base(b), new_allocated);
} else {
new_base = xmalloc(new_allocated);
ofpbuf_copy__(b, new_base, new_headroom, new_tailroom);
free(ofpbuf_base(b));
}
break;
case OFPBUF_STACK:
OVS_NOT_REACHED();
case OFPBUF_STUB:
b->source = OFPBUF_MALLOC;
new_base = xmalloc(new_allocated);
ofpbuf_copy__(b, new_base, new_headroom, new_tailroom);
break;
default:
OVS_NOT_REACHED();
}
b->allocated = new_allocated;
ofpbuf_set_base(b, new_base);
new_data = (char *) new_base + new_headroom;
if (ofpbuf_data(b) != new_data) {
if (b->frame) {
uintptr_t data_delta = (char *) new_data - (char *) ofpbuf_data(b);
b->frame = (char *) b->frame + data_delta;
}
ofpbuf_set_data(b, new_data);
}
}
/* Returns the data in 'b' as a block of malloc()'d memory and frees the buffer
* within 'b'. (If 'b' itself was dynamically allocated, e.g. with
* ofpbuf_new(), then it should still be freed with, e.g., ofpbuf_delete().) */
void *
ofpbuf_steal_data(struct ofpbuf *b)
{
void *p;
ovs_assert(b->source != OFPBUF_DPDK);
if (b->source == OFPBUF_MALLOC && ofpbuf_data(b) == ofpbuf_base(b)) {
p = ofpbuf_data(b);
} else {
p = xmemdup(ofpbuf_data(b), ofpbuf_size(b));
if (b->source == OFPBUF_MALLOC) {
free(ofpbuf_base(b));
}
}
ofpbuf_set_base(b, NULL);
ofpbuf_set_data(b, NULL);
return p;
}
static void
ofpbuf_copy__(struct ofpbuf *b, uint8_t *new_base,
size_t new_headroom, size_t new_tailroom)
{
const uint8_t *old_base = ofpbuf_base(b);
size_t old_headroom = ofpbuf_headroom(b);
size_t old_tailroom = ofpbuf_tailroom(b);
size_t copy_headroom = MIN(old_headroom, new_headroom);
size_t copy_tailroom = MIN(old_tailroom, new_tailroom);
memcpy(&new_base[new_headroom - copy_headroom],
&old_base[old_headroom - copy_headroom],
copy_headroom + ofpbuf_size(b) + copy_tailroom);
}
/* Frees memory that 'b' points to. */
void
ofpbuf_uninit(struct ofpbuf *b)
{
if (b) {
if (b->source == OFPBUF_MALLOC) {
free(ofpbuf_base(b));
} else if (b->source == OFPBUF_DPDK) {
#ifdef DPDK_NETDEV
/* If this ofpbuf was allocated by DPDK it must have been
* created as a dpif_packet */
free_dpdk_buf((struct dpif_packet*) b);
#else
ovs_assert(b->source != OFPBUF_DPDK);
#endif
}
}
}
static int
nl_sock_recv__(struct nl_sock *sock, struct ofpbuf *buf, bool wait)
{
/* We can't accurately predict the size of the data to be received. The
* caller is supposed to have allocated enough space in 'buf' to handle the
* "typical" case. To handle exceptions, we make available enough space in
* 'tail' to allow Netlink messages to be up to 64 kB long (a reasonable
* figure since that's the maximum length of a Netlink attribute). */
struct nlmsghdr *nlmsghdr;
#ifdef _WIN32
#define MAX_STACK_LENGTH 81920
uint8_t tail[MAX_STACK_LENGTH];
#else
uint8_t tail[65536];
#endif
struct iovec iov[2];
struct msghdr msg;
ssize_t retval;
int error;
ovs_assert(buf->allocated >= sizeof *nlmsghdr);
ofpbuf_clear(buf);
iov[0].iov_base = ofpbuf_base(buf);
iov[0].iov_len = buf->allocated;
iov[1].iov_base = tail;
iov[1].iov_len = sizeof tail;
memset(&msg, 0, sizeof msg);
msg.msg_iov = iov;
msg.msg_iovlen = 2;
/* Receive a Netlink message from the kernel.
*
* This works around a kernel bug in which the kernel returns an error code
* as if it were the number of bytes read. It doesn't actually modify
* anything in the receive buffer in that case, so we can initialize the
* Netlink header with an impossible message length and then, upon success,
* check whether it changed. */
nlmsghdr = ofpbuf_base(buf);
do {
nlmsghdr->nlmsg_len = UINT32_MAX;
#ifdef _WIN32
boolean result = false;
DWORD last_error = 0;
result = ReadFile(sock->handle, tail, MAX_STACK_LENGTH, &retval, NULL);
last_error = GetLastError();
if (last_error != ERROR_SUCCESS && !result) {
retval = -1;
errno = EAGAIN;
} else {
ofpbuf_put(buf, tail, retval);
}
#else
retval = recvmsg(sock->fd, &msg, wait ? 0 : MSG_DONTWAIT);
#endif
error = (retval < 0 ? errno
: retval == 0 ? ECONNRESET /* not possible? */
: nlmsghdr->nlmsg_len != UINT32_MAX ? 0
: retval);
} while (error == EINTR);
if (error) {
if (error == ENOBUFS) {
/* Socket receive buffer overflow dropped one or more messages that
* the kernel tried to send to us. */
COVERAGE_INC(netlink_overflow);
}
return error;
}
if (msg.msg_flags & MSG_TRUNC) {
VLOG_ERR_RL(&rl, "truncated message (longer than %"PRIuSIZE" bytes)",
sizeof tail);
return E2BIG;
}
if (retval < sizeof *nlmsghdr
|| nlmsghdr->nlmsg_len < sizeof *nlmsghdr
|| nlmsghdr->nlmsg_len > retval) {
VLOG_ERR_RL(&rl, "received invalid nlmsg (%"PRIuSIZE" bytes < %"PRIuSIZE")",
retval, sizeof *nlmsghdr);
return EPROTO;
}
#ifndef _WIN32
ofpbuf_set_size(buf, MIN(retval, buf->allocated));
if (retval > buf->allocated) {
COVERAGE_INC(netlink_recv_jumbo);
ofpbuf_put(buf, tail, retval - buf->allocated);
}
#endif
log_nlmsg(__func__, 0, ofpbuf_data(buf), ofpbuf_size(buf), sock->protocol);
COVERAGE_INC(netlink_received);
return 0;
}
