Solve http buffering issues and hangs once and for all (hopefully!) by

simply not trying to return exactly what the caller asked for - just
return whatever we got and let the caller be the judge of whether it
was enough.  If an error occurs or the connection times out after we
already received some data, return a short read, under the assumption
that the next call will fail or time out before we read anything.

As it turns out, none of the code that calls fetch_read() assumes an
all-or-nothing result anyway, except for a couple of lines where we
read the CR LF at the end of a hunk in HTTP hunked encoding, so the
changes outside of fetch_read() and http_readfn() are minimal.

While there, replace select(2) with poll(2).

MFC after:	3 days

1 files changed, 40 insertions, 93 deletions

diff --git a/lib/libfetch/common.c b/lib/libfetch/common.c
index fbacb02..ab81b58 100644
--- a/lib/libfetch/common.c
+++ b/lib/libfetch/common.c
@@ -41,6 +41,7 @@ __FBSDID("$FreeBSD$");
 #include <errno.h>
 #include <fcntl.h>
 #include <netdb.h>
+#include <poll.h>
 #include <pwd.h>
 #include <stdarg.h>
 #include <stdlib.h>
@@ -641,7 +642,7 @@ fetch_ssl_verify_hname(X509 *cert, const char *host)
 	struct addrinfo *ip;
 	STACK_OF(GENERAL_NAME) *altnames;
 	X509_NAME *subject;
-	int ret;	
+	int ret;
 
 	ret = 0;
 	ip = fetch_ssl_get_numeric_addrinfo(host, strlen(host));
@@ -913,33 +914,6 @@ fetch_ssl_read(SSL *ssl, char *buf, size_t len)
 }
 #endif
 
-/*
- * Cache some data that was read from a socket but cannot be immediately
- * returned because of an interrupted system call.
- */
-static int
-fetch_cache_data(conn_t *conn, char *src, size_t nbytes)
-{
-	char *tmp;
-
-	if (conn->cache.size < nbytes) {
-		tmp = realloc(conn->cache.buf, nbytes);
-		if (tmp == NULL) {
-			fetch_syserr();
-			return (-1);
-		}
-		conn->cache.buf = tmp;
-		conn->cache.size = nbytes;
-	}
-
-	memcpy(conn->cache.buf, src, nbytes);
-	conn->cache.len = nbytes;
-	conn->cache.pos = 0;
-
-	return (0);
-}
-
-
 static ssize_t
 fetch_socket_read(int sd, char *buf, size_t len)
 {
@@ -962,46 +936,31 @@ ssize_t
 fetch_read(conn_t *conn, char *buf, size_t len)
 {
 	struct timeval now, timeout, delta;
-	fd_set readfds;
-	ssize_t rlen, total;
-	char *start;
+	struct pollfd pfd;
+	ssize_t rlen;
+	int deltams;
 
 	if (fetchTimeout > 0) {
 		gettimeofday(&timeout, NULL);
 		timeout.tv_sec += fetchTimeout;
 	}
 
-	total = 0;
-	start = buf;
+	deltams = INFTIM;
+	memset(&pfd, 0, sizeof pfd);
+	pfd.fd = conn->sd;
+	pfd.events = POLLIN | POLLERR;
 
-	if (conn->cache.len > 0) {
-		/*
-		 * The last invocation of fetch_read was interrupted by a
-		 * signal after some data had been read from the socket. Copy
-		 * the cached data into the supplied buffer before trying to
-		 * read from the socket again.
-		 */
-		total = (conn->cache.len < len) ? conn->cache.len : len;
-		memcpy(buf, conn->cache.buf, total);
-
-		conn->cache.len -= total;
-		conn->cache.pos += total;
-		len -= total;
-		buf += total;
-	}
-
-	while (len > 0) {
+	for (;;) {
 		/*
 		 * The socket is non-blocking.  Instead of the canonical
-		 * select() -> read(), we do the following:
+		 * poll() -> read(), we do the following:
 		 *
 		 * 1) call read() or SSL_read().
-		 * 2) if an error occurred, return -1.
-		 * 3) if we received data but we still expect more,
-		 *    update our counters and loop.
+		 * 2) if we received some data, return it.
+		 * 3) if an error occurred, return -1.
 		 * 4) if read() or SSL_read() signaled EOF, return.
 		 * 5) if we did not receive any data but we're not at EOF,
-		 *    call select().
+		 *    call poll().
 		 *
 		 * In the SSL case, this is necessary because if we
 		 * receive a close notification, we have to call
@@ -1017,46 +976,34 @@ fetch_read(conn_t *conn, char *buf, size_t len)
 		else
 #endif
 			rlen = fetch_socket_read(conn->sd, buf, len);
-		if (rlen == 0) {
+		if (rlen > 0) {
 			break;
-		} else if (rlen > 0) {
-			len -= rlen;
-			buf += rlen;
-			total += rlen;
-			continue;
 		} else if (rlen == FETCH_READ_ERROR) {
 			if (errno == EINTR)
-				fetch_cache_data(conn, start, total);
+				break;
 			return (-1);
 		}
-		// assert(rlen == FETCH_READ_WAIT);
-		FD_ZERO(&readfds);
-		while (!FD_ISSET(conn->sd, &readfds)) {
-			FD_SET(conn->sd, &readfds);
-			if (fetchTimeout > 0) {
-				gettimeofday(&now, NULL);
-				if (!timercmp(&timeout, &now, >)) {
-					errno = ETIMEDOUT;
-					fetch_syserr();
-					return (-1);
-				}
-				timersub(&timeout, &now, &delta);
-			}
-			errno = 0;
-			if (select(conn->sd + 1, &readfds, NULL, NULL,
-				fetchTimeout > 0 ? &delta : NULL) < 0) {
-				if (errno == EINTR) {
-					if (fetchRestartCalls)
-						continue;
-					/* Save anything that was read. */
-					fetch_cache_data(conn, start, total);
-				}
+		if (fetchTimeout > 0) {
+			gettimeofday(&now, NULL);
+			if (!timercmp(&timeout, &now, >)) {
+				errno = ETIMEDOUT;
 				fetch_syserr();
 				return (-1);
 			}
+			timersub(&timeout, &now, &delta);
+			deltams = delta.tv_sec * 1000 +
+			    delta.tv_usec / 1000;;
+		}
+		errno = 0;
+		pfd.revents = 0;
+		if (poll(&pfd, 1, deltams) < 0) {
+			if (errno == EINTR && fetchRestartCalls)
+				continue;
+			fetch_syserr();
+			return (-1);
 		}
 	}
-	return (total);
+	return (rlen);
 }
 
 
@@ -1130,20 +1077,21 @@ ssize_t
 fetch_writev(conn_t *conn, struct iovec *iov, int iovcnt)
 {
 	struct timeval now, timeout, delta;
-	fd_set writefds;
+	struct pollfd pfd;
 	ssize_t wlen, total;
-	int r;
+	int deltams, r;
 
+	memset(&pfd, 0, sizeof pfd);
 	if (fetchTimeout) {
-		FD_ZERO(&writefds);
+		pfd.fd = conn->sd;
+		pfd.events = POLLOUT | POLLERR;
 		gettimeofday(&timeout, NULL);
 		timeout.tv_sec += fetchTimeout;
 	}
 
 	total = 0;
 	while (iovcnt > 0) {
-		while (fetchTimeout && !FD_ISSET(conn->sd, &writefds)) {
-			FD_SET(conn->sd, &writefds);
+		while (fetchTimeout && pfd.revents == 0) {
 			gettimeofday(&now, NULL);
 			delta.tv_sec = timeout.tv_sec - now.tv_sec;
 			delta.tv_usec = timeout.tv_usec - now.tv_usec;
@@ -1156,9 +1104,9 @@ fetch_writev(conn_t *conn, struct iovec *iov, int iovcnt)
 				fetch_syserr();
 				return (-1);
 			}
+			deltams = delta.tv_sec * 1000 + delta.tv_usec / 1000;;
 			errno = 0;
-			r = select(conn->sd + 1, NULL, &writefds, NULL, &delta);
-			if (r == -1) {
+			if ((r = poll(&pfd, 1, deltams)) == -1) {
 				if (errno == EINTR && fetchRestartCalls)
 					continue;
 				return (-1);
@@ -1250,7 +1198,6 @@ fetch_close(conn_t *conn)
 	}
 #endif
 	ret = close(conn->sd);
-	free(conn->cache.buf);
 	free(conn->buf);
 	free(conn);
 	return (ret);