From 98453c81afc905b0713ffd1cd53ee3aadaccf610 Mon Sep 17 00:00:00 2001 From: Maxim Sobolev Date: Sat, 4 Jun 2011 16:01:30 +0000 Subject: [PATCH] Read from the socket using the same max buffer size as we use while sending. What happens otherwise is that the sender splits all the traffic into 32k chunks, while the receiver is waiting for the whole packet. Then for a certain packet sizes, particularly 66607 bytes in my case, the communication stucks to secondary is expecting to read one chunk of 66607 bytes, while primary is sending two chunks of 32768 bytes and third chunk of 1071. Probably due to TCP windowing and buffering the final chunk gets stuck somewhere, so neither server not client can make any progress. This patch also protect from short reads, as according to the manual page there are some cases when MSG_WAITALL can give less data than expected. MFC after: 3 days --- sbin/hastd/proto_common.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/sbin/hastd/proto_common.c b/sbin/hastd/proto_common.c index 183b4e4fdcd2..1f541da7576c 100644 --- a/sbin/hastd/proto_common.c +++ b/sbin/hastd/proto_common.c @@ -194,6 +194,8 @@ int proto_common_recv(int sock, unsigned char *data, size_t size, int *fdp) { ssize_t done; + size_t total_done, recvsize; + unsigned char *dp; PJDLOG_ASSERT(sock >= 0); @@ -210,9 +212,19 @@ proto_common_recv(int sock, unsigned char *data, size_t size, int *fdp) PJDLOG_ASSERT(data != NULL); PJDLOG_ASSERT(size > 0); + total_done = 0; + dp = data; do { - done = recv(sock, data, size, MSG_WAITALL); - } while (done == -1 && errno == EINTR); + recvsize = size - total_done; + recvsize = recvsize < MAX_SEND_SIZE ? recvsize : MAX_SEND_SIZE; + done = recv(sock, dp, recvsize, MSG_WAITALL); + if (done == -1 && errno == EINTR) + continue; + if (done <= 0) + break; + total_done += done; + dp += done; + } while (total_done < size); if (done == 0) { return (ENOTCONN); } else if (done < 0) {