From 966b4ec5e5cc2a6f49b7559461eb99a61e9ec227 Mon Sep 17 00:00:00 2001
From: Aaron Myers <atm@apache.org>
Date: Fri, 17 May 2013 00:11:11 +0000
Subject: [PATCH] HADOOP-9566. Performing direct read using libhdfs sometimes
 raises SIGPIPE (which in turn throws SIGABRT) causing client crashes.
 Contributed by Colin Patrick McCabe.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1483613 13f79535-47bb-0310-9956-ffa450edef68
---
 .../hadoop-common/CHANGES.txt                 |  4 +++
 .../org/apache/hadoop/net/unix/DomainSocket.c | 30 ++++++++++++++++---
 2 files changed, 30 insertions(+), 4 deletions(-)

diff --git a/hadoop-common-project/hadoop-common/CHANGES.txt b/hadoop-common-project/hadoop-common/CHANGES.txt
index 36526b3cf3b..0f659b54b09 100644
--- a/hadoop-common-project/hadoop-common/CHANGES.txt
+++ b/hadoop-common-project/hadoop-common/CHANGES.txt
@@ -185,6 +185,10 @@ Release 2.0.5-beta - UNRELEASED
     HADOOP-9563. Fix incompatibility introduced by HADOOP-9523.
     (Tian Hong Wang via suresh)
 
+    HADOOP-9566. Performing direct read using libhdfs sometimes raises SIGPIPE
+    (which in turn throws SIGABRT) causing client crashes. (Colin Patrick
+    McCabe via atm)
+
 Release 2.0.4-alpha - 2013-04-25 
 
   INCOMPATIBLE CHANGES
diff --git a/hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/net/unix/DomainSocket.c b/hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/net/unix/DomainSocket.c
index 1c2ea698a00..cf34c525323 100644
--- a/hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/net/unix/DomainSocket.c
+++ b/hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/net/unix/DomainSocket.c
@@ -16,8 +16,7 @@
  * limitations under the License.
  */
 
-#define _GNU_SOURCE
-
+#include "config.h"
 #include "exception.h"
 #include "org/apache/hadoop/io/nativeio/file_descriptor.h"
 #include "org_apache_hadoop.h"
@@ -31,6 +30,7 @@
 #include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
+#include <string.h>
 #include <sys/ioctl.h> /* for FIONREAD */
 #include <sys/socket.h>
 #include <sys/stat.h>
@@ -47,6 +47,15 @@
 #define DEFAULT_SEND_TIMEOUT 120000
 #define LISTEN_BACKLOG 128
 
+/* In Linux, you can pass the MSG_NOSIGNAL flag to send, sendto, etc. to prevent
+ * those functions from generating SIGPIPE.  HDFS-4831 for details.
+ */
+#ifdef MSG_NOSIGNAL
+#define PLATFORM_SEND_FLAGS MSG_NOSIGNAL
+#else
+#define PLATFORM_SEND_FLAGS 0
+#endif
+
 /**
  * Can't pass more than this number of file descriptors in a single message.
  */
@@ -176,6 +185,19 @@ static jthrowable setup(JNIEnv *env, int *ofd, jobject jpath, int doConnect)
         "is %zd bytes.", sizeof(addr.sun_path) - 1);
     goto done;
   }
+#ifdef SO_NOSIGPIPE
+  /* On MacOS and some BSDs, SO_NOSIGPIPE will keep send and sendto from causing
+   * EPIPE.  Note: this will NOT help when using write or writev, only with
+   * send, sendto, sendmsg, etc.  See HDFS-4831.
+   */
+  ret = 1;
+  if (setsockopt(fd, SOL_SOCKET, SO_NOSIGPIPE, (void *)&ret, sizeof(ret))) {
+    ret = errno;
+    jthr = newSocketException(env, ret,
+        "error setting SO_NOSIGPIPE on socket: error %s", terror(ret));
+    goto done;
+  }
+#endif
   if (doConnect) {
     RETRY_ON_EINTR(ret, connect(fd, 
         (struct sockaddr*)&addr, sizeof(addr)));
@@ -583,7 +605,7 @@ static jthrowable write_fully(JNIEnv *env, int fd, int8_t *buf, int amt)
   int err, res;
 
   while (amt > 0) {
-    res = write(fd, buf, amt);
+    res = send(fd, buf, amt, PLATFORM_SEND_FLAGS);
     if (res < 0) {
       err = errno;
       if (err == EINTR) {
@@ -685,7 +707,7 @@ jint offset, jint length)
       goto done;
     }
   }
-  RETRY_ON_EINTR(ret, sendmsg(fd, &socketMsg, 0));
+  RETRY_ON_EINTR(ret, sendmsg(fd, &socketMsg, PLATFORM_SEND_FLAGS));
   if (ret < 0) {
     ret = errno;
     jthr = newSocketException(env, ret, "sendmsg(2) error: %s", terror(ret));