From: Yan, Zheng <zyan@redhat.com>
Date: Fri, 18 Dec 2015 14:04:08 +0000 (+0800)
Subject: msg/simple/Pipe: avoid infinite loop in Pipe::do_recv()
X-Git-Tag: v11.0.0~668
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=63e44e32974c9bae17bb1bfd4261dcb024ad845c;p=ceph.git

msg/simple/Pipe: avoid infinite loop in Pipe::do_recv()

I found that, sometimes an OSD thread uses 100% CPU after cutting network
between OSD and client. recv(2) in Pipe::do_recv() keeps returning -EAGAIN,
which causes infinite loop. the call trace is:

 Pipe::do_recv (...)
 Pipe::buffered_recv (...)
 Pipe::tcp_read_nonblocking (...)
 Pipe::tcp_read (...)

Pipe::tcp_read() first calls Pipe::tcp_read_wait() to check if data is
avaliable. If there are prefetched data, Pipe::tcp_read_wait() return
immediately. Pipe::buffered_recv() is called, which reads data from the
prefetched data. If prefetched data isn't enough, Pipe::buffered_recv()
calls Pipe::do_recv() to read data from socket. But it's possble that
socket has no data at this time, so Pipe::do_recv() keeps retry.

The fix is simple, just not retry when recv(2) return -EAGAIN.

Fixes: #14120
Signed-off-by: Yan, Zheng <zyan@redhat.com>
---

diff --git a/src/msg/simple/Pipe.cc b/src/msg/simple/Pipe.cc
index 74efa72015c0..53781b82b173 100644
--- a/src/msg/simple/Pipe.cc
+++ b/src/msg/simple/Pipe.cc
@@ -2513,7 +2513,7 @@ ssize_t Pipe::do_recv(char *buf, size_t len, int flags)
 again:
   ssize_t got = ::recv( sd, buf, len, flags );
   if (got < 0) {
-    if (errno == EAGAIN || errno == EINTR) {
+    if (errno == EINTR) {
       goto again;
     }
     ldout(msgr->cct, 10) << __func__ << " socket " << sd << " returned "