From: Sean Hefty <sean.hefty@intel.com>
Date: Sat, 26 May 2012 07:02:47 +0000 (-0700)
Subject: new
X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=4ac6e78f308507d050e4be8071f1197b52661ce5;p=~shefty%2Flibrdmacm.git

new
---

diff --git a/meta b/meta
index 1af85bfc..21563d55 100644
--- a/meta
+++ b/meta
@@ -1,6 +1,6 @@
 Version: 1
-Previous: 8e67fb9509f9aee8102f4ab7e75937764849883f
-Head: b6f44cc4d1cce1b5580696e1fa60245b83c9382c
+Previous: 619f8a33a9b674174e7d9fe3c1d4a8148d28e37d
+Head: ffa5cddcb740eb5e392038beeec086858082ba0f
 Applied:
   destroy_cqs: 66410c265b2661e16b7f6126bcc4ea5c7e1af1bd
   conn-err: ac46465ae6061b06f005ba9b49b371c9eb0d2325
@@ -17,6 +17,7 @@ Applied:
   rs-with-acm: fadeaad64ff2d89580b571bee92bf9ecfa7b7396
   pre-socket: 8cb5871f6da8f40d9315e4dc57fdbeb9c3ae6de7
   rs-states: b6f44cc4d1cce1b5580696e1fa60245b83c9382c
+  rs-test-nonblock: ffa5cddcb740eb5e392038beeec086858082ba0f
 Unapplied:
   rs-recv-hang: ed95ce9c3c8c5a2397f546b0a538d8ecd22ed41f
 Hidden:
diff --git a/patches/rs-test-nonblock b/patches/rs-test-nonblock
new file mode 100644
index 00000000..f6917f36
--- /dev/null
+++ b/patches/rs-test-nonblock
@@ -0,0 +1,27 @@
+Bottom: 21cab4e0e2179a70e55cee483fbfdabf4510108a
+Top:    21cab4e0e2179a70e55cee483fbfdabf4510108a
+Author: Sean Hefty <sean.hefty@intel.com>
+Date:   2012-05-26 00:02:47 -0700
+
+rsocket: Merge nonblock test with test() routine in rs_process_cq
+
+rs_process_cq takes the following 2 parameters: nonblock and test().
+These are used to control the operation of rs_process_cq.  If
+nonblock is true, rs_process_cq will exit without arming the CQ or
+waiting on a CQ event.  rs_process_cq() will also exit if test()
+returns true.  The only difference in the operation is the return
+value that rs_process_cq() returns.
+
+We can simplify the code by merging the nonblock test into the
+caller's provided test() routine.  The test() routine simply needs
+to return the correct value for rs_process_cq().  This will also
+simplify fixing an issue where a caller may block indefinitely
+in send() or recv() after an rsocket has been disconnected.  That
+fix is in a subsequent patch.
+
+Signed-off-by: Sean Hefty <sean.hefty@intel.com>
+
+
+---
+
+