]> git.openfabrics.org - ~shefty/librdmacm.git/commitdiff
new
authorSean Hefty <sean.hefty@intel.com>
Fri, 27 Jul 2012 17:46:42 +0000 (10:46 -0700)
committerSean Hefty <sean.hefty@intel.com>
Fri, 27 Jul 2012 17:46:42 +0000 (10:46 -0700)
meta
patches/fast-disc [new file with mode: 0644]

diff --git a/meta b/meta
index ea221b225d6e91ded3cee9fe30c58fb5264ddd43..152fb14d0934fdca08daf2e33e10ac222fbc5ef2 100644 (file)
--- a/meta
+++ b/meta
@@ -1,6 +1,6 @@
 Version: 1
-Previous: 590dcd2580933003e4fb7816ba23343bfcedf385
-Head: 59cddb8c2cd143ffe205716eef276b9c502e1eb5
+Previous: f6ea75d1fe6bed58f49784f07cca5bf655ed6cf6
+Head: 8ba3be5f12ffff6818c394d0b0589cb713f68841
 Applied:
   cma-rm-pd: 2ffda7f2991395570b9e776ff5ae256ca9684771
   transpose: 3e52eb22f44eafaefa95c4674bc5665a94e15694
@@ -9,6 +9,7 @@ Applied:
   init-getname: 7d988863b218d1b66e3739ec4b6f51acc72b2334
   rs-ftp: 28e0744eb89227fbeded485fbad64010b9edf0f6
   check-id: 59cddb8c2cd143ffe205716eef276b9c502e1eb5
+  fast-disc: 8ba3be5f12ffff6818c394d0b0589cb713f68841
 Unapplied:
   dbg: 0c269855776d3001e37da8c8afe283c20e1d6cd6
   waitall-buggy: c49c6b56c55385774065f5aa2704078e6ae0ceb8
diff --git a/patches/fast-disc b/patches/fast-disc
new file mode 100644 (file)
index 0000000..aa968a2
--- /dev/null
@@ -0,0 +1,23 @@
+Bottom: b5df9cc04d6e7629ae937bfe9bbc3c7da5193ea0
+Top:    b5df9cc04d6e7629ae937bfe9bbc3c7da5193ea0
+Author: Sean Hefty <sean.hefty@intel.com>
+Date:   2012-07-27 10:46:42 -0700
+
+rsocket: Improve disconnect time under normal conditions
+
+When both sides of a connection attempt to close at the same
+time, one of the two sides can easily get an error when sending
+a disconnect message.  This results in that side hanging
+during close until the send times out.  (The time out is caused
+by the remote side destroying its QP.)
+
+We can reduce the chance of this occurring by immediately
+assuming that the disconnect has been successful once we've
+received the remote side's disconnect message.
+
+Signed-off-by: Sean Hefty <sean.hefty@intel.com>
+
+
+---
+
+