From: Sage Weil <sage@redhat.com>
Date: Wed, 23 Nov 2016 22:41:44 +0000 (-0500)
Subject: os/bluestore: set csum order based on alloc hints
X-Git-Tag: v11.1.0~127^2~1
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=ae772b8d0d9fc0186759b9e2051633ff1127b2b0;p=ceph.git

os/bluestore: set csum order based on alloc hints

If the users hints that this object is sequentially read and written we
can use a larger csum order.

Signed-off-by: Sage Weil <sage@redhat.com>
---

diff --git a/src/os/bluestore/BlueStore.cc b/src/os/bluestore/BlueStore.cc
index ace6ee33560dd..31075d4969b9e 100644
--- a/src/os/bluestore/BlueStore.cc
+++ b/src/os/bluestore/BlueStore.cc
@@ -8049,7 +8049,8 @@ int BlueStore::_do_write(
 			CEPH_OSD_ALLOC_HINT_FLAG_APPEND_ONLY)) &&
       (alloc_hints & CEPH_OSD_ALLOC_HINT_FLAG_RANDOM_WRITE) == 0) {
     dout(20) << __func__ << " will prefer large blob and csum sizes" << dendl;
-    wctx.csum_order = min_alloc_size_order;
+    wctx.csum_order = std::max(min_alloc_size_order,
+			       (size_t)ctzl(o->onode.expected_write_size));
     if (wctx.compress) {
       wctx.target_blob_size = select_option(
         "compression_max_blob_size",