From: Sage Weil <sage.weil@dreamhost.com>
Date: Tue, 26 Apr 2011 19:10:12 +0000 (-0700)
Subject: osd: add RWORDERED osd op flag
X-Git-Tag: v0.27.1~7
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=b08ee2c6341a1a52356972f8e48bd8b3de9136bf;p=ceph.git

osd: add RWORDERED osd op flag

Order this op wrt reads the same way a read-modify-write would be.
(Otherwise we may get a fast/stale read result on a not-yet-complete
write.)

This fixes a problem where the Filer was marking a probe stat as a write
to get this same effect, but the OSD would EINVAL if it was a snapped
object (which happens in certain cases where the MDS is recovering the
file size of a snapped file).

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

diff --git a/src/include/rados.h b/src/include/rados.h
index 0ed075f553c3..09dc3e097dd1 100644
--- a/src/include/rados.h
+++ b/src/include/rados.h
@@ -326,6 +326,7 @@ enum {
 	CEPH_OSD_FLAG_EXEC =           0x0800,  /* op may exec */
 	CEPH_OSD_FLAG_EXEC_PUBLIC =    0x1000,  /* op may exec (public) */
 	CEPH_OSD_FLAG_LOCALIZE_READS = 0x2000,  /* read from nearby replica, if any */
+	CEPH_OSD_FLAG_RWORDERED =      0x4000,  /* order wrt concurrent reads */
 };
 
 enum {
diff --git a/src/osd/ReplicatedPG.cc b/src/osd/ReplicatedPG.cc
index 596ecb047a73..ba76b0512b81 100644
--- a/src/osd/ReplicatedPG.cc
+++ b/src/osd/ReplicatedPG.cc
@@ -397,7 +397,8 @@ void ReplicatedPG::do_op(MOSDOp *op)
   bool ok;
   dout(10) << "do_op mode is " << mode << dendl;
   assert(!mode.wake);   // we should never have woken waiters here.
-  if (op->may_read() && op->may_write())
+  if ((op->may_read() && op->may_write()) ||
+      (op->get_flags() & CEPH_OSD_FLAG_RWORDERED))
     ok = mode.try_rmw(client);
   else if (op->may_write())
     ok = mode.try_write(client);
diff --git a/src/osdc/Filer.cc b/src/osdc/Filer.cc
index 4690ce9b6e88..335a879fd0a2 100644
--- a/src/osdc/Filer.cc
+++ b/src/osdc/Filer.cc
@@ -108,7 +108,7 @@ void Filer::_probe(Probe *probe)
     dout(10) << "_probe  probing " << p->oid << dendl;
     C_Probe *c = new C_Probe(this, probe, p->oid);
     probe->ops[p->oid] = objecter->stat(p->oid, p->oloc, probe->snapid, &c->size, &c->mtime, 
-					probe->flags | CEPH_OSD_FLAG_WRITE, c);
+					probe->flags | CEPH_OSD_FLAG_RWORDERED, c);
   }
 }