Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client
Pull Ceph fixes from Sage Weil:
"Just two small items from Ilya:
The first patch fixes the RBD readahead to grab full objects. The
second fixes the write ops to prevent undue promotion when a cache
tier is configured on the server side"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
rbd: use writefull op for object size writes
rbd: set max_sectors explicitly
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index d93a037..f5e49b6 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -1863,9 +1863,11 @@
rbd_osd_read_callback(obj_request);
break;
case CEPH_OSD_OP_SETALLOCHINT:
- rbd_assert(osd_req->r_ops[1].op == CEPH_OSD_OP_WRITE);
+ rbd_assert(osd_req->r_ops[1].op == CEPH_OSD_OP_WRITE ||
+ osd_req->r_ops[1].op == CEPH_OSD_OP_WRITEFULL);
/* fall through */
case CEPH_OSD_OP_WRITE:
+ case CEPH_OSD_OP_WRITEFULL:
rbd_osd_write_callback(obj_request);
break;
case CEPH_OSD_OP_STAT:
@@ -2401,7 +2403,10 @@
opcode = CEPH_OSD_OP_ZERO;
}
} else if (op_type == OBJ_OP_WRITE) {
- opcode = CEPH_OSD_OP_WRITE;
+ if (!offset && length == object_size)
+ opcode = CEPH_OSD_OP_WRITEFULL;
+ else
+ opcode = CEPH_OSD_OP_WRITE;
osd_req_op_alloc_hint_init(osd_request, num_ops,
object_size, object_size);
num_ops++;
@@ -3760,6 +3765,7 @@
/* set io sizes to object size */
segment_size = rbd_obj_bytes(&rbd_dev->header);
blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
+ q->limits.max_sectors = queue_max_hw_sectors(q);
blk_queue_max_segments(q, segment_size / SECTOR_SIZE);
blk_queue_max_segment_size(q, segment_size);
blk_queue_io_min(q, segment_size);
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 80b94e3..f79ccac 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -285,6 +285,7 @@
switch (op->op) {
case CEPH_OSD_OP_READ:
case CEPH_OSD_OP_WRITE:
+ case CEPH_OSD_OP_WRITEFULL:
ceph_osd_data_release(&op->extent.osd_data);
break;
case CEPH_OSD_OP_CALL:
@@ -485,13 +486,14 @@
size_t payload_len = 0;
BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
- opcode != CEPH_OSD_OP_ZERO && opcode != CEPH_OSD_OP_TRUNCATE);
+ opcode != CEPH_OSD_OP_WRITEFULL && opcode != CEPH_OSD_OP_ZERO &&
+ opcode != CEPH_OSD_OP_TRUNCATE);
op->extent.offset = offset;
op->extent.length = length;
op->extent.truncate_size = truncate_size;
op->extent.truncate_seq = truncate_seq;
- if (opcode == CEPH_OSD_OP_WRITE)
+ if (opcode == CEPH_OSD_OP_WRITE || opcode == CEPH_OSD_OP_WRITEFULL)
payload_len += length;
op->payload_len = payload_len;
@@ -670,9 +672,11 @@
break;
case CEPH_OSD_OP_READ:
case CEPH_OSD_OP_WRITE:
+ case CEPH_OSD_OP_WRITEFULL:
case CEPH_OSD_OP_ZERO:
case CEPH_OSD_OP_TRUNCATE:
- if (src->op == CEPH_OSD_OP_WRITE)
+ if (src->op == CEPH_OSD_OP_WRITE ||
+ src->op == CEPH_OSD_OP_WRITEFULL)
request_data_len = src->extent.length;
dst->extent.offset = cpu_to_le64(src->extent.offset);
dst->extent.length = cpu_to_le64(src->extent.length);
@@ -681,7 +685,8 @@
dst->extent.truncate_seq =
cpu_to_le32(src->extent.truncate_seq);
osd_data = &src->extent.osd_data;
- if (src->op == CEPH_OSD_OP_WRITE)
+ if (src->op == CEPH_OSD_OP_WRITE ||
+ src->op == CEPH_OSD_OP_WRITEFULL)
ceph_osdc_msg_data_add(req->r_request, osd_data);
else
ceph_osdc_msg_data_add(req->r_reply, osd_data);