2010-04-07 00:14:15 +02:00
|
|
|
#include <linux/ceph/ceph_debug.h>
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-04-07 00:14:15 +02:00
|
|
|
#include <linux/module.h>
|
2009-10-06 20:31:10 +02:00
|
|
|
#include <linux/err.h>
|
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/uaccess.h>
|
2010-04-07 00:01:27 +02:00
|
|
|
#ifdef CONFIG_BLOCK
|
|
|
|
#include <linux/bio.h>
|
|
|
|
#endif
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-04-07 00:14:15 +02:00
|
|
|
#include <linux/ceph/libceph.h>
|
|
|
|
#include <linux/ceph/osd_client.h>
|
|
|
|
#include <linux/ceph/messenger.h>
|
|
|
|
#include <linux/ceph/decode.h>
|
|
|
|
#include <linux/ceph/auth.h>
|
|
|
|
#include <linux/ceph/pagelist.h>
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-03-01 22:02:00 +01:00
|
|
|
#define OSD_OP_FRONT_LEN 4096
|
|
|
|
#define OSD_OPREPLY_FRONT_LEN 512
|
2010-01-14 02:03:23 +01:00
|
|
|
|
2010-05-20 10:40:19 +02:00
|
|
|
static const struct ceph_connection_operations osd_con_ops;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
static void send_queued(struct ceph_osd_client *osdc);
|
|
|
|
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
static int op_needs_trail(int op)
|
|
|
|
{
|
|
|
|
switch (op) {
|
|
|
|
case CEPH_OSD_OP_GETXATTR:
|
|
|
|
case CEPH_OSD_OP_SETXATTR:
|
|
|
|
case CEPH_OSD_OP_CMPXATTR:
|
|
|
|
case CEPH_OSD_OP_CALL:
|
|
|
|
return 1;
|
|
|
|
default:
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int op_has_extent(int op)
|
|
|
|
{
|
|
|
|
return (op == CEPH_OSD_OP_READ ||
|
|
|
|
op == CEPH_OSD_OP_WRITE);
|
|
|
|
}
|
|
|
|
|
2010-04-06 23:51:47 +02:00
|
|
|
void ceph_calc_raw_layout(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_file_layout *layout,
|
|
|
|
u64 snapid,
|
2010-04-07 00:01:27 +02:00
|
|
|
u64 off, u64 *plen, u64 *bno,
|
|
|
|
struct ceph_osd_request *req,
|
|
|
|
struct ceph_osd_req_op *op)
|
2010-04-06 23:51:47 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
|
2010-04-07 00:01:27 +02:00
|
|
|
u64 orig_len = *plen;
|
2010-04-06 23:51:47 +02:00
|
|
|
u64 objoff, objlen; /* extent in object */
|
|
|
|
|
|
|
|
reqhead->snapid = cpu_to_le64(snapid);
|
|
|
|
|
|
|
|
/* object extent? */
|
2010-04-07 00:01:27 +02:00
|
|
|
ceph_calc_file_object_mapping(layout, off, plen, bno,
|
2010-04-06 23:51:47 +02:00
|
|
|
&objoff, &objlen);
|
2010-04-07 00:01:27 +02:00
|
|
|
if (*plen < orig_len)
|
2010-04-06 23:51:47 +02:00
|
|
|
dout(" skipping last %llu, final file extent %llu~%llu\n",
|
2010-04-07 00:01:27 +02:00
|
|
|
orig_len - *plen, off, *plen);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
if (op_has_extent(op->op)) {
|
|
|
|
op->extent.offset = objoff;
|
|
|
|
op->extent.length = objlen;
|
|
|
|
}
|
|
|
|
req->r_num_pages = calc_pages_for(off, *plen);
|
2010-11-09 21:43:12 +01:00
|
|
|
req->r_page_alignment = off & ~PAGE_MASK;
|
2010-04-07 00:14:15 +02:00
|
|
|
if (op->op == CEPH_OSD_OP_WRITE)
|
|
|
|
op->payload_len = *plen;
|
2010-04-06 23:51:47 +02:00
|
|
|
|
|
|
|
dout("calc_layout bno=%llx %llu~%llu (%d pages)\n",
|
|
|
|
*bno, objoff, objlen, req->r_num_pages);
|
|
|
|
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_calc_raw_layout);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
/*
|
|
|
|
* Implement client access to distributed object storage cluster.
|
|
|
|
*
|
|
|
|
* All data objects are stored within a cluster/cloud of OSDs, or
|
|
|
|
* "object storage devices." (Note that Ceph OSDs have _nothing_ to
|
|
|
|
* do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply
|
|
|
|
* remote daemons serving up and coordinating consistent and safe
|
|
|
|
* access to storage.
|
|
|
|
*
|
|
|
|
* Cluster membership and the mapping of data objects onto storage devices
|
|
|
|
* are described by the osd map.
|
|
|
|
*
|
|
|
|
* We keep track of pending OSD requests (read, write), resubmit
|
|
|
|
* requests to different OSDs when the cluster topology/data layout
|
|
|
|
* change, or retry the affected requests when the communications
|
|
|
|
* channel with an OSD is reset.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* calculate the mapping of a file extent onto an object, and fill out the
|
|
|
|
* request accordingly. shorten extent as necessary if it crosses an
|
|
|
|
* object boundary.
|
|
|
|
*
|
|
|
|
* fill osd op in request message.
|
|
|
|
*/
|
|
|
|
static void calc_layout(struct ceph_osd_client *osdc,
|
2010-04-06 23:51:47 +02:00
|
|
|
struct ceph_vino vino,
|
|
|
|
struct ceph_file_layout *layout,
|
2009-10-06 20:31:10 +02:00
|
|
|
u64 off, u64 *plen,
|
2010-04-07 00:01:27 +02:00
|
|
|
struct ceph_osd_request *req,
|
|
|
|
struct ceph_osd_req_op *op)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
u64 bno;
|
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
ceph_calc_raw_layout(osdc, layout, vino.snap, off,
|
|
|
|
plen, &bno, req, op);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno);
|
|
|
|
req->r_oid_len = strlen(req->r_oid);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* requests
|
|
|
|
*/
|
2009-12-07 22:37:03 +01:00
|
|
|
void ceph_osdc_release_request(struct kref *kref)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
2009-12-07 22:37:03 +01:00
|
|
|
struct ceph_osd_request *req = container_of(kref,
|
|
|
|
struct ceph_osd_request,
|
|
|
|
r_kref);
|
|
|
|
|
|
|
|
if (req->r_request)
|
|
|
|
ceph_msg_put(req->r_request);
|
|
|
|
if (req->r_reply)
|
|
|
|
ceph_msg_put(req->r_reply);
|
2010-01-14 02:03:23 +01:00
|
|
|
if (req->r_con_filling_msg) {
|
2009-12-22 19:45:45 +01:00
|
|
|
dout("release_request revoking pages %p from con %p\n",
|
2010-01-14 02:03:23 +01:00
|
|
|
req->r_pages, req->r_con_filling_msg);
|
|
|
|
ceph_con_revoke_message(req->r_con_filling_msg,
|
|
|
|
req->r_reply);
|
|
|
|
ceph_con_put(req->r_con_filling_msg);
|
2009-12-22 19:45:45 +01:00
|
|
|
}
|
2009-12-07 22:37:03 +01:00
|
|
|
if (req->r_own_pages)
|
|
|
|
ceph_release_page_vector(req->r_pages,
|
|
|
|
req->r_num_pages);
|
2010-04-07 00:01:27 +02:00
|
|
|
#ifdef CONFIG_BLOCK
|
|
|
|
if (req->r_bio)
|
|
|
|
bio_put(req->r_bio);
|
|
|
|
#endif
|
2009-12-07 22:37:03 +01:00
|
|
|
ceph_put_snap_context(req->r_snapc);
|
2010-04-07 00:01:27 +02:00
|
|
|
if (req->r_trail) {
|
|
|
|
ceph_pagelist_release(req->r_trail);
|
|
|
|
kfree(req->r_trail);
|
|
|
|
}
|
2009-12-07 22:37:03 +01:00
|
|
|
if (req->r_mempool)
|
|
|
|
mempool_free(req, req->r_osdc->req_mempool);
|
|
|
|
else
|
|
|
|
kfree(req);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_release_request);
|
2010-04-07 00:01:27 +02:00
|
|
|
|
|
|
|
static int get_num_ops(struct ceph_osd_req_op *ops, int *needs_trail)
|
|
|
|
{
|
|
|
|
int i = 0;
|
|
|
|
|
|
|
|
if (needs_trail)
|
|
|
|
*needs_trail = 0;
|
|
|
|
while (ops[i].op) {
|
|
|
|
if (needs_trail && op_needs_trail(ops[i].op))
|
|
|
|
*needs_trail = 1;
|
|
|
|
i++;
|
|
|
|
}
|
|
|
|
|
|
|
|
return i;
|
|
|
|
}
|
|
|
|
|
2010-04-06 23:51:47 +02:00
|
|
|
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
|
|
|
|
int flags,
|
2009-10-06 20:31:10 +02:00
|
|
|
struct ceph_snap_context *snapc,
|
2010-04-07 00:01:27 +02:00
|
|
|
struct ceph_osd_req_op *ops,
|
2010-04-06 23:51:47 +02:00
|
|
|
bool use_mempool,
|
|
|
|
gfp_t gfp_flags,
|
2010-04-07 00:01:27 +02:00
|
|
|
struct page **pages,
|
|
|
|
struct bio *bio)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
struct ceph_msg *msg;
|
2010-04-07 00:01:27 +02:00
|
|
|
int needs_trail;
|
|
|
|
int num_op = get_num_ops(ops, &needs_trail);
|
|
|
|
size_t msg_size = sizeof(struct ceph_osd_request_head);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
msg_size += num_op*sizeof(struct ceph_osd_op);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
if (use_mempool) {
|
2010-04-06 23:51:47 +02:00
|
|
|
req = mempool_alloc(osdc->req_mempool, gfp_flags);
|
2009-10-06 20:31:10 +02:00
|
|
|
memset(req, 0, sizeof(*req));
|
|
|
|
} else {
|
2010-04-06 23:51:47 +02:00
|
|
|
req = kzalloc(sizeof(*req), gfp_flags);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
if (req == NULL)
|
2010-04-02 01:06:19 +02:00
|
|
|
return NULL;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
req->r_osdc = osdc;
|
|
|
|
req->r_mempool = use_mempool;
|
2010-04-07 00:01:27 +02:00
|
|
|
|
2009-12-07 22:37:03 +01:00
|
|
|
kref_init(&req->r_kref);
|
2009-10-06 20:31:10 +02:00
|
|
|
init_completion(&req->r_completion);
|
|
|
|
init_completion(&req->r_safe_completion);
|
|
|
|
INIT_LIST_HEAD(&req->r_unsafe_item);
|
|
|
|
req->r_flags = flags;
|
|
|
|
|
|
|
|
WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0);
|
|
|
|
|
2010-03-01 22:02:00 +01:00
|
|
|
/* create reply message */
|
|
|
|
if (use_mempool)
|
|
|
|
msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
|
|
|
|
else
|
|
|
|
msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
|
2010-04-06 23:51:47 +02:00
|
|
|
OSD_OPREPLY_FRONT_LEN, gfp_flags);
|
2010-04-02 01:06:19 +02:00
|
|
|
if (!msg) {
|
2010-03-01 22:02:00 +01:00
|
|
|
ceph_osdc_put_request(req);
|
2010-04-02 01:06:19 +02:00
|
|
|
return NULL;
|
2010-03-01 22:02:00 +01:00
|
|
|
}
|
|
|
|
req->r_reply = msg;
|
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
/* allocate space for the trailing data */
|
|
|
|
if (needs_trail) {
|
|
|
|
req->r_trail = kmalloc(sizeof(struct ceph_pagelist), gfp_flags);
|
|
|
|
if (!req->r_trail) {
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
ceph_pagelist_init(req->r_trail);
|
|
|
|
}
|
2010-03-01 22:02:00 +01:00
|
|
|
/* create request message; allow space for oid */
|
2009-10-06 20:31:10 +02:00
|
|
|
msg_size += 40;
|
|
|
|
if (snapc)
|
|
|
|
msg_size += sizeof(u64) * snapc->num_snaps;
|
|
|
|
if (use_mempool)
|
2009-10-15 02:36:07 +02:00
|
|
|
msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
|
2009-10-06 20:31:10 +02:00
|
|
|
else
|
2010-04-06 23:51:47 +02:00
|
|
|
msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags);
|
2010-04-02 01:06:19 +02:00
|
|
|
if (!msg) {
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_osdc_put_request(req);
|
2010-04-02 01:06:19 +02:00
|
|
|
return NULL;
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
2010-04-07 00:01:27 +02:00
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP);
|
|
|
|
memset(msg->front.iov_base, 0, msg->front.iov_len);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
|
|
|
req->r_request = msg;
|
|
|
|
req->r_pages = pages;
|
2010-04-07 00:01:27 +02:00
|
|
|
#ifdef CONFIG_BLOCK
|
|
|
|
if (bio) {
|
|
|
|
req->r_bio = bio;
|
|
|
|
bio_get(req->r_bio);
|
|
|
|
}
|
|
|
|
#endif
|
2010-04-06 23:51:47 +02:00
|
|
|
|
|
|
|
return req;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_alloc_request);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
static void osd_req_encode_op(struct ceph_osd_request *req,
|
|
|
|
struct ceph_osd_op *dst,
|
|
|
|
struct ceph_osd_req_op *src)
|
|
|
|
{
|
|
|
|
dst->op = cpu_to_le16(src->op);
|
|
|
|
|
|
|
|
switch (dst->op) {
|
|
|
|
case CEPH_OSD_OP_READ:
|
|
|
|
case CEPH_OSD_OP_WRITE:
|
|
|
|
dst->extent.offset =
|
|
|
|
cpu_to_le64(src->extent.offset);
|
|
|
|
dst->extent.length =
|
|
|
|
cpu_to_le64(src->extent.length);
|
|
|
|
dst->extent.truncate_size =
|
|
|
|
cpu_to_le64(src->extent.truncate_size);
|
|
|
|
dst->extent.truncate_seq =
|
|
|
|
cpu_to_le32(src->extent.truncate_seq);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case CEPH_OSD_OP_GETXATTR:
|
|
|
|
case CEPH_OSD_OP_SETXATTR:
|
|
|
|
case CEPH_OSD_OP_CMPXATTR:
|
|
|
|
BUG_ON(!req->r_trail);
|
|
|
|
|
|
|
|
dst->xattr.name_len = cpu_to_le32(src->xattr.name_len);
|
|
|
|
dst->xattr.value_len = cpu_to_le32(src->xattr.value_len);
|
|
|
|
dst->xattr.cmp_op = src->xattr.cmp_op;
|
|
|
|
dst->xattr.cmp_mode = src->xattr.cmp_mode;
|
|
|
|
ceph_pagelist_append(req->r_trail, src->xattr.name,
|
|
|
|
src->xattr.name_len);
|
|
|
|
ceph_pagelist_append(req->r_trail, src->xattr.val,
|
|
|
|
src->xattr.value_len);
|
|
|
|
break;
|
2010-05-19 01:38:08 +02:00
|
|
|
case CEPH_OSD_OP_CALL:
|
|
|
|
BUG_ON(!req->r_trail);
|
|
|
|
|
|
|
|
dst->cls.class_len = src->cls.class_len;
|
|
|
|
dst->cls.method_len = src->cls.method_len;
|
|
|
|
dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
|
|
|
|
|
|
|
|
ceph_pagelist_append(req->r_trail, src->cls.class_name,
|
|
|
|
src->cls.class_len);
|
|
|
|
ceph_pagelist_append(req->r_trail, src->cls.method_name,
|
|
|
|
src->cls.method_len);
|
|
|
|
ceph_pagelist_append(req->r_trail, src->cls.indata,
|
|
|
|
src->cls.indata_len);
|
|
|
|
break;
|
|
|
|
case CEPH_OSD_OP_ROLLBACK:
|
|
|
|
dst->snap.snapid = cpu_to_le64(src->snap.snapid);
|
|
|
|
break;
|
2010-04-07 00:01:27 +02:00
|
|
|
case CEPH_OSD_OP_STARTSYNC:
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
pr_err("unrecognized osd opcode %d\n", dst->op);
|
|
|
|
WARN_ON(1);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
dst->payload_len = cpu_to_le32(src->payload_len);
|
|
|
|
}
|
|
|
|
|
2010-04-06 23:51:47 +02:00
|
|
|
/*
|
|
|
|
* build new request AND message
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
void ceph_osdc_build_request(struct ceph_osd_request *req,
|
2010-04-07 00:01:27 +02:00
|
|
|
u64 off, u64 *plen,
|
|
|
|
struct ceph_osd_req_op *src_ops,
|
|
|
|
struct ceph_snap_context *snapc,
|
|
|
|
struct timespec *mtime,
|
|
|
|
const char *oid,
|
|
|
|
int oid_len)
|
2010-04-06 23:51:47 +02:00
|
|
|
{
|
|
|
|
struct ceph_msg *msg = req->r_request;
|
|
|
|
struct ceph_osd_request_head *head;
|
2010-04-07 00:01:27 +02:00
|
|
|
struct ceph_osd_req_op *src_op;
|
2010-04-06 23:51:47 +02:00
|
|
|
struct ceph_osd_op *op;
|
|
|
|
void *p;
|
2010-04-07 00:01:27 +02:00
|
|
|
int num_op = get_num_ops(src_ops, NULL);
|
2010-04-06 23:51:47 +02:00
|
|
|
size_t msg_size = sizeof(*head) + num_op*sizeof(*op);
|
|
|
|
int flags = req->r_flags;
|
2010-04-07 00:01:27 +02:00
|
|
|
u64 data_len = 0;
|
|
|
|
int i;
|
2010-04-06 23:51:47 +02:00
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
head = msg->front.iov_base;
|
|
|
|
op = (void *)(head + 1);
|
|
|
|
p = (void *)(op + num_op);
|
|
|
|
|
|
|
|
req->r_snapc = ceph_get_snap_context(snapc);
|
|
|
|
|
|
|
|
head->client_inc = cpu_to_le32(1); /* always, for now. */
|
|
|
|
head->flags = cpu_to_le32(flags);
|
|
|
|
if (flags & CEPH_OSD_FLAG_WRITE)
|
|
|
|
ceph_encode_timespec(&head->mtime, mtime);
|
|
|
|
head->num_ops = cpu_to_le16(num_op);
|
|
|
|
|
|
|
|
|
|
|
|
/* fill in oid */
|
2010-04-06 23:51:47 +02:00
|
|
|
head->object_len = cpu_to_le32(oid_len);
|
|
|
|
memcpy(p, oid, oid_len);
|
|
|
|
p += oid_len;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
src_op = src_ops;
|
|
|
|
while (src_op->op) {
|
|
|
|
osd_req_encode_op(req, op, src_op);
|
|
|
|
src_op++;
|
2009-10-06 20:31:10 +02:00
|
|
|
op++;
|
|
|
|
}
|
2010-04-07 00:01:27 +02:00
|
|
|
|
|
|
|
if (req->r_trail)
|
|
|
|
data_len += req->r_trail->length;
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
if (snapc) {
|
|
|
|
head->snap_seq = cpu_to_le64(snapc->seq);
|
|
|
|
head->num_snaps = cpu_to_le32(snapc->num_snaps);
|
|
|
|
for (i = 0; i < snapc->num_snaps; i++) {
|
|
|
|
put_unaligned_le64(snapc->snaps[i], p);
|
|
|
|
p += sizeof(u64);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
if (flags & CEPH_OSD_FLAG_WRITE) {
|
|
|
|
req->r_request->hdr.data_off = cpu_to_le16(off);
|
|
|
|
req->r_request->hdr.data_len = cpu_to_le32(*plen + data_len);
|
|
|
|
} else if (data_len) {
|
|
|
|
req->r_request->hdr.data_off = 0;
|
|
|
|
req->r_request->hdr.data_len = cpu_to_le32(data_len);
|
|
|
|
}
|
|
|
|
|
2010-11-09 21:40:00 +01:00
|
|
|
req->r_request->page_alignment = req->r_page_alignment;
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
|
2010-03-02 00:26:41 +01:00
|
|
|
msg_size = p - msg->front.iov_base;
|
|
|
|
msg->front.iov_len = msg_size;
|
|
|
|
msg->hdr.front_len = cpu_to_le32(msg_size);
|
2010-04-06 23:51:47 +02:00
|
|
|
return;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_build_request);
|
2010-04-06 23:51:47 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* build new request AND message, calculate layout, and adjust file
|
|
|
|
* extent as needed.
|
|
|
|
*
|
|
|
|
* if the file was recently truncated, we include information about its
|
|
|
|
* old and new size so that the object can be updated appropriately. (we
|
|
|
|
* avoid synchronously deleting truncated objects because it's slow.)
|
|
|
|
*
|
|
|
|
* if @do_sync, include a 'startsync' command so that the osd will flush
|
|
|
|
* data quickly.
|
|
|
|
*/
|
|
|
|
struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_file_layout *layout,
|
|
|
|
struct ceph_vino vino,
|
|
|
|
u64 off, u64 *plen,
|
|
|
|
int opcode, int flags,
|
|
|
|
struct ceph_snap_context *snapc,
|
|
|
|
int do_sync,
|
|
|
|
u32 truncate_seq,
|
|
|
|
u64 truncate_size,
|
|
|
|
struct timespec *mtime,
|
2010-11-09 21:43:12 +01:00
|
|
|
bool use_mempool, int num_reply,
|
|
|
|
int page_align)
|
2010-04-06 23:51:47 +02:00
|
|
|
{
|
2010-04-07 00:01:27 +02:00
|
|
|
struct ceph_osd_req_op ops[3];
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
|
|
|
|
ops[0].op = opcode;
|
|
|
|
ops[0].extent.truncate_seq = truncate_seq;
|
|
|
|
ops[0].extent.truncate_size = truncate_size;
|
|
|
|
ops[0].payload_len = 0;
|
|
|
|
|
|
|
|
if (do_sync) {
|
|
|
|
ops[1].op = CEPH_OSD_OP_STARTSYNC;
|
|
|
|
ops[1].payload_len = 0;
|
|
|
|
ops[2].op = 0;
|
|
|
|
} else
|
|
|
|
ops[1].op = 0;
|
|
|
|
|
|
|
|
req = ceph_osdc_alloc_request(osdc, flags,
|
|
|
|
snapc, ops,
|
2010-04-06 23:51:47 +02:00
|
|
|
use_mempool,
|
2010-04-07 00:01:27 +02:00
|
|
|
GFP_NOFS, NULL, NULL);
|
2010-04-06 23:51:47 +02:00
|
|
|
if (IS_ERR(req))
|
|
|
|
return req;
|
|
|
|
|
|
|
|
/* calculate max write size */
|
2010-04-07 00:01:27 +02:00
|
|
|
calc_layout(osdc, vino, layout, off, plen, req, ops);
|
2010-04-06 23:51:47 +02:00
|
|
|
req->r_file_layout = *layout; /* keep a copy */
|
|
|
|
|
2010-11-09 21:43:12 +01:00
|
|
|
/* in case it differs from natural alignment that calc_layout
|
|
|
|
filled in for us */
|
|
|
|
req->r_page_alignment = page_align;
|
|
|
|
|
2010-04-07 00:01:27 +02:00
|
|
|
ceph_osdc_build_request(req, off, plen, ops,
|
|
|
|
snapc,
|
2010-04-06 23:51:47 +02:00
|
|
|
mtime,
|
|
|
|
req->r_oid, req->r_oid_len);
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
return req;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_new_request);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We keep osd requests in an rbtree, sorted by ->r_tid.
|
|
|
|
*/
|
|
|
|
static void __insert_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *new)
|
|
|
|
{
|
|
|
|
struct rb_node **p = &osdc->requests.rb_node;
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
struct ceph_osd_request *req = NULL;
|
|
|
|
|
|
|
|
while (*p) {
|
|
|
|
parent = *p;
|
|
|
|
req = rb_entry(parent, struct ceph_osd_request, r_node);
|
|
|
|
if (new->r_tid < req->r_tid)
|
|
|
|
p = &(*p)->rb_left;
|
|
|
|
else if (new->r_tid > req->r_tid)
|
|
|
|
p = &(*p)->rb_right;
|
|
|
|
else
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(&new->r_node, parent, p);
|
|
|
|
rb_insert_color(&new->r_node, &osdc->requests);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
|
|
|
|
u64 tid)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
struct rb_node *n = osdc->requests.rb_node;
|
|
|
|
|
|
|
|
while (n) {
|
|
|
|
req = rb_entry(n, struct ceph_osd_request, r_node);
|
|
|
|
if (tid < req->r_tid)
|
|
|
|
n = n->rb_left;
|
|
|
|
else if (tid > req->r_tid)
|
|
|
|
n = n->rb_right;
|
|
|
|
else
|
|
|
|
return req;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct ceph_osd_request *
|
|
|
|
__lookup_request_ge(struct ceph_osd_client *osdc,
|
|
|
|
u64 tid)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
struct rb_node *n = osdc->requests.rb_node;
|
|
|
|
|
|
|
|
while (n) {
|
|
|
|
req = rb_entry(n, struct ceph_osd_request, r_node);
|
|
|
|
if (tid < req->r_tid) {
|
|
|
|
if (!n->rb_left)
|
|
|
|
return req;
|
|
|
|
n = n->rb_left;
|
|
|
|
} else if (tid > req->r_tid) {
|
|
|
|
n = n->rb_right;
|
|
|
|
} else {
|
|
|
|
return req;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
/*
|
|
|
|
* Resubmit requests pending on the given osd.
|
|
|
|
*/
|
|
|
|
static void __kick_osd_requests(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd *osd)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
dout("__kick_osd_requests osd%d\n", osd->o_osd);
|
|
|
|
err = __reset_osd(osdc, osd);
|
|
|
|
if (err == -EAGAIN)
|
|
|
|
return;
|
|
|
|
|
|
|
|
list_for_each_entry(req, &osd->o_requests, r_osd_item) {
|
|
|
|
list_move(&req->r_req_lru_item, &osdc->req_unsent);
|
|
|
|
dout("requeued %p tid %llu osd%d\n", req, req->r_tid,
|
|
|
|
osd->o_osd);
|
|
|
|
req->r_flags |= CEPH_OSD_FLAG_RETRY;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kick_osd_requests(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd *kickosd)
|
|
|
|
{
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
__kick_osd_requests(osdc, kickosd);
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
2009-10-09 19:29:18 +02:00
|
|
|
* If the osd connection drops, we need to resubmit all requests.
|
2009-10-06 20:31:10 +02:00
|
|
|
*/
|
|
|
|
static void osd_reset(struct ceph_connection *con)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
|
|
|
struct ceph_osd_client *osdc;
|
|
|
|
|
|
|
|
if (!osd)
|
|
|
|
return;
|
|
|
|
dout("osd_reset osd%d\n", osd->o_osd);
|
|
|
|
osdc = osd->o_osdc;
|
|
|
|
down_read(&osdc->map_sem);
|
2011-01-18 05:34:08 +01:00
|
|
|
kick_osd_requests(osdc, osd);
|
|
|
|
send_queued(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
up_read(&osdc->map_sem);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Track open sessions with osds.
|
|
|
|
*/
|
|
|
|
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd;
|
|
|
|
|
|
|
|
osd = kzalloc(sizeof(*osd), GFP_NOFS);
|
|
|
|
if (!osd)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
atomic_set(&osd->o_ref, 1);
|
|
|
|
osd->o_osdc = osdc;
|
|
|
|
INIT_LIST_HEAD(&osd->o_requests);
|
2010-02-03 20:00:26 +01:00
|
|
|
INIT_LIST_HEAD(&osd->o_osd_lru);
|
2009-10-06 20:31:10 +02:00
|
|
|
osd->o_incarnation = 1;
|
|
|
|
|
|
|
|
ceph_con_init(osdc->client->msgr, &osd->o_con);
|
|
|
|
osd->o_con.private = osd;
|
|
|
|
osd->o_con.ops = &osd_con_ops;
|
|
|
|
osd->o_con.peer_name.type = CEPH_ENTITY_TYPE_OSD;
|
2009-11-19 01:19:57 +01:00
|
|
|
|
2010-02-27 00:32:31 +01:00
|
|
|
INIT_LIST_HEAD(&osd->o_keepalive_item);
|
2009-10-06 20:31:10 +02:00
|
|
|
return osd;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct ceph_osd *get_osd(struct ceph_osd *osd)
|
|
|
|
{
|
|
|
|
if (atomic_inc_not_zero(&osd->o_ref)) {
|
|
|
|
dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
|
|
|
|
atomic_read(&osd->o_ref));
|
|
|
|
return osd;
|
|
|
|
} else {
|
|
|
|
dout("get_osd %p FAIL\n", osd);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void put_osd(struct ceph_osd *osd)
|
|
|
|
{
|
|
|
|
dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
|
|
|
|
atomic_read(&osd->o_ref) - 1);
|
2010-05-27 23:15:49 +02:00
|
|
|
if (atomic_dec_and_test(&osd->o_ref)) {
|
|
|
|
struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;
|
|
|
|
|
|
|
|
if (osd->o_authorizer)
|
|
|
|
ac->ops->destroy_authorizer(ac, osd->o_authorizer);
|
2009-10-06 20:31:10 +02:00
|
|
|
kfree(osd);
|
2010-05-27 23:15:49 +02:00
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* remove an osd from our map
|
|
|
|
*/
|
2010-02-03 20:00:26 +01:00
|
|
|
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
2010-02-03 20:00:26 +01:00
|
|
|
dout("__remove_osd %p\n", osd);
|
2009-10-06 20:31:10 +02:00
|
|
|
BUG_ON(!list_empty(&osd->o_requests));
|
|
|
|
rb_erase(&osd->o_node, &osdc->osds);
|
2010-02-03 20:00:26 +01:00
|
|
|
list_del_init(&osd->o_osd_lru);
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_con_close(&osd->o_con);
|
|
|
|
put_osd(osd);
|
|
|
|
}
|
|
|
|
|
2010-02-03 20:00:26 +01:00
|
|
|
static void __move_osd_to_lru(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd *osd)
|
|
|
|
{
|
|
|
|
dout("__move_osd_to_lru %p\n", osd);
|
|
|
|
BUG_ON(!list_empty(&osd->o_osd_lru));
|
|
|
|
list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
|
2010-04-07 00:14:15 +02:00
|
|
|
osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
|
2010-02-03 20:00:26 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void __remove_osd_from_lru(struct ceph_osd *osd)
|
|
|
|
{
|
|
|
|
dout("__remove_osd_from_lru %p\n", osd);
|
|
|
|
if (!list_empty(&osd->o_osd_lru))
|
|
|
|
list_del_init(&osd->o_osd_lru);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void remove_old_osds(struct ceph_osd_client *osdc, int remove_all)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd, *nosd;
|
|
|
|
|
|
|
|
dout("__remove_old_osds %p\n", osdc);
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
|
|
|
|
if (!remove_all && time_before(jiffies, osd->lru_ttl))
|
|
|
|
break;
|
|
|
|
__remove_osd(osdc, osd);
|
|
|
|
}
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
}
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
/*
|
|
|
|
* reset osd connect
|
|
|
|
*/
|
2010-02-03 20:00:26 +01:00
|
|
|
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
ceph: avoid reopening osd connections when address hasn't changed
We get a fault callback on _every_ tcp connection fault. Normally, we
want to reopen the connection when that happens. If the address we have
is bad, however, and connection attempts always result in a connection
refused or similar error, explicitly closing and reopening the msgr
connection just prevents the messenger's backoff logic from kicking in.
The result can be a console full of
[ 3974.417106] ceph: osd11 10.3.14.138:6800 connection failed
[ 3974.423295] ceph: osd11 10.3.14.138:6800 connection failed
[ 3974.429709] ceph: osd11 10.3.14.138:6800 connection failed
Instead, if we get a fault, and have outstanding requests, but the osd
address hasn't changed and the connection never successfully connected in
the first place, do nothing to the osd connection. The messenger layer
will back off and retry periodically, because we never connected and thus
the lossy bit is not set.
Instead, touch each request's r_stamp so that handle_timeout can tell the
request is still alive and kicking.
Signed-off-by: Sage Weil <sage@newdream.net>
2010-03-22 22:51:18 +01:00
|
|
|
struct ceph_osd_request *req;
|
2009-10-06 20:31:10 +02:00
|
|
|
int ret = 0;
|
|
|
|
|
2010-02-03 20:00:26 +01:00
|
|
|
dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
|
2009-10-06 20:31:10 +02:00
|
|
|
if (list_empty(&osd->o_requests)) {
|
2010-02-03 20:00:26 +01:00
|
|
|
__remove_osd(osdc, osd);
|
ceph: avoid reopening osd connections when address hasn't changed
We get a fault callback on _every_ tcp connection fault. Normally, we
want to reopen the connection when that happens. If the address we have
is bad, however, and connection attempts always result in a connection
refused or similar error, explicitly closing and reopening the msgr
connection just prevents the messenger's backoff logic from kicking in.
The result can be a console full of
[ 3974.417106] ceph: osd11 10.3.14.138:6800 connection failed
[ 3974.423295] ceph: osd11 10.3.14.138:6800 connection failed
[ 3974.429709] ceph: osd11 10.3.14.138:6800 connection failed
Instead, if we get a fault, and have outstanding requests, but the osd
address hasn't changed and the connection never successfully connected in
the first place, do nothing to the osd connection. The messenger layer
will back off and retry periodically, because we never connected and thus
the lossy bit is not set.
Instead, touch each request's r_stamp so that handle_timeout can tell the
request is still alive and kicking.
Signed-off-by: Sage Weil <sage@newdream.net>
2010-03-22 22:51:18 +01:00
|
|
|
} else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd],
|
|
|
|
&osd->o_con.peer_addr,
|
|
|
|
sizeof(osd->o_con.peer_addr)) == 0 &&
|
|
|
|
!ceph_con_opened(&osd->o_con)) {
|
|
|
|
dout(" osd addr hasn't changed and connection never opened,"
|
|
|
|
" letting msgr retry");
|
|
|
|
/* touch each r_stamp for handle_timeout()'s benfit */
|
|
|
|
list_for_each_entry(req, &osd->o_requests, r_osd_item)
|
|
|
|
req->r_stamp = jiffies;
|
|
|
|
ret = -EAGAIN;
|
2009-10-06 20:31:10 +02:00
|
|
|
} else {
|
|
|
|
ceph_con_close(&osd->o_con);
|
|
|
|
ceph_con_open(&osd->o_con, &osdc->osdmap->osd_addr[osd->o_osd]);
|
|
|
|
osd->o_incarnation++;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
|
|
|
|
{
|
|
|
|
struct rb_node **p = &osdc->osds.rb_node;
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
struct ceph_osd *osd = NULL;
|
|
|
|
|
|
|
|
while (*p) {
|
|
|
|
parent = *p;
|
|
|
|
osd = rb_entry(parent, struct ceph_osd, o_node);
|
|
|
|
if (new->o_osd < osd->o_osd)
|
|
|
|
p = &(*p)->rb_left;
|
|
|
|
else if (new->o_osd > osd->o_osd)
|
|
|
|
p = &(*p)->rb_right;
|
|
|
|
else
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(&new->o_node, parent, p);
|
|
|
|
rb_insert_color(&new->o_node, &osdc->osds);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd;
|
|
|
|
struct rb_node *n = osdc->osds.rb_node;
|
|
|
|
|
|
|
|
while (n) {
|
|
|
|
osd = rb_entry(n, struct ceph_osd, o_node);
|
|
|
|
if (o < osd->o_osd)
|
|
|
|
n = n->rb_left;
|
|
|
|
else if (o > osd->o_osd)
|
|
|
|
n = n->rb_right;
|
|
|
|
else
|
|
|
|
return osd;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2010-02-27 00:32:31 +01:00
|
|
|
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
schedule_delayed_work(&osdc->timeout_work,
|
2010-04-07 00:14:15 +02:00
|
|
|
osdc->client->options->osd_keepalive_timeout * HZ);
|
2010-02-27 00:32:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
cancel_delayed_work(&osdc->timeout_work);
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Register request, assign tid. If this is the first request, set up
|
|
|
|
* the timeout event.
|
|
|
|
*/
|
|
|
|
static void register_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req)
|
|
|
|
{
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
req->r_tid = ++osdc->last_tid;
|
2009-12-22 20:24:33 +01:00
|
|
|
req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
|
2010-02-27 00:32:31 +01:00
|
|
|
INIT_LIST_HEAD(&req->r_req_lru_item);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
dout("register_request %p tid %lld\n", req, req->r_tid);
|
|
|
|
__insert_request(osdc, req);
|
|
|
|
ceph_osdc_get_request(req);
|
|
|
|
osdc->num_requests++;
|
|
|
|
|
|
|
|
if (osdc->num_requests == 1) {
|
2010-02-27 00:32:31 +01:00
|
|
|
dout(" first request, scheduling timeout\n");
|
|
|
|
__schedule_osd_timeout(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* called under osdc->request_mutex
|
|
|
|
*/
|
|
|
|
static void __unregister_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req)
|
|
|
|
{
|
|
|
|
dout("__unregister_request %p tid %lld\n", req, req->r_tid);
|
|
|
|
rb_erase(&req->r_node, &osdc->requests);
|
|
|
|
osdc->num_requests--;
|
|
|
|
|
2009-10-09 01:57:16 +02:00
|
|
|
if (req->r_osd) {
|
|
|
|
/* make sure the original request isn't in flight. */
|
|
|
|
ceph_con_revoke(&req->r_osd->o_con, req->r_request);
|
|
|
|
|
|
|
|
list_del_init(&req->r_osd_item);
|
|
|
|
if (list_empty(&req->r_osd->o_requests))
|
2010-02-03 20:00:26 +01:00
|
|
|
__move_osd_to_lru(osdc, req->r_osd);
|
2009-10-09 01:57:16 +02:00
|
|
|
req->r_osd = NULL;
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
|
2010-02-27 00:32:31 +01:00
|
|
|
list_del_init(&req->r_req_lru_item);
|
|
|
|
if (osdc->num_requests == 0) {
|
|
|
|
dout(" no requests, canceling timeout\n");
|
|
|
|
__cancel_osd_timeout(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Cancel a previously queued request message
|
|
|
|
*/
|
|
|
|
static void __cancel_request(struct ceph_osd_request *req)
|
|
|
|
{
|
2010-09-27 19:18:52 +02:00
|
|
|
if (req->r_sent && req->r_osd) {
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_con_revoke(&req->r_osd->o_con, req->r_request);
|
|
|
|
req->r_sent = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pick an osd (the first 'up' osd in the pg), allocate the osd struct
|
|
|
|
* (as needed), and set the request r_osd appropriately. If there is
|
2011-01-18 05:34:08 +01:00
|
|
|
* no up osd, set r_osd to NULL. Move the request to the appropiate list
|
|
|
|
* (unsent, homeless) or leave on in-flight lru.
|
2009-10-06 20:31:10 +02:00
|
|
|
*
|
|
|
|
* Return 0 if unchanged, 1 if changed, or negative on error.
|
|
|
|
*
|
|
|
|
* Caller should hold map_sem for read and request_mutex.
|
|
|
|
*/
|
2011-01-18 05:34:08 +01:00
|
|
|
static int __map_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
|
2009-11-04 20:39:12 +01:00
|
|
|
struct ceph_pg pgid;
|
2010-05-10 19:24:48 +02:00
|
|
|
int acting[CEPH_PG_MAX_SIZE];
|
|
|
|
int o = -1, num = 0;
|
2009-10-06 20:31:10 +02:00
|
|
|
int err;
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
dout("map_request %p tid %lld\n", req, req->r_tid);
|
2009-10-06 20:31:10 +02:00
|
|
|
err = ceph_calc_object_layout(&reqhead->layout, req->r_oid,
|
|
|
|
&req->r_file_layout, osdc->osdmap);
|
2011-01-18 05:34:08 +01:00
|
|
|
if (err) {
|
|
|
|
list_move(&req->r_req_lru_item, &osdc->req_notarget);
|
2009-10-06 20:31:10 +02:00
|
|
|
return err;
|
2011-01-18 05:34:08 +01:00
|
|
|
}
|
2009-11-04 20:39:12 +01:00
|
|
|
pgid = reqhead->layout.ol_pgid;
|
2010-01-09 00:58:25 +01:00
|
|
|
req->r_pgid = pgid;
|
|
|
|
|
2010-05-10 19:24:48 +02:00
|
|
|
err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
|
|
|
|
if (err > 0) {
|
|
|
|
o = acting[0];
|
|
|
|
num = err;
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
if ((req->r_osd && req->r_osd->o_osd == o &&
|
2010-05-10 19:24:48 +02:00
|
|
|
req->r_sent >= req->r_osd->o_incarnation &&
|
|
|
|
req->r_num_pg_osds == num &&
|
|
|
|
memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
|
2009-10-06 20:31:10 +02:00
|
|
|
(req->r_osd == NULL && o == -1))
|
|
|
|
return 0; /* no change */
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
dout("map_request tid %llu pgid %d.%x osd%d (was osd%d)\n",
|
2009-11-04 20:39:12 +01:00
|
|
|
req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o,
|
2009-10-06 20:31:10 +02:00
|
|
|
req->r_osd ? req->r_osd->o_osd : -1);
|
|
|
|
|
2010-05-10 19:24:48 +02:00
|
|
|
/* record full pg acting set */
|
|
|
|
memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
|
|
|
|
req->r_num_pg_osds = num;
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
if (req->r_osd) {
|
|
|
|
__cancel_request(req);
|
|
|
|
list_del_init(&req->r_osd_item);
|
|
|
|
req->r_osd = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
req->r_osd = __lookup_osd(osdc, o);
|
|
|
|
if (!req->r_osd && o >= 0) {
|
2010-02-26 18:37:33 +01:00
|
|
|
err = -ENOMEM;
|
|
|
|
req->r_osd = create_osd(osdc);
|
2011-01-18 05:34:08 +01:00
|
|
|
if (!req->r_osd) {
|
|
|
|
list_move(&req->r_req_lru_item, &osdc->req_notarget);
|
2010-02-26 18:37:33 +01:00
|
|
|
goto out;
|
2011-01-18 05:34:08 +01:00
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
dout("map_request osd %p is osd%d\n", req->r_osd, o);
|
2009-10-06 20:31:10 +02:00
|
|
|
req->r_osd->o_osd = o;
|
|
|
|
req->r_osd->o_con.peer_name.num = cpu_to_le64(o);
|
|
|
|
__insert_osd(osdc, req->r_osd);
|
|
|
|
|
|
|
|
ceph_con_open(&req->r_osd->o_con, &osdc->osdmap->osd_addr[o]);
|
|
|
|
}
|
|
|
|
|
2010-02-03 20:00:26 +01:00
|
|
|
if (req->r_osd) {
|
|
|
|
__remove_osd_from_lru(req->r_osd);
|
2009-10-06 20:31:10 +02:00
|
|
|
list_add(&req->r_osd_item, &req->r_osd->o_requests);
|
2011-01-18 05:34:08 +01:00
|
|
|
list_move(&req->r_req_lru_item, &osdc->req_unsent);
|
|
|
|
} else {
|
|
|
|
list_move(&req->r_req_lru_item, &osdc->req_notarget);
|
2010-02-03 20:00:26 +01:00
|
|
|
}
|
2010-05-10 19:24:48 +02:00
|
|
|
err = 1; /* osd or pg changed */
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* caller should hold map_sem (for read) and request_mutex
|
|
|
|
*/
|
|
|
|
static int __send_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request_head *reqhead;
|
|
|
|
|
|
|
|
dout("send_request %p tid %llu to osd%d flags %d\n",
|
|
|
|
req, req->r_tid, req->r_osd->o_osd, req->r_flags);
|
|
|
|
|
|
|
|
reqhead = req->r_request->front.iov_base;
|
|
|
|
reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch);
|
|
|
|
reqhead->flags |= cpu_to_le32(req->r_flags); /* e.g., RETRY */
|
|
|
|
reqhead->reassert_version = req->r_reassert_version;
|
|
|
|
|
2010-03-22 22:42:30 +01:00
|
|
|
req->r_stamp = jiffies;
|
2010-08-23 06:34:27 +02:00
|
|
|
list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
ceph_msg_get(req->r_request); /* send consumes a ref */
|
|
|
|
ceph_con_send(&req->r_osd->o_con, req->r_request);
|
|
|
|
req->r_sent = req->r_osd->o_incarnation;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
/*
|
|
|
|
* Send any requests in the queue (req_unsent).
|
|
|
|
*/
|
|
|
|
static void send_queued(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req, *tmp;
|
|
|
|
|
|
|
|
dout("send_queued\n");
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) {
|
|
|
|
__send_request(osdc, req);
|
|
|
|
}
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
}
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
/*
|
|
|
|
* Timeout callback, called every N seconds when 1 or more osd
|
|
|
|
* requests has been active for more than N seconds. When this
|
|
|
|
* happens, we ping all OSDs with requests who have timed out to
|
|
|
|
* ensure any communications channel reset is detected. Reset the
|
|
|
|
* request timeouts another N seconds in the future as we go.
|
|
|
|
* Reschedule the timeout event another N seconds in future (unless
|
|
|
|
* there are no open requests).
|
|
|
|
*/
|
|
|
|
static void handle_timeout(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct ceph_osd_client *osdc =
|
|
|
|
container_of(work, struct ceph_osd_client, timeout_work.work);
|
2010-02-27 00:32:31 +01:00
|
|
|
struct ceph_osd_request *req, *last_req = NULL;
|
2009-10-06 20:31:10 +02:00
|
|
|
struct ceph_osd *osd;
|
2010-04-07 00:14:15 +02:00
|
|
|
unsigned long timeout = osdc->client->options->osd_timeout * HZ;
|
2010-02-27 00:32:31 +01:00
|
|
|
unsigned long keepalive =
|
2010-04-07 00:14:15 +02:00
|
|
|
osdc->client->options->osd_keepalive_timeout * HZ;
|
2010-03-22 22:42:30 +01:00
|
|
|
unsigned long last_stamp = 0;
|
2010-02-27 00:32:31 +01:00
|
|
|
struct list_head slow_osds;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
dout("timeout\n");
|
|
|
|
down_read(&osdc->map_sem);
|
|
|
|
|
|
|
|
ceph_monc_request_next_osdmap(&osdc->client->monc);
|
|
|
|
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
|
2010-02-27 00:32:31 +01:00
|
|
|
/*
|
|
|
|
* reset osds that appear to be _really_ unresponsive. this
|
|
|
|
* is a failsafe measure.. we really shouldn't be getting to
|
|
|
|
* this point if the system is working properly. the monitors
|
|
|
|
* should mark the osd as failed and we should find out about
|
|
|
|
* it from an updated osd map.
|
|
|
|
*/
|
2010-04-21 20:09:38 +02:00
|
|
|
while (timeout && !list_empty(&osdc->req_lru)) {
|
2010-02-27 00:32:31 +01:00
|
|
|
req = list_entry(osdc->req_lru.next, struct ceph_osd_request,
|
|
|
|
r_req_lru_item);
|
|
|
|
|
2010-03-22 22:42:30 +01:00
|
|
|
if (time_before(jiffies, req->r_stamp + timeout))
|
2010-02-27 00:32:31 +01:00
|
|
|
break;
|
|
|
|
|
2010-03-22 22:42:30 +01:00
|
|
|
BUG_ON(req == last_req && req->r_stamp == last_stamp);
|
2010-02-27 00:32:31 +01:00
|
|
|
last_req = req;
|
2010-03-22 22:42:30 +01:00
|
|
|
last_stamp = req->r_stamp;
|
2010-02-27 00:32:31 +01:00
|
|
|
|
|
|
|
osd = req->r_osd;
|
|
|
|
BUG_ON(!osd);
|
|
|
|
pr_warning(" tid %llu timed out on osd%d, will reset osd\n",
|
|
|
|
req->r_tid, osd->o_osd);
|
2011-01-18 05:34:08 +01:00
|
|
|
__kick_osd_requests(osdc, osd);
|
2010-02-27 00:32:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ping osds that are a bit slow. this ensures that if there
|
|
|
|
* is a break in the TCP connection we will notice, and reopen
|
|
|
|
* a connection with that osd (from the fault callback).
|
|
|
|
*/
|
|
|
|
INIT_LIST_HEAD(&slow_osds);
|
|
|
|
list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
|
2010-03-22 22:42:30 +01:00
|
|
|
if (time_before(jiffies, req->r_stamp + keepalive))
|
2010-02-27 00:32:31 +01:00
|
|
|
break;
|
|
|
|
|
|
|
|
osd = req->r_osd;
|
|
|
|
BUG_ON(!osd);
|
|
|
|
dout(" tid %llu is slow, will send keepalive on osd%d\n",
|
2009-10-06 20:31:10 +02:00
|
|
|
req->r_tid, osd->o_osd);
|
2010-02-27 00:32:31 +01:00
|
|
|
list_move_tail(&osd->o_keepalive_item, &slow_osds);
|
|
|
|
}
|
|
|
|
while (!list_empty(&slow_osds)) {
|
|
|
|
osd = list_entry(slow_osds.next, struct ceph_osd,
|
|
|
|
o_keepalive_item);
|
|
|
|
list_del_init(&osd->o_keepalive_item);
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_con_keepalive(&osd->o_con);
|
|
|
|
}
|
|
|
|
|
2010-02-27 00:32:31 +01:00
|
|
|
__schedule_osd_timeout(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
mutex_unlock(&osdc->request_mutex);
|
2011-01-18 05:34:08 +01:00
|
|
|
send_queued(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
up_read(&osdc->map_sem);
|
|
|
|
}
|
|
|
|
|
2010-02-03 20:00:26 +01:00
|
|
|
static void handle_osds_timeout(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct ceph_osd_client *osdc =
|
|
|
|
container_of(work, struct ceph_osd_client,
|
|
|
|
osds_timeout_work.work);
|
|
|
|
unsigned long delay =
|
2010-04-07 00:14:15 +02:00
|
|
|
osdc->client->options->osd_idle_ttl * HZ >> 2;
|
2010-02-03 20:00:26 +01:00
|
|
|
|
|
|
|
dout("osds timeout\n");
|
|
|
|
down_read(&osdc->map_sem);
|
|
|
|
remove_old_osds(osdc, 0);
|
|
|
|
up_read(&osdc->map_sem);
|
|
|
|
|
|
|
|
schedule_delayed_work(&osdc->osds_timeout_work,
|
|
|
|
round_jiffies_relative(delay));
|
|
|
|
}
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
/*
|
|
|
|
* handle osd op reply. either call the callback if it is specified,
|
|
|
|
* or do the completion to wake up the waiting thread.
|
|
|
|
*/
|
2009-12-22 19:45:45 +01:00
|
|
|
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
|
|
|
|
struct ceph_connection *con)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd_reply_head *rhead = msg->front.iov_base;
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
u64 tid;
|
|
|
|
int numops, object_len, flags;
|
2010-05-11 18:53:18 +02:00
|
|
|
s32 result;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2009-12-22 20:24:33 +01:00
|
|
|
tid = le64_to_cpu(msg->hdr.tid);
|
2009-10-06 20:31:10 +02:00
|
|
|
if (msg->front.iov_len < sizeof(*rhead))
|
|
|
|
goto bad;
|
|
|
|
numops = le32_to_cpu(rhead->num_ops);
|
|
|
|
object_len = le32_to_cpu(rhead->object_len);
|
2010-05-11 18:53:18 +02:00
|
|
|
result = le32_to_cpu(rhead->result);
|
2009-10-06 20:31:10 +02:00
|
|
|
if (msg->front.iov_len != sizeof(*rhead) + object_len +
|
|
|
|
numops * sizeof(struct ceph_osd_op))
|
|
|
|
goto bad;
|
2010-05-11 18:53:18 +02:00
|
|
|
dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/* lookup */
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
req = __lookup_request(osdc, tid);
|
|
|
|
if (req == NULL) {
|
|
|
|
dout("handle_reply tid %llu dne\n", tid);
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
ceph_osdc_get_request(req);
|
|
|
|
flags = le32_to_cpu(rhead->flags);
|
|
|
|
|
2009-12-22 19:45:45 +01:00
|
|
|
/*
|
2010-01-14 02:03:23 +01:00
|
|
|
* if this connection filled our message, drop our reference now, to
|
2009-12-22 19:45:45 +01:00
|
|
|
* avoid a (safe but slower) revoke later.
|
|
|
|
*/
|
2010-01-14 02:03:23 +01:00
|
|
|
if (req->r_con_filling_msg == con && req->r_reply == msg) {
|
2010-03-01 22:02:00 +01:00
|
|
|
dout(" dropping con_filling_msg ref %p\n", con);
|
2010-01-14 02:03:23 +01:00
|
|
|
req->r_con_filling_msg = NULL;
|
2009-12-22 19:45:45 +01:00
|
|
|
ceph_con_put(con);
|
|
|
|
}
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
if (!req->r_got_reply) {
|
|
|
|
unsigned bytes;
|
|
|
|
|
|
|
|
req->r_result = le32_to_cpu(rhead->result);
|
|
|
|
bytes = le32_to_cpu(msg->hdr.data_len);
|
|
|
|
dout("handle_reply result %d bytes %d\n", req->r_result,
|
|
|
|
bytes);
|
|
|
|
if (req->r_result == 0)
|
|
|
|
req->r_result = bytes;
|
|
|
|
|
|
|
|
/* in case this is a write and we need to replay, */
|
|
|
|
req->r_reassert_version = rhead->reassert_version;
|
|
|
|
|
|
|
|
req->r_got_reply = 1;
|
|
|
|
} else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
|
|
|
|
dout("handle_reply tid %llu dup ack\n", tid);
|
2009-12-01 21:23:54 +01:00
|
|
|
mutex_unlock(&osdc->request_mutex);
|
2009-10-06 20:31:10 +02:00
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
|
|
|
|
dout("handle_reply tid %llu flags %d\n", tid, flags);
|
|
|
|
|
|
|
|
/* either this is a read, or we got the safe response */
|
2010-05-11 18:53:18 +02:00
|
|
|
if (result < 0 ||
|
|
|
|
(flags & CEPH_OSD_FLAG_ONDISK) ||
|
2009-10-06 20:31:10 +02:00
|
|
|
((flags & CEPH_OSD_FLAG_WRITE) == 0))
|
|
|
|
__unregister_request(osdc, req);
|
|
|
|
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
|
|
|
|
if (req->r_callback)
|
|
|
|
req->r_callback(req, msg);
|
|
|
|
else
|
2010-07-27 22:11:08 +02:00
|
|
|
complete_all(&req->r_completion);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
if (flags & CEPH_OSD_FLAG_ONDISK) {
|
|
|
|
if (req->r_safe_callback)
|
|
|
|
req->r_safe_callback(req, msg);
|
2010-07-27 22:11:08 +02:00
|
|
|
complete_all(&req->r_safe_completion); /* fsync waiter */
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
done:
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
return;
|
|
|
|
|
|
|
|
bad:
|
|
|
|
pr_err("corrupt osd_op_reply got %d %d expected %d\n",
|
|
|
|
(int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len),
|
|
|
|
(int)sizeof(*rhead));
|
2009-12-15 00:13:47 +01:00
|
|
|
ceph_msg_dump(msg);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
static void reset_changed_osds(struct ceph_osd_client *osdc)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct rb_node *p, *n;
|
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
for (p = rb_first(&osdc->osds); p; p = n) {
|
|
|
|
struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
n = rb_next(p);
|
|
|
|
if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
|
|
|
|
memcmp(&osd->o_con.peer_addr,
|
|
|
|
ceph_osd_addr(osdc->osdmap,
|
|
|
|
osd->o_osd),
|
|
|
|
sizeof(struct ceph_entity_addr)) != 0)
|
|
|
|
__reset_osd(osdc, osd);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
2010-02-27 00:32:31 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2011-01-18 05:34:08 +01:00
|
|
|
* Requeue requests whose mapping to an OSD has changed. If requests map to
|
|
|
|
* no osd, request a new map.
|
2010-02-27 00:32:31 +01:00
|
|
|
*
|
|
|
|
* Caller should hold map_sem for read and request_mutex.
|
|
|
|
*/
|
2011-01-18 05:34:08 +01:00
|
|
|
static void kick_requests(struct ceph_osd_client *osdc)
|
2010-02-27 00:32:31 +01:00
|
|
|
{
|
2011-01-18 05:34:08 +01:00
|
|
|
struct ceph_osd_request *req;
|
|
|
|
struct rb_node *p;
|
|
|
|
int needmap = 0;
|
|
|
|
int err;
|
2010-02-27 00:32:31 +01:00
|
|
|
|
2011-01-18 05:34:08 +01:00
|
|
|
dout("kick_requests\n");
|
2010-02-27 00:32:31 +01:00
|
|
|
mutex_lock(&osdc->request_mutex);
|
2011-01-18 05:34:08 +01:00
|
|
|
for (p = rb_first(&osdc->requests); p; p = rb_next(p)) {
|
|
|
|
req = rb_entry(p, struct ceph_osd_request, r_node);
|
|
|
|
err = __map_request(osdc, req);
|
|
|
|
if (err < 0)
|
|
|
|
continue; /* error */
|
|
|
|
if (req->r_osd == NULL) {
|
|
|
|
dout("%p tid %llu maps to no osd\n", req, req->r_tid);
|
|
|
|
needmap++; /* request a newer map */
|
|
|
|
} else if (err > 0) {
|
|
|
|
dout("%p tid %llu requeued on osd%d\n", req, req->r_tid,
|
|
|
|
req->r_osd ? req->r_osd->o_osd : -1);
|
|
|
|
req->r_flags |= CEPH_OSD_FLAG_RETRY;
|
|
|
|
}
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
|
|
|
|
if (needmap) {
|
|
|
|
dout("%d requests for down osds, need new map\n", needmap);
|
|
|
|
ceph_monc_request_next_osdmap(&osdc->client->monc);
|
|
|
|
}
|
2010-02-27 00:32:31 +01:00
|
|
|
}
|
2011-01-18 05:34:08 +01:00
|
|
|
|
|
|
|
|
2009-10-06 20:31:10 +02:00
|
|
|
/*
|
|
|
|
* Process updated osd map.
|
|
|
|
*
|
|
|
|
* The message contains any number of incremental and full maps, normally
|
|
|
|
* indicating some sort of topology change in the cluster. Kick requests
|
|
|
|
* off to different OSDs as needed.
|
|
|
|
*/
|
|
|
|
void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
|
|
|
|
{
|
|
|
|
void *p, *end, *next;
|
|
|
|
u32 nr_maps, maplen;
|
|
|
|
u32 epoch;
|
|
|
|
struct ceph_osdmap *newmap = NULL, *oldmap;
|
|
|
|
int err;
|
|
|
|
struct ceph_fsid fsid;
|
|
|
|
|
|
|
|
dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
|
|
|
|
p = msg->front.iov_base;
|
|
|
|
end = p + msg->front.iov_len;
|
|
|
|
|
|
|
|
/* verify fsid */
|
|
|
|
ceph_decode_need(&p, end, sizeof(fsid), bad);
|
|
|
|
ceph_decode_copy(&p, &fsid, sizeof(fsid));
|
2009-11-19 01:50:41 +01:00
|
|
|
if (ceph_check_fsid(osdc->client, &fsid) < 0)
|
|
|
|
return;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
down_write(&osdc->map_sem);
|
|
|
|
|
|
|
|
/* incremental maps */
|
|
|
|
ceph_decode_32_safe(&p, end, nr_maps, bad);
|
|
|
|
dout(" %d inc maps\n", nr_maps);
|
|
|
|
while (nr_maps > 0) {
|
|
|
|
ceph_decode_need(&p, end, 2*sizeof(u32), bad);
|
2009-10-14 18:59:09 +02:00
|
|
|
epoch = ceph_decode_32(&p);
|
|
|
|
maplen = ceph_decode_32(&p);
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_decode_need(&p, end, maplen, bad);
|
|
|
|
next = p + maplen;
|
|
|
|
if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
|
|
|
|
dout("applying incremental map %u len %d\n",
|
|
|
|
epoch, maplen);
|
|
|
|
newmap = osdmap_apply_incremental(&p, next,
|
|
|
|
osdc->osdmap,
|
|
|
|
osdc->client->msgr);
|
|
|
|
if (IS_ERR(newmap)) {
|
|
|
|
err = PTR_ERR(newmap);
|
|
|
|
goto bad;
|
|
|
|
}
|
2009-12-21 23:49:37 +01:00
|
|
|
BUG_ON(!newmap);
|
2009-10-06 20:31:10 +02:00
|
|
|
if (newmap != osdc->osdmap) {
|
|
|
|
ceph_osdmap_destroy(osdc->osdmap);
|
|
|
|
osdc->osdmap = newmap;
|
|
|
|
}
|
2011-01-18 05:34:08 +01:00
|
|
|
kick_requests(osdc);
|
|
|
|
reset_changed_osds(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
} else {
|
|
|
|
dout("ignoring incremental map %u len %d\n",
|
|
|
|
epoch, maplen);
|
|
|
|
}
|
|
|
|
p = next;
|
|
|
|
nr_maps--;
|
|
|
|
}
|
|
|
|
if (newmap)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
/* full maps */
|
|
|
|
ceph_decode_32_safe(&p, end, nr_maps, bad);
|
|
|
|
dout(" %d full maps\n", nr_maps);
|
|
|
|
while (nr_maps) {
|
|
|
|
ceph_decode_need(&p, end, 2*sizeof(u32), bad);
|
2009-10-14 18:59:09 +02:00
|
|
|
epoch = ceph_decode_32(&p);
|
|
|
|
maplen = ceph_decode_32(&p);
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_decode_need(&p, end, maplen, bad);
|
|
|
|
if (nr_maps > 1) {
|
|
|
|
dout("skipping non-latest full map %u len %d\n",
|
|
|
|
epoch, maplen);
|
|
|
|
} else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
|
|
|
|
dout("skipping full map %u len %d, "
|
|
|
|
"older than our %u\n", epoch, maplen,
|
|
|
|
osdc->osdmap->epoch);
|
|
|
|
} else {
|
|
|
|
dout("taking full map %u len %d\n", epoch, maplen);
|
|
|
|
newmap = osdmap_decode(&p, p+maplen);
|
|
|
|
if (IS_ERR(newmap)) {
|
|
|
|
err = PTR_ERR(newmap);
|
|
|
|
goto bad;
|
|
|
|
}
|
2009-12-21 23:49:37 +01:00
|
|
|
BUG_ON(!newmap);
|
2009-10-06 20:31:10 +02:00
|
|
|
oldmap = osdc->osdmap;
|
|
|
|
osdc->osdmap = newmap;
|
|
|
|
if (oldmap)
|
|
|
|
ceph_osdmap_destroy(oldmap);
|
2011-01-18 05:34:08 +01:00
|
|
|
kick_requests(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
p += maplen;
|
|
|
|
nr_maps--;
|
|
|
|
}
|
|
|
|
|
|
|
|
done:
|
|
|
|
downgrade_write(&osdc->map_sem);
|
|
|
|
ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
|
2011-01-18 05:34:08 +01:00
|
|
|
send_queued(osdc);
|
2009-10-06 20:31:10 +02:00
|
|
|
up_read(&osdc->map_sem);
|
2010-07-27 22:11:08 +02:00
|
|
|
wake_up_all(&osdc->client->auth_wq);
|
2009-10-06 20:31:10 +02:00
|
|
|
return;
|
|
|
|
|
|
|
|
bad:
|
|
|
|
pr_err("osdc handle_map corrupt msg\n");
|
2009-12-15 00:13:47 +01:00
|
|
|
ceph_msg_dump(msg);
|
2009-10-06 20:31:10 +02:00
|
|
|
up_write(&osdc->map_sem);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Register request, send initial attempt.
|
|
|
|
*/
|
|
|
|
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req,
|
|
|
|
bool nofail)
|
|
|
|
{
|
2009-10-09 01:55:47 +02:00
|
|
|
int rc = 0;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
req->r_request->pages = req->r_pages;
|
|
|
|
req->r_request->nr_pages = req->r_num_pages;
|
2010-04-07 00:01:27 +02:00
|
|
|
#ifdef CONFIG_BLOCK
|
|
|
|
req->r_request->bio = req->r_bio;
|
|
|
|
#endif
|
|
|
|
req->r_request->trail = req->r_trail;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
register_request(osdc, req);
|
|
|
|
|
|
|
|
down_read(&osdc->map_sem);
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
2009-10-09 01:55:47 +02:00
|
|
|
/*
|
|
|
|
* a racing kick_requests() may have sent the message for us
|
|
|
|
* while we dropped request_mutex above, so only send now if
|
|
|
|
* the request still han't been touched yet.
|
|
|
|
*/
|
|
|
|
if (req->r_sent == 0) {
|
2011-01-18 05:34:08 +01:00
|
|
|
rc = __map_request(osdc, req);
|
|
|
|
if (rc < 0)
|
|
|
|
return rc;
|
|
|
|
if (req->r_osd == NULL) {
|
|
|
|
dout("send_request %p no up osds in pg\n", req);
|
|
|
|
ceph_monc_request_next_osdmap(&osdc->client->monc);
|
|
|
|
} else {
|
|
|
|
rc = __send_request(osdc, req);
|
|
|
|
if (rc) {
|
|
|
|
if (nofail) {
|
|
|
|
dout("osdc_start_request failed send, "
|
|
|
|
" will retry %lld\n", req->r_tid);
|
|
|
|
rc = 0;
|
|
|
|
} else {
|
|
|
|
__unregister_request(osdc, req);
|
|
|
|
}
|
2009-10-09 01:55:47 +02:00
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
up_read(&osdc->map_sem);
|
|
|
|
return rc;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_start_request);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* wait for a request to complete
|
|
|
|
*/
|
|
|
|
int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_osd_request *req)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
rc = wait_for_completion_interruptible(&req->r_completion);
|
|
|
|
if (rc < 0) {
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
__cancel_request(req);
|
2009-12-22 19:29:39 +01:00
|
|
|
__unregister_request(osdc, req);
|
2009-10-06 20:31:10 +02:00
|
|
|
mutex_unlock(&osdc->request_mutex);
|
2009-12-22 19:29:39 +01:00
|
|
|
dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
|
2009-10-06 20:31:10 +02:00
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
|
|
|
|
return req->r_result;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_wait_request);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* sync - wait for all in-flight requests to flush. avoid starvation.
|
|
|
|
*/
|
|
|
|
void ceph_osdc_sync(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
u64 last_tid, next_tid = 0;
|
|
|
|
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
last_tid = osdc->last_tid;
|
|
|
|
while (1) {
|
|
|
|
req = __lookup_request_ge(osdc, next_tid);
|
|
|
|
if (!req)
|
|
|
|
break;
|
|
|
|
if (req->r_tid > last_tid)
|
|
|
|
break;
|
|
|
|
|
|
|
|
next_tid = req->r_tid + 1;
|
|
|
|
if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
ceph_osdc_get_request(req);
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
dout("sync waiting on tid %llu (last is %llu)\n",
|
|
|
|
req->r_tid, last_tid);
|
|
|
|
wait_for_completion(&req->r_safe_completion);
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
}
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
|
|
|
dout("sync done (thru tid %llu)\n", last_tid);
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_sync);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* init, shutdown
|
|
|
|
*/
|
|
|
|
int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
|
|
|
|
dout("init\n");
|
|
|
|
osdc->client = client;
|
|
|
|
osdc->osdmap = NULL;
|
|
|
|
init_rwsem(&osdc->map_sem);
|
|
|
|
init_completion(&osdc->map_waiters);
|
|
|
|
osdc->last_requested_map = 0;
|
|
|
|
mutex_init(&osdc->request_mutex);
|
|
|
|
osdc->last_tid = 0;
|
|
|
|
osdc->osds = RB_ROOT;
|
2010-02-03 20:00:26 +01:00
|
|
|
INIT_LIST_HEAD(&osdc->osd_lru);
|
2009-10-06 20:31:10 +02:00
|
|
|
osdc->requests = RB_ROOT;
|
2010-02-27 00:32:31 +01:00
|
|
|
INIT_LIST_HEAD(&osdc->req_lru);
|
2011-01-18 05:34:08 +01:00
|
|
|
INIT_LIST_HEAD(&osdc->req_unsent);
|
|
|
|
INIT_LIST_HEAD(&osdc->req_notarget);
|
2009-10-06 20:31:10 +02:00
|
|
|
osdc->num_requests = 0;
|
|
|
|
INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
|
2010-02-03 20:00:26 +01:00
|
|
|
INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
|
|
|
|
|
|
|
|
schedule_delayed_work(&osdc->osds_timeout_work,
|
2010-04-07 00:14:15 +02:00
|
|
|
round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2009-11-18 23:52:18 +01:00
|
|
|
err = -ENOMEM;
|
2009-10-06 20:31:10 +02:00
|
|
|
osdc->req_mempool = mempool_create_kmalloc_pool(10,
|
|
|
|
sizeof(struct ceph_osd_request));
|
|
|
|
if (!osdc->req_mempool)
|
2009-11-18 23:52:18 +01:00
|
|
|
goto out;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-04-24 18:56:35 +02:00
|
|
|
err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true,
|
|
|
|
"osd_op");
|
2009-10-06 20:31:10 +02:00
|
|
|
if (err < 0)
|
2009-11-18 23:52:18 +01:00
|
|
|
goto out_mempool;
|
2010-03-01 22:02:00 +01:00
|
|
|
err = ceph_msgpool_init(&osdc->msgpool_op_reply,
|
2010-04-24 18:56:35 +02:00
|
|
|
OSD_OPREPLY_FRONT_LEN, 10, true,
|
|
|
|
"osd_op_reply");
|
2010-03-01 22:02:00 +01:00
|
|
|
if (err < 0)
|
|
|
|
goto out_msgpool;
|
2009-10-06 20:31:10 +02:00
|
|
|
return 0;
|
2009-11-18 23:52:18 +01:00
|
|
|
|
2010-03-01 22:02:00 +01:00
|
|
|
out_msgpool:
|
|
|
|
ceph_msgpool_destroy(&osdc->msgpool_op);
|
2009-11-18 23:52:18 +01:00
|
|
|
out_mempool:
|
|
|
|
mempool_destroy(osdc->req_mempool);
|
|
|
|
out:
|
|
|
|
return err;
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_init);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
void ceph_osdc_stop(struct ceph_osd_client *osdc)
|
|
|
|
{
|
|
|
|
cancel_delayed_work_sync(&osdc->timeout_work);
|
2010-02-03 20:00:26 +01:00
|
|
|
cancel_delayed_work_sync(&osdc->osds_timeout_work);
|
2009-10-06 20:31:10 +02:00
|
|
|
if (osdc->osdmap) {
|
|
|
|
ceph_osdmap_destroy(osdc->osdmap);
|
|
|
|
osdc->osdmap = NULL;
|
|
|
|
}
|
2010-02-03 20:00:26 +01:00
|
|
|
remove_old_osds(osdc, 1);
|
2009-10-06 20:31:10 +02:00
|
|
|
mempool_destroy(osdc->req_mempool);
|
|
|
|
ceph_msgpool_destroy(&osdc->msgpool_op);
|
2010-03-01 22:02:00 +01:00
|
|
|
ceph_msgpool_destroy(&osdc->msgpool_op_reply);
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_stop);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Read some contiguous pages. If we cross a stripe boundary, shorten
|
|
|
|
* *plen. Return number of bytes read, or error.
|
|
|
|
*/
|
|
|
|
int ceph_osdc_readpages(struct ceph_osd_client *osdc,
|
|
|
|
struct ceph_vino vino, struct ceph_file_layout *layout,
|
|
|
|
u64 off, u64 *plen,
|
|
|
|
u32 truncate_seq, u64 truncate_size,
|
2010-11-09 21:43:12 +01:00
|
|
|
struct page **pages, int num_pages, int page_align)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
|
|
|
|
vino.snap, off, *plen);
|
|
|
|
req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
|
|
|
|
CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
|
|
|
|
NULL, 0, truncate_seq, truncate_size, NULL,
|
2010-11-09 21:43:12 +01:00
|
|
|
false, 1, page_align);
|
2010-04-02 01:06:19 +02:00
|
|
|
if (!req)
|
|
|
|
return -ENOMEM;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/* it may be a short read due to an object boundary */
|
|
|
|
req->r_pages = pages;
|
|
|
|
|
2010-11-09 21:43:12 +01:00
|
|
|
dout("readpages final extent is %llu~%llu (%d pages align %d)\n",
|
|
|
|
off, *plen, req->r_num_pages, page_align);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
rc = ceph_osdc_start_request(osdc, req, false);
|
|
|
|
if (!rc)
|
|
|
|
rc = ceph_osdc_wait_request(osdc, req);
|
|
|
|
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
dout("readpages result %d\n", rc);
|
|
|
|
return rc;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_readpages);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* do a synchronous write on N pages
|
|
|
|
*/
|
|
|
|
int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
|
|
|
|
struct ceph_file_layout *layout,
|
|
|
|
struct ceph_snap_context *snapc,
|
|
|
|
u64 off, u64 len,
|
|
|
|
u32 truncate_seq, u64 truncate_size,
|
|
|
|
struct timespec *mtime,
|
|
|
|
struct page **pages, int num_pages,
|
|
|
|
int flags, int do_sync, bool nofail)
|
|
|
|
{
|
|
|
|
struct ceph_osd_request *req;
|
|
|
|
int rc = 0;
|
2010-11-09 21:43:12 +01:00
|
|
|
int page_align = off & ~PAGE_MASK;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
BUG_ON(vino.snap != CEPH_NOSNAP);
|
|
|
|
req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
|
|
|
|
CEPH_OSD_OP_WRITE,
|
|
|
|
flags | CEPH_OSD_FLAG_ONDISK |
|
|
|
|
CEPH_OSD_FLAG_WRITE,
|
|
|
|
snapc, do_sync,
|
|
|
|
truncate_seq, truncate_size, mtime,
|
2010-11-09 21:43:12 +01:00
|
|
|
nofail, 1, page_align);
|
2010-04-02 01:06:19 +02:00
|
|
|
if (!req)
|
|
|
|
return -ENOMEM;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/* it may be a short write due to an object boundary */
|
|
|
|
req->r_pages = pages;
|
|
|
|
dout("writepages %llu~%llu (%d pages)\n", off, len,
|
|
|
|
req->r_num_pages);
|
|
|
|
|
|
|
|
rc = ceph_osdc_start_request(osdc, req, nofail);
|
|
|
|
if (!rc)
|
|
|
|
rc = ceph_osdc_wait_request(osdc, req);
|
|
|
|
|
|
|
|
ceph_osdc_put_request(req);
|
|
|
|
if (rc == 0)
|
|
|
|
rc = len;
|
|
|
|
dout("writepages result %d\n", rc);
|
|
|
|
return rc;
|
|
|
|
}
|
2010-04-07 00:14:15 +02:00
|
|
|
EXPORT_SYMBOL(ceph_osdc_writepages);
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* handle incoming message
|
|
|
|
*/
|
|
|
|
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
2009-11-21 16:53:16 +01:00
|
|
|
struct ceph_osd_client *osdc;
|
2009-10-06 20:31:10 +02:00
|
|
|
int type = le16_to_cpu(msg->hdr.type);
|
|
|
|
|
|
|
|
if (!osd)
|
2010-06-13 19:27:53 +02:00
|
|
|
goto out;
|
2009-11-21 16:53:16 +01:00
|
|
|
osdc = osd->o_osdc;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case CEPH_MSG_OSD_MAP:
|
|
|
|
ceph_osdc_handle_map(osdc, msg);
|
|
|
|
break;
|
|
|
|
case CEPH_MSG_OSD_OPREPLY:
|
2009-12-22 19:45:45 +01:00
|
|
|
handle_reply(osdc, msg, con);
|
2009-10-06 20:31:10 +02:00
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
|
|
|
pr_err("received unknown message type %d %s\n", type,
|
|
|
|
ceph_msg_type_name(type));
|
|
|
|
}
|
2010-06-13 19:27:53 +02:00
|
|
|
out:
|
2009-10-06 20:31:10 +02:00
|
|
|
ceph_msg_put(msg);
|
|
|
|
}
|
|
|
|
|
2010-02-20 06:43:23 +01:00
|
|
|
/*
|
2010-03-04 19:22:59 +01:00
|
|
|
* lookup and return message for incoming reply. set up reply message
|
|
|
|
* pages.
|
2010-02-20 06:43:23 +01:00
|
|
|
*/
|
|
|
|
static struct ceph_msg *get_reply(struct ceph_connection *con,
|
2010-01-08 22:58:34 +01:00
|
|
|
struct ceph_msg_header *hdr,
|
|
|
|
int *skip)
|
2009-10-06 20:31:10 +02:00
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
|
|
|
struct ceph_osd_client *osdc = osd->o_osdc;
|
2010-01-08 22:58:34 +01:00
|
|
|
struct ceph_msg *m;
|
2010-01-11 23:47:13 +01:00
|
|
|
struct ceph_osd_request *req;
|
2010-02-20 06:43:23 +01:00
|
|
|
int front = le32_to_cpu(hdr->front_len);
|
|
|
|
int data_len = le32_to_cpu(hdr->data_len);
|
2010-01-11 23:47:13 +01:00
|
|
|
u64 tid;
|
2009-10-06 20:31:10 +02:00
|
|
|
|
2010-01-11 23:47:13 +01:00
|
|
|
tid = le64_to_cpu(hdr->tid);
|
|
|
|
mutex_lock(&osdc->request_mutex);
|
|
|
|
req = __lookup_request(osdc, tid);
|
|
|
|
if (!req) {
|
|
|
|
*skip = 1;
|
|
|
|
m = NULL;
|
2010-03-01 22:02:00 +01:00
|
|
|
pr_info("get_reply unknown tid %llu from osd%d\n", tid,
|
2010-02-20 06:43:23 +01:00
|
|
|
osd->o_osd);
|
2010-01-11 23:47:13 +01:00
|
|
|
goto out;
|
|
|
|
}
|
2010-03-01 22:02:00 +01:00
|
|
|
|
|
|
|
if (req->r_con_filling_msg) {
|
|
|
|
dout("get_reply revoking msg %p from old con %p\n",
|
|
|
|
req->r_reply, req->r_con_filling_msg);
|
|
|
|
ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply);
|
|
|
|
ceph_con_put(req->r_con_filling_msg);
|
2010-03-25 05:30:19 +01:00
|
|
|
req->r_con_filling_msg = NULL;
|
2010-01-11 23:47:13 +01:00
|
|
|
}
|
|
|
|
|
2010-03-01 22:02:00 +01:00
|
|
|
if (front > req->r_reply->front.iov_len) {
|
|
|
|
pr_warning("get_reply front %d > preallocated %d\n",
|
|
|
|
front, (int)req->r_reply->front.iov_len);
|
2010-04-06 23:33:58 +02:00
|
|
|
m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS);
|
2010-04-02 01:06:19 +02:00
|
|
|
if (!m)
|
2010-03-01 22:02:00 +01:00
|
|
|
goto out;
|
|
|
|
ceph_msg_put(req->r_reply);
|
|
|
|
req->r_reply = m;
|
|
|
|
}
|
|
|
|
m = ceph_msg_get(req->r_reply);
|
|
|
|
|
2010-01-11 23:47:13 +01:00
|
|
|
if (data_len > 0) {
|
2010-11-09 21:43:12 +01:00
|
|
|
int want = calc_pages_for(req->r_page_alignment, data_len);
|
2010-03-04 19:22:59 +01:00
|
|
|
|
|
|
|
if (unlikely(req->r_num_pages < want)) {
|
|
|
|
pr_warning("tid %lld reply %d > expected %d pages\n",
|
|
|
|
tid, want, m->nr_pages);
|
2010-01-11 23:47:13 +01:00
|
|
|
*skip = 1;
|
|
|
|
ceph_msg_put(m);
|
2010-04-02 01:06:19 +02:00
|
|
|
m = NULL;
|
2010-03-04 19:22:59 +01:00
|
|
|
goto out;
|
2010-01-11 23:47:13 +01:00
|
|
|
}
|
2010-03-04 19:22:59 +01:00
|
|
|
m->pages = req->r_pages;
|
|
|
|
m->nr_pages = req->r_num_pages;
|
2010-11-09 21:40:00 +01:00
|
|
|
m->page_alignment = req->r_page_alignment;
|
2010-04-07 00:01:27 +02:00
|
|
|
#ifdef CONFIG_BLOCK
|
|
|
|
m->bio = req->r_bio;
|
|
|
|
#endif
|
2010-01-11 23:47:13 +01:00
|
|
|
}
|
2010-02-20 06:43:23 +01:00
|
|
|
*skip = 0;
|
2010-03-01 22:02:00 +01:00
|
|
|
req->r_con_filling_msg = ceph_con_get(con);
|
|
|
|
dout("get_reply tid %lld %p\n", tid, m);
|
2010-01-11 23:47:13 +01:00
|
|
|
|
|
|
|
out:
|
|
|
|
mutex_unlock(&osdc->request_mutex);
|
2010-01-08 22:58:34 +01:00
|
|
|
return m;
|
2010-02-20 06:43:23 +01:00
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct ceph_msg *alloc_msg(struct ceph_connection *con,
|
|
|
|
struct ceph_msg_header *hdr,
|
|
|
|
int *skip)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
|
|
|
int type = le16_to_cpu(hdr->type);
|
|
|
|
int front = le32_to_cpu(hdr->front_len);
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case CEPH_MSG_OSD_MAP:
|
2010-04-06 23:33:58 +02:00
|
|
|
return ceph_msg_new(type, front, GFP_NOFS);
|
2010-02-20 06:43:23 +01:00
|
|
|
case CEPH_MSG_OSD_OPREPLY:
|
|
|
|
return get_reply(con, hdr, skip);
|
|
|
|
default:
|
|
|
|
pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
|
|
|
|
osd->o_osd);
|
|
|
|
*skip = 1;
|
|
|
|
return NULL;
|
|
|
|
}
|
2009-10-06 20:31:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Wrappers to refcount containing ceph_osd struct
|
|
|
|
*/
|
|
|
|
static struct ceph_connection *get_osd_con(struct ceph_connection *con)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
|
|
|
if (get_osd(osd))
|
|
|
|
return con;
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void put_osd_con(struct ceph_connection *con)
|
|
|
|
{
|
|
|
|
struct ceph_osd *osd = con->private;
|
|
|
|
put_osd(osd);
|
|
|
|
}
|
|
|
|
|
2009-11-19 01:19:57 +01:00
|
|
|
/*
|
|
|
|
* authentication
|
|
|
|
*/
|
|
|
|
static int get_authorizer(struct ceph_connection *con,
|
2010-08-03 19:25:11 +02:00
|
|
|
void **buf, int *len, int *proto,
|
|
|
|
void **reply_buf, int *reply_len, int force_new)
|
2009-11-19 01:19:57 +01:00
|
|
|
{
|
|
|
|
struct ceph_osd *o = con->private;
|
|
|
|
struct ceph_osd_client *osdc = o->o_osdc;
|
|
|
|
struct ceph_auth_client *ac = osdc->client->monc.auth;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (force_new && o->o_authorizer) {
|
|
|
|
ac->ops->destroy_authorizer(ac, o->o_authorizer);
|
|
|
|
o->o_authorizer = NULL;
|
|
|
|
}
|
|
|
|
if (o->o_authorizer == NULL) {
|
|
|
|
ret = ac->ops->create_authorizer(
|
|
|
|
ac, CEPH_ENTITY_TYPE_OSD,
|
|
|
|
&o->o_authorizer,
|
|
|
|
&o->o_authorizer_buf,
|
|
|
|
&o->o_authorizer_buf_len,
|
|
|
|
&o->o_authorizer_reply_buf,
|
|
|
|
&o->o_authorizer_reply_buf_len);
|
|
|
|
if (ret)
|
2010-08-03 19:25:11 +02:00
|
|
|
return ret;
|
2009-11-19 01:19:57 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
*proto = ac->protocol;
|
|
|
|
*buf = o->o_authorizer_buf;
|
|
|
|
*len = o->o_authorizer_buf_len;
|
|
|
|
*reply_buf = o->o_authorizer_reply_buf;
|
|
|
|
*reply_len = o->o_authorizer_reply_buf_len;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int verify_authorizer_reply(struct ceph_connection *con, int len)
|
|
|
|
{
|
|
|
|
struct ceph_osd *o = con->private;
|
|
|
|
struct ceph_osd_client *osdc = o->o_osdc;
|
|
|
|
struct ceph_auth_client *ac = osdc->client->monc.auth;
|
|
|
|
|
|
|
|
return ac->ops->verify_authorizer_reply(ac, o->o_authorizer, len);
|
|
|
|
}
|
|
|
|
|
2010-02-03 01:21:06 +01:00
|
|
|
static int invalidate_authorizer(struct ceph_connection *con)
|
|
|
|
{
|
|
|
|
struct ceph_osd *o = con->private;
|
|
|
|
struct ceph_osd_client *osdc = o->o_osdc;
|
|
|
|
struct ceph_auth_client *ac = osdc->client->monc.auth;
|
|
|
|
|
|
|
|
if (ac->ops->invalidate_authorizer)
|
|
|
|
ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
|
|
|
|
|
|
|
|
return ceph_monc_validate_auth(&osdc->client->monc);
|
|
|
|
}
|
2009-11-19 01:19:57 +01:00
|
|
|
|
2010-05-20 10:40:19 +02:00
|
|
|
static const struct ceph_connection_operations osd_con_ops = {
|
2009-10-06 20:31:10 +02:00
|
|
|
.get = get_osd_con,
|
|
|
|
.put = put_osd_con,
|
|
|
|
.dispatch = dispatch,
|
2009-11-19 01:19:57 +01:00
|
|
|
.get_authorizer = get_authorizer,
|
|
|
|
.verify_authorizer_reply = verify_authorizer_reply,
|
2010-02-03 01:21:06 +01:00
|
|
|
.invalidate_authorizer = invalidate_authorizer,
|
2009-10-06 20:31:10 +02:00
|
|
|
.alloc_msg = alloc_msg,
|
2009-10-09 19:29:18 +02:00
|
|
|
.fault = osd_reset,
|
2009-10-06 20:31:10 +02:00
|
|
|
};
|