3 files changed, 35 insertions, 27 deletions

diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index 0e8157ee5d43..b6ff4a1519ab 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -97,11 +97,9 @@ struct workqueue_struct *ceph_msgr_wq;
 int ceph_msgr_init(void)
 {
         ceph_msgr_wq = create_workqueue("ceph-msgr");
-        if (IS_ERR(ceph_msgr_wq)) {
-                int ret = PTR_ERR(ceph_msgr_wq);
-                pr_err("msgr_init failed to create workqueue: %d\n", ret);
-                ceph_msgr_wq = NULL;
-                return ret;
+        if (!ceph_msgr_wq) {
+                pr_err("msgr_init failed to create workqueue\n");
+                return -ENOMEM;
         }
         return 0;
 }
@@ -540,8 +538,7 @@ static void prepare_write_message(struct ceph_connection *con)
                 /* initialize page iterator */
                 con->out_msg_pos.page = 0;
                 if (m->pages)
-                        con->out_msg_pos.page_pos =
-                                le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK;
+                        con->out_msg_pos.page_pos = m->page_alignment;
                 else
                         con->out_msg_pos.page_pos = 0;
                 con->out_msg_pos.data_pos = 0;
@@ -1491,7 +1488,7 @@ static int read_partial_message(struct ceph_connection *con)
         struct ceph_msg *m = con->in_msg;
         int ret;
         int to, left;
-        unsigned front_len, middle_len, data_len, data_off;
+        unsigned front_len, middle_len, data_len;
         int datacrc = con->msgr->nocrc;
         int skip;
         u64 seq;
@@ -1527,19 +1524,17 @@ static int read_partial_message(struct ceph_connection *con)
         data_len = le32_to_cpu(con->in_hdr.data_len);
         if (data_len > CEPH_MSG_MAX_DATA_LEN)
                 return -EIO;
-        data_off = le16_to_cpu(con->in_hdr.data_off);
 
         /* verify seq# */
         seq = le64_to_cpu(con->in_hdr.seq);
         if ((s64)seq - (s64)con->in_seq < 1) {
-                pr_info("skipping %s%lld %s seq %lld, expected %lld\n",
+                pr_info("skipping %s%lld %s seq %lld expected %lld\n",
                         ENTITY_NAME(con->peer_name),
                         ceph_pr_addr(&con->peer_addr.in_addr),
                         seq, con->in_seq + 1);
                 con->in_base_pos = -front_len - middle_len - data_len -
                         sizeof(m->footer);
                 con->in_tag = CEPH_MSGR_TAG_READY;
-                con->in_seq++;
                 return 0;
         } else if ((s64)seq - (s64)con->in_seq > 1) {
                 pr_err("read_partial_message bad seq %lld expected %lld\n",
@@ -1576,7 +1571,7 @@ static int read_partial_message(struct ceph_connection *con)
 
                 con->in_msg_pos.page = 0;
                 if (m->pages)
-                        con->in_msg_pos.page_pos = data_off & ~PAGE_MASK;
+                        con->in_msg_pos.page_pos = m->page_alignment;
                 else
                         con->in_msg_pos.page_pos = 0;
                 con->in_msg_pos.data_pos = 0;
@@ -2301,6 +2296,7 @@ struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags)
 
         /* data */
         m->nr_pages = 0;
+        m->page_alignment = 0;
         m->pages = NULL;
         m->pagelist = NULL;
         m->bio = NULL;
@@ -2370,6 +2366,7 @@ static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
                                type, front_len);
                         return NULL;
                 }
+                msg->page_alignment = le16_to_cpu(hdr->data_off);
         }
         memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr));
 
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 79391994b3ed..3e20a122ffa2 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -71,6 +71,7 @@ void ceph_calc_raw_layout(struct ceph_osd_client *osdc,
                 op->extent.length = objlen;
         }
         req->r_num_pages = calc_pages_for(off, *plen);
+        req->r_page_alignment = off & ~PAGE_MASK;
         if (op->op == CEPH_OSD_OP_WRITE)
                 op->payload_len = *plen;
 
@@ -390,6 +391,8 @@ void ceph_osdc_build_request(struct ceph_osd_request *req,
                 req->r_request->hdr.data_len = cpu_to_le32(data_len);
         }
 
+        req->r_request->page_alignment = req->r_page_alignment;
+
         BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
         msg_size = p - msg->front.iov_base;
         msg->front.iov_len = msg_size;
@@ -419,7 +422,8 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
                                                u32 truncate_seq,
                                                u64 truncate_size,
                                                struct timespec *mtime,
-                                               bool use_mempool, int num_reply)
+                                               bool use_mempool, int num_reply,
+                                               int page_align)
 {
         struct ceph_osd_req_op ops[3];
         struct ceph_osd_request *req;
@@ -447,6 +451,10 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
         calc_layout(osdc, vino, layout, off, plen, req, ops);
         req->r_file_layout = *layout;  /* keep a copy */
 
+        /* in case it differs from natural alignment that calc_layout
+           filled in for us */
+        req->r_page_alignment = page_align;
+
         ceph_osdc_build_request(req, off, plen, ops,
                                 snapc,
                                 mtime,
@@ -1489,7 +1497,7 @@ 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,
-                        struct page **pages, int num_pages)
+                        struct page **pages, int num_pages, int page_align)
 {
         struct ceph_osd_request *req;
         int rc = 0;
@@ -1499,15 +1507,15 @@ int ceph_osdc_readpages(struct ceph_osd_client *osdc,
         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,
-                                    false, 1);
+                                    false, 1, page_align);
         if (!req)
                 return -ENOMEM;
 
         /* it may be a short read due to an object boundary */
         req->r_pages = pages;
 
-        dout("readpages  final extent is %llu~%llu (%d pages)\n",
-             off, *plen, req->r_num_pages);
+        dout("readpages  final extent is %llu~%llu (%d pages align %d)\n",
+             off, *plen, req->r_num_pages, page_align);
 
         rc = ceph_osdc_start_request(osdc, req, false);
         if (!rc)
@@ -1533,6 +1541,7 @@ int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
 {
         struct ceph_osd_request *req;
         int rc = 0;
+        int page_align = off & ~PAGE_MASK;
 
         BUG_ON(vino.snap != CEPH_NOSNAP);
         req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
@@ -1541,7 +1550,7 @@ int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
                                             CEPH_OSD_FLAG_WRITE,
                                     snapc, do_sync,
                                     truncate_seq, truncate_size, mtime,
-                                    nofail, 1);
+                                    nofail, 1, page_align);
         if (!req)
                 return -ENOMEM;
 
@@ -1638,8 +1647,7 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
         m = ceph_msg_get(req->r_reply);
 
         if (data_len > 0) {
-                unsigned data_off = le16_to_cpu(hdr->data_off);
-                int want = calc_pages_for(data_off & ~PAGE_MASK, data_len);
+                int want = calc_pages_for(req->r_page_alignment, data_len);
 
                 if (unlikely(req->r_num_pages < want)) {
                         pr_warning("tid %lld reply %d > expected %d pages\n",
@@ -1651,6 +1659,7 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
                 }
                 m->pages = req->r_pages;
                 m->nr_pages = req->r_num_pages;
+                m->page_alignment = req->r_page_alignment;
 #ifdef CONFIG_BLOCK
                 m->bio = req->r_bio;
 #endif
diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c
index 54caf0687155..1a040e64c69f 100644
--- a/net/ceph/pagevec.c
+++ b/net/ceph/pagevec.c
@@ -13,8 +13,7 @@
  * build a vector of user pages
  */
 struct page **ceph_get_direct_page_vector(const char __user *data,
-                                                 int num_pages,
-                                                 loff_t off, size_t len)
+                                          int num_pages, bool write_page)
 {
         struct page **pages;
         int rc;
@@ -25,24 +24,27 @@ struct page **ceph_get_direct_page_vector(const char __user *data,
 
         down_read(&current->mm->mmap_sem);
         rc = get_user_pages(current, current->mm, (unsigned long)data,
-                            num_pages, 0, 0, pages, NULL);
+                            num_pages, write_page, 0, pages, NULL);
         up_read(&current->mm->mmap_sem);
-        if (rc < 0)
+        if (rc < num_pages)
                 goto fail;
         return pages;
 
 fail:
-        kfree(pages);
+        ceph_put_page_vector(pages, rc > 0 ? rc : 0, false);
         return ERR_PTR(rc);
 }
 EXPORT_SYMBOL(ceph_get_direct_page_vector);
 
-void ceph_put_page_vector(struct page **pages, int num_pages)
+void ceph_put_page_vector(struct page **pages, int num_pages, bool dirty)
 {
         int i;
 
-        for (i = 0; i < num_pages; i++)
+        for (i = 0; i < num_pages; i++) {
+                if (dirty)
+                        set_page_dirty_lock(pages[i]);
                 put_page(pages[i]);
+        }
         kfree(pages);
 }
 EXPORT_SYMBOL(ceph_put_page_vector);