libceph, rbd: new bio handling code (aka don't clone bios)

The reason we clone bios is to be able to give each object request
(and consequently each ceph_osd_data/ceph_msg_data item) its own
pointer to a (list of) bio(s).  The messenger then initializes its
cursor with cloned bio's ->bi_iter, so it knows where to start reading
from/writing to.  That's all the cloned bios are used for: to determine
each object request's starting position in the provided data buffer.

Introduce ceph_bio_iter to do exactly that -- store position within bio
list (i.e. pointer to bio) + position within that bio (i.e. bvec_iter).

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>

2 files changed, 57 insertions, 13 deletions

diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h
index ead9d85f1c11..d7b9605fd51d 100644
--- a/include/linux/ceph/messenger.h
+++ b/include/linux/ceph/messenger.h
@@ -93,14 +93,60 @@ static __inline__ bool ceph_msg_data_type_valid(enum ceph_msg_data_type type)
         }
 }
 
+#ifdef CONFIG_BLOCK
+
+struct ceph_bio_iter {
+        struct bio *bio;
+        struct bvec_iter iter;
+};
+
+#define __ceph_bio_iter_advance_step(it, n, STEP) do {                        \
+        unsigned int __n = (n), __cur_n;                                      \
+                                                                              \
+        while (__n) {                                                         \
+                BUG_ON(!(it)->iter.bi_size);                                  \
+                __cur_n = min((it)->iter.bi_size, __n);                       \
+                (void)(STEP);                                                 \
+                bio_advance_iter((it)->bio, &(it)->iter, __cur_n);            \
+                if (!(it)->iter.bi_size && (it)->bio->bi_next) {              \
+                        dout("__ceph_bio_iter_advance_step next bio\n");      \
+                        (it)->bio = (it)->bio->bi_next;                       \
+                        (it)->iter = (it)->bio->bi_iter;                      \
+                }                                                             \
+                __n -= __cur_n;                                               \
+        }                                                                     \
+} while (0)
+
+/*
+ * Advance @it by @n bytes.
+ */
+#define ceph_bio_iter_advance(it, n)                                          \
+        __ceph_bio_iter_advance_step(it, n, 0)
+
+/*
+ * Advance @it by @n bytes, executing BVEC_STEP for each bio_vec.
+ */
+#define ceph_bio_iter_advance_step(it, n, BVEC_STEP)                          \
+        __ceph_bio_iter_advance_step(it, n, ({                                \
+                struct bio_vec bv;                                            \
+                struct bvec_iter __cur_iter;                                  \
+                                                                              \
+                __cur_iter = (it)->iter;                                      \
+                __cur_iter.bi_size = __cur_n;                                 \
+                __bio_for_each_segment(bv, (it)->bio, __cur_iter, __cur_iter) \
+                        (void)(BVEC_STEP);                                    \
+        }))
+
+#endif /* CONFIG_BLOCK */
+
 struct ceph_msg_data {
         struct list_head                links;  /* ceph_msg->data */
         enum ceph_msg_data_type         type;
         union {
 #ifdef CONFIG_BLOCK
                 struct {
-                        struct bio      *bio;
-                        size_t          bio_length;
+                        struct ceph_bio_iter    bio_pos;
+                        u32                     bio_length;
                 };
 #endif /* CONFIG_BLOCK */
                 struct {
@@ -122,10 +168,7 @@ struct ceph_msg_data_cursor {
         bool                    need_crc;       /* crc update needed */
         union {
 #ifdef CONFIG_BLOCK
-                struct {                                /* bio */
-                        struct bio      *bio;           /* bio from list */
-                        struct bvec_iter bvec_iter;
-                };
+                struct ceph_bio_iter    bio_iter;
 #endif /* CONFIG_BLOCK */
                 struct {                                /* pages */
                         unsigned int    page_offset;    /* offset in page */
@@ -290,8 +333,8 @@ extern void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
 extern void ceph_msg_data_add_pagelist(struct ceph_msg *msg,
                                 struct ceph_pagelist *pagelist);
 #ifdef CONFIG_BLOCK
-extern void ceph_msg_data_add_bio(struct ceph_msg *msg, struct bio *bio,
-                                size_t length);
+void ceph_msg_data_add_bio(struct ceph_msg *msg, struct ceph_bio_iter *bio_pos,
+                           u32 length);
 #endif /* CONFIG_BLOCK */
 
 extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags,
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index 52fb37d1c2a5..315691490cb0 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -72,8 +72,8 @@ struct ceph_osd_data {
                 struct ceph_pagelist    *pagelist;
 #ifdef CONFIG_BLOCK
                 struct {
-                        struct bio      *bio;           /* list of bios */
-                        size_t          bio_length;     /* total in list */
+                        struct ceph_bio_iter    bio_pos;
+                        u32                     bio_length;
                 };
 #endif /* CONFIG_BLOCK */
         };
@@ -405,9 +405,10 @@ extern void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *,
                                         unsigned int which,
                                         struct ceph_pagelist *pagelist);
 #ifdef CONFIG_BLOCK
-extern void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *,
-                                        unsigned int which,
-                                        struct bio *bio, size_t bio_length);
+void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
+                                    unsigned int which,
+                                    struct ceph_bio_iter *bio_pos,
+                                    u32 bio_length);
 #endif /* CONFIG_BLOCK */
 
 extern void osd_req_op_cls_request_data_pagelist(struct ceph_osd_request *,