From dccbf08005df800f5c8e948ab6132ed5536134bc Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Sat, 17 Feb 2018 09:29:58 +0100 Subject: libceph, ceph: change ceph_calc_file_object_mapping() signature - make it void - xlen (object extent length) out parameter should be u32 because only a single stripe unit is mapped at a time Signed-off-by: Ilya Dryomov Reviewed-by: Alex Elder --- include/linux/ceph/osdmap.h | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index d41fad99c0fa..92314035dac1 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -280,10 +280,9 @@ bool ceph_osds_changed(const struct ceph_osds *old_acting, const struct ceph_osds *new_acting, bool any_change); -/* calculate mapping of a file extent to an object */ -extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout, - u64 off, u64 len, - u64 *bno, u64 *oxoff, u64 *oxlen); +void ceph_calc_file_object_mapping(struct ceph_file_layout *l, + u64 off, u64 len, + u64 *objno, u64 *objoff, u32 *xlen); int __ceph_object_locator_to_pg(struct ceph_pg_pool_info *pi, const struct ceph_object_id *oid, -- cgit v1.2.3 From 5359a17d2706b86da2af83027343d5eb256f7670 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Sat, 20 Jan 2018 10:30:10 +0100 Subject: 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 --- include/linux/ceph/messenger.h | 59 +++++++++++++++++++++++++++++++++++------ include/linux/ceph/osd_client.h | 11 ++++---- 2 files changed, 57 insertions(+), 13 deletions(-) (limited to 'include/linux') 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 *, -- cgit v1.2.3 From b9e281c2b38804984d619e1d9efc4b9020bcb291 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Sat, 20 Jan 2018 10:30:11 +0100 Subject: libceph: introduce BVECS data type In preparation for rbd "fancy" striping, introduce ceph_bvec_iter for working with bio_vec array data buffers. The wrappers are trivial, but make it look similar to ceph_bio_iter. Signed-off-by: Ilya Dryomov --- include/linux/ceph/messenger.h | 42 +++++++++++++++++++++++++++++++++++++++++ include/linux/ceph/osd_client.h | 8 ++++++++ 2 files changed, 50 insertions(+) (limited to 'include/linux') diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h index d7b9605fd51d..c7dfcb8a1fb2 100644 --- a/include/linux/ceph/messenger.h +++ b/include/linux/ceph/messenger.h @@ -76,6 +76,7 @@ enum ceph_msg_data_type { #ifdef CONFIG_BLOCK CEPH_MSG_DATA_BIO, /* data source/destination is a bio list */ #endif /* CONFIG_BLOCK */ + CEPH_MSG_DATA_BVECS, /* data source/destination is a bio_vec array */ }; static __inline__ bool ceph_msg_data_type_valid(enum ceph_msg_data_type type) @@ -87,6 +88,7 @@ static __inline__ bool ceph_msg_data_type_valid(enum ceph_msg_data_type type) #ifdef CONFIG_BLOCK case CEPH_MSG_DATA_BIO: #endif /* CONFIG_BLOCK */ + case CEPH_MSG_DATA_BVECS: return true; default: return false; @@ -139,6 +141,42 @@ struct ceph_bio_iter { #endif /* CONFIG_BLOCK */ +struct ceph_bvec_iter { + struct bio_vec *bvecs; + struct bvec_iter iter; +}; + +#define __ceph_bvec_iter_advance_step(it, n, STEP) do { \ + BUG_ON((n) > (it)->iter.bi_size); \ + (void)(STEP); \ + bvec_iter_advance((it)->bvecs, &(it)->iter, (n)); \ +} while (0) + +/* + * Advance @it by @n bytes. + */ +#define ceph_bvec_iter_advance(it, n) \ + __ceph_bvec_iter_advance_step(it, n, 0) + +/* + * Advance @it by @n bytes, executing BVEC_STEP for each bio_vec. + */ +#define ceph_bvec_iter_advance_step(it, n, BVEC_STEP) \ + __ceph_bvec_iter_advance_step(it, n, ({ \ + struct bio_vec bv; \ + struct bvec_iter __cur_iter; \ + \ + __cur_iter = (it)->iter; \ + __cur_iter.bi_size = (n); \ + for_each_bvec(bv, (it)->bvecs, __cur_iter, __cur_iter) \ + (void)(BVEC_STEP); \ + })) + +#define ceph_bvec_iter_shorten(it, n) do { \ + BUG_ON((n) > (it)->iter.bi_size); \ + (it)->iter.bi_size = (n); \ +} while (0) + struct ceph_msg_data { struct list_head links; /* ceph_msg->data */ enum ceph_msg_data_type type; @@ -149,6 +187,7 @@ struct ceph_msg_data { u32 bio_length; }; #endif /* CONFIG_BLOCK */ + struct ceph_bvec_iter bvec_pos; struct { struct page **pages; /* NOT OWNER. */ size_t length; /* total # bytes */ @@ -170,6 +209,7 @@ struct ceph_msg_data_cursor { #ifdef CONFIG_BLOCK struct ceph_bio_iter bio_iter; #endif /* CONFIG_BLOCK */ + struct bvec_iter bvec_iter; struct { /* pages */ unsigned int page_offset; /* offset in page */ unsigned short page_index; /* index in array */ @@ -336,6 +376,8 @@ extern void ceph_msg_data_add_pagelist(struct ceph_msg *msg, void ceph_msg_data_add_bio(struct ceph_msg *msg, struct ceph_bio_iter *bio_pos, u32 length); #endif /* CONFIG_BLOCK */ +void ceph_msg_data_add_bvecs(struct ceph_msg *msg, + struct ceph_bvec_iter *bvec_pos); extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags, bool can_fail); diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 315691490cb0..528ccc943cee 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -57,6 +57,7 @@ enum ceph_osd_data_type { #ifdef CONFIG_BLOCK CEPH_OSD_DATA_TYPE_BIO, #endif /* CONFIG_BLOCK */ + CEPH_OSD_DATA_TYPE_BVECS, }; struct ceph_osd_data { @@ -76,6 +77,7 @@ struct ceph_osd_data { u32 bio_length; }; #endif /* CONFIG_BLOCK */ + struct ceph_bvec_iter bvec_pos; }; }; @@ -410,6 +412,9 @@ void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req, struct ceph_bio_iter *bio_pos, u32 bio_length); #endif /* CONFIG_BLOCK */ +void osd_req_op_extent_osd_data_bvec_pos(struct ceph_osd_request *osd_req, + unsigned int which, + struct ceph_bvec_iter *bvec_pos); extern void osd_req_op_cls_request_data_pagelist(struct ceph_osd_request *, unsigned int which, @@ -419,6 +424,9 @@ extern void osd_req_op_cls_request_data_pages(struct ceph_osd_request *, struct page **pages, u64 length, u32 alignment, bool pages_from_pool, bool own_pages); +void osd_req_op_cls_request_data_bvecs(struct ceph_osd_request *osd_req, + unsigned int which, + struct bio_vec *bvecs, u32 bytes); extern void osd_req_op_cls_response_data_pages(struct ceph_osd_request *, unsigned int which, struct page **pages, u64 length, -- cgit v1.2.3 From ed0811d2d243c4195580a9671266031907c02ca7 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Fri, 2 Feb 2018 15:23:22 +0100 Subject: libceph: striping framework implementation Signed-off-by: Ilya Dryomov --- include/linux/ceph/striper.h | 65 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 65 insertions(+) create mode 100644 include/linux/ceph/striper.h (limited to 'include/linux') diff --git a/include/linux/ceph/striper.h b/include/linux/ceph/striper.h new file mode 100644 index 000000000000..74134ee5fdc8 --- /dev/null +++ b/include/linux/ceph/striper.h @@ -0,0 +1,65 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_CEPH_STRIPER_H +#define _LINUX_CEPH_STRIPER_H + +#include +#include + +struct ceph_file_layout; + +struct ceph_object_extent { + struct list_head oe_item; + u64 oe_objno; + u64 oe_off; + u64 oe_len; +}; + +static inline void ceph_object_extent_init(struct ceph_object_extent *ex) +{ + INIT_LIST_HEAD(&ex->oe_item); +} + +/* + * Called for each mapped stripe unit. + * + * @bytes: number of bytes mapped, i.e. the minimum of the full length + * requested (file extent length) or the remainder of the stripe + * unit within an object + */ +typedef void (*ceph_object_extent_fn_t)(struct ceph_object_extent *ex, + u32 bytes, void *arg); + +int ceph_file_to_extents(struct ceph_file_layout *l, u64 off, u64 len, + struct list_head *object_extents, + struct ceph_object_extent *alloc_fn(void *arg), + void *alloc_arg, + ceph_object_extent_fn_t action_fn, + void *action_arg); +int ceph_iterate_extents(struct ceph_file_layout *l, u64 off, u64 len, + struct list_head *object_extents, + ceph_object_extent_fn_t action_fn, + void *action_arg); + +struct ceph_file_extent { + u64 fe_off; + u64 fe_len; +}; + +static inline u64 ceph_file_extents_bytes(struct ceph_file_extent *file_extents, + u32 num_file_extents) +{ + u64 bytes = 0; + u32 i; + + for (i = 0; i < num_file_extents; i++) + bytes += file_extents[i].fe_len; + + return bytes; +} + +int ceph_extent_to_file(struct ceph_file_layout *l, + u64 objno, u64 objoff, u64 objlen, + struct ceph_file_extent **file_extents, + u32 *num_file_extents); + +#endif -- cgit v1.2.3 From 08c1ac508b6dc20ac866e7cdb7279245437c7d26 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Sat, 17 Feb 2018 10:41:20 +0100 Subject: libceph, ceph: move ceph_calc_file_object_mapping() to striper.c ceph_calc_file_object_mapping() has nothing to do with osdmaps. Signed-off-by: Ilya Dryomov --- include/linux/ceph/osdmap.h | 5 ----- include/linux/ceph/striper.h | 4 ++++ 2 files changed, 4 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 92314035dac1..e71fb222c7c3 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -5,7 +5,6 @@ #include #include #include -#include #include /* @@ -280,10 +279,6 @@ bool ceph_osds_changed(const struct ceph_osds *old_acting, const struct ceph_osds *new_acting, bool any_change); -void ceph_calc_file_object_mapping(struct ceph_file_layout *l, - u64 off, u64 len, - u64 *objno, u64 *objoff, u32 *xlen); - int __ceph_object_locator_to_pg(struct ceph_pg_pool_info *pi, const struct ceph_object_id *oid, const struct ceph_object_locator *oloc, diff --git a/include/linux/ceph/striper.h b/include/linux/ceph/striper.h index 74134ee5fdc8..cbd0d24b7148 100644 --- a/include/linux/ceph/striper.h +++ b/include/linux/ceph/striper.h @@ -7,6 +7,10 @@ struct ceph_file_layout; +void ceph_calc_file_object_mapping(struct ceph_file_layout *l, + u64 off, u64 len, + u64 *objno, u64 *objoff, u32 *xlen); + struct ceph_object_extent { struct list_head oe_item; u64 oe_objno; -- cgit v1.2.3 From bb48bd4dc45f9ee1e44d8e9fcb01023e0d0ba80d Mon Sep 17 00:00:00 2001 From: Chengguang Xu Date: Tue, 13 Mar 2018 10:42:44 +0800 Subject: ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov --- include/linux/ceph/libceph.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/linux') diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h index c2ec44cf5098..49c93b9308d7 100644 --- a/include/linux/ceph/libceph.h +++ b/include/linux/ceph/libceph.h @@ -262,6 +262,7 @@ extern struct kmem_cache *ceph_cap_cachep; extern struct kmem_cache *ceph_cap_flush_cachep; extern struct kmem_cache *ceph_dentry_cachep; extern struct kmem_cache *ceph_file_cachep; +extern struct kmem_cache *ceph_dir_file_cachep; /* ceph_common.c */ extern bool libceph_compatible(void *data); -- cgit v1.2.3 From fb18a57568c2b84cd611e242c0f6fa97b45e4907 Mon Sep 17 00:00:00 2001 From: Luis Henriques Date: Fri, 5 Jan 2018 10:47:18 +0000 Subject: ceph: quota: add initial infrastructure to support cephfs quotas This patch adds the infrastructure required to support cephfs quotas as it is currently implemented in the ceph fuse client. Cephfs quotas can be set on any directory, and can restrict the number of bytes or the number of files stored beneath that point in the directory hierarchy. Quotas are set using the extended attributes 'ceph.quota.max_files' and 'ceph.quota.max_bytes', and can be removed by setting these attributes to '0'. Link: http://tracker.ceph.com/issues/22372 Signed-off-by: Luis Henriques Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov --- include/linux/ceph/ceph_features.h | 1 + include/linux/ceph/ceph_fs.h | 17 +++++++++++++++++ 2 files changed, 18 insertions(+) (limited to 'include/linux') diff --git a/include/linux/ceph/ceph_features.h b/include/linux/ceph/ceph_features.h index 59042d5ac520..3901927cf6a0 100644 --- a/include/linux/ceph/ceph_features.h +++ b/include/linux/ceph/ceph_features.h @@ -204,6 +204,7 @@ DEFINE_CEPH_FEATURE_DEPRECATED(63, 1, RESERVED_BROKEN, LUMINOUS) // client-facin CEPH_FEATURE_OSD_PRIMARY_AFFINITY | \ CEPH_FEATURE_MSGR_KEEPALIVE2 | \ CEPH_FEATURE_OSD_POOLRESEND | \ + CEPH_FEATURE_MDS_QUOTA | \ CEPH_FEATURE_CRUSH_V4 | \ CEPH_FEATURE_NEW_OSDOP_ENCODING | \ CEPH_FEATURE_SERVER_JEWEL | \ diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index 88dd51381aaf..7ecfc88314d8 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -134,6 +134,7 @@ struct ceph_dir_layout { #define CEPH_MSG_CLIENT_LEASE 0x311 #define CEPH_MSG_CLIENT_SNAP 0x312 #define CEPH_MSG_CLIENT_CAPRELEASE 0x313 +#define CEPH_MSG_CLIENT_QUOTA 0x314 /* pool ops */ #define CEPH_MSG_POOLOP_REPLY 48 @@ -807,4 +808,20 @@ struct ceph_mds_snap_realm { } __attribute__ ((packed)); /* followed by my snap list, then prior parent snap list */ +/* + * quotas + */ +struct ceph_mds_quota { + __le64 ino; /* ino */ + struct ceph_timespec rctime; + __le64 rbytes; /* dir stats */ + __le64 rfiles; + __le64 rsubdirs; + __u8 struct_v; /* compat */ + __u8 struct_compat; + __le32 struct_len; + __le64 max_bytes; /* quota max. bytes */ + __le64 max_files; /* quota max. files */ +} __attribute__ ((packed)); + #endif -- cgit v1.2.3