From 7627151ea30bce2051e3cb27d7bb2c30083f86a5 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 3 Feb 2016 21:24:49 +0800 Subject: [PATCH] libceph: define new ceph_file_layout structure Define new ceph_file_layout structure and rename old ceph_file_layout to ceph_file_layout_legacy. This is preparation for adding namespace to ceph_file_layout structure. Signed-off-by: Yan, Zheng --- drivers/block/rbd.c | 14 +++++----- fs/ceph/addr.c | 18 ++++++------- fs/ceph/caps.c | 5 +++- fs/ceph/file.c | 6 ++--- fs/ceph/inode.c | 7 ++--- fs/ceph/ioctl.c | 22 ++++++++-------- fs/ceph/mds_client.h | 2 +- fs/ceph/xattr.c | 28 ++++++++------------ include/linux/ceph/ceph_fs.h | 50 +++++++++++++++--------------------- net/ceph/ceph_fs.c | 30 +++++++++++++++++++--- net/ceph/osd_client.c | 4 +-- net/ceph/osdmap.c | 6 ++--- 12 files changed, 103 insertions(+), 89 deletions(-) diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 81666a56415e..cc272ed46cfd 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1937,7 +1937,7 @@ static struct ceph_osd_request *rbd_osd_req_create( osd_req->r_callback = rbd_osd_req_callback; osd_req->r_priv = obj_request; - osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + osd_req->r_base_oloc.pool = rbd_dev->layout.pool_id; if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", obj_request->object_name)) goto fail; @@ -1991,7 +1991,7 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) osd_req->r_callback = rbd_osd_req_callback; osd_req->r_priv = obj_request; - osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + osd_req->r_base_oloc.pool = rbd_dev->layout.pool_id; if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", obj_request->object_name)) goto fail; @@ -3995,10 +3995,10 @@ static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc, /* Initialize the layout used for all rbd requests */ - rbd_dev->layout.fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); - rbd_dev->layout.fl_stripe_count = cpu_to_le32(1); - rbd_dev->layout.fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); - rbd_dev->layout.fl_pg_pool = cpu_to_le32((u32) spec->pool_id); + rbd_dev->layout.stripe_unit = 1 << RBD_MAX_OBJ_ORDER; + rbd_dev->layout.stripe_count = 1; + rbd_dev->layout.object_size = 1 << RBD_MAX_OBJ_ORDER; + rbd_dev->layout.pool_id = spec->pool_id; /* * If this is a mapping rbd_dev (as opposed to a parent one), @@ -5187,7 +5187,7 @@ static int rbd_dev_header_name(struct rbd_device *rbd_dev) rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); - rbd_dev->header_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + rbd_dev->header_oloc.pool = rbd_dev->layout.pool_id; if (rbd_dev->image_format == 1) ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s", spec->image_name, RBD_SUFFIX); diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 26a9d10d75e9..3f8efd866fec 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1730,7 +1730,7 @@ enum { POOL_WRITE = 2, }; -static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) +static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) { struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); struct ceph_mds_client *mdsc = fsc->mdsc; @@ -1757,7 +1757,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) if (*p) goto out; - dout("__ceph_pool_perm_get pool %u no perm cached\n", pool); + dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool); down_write(&mdsc->pool_perm_rwsem); parent = NULL; @@ -1860,13 +1860,13 @@ out_unlock: out: if (!err) err = have; - dout("__ceph_pool_perm_get pool %u result = %d\n", pool, err); + dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err); return err; } int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) { - u32 pool; + s64 pool; int ret, flags; /* does not support pool namespace yet */ @@ -1879,17 +1879,17 @@ int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) spin_lock(&ci->i_ceph_lock); flags = ci->i_ceph_flags; - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; spin_unlock(&ci->i_ceph_lock); check: if (flags & CEPH_I_POOL_PERM) { if ((need & CEPH_CAP_FILE_RD) && !(flags & CEPH_I_POOL_RD)) { - dout("ceph_pool_perm_check pool %u no read perm\n", + dout("ceph_pool_perm_check pool %lld no read perm\n", pool); return -EPERM; } if ((need & CEPH_CAP_FILE_WR) && !(flags & CEPH_I_POOL_WR)) { - dout("ceph_pool_perm_check pool %u no write perm\n", + dout("ceph_pool_perm_check pool %lld no write perm\n", pool); return -EPERM; } @@ -1907,10 +1907,10 @@ check: flags |= CEPH_I_POOL_WR; spin_lock(&ci->i_ceph_lock); - if (pool == ceph_file_layout_pg_pool(ci->i_layout)) { + if (pool == ci->i_layout.pool_id) { ci->i_ceph_flags = flags; } else { - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; flags = ci->i_ceph_flags; } spin_unlock(&ci->i_ceph_lock); diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 6f60d0a3d0f9..f24722dce167 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2895,8 +2895,11 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (newcaps & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR)) { /* file layout may have changed */ - ci->i_layout = grant->layout; + s64 old_pool = ci->i_layout.pool_id; + ceph_file_layout_from_legacy(&ci->i_layout, &grant->layout); ci->i_pool_ns_len = pool_ns_len; + if (ci->i_layout.pool_id != old_pool) + ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; /* size/truncate_seq? */ queue_trunc = ceph_fill_file_size(inode, issued, diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 0daaf7ceedc5..cba5dcf49a65 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -1583,9 +1583,9 @@ static int ceph_zero_objects(struct inode *inode, loff_t offset, loff_t length) { int ret = 0; struct ceph_inode_info *ci = ceph_inode(inode); - s32 stripe_unit = ceph_file_layout_su(ci->i_layout); - s32 stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - s32 object_size = ceph_file_layout_object_size(ci->i_layout); + s32 stripe_unit = ci->i_layout.stripe_unit; + s32 stripe_count = ci->i_layout.stripe_count; + s32 object_size = ci->i_layout.object_size; u64 object_set_size = object_size * stripe_count; u64 nearly, t; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index f059b5997072..6c5903e0a976 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -814,10 +814,11 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (new_version || (new_issued & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { - if (ci->i_layout.fl_pg_pool != info->layout.fl_pg_pool) - ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; - ci->i_layout = info->layout; + s64 old_pool = ci->i_layout.pool_id; + ceph_file_layout_from_legacy(&ci->i_layout, &info->layout); ci->i_pool_ns_len = iinfo->pool_ns_len; + if (ci->i_layout.pool_id != old_pool) + ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; queue_trunc = ceph_fill_file_size(inode, issued, le32_to_cpu(info->truncate_seq), diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 0946f2d4a81f..843dd31a02cd 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -21,10 +21,10 @@ static long ceph_ioctl_get_layout(struct file *file, void __user *arg) err = ceph_do_getattr(file_inode(file), CEPH_STAT_CAP_LAYOUT, false); if (!err) { - l.stripe_unit = ceph_file_layout_su(ci->i_layout); - l.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - l.object_size = ceph_file_layout_object_size(ci->i_layout); - l.data_pool = le32_to_cpu(ci->i_layout.fl_pg_pool); + l.stripe_unit = ci->i_layout.stripe_unit; + l.stripe_count = ci->i_layout.stripe_count; + l.object_size = ci->i_layout.object_size; + l.data_pool = ci->i_layout.pool_id; l.preferred_osd = (s32)-1; if (copy_to_user(arg, &l, sizeof(l))) return -EFAULT; @@ -82,19 +82,19 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) if (l.stripe_count) nl.stripe_count = l.stripe_count; else - nl.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); + nl.stripe_count = ci->i_layout.stripe_count; if (l.stripe_unit) nl.stripe_unit = l.stripe_unit; else - nl.stripe_unit = ceph_file_layout_su(ci->i_layout); + nl.stripe_unit = ci->i_layout.stripe_unit; if (l.object_size) nl.object_size = l.object_size; else - nl.object_size = ceph_file_layout_object_size(ci->i_layout); + nl.object_size = ci->i_layout.object_size; if (l.data_pool) nl.data_pool = l.data_pool; else - nl.data_pool = ceph_file_layout_pg_pool(ci->i_layout); + nl.data_pool = ci->i_layout.pool_id; /* this is obsolete, and always -1 */ nl.preferred_osd = le64_to_cpu(-1); @@ -202,8 +202,8 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) return -EIO; } dl.file_offset -= dl.object_offset; - dl.object_size = ceph_file_layout_object_size(ci->i_layout); - dl.block_size = ceph_file_layout_su(ci->i_layout); + dl.object_size = ci->i_layout.object_size; + dl.block_size = ci->i_layout.stripe_unit; /* block_offset = object_offset % block_size */ tmp = dl.object_offset; @@ -212,7 +212,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) snprintf(dl.object_name, sizeof(dl.object_name), "%llx.%08llx", ceph_ino(inode), dl.object_no); - oloc.pool = ceph_file_layout_pg_pool(ci->i_layout); + oloc.pool = ci->i_layout.pool_id; ceph_oid_printf(&oid, "%s", dl.object_name); r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index e7d38aac7109..75ecf967d0b7 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -275,8 +275,8 @@ struct ceph_mds_request { struct ceph_pool_perm { struct rb_node node; - u32 pool; int perm; + s64 pool; }; /* diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 4870b29df224..5377c9c7a0c5 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -72,7 +72,7 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + s64 pool = ci->i_layout.pool_id; const char *pool_name; char buf[128]; @@ -82,10 +82,9 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, if (pool_name) { size_t len = strlen(pool_name); ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); + "stripe_unit=%u stripe_count=%u object_size=%u pool=", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size); if (!size) { ret += len; } else if (ret + len > size) { @@ -97,11 +96,9 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, } } else { ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout), - (unsigned long long)pool); + "stripe_unit=%u stripe_count=%u object_size=%u pool=%lld", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size, (unsigned long long)pool); if (size) { if (ret <= size) memcpy(val, buf, ret); @@ -116,22 +113,19 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, static size_t ceph_vxattrcb_layout_stripe_unit(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_unit); } static size_t ceph_vxattrcb_layout_stripe_count(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_count); } static size_t ceph_vxattrcb_layout_object_size(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.object_size); } static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, @@ -140,7 +134,7 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + s64 pool = ci->i_layout.pool_id; const char *pool_name; down_read(&osdc->lock); diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index dfce616002ad..e5a5fb9ca3f5 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -34,9 +34,9 @@ #define CEPH_MAX_MON 31 /* - * ceph_file_layout - describe data layout for a file/inode + * legacy ceph_file_layoute */ -struct ceph_file_layout { +struct ceph_file_layout_legacy { /* file -> object mapping */ __le32 fl_stripe_unit; /* stripe unit, in bytes. must be multiple of page size. */ @@ -53,33 +53,25 @@ struct ceph_file_layout { __le32 fl_pg_pool; /* namespace, crush ruleset, rep level */ } __attribute__ ((packed)); -#define ceph_file_layout_su(l) ((__s32)le32_to_cpu((l).fl_stripe_unit)) -#define ceph_file_layout_stripe_count(l) \ - ((__s32)le32_to_cpu((l).fl_stripe_count)) -#define ceph_file_layout_object_size(l) ((__s32)le32_to_cpu((l).fl_object_size)) -#define ceph_file_layout_cas_hash(l) ((__s32)le32_to_cpu((l).fl_cas_hash)) -#define ceph_file_layout_object_su(l) \ - ((__s32)le32_to_cpu((l).fl_object_stripe_unit)) -#define ceph_file_layout_pg_pool(l) \ - ((__s32)le32_to_cpu((l).fl_pg_pool)) - -static inline unsigned ceph_file_layout_stripe_width(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_stripe_unit) * - le32_to_cpu(l->fl_stripe_count); -} - -/* "period" == bytes before i start on a new set of objects */ -static inline unsigned ceph_file_layout_period(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_object_size) * - le32_to_cpu(l->fl_stripe_count); -} +/* + * ceph_file_layout - describe data layout for a file/inode + */ +struct ceph_file_layout { + /* file -> object mapping */ + u32 stripe_unit; /* stripe unit, in bytes */ + u32 stripe_count; /* over this many objects */ + u32 object_size; /* until objects are this big */ + s64 pool_id; /* rados pool id */ +}; + +extern int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); +extern void ceph_file_layout_from_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy); +extern void ceph_file_layout_to_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy); #define CEPH_MIN_STRIPE_UNIT 65536 -int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); - struct ceph_dir_layout { __u8 dl_dir_hash; /* see ceph_hash.h for ids */ __u8 dl_unused1; @@ -399,7 +391,7 @@ union ceph_mds_request_args { __le32 flags; } __attribute__ ((packed)) setxattr; struct { - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; } __attribute__ ((packed)) setlayout; struct { __u8 rule; /* currently fcntl or flock */ @@ -478,7 +470,7 @@ struct ceph_mds_reply_inode { __le64 version; /* inode version */ __le64 xattr_version; /* version for xattr blob */ struct ceph_mds_reply_cap cap; /* caps issued for this inode */ - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; struct ceph_timespec ctime, mtime, atime; __le32 time_warp_seq; __le64 size, max_size, truncate_size; @@ -673,7 +665,7 @@ struct ceph_mds_caps { __le64 size, max_size, truncate_size; __le32 truncate_seq; struct ceph_timespec mtime, atime, ctime; - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; __le32 time_warp_seq; } __attribute__ ((packed)); diff --git a/net/ceph/ceph_fs.c b/net/ceph/ceph_fs.c index 41466ccb972a..7d54e944de5e 100644 --- a/net/ceph/ceph_fs.c +++ b/net/ceph/ceph_fs.c @@ -9,9 +9,9 @@ */ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) { - __u32 su = le32_to_cpu(layout->fl_stripe_unit); - __u32 sc = le32_to_cpu(layout->fl_stripe_count); - __u32 os = le32_to_cpu(layout->fl_object_size); + __u32 su = layout->stripe_unit; + __u32 sc = layout->stripe_count; + __u32 os = layout->object_size; /* stripe unit, object size must be non-zero, 64k increment */ if (!su || (su & (CEPH_MIN_STRIPE_UNIT-1))) @@ -27,6 +27,30 @@ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) return 1; } +void ceph_file_layout_from_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy) +{ + fl->stripe_unit = le32_to_cpu(legacy->fl_stripe_unit); + fl->stripe_count = le32_to_cpu(legacy->fl_stripe_count); + fl->object_size = le32_to_cpu(legacy->fl_object_size); + fl->pool_id = le32_to_cpu(legacy->fl_pg_pool); + if (fl->pool_id == 0) + fl->pool_id = -1; +} +EXPORT_SYMBOL(ceph_file_layout_from_legacy); + +void ceph_file_layout_to_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy) +{ + legacy->fl_stripe_unit = cpu_to_le32(fl->stripe_unit); + legacy->fl_stripe_count = cpu_to_le32(fl->stripe_count); + legacy->fl_object_size = cpu_to_le32(fl->object_size); + if (fl->pool_id >= 0) + legacy->fl_pg_pool = cpu_to_le32(fl->pool_id); + else + legacy->fl_pg_pool = 0; +} +EXPORT_SYMBOL(ceph_file_layout_to_legacy); int ceph_flags_to_mode(int flags) { diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 89469592076c..23efcac80072 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -932,7 +932,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) { osd_req_op_init(req, which, opcode, 0); } else { - u32 object_size = le32_to_cpu(layout->fl_object_size); + u32 object_size = layout->object_size; u32 object_base = off - objoff; if (!(truncate_seq == 1 && truncate_size == -1ULL)) { if (truncate_size <= object_base) { @@ -948,7 +948,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, } req->r_flags = flags; - req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout); + req->r_base_oloc.pool = layout->pool_id; ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum); req->r_snapid = vino.snap; diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 7e480bf75bcf..5947b2e9fb8e 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1770,9 +1770,9 @@ int ceph_calc_file_object_mapping(struct ceph_file_layout *layout, u64 *ono, u64 *oxoff, u64 *oxlen) { - u32 osize = le32_to_cpu(layout->fl_object_size); - u32 su = le32_to_cpu(layout->fl_stripe_unit); - u32 sc = le32_to_cpu(layout->fl_stripe_count); + u32 osize = layout->object_size; + u32 su = layout->stripe_unit; + u32 sc = layout->stripe_count; u32 bl, stripeno, stripepos, objsetno; u32 su_per_object; u64 t, su_offset; -- 2.39.2