Hi, in this thread comes the third version of the patches implementing quotas for OCFS2 (sorry for spamming ocfs2-devel so often with such a big pile of patches ;). I've now addressed all Joel's comments (feature bits added among supported bits in the last patch, hopefully fixed an oops when enabling quota fails on mount), also fixed a bug Tristan reported (message about exceeding softlimit was not always properly printed), fixed a bug in node recovery leading to possible oops when filesystem was without quotas (reported by Andrew Beekhof) and also reported a few style things and compilation problems reported by Andrew Morton and other people getting the patches via -mm kernel. As always, comments are welcome (but I won't respond quickly because I'll be mostly out of email for next two weeks). Honza
Jan Kara
2008-Oct-31 06:38 UTC
[Ocfs2-devel] [PATCH 01/27] quota: Add callbacks for allocating and destroying dquot structures
Some filesystems would like to keep private information together with each dquot. Add callbacks alloc_dquot and destroy_dquot allowing filesystem to allocate larger dquots from their private slab in a similar fashion we currently allocate inodes. Signed-off-by: Jan Kara <jack at suse.cz> --- fs/dquot.c | 27 ++++++++++++++++++++++----- include/linux/quota.h | 2 ++ 2 files changed, 24 insertions(+), 5 deletions(-) diff --git a/fs/dquot.c b/fs/dquot.c index 5e95261..5512e38 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -415,6 +415,16 @@ out_dqlock: return ret; } +static void dquot_destroy(struct dquot *dquot) +{ + kmem_cache_free(dquot_cachep, dquot); +} + +static inline void do_destroy_dquot(struct dquot *dquot) +{ + dquot->dq_sb->dq_op->destroy_dquot(dquot); +} + /* Invalidate all dquots on the list. Note that this function is called after * quota is disabled and pointers from inodes removed so there cannot be new * quota users. There can still be some users of quotas due to inodes being @@ -463,7 +473,7 @@ restart: remove_dquot_hash(dquot); remove_free_dquot(dquot); remove_inuse(dquot); - kmem_cache_free(dquot_cachep, dquot); + do_destroy_dquot(dquot); } spin_unlock(&dq_list_lock); } @@ -527,7 +537,7 @@ static void prune_dqcache(int count) remove_dquot_hash(dquot); remove_free_dquot(dquot); remove_inuse(dquot); - kmem_cache_free(dquot_cachep, dquot); + do_destroy_dquot(dquot); count--; head = free_dquots.prev; } @@ -625,11 +635,16 @@ we_slept: spin_unlock(&dq_list_lock); } +static struct dquot *dquot_alloc(struct super_block *sb, int type) +{ + return kmem_cache_zalloc(dquot_cachep, GFP_NOFS); +} + static struct dquot *get_empty_dquot(struct super_block *sb, int type) { struct dquot *dquot; - dquot = kmem_cache_zalloc(dquot_cachep, GFP_NOFS); + dquot = sb->dq_op->alloc_dquot(sb, type); if(!dquot) return NODQUOT; @@ -682,7 +697,7 @@ we_slept: dqstats.lookups++; spin_unlock(&dq_list_lock); if (empty) - kmem_cache_free(dquot_cachep, empty); + do_destroy_dquot(empty); } /* Wait for dq_lock - after this we know that either dquot_release() is already * finished or it will be canceled due to dq_count > 1 test */ @@ -1533,7 +1548,9 @@ struct dquot_operations dquot_operations = { .acquire_dquot = dquot_acquire, .release_dquot = dquot_release, .mark_dirty = dquot_mark_dquot_dirty, - .write_info = dquot_commit_info + .write_info = dquot_commit_info, + .alloc_dquot = dquot_alloc, + .destroy_dquot = dquot_destroy, }; static inline void set_enable_flags(struct quota_info *dqopt, int type) diff --git a/include/linux/quota.h b/include/linux/quota.h index 40401b5..3ce708c 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -292,6 +292,8 @@ struct dquot_operations { int (*free_inode) (const struct inode *, unsigned long); int (*transfer) (struct inode *, struct iattr *); int (*write_dquot) (struct dquot *); /* Ordinary dquot write */ + struct dquot *(*alloc_dquot)(struct super_block *, int); /* Allocate memory for new dquot */ + void (*destroy_dquot)(struct dquot *); /* Free memory for dquot */ int (*acquire_dquot) (struct dquot *); /* Quota is going to be created on disk */ int (*release_dquot) (struct dquot *); /* Quota is going to be deleted from disk */ int (*mark_dirty) (struct dquot *); /* Dquot is marked dirty */ -- 1.5.2.4
Jan Kara
2008-Oct-31 06:38 UTC
[Ocfs2-devel] [PATCH 02/27] quota: Increase size of variables for limits and inode usage
So far quota was fine with quota block limits and inode limits/numbers in a 32-bit type. Now with rapid increase in storage sizes there are coming requests to be able to handle quota limits above 4TB / more that 2^32 inodes. So bump up sizes of types in mem_dqblk structure to 64-bits to be able to handle this. Also update inode allocation / checking functions to use qsize_t and make global structure keep quota limits in bytes so that things are consistent. Signed-off-by: Jan Kara <jack at suse.cz> --- fs/dquot.c | 50 ++++++++++++++++++++++++++------------------- fs/quota_v1.c | 25 +++++++++++++++++----- fs/quota_v2.c | 21 +++++++++++++++--- include/linux/quota.h | 28 +++++++++++-------------- include/linux/quotaops.h | 4 +- 5 files changed, 79 insertions(+), 49 deletions(-) diff --git a/fs/dquot.c b/fs/dquot.c index 5512e38..59e8aea 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -835,7 +835,7 @@ static void drop_dquot_ref(struct super_block *sb, int type) } } -static inline void dquot_incr_inodes(struct dquot *dquot, unsigned long number) +static inline void dquot_incr_inodes(struct dquot *dquot, qsize_t number) { dquot->dq_dqb.dqb_curinodes += number; } @@ -845,7 +845,7 @@ static inline void dquot_incr_space(struct dquot *dquot, qsize_t number) dquot->dq_dqb.dqb_curspace += number; } -static inline void dquot_decr_inodes(struct dquot *dquot, unsigned long number) +static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number) { if (dquot->dq_dqb.dqb_curinodes > number) dquot->dq_dqb.dqb_curinodes -= number; @@ -862,7 +862,7 @@ static inline void dquot_decr_space(struct dquot *dquot, qsize_t number) dquot->dq_dqb.dqb_curspace -= number; else dquot->dq_dqb.dqb_curspace = 0; - if (toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit) + if (dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit) dquot->dq_dqb.dqb_btime = (time_t) 0; clear_bit(DQ_BLKS_B, &dquot->dq_flags); } @@ -1038,7 +1038,7 @@ static inline char ignore_hardlimit(struct dquot *dquot) } /* needs dq_data_lock */ -static int check_idq(struct dquot *dquot, ulong inodes, char *warntype) +static int check_idq(struct dquot *dquot, qsize_t inodes, char *warntype) { *warntype = QUOTA_NL_NOWARN; if (inodes <= 0 || test_bit(DQ_FAKE_B, &dquot->dq_flags)) @@ -1077,7 +1077,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war return QUOTA_OK; if (dquot->dq_dqb.dqb_bhardlimit && - toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bhardlimit && + dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bhardlimit && !ignore_hardlimit(dquot)) { if (!prealloc) *warntype = QUOTA_NL_BHARDWARN; @@ -1085,7 +1085,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war } if (dquot->dq_dqb.dqb_bsoftlimit && - toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit && + dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit && dquot->dq_dqb.dqb_btime && get_seconds() >= dquot->dq_dqb.dqb_btime && !ignore_hardlimit(dquot)) { if (!prealloc) @@ -1094,7 +1094,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war } if (dquot->dq_dqb.dqb_bsoftlimit && - toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit && + dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit && dquot->dq_dqb.dqb_btime == 0) { if (!prealloc) { *warntype = QUOTA_NL_BSOFTWARN; @@ -1111,7 +1111,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war return QUOTA_OK; } -static int info_idq_free(struct dquot *dquot, ulong inodes) +static int info_idq_free(struct dquot *dquot, qsize_t inodes) { if (test_bit(DQ_FAKE_B, &dquot->dq_flags) || dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit) @@ -1128,15 +1128,13 @@ static int info_idq_free(struct dquot *dquot, ulong inodes) static int info_bdq_free(struct dquot *dquot, qsize_t space) { if (test_bit(DQ_FAKE_B, &dquot->dq_flags) || - toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit) + dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit) return QUOTA_NL_NOWARN; - if (toqb(dquot->dq_dqb.dqb_curspace - space) <- dquot->dq_dqb.dqb_bsoftlimit) + if (dquot->dq_dqb.dqb_curspace - space <= dquot->dq_dqb.dqb_bsoftlimit) return QUOTA_NL_BSOFTBELOW; - if (toqb(dquot->dq_dqb.dqb_curspace) >= dquot->dq_dqb.dqb_bhardlimit && - toqb(dquot->dq_dqb.dqb_curspace - space) < - dquot->dq_dqb.dqb_bhardlimit) + if (dquot->dq_dqb.dqb_curspace >= dquot->dq_dqb.dqb_bhardlimit && + dquot->dq_dqb.dqb_curspace - space < dquot->dq_dqb.dqb_bhardlimit) return QUOTA_NL_BHARDBELOW; return QUOTA_NL_NOWARN; } @@ -1279,7 +1277,7 @@ warn_put_all: /* * This operation can block, but only after everything is updated */ -int dquot_alloc_inode(const struct inode *inode, unsigned long number) +int dquot_alloc_inode(const struct inode *inode, qsize_t number) { int cnt, ret = NO_QUOTA; char warntype[MAXQUOTAS]; @@ -1364,7 +1362,7 @@ out_sub: /* * This operation can block, but only after everything is updated */ -int dquot_free_inode(const struct inode *inode, unsigned long number) +int dquot_free_inode(const struct inode *inode, qsize_t number) { unsigned int cnt; char warntype[MAXQUOTAS]; @@ -1883,14 +1881,24 @@ int vfs_dq_quota_on_remount(struct super_block *sb) return ret; } +static inline qsize_t qbtos(qsize_t blocks) +{ + return blocks << QIF_DQBLKSIZE_BITS; +} + +static inline qsize_t stoqb(qsize_t space) +{ + return (space + QIF_DQBLKSIZE - 1) >> QIF_DQBLKSIZE_BITS; +} + /* Generic routine for getting common part of quota structure */ static void do_get_dqblk(struct dquot *dquot, struct if_dqblk *di) { struct mem_dqblk *dm = &dquot->dq_dqb; spin_lock(&dq_data_lock); - di->dqb_bhardlimit = dm->dqb_bhardlimit; - di->dqb_bsoftlimit = dm->dqb_bsoftlimit; + di->dqb_bhardlimit = stoqb(dm->dqb_bhardlimit); + di->dqb_bsoftlimit = stoqb(dm->dqb_bsoftlimit); di->dqb_curspace = dm->dqb_curspace; di->dqb_ihardlimit = dm->dqb_ihardlimit; di->dqb_isoftlimit = dm->dqb_isoftlimit; @@ -1937,8 +1945,8 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) check_blim = 1; } if (di->dqb_valid & QIF_BLIMITS) { - dm->dqb_bsoftlimit = di->dqb_bsoftlimit; - dm->dqb_bhardlimit = di->dqb_bhardlimit; + dm->dqb_bsoftlimit = qbtos(di->dqb_bsoftlimit); + dm->dqb_bhardlimit = qbtos(di->dqb_bhardlimit); check_blim = 1; } if (di->dqb_valid & QIF_INODES) { @@ -1956,7 +1964,7 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) dm->dqb_itime = di->dqb_itime; if (check_blim) { - if (!dm->dqb_bsoftlimit || toqb(dm->dqb_curspace) < dm->dqb_bsoftlimit) { + if (!dm->dqb_bsoftlimit || dm->dqb_curspace < dm->dqb_bsoftlimit) { dm->dqb_btime = 0; clear_bit(DQ_BLKS_B, &dquot->dq_flags); } diff --git a/fs/quota_v1.c b/fs/quota_v1.c index 5ae15b1..3e078ee 100644 --- a/fs/quota_v1.c +++ b/fs/quota_v1.c @@ -14,14 +14,27 @@ MODULE_AUTHOR("Jan Kara"); MODULE_DESCRIPTION("Old quota format support"); MODULE_LICENSE("GPL"); +#define QUOTABLOCK_BITS 10 +#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) + +static inline qsize_t v1_stoqb(qsize_t space) +{ + return (space + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS; +} + +static inline qsize_t v1_qbtos(qsize_t blocks) +{ + return blocks << QUOTABLOCK_BITS; +} + static void v1_disk2mem_dqblk(struct mem_dqblk *m, struct v1_disk_dqblk *d) { m->dqb_ihardlimit = d->dqb_ihardlimit; m->dqb_isoftlimit = d->dqb_isoftlimit; m->dqb_curinodes = d->dqb_curinodes; - m->dqb_bhardlimit = d->dqb_bhardlimit; - m->dqb_bsoftlimit = d->dqb_bsoftlimit; - m->dqb_curspace = ((qsize_t)d->dqb_curblocks) << QUOTABLOCK_BITS; + m->dqb_bhardlimit = v1_qbtos(d->dqb_bhardlimit); + m->dqb_bsoftlimit = v1_qbtos(d->dqb_bsoftlimit); + m->dqb_curspace = v1_qbtos(d->dqb_curblocks); m->dqb_itime = d->dqb_itime; m->dqb_btime = d->dqb_btime; } @@ -31,9 +44,9 @@ static void v1_mem2disk_dqblk(struct v1_disk_dqblk *d, struct mem_dqblk *m) d->dqb_ihardlimit = m->dqb_ihardlimit; d->dqb_isoftlimit = m->dqb_isoftlimit; d->dqb_curinodes = m->dqb_curinodes; - d->dqb_bhardlimit = m->dqb_bhardlimit; - d->dqb_bsoftlimit = m->dqb_bsoftlimit; - d->dqb_curblocks = toqb(m->dqb_curspace); + d->dqb_bhardlimit = v1_stoqb(m->dqb_bhardlimit); + d->dqb_bsoftlimit = v1_stoqb(m->dqb_bsoftlimit); + d->dqb_curblocks = v1_stoqb(m->dqb_curspace); d->dqb_itime = m->dqb_itime; d->dqb_btime = m->dqb_btime; } diff --git a/fs/quota_v2.c b/fs/quota_v2.c index b53827d..51c4717 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -26,6 +26,19 @@ typedef char *dqbuf_t; #define GETIDINDEX(id, depth) (((id) >> ((V2_DQTREEDEPTH-(depth)-1)*8)) & 0xff) #define GETENTRIES(buf) ((struct v2_disk_dqblk *)(((char *)buf)+sizeof(struct v2_disk_dqdbheader))) +#define QUOTABLOCK_BITS 10 +#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) + +static inline qsize_t v2_stoqb(qsize_t space) +{ + return (space + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS; +} + +static inline qsize_t v2_qbtos(qsize_t blocks) +{ + return blocks << QUOTABLOCK_BITS; +} + /* Check whether given file is really vfsv0 quotafile */ static int v2_check_quota_file(struct super_block *sb, int type) { @@ -104,8 +117,8 @@ static void disk2memdqb(struct mem_dqblk *m, struct v2_disk_dqblk *d) m->dqb_isoftlimit = le32_to_cpu(d->dqb_isoftlimit); m->dqb_curinodes = le32_to_cpu(d->dqb_curinodes); m->dqb_itime = le64_to_cpu(d->dqb_itime); - m->dqb_bhardlimit = le32_to_cpu(d->dqb_bhardlimit); - m->dqb_bsoftlimit = le32_to_cpu(d->dqb_bsoftlimit); + m->dqb_bhardlimit = v2_qbtos(le32_to_cpu(d->dqb_bhardlimit)); + m->dqb_bsoftlimit = v2_qbtos(le32_to_cpu(d->dqb_bsoftlimit)); m->dqb_curspace = le64_to_cpu(d->dqb_curspace); m->dqb_btime = le64_to_cpu(d->dqb_btime); } @@ -116,8 +129,8 @@ static void mem2diskdqb(struct v2_disk_dqblk *d, struct mem_dqblk *m, qid_t id) d->dqb_isoftlimit = cpu_to_le32(m->dqb_isoftlimit); d->dqb_curinodes = cpu_to_le32(m->dqb_curinodes); d->dqb_itime = cpu_to_le64(m->dqb_itime); - d->dqb_bhardlimit = cpu_to_le32(m->dqb_bhardlimit); - d->dqb_bsoftlimit = cpu_to_le32(m->dqb_bsoftlimit); + d->dqb_bhardlimit = cpu_to_le32(v2_qbtos(m->dqb_bhardlimit)); + d->dqb_bsoftlimit = cpu_to_le32(v2_qbtos(m->dqb_bsoftlimit)); d->dqb_curspace = cpu_to_le64(m->dqb_curspace); d->dqb_btime = cpu_to_le64(m->dqb_btime); d->dqb_id = cpu_to_le32(id); diff --git a/include/linux/quota.h b/include/linux/quota.h index 3ce708c..9ea4683 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -39,15 +39,6 @@ #define __DQUOT_VERSION__ "dquot_6.5.1" #define __DQUOT_NUM_VERSION__ 6*10000+5*100+1 -/* Size of blocks in which are counted size limits */ -#define QUOTABLOCK_BITS 10 -#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) - -/* Conversion routines from and to quota blocks */ -#define qb2kb(x) ((x) << (QUOTABLOCK_BITS-10)) -#define kb2qb(x) ((x) >> (QUOTABLOCK_BITS-10)) -#define toqb(x) (((x) + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS) - #define MAXQUOTAS 2 #define USRQUOTA 0 /* element used for user quotas */ #define GRPQUOTA 1 /* element used for group quotas */ @@ -80,6 +71,11 @@ #define Q_GETQUOTA 0x800007 /* get user quota structure */ #define Q_SETQUOTA 0x800008 /* set user quota structure */ +/* Size of block in which space limits are passed through the quota + * interface */ +#define QIF_DQBLKSIZE_BITS 10 +#define QIF_DQBLKSIZE (1 << QIF_DQBLKSIZE_BITS) + /* * Quota structure used for communication with userspace via quotactl * Following flags are used to specify which fields are valid @@ -187,12 +183,12 @@ extern spinlock_t dq_data_lock; * Data for one user/group kept in memory */ struct mem_dqblk { - __u32 dqb_bhardlimit; /* absolute limit on disk blks alloc */ - __u32 dqb_bsoftlimit; /* preferred limit on disk blks */ + qsize_t dqb_bhardlimit; /* absolute limit on disk blks alloc */ + qsize_t dqb_bsoftlimit; /* preferred limit on disk blks */ qsize_t dqb_curspace; /* current used space */ - __u32 dqb_ihardlimit; /* absolute limit on allocated inodes */ - __u32 dqb_isoftlimit; /* preferred inode limit */ - __u32 dqb_curinodes; /* current # allocated inodes */ + qsize_t dqb_ihardlimit; /* absolute limit on allocated inodes */ + qsize_t dqb_isoftlimit; /* preferred inode limit */ + qsize_t dqb_curinodes; /* current # allocated inodes */ time_t dqb_btime; /* time limit for excessive disk use */ time_t dqb_itime; /* time limit for excessive inode use */ }; @@ -287,9 +283,9 @@ struct dquot_operations { int (*initialize) (struct inode *, int); int (*drop) (struct inode *); int (*alloc_space) (struct inode *, qsize_t, int); - int (*alloc_inode) (const struct inode *, unsigned long); + int (*alloc_inode) (const struct inode *, qsize_t); int (*free_space) (struct inode *, qsize_t); - int (*free_inode) (const struct inode *, unsigned long); + int (*free_inode) (const struct inode *, qsize_t); int (*transfer) (struct inode *, struct iattr *); int (*write_dquot) (struct dquot *); /* Ordinary dquot write */ struct dquot *(*alloc_dquot)(struct super_block *, int); /* Allocate memory for new dquot */ diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index a558a4c..adcc7ba 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -26,10 +26,10 @@ int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); -int dquot_alloc_inode(const struct inode *inode, unsigned long number); +int dquot_alloc_inode(const struct inode *inode, qsize_t number); int dquot_free_space(struct inode *inode, qsize_t number); -int dquot_free_inode(const struct inode *inode, unsigned long number); +int dquot_free_inode(const struct inode *inode, qsize_t number); int dquot_transfer(struct inode *inode, struct iattr *iattr); int dquot_commit(struct dquot *dquot); -- 1.5.2.4
Jan Kara
2008-Oct-31 06:38 UTC
[Ocfs2-devel] [PATCH 07/27] ext4: Use sb_any_quota_loaded() instead of sb_any_quota_enabled()
Signed-off-by: Jan Kara <jack at suse.cz> --- fs/ext4/super.c | 11 ++++------- 1 files changed, 4 insertions(+), 7 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index bdddea1..be34b66 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1139,8 +1139,7 @@ static int parse_options(char *options, struct super_block *sb, case Opt_grpjquota: qtype = GRPQUOTA; set_qf_name: - if ((sb_any_quota_enabled(sb) || - sb_any_quota_suspended(sb)) && + if (sb_any_quota_loaded(sb) && !sbi->s_qf_names[qtype]) { printk(KERN_ERR "EXT4-fs: Cannot change journaled " @@ -1179,8 +1178,7 @@ set_qf_name: case Opt_offgrpjquota: qtype = GRPQUOTA; clear_qf_name: - if ((sb_any_quota_enabled(sb) || - sb_any_quota_suspended(sb)) && + if (sb_any_quota_loaded(sb) && sbi->s_qf_names[qtype]) { printk(KERN_ERR "EXT4-fs: Cannot change " "journaled quota options when " @@ -1199,8 +1197,7 @@ clear_qf_name: case Opt_jqfmt_vfsv0: qfmt = QFMT_VFS_V0; set_qf_format: - if ((sb_any_quota_enabled(sb) || - sb_any_quota_suspended(sb)) && + if (sb_any_quota_loaded(sb) && sbi->s_jquota_fmt != qfmt) { printk(KERN_ERR "EXT4-fs: Cannot change " "journaled quota options when " @@ -1219,7 +1216,7 @@ set_qf_format: set_opt(sbi->s_mount_opt, GRPQUOTA); break; case Opt_noquota: - if (sb_any_quota_enabled(sb)) { + if (sb_any_quota_loaded(sb)) { printk(KERN_ERR "EXT4-fs: Cannot change quota " "options when quota turned on.\n"); return 0; -- 1.5.2.4
Increase reported version number of quota support since quota core has changed significantly. Also remove __DQUOT_NUM_VERSION__ since nobody uses it. Signed-off-by: Jan Kara <jack at suse.cz> --- include/linux/quota.h | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/include/linux/quota.h b/include/linux/quota.h index 6d98885..ec82beb 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -36,8 +36,7 @@ #include <linux/errno.h> #include <linux/types.h> -#define __DQUOT_VERSION__ "dquot_6.5.1" -#define __DQUOT_NUM_VERSION__ 6*10000+5*100+1 +#define __DQUOT_VERSION__ "dquot_6.5.2" #define MAXQUOTAS 2 #define USRQUOTA 0 /* element used for user quotas */ -- 1.5.2.4
Jan Kara
2008-Oct-31 06:39 UTC
[Ocfs2-devel] [PATCH 17/27] quota: Add helpers to allow ocfs2 specific quota initialization, freeing and recovery
OCFS2 needs to peek whether quota structure is already in memory so that it can avoid expensive cluster locking in that case. Similarly when freeing dquots, it checks whether it is the last quota structure user or not. Finally, it needs to get reference to dquot structure for specified id and quota type when recovering quota file after crash. Signed-off-by: Jan Kara <jack at suse.cz> --- fs/dquot.c | 38 ++++++++++++++++++++++++++++++++------ include/linux/quotaops.h | 4 ++++ 2 files changed, 36 insertions(+), 6 deletions(-) diff --git a/fs/dquot.c b/fs/dquot.c index 6784892..8774be6 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -211,8 +211,6 @@ static struct hlist_head *dquot_hash; struct dqstats dqstats; -static void dqput(struct dquot *dquot); - static inline unsigned int hashfn(const struct super_block *sb, unsigned int id, int type) { @@ -568,7 +566,7 @@ static struct shrinker dqcache_shrinker = { * NOTE: If you change this function please check whether dqput_blocks() works right... * MUST be called with either dqptr_sem or dqonoff_mutex held */ -static void dqput(struct dquot *dquot) +void dqput(struct dquot *dquot) { int ret; @@ -662,10 +660,28 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type) } /* + * Check whether dquot is in memory. + * MUST be called with either dqptr_sem or dqonoff_mutex held + */ +int dquot_is_cached(struct super_block *sb, unsigned int id, int type) +{ + unsigned int hashent = hashfn(sb, id, type); + int ret = 0; + + if (!sb_has_quota_active(sb, type)) + return 0; + spin_lock(&dq_list_lock); + if (find_dquot(hashent, sb, id, type) != NODQUOT) + ret = 1; + spin_unlock(&dq_list_lock); + return ret; +} + +/* * Get reference to dquot * MUST be called with either dqptr_sem or dqonoff_mutex held */ -static struct dquot *dqget(struct super_block *sb, unsigned int id, int type) +struct dquot *dqget(struct super_block *sb, unsigned int id, int type) { unsigned int hashent = hashfn(sb, id, type); struct dquot *dquot, *empty = NODQUOT; @@ -1184,17 +1200,23 @@ out_err: * Release all quotas referenced by inode * Transaction must be started at an entry */ -int dquot_drop(struct inode *inode) +int dquot_drop_locked(struct inode *inode) { int cnt; - down_write(&sb_dqopt(inode->i_sb)->dqptr_sem); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (inode->i_dquot[cnt] != NODQUOT) { dqput(inode->i_dquot[cnt]); inode->i_dquot[cnt] = NODQUOT; } } + return 0; +} + +int dquot_drop(struct inode *inode) +{ + down_write(&sb_dqopt(inode->i_sb)->dqptr_sem); + dquot_drop_locked(inode); up_write(&sb_dqopt(inode->i_sb)->dqptr_sem); return 0; } @@ -2308,7 +2330,11 @@ EXPORT_SYMBOL(dquot_release); EXPORT_SYMBOL(dquot_mark_dquot_dirty); EXPORT_SYMBOL(dquot_initialize); EXPORT_SYMBOL(dquot_drop); +EXPORT_SYMBOL(dquot_drop_locked); EXPORT_SYMBOL(vfs_dq_drop); +EXPORT_SYMBOL(dqget); +EXPORT_SYMBOL(dqput); +EXPORT_SYMBOL(dquot_is_cached); EXPORT_SYMBOL(dquot_alloc_space); EXPORT_SYMBOL(dquot_alloc_inode); EXPORT_SYMBOL(dquot_free_space); diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index e840ca5..e3a1027 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -24,6 +24,10 @@ void sync_dquots(struct super_block *sb, int type); int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); +int dquot_drop_locked(struct inode *inode); +struct dquot *dqget(struct super_block *sb, unsigned int id, int type); +void dqput(struct dquot *dquot); +int dquot_is_cached(struct super_block *sb, unsigned int id, int type); int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); int dquot_alloc_inode(const struct inode *inode, qsize_t number); -- 1.5.2.4
Jan Kara
2008-Oct-31 06:39 UTC
[Ocfs2-devel] [PATCH 18/27] quota: Implement function for scanning active dquots
OCFS2 needs to scan all active dquots once in a while and sync quota information among cluster nodes. Provide a helper function for it so that it does not have to reimplement internally a list which VFS already has. Moreover this function is probably going to be useful for other clustered filesystems if they decide to use VFS quotas. Signed-off-by: Jan Kara <jack at suse.cz> --- fs/dquot.c | 36 ++++++++++++++++++++++++++++++++++++ include/linux/quotaops.h | 3 +++ 2 files changed, 39 insertions(+), 0 deletions(-) diff --git a/fs/dquot.c b/fs/dquot.c index 8774be6..6f7df91 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -476,6 +476,41 @@ restart: spin_unlock(&dq_list_lock); } +/* Call callback for every active dquot on given filesystem */ +int dquot_scan_active(struct super_block *sb, + int (*fn)(struct dquot *dquot, unsigned long priv), + unsigned long priv) +{ + struct dquot *dquot, *old_dquot = NULL; + int ret = 0; + + mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); + spin_lock(&dq_list_lock); + list_for_each_entry(dquot, &inuse_list, dq_inuse) { + if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) + continue; + if (dquot->dq_sb != sb) + continue; + /* Now we have active dquot so we can just increase use count */ + atomic_inc(&dquot->dq_count); + dqstats.lookups++; + spin_unlock(&dq_list_lock); + dqput(old_dquot); + old_dquot = dquot; + ret = fn(dquot, priv); + if (ret < 0) + goto out; + spin_lock(&dq_list_lock); + /* We are safe to continue now because our dquot could not + * be moved out of the inuse list while we hold the reference */ + } + spin_unlock(&dq_list_lock); +out: + dqput(old_dquot); + mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); + return ret; +} + int vfs_quota_sync(struct super_block *sb, int type) { struct list_head *dirty; @@ -2318,6 +2353,7 @@ EXPORT_SYMBOL(vfs_quota_on_path); EXPORT_SYMBOL(vfs_quota_on_mount); EXPORT_SYMBOL(vfs_quota_disable); EXPORT_SYMBOL(vfs_quota_off); +EXPORT_SYMBOL(dquot_scan_active); EXPORT_SYMBOL(vfs_quota_sync); EXPORT_SYMBOL(vfs_get_dqinfo); EXPORT_SYMBOL(vfs_set_dqinfo); diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index e3a1027..f491394 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -28,6 +28,9 @@ int dquot_drop_locked(struct inode *inode); struct dquot *dqget(struct super_block *sb, unsigned int id, int type); void dqput(struct dquot *dquot); int dquot_is_cached(struct super_block *sb, unsigned int id, int type); +int dquot_scan_active(struct super_block *sb, + int (*fn)(struct dquot *dquot, unsigned long priv), + unsigned long priv); int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); int dquot_alloc_inode(const struct inode *inode, qsize_t number); -- 1.5.2.4
On Fri, 2008-10-31 at 07:38 +0100, Jan Kara wrote:> Hi, > > in this thread comes the third version of the patches implementing quotas for > OCFS2 (sorry for spamming ocfs2-devel so often with such a big pile of patches > ;). I've now addressed all Joel's comments (feature bits added among supported > bits in the last patch, hopefully fixed an oops when enabling quota fails on > mount), also fixed a bug Tristan reported (message about exceeding softlimit > was not always properly printed), fixed a bug in node recovery leading to > possible oops when filesystem was without quotas (reported by Andrew Beekhof) > and also reported a few style things and compilation problems reported > by Andrew Morton and other people getting the patches via -mm kernel. As > always, comments are welcome (but I won't respond quickly because I'll be > mostly out of email for next two weeks). > HonzaThanks for the quick 3rd version of quota patches on ocfs2,is this patch set based on mark's kernel tree or mainline? i'll soon check my reported bug on your provided kernel:) BTW, I've worked out a first version of sanity checker for quota on ocfs2,now its only on a single-node mode,will try to work on mutiple-nodes tests later for node crash,heavy parallel quota usage and negative quota usage test etc.maybe by the help of openmpi tools:) I've attached the singe-node test script by this mail,you may have it for sanity check if you wish,it includes one user/one group/multiple groups test for all quota related generic functions(also soft/hard/grace time limitaion for inodes and space usage tested here),you also can tune the workload extremely high to expect it as a stress puncher. it can be used like, ./quota_tests.sh -o quotalog -d /dev/sdc5 -t ocfs2 -i small /quota (here -i specify the workload, small means for a generic test,medium and large increase the stress sharply:) I've tested you 2rd version of patch set by specifying a medium workload,the result was good to me:) The testing status and testing tools will be kept on http://oss.oracle.com/osswiki/QuotaTest Regards, Tristan.> > _______________________________________________ > Ocfs2-devel mailing list > Ocfs2-devel at oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-devel-------------- next part -------------- A non-text attachment was scrubbed... Name: quota_tests.sh Type: application/x-shellscript Size: 25591 bytes Desc: not available Url : http://oss.oracle.com/pipermail/ocfs2-devel/attachments/20081031/b74cde20/attachment-0001.bin
On Fri, 2008-10-31 at 17:06 +0800, tristan.ye wrote:> On Fri, 2008-10-31 at 07:38 +0100, Jan Kara wrote: > > Hi, > > > > in this thread comes the third version of the patches implementing quotas for > > OCFS2 (sorry for spamming ocfs2-devel so often with such a big pile of patches > > ;). I've now addressed all Joel's comments (feature bits added among supported > > bits in the last patch, hopefully fixed an oops when enabling quota fails on > > mount), also fixed a bug Tristan reported (message about exceeding softlimit > > was not always properly printed), fixed a bug in node recovery leading to > > possible oops when filesystem was without quotas (reported by Andrew Beekhof) > > and also reported a few style things and compilation problems reported > > by Andrew Morton and other people getting the patches via -mm kernel. As > > always, comments are welcome (but I won't respond quickly because I'll be > > mostly out of email for next two weeks). > > Honza > > Thanks for the quick 3rd version of quota patches on ocfs2,is this patch > set based on mark's kernel tree or mainline? i'll soon check my reported > bug on your provided kernel:) > > BTW, I've worked out a first version of sanity checker for quota on > ocfs2,now its only on a single-node mode,will try to work on > mutiple-nodes tests later for node crash,heavy parallel quota usage and > negative quota usage test etc.maybe by the help of openmpi tools:) > > I've attached the singe-node test script by this mail,you may have it > for sanity check if you wish,it includes one user/one group/multiple > groups test for all quota related generic functions(also soft/hard/grace > time limitaion for inodes and space usage tested here),you also can tune > the workload extremely high to expect it as a stress puncher. > > it can be used like, > > ./quota_tests.sh -o quotalog -d /dev/sdc5 -t ocfs2 -i small /quota (here > -i specify the workload, small means for a generic test,medium and large > increase the stress sharply:) > > I've tested you 2rd version of patch set by specifying a medium > workload,the result was good to me:) > > The testing status and testing tools will be kept on > http://oss.oracle.com/osswiki/QuotaTestSorry, The right URL of testing wiki shoud be http://oss.oracle.com/osswiki/OCFS2/QuotaTest , the former one is going to be obsolete since it was under a wrong namespace. Regards, Tristan.> > > Regards, > > Tristan. > > > > > _______________________________________________ > > Ocfs2-devel mailing list > > Ocfs2-devel at oss.oracle.com > > http://oss.oracle.com/mailman/listinfo/ocfs2-devel