From 74f783af95c982aef6d3a1415275650dcf511666 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 19 Aug 2008 14:51:22 +0200 Subject: 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 Signed-off-by: Mark Fasheh --- include/linux/quota.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/linux/quota.h b/include/linux/quota.h index 40401b554484..3ce708c2cb3c 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 */ -- cgit v1.2.3 From 12095460f7f315f8ef67a55b2194195d325d48d7 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 20 Aug 2008 14:45:12 +0200 Subject: 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 Signed-off-by: Mark Fasheh --- 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(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 1b5fc4b7fbeb..c02223b6aeb2 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 5ae15b13eeb0..3e078eee5644 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 b53827dc02d9..51c4717f7c6a 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 3ce708c2cb3c..9ea468363f9f 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 a558a4c1d35a..adcc7ba3accb 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); -- cgit v1.2.3 From e4bc7b4b7ff783779b6928d55a9308910bf180a3 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 20 Aug 2008 16:21:01 +0200 Subject: quota: Make _SUSPENDED just a flag Upto now, DQUOT_USR_SUSPENDED behaved like a state - i.e., either quota was enabled or suspended or none. Now allowed states are 0, ENABLED, ENABLED | SUSPENDED. This will be useful later when we implement separate enabling of quota usage tracking and limits enforcement because we need to keep track of a state which has been suspended. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 10 ++++++---- include/linux/quotaops.h | 6 ++++-- 2 files changed, 10 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index c88330602ddd..22340c610e1a 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -1570,18 +1570,20 @@ static inline void reset_enable_flags(struct quota_info *dqopt, int type, { switch (type) { case USRQUOTA: - dqopt->flags &= ~DQUOT_USR_ENABLED; if (remount) dqopt->flags |= DQUOT_USR_SUSPENDED; - else + else { + dqopt->flags &= ~DQUOT_USR_ENABLED; dqopt->flags &= ~DQUOT_USR_SUSPENDED; + } break; case GRPQUOTA: - dqopt->flags &= ~DQUOT_GRP_ENABLED; if (remount) dqopt->flags |= DQUOT_GRP_SUSPENDED; - else + else { + dqopt->flags &= ~DQUOT_GRP_ENABLED; dqopt->flags &= ~DQUOT_GRP_SUSPENDED; + } break; } } diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index adcc7ba3accb..ffd97071cd1e 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -67,8 +67,10 @@ static inline struct mem_dqinfo *sb_dqinfo(struct super_block *sb, int type) static inline int sb_has_quota_enabled(struct super_block *sb, int type) { if (type == USRQUOTA) - return sb_dqopt(sb)->flags & DQUOT_USR_ENABLED; - return sb_dqopt(sb)->flags & DQUOT_GRP_ENABLED; + return (sb_dqopt(sb)->flags & DQUOT_USR_ENABLED) + && !(sb_dqopt(sb)->flags & DQUOT_USR_SUSPENDED); + return (sb_dqopt(sb)->flags & DQUOT_GRP_ENABLED) + && !(sb_dqopt(sb)->flags & DQUOT_GROUP_SUSPENDED); } static inline int sb_any_quota_enabled(struct super_block *sb) -- cgit v1.2.3 From f55abc0fb9c3189de3da829adf3220322c0da43e Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 20 Aug 2008 17:50:32 +0200 Subject: quota: Allow to separately enable quota accounting and enforcing limits Split DQUOT_USR_ENABLED (and DQUOT_GRP_ENABLED) into DQUOT_USR_USAGE_ENABLED and DQUOT_USR_LIMITS_ENABLED. This way we are able to separately enable / disable whether we should: 1) ignore quotas completely 2) just keep uptodate information about usage 3) actually enforce quota limits This is going to be useful when quota is treated as filesystem metadata - we then want to keep quota information uptodate all the time and just enable / disable limits enforcement. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 222 +++++++++++++++++++++++++++++------------------ fs/quota.c | 8 +- include/linux/quota.h | 30 ++++++- include/linux/quotaops.h | 91 ++++++++++++++----- 4 files changed, 239 insertions(+), 112 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 22340c610e1a..7569633efe0e 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -489,7 +489,7 @@ int vfs_quota_sync(struct super_block *sb, int type) for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (type != -1 && cnt != type) continue; - if (!sb_has_quota_enabled(sb, cnt)) + if (!sb_has_quota_active(sb, cnt)) continue; spin_lock(&dq_list_lock); dirty = &dqopt->info[cnt].dqi_dirty_list; @@ -514,8 +514,8 @@ int vfs_quota_sync(struct super_block *sb, int type) } for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if ((cnt == type || type == -1) && sb_has_quota_enabled(sb, cnt) - && info_dirty(&dqopt->info[cnt])) + if ((cnt == type || type == -1) && sb_has_quota_active(sb, cnt) + && info_dirty(&dqopt->info[cnt])) sb->dq_op->write_info(sb, cnt); spin_lock(&dq_list_lock); dqstats.syncs++; @@ -594,7 +594,7 @@ we_slept: /* We have more than one user... nothing to do */ atomic_dec(&dquot->dq_count); /* Releasing dquot during quotaoff phase? */ - if (!sb_has_quota_enabled(dquot->dq_sb, dquot->dq_type) && + if (!sb_has_quota_active(dquot->dq_sb, dquot->dq_type) && atomic_read(&dquot->dq_count) == 1) wake_up(&dquot->dq_wait_unused); spin_unlock(&dq_list_lock); @@ -670,7 +670,7 @@ static struct dquot *dqget(struct super_block *sb, unsigned int id, int type) unsigned int hashent = hashfn(sb, id, type); struct dquot *dquot, *empty = NODQUOT; - if (!sb_has_quota_enabled(sb, type)) + if (!sb_has_quota_active(sb, type)) return NODQUOT; we_slept: spin_lock(&dq_list_lock); @@ -1041,7 +1041,8 @@ static inline char ignore_hardlimit(struct dquot *dquot) static int check_idq(struct dquot *dquot, qsize_t inodes, char *warntype) { *warntype = QUOTA_NL_NOWARN; - if (test_bit(DQ_FAKE_B, &dquot->dq_flags)) + if (!sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type) || + test_bit(DQ_FAKE_B, &dquot->dq_flags)) return QUOTA_OK; if (dquot->dq_dqb.dqb_ihardlimit && @@ -1073,7 +1074,8 @@ static int check_idq(struct dquot *dquot, qsize_t inodes, char *warntype) static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *warntype) { *warntype = QUOTA_NL_NOWARN; - if (test_bit(DQ_FAKE_B, &dquot->dq_flags)) + if (!sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type) || + test_bit(DQ_FAKE_B, &dquot->dq_flags)) return QUOTA_OK; if (dquot->dq_dqb.dqb_bhardlimit && @@ -1114,7 +1116,8 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war 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) + dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit || + !sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type)) return QUOTA_NL_NOWARN; if (dquot->dq_dqb.dqb_curinodes - inodes <= dquot->dq_dqb.dqb_isoftlimit) @@ -1508,7 +1511,7 @@ warn_put_all: /* Wrapper for transferring ownership of an inode */ int vfs_dq_transfer(struct inode *inode, struct iattr *iattr) { - if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) { + if (sb_any_quota_active(inode->i_sb) && !IS_NOQUOTA(inode)) { vfs_dq_init(inode); if (inode->i_sb->dq_op->transfer(inode, iattr) == NO_QUOTA) return 1; @@ -1551,53 +1554,22 @@ struct dquot_operations dquot_operations = { .destroy_dquot = dquot_destroy, }; -static inline void set_enable_flags(struct quota_info *dqopt, int type) -{ - switch (type) { - case USRQUOTA: - dqopt->flags |= DQUOT_USR_ENABLED; - dqopt->flags &= ~DQUOT_USR_SUSPENDED; - break; - case GRPQUOTA: - dqopt->flags |= DQUOT_GRP_ENABLED; - dqopt->flags &= ~DQUOT_GRP_SUSPENDED; - break; - } -} - -static inline void reset_enable_flags(struct quota_info *dqopt, int type, - int remount) -{ - switch (type) { - case USRQUOTA: - if (remount) - dqopt->flags |= DQUOT_USR_SUSPENDED; - else { - dqopt->flags &= ~DQUOT_USR_ENABLED; - dqopt->flags &= ~DQUOT_USR_SUSPENDED; - } - break; - case GRPQUOTA: - if (remount) - dqopt->flags |= DQUOT_GRP_SUSPENDED; - else { - dqopt->flags &= ~DQUOT_GRP_ENABLED; - dqopt->flags &= ~DQUOT_GRP_SUSPENDED; - } - break; - } -} - - /* * Turn quota off on a device. type == -1 ==> quotaoff for all types (umount) */ -int vfs_quota_off(struct super_block *sb, int type, int remount) +int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags) { int cnt, ret = 0; struct quota_info *dqopt = sb_dqopt(sb); struct inode *toputinode[MAXQUOTAS]; + /* Cannot turn off usage accounting without turning off limits, or + * suspend quotas and simultaneously turn quotas off. */ + if ((flags & DQUOT_USAGE_ENABLED && !(flags & DQUOT_LIMITS_ENABLED)) + || (flags & DQUOT_SUSPENDED && flags & (DQUOT_LIMITS_ENABLED | + DQUOT_USAGE_ENABLED))) + return -EINVAL; + /* We need to serialize quota_off() for device */ mutex_lock(&dqopt->dqonoff_mutex); @@ -1606,7 +1578,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) * sometimes we are called when fill_super() failed and calling * sync_fs() in such cases does no good. */ - if (!sb_any_quota_enabled(sb) && !sb_any_quota_suspended(sb)) { + if (!sb_any_quota_loaded(sb)) { mutex_unlock(&dqopt->dqonoff_mutex); return 0; } @@ -1614,17 +1586,28 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) toputinode[cnt] = NULL; if (type != -1 && cnt != type) continue; - /* If we keep inodes of quota files after remount and quotaoff - * is called, drop kept inodes. */ - if (!remount && sb_has_quota_suspended(sb, cnt)) { - iput(dqopt->files[cnt]); - dqopt->files[cnt] = NULL; - reset_enable_flags(dqopt, cnt, 0); + if (!sb_has_quota_loaded(sb, cnt)) continue; + + if (flags & DQUOT_SUSPENDED) { + dqopt->flags |= + dquot_state_flag(DQUOT_SUSPENDED, cnt); + } else { + dqopt->flags &= ~dquot_state_flag(flags, cnt); + /* Turning off suspended quotas? */ + if (!sb_has_quota_loaded(sb, cnt) && + sb_has_quota_suspended(sb, cnt)) { + dqopt->flags &= ~dquot_state_flag( + DQUOT_SUSPENDED, cnt); + iput(dqopt->files[cnt]); + dqopt->files[cnt] = NULL; + continue; + } } - if (!sb_has_quota_enabled(sb, cnt)) + + /* We still have to keep quota loaded? */ + if (sb_has_quota_loaded(sb, cnt) && !(flags & DQUOT_SUSPENDED)) continue; - reset_enable_flags(dqopt, cnt, remount); /* Note: these are blocking operations */ drop_dquot_ref(sb, cnt); @@ -1640,7 +1623,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) put_quota_format(dqopt->info[cnt].dqi_format); toputinode[cnt] = dqopt->files[cnt]; - if (!remount) + if (!sb_has_quota_loaded(sb, cnt)) dqopt->files[cnt] = NULL; dqopt->info[cnt].dqi_flags = 0; dqopt->info[cnt].dqi_igrace = 0; @@ -1663,7 +1646,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) mutex_lock(&dqopt->dqonoff_mutex); /* If quota was reenabled in the meantime, we have * nothing to do */ - if (!sb_has_quota_enabled(sb, cnt)) { + if (!sb_has_quota_loaded(sb, cnt)) { mutex_lock_nested(&toputinode[cnt]->i_mutex, I_MUTEX_QUOTA); toputinode[cnt]->i_flags &= ~(S_IMMUTABLE | S_NOATIME | S_NOQUOTA); @@ -1673,10 +1656,13 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) } mutex_unlock(&dqopt->dqonoff_mutex); /* On remount RO, we keep the inode pointer so that we - * can reenable quota on the subsequent remount RW. - * But we have better not keep inode pointer when there - * is pending delete on the quota file... */ - if (!remount) + * can reenable quota on the subsequent remount RW. We + * have to check 'flags' variable and not use sb_has_ + * function because another quotaon / quotaoff could + * change global state before we got here. We refuse + * to suspend quotas when there is pending delete on + * the quota file... */ + if (!(flags & DQUOT_SUSPENDED)) iput(toputinode[cnt]); else if (!toputinode[cnt]->i_nlink) ret = -EBUSY; @@ -1686,12 +1672,22 @@ int vfs_quota_off(struct super_block *sb, int type, int remount) return ret; } +int vfs_quota_off(struct super_block *sb, int type, int remount) +{ + return vfs_quota_disable(sb, type, remount ? DQUOT_SUSPENDED : + (DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED)); +} + /* * Turn quotas on on a device */ -/* Helper function when we already have the inode */ -static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) +/* + * Helper function to turn quotas on when we already have the inode of + * quota file and no quota information is loaded. + */ +static int vfs_load_quota_inode(struct inode *inode, int type, int format_id, + unsigned int flags) { struct quota_format_type *fmt = find_quota_format(format_id); struct super_block *sb = inode->i_sb; @@ -1713,6 +1709,11 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) error = -EINVAL; goto out_fmt; } + /* Usage always has to be set... */ + if (!(flags & DQUOT_USAGE_ENABLED)) { + error = -EINVAL; + goto out_fmt; + } /* As we bypass the pagecache we must now flush the inode so that * we see all the changes from userspace... */ @@ -1721,8 +1722,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) invalidate_bdev(sb->s_bdev); mutex_lock(&inode->i_mutex); mutex_lock(&dqopt->dqonoff_mutex); - if (sb_has_quota_enabled(sb, type) || - sb_has_quota_suspended(sb, type)) { + if (sb_has_quota_loaded(sb, type)) { error = -EBUSY; goto out_lock; } @@ -1754,7 +1754,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) } mutex_unlock(&dqopt->dqio_mutex); mutex_unlock(&inode->i_mutex); - set_enable_flags(dqopt, type); + dqopt->flags |= dquot_state_flag(flags, type); add_dquot_ref(sb, type); mutex_unlock(&dqopt->dqonoff_mutex); @@ -1787,20 +1787,23 @@ static int vfs_quota_on_remount(struct super_block *sb, int type) struct quota_info *dqopt = sb_dqopt(sb); struct inode *inode; int ret; + unsigned int flags; mutex_lock(&dqopt->dqonoff_mutex); if (!sb_has_quota_suspended(sb, type)) { mutex_unlock(&dqopt->dqonoff_mutex); return 0; } - BUG_ON(sb_has_quota_enabled(sb, type)); - inode = dqopt->files[type]; dqopt->files[type] = NULL; - reset_enable_flags(dqopt, type, 0); + flags = dqopt->flags & dquot_state_flag(DQUOT_USAGE_ENABLED | + DQUOT_LIMITS_ENABLED, type); + dqopt->flags &= ~dquot_state_flag(DQUOT_STATE_FLAGS, type); mutex_unlock(&dqopt->dqonoff_mutex); - ret = vfs_quota_on_inode(inode, type, dqopt->info[type].dqi_fmt_id); + flags = dquot_generic_flag(flags, type); + ret = vfs_load_quota_inode(inode, type, dqopt->info[type].dqi_fmt_id, + flags); iput(inode); return ret; @@ -1816,12 +1819,12 @@ int vfs_quota_on_path(struct super_block *sb, int type, int format_id, if (path->mnt->mnt_sb != sb) error = -EXDEV; else - error = vfs_quota_on_inode(path->dentry->d_inode, type, - format_id); + error = vfs_load_quota_inode(path->dentry->d_inode, type, + format_id, DQUOT_USAGE_ENABLED | + DQUOT_LIMITS_ENABLED); return error; } -/* Actual function called from quotactl() */ int vfs_quota_on(struct super_block *sb, int type, int format_id, char *name, int remount) { @@ -1839,6 +1842,50 @@ int vfs_quota_on(struct super_block *sb, int type, int format_id, char *name, return error; } +/* + * More powerful function for turning on quotas allowing setting + * of individual quota flags + */ +int vfs_quota_enable(struct inode *inode, int type, int format_id, + unsigned int flags) +{ + int ret = 0; + struct super_block *sb = inode->i_sb; + struct quota_info *dqopt = sb_dqopt(sb); + + /* Just unsuspend quotas? */ + if (flags & DQUOT_SUSPENDED) + return vfs_quota_on_remount(sb, type); + if (!flags) + return 0; + /* Just updating flags needed? */ + if (sb_has_quota_loaded(sb, type)) { + mutex_lock(&dqopt->dqonoff_mutex); + /* Now do a reliable test... */ + if (!sb_has_quota_loaded(sb, type)) { + mutex_unlock(&dqopt->dqonoff_mutex); + goto load_quota; + } + if (flags & DQUOT_USAGE_ENABLED && + sb_has_quota_usage_enabled(sb, type)) { + ret = -EBUSY; + goto out_lock; + } + if (flags & DQUOT_LIMITS_ENABLED && + sb_has_quota_limits_enabled(sb, type)) { + ret = -EBUSY; + goto out_lock; + } + sb_dqopt(sb)->flags |= dquot_state_flag(flags, type); +out_lock: + mutex_unlock(&dqopt->dqonoff_mutex); + return ret; + } + +load_quota: + return vfs_load_quota_inode(inode, type, format_id, flags); +} + /* * This function is used when filesystem needs to initialize quotas * during mount time. @@ -1860,7 +1907,8 @@ int vfs_quota_on_mount(struct super_block *sb, char *qf_name, error = security_quota_on(dentry); if (!error) - error = vfs_quota_on_inode(dentry->d_inode, type, format_id); + error = vfs_load_quota_inode(dentry->d_inode, type, format_id, + DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED); out: dput(dentry); @@ -1997,12 +2045,14 @@ int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d int rc; mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); - if (!(dquot = dqget(sb, id, type))) { - mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); - return -ESRCH; + dquot = dqget(sb, id, type); + if (!dquot) { + rc = -ESRCH; + goto out; } rc = do_set_dqblk(dquot, di); dqput(dquot); +out: mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); return rc; } @@ -2013,7 +2063,7 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii) struct mem_dqinfo *mi; mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); - if (!sb_has_quota_enabled(sb, type)) { + if (!sb_has_quota_active(sb, type)) { mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); return -ESRCH; } @@ -2032,11 +2082,12 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii) int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii) { struct mem_dqinfo *mi; + int err = 0; mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); - if (!sb_has_quota_enabled(sb, type)) { - mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); - return -ESRCH; + if (!sb_has_quota_active(sb, type)) { + err = -ESRCH; + goto out; } mi = sb_dqopt(sb)->info + type; spin_lock(&dq_data_lock); @@ -2050,8 +2101,9 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii) mark_info_dirty(sb, type); /* Force write to disk */ sb->dq_op->write_info(sb, type); +out: mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); - return 0; + return err; } struct quotactl_ops vfs_quotactl_ops = { @@ -2213,9 +2265,11 @@ EXPORT_SYMBOL(register_quota_format); EXPORT_SYMBOL(unregister_quota_format); EXPORT_SYMBOL(dqstats); EXPORT_SYMBOL(dq_data_lock); +EXPORT_SYMBOL(vfs_quota_enable); EXPORT_SYMBOL(vfs_quota_on); EXPORT_SYMBOL(vfs_quota_on_path); EXPORT_SYMBOL(vfs_quota_on_mount); +EXPORT_SYMBOL(vfs_quota_disable); EXPORT_SYMBOL(vfs_quota_off); EXPORT_SYMBOL(vfs_quota_sync); EXPORT_SYMBOL(vfs_get_dqinfo); diff --git a/fs/quota.c b/fs/quota.c index b7fe44e01618..8678d9f35ee9 100644 --- a/fs/quota.c +++ b/fs/quota.c @@ -73,7 +73,7 @@ static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid case Q_SETQUOTA: case Q_GETQUOTA: /* This is just informative test so we are satisfied without a lock */ - if (!sb_has_quota_enabled(sb, type)) + if (!sb_has_quota_active(sb, type)) return -ESRCH; } @@ -175,7 +175,7 @@ static void quota_sync_sb(struct super_block *sb, int type) for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (type != -1 && cnt != type) continue; - if (!sb_has_quota_enabled(sb, cnt)) + if (!sb_has_quota_active(sb, cnt)) continue; mutex_lock_nested(&sb_dqopt(sb)->files[cnt]->i_mutex, I_MUTEX_QUOTA); truncate_inode_pages(&sb_dqopt(sb)->files[cnt]->i_data, 0); @@ -201,7 +201,7 @@ restart: for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (type != -1 && type != cnt) continue; - if (!sb_has_quota_enabled(sb, cnt)) + if (!sb_has_quota_active(sb, cnt)) continue; if (!info_dirty(&sb_dqopt(sb)->info[cnt]) && list_empty(&sb_dqopt(sb)->info[cnt].dqi_dirty_list)) @@ -245,7 +245,7 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id, void __u32 fmt; down_read(&sb_dqopt(sb)->dqptr_sem); - if (!sb_has_quota_enabled(sb, type)) { + if (!sb_has_quota_active(sb, type)) { up_read(&sb_dqopt(sb)->dqptr_sem); return -ESRCH; } diff --git a/include/linux/quota.h b/include/linux/quota.h index 9ea468363f9f..93717abcd35b 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -318,12 +318,34 @@ struct quota_format_type { struct quota_format_type *qf_next; }; -#define DQUOT_USR_ENABLED 0x01 /* User diskquotas enabled */ -#define DQUOT_GRP_ENABLED 0x02 /* Group diskquotas enabled */ -#define DQUOT_USR_SUSPENDED 0x04 /* User diskquotas are off, but +/* Quota state flags - they actually come in two flavors - for users and groups */ +enum { + _DQUOT_USAGE_ENABLED = 0, /* Track disk usage for users */ + _DQUOT_LIMITS_ENABLED, /* Enforce quota limits for users */ + _DQUOT_SUSPENDED, /* User diskquotas are off, but * we have necessary info in * memory to turn them on */ -#define DQUOT_GRP_SUSPENDED 0x08 /* The same for group quotas */ + _DQUOT_STATE_FLAGS +}; +#define DQUOT_USAGE_ENABLED (1 << _DQUOT_USAGE_ENABLED) +#define DQUOT_LIMITS_ENABLED (1 << _DQUOT_LIMITS_ENABLED) +#define DQUOT_SUSPENDED (1 << _DQUOT_SUSPENDED) +#define DQUOT_STATE_FLAGS (DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED | \ + DQUOT_SUSPENDED) + +static inline unsigned int dquot_state_flag(unsigned int flags, int type) +{ + if (type == USRQUOTA) + return flags; + return flags << _DQUOT_STATE_FLAGS; +} + +static inline unsigned int dquot_generic_flag(unsigned int flags, int type) +{ + if (type == USRQUOTA) + return flags; + return flags >> _DQUOT_STATE_FLAGS; +} struct quota_info { unsigned int flags; /* Flags for diskquotas on this device */ diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index ffd97071cd1e..3b3346fa657c 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -40,11 +40,14 @@ int dquot_mark_dquot_dirty(struct dquot *dquot); int vfs_quota_on(struct super_block *sb, int type, int format_id, char *path, int remount); +int vfs_quota_enable(struct inode *inode, int type, int format_id, + unsigned int flags); int vfs_quota_on_path(struct super_block *sb, int type, int format_id, struct path *path); int vfs_quota_on_mount(struct super_block *sb, char *qf_name, int format_id, int type); int vfs_quota_off(struct super_block *sb, int type, int remount); +int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags); int vfs_quota_sync(struct super_block *sb, int type); int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); @@ -64,26 +67,22 @@ static inline struct mem_dqinfo *sb_dqinfo(struct super_block *sb, int type) * Functions for checking status of quota */ -static inline int sb_has_quota_enabled(struct super_block *sb, int type) +static inline int sb_has_quota_usage_enabled(struct super_block *sb, int type) { - if (type == USRQUOTA) - return (sb_dqopt(sb)->flags & DQUOT_USR_ENABLED) - && !(sb_dqopt(sb)->flags & DQUOT_USR_SUSPENDED); - return (sb_dqopt(sb)->flags & DQUOT_GRP_ENABLED) - && !(sb_dqopt(sb)->flags & DQUOT_GROUP_SUSPENDED); + return sb_dqopt(sb)->flags & + dquot_state_flag(DQUOT_USAGE_ENABLED, type); } -static inline int sb_any_quota_enabled(struct super_block *sb) +static inline int sb_has_quota_limits_enabled(struct super_block *sb, int type) { - return sb_has_quota_enabled(sb, USRQUOTA) || - sb_has_quota_enabled(sb, GRPQUOTA); + return sb_dqopt(sb)->flags & + dquot_state_flag(DQUOT_LIMITS_ENABLED, type); } static inline int sb_has_quota_suspended(struct super_block *sb, int type) { - if (type == USRQUOTA) - return sb_dqopt(sb)->flags & DQUOT_USR_SUSPENDED; - return sb_dqopt(sb)->flags & DQUOT_GRP_SUSPENDED; + return sb_dqopt(sb)->flags & + dquot_state_flag(DQUOT_SUSPENDED, type); } static inline int sb_any_quota_suspended(struct super_block *sb) @@ -92,6 +91,34 @@ static inline int sb_any_quota_suspended(struct super_block *sb) sb_has_quota_suspended(sb, GRPQUOTA); } +/* Does kernel know about any quota information for given sb + type? */ +static inline int sb_has_quota_loaded(struct super_block *sb, int type) +{ + /* Currently if anything is on, then quota usage is on as well */ + return sb_has_quota_usage_enabled(sb, type); +} + +static inline int sb_any_quota_loaded(struct super_block *sb) +{ + return sb_has_quota_loaded(sb, USRQUOTA) || + sb_has_quota_loaded(sb, GRPQUOTA); +} + +static inline int sb_has_quota_active(struct super_block *sb, int type) +{ + return sb_has_quota_loaded(sb, type) && + !sb_has_quota_suspended(sb, type); +} + +static inline int sb_any_quota_active(struct super_block *sb) +{ + return sb_has_quota_active(sb, USRQUOTA) || + sb_has_quota_active(sb, GRPQUOTA); +} + +/* For backward compatibility until we remove all users */ +#define sb_any_quota_enabled(sb) sb_any_quota_active(sb) + /* * Operations supported for diskquotas. */ @@ -106,7 +133,7 @@ extern struct quotactl_ops vfs_quotactl_ops; static inline void vfs_dq_init(struct inode *inode) { BUG_ON(!inode->i_sb); - if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) + if (sb_any_quota_active(inode->i_sb) && !IS_NOQUOTA(inode)) inode->i_sb->dq_op->initialize(inode, -1); } @@ -114,7 +141,7 @@ static inline void vfs_dq_init(struct inode *inode) * a transaction (deadlocks possible otherwise) */ static inline int vfs_dq_prealloc_space_nodirty(struct inode *inode, qsize_t nr) { - if (sb_any_quota_enabled(inode->i_sb)) { + if (sb_any_quota_active(inode->i_sb)) { /* Used space is updated in alloc_space() */ if (inode->i_sb->dq_op->alloc_space(inode, nr, 1) == NO_QUOTA) return 1; @@ -134,7 +161,7 @@ static inline int vfs_dq_prealloc_space(struct inode *inode, qsize_t nr) static inline int vfs_dq_alloc_space_nodirty(struct inode *inode, qsize_t nr) { - if (sb_any_quota_enabled(inode->i_sb)) { + if (sb_any_quota_active(inode->i_sb)) { /* Used space is updated in alloc_space() */ if (inode->i_sb->dq_op->alloc_space(inode, nr, 0) == NO_QUOTA) return 1; @@ -154,7 +181,7 @@ static inline int vfs_dq_alloc_space(struct inode *inode, qsize_t nr) static inline int vfs_dq_alloc_inode(struct inode *inode) { - if (sb_any_quota_enabled(inode->i_sb)) { + if (sb_any_quota_active(inode->i_sb)) { vfs_dq_init(inode); if (inode->i_sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) return 1; @@ -164,7 +191,7 @@ static inline int vfs_dq_alloc_inode(struct inode *inode) static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { - if (sb_any_quota_enabled(inode->i_sb)) + if (sb_any_quota_active(inode->i_sb)) inode->i_sb->dq_op->free_space(inode, nr); else inode_sub_bytes(inode, nr); @@ -178,7 +205,7 @@ static inline void vfs_dq_free_space(struct inode *inode, qsize_t nr) static inline void vfs_dq_free_inode(struct inode *inode) { - if (sb_any_quota_enabled(inode->i_sb)) + if (sb_any_quota_active(inode->i_sb)) inode->i_sb->dq_op->free_inode(inode, 1); } @@ -199,12 +226,12 @@ static inline int vfs_dq_off(struct super_block *sb, int remount) #else -static inline int sb_has_quota_enabled(struct super_block *sb, int type) +static inline int sb_has_quota_usage_enabled(struct super_block *sb, int type) { return 0; } -static inline int sb_any_quota_enabled(struct super_block *sb) +static inline int sb_has_quota_limits_enabled(struct super_block *sb, int type) { return 0; } @@ -219,6 +246,30 @@ static inline int sb_any_quota_suspended(struct super_block *sb) return 0; } +/* Does kernel know about any quota information for given sb + type? */ +static inline int sb_has_quota_loaded(struct super_block *sb, int type) +{ + return 0; +} + +static inline int sb_any_quota_loaded(struct super_block *sb) +{ + return 0; +} + +static inline int sb_has_quota_active(struct super_block *sb, int type) +{ + return 0; +} + +static inline int sb_any_quota_active(struct super_block *sb) +{ + return 0; +} + +/* For backward compatibility until we remove all users */ +#define sb_any_quota_enabled(sb) sb_any_quota_active(sb) + /* * NO-OP when quota not configured. */ -- cgit v1.2.3 From dcb30695f2cac86b71417629a6fe8042b4fe2ab2 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 20 Aug 2008 18:30:40 +0200 Subject: quota: Remove compatibility function sb_any_quota_enabled() Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- include/linux/quotaops.h | 6 ------ 1 file changed, 6 deletions(-) (limited to 'include') diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 3b3346fa657c..e840ca523175 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -116,9 +116,6 @@ static inline int sb_any_quota_active(struct super_block *sb) sb_has_quota_active(sb, GRPQUOTA); } -/* For backward compatibility until we remove all users */ -#define sb_any_quota_enabled(sb) sb_any_quota_active(sb) - /* * Operations supported for diskquotas. */ @@ -267,9 +264,6 @@ static inline int sb_any_quota_active(struct super_block *sb) return 0; } -/* For backward compatibility until we remove all users */ -#define sb_any_quota_enabled(sb) sb_any_quota_active(sb) - /* * NO-OP when quota not configured. */ -- cgit v1.2.3 From ca785ec66b991e9ca74dd9840fc014487ad095e1 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 30 Sep 2008 17:53:37 +0200 Subject: quota: Introduce DQUOT_QUOTA_SYS_FILE flag If filesystem can handle quota files as system files hidden from users, we can skip a lot of cache invalidation, syncing, inode flags setting etc. when turning quotas on, off and quota_sync. Allow filesystem to indicate that it is hiding quota files from users by DQUOT_QUOTA_SYS_FILE flag. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 45 ++++++++++++++++++++++++++++++--------------- fs/quota.c | 3 +++ include/linux/quota.h | 7 +++++++ 3 files changed, 40 insertions(+), 15 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 7569633efe0e..74185c34a4f0 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -1631,6 +1631,11 @@ int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags) dqopt->ops[cnt] = NULL; } mutex_unlock(&dqopt->dqonoff_mutex); + + /* Skip syncing and setting flags if quota files are hidden */ + if (dqopt->flags & DQUOT_QUOTA_SYS_FILE) + goto put_inodes; + /* Sync the superblock so that buffers with quota data are written to * disk (and so userspace sees correct data afterwards). */ if (sb->s_op->sync_fs) @@ -1655,6 +1660,12 @@ int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags) mark_inode_dirty(toputinode[cnt]); } mutex_unlock(&dqopt->dqonoff_mutex); + } + if (sb->s_bdev) + invalidate_bdev(sb->s_bdev); +put_inodes: + for (cnt = 0; cnt < MAXQUOTAS; cnt++) + if (toputinode[cnt]) { /* On remount RO, we keep the inode pointer so that we * can reenable quota on the subsequent remount RW. We * have to check 'flags' variable and not use sb_has_ @@ -1667,8 +1678,6 @@ int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags) else if (!toputinode[cnt]->i_nlink) ret = -EBUSY; } - if (sb->s_bdev) - invalidate_bdev(sb->s_bdev); return ret; } @@ -1715,25 +1724,31 @@ static int vfs_load_quota_inode(struct inode *inode, int type, int format_id, goto out_fmt; } - /* As we bypass the pagecache we must now flush the inode so that - * we see all the changes from userspace... */ - write_inode_now(inode, 1); - /* And now flush the block cache so that kernel sees the changes */ - invalidate_bdev(sb->s_bdev); + if (!(dqopt->flags & DQUOT_QUOTA_SYS_FILE)) { + /* As we bypass the pagecache we must now flush the inode so + * that we see all the changes from userspace... */ + write_inode_now(inode, 1); + /* And now flush the block cache so that kernel sees the + * changes */ + invalidate_bdev(sb->s_bdev); + } mutex_lock(&inode->i_mutex); mutex_lock(&dqopt->dqonoff_mutex); if (sb_has_quota_loaded(sb, type)) { error = -EBUSY; goto out_lock; } - /* We don't want quota and atime on quota files (deadlocks possible) - * Also nobody should write to the file - we use special IO operations - * which ignore the immutable bit. */ - down_write(&dqopt->dqptr_sem); - oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE | S_NOQUOTA); - inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE; - up_write(&dqopt->dqptr_sem); - sb->dq_op->drop(inode); + + if (!(dqopt->flags & DQUOT_QUOTA_SYS_FILE)) { + /* We don't want quota and atime on quota files (deadlocks + * possible) Also nobody should write to the file - we use + * special IO operations which ignore the immutable bit. */ + down_write(&dqopt->dqptr_sem); + oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE | S_NOQUOTA); + inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE; + up_write(&dqopt->dqptr_sem); + sb->dq_op->drop(inode); + } error = -EIO; dqopt->files[type] = igrab(inode); diff --git a/fs/quota.c b/fs/quota.c index 8678d9f35ee9..4a8c94f05f76 100644 --- a/fs/quota.c +++ b/fs/quota.c @@ -160,6 +160,9 @@ static void quota_sync_sb(struct super_block *sb, int type) int cnt; sb->s_qcop->quota_sync(sb, type); + + if (sb_dqopt(sb)->flags & DQUOT_QUOTA_SYS_FILE) + return; /* This is not very clever (and fast) but currently I don't know about * any other simple way of getting quota data to disk and we must get * them there for userspace to be visible... */ diff --git a/include/linux/quota.h b/include/linux/quota.h index 93717abcd35b..80b8807b4988 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -332,6 +332,13 @@ enum { #define DQUOT_SUSPENDED (1 << _DQUOT_SUSPENDED) #define DQUOT_STATE_FLAGS (DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED | \ DQUOT_SUSPENDED) +/* Other quota flags */ +#define DQUOT_QUOTA_SYS_FILE (1 << 6) /* Quota file is a special + * system file and user cannot + * touch it. Filesystem is + * responsible for setting + * S_NOQUOTA, S_NOATIME flags + */ static inline unsigned int dquot_state_flag(unsigned int flags, int type) { -- cgit v1.2.3 From cf770c137122b78470a67ebd5498947869a09197 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Sun, 21 Sep 2008 23:17:53 +0200 Subject: quota: Move quotaio_v[12].h from include/linux/ to fs/ Since these include files are used only by implementation of quota formats, there's no need to have them in include/linux/. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/quota_v1.c | 3 +- fs/quota_v2.c | 7 ++-- fs/quotaio_v1.h | 33 +++++++++++++++++++ fs/quotaio_v2.h | 79 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/Kbuild | 2 -- include/linux/quotaio_v1.h | 33 ------------------- include/linux/quotaio_v2.h | 79 ---------------------------------------------- 7 files changed, 118 insertions(+), 118 deletions(-) create mode 100644 fs/quotaio_v1.h create mode 100644 fs/quotaio_v2.h delete mode 100644 include/linux/quotaio_v1.h delete mode 100644 include/linux/quotaio_v2.h (limited to 'include') diff --git a/fs/quota_v1.c b/fs/quota_v1.c index 3e078eee5644..b4af1c69ad16 100644 --- a/fs/quota_v1.c +++ b/fs/quota_v1.c @@ -3,13 +3,14 @@ #include #include #include -#include #include #include #include #include +#include "quotaio_v1.h" + MODULE_AUTHOR("Jan Kara"); MODULE_DESCRIPTION("Old quota format support"); MODULE_LICENSE("GPL"); diff --git a/fs/quota_v2.c b/fs/quota_v2.c index 51c4717f7c6a..a21d1a7c356a 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -6,7 +6,6 @@ #include #include #include -#include #include #include #include @@ -15,6 +14,8 @@ #include +#include "quotaio_v2.h" + MODULE_AUTHOR("Jan Kara"); MODULE_DESCRIPTION("Quota format v2 support"); MODULE_LICENSE("GPL"); @@ -129,8 +130,8 @@ 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(v2_qbtos(m->dqb_bhardlimit)); - d->dqb_bsoftlimit = cpu_to_le32(v2_qbtos(m->dqb_bsoftlimit)); + d->dqb_bhardlimit = cpu_to_le32(v2_stoqb(m->dqb_bhardlimit)); + d->dqb_bsoftlimit = cpu_to_le32(v2_stoqb(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/fs/quotaio_v1.h b/fs/quotaio_v1.h new file mode 100644 index 000000000000..746654b5de70 --- /dev/null +++ b/fs/quotaio_v1.h @@ -0,0 +1,33 @@ +#ifndef _LINUX_QUOTAIO_V1_H +#define _LINUX_QUOTAIO_V1_H + +#include + +/* + * The following constants define the amount of time given a user + * before the soft limits are treated as hard limits (usually resulting + * in an allocation failure). The timer is started when the user crosses + * their soft limit, it is reset when they go below their soft limit. + */ +#define MAX_IQ_TIME 604800 /* (7*24*60*60) 1 week */ +#define MAX_DQ_TIME 604800 /* (7*24*60*60) 1 week */ + +/* + * The following structure defines the format of the disk quota file + * (as it appears on disk) - the file is an array of these structures + * indexed by user or group number. + */ +struct v1_disk_dqblk { + __u32 dqb_bhardlimit; /* absolute limit on disk blks alloc */ + __u32 dqb_bsoftlimit; /* preferred limit on disk blks */ + __u32 dqb_curblocks; /* current block count */ + __u32 dqb_ihardlimit; /* absolute limit on allocated inodes */ + __u32 dqb_isoftlimit; /* preferred inode limit */ + __u32 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 */ +}; + +#define v1_dqoff(UID) ((loff_t)((UID) * sizeof (struct v1_disk_dqblk))) + +#endif /* _LINUX_QUOTAIO_V1_H */ diff --git a/fs/quotaio_v2.h b/fs/quotaio_v2.h new file mode 100644 index 000000000000..303d7cbe30d4 --- /dev/null +++ b/fs/quotaio_v2.h @@ -0,0 +1,79 @@ +/* + * Definitions of structures for vfsv0 quota format + */ + +#ifndef _LINUX_QUOTAIO_V2_H +#define _LINUX_QUOTAIO_V2_H + +#include +#include + +/* + * Definitions of magics and versions of current quota files + */ +#define V2_INITQMAGICS {\ + 0xd9c01f11, /* USRQUOTA */\ + 0xd9c01927 /* GRPQUOTA */\ +} + +#define V2_INITQVERSIONS {\ + 0, /* USRQUOTA */\ + 0 /* GRPQUOTA */\ +} + +/* + * The following structure defines the format of the disk quota file + * (as it appears on disk) - the file is a radix tree whose leaves point + * to blocks of these structures. + */ +struct v2_disk_dqblk { + __le32 dqb_id; /* id this quota applies to */ + __le32 dqb_ihardlimit; /* absolute limit on allocated inodes */ + __le32 dqb_isoftlimit; /* preferred inode limit */ + __le32 dqb_curinodes; /* current # allocated inodes */ + __le32 dqb_bhardlimit; /* absolute limit on disk space (in QUOTABLOCK_SIZE) */ + __le32 dqb_bsoftlimit; /* preferred limit on disk space (in QUOTABLOCK_SIZE) */ + __le64 dqb_curspace; /* current space occupied (in bytes) */ + __le64 dqb_btime; /* time limit for excessive disk use */ + __le64 dqb_itime; /* time limit for excessive inode use */ +}; + +/* + * Here are header structures as written on disk and their in-memory copies + */ +/* First generic header */ +struct v2_disk_dqheader { + __le32 dqh_magic; /* Magic number identifying file */ + __le32 dqh_version; /* File version */ +}; + +/* Header with type and version specific information */ +struct v2_disk_dqinfo { + __le32 dqi_bgrace; /* Time before block soft limit becomes hard limit */ + __le32 dqi_igrace; /* Time before inode soft limit becomes hard limit */ + __le32 dqi_flags; /* Flags for quotafile (DQF_*) */ + __le32 dqi_blocks; /* Number of blocks in file */ + __le32 dqi_free_blk; /* Number of first free block in the list */ + __le32 dqi_free_entry; /* Number of block with at least one free entry */ +}; + +/* + * Structure of header of block with quota structures. It is padded to 16 bytes so + * there will be space for exactly 21 quota-entries in a block + */ +struct v2_disk_dqdbheader { + __le32 dqdh_next_free; /* Number of next block with free entry */ + __le32 dqdh_prev_free; /* Number of previous block with free entry */ + __le16 dqdh_entries; /* Number of valid entries in block */ + __le16 dqdh_pad1; + __le32 dqdh_pad2; +}; + +#define V2_DQINFOOFF sizeof(struct v2_disk_dqheader) /* Offset of info header in file */ +#define V2_DQBLKSIZE_BITS 10 +#define V2_DQBLKSIZE (1 << V2_DQBLKSIZE_BITS) /* Size of block with quota structures */ +#define V2_DQTREEOFF 1 /* Offset of tree in file in blocks */ +#define V2_DQTREEDEPTH 4 /* Depth of quota tree */ +#define V2_DQSTRINBLK ((V2_DQBLKSIZE - sizeof(struct v2_disk_dqdbheader)) / sizeof(struct v2_disk_dqblk)) /* Number of entries in one blocks */ + +#endif /* _LINUX_QUOTAIO_V2_H */ diff --git a/include/linux/Kbuild b/include/linux/Kbuild index 95ac82340c3b..900a787cbae9 100644 --- a/include/linux/Kbuild +++ b/include/linux/Kbuild @@ -134,8 +134,6 @@ header-y += posix_types.h header-y += ppdev.h header-y += prctl.h header-y += qnxtypes.h -header-y += quotaio_v1.h -header-y += quotaio_v2.h header-y += radeonfb.h header-y += raw.h header-y += resource.h diff --git a/include/linux/quotaio_v1.h b/include/linux/quotaio_v1.h deleted file mode 100644 index 746654b5de70..000000000000 --- a/include/linux/quotaio_v1.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef _LINUX_QUOTAIO_V1_H -#define _LINUX_QUOTAIO_V1_H - -#include - -/* - * The following constants define the amount of time given a user - * before the soft limits are treated as hard limits (usually resulting - * in an allocation failure). The timer is started when the user crosses - * their soft limit, it is reset when they go below their soft limit. - */ -#define MAX_IQ_TIME 604800 /* (7*24*60*60) 1 week */ -#define MAX_DQ_TIME 604800 /* (7*24*60*60) 1 week */ - -/* - * The following structure defines the format of the disk quota file - * (as it appears on disk) - the file is an array of these structures - * indexed by user or group number. - */ -struct v1_disk_dqblk { - __u32 dqb_bhardlimit; /* absolute limit on disk blks alloc */ - __u32 dqb_bsoftlimit; /* preferred limit on disk blks */ - __u32 dqb_curblocks; /* current block count */ - __u32 dqb_ihardlimit; /* absolute limit on allocated inodes */ - __u32 dqb_isoftlimit; /* preferred inode limit */ - __u32 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 */ -}; - -#define v1_dqoff(UID) ((loff_t)((UID) * sizeof (struct v1_disk_dqblk))) - -#endif /* _LINUX_QUOTAIO_V1_H */ diff --git a/include/linux/quotaio_v2.h b/include/linux/quotaio_v2.h deleted file mode 100644 index 303d7cbe30d4..000000000000 --- a/include/linux/quotaio_v2.h +++ /dev/null @@ -1,79 +0,0 @@ -/* - * Definitions of structures for vfsv0 quota format - */ - -#ifndef _LINUX_QUOTAIO_V2_H -#define _LINUX_QUOTAIO_V2_H - -#include -#include - -/* - * Definitions of magics and versions of current quota files - */ -#define V2_INITQMAGICS {\ - 0xd9c01f11, /* USRQUOTA */\ - 0xd9c01927 /* GRPQUOTA */\ -} - -#define V2_INITQVERSIONS {\ - 0, /* USRQUOTA */\ - 0 /* GRPQUOTA */\ -} - -/* - * The following structure defines the format of the disk quota file - * (as it appears on disk) - the file is a radix tree whose leaves point - * to blocks of these structures. - */ -struct v2_disk_dqblk { - __le32 dqb_id; /* id this quota applies to */ - __le32 dqb_ihardlimit; /* absolute limit on allocated inodes */ - __le32 dqb_isoftlimit; /* preferred inode limit */ - __le32 dqb_curinodes; /* current # allocated inodes */ - __le32 dqb_bhardlimit; /* absolute limit on disk space (in QUOTABLOCK_SIZE) */ - __le32 dqb_bsoftlimit; /* preferred limit on disk space (in QUOTABLOCK_SIZE) */ - __le64 dqb_curspace; /* current space occupied (in bytes) */ - __le64 dqb_btime; /* time limit for excessive disk use */ - __le64 dqb_itime; /* time limit for excessive inode use */ -}; - -/* - * Here are header structures as written on disk and their in-memory copies - */ -/* First generic header */ -struct v2_disk_dqheader { - __le32 dqh_magic; /* Magic number identifying file */ - __le32 dqh_version; /* File version */ -}; - -/* Header with type and version specific information */ -struct v2_disk_dqinfo { - __le32 dqi_bgrace; /* Time before block soft limit becomes hard limit */ - __le32 dqi_igrace; /* Time before inode soft limit becomes hard limit */ - __le32 dqi_flags; /* Flags for quotafile (DQF_*) */ - __le32 dqi_blocks; /* Number of blocks in file */ - __le32 dqi_free_blk; /* Number of first free block in the list */ - __le32 dqi_free_entry; /* Number of block with at least one free entry */ -}; - -/* - * Structure of header of block with quota structures. It is padded to 16 bytes so - * there will be space for exactly 21 quota-entries in a block - */ -struct v2_disk_dqdbheader { - __le32 dqdh_next_free; /* Number of next block with free entry */ - __le32 dqdh_prev_free; /* Number of previous block with free entry */ - __le16 dqdh_entries; /* Number of valid entries in block */ - __le16 dqdh_pad1; - __le32 dqdh_pad2; -}; - -#define V2_DQINFOOFF sizeof(struct v2_disk_dqheader) /* Offset of info header in file */ -#define V2_DQBLKSIZE_BITS 10 -#define V2_DQBLKSIZE (1 << V2_DQBLKSIZE_BITS) /* Size of block with quota structures */ -#define V2_DQTREEOFF 1 /* Offset of tree in file in blocks */ -#define V2_DQTREEDEPTH 4 /* Depth of quota tree */ -#define V2_DQSTRINBLK ((V2_DQBLKSIZE - sizeof(struct v2_disk_dqdbheader)) / sizeof(struct v2_disk_dqblk)) /* Number of entries in one blocks */ - -#endif /* _LINUX_QUOTAIO_V2_H */ -- cgit v1.2.3 From 1ccd14b9c271c1ac6eec5c5ec5def433100e7248 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 22 Sep 2008 05:54:49 +0200 Subject: quota: Split off quota tree handling into a separate file There is going to be a new version of quota format having 64-bit quota limits and a new quota format for OCFS2. They are both going to use the same tree structure as VFSv0 quota format. So split out tree handling into a separate file and make size of leaf blocks, amount of space usable in each block (needed for checksumming) and structures contained in them configurable so that the code can be shared. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/Kconfig | 5 + fs/Makefile | 1 + fs/quota_tree.c | 645 ++++++++++++++++++++++++++++++++++++++++++++ fs/quota_tree.h | 25 ++ fs/quota_v2.c | 596 ++++------------------------------------ fs/quotaio_v2.h | 33 +-- include/linux/dqblk_qtree.h | 56 ++++ include/linux/dqblk_v2.h | 19 +- 8 files changed, 799 insertions(+), 581 deletions(-) create mode 100644 fs/quota_tree.c create mode 100644 fs/quota_tree.h create mode 100644 include/linux/dqblk_qtree.h (limited to 'include') diff --git a/fs/Kconfig b/fs/Kconfig index b93425ad15de..c1ce3d8831d8 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -302,6 +302,10 @@ config PRINT_QUOTA_WARNING Note that this behavior is currently deprecated and may go away in future. Please use notification via netlink socket instead. +# Generic support for tree structured quota files. Seleted when needed. +config QUOTA_TREE + tristate + config QFMT_V1 tristate "Old quota format support" depends on QUOTA @@ -313,6 +317,7 @@ config QFMT_V1 config QFMT_V2 tristate "Quota format v2 support" depends on QUOTA + select QUOTA_TREE help This quota format allows using quotas with 32-bit UIDs/GIDs. If you need this functionality say Y here. diff --git a/fs/Makefile b/fs/Makefile index e6f423d1d228..c830611550d3 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -54,6 +54,7 @@ obj-$(CONFIG_GENERIC_ACL) += generic_acl.o obj-$(CONFIG_QUOTA) += dquot.o obj-$(CONFIG_QFMT_V1) += quota_v1.o obj-$(CONFIG_QFMT_V2) += quota_v2.o +obj-$(CONFIG_QUOTA_TREE) += quota_tree.o obj-$(CONFIG_QUOTACTL) += quota.o obj-$(CONFIG_PROC_FS) += proc/ diff --git a/fs/quota_tree.c b/fs/quota_tree.c new file mode 100644 index 000000000000..953404c95b17 --- /dev/null +++ b/fs/quota_tree.c @@ -0,0 +1,645 @@ +/* + * vfsv0 quota IO operations on file + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "quota_tree.h" + +MODULE_AUTHOR("Jan Kara"); +MODULE_DESCRIPTION("Quota trie support"); +MODULE_LICENSE("GPL"); + +#define __QUOTA_QT_PARANOIA + +typedef char *dqbuf_t; + +static int get_index(struct qtree_mem_dqinfo *info, qid_t id, int depth) +{ + unsigned int epb = info->dqi_usable_bs >> 2; + + depth = info->dqi_qtree_depth - depth - 1; + while (depth--) + id /= epb; + return id % epb; +} + +/* Number of entries in one blocks */ +static inline int qtree_dqstr_in_blk(struct qtree_mem_dqinfo *info) +{ + return (info->dqi_usable_bs - sizeof(struct qt_disk_dqdbheader)) + / info->dqi_entry_size; +} + +static dqbuf_t getdqbuf(size_t size) +{ + dqbuf_t buf = kmalloc(size, GFP_NOFS); + if (!buf) + printk(KERN_WARNING "VFS: Not enough memory for quota buffers.\n"); + return buf; +} + +static inline void freedqbuf(dqbuf_t buf) +{ + kfree(buf); +} + +static inline ssize_t read_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf) +{ + struct super_block *sb = info->dqi_sb; + + memset(buf, 0, info->dqi_usable_bs); + return sb->s_op->quota_read(sb, info->dqi_type, (char *)buf, + info->dqi_usable_bs, blk << info->dqi_blocksize_bits); +} + +static inline ssize_t write_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf) +{ + struct super_block *sb = info->dqi_sb; + + return sb->s_op->quota_write(sb, info->dqi_type, (char *)buf, + info->dqi_usable_bs, blk << info->dqi_blocksize_bits); +} + +/* Remove empty block from list and return it */ +static int get_free_dqblk(struct qtree_mem_dqinfo *info) +{ + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; + int ret, blk; + + if (!buf) + return -ENOMEM; + if (info->dqi_free_blk) { + blk = info->dqi_free_blk; + ret = read_blk(info, blk, buf); + if (ret < 0) + goto out_buf; + info->dqi_free_blk = le32_to_cpu(dh->dqdh_next_free); + } + else { + memset(buf, 0, info->dqi_usable_bs); + /* Assure block allocation... */ + ret = write_blk(info, info->dqi_blocks, buf); + if (ret < 0) + goto out_buf; + blk = info->dqi_blocks++; + } + mark_info_dirty(info->dqi_sb, info->dqi_type); + ret = blk; +out_buf: + freedqbuf(buf); + return ret; +} + +/* Insert empty block to the list */ +static int put_free_dqblk(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk) +{ + struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; + int err; + + dh->dqdh_next_free = cpu_to_le32(info->dqi_free_blk); + dh->dqdh_prev_free = cpu_to_le32(0); + dh->dqdh_entries = cpu_to_le16(0); + err = write_blk(info, blk, buf); + if (err < 0) + return err; + info->dqi_free_blk = blk; + mark_info_dirty(info->dqi_sb, info->dqi_type); + return 0; +} + +/* Remove given block from the list of blocks with free entries */ +static int remove_free_dqentry(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk) +{ + dqbuf_t tmpbuf = getdqbuf(info->dqi_usable_bs); + struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; + uint nextblk = le32_to_cpu(dh->dqdh_next_free); + uint prevblk = le32_to_cpu(dh->dqdh_prev_free); + int err; + + if (!tmpbuf) + return -ENOMEM; + if (nextblk) { + err = read_blk(info, nextblk, tmpbuf); + if (err < 0) + goto out_buf; + ((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = + dh->dqdh_prev_free; + err = write_blk(info, nextblk, tmpbuf); + if (err < 0) + goto out_buf; + } + if (prevblk) { + err = read_blk(info, prevblk, tmpbuf); + if (err < 0) + goto out_buf; + ((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_next_free = + dh->dqdh_next_free; + err = write_blk(info, prevblk, tmpbuf); + if (err < 0) + goto out_buf; + } else { + info->dqi_free_entry = nextblk; + mark_info_dirty(info->dqi_sb, info->dqi_type); + } + freedqbuf(tmpbuf); + dh->dqdh_next_free = dh->dqdh_prev_free = cpu_to_le32(0); + /* No matter whether write succeeds block is out of list */ + if (write_blk(info, blk, buf) < 0) + printk(KERN_ERR "VFS: Can't write block (%u) with free entries.\n", blk); + return 0; +out_buf: + freedqbuf(tmpbuf); + return err; +} + +/* Insert given block to the beginning of list with free entries */ +static int insert_free_dqentry(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk) +{ + dqbuf_t tmpbuf = getdqbuf(info->dqi_usable_bs); + struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; + int err; + + if (!tmpbuf) + return -ENOMEM; + dh->dqdh_next_free = cpu_to_le32(info->dqi_free_entry); + dh->dqdh_prev_free = cpu_to_le32(0); + err = write_blk(info, blk, buf); + if (err < 0) + goto out_buf; + if (info->dqi_free_entry) { + err = read_blk(info, info->dqi_free_entry, tmpbuf); + if (err < 0) + goto out_buf; + ((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = + cpu_to_le32(blk); + err = write_blk(info, info->dqi_free_entry, tmpbuf); + if (err < 0) + goto out_buf; + } + freedqbuf(tmpbuf); + info->dqi_free_entry = blk; + mark_info_dirty(info->dqi_sb, info->dqi_type); + return 0; +out_buf: + freedqbuf(tmpbuf); + return err; +} + +/* Is the entry in the block free? */ +int qtree_entry_unused(struct qtree_mem_dqinfo *info, char *disk) +{ + int i; + + for (i = 0; i < info->dqi_entry_size; i++) + if (disk[i]) + return 0; + return 1; +} +EXPORT_SYMBOL(qtree_entry_unused); + +/* Find space for dquot */ +static uint find_free_dqentry(struct qtree_mem_dqinfo *info, + struct dquot *dquot, int *err) +{ + uint blk, i; + struct qt_disk_dqdbheader *dh; + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + char *ddquot; + + *err = 0; + if (!buf) { + *err = -ENOMEM; + return 0; + } + dh = (struct qt_disk_dqdbheader *)buf; + if (info->dqi_free_entry) { + blk = info->dqi_free_entry; + *err = read_blk(info, blk, buf); + if (*err < 0) + goto out_buf; + } else { + blk = get_free_dqblk(info); + if ((int)blk < 0) { + *err = blk; + freedqbuf(buf); + return 0; + } + memset(buf, 0, info->dqi_usable_bs); + /* This is enough as block is already zeroed and entry list is empty... */ + info->dqi_free_entry = blk; + mark_info_dirty(dquot->dq_sb, dquot->dq_type); + } + /* Block will be full? */ + if (le16_to_cpu(dh->dqdh_entries) + 1 >= qtree_dqstr_in_blk(info)) { + *err = remove_free_dqentry(info, buf, blk); + if (*err < 0) { + printk(KERN_ERR "VFS: find_free_dqentry(): Can't " + "remove block (%u) from entry free list.\n", + blk); + goto out_buf; + } + } + le16_add_cpu(&dh->dqdh_entries, 1); + /* Find free structure in block */ + for (i = 0, ddquot = ((char *)buf) + sizeof(struct qt_disk_dqdbheader); + i < qtree_dqstr_in_blk(info) && !qtree_entry_unused(info, ddquot); + i++, ddquot += info->dqi_entry_size); +#ifdef __QUOTA_QT_PARANOIA + if (i == qtree_dqstr_in_blk(info)) { + printk(KERN_ERR "VFS: find_free_dqentry(): Data block full " + "but it shouldn't.\n"); + *err = -EIO; + goto out_buf; + } +#endif + *err = write_blk(info, blk, buf); + if (*err < 0) { + printk(KERN_ERR "VFS: find_free_dqentry(): Can't write quota " + "data block %u.\n", blk); + goto out_buf; + } + dquot->dq_off = (blk << info->dqi_blocksize_bits) + + sizeof(struct qt_disk_dqdbheader) + + i * info->dqi_entry_size; + freedqbuf(buf); + return blk; +out_buf: + freedqbuf(buf); + return 0; +} + +/* Insert reference to structure into the trie */ +static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, + uint *treeblk, int depth) +{ + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + int ret = 0, newson = 0, newact = 0; + __le32 *ref; + uint newblk; + + if (!buf) + return -ENOMEM; + if (!*treeblk) { + ret = get_free_dqblk(info); + if (ret < 0) + goto out_buf; + *treeblk = ret; + memset(buf, 0, info->dqi_usable_bs); + newact = 1; + } else { + ret = read_blk(info, *treeblk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't read tree quota block " + "%u.\n", *treeblk); + goto out_buf; + } + } + ref = (__le32 *)buf; + newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]); + if (!newblk) + newson = 1; + if (depth == info->dqi_qtree_depth - 1) { +#ifdef __QUOTA_QT_PARANOIA + if (newblk) { + printk(KERN_ERR "VFS: Inserting already present quota " + "entry (block %u).\n", + le32_to_cpu(ref[get_index(info, + dquot->dq_id, depth)])); + ret = -EIO; + goto out_buf; + } +#endif + newblk = find_free_dqentry(info, dquot, &ret); + } else { + ret = do_insert_tree(info, dquot, &newblk, depth+1); + } + if (newson && ret >= 0) { + ref[get_index(info, dquot->dq_id, depth)] = + cpu_to_le32(newblk); + ret = write_blk(info, *treeblk, buf); + } else if (newact && ret < 0) { + put_free_dqblk(info, buf, *treeblk); + } +out_buf: + freedqbuf(buf); + return ret; +} + +/* Wrapper for inserting quota structure into tree */ +static inline int dq_insert_tree(struct qtree_mem_dqinfo *info, + struct dquot *dquot) +{ + int tmp = QT_TREEOFF; + return do_insert_tree(info, dquot, &tmp, 0); +} + +/* + * We don't have to be afraid of deadlocks as we never have quotas on quota files... + */ +int qtree_write_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) +{ + int type = dquot->dq_type; + struct super_block *sb = dquot->dq_sb; + ssize_t ret; + dqbuf_t ddquot = getdqbuf(info->dqi_entry_size); + + if (!ddquot) + return -ENOMEM; + + /* dq_off is guarded by dqio_mutex */ + if (!dquot->dq_off) { + ret = dq_insert_tree(info, dquot); + if (ret < 0) { + printk(KERN_ERR "VFS: Error %zd occurred while " + "creating quota.\n", ret); + freedqbuf(ddquot); + return ret; + } + } + spin_lock(&dq_data_lock); + info->dqi_ops->mem2disk_dqblk(ddquot, dquot); + spin_unlock(&dq_data_lock); + ret = sb->s_op->quota_write(sb, type, (char *)ddquot, + info->dqi_entry_size, dquot->dq_off); + if (ret != info->dqi_entry_size) { + printk(KERN_WARNING "VFS: dquota write failed on dev %s\n", + sb->s_id); + if (ret >= 0) + ret = -ENOSPC; + } else { + ret = 0; + } + dqstats.writes++; + freedqbuf(ddquot); + + return ret; +} +EXPORT_SYMBOL(qtree_write_dquot); + +/* Free dquot entry in data block */ +static int free_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot, + uint blk) +{ + struct qt_disk_dqdbheader *dh; + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + int ret = 0; + + if (!buf) + return -ENOMEM; + if (dquot->dq_off >> info->dqi_blocksize_bits != blk) { + printk(KERN_ERR "VFS: Quota structure has offset to other " + "block (%u) than it should (%u).\n", blk, + (uint)(dquot->dq_off >> info->dqi_blocksize_bits)); + goto out_buf; + } + ret = read_blk(info, blk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't read quota data block %u\n", blk); + goto out_buf; + } + dh = (struct qt_disk_dqdbheader *)buf; + le16_add_cpu(&dh->dqdh_entries, -1); + if (!le16_to_cpu(dh->dqdh_entries)) { /* Block got free? */ + ret = remove_free_dqentry(info, buf, blk); + if (ret >= 0) + ret = put_free_dqblk(info, buf, blk); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't move quota data block (%u) " + "to free list.\n", blk); + goto out_buf; + } + } else { + memset(buf + + (dquot->dq_off & ((1 << info->dqi_blocksize_bits) - 1)), + 0, info->dqi_entry_size); + if (le16_to_cpu(dh->dqdh_entries) == + qtree_dqstr_in_blk(info) - 1) { + /* Insert will write block itself */ + ret = insert_free_dqentry(info, buf, blk); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't insert quota data " + "block (%u) to free entry list.\n", blk); + goto out_buf; + } + } else { + ret = write_blk(info, blk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't write quota data " + "block %u\n", blk); + goto out_buf; + } + } + } + dquot->dq_off = 0; /* Quota is now unattached */ +out_buf: + freedqbuf(buf); + return ret; +} + +/* Remove reference to dquot from tree */ +static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, + uint *blk, int depth) +{ + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + int ret = 0; + uint newblk; + __le32 *ref = (__le32 *)buf; + + if (!buf) + return -ENOMEM; + ret = read_blk(info, *blk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't read quota data block %u\n", *blk); + goto out_buf; + } + newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]); + if (depth == info->dqi_qtree_depth - 1) { + ret = free_dqentry(info, dquot, newblk); + newblk = 0; + } else { + ret = remove_tree(info, dquot, &newblk, depth+1); + } + if (ret >= 0 && !newblk) { + int i; + ref[get_index(info, dquot->dq_id, depth)] = cpu_to_le32(0); + /* Block got empty? */ + for (i = 0; + i < (info->dqi_usable_bs >> 2) && !ref[i]; + i++); + /* Don't put the root block into the free block list */ + if (i == (info->dqi_usable_bs >> 2) + && *blk != QT_TREEOFF) { + put_free_dqblk(info, buf, *blk); + *blk = 0; + } else { + ret = write_blk(info, *blk, buf); + if (ret < 0) + printk(KERN_ERR "VFS: Can't write quota tree " + "block %u.\n", *blk); + } + } +out_buf: + freedqbuf(buf); + return ret; +} + +/* Delete dquot from tree */ +int qtree_delete_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) +{ + uint tmp = QT_TREEOFF; + + if (!dquot->dq_off) /* Even not allocated? */ + return 0; + return remove_tree(info, dquot, &tmp, 0); +} +EXPORT_SYMBOL(qtree_delete_dquot); + +/* Find entry in block */ +static loff_t find_block_dqentry(struct qtree_mem_dqinfo *info, + struct dquot *dquot, uint blk) +{ + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + loff_t ret = 0; + int i; + char *ddquot; + + if (!buf) + return -ENOMEM; + ret = read_blk(info, blk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk); + goto out_buf; + } + for (i = 0, ddquot = ((char *)buf) + sizeof(struct qt_disk_dqdbheader); + i < qtree_dqstr_in_blk(info) && !info->dqi_ops->is_id(ddquot, dquot); + i++, ddquot += info->dqi_entry_size); + if (i == qtree_dqstr_in_blk(info)) { + printk(KERN_ERR "VFS: Quota for id %u referenced " + "but not present.\n", dquot->dq_id); + ret = -EIO; + goto out_buf; + } else { + ret = (blk << info->dqi_blocksize_bits) + sizeof(struct + qt_disk_dqdbheader) + i * info->dqi_entry_size; + } +out_buf: + freedqbuf(buf); + return ret; +} + +/* Find entry for given id in the tree */ +static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info, + struct dquot *dquot, uint blk, int depth) +{ + dqbuf_t buf = getdqbuf(info->dqi_usable_bs); + loff_t ret = 0; + __le32 *ref = (__le32 *)buf; + + if (!buf) + return -ENOMEM; + ret = read_blk(info, blk, buf); + if (ret < 0) { + printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk); + goto out_buf; + } + ret = 0; + blk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]); + if (!blk) /* No reference? */ + goto out_buf; + if (depth < info->dqi_qtree_depth - 1) + ret = find_tree_dqentry(info, dquot, blk, depth+1); + else + ret = find_block_dqentry(info, dquot, blk); +out_buf: + freedqbuf(buf); + return ret; +} + +/* Find entry for given id in the tree - wrapper function */ +static inline loff_t find_dqentry(struct qtree_mem_dqinfo *info, + struct dquot *dquot) +{ + return find_tree_dqentry(info, dquot, QT_TREEOFF, 0); +} + +int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) +{ + int type = dquot->dq_type; + struct super_block *sb = dquot->dq_sb; + loff_t offset; + dqbuf_t ddquot; + int ret = 0; + +#ifdef __QUOTA_QT_PARANOIA + /* Invalidated quota? */ + if (!sb_dqopt(dquot->dq_sb)->files[type]) { + printk(KERN_ERR "VFS: Quota invalidated while reading!\n"); + return -EIO; + } +#endif + /* Do we know offset of the dquot entry in the quota file? */ + if (!dquot->dq_off) { + offset = find_dqentry(info, dquot); + if (offset <= 0) { /* Entry not present? */ + if (offset < 0) + printk(KERN_ERR "VFS: Can't read quota " + "structure for id %u.\n", dquot->dq_id); + dquot->dq_off = 0; + set_bit(DQ_FAKE_B, &dquot->dq_flags); + memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk)); + ret = offset; + goto out; + } + dquot->dq_off = offset; + } + ddquot = getdqbuf(info->dqi_entry_size); + if (!ddquot) + return -ENOMEM; + ret = sb->s_op->quota_read(sb, type, (char *)ddquot, + info->dqi_entry_size, dquot->dq_off); + if (ret != info->dqi_entry_size) { + if (ret >= 0) + ret = -EIO; + printk(KERN_ERR "VFS: Error while reading quota " + "structure for id %u.\n", dquot->dq_id); + set_bit(DQ_FAKE_B, &dquot->dq_flags); + memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk)); + freedqbuf(ddquot); + goto out; + } + spin_lock(&dq_data_lock); + info->dqi_ops->disk2mem_dqblk(dquot, ddquot); + if (!dquot->dq_dqb.dqb_bhardlimit && + !dquot->dq_dqb.dqb_bsoftlimit && + !dquot->dq_dqb.dqb_ihardlimit && + !dquot->dq_dqb.dqb_isoftlimit) + set_bit(DQ_FAKE_B, &dquot->dq_flags); + spin_unlock(&dq_data_lock); + freedqbuf(ddquot); +out: + dqstats.reads++; + return ret; +} +EXPORT_SYMBOL(qtree_read_dquot); + +/* Check whether dquot should not be deleted. We know we are + * the only one operating on dquot (thanks to dq_lock) */ +int qtree_release_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) +{ + if (test_bit(DQ_FAKE_B, &dquot->dq_flags) && !(dquot->dq_dqb.dqb_curinodes | dquot->dq_dqb.dqb_curspace)) + return qtree_delete_dquot(info, dquot); + return 0; +} +EXPORT_SYMBOL(qtree_release_dquot); diff --git a/fs/quota_tree.h b/fs/quota_tree.h new file mode 100644 index 000000000000..a1ab8db81a51 --- /dev/null +++ b/fs/quota_tree.h @@ -0,0 +1,25 @@ +/* + * Definitions of structures for vfsv0 quota format + */ + +#ifndef _LINUX_QUOTA_TREE_H +#define _LINUX_QUOTA_TREE_H + +#include +#include + +/* + * Structure of header of block with quota structures. It is padded to 16 bytes so + * there will be space for exactly 21 quota-entries in a block + */ +struct qt_disk_dqdbheader { + __le32 dqdh_next_free; /* Number of next block with free entry */ + __le32 dqdh_prev_free; /* Number of previous block with free entry */ + __le16 dqdh_entries; /* Number of valid entries in block */ + __le16 dqdh_pad1; + __le32 dqdh_pad2; +}; + +#define QT_TREEOFF 1 /* Offset of tree in file in blocks */ + +#endif /* _LINUX_QUOTAIO_TREE_H */ diff --git a/fs/quota_v2.c b/fs/quota_v2.c index a21d1a7c356a..a87f1028a425 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -14,6 +14,7 @@ #include +#include "quota_tree.h" #include "quotaio_v2.h" MODULE_AUTHOR("Jan Kara"); @@ -22,10 +23,15 @@ MODULE_LICENSE("GPL"); #define __QUOTA_V2_PARANOIA -typedef char *dqbuf_t; +static void v2_mem2diskdqb(void *dp, struct dquot *dquot); +static void v2_disk2memdqb(struct dquot *dquot, void *dp); +static int v2_is_id(void *dp, struct dquot *dquot); -#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))) +static struct qtree_fmt_operations v2_qtree_ops = { + .mem2disk_dqblk = v2_mem2diskdqb, + .disk2mem_dqblk = v2_disk2memdqb, + .is_id = v2_is_id, +}; #define QUOTABLOCK_BITS 10 #define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) @@ -64,7 +70,7 @@ static int v2_check_quota_file(struct super_block *sb, int type) static int v2_read_file_info(struct super_block *sb, int type) { struct v2_disk_dqinfo dinfo; - struct mem_dqinfo *info = sb_dqopt(sb)->info+type; + struct mem_dqinfo *info = sb_dqinfo(sb, type); ssize_t size; size = sb->s_op->quota_read(sb, type, (char *)&dinfo, @@ -80,9 +86,16 @@ static int v2_read_file_info(struct super_block *sb, int type) info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace); info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace); info->dqi_flags = le32_to_cpu(dinfo.dqi_flags); - info->u.v2_i.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks); - info->u.v2_i.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk); - info->u.v2_i.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry); + info->u.v2_i.i.dqi_sb = sb; + info->u.v2_i.i.dqi_type = type; + info->u.v2_i.i.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks); + info->u.v2_i.i.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk); + info->u.v2_i.i.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry); + info->u.v2_i.i.dqi_blocksize_bits = V2_DQBLKSIZE_BITS; + info->u.v2_i.i.dqi_usable_bs = 1 << V2_DQBLKSIZE_BITS; + info->u.v2_i.i.dqi_qtree_depth = qtree_depth(&info->u.v2_i.i); + info->u.v2_i.i.dqi_entry_size = sizeof(struct v2_disk_dqblk); + info->u.v2_i.i.dqi_ops = &v2_qtree_ops; return 0; } @@ -90,7 +103,7 @@ static int v2_read_file_info(struct super_block *sb, int type) static int v2_write_file_info(struct super_block *sb, int type) { struct v2_disk_dqinfo dinfo; - struct mem_dqinfo *info = sb_dqopt(sb)->info+type; + struct mem_dqinfo *info = sb_dqinfo(sb, type); ssize_t size; spin_lock(&dq_data_lock); @@ -99,9 +112,9 @@ static int v2_write_file_info(struct super_block *sb, int type) dinfo.dqi_igrace = cpu_to_le32(info->dqi_igrace); dinfo.dqi_flags = cpu_to_le32(info->dqi_flags & DQF_MASK); spin_unlock(&dq_data_lock); - dinfo.dqi_blocks = cpu_to_le32(info->u.v2_i.dqi_blocks); - dinfo.dqi_free_blk = cpu_to_le32(info->u.v2_i.dqi_free_blk); - dinfo.dqi_free_entry = cpu_to_le32(info->u.v2_i.dqi_free_entry); + dinfo.dqi_blocks = cpu_to_le32(info->u.v2_i.i.dqi_blocks); + dinfo.dqi_free_blk = cpu_to_le32(info->u.v2_i.i.dqi_free_blk); + dinfo.dqi_free_entry = cpu_to_le32(info->u.v2_i.i.dqi_free_entry); size = sb->s_op->quota_write(sb, type, (char *)&dinfo, sizeof(struct v2_disk_dqinfo), V2_DQINFOOFF); if (size != sizeof(struct v2_disk_dqinfo)) { @@ -112,8 +125,11 @@ static int v2_write_file_info(struct super_block *sb, int type) return 0; } -static void disk2memdqb(struct mem_dqblk *m, struct v2_disk_dqblk *d) +static void v2_disk2memdqb(struct dquot *dquot, void *dp) { + struct v2_disk_dqblk *d = dp, empty; + struct mem_dqblk *m = &dquot->dq_dqb; + m->dqb_ihardlimit = le32_to_cpu(d->dqb_ihardlimit); m->dqb_isoftlimit = le32_to_cpu(d->dqb_isoftlimit); m->dqb_curinodes = le32_to_cpu(d->dqb_curinodes); @@ -122,10 +138,20 @@ static void disk2memdqb(struct mem_dqblk *m, struct v2_disk_dqblk *d) 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); + /* We need to escape back all-zero structure */ + memset(&empty, 0, sizeof(struct v2_disk_dqblk)); + empty.dqb_itime = cpu_to_le64(1); + if (!memcmp(&empty, dp, sizeof(struct v2_disk_dqblk))) + m->dqb_itime = 0; } -static void mem2diskdqb(struct v2_disk_dqblk *d, struct mem_dqblk *m, qid_t id) +static void v2_mem2diskdqb(void *dp, struct dquot *dquot) { + struct v2_disk_dqblk *d = dp; + struct mem_dqblk *m = &dquot->dq_dqb; + struct qtree_mem_dqinfo *info = + &sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i; + d->dqb_ihardlimit = cpu_to_le32(m->dqb_ihardlimit); d->dqb_isoftlimit = cpu_to_le32(m->dqb_isoftlimit); d->dqb_curinodes = cpu_to_le32(m->dqb_curinodes); @@ -134,553 +160,35 @@ static void mem2diskdqb(struct v2_disk_dqblk *d, struct mem_dqblk *m, qid_t id) d->dqb_bsoftlimit = cpu_to_le32(v2_stoqb(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); -} - -static dqbuf_t getdqbuf(void) -{ - dqbuf_t buf = kmalloc(V2_DQBLKSIZE, GFP_NOFS); - if (!buf) - printk(KERN_WARNING "VFS: Not enough memory for quota buffers.\n"); - return buf; -} - -static inline void freedqbuf(dqbuf_t buf) -{ - kfree(buf); -} - -static inline ssize_t read_blk(struct super_block *sb, int type, uint blk, dqbuf_t buf) -{ - memset(buf, 0, V2_DQBLKSIZE); - return sb->s_op->quota_read(sb, type, (char *)buf, - V2_DQBLKSIZE, blk << V2_DQBLKSIZE_BITS); -} - -static inline ssize_t write_blk(struct super_block *sb, int type, uint blk, dqbuf_t buf) -{ - return sb->s_op->quota_write(sb, type, (char *)buf, - V2_DQBLKSIZE, blk << V2_DQBLKSIZE_BITS); -} - -/* Remove empty block from list and return it */ -static int get_free_dqblk(struct super_block *sb, int type) -{ - dqbuf_t buf = getdqbuf(); - struct mem_dqinfo *info = sb_dqinfo(sb, type); - struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf; - int ret, blk; - - if (!buf) - return -ENOMEM; - if (info->u.v2_i.dqi_free_blk) { - blk = info->u.v2_i.dqi_free_blk; - if ((ret = read_blk(sb, type, blk, buf)) < 0) - goto out_buf; - info->u.v2_i.dqi_free_blk = le32_to_cpu(dh->dqdh_next_free); - } - else { - memset(buf, 0, V2_DQBLKSIZE); - /* Assure block allocation... */ - if ((ret = write_blk(sb, type, info->u.v2_i.dqi_blocks, buf)) < 0) - goto out_buf; - blk = info->u.v2_i.dqi_blocks++; - } - mark_info_dirty(sb, type); - ret = blk; -out_buf: - freedqbuf(buf); - return ret; -} - -/* Insert empty block to the list */ -static int put_free_dqblk(struct super_block *sb, int type, dqbuf_t buf, uint blk) -{ - struct mem_dqinfo *info = sb_dqinfo(sb, type); - struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf; - int err; - - dh->dqdh_next_free = cpu_to_le32(info->u.v2_i.dqi_free_blk); - dh->dqdh_prev_free = cpu_to_le32(0); - dh->dqdh_entries = cpu_to_le16(0); - info->u.v2_i.dqi_free_blk = blk; - mark_info_dirty(sb, type); - /* Some strange block. We had better leave it... */ - if ((err = write_blk(sb, type, blk, buf)) < 0) - return err; - return 0; + d->dqb_id = cpu_to_le32(dquot->dq_id); + if (qtree_entry_unused(info, dp)) + d->dqb_itime = cpu_to_le64(1); } -/* Remove given block from the list of blocks with free entries */ -static int remove_free_dqentry(struct super_block *sb, int type, dqbuf_t buf, uint blk) +static int v2_is_id(void *dp, struct dquot *dquot) { - dqbuf_t tmpbuf = getdqbuf(); - struct mem_dqinfo *info = sb_dqinfo(sb, type); - struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf; - uint nextblk = le32_to_cpu(dh->dqdh_next_free), prevblk = le32_to_cpu(dh->dqdh_prev_free); - int err; + struct v2_disk_dqblk *d = dp; + struct qtree_mem_dqinfo *info = + &sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i; - if (!tmpbuf) - return -ENOMEM; - if (nextblk) { - if ((err = read_blk(sb, type, nextblk, tmpbuf)) < 0) - goto out_buf; - ((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = dh->dqdh_prev_free; - if ((err = write_blk(sb, type, nextblk, tmpbuf)) < 0) - goto out_buf; - } - if (prevblk) { - if ((err = read_blk(sb, type, prevblk, tmpbuf)) < 0) - goto out_buf; - ((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_next_free = dh->dqdh_next_free; - if ((err = write_blk(sb, type, prevblk, tmpbuf)) < 0) - goto out_buf; - } - else { - info->u.v2_i.dqi_free_entry = nextblk; - mark_info_dirty(sb, type); - } - freedqbuf(tmpbuf); - dh->dqdh_next_free = dh->dqdh_prev_free = cpu_to_le32(0); - /* No matter whether write succeeds block is out of list */ - if (write_blk(sb, type, blk, buf) < 0) - printk(KERN_ERR "VFS: Can't write block (%u) with free entries.\n", blk); - return 0; -out_buf: - freedqbuf(tmpbuf); - return err; -} - -/* Insert given block to the beginning of list with free entries */ -static int insert_free_dqentry(struct super_block *sb, int type, dqbuf_t buf, uint blk) -{ - dqbuf_t tmpbuf = getdqbuf(); - struct mem_dqinfo *info = sb_dqinfo(sb, type); - struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf; - int err; - - if (!tmpbuf) - return -ENOMEM; - dh->dqdh_next_free = cpu_to_le32(info->u.v2_i.dqi_free_entry); - dh->dqdh_prev_free = cpu_to_le32(0); - if ((err = write_blk(sb, type, blk, buf)) < 0) - goto out_buf; - if (info->u.v2_i.dqi_free_entry) { - if ((err = read_blk(sb, type, info->u.v2_i.dqi_free_entry, tmpbuf)) < 0) - goto out_buf; - ((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = cpu_to_le32(blk); - if ((err = write_blk(sb, type, info->u.v2_i.dqi_free_entry, tmpbuf)) < 0) - goto out_buf; - } - freedqbuf(tmpbuf); - info->u.v2_i.dqi_free_entry = blk; - mark_info_dirty(sb, type); - return 0; -out_buf: - freedqbuf(tmpbuf); - return err; -} - -/* Find space for dquot */ -static uint find_free_dqentry(struct dquot *dquot, int *err) -{ - struct super_block *sb = dquot->dq_sb; - struct mem_dqinfo *info = sb_dqopt(sb)->info+dquot->dq_type; - uint blk, i; - struct v2_disk_dqdbheader *dh; - struct v2_disk_dqblk *ddquot; - struct v2_disk_dqblk fakedquot; - dqbuf_t buf; - - *err = 0; - if (!(buf = getdqbuf())) { - *err = -ENOMEM; + if (qtree_entry_unused(info, dp)) return 0; - } - dh = (struct v2_disk_dqdbheader *)buf; - ddquot = GETENTRIES(buf); - if (info->u.v2_i.dqi_free_entry) { - blk = info->u.v2_i.dqi_free_entry; - if ((*err = read_blk(sb, dquot->dq_type, blk, buf)) < 0) - goto out_buf; - } - else { - blk = get_free_dqblk(sb, dquot->dq_type); - if ((int)blk < 0) { - *err = blk; - freedqbuf(buf); - return 0; - } - memset(buf, 0, V2_DQBLKSIZE); - /* This is enough as block is already zeroed and entry list is empty... */ - info->u.v2_i.dqi_free_entry = blk; - mark_info_dirty(sb, dquot->dq_type); - } - if (le16_to_cpu(dh->dqdh_entries)+1 >= V2_DQSTRINBLK) /* Block will be full? */ - if ((*err = remove_free_dqentry(sb, dquot->dq_type, buf, blk)) < 0) { - printk(KERN_ERR "VFS: find_free_dqentry(): Can't remove block (%u) from entry free list.\n", blk); - goto out_buf; - } - le16_add_cpu(&dh->dqdh_entries, 1); - memset(&fakedquot, 0, sizeof(struct v2_disk_dqblk)); - /* Find free structure in block */ - for (i = 0; i < V2_DQSTRINBLK && memcmp(&fakedquot, ddquot+i, sizeof(struct v2_disk_dqblk)); i++); -#ifdef __QUOTA_V2_PARANOIA - if (i == V2_DQSTRINBLK) { - printk(KERN_ERR "VFS: find_free_dqentry(): Data block full but it shouldn't.\n"); - *err = -EIO; - goto out_buf; - } -#endif - if ((*err = write_blk(sb, dquot->dq_type, blk, buf)) < 0) { - printk(KERN_ERR "VFS: find_free_dqentry(): Can't write quota data block %u.\n", blk); - goto out_buf; - } - dquot->dq_off = (blk<dqb_id) == dquot->dq_id; } -/* Insert reference to structure into the trie */ -static int do_insert_tree(struct dquot *dquot, uint *treeblk, int depth) -{ - struct super_block *sb = dquot->dq_sb; - dqbuf_t buf; - int ret = 0, newson = 0, newact = 0; - __le32 *ref; - uint newblk; - - if (!(buf = getdqbuf())) - return -ENOMEM; - if (!*treeblk) { - ret = get_free_dqblk(sb, dquot->dq_type); - if (ret < 0) - goto out_buf; - *treeblk = ret; - memset(buf, 0, V2_DQBLKSIZE); - newact = 1; - } - else { - if ((ret = read_blk(sb, dquot->dq_type, *treeblk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't read tree quota block %u.\n", *treeblk); - goto out_buf; - } - } - ref = (__le32 *)buf; - newblk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]); - if (!newblk) - newson = 1; - if (depth == V2_DQTREEDEPTH-1) { -#ifdef __QUOTA_V2_PARANOIA - if (newblk) { - printk(KERN_ERR "VFS: Inserting already present quota entry (block %u).\n", le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)])); - ret = -EIO; - goto out_buf; - } -#endif - newblk = find_free_dqentry(dquot, &ret); - } - else - ret = do_insert_tree(dquot, &newblk, depth+1); - if (newson && ret >= 0) { - ref[GETIDINDEX(dquot->dq_id, depth)] = cpu_to_le32(newblk); - ret = write_blk(sb, dquot->dq_type, *treeblk, buf); - } - else if (newact && ret < 0) - put_free_dqblk(sb, dquot->dq_type, buf, *treeblk); -out_buf: - freedqbuf(buf); - return ret; -} - -/* Wrapper for inserting quota structure into tree */ -static inline int dq_insert_tree(struct dquot *dquot) +static int v2_read_dquot(struct dquot *dquot) { - int tmp = V2_DQTREEOFF; - return do_insert_tree(dquot, &tmp, 0); + return qtree_read_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); } -/* - * We don't have to be afraid of deadlocks as we never have quotas on quota files... - */ static int v2_write_dquot(struct dquot *dquot) { - int type = dquot->dq_type; - ssize_t ret; - struct v2_disk_dqblk ddquot, empty; - - /* dq_off is guarded by dqio_mutex */ - if (!dquot->dq_off) - if ((ret = dq_insert_tree(dquot)) < 0) { - printk(KERN_ERR "VFS: Error %zd occurred while creating quota.\n", ret); - return ret; - } - spin_lock(&dq_data_lock); - mem2diskdqb(&ddquot, &dquot->dq_dqb, dquot->dq_id); - /* Argh... We may need to write structure full of zeroes but that would be - * treated as an empty place by the rest of the code. Format change would - * be definitely cleaner but the problems probably are not worth it */ - memset(&empty, 0, sizeof(struct v2_disk_dqblk)); - if (!memcmp(&empty, &ddquot, sizeof(struct v2_disk_dqblk))) - ddquot.dqb_itime = cpu_to_le64(1); - spin_unlock(&dq_data_lock); - ret = dquot->dq_sb->s_op->quota_write(dquot->dq_sb, type, - (char *)&ddquot, sizeof(struct v2_disk_dqblk), dquot->dq_off); - if (ret != sizeof(struct v2_disk_dqblk)) { - printk(KERN_WARNING "VFS: dquota write failed on dev %s\n", dquot->dq_sb->s_id); - if (ret >= 0) - ret = -ENOSPC; - } - else - ret = 0; - dqstats.writes++; - - return ret; + return qtree_write_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); } -/* Free dquot entry in data block */ -static int free_dqentry(struct dquot *dquot, uint blk) -{ - struct super_block *sb = dquot->dq_sb; - int type = dquot->dq_type; - struct v2_disk_dqdbheader *dh; - dqbuf_t buf = getdqbuf(); - int ret = 0; - - if (!buf) - return -ENOMEM; - if (dquot->dq_off >> V2_DQBLKSIZE_BITS != blk) { - printk(KERN_ERR "VFS: Quota structure has offset to other " - "block (%u) than it should (%u).\n", blk, - (uint)(dquot->dq_off >> V2_DQBLKSIZE_BITS)); - goto out_buf; - } - if ((ret = read_blk(sb, type, blk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't read quota data block %u\n", blk); - goto out_buf; - } - dh = (struct v2_disk_dqdbheader *)buf; - le16_add_cpu(&dh->dqdh_entries, -1); - if (!le16_to_cpu(dh->dqdh_entries)) { /* Block got free? */ - if ((ret = remove_free_dqentry(sb, type, buf, blk)) < 0 || - (ret = put_free_dqblk(sb, type, buf, blk)) < 0) { - printk(KERN_ERR "VFS: Can't move quota data block (%u) " - "to free list.\n", blk); - goto out_buf; - } - } - else { - memset(buf+(dquot->dq_off & ((1 << V2_DQBLKSIZE_BITS)-1)), 0, - sizeof(struct v2_disk_dqblk)); - if (le16_to_cpu(dh->dqdh_entries) == V2_DQSTRINBLK-1) { - /* Insert will write block itself */ - if ((ret = insert_free_dqentry(sb, type, buf, blk)) < 0) { - printk(KERN_ERR "VFS: Can't insert quota data block (%u) to free entry list.\n", blk); - goto out_buf; - } - } - else - if ((ret = write_blk(sb, type, blk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't write quota data " - "block %u\n", blk); - goto out_buf; - } - } - dquot->dq_off = 0; /* Quota is now unattached */ -out_buf: - freedqbuf(buf); - return ret; -} - -/* Remove reference to dquot from tree */ -static int remove_tree(struct dquot *dquot, uint *blk, int depth) -{ - struct super_block *sb = dquot->dq_sb; - int type = dquot->dq_type; - dqbuf_t buf = getdqbuf(); - int ret = 0; - uint newblk; - __le32 *ref = (__le32 *)buf; - - if (!buf) - return -ENOMEM; - if ((ret = read_blk(sb, type, *blk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't read quota data block %u\n", *blk); - goto out_buf; - } - newblk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]); - if (depth == V2_DQTREEDEPTH-1) { - ret = free_dqentry(dquot, newblk); - newblk = 0; - } - else - ret = remove_tree(dquot, &newblk, depth+1); - if (ret >= 0 && !newblk) { - int i; - ref[GETIDINDEX(dquot->dq_id, depth)] = cpu_to_le32(0); - for (i = 0; i < V2_DQBLKSIZE && !buf[i]; i++); /* Block got empty? */ - /* Don't put the root block into the free block list */ - if (i == V2_DQBLKSIZE && *blk != V2_DQTREEOFF) { - put_free_dqblk(sb, type, buf, *blk); - *blk = 0; - } - else - if ((ret = write_blk(sb, type, *blk, buf)) < 0) - printk(KERN_ERR "VFS: Can't write quota tree " - "block %u.\n", *blk); - } -out_buf: - freedqbuf(buf); - return ret; -} - -/* Delete dquot from tree */ -static int v2_delete_dquot(struct dquot *dquot) -{ - uint tmp = V2_DQTREEOFF; - - if (!dquot->dq_off) /* Even not allocated? */ - return 0; - return remove_tree(dquot, &tmp, 0); -} - -/* Find entry in block */ -static loff_t find_block_dqentry(struct dquot *dquot, uint blk) -{ - dqbuf_t buf = getdqbuf(); - loff_t ret = 0; - int i; - struct v2_disk_dqblk *ddquot = GETENTRIES(buf); - - if (!buf) - return -ENOMEM; - if ((ret = read_blk(dquot->dq_sb, dquot->dq_type, blk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk); - goto out_buf; - } - if (dquot->dq_id) - for (i = 0; i < V2_DQSTRINBLK && - le32_to_cpu(ddquot[i].dqb_id) != dquot->dq_id; i++); - else { /* ID 0 as a bit more complicated searching... */ - struct v2_disk_dqblk fakedquot; - - memset(&fakedquot, 0, sizeof(struct v2_disk_dqblk)); - for (i = 0; i < V2_DQSTRINBLK; i++) - if (!le32_to_cpu(ddquot[i].dqb_id) && - memcmp(&fakedquot, ddquot+i, sizeof(struct v2_disk_dqblk))) - break; - } - if (i == V2_DQSTRINBLK) { - printk(KERN_ERR "VFS: Quota for id %u referenced " - "but not present.\n", dquot->dq_id); - ret = -EIO; - goto out_buf; - } - else - ret = (blk << V2_DQBLKSIZE_BITS) + sizeof(struct - v2_disk_dqdbheader) + i * sizeof(struct v2_disk_dqblk); -out_buf: - freedqbuf(buf); - return ret; -} - -/* Find entry for given id in the tree */ -static loff_t find_tree_dqentry(struct dquot *dquot, uint blk, int depth) -{ - dqbuf_t buf = getdqbuf(); - loff_t ret = 0; - __le32 *ref = (__le32 *)buf; - - if (!buf) - return -ENOMEM; - if ((ret = read_blk(dquot->dq_sb, dquot->dq_type, blk, buf)) < 0) { - printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk); - goto out_buf; - } - ret = 0; - blk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]); - if (!blk) /* No reference? */ - goto out_buf; - if (depth < V2_DQTREEDEPTH-1) - ret = find_tree_dqentry(dquot, blk, depth+1); - else - ret = find_block_dqentry(dquot, blk); -out_buf: - freedqbuf(buf); - return ret; -} - -/* Find entry for given id in the tree - wrapper function */ -static inline loff_t find_dqentry(struct dquot *dquot) -{ - return find_tree_dqentry(dquot, V2_DQTREEOFF, 0); -} - -static int v2_read_dquot(struct dquot *dquot) -{ - int type = dquot->dq_type; - loff_t offset; - struct v2_disk_dqblk ddquot, empty; - int ret = 0; - -#ifdef __QUOTA_V2_PARANOIA - /* Invalidated quota? */ - if (!dquot->dq_sb || !sb_dqopt(dquot->dq_sb)->files[type]) { - printk(KERN_ERR "VFS: Quota invalidated while reading!\n"); - return -EIO; - } -#endif - offset = find_dqentry(dquot); - if (offset <= 0) { /* Entry not present? */ - if (offset < 0) - printk(KERN_ERR "VFS: Can't read quota " - "structure for id %u.\n", dquot->dq_id); - dquot->dq_off = 0; - set_bit(DQ_FAKE_B, &dquot->dq_flags); - memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk)); - ret = offset; - } - else { - dquot->dq_off = offset; - if ((ret = dquot->dq_sb->s_op->quota_read(dquot->dq_sb, type, - (char *)&ddquot, sizeof(struct v2_disk_dqblk), offset)) - != sizeof(struct v2_disk_dqblk)) { - if (ret >= 0) - ret = -EIO; - printk(KERN_ERR "VFS: Error while reading quota " - "structure for id %u.\n", dquot->dq_id); - memset(&ddquot, 0, sizeof(struct v2_disk_dqblk)); - } - else { - ret = 0; - /* We need to escape back all-zero structure */ - memset(&empty, 0, sizeof(struct v2_disk_dqblk)); - empty.dqb_itime = cpu_to_le64(1); - if (!memcmp(&empty, &ddquot, sizeof(struct v2_disk_dqblk))) - ddquot.dqb_itime = 0; - } - disk2memdqb(&dquot->dq_dqb, &ddquot); - if (!dquot->dq_dqb.dqb_bhardlimit && - !dquot->dq_dqb.dqb_bsoftlimit && - !dquot->dq_dqb.dqb_ihardlimit && - !dquot->dq_dqb.dqb_isoftlimit) - set_bit(DQ_FAKE_B, &dquot->dq_flags); - } - dqstats.reads++; - - return ret; -} - -/* Check whether dquot should not be deleted. We know we are - * the only one operating on dquot (thanks to dq_lock) */ static int v2_release_dquot(struct dquot *dquot) { - if (test_bit(DQ_FAKE_B, &dquot->dq_flags) && !(dquot->dq_dqb.dqb_curinodes | dquot->dq_dqb.dqb_curspace)) - return v2_delete_dquot(dquot); - return 0; + return qtree_release_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); } static struct quota_format_ops v2_format_ops = { diff --git a/fs/quotaio_v2.h b/fs/quotaio_v2.h index 303d7cbe30d4..530fe580685c 100644 --- a/fs/quotaio_v2.h +++ b/fs/quotaio_v2.h @@ -21,6 +21,12 @@ 0 /* GRPQUOTA */\ } +/* First generic header */ +struct v2_disk_dqheader { + __le32 dqh_magic; /* Magic number identifying file */ + __le32 dqh_version; /* File version */ +}; + /* * The following structure defines the format of the disk quota file * (as it appears on disk) - the file is a radix tree whose leaves point @@ -38,15 +44,6 @@ struct v2_disk_dqblk { __le64 dqb_itime; /* time limit for excessive inode use */ }; -/* - * Here are header structures as written on disk and their in-memory copies - */ -/* First generic header */ -struct v2_disk_dqheader { - __le32 dqh_magic; /* Magic number identifying file */ - __le32 dqh_version; /* File version */ -}; - /* Header with type and version specific information */ struct v2_disk_dqinfo { __le32 dqi_bgrace; /* Time before block soft limit becomes hard limit */ @@ -57,23 +54,7 @@ struct v2_disk_dqinfo { __le32 dqi_free_entry; /* Number of block with at least one free entry */ }; -/* - * Structure of header of block with quota structures. It is padded to 16 bytes so - * there will be space for exactly 21 quota-entries in a block - */ -struct v2_disk_dqdbheader { - __le32 dqdh_next_free; /* Number of next block with free entry */ - __le32 dqdh_prev_free; /* Number of previous block with free entry */ - __le16 dqdh_entries; /* Number of valid entries in block */ - __le16 dqdh_pad1; - __le32 dqdh_pad2; -}; - #define V2_DQINFOOFF sizeof(struct v2_disk_dqheader) /* Offset of info header in file */ -#define V2_DQBLKSIZE_BITS 10 -#define V2_DQBLKSIZE (1 << V2_DQBLKSIZE_BITS) /* Size of block with quota structures */ -#define V2_DQTREEOFF 1 /* Offset of tree in file in blocks */ -#define V2_DQTREEDEPTH 4 /* Depth of quota tree */ -#define V2_DQSTRINBLK ((V2_DQBLKSIZE - sizeof(struct v2_disk_dqdbheader)) / sizeof(struct v2_disk_dqblk)) /* Number of entries in one blocks */ +#define V2_DQBLKSIZE_BITS 10 /* Size of leaf block in tree */ #endif /* _LINUX_QUOTAIO_V2_H */ diff --git a/include/linux/dqblk_qtree.h b/include/linux/dqblk_qtree.h new file mode 100644 index 000000000000..82a16527b367 --- /dev/null +++ b/include/linux/dqblk_qtree.h @@ -0,0 +1,56 @@ +/* + * Definitions of structures and functions for quota formats using trie + */ + +#ifndef _LINUX_DQBLK_QTREE_H +#define _LINUX_DQBLK_QTREE_H + +#include + +/* Numbers of blocks needed for updates - we count with the smallest + * possible block size (1024) */ +#define QTREE_INIT_ALLOC 4 +#define QTREE_INIT_REWRITE 2 +#define QTREE_DEL_ALLOC 0 +#define QTREE_DEL_REWRITE 6 + +struct dquot; + +/* Operations */ +struct qtree_fmt_operations { + void (*mem2disk_dqblk)(void *disk, struct dquot *dquot); /* Convert given entry from in memory format to disk one */ + void (*disk2mem_dqblk)(struct dquot *dquot, void *disk); /* Convert given entry from disk format to in memory one */ + int (*is_id)(void *disk, struct dquot *dquot); /* Is this structure for given id? */ +}; + +/* Inmemory copy of version specific information */ +struct qtree_mem_dqinfo { + struct super_block *dqi_sb; /* Sb quota is on */ + int dqi_type; /* Quota type */ + unsigned int dqi_blocks; /* # of blocks in quota file */ + unsigned int dqi_free_blk; /* First block in list of free blocks */ + unsigned int dqi_free_entry; /* First block with free entry */ + unsigned int dqi_blocksize_bits; /* Block size of quota file */ + unsigned int dqi_entry_size; /* Size of quota entry in quota file */ + unsigned int dqi_usable_bs; /* Space usable in block for quota data */ + unsigned int dqi_qtree_depth; /* Precomputed depth of quota tree */ + struct qtree_fmt_operations *dqi_ops; /* Operations for entry manipulation */ +}; + +int qtree_write_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot); +int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot); +int qtree_delete_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot); +int qtree_release_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot); +int qtree_entry_unused(struct qtree_mem_dqinfo *info, char *disk); +static inline int qtree_depth(struct qtree_mem_dqinfo *info) +{ + unsigned int epb = info->dqi_usable_bs >> 2; + unsigned long long entries = epb; + int i; + + for (i = 1; entries < (1ULL << 32); i++) + entries *= epb; + return i; +} + +#endif /* _LINUX_DQBLK_QTREE_H */ diff --git a/include/linux/dqblk_v2.h b/include/linux/dqblk_v2.h index 4f853322cb7f..e5e22a787d58 100644 --- a/include/linux/dqblk_v2.h +++ b/include/linux/dqblk_v2.h @@ -1,26 +1,23 @@ /* - * Definitions of structures for vfsv0 quota format + * Definitions for vfsv0 quota format */ #ifndef _LINUX_DQBLK_V2_H #define _LINUX_DQBLK_V2_H -#include +#include -/* id numbers of quota format */ +/* Id number of quota format */ #define QFMT_VFS_V0 2 /* Numbers of blocks needed for updates */ -#define V2_INIT_ALLOC 4 -#define V2_INIT_REWRITE 2 -#define V2_DEL_ALLOC 0 -#define V2_DEL_REWRITE 6 +#define V2_INIT_ALLOC QTREE_INIT_ALLOC +#define V2_INIT_REWRITE QTREE_INIT_REWRITE +#define V2_DEL_ALLOC QTREE_DEL_ALLOC +#define V2_DEL_REWRITE QTREE_DEL_REWRITE -/* Inmemory copy of version specific information */ struct v2_mem_dqinfo { - unsigned int dqi_blocks; - unsigned int dqi_free_blk; - unsigned int dqi_free_entry; + struct qtree_mem_dqinfo i; }; #endif /* _LINUX_DQBLK_V2_H */ -- cgit v1.2.3 From e3d4d56b9715e40ded2a84d0d4fa7f3b6c58983c Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 2 Oct 2008 18:44:14 +0200 Subject: quota: Convert union in mem_dqinfo to a pointer Coming quota support for OCFS2 is going to need quite a bit of additional per-sb quota information. Moreover having fs.h include all the types needed for this structure would be a pain in the a**. So remove the union from mem_dqinfo and add a private pointer for filesystem's use. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/quota_v2.c | 53 +++++++++++++++++++++++++++++++----------------- include/linux/dqblk_v1.h | 4 ---- include/linux/dqblk_v2.h | 4 ---- include/linux/quota.h | 5 +---- 4 files changed, 35 insertions(+), 31 deletions(-) (limited to 'include') diff --git a/fs/quota_v2.c b/fs/quota_v2.c index a87f1028a425..b618b563635c 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -71,6 +71,7 @@ static int v2_read_file_info(struct super_block *sb, int type) { struct v2_disk_dqinfo dinfo; struct mem_dqinfo *info = sb_dqinfo(sb, type); + struct qtree_mem_dqinfo *qinfo; ssize_t size; size = sb->s_op->quota_read(sb, type, (char *)&dinfo, @@ -80,22 +81,29 @@ static int v2_read_file_info(struct super_block *sb, int type) sb->s_id); return -1; } + info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_NOFS); + if (!info->dqi_priv) { + printk(KERN_WARNING + "Not enough memory for quota information structure.\n"); + return -1; + } + qinfo = info->dqi_priv; /* limits are stored as unsigned 32-bit data */ info->dqi_maxblimit = 0xffffffff; info->dqi_maxilimit = 0xffffffff; info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace); info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace); info->dqi_flags = le32_to_cpu(dinfo.dqi_flags); - info->u.v2_i.i.dqi_sb = sb; - info->u.v2_i.i.dqi_type = type; - info->u.v2_i.i.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks); - info->u.v2_i.i.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk); - info->u.v2_i.i.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry); - info->u.v2_i.i.dqi_blocksize_bits = V2_DQBLKSIZE_BITS; - info->u.v2_i.i.dqi_usable_bs = 1 << V2_DQBLKSIZE_BITS; - info->u.v2_i.i.dqi_qtree_depth = qtree_depth(&info->u.v2_i.i); - info->u.v2_i.i.dqi_entry_size = sizeof(struct v2_disk_dqblk); - info->u.v2_i.i.dqi_ops = &v2_qtree_ops; + qinfo->dqi_sb = sb; + qinfo->dqi_type = type; + qinfo->dqi_blocks = le32_to_cpu(dinfo.dqi_blocks); + qinfo->dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk); + qinfo->dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry); + qinfo->dqi_blocksize_bits = V2_DQBLKSIZE_BITS; + qinfo->dqi_usable_bs = 1 << V2_DQBLKSIZE_BITS; + qinfo->dqi_qtree_depth = qtree_depth(qinfo); + qinfo->dqi_entry_size = sizeof(struct v2_disk_dqblk); + qinfo->dqi_ops = &v2_qtree_ops; return 0; } @@ -104,6 +112,7 @@ static int v2_write_file_info(struct super_block *sb, int type) { struct v2_disk_dqinfo dinfo; struct mem_dqinfo *info = sb_dqinfo(sb, type); + struct qtree_mem_dqinfo *qinfo = info->dqi_priv; ssize_t size; spin_lock(&dq_data_lock); @@ -112,9 +121,9 @@ static int v2_write_file_info(struct super_block *sb, int type) dinfo.dqi_igrace = cpu_to_le32(info->dqi_igrace); dinfo.dqi_flags = cpu_to_le32(info->dqi_flags & DQF_MASK); spin_unlock(&dq_data_lock); - dinfo.dqi_blocks = cpu_to_le32(info->u.v2_i.i.dqi_blocks); - dinfo.dqi_free_blk = cpu_to_le32(info->u.v2_i.i.dqi_free_blk); - dinfo.dqi_free_entry = cpu_to_le32(info->u.v2_i.i.dqi_free_entry); + dinfo.dqi_blocks = cpu_to_le32(qinfo->dqi_blocks); + dinfo.dqi_free_blk = cpu_to_le32(qinfo->dqi_free_blk); + dinfo.dqi_free_entry = cpu_to_le32(qinfo->dqi_free_entry); size = sb->s_op->quota_write(sb, type, (char *)&dinfo, sizeof(struct v2_disk_dqinfo), V2_DQINFOOFF); if (size != sizeof(struct v2_disk_dqinfo)) { @@ -150,7 +159,7 @@ static void v2_mem2diskdqb(void *dp, struct dquot *dquot) struct v2_disk_dqblk *d = dp; struct mem_dqblk *m = &dquot->dq_dqb; struct qtree_mem_dqinfo *info = - &sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i; + sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv; d->dqb_ihardlimit = cpu_to_le32(m->dqb_ihardlimit); d->dqb_isoftlimit = cpu_to_le32(m->dqb_isoftlimit); @@ -169,7 +178,7 @@ static int v2_is_id(void *dp, struct dquot *dquot) { struct v2_disk_dqblk *d = dp; struct qtree_mem_dqinfo *info = - &sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i; + sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv; if (qtree_entry_unused(info, dp)) return 0; @@ -178,24 +187,30 @@ static int v2_is_id(void *dp, struct dquot *dquot) static int v2_read_dquot(struct dquot *dquot) { - return qtree_read_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); + return qtree_read_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot); } static int v2_write_dquot(struct dquot *dquot) { - return qtree_write_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); + return qtree_write_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot); } static int v2_release_dquot(struct dquot *dquot) { - return qtree_release_dquot(&sb_dqinfo(dquot->dq_sb, dquot->dq_type)->u.v2_i.i, dquot); + return qtree_release_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot); +} + +static int v2_free_file_info(struct super_block *sb, int type) +{ + kfree(sb_dqinfo(sb, type)->dqi_priv); + return 0; } static struct quota_format_ops v2_format_ops = { .check_quota_file = v2_check_quota_file, .read_file_info = v2_read_file_info, .write_file_info = v2_write_file_info, - .free_file_info = NULL, + .free_file_info = v2_free_file_info, .read_dqblk = v2_read_dquot, .commit_dqblk = v2_write_dquot, .release_dqblk = v2_release_dquot, diff --git a/include/linux/dqblk_v1.h b/include/linux/dqblk_v1.h index 57f1250d5a52..9cea901f5bba 100644 --- a/include/linux/dqblk_v1.h +++ b/include/linux/dqblk_v1.h @@ -17,8 +17,4 @@ #define V1_DEL_ALLOC 0 #define V1_DEL_REWRITE 2 -/* Special information about quotafile */ -struct v1_mem_dqinfo { -}; - #endif /* _LINUX_DQBLK_V1_H */ diff --git a/include/linux/dqblk_v2.h b/include/linux/dqblk_v2.h index e5e22a787d58..ff8af1b4bda7 100644 --- a/include/linux/dqblk_v2.h +++ b/include/linux/dqblk_v2.h @@ -16,8 +16,4 @@ #define V2_DEL_ALLOC QTREE_DEL_ALLOC #define V2_DEL_REWRITE QTREE_DEL_REWRITE -struct v2_mem_dqinfo { - struct qtree_mem_dqinfo i; -}; - #endif /* _LINUX_DQBLK_V2_H */ diff --git a/include/linux/quota.h b/include/linux/quota.h index 80b8807b4988..e51dfdc0aef0 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -208,10 +208,7 @@ struct mem_dqinfo { unsigned int dqi_igrace; qsize_t dqi_maxblimit; qsize_t dqi_maxilimit; - union { - struct v1_mem_dqinfo v1_i; - struct v2_mem_dqinfo v2_i; - } u; + void *dqi_priv; }; struct super_block; -- cgit v1.2.3 From db49d2df489f727096438706a5428115e84a3f0d Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 1 Oct 2008 18:21:39 +0200 Subject: quota: Allow negative usage of space and inodes For clustered filesystems, it can happen that space / inode usage goes negative temporarily (because some node is allocating another node is freeing and they are not completely in sync). So let quota code allow this and change qsize_t so a signed type so that we don't underflow the variables. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 6 ++++-- include/linux/quota.h | 3 ++- 2 files changed, 6 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 74185c34a4f0..9c78ffe1aad2 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -847,7 +847,8 @@ static inline void dquot_incr_space(struct dquot *dquot, qsize_t number) static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number) { - if (dquot->dq_dqb.dqb_curinodes > number) + if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE || + dquot->dq_dqb.dqb_curinodes >= number) dquot->dq_dqb.dqb_curinodes -= number; else dquot->dq_dqb.dqb_curinodes = 0; @@ -858,7 +859,8 @@ static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number) static inline void dquot_decr_space(struct dquot *dquot, qsize_t number) { - if (dquot->dq_dqb.dqb_curspace > number) + if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE || + dquot->dq_dqb.dqb_curspace >= number) dquot->dq_dqb.dqb_curspace -= number; else dquot->dq_dqb.dqb_curspace = 0; diff --git a/include/linux/quota.h b/include/linux/quota.h index e51dfdc0aef0..75bf761caef2 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -168,7 +168,7 @@ enum { #include typedef __kernel_uid32_t qid_t; /* Type in which we store ids in memory */ -typedef __u64 qsize_t; /* Type in which we store sizes */ +typedef long long qsize_t; /* Type in which we store sizes */ extern spinlock_t dq_data_lock; @@ -336,6 +336,7 @@ enum { * responsible for setting * S_NOQUOTA, S_NOATIME flags */ +#define DQUOT_NEGATIVE_USAGE (1 << 7) /* Allow negative quota usage */ static inline unsigned int dquot_state_flag(unsigned int flags, int type) { -- cgit v1.2.3 From 4d59bce4f9eaf26d6d9046b56a2f1c0c7f20981d Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 2 Oct 2008 16:48:10 +0200 Subject: quota: Keep which entries were set by SETQUOTA quotactl Quota in a clustered environment needs to synchronize quota information among cluster nodes. This means we have to occasionally update some information in dquot from disk / network. On the other hand we have to be careful not to overwrite changes administrator did via SETQUOTA. So indicate in dquot->dq_flags which entries have been set by SETQUOTA and quota format can clear these flags when it properly propagated the changes. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 12 ++++++++++-- include/linux/quota.h | 26 ++++++++++++++++++++------ 2 files changed, 30 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 9c78ffe1aad2..89226726daa8 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -2010,25 +2010,33 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) if (di->dqb_valid & QIF_SPACE) { dm->dqb_curspace = di->dqb_curspace; check_blim = 1; + __set_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags); } if (di->dqb_valid & QIF_BLIMITS) { dm->dqb_bsoftlimit = qbtos(di->dqb_bsoftlimit); dm->dqb_bhardlimit = qbtos(di->dqb_bhardlimit); check_blim = 1; + __set_bit(DQ_LASTSET_B + QIF_BLIMITS_B, &dquot->dq_flags); } if (di->dqb_valid & QIF_INODES) { dm->dqb_curinodes = di->dqb_curinodes; check_ilim = 1; + __set_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags); } if (di->dqb_valid & QIF_ILIMITS) { dm->dqb_isoftlimit = di->dqb_isoftlimit; dm->dqb_ihardlimit = di->dqb_ihardlimit; check_ilim = 1; + __set_bit(DQ_LASTSET_B + QIF_ILIMITS_B, &dquot->dq_flags); } - if (di->dqb_valid & QIF_BTIME) + if (di->dqb_valid & QIF_BTIME) { dm->dqb_btime = di->dqb_btime; - if (di->dqb_valid & QIF_ITIME) + __set_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags); + } + if (di->dqb_valid & QIF_ITIME) { dm->dqb_itime = di->dqb_itime; + __set_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags); + } if (check_blim) { if (!dm->dqb_bsoftlimit || dm->dqb_curspace < dm->dqb_bsoftlimit) { diff --git a/include/linux/quota.h b/include/linux/quota.h index 75bf761caef2..6d98885c16da 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -80,12 +80,21 @@ * Quota structure used for communication with userspace via quotactl * Following flags are used to specify which fields are valid */ -#define QIF_BLIMITS 1 -#define QIF_SPACE 2 -#define QIF_ILIMITS 4 -#define QIF_INODES 8 -#define QIF_BTIME 16 -#define QIF_ITIME 32 +enum { + QIF_BLIMITS_B = 0, + QIF_SPACE_B, + QIF_ILIMITS_B, + QIF_INODES_B, + QIF_BTIME_B, + QIF_ITIME_B, +}; + +#define QIF_BLIMITS (1 << QIF_BLIMITS_B) +#define QIF_SPACE (1 << QIF_SPACE_B) +#define QIF_ILIMITS (1 << QIF_ILIMITS_B) +#define QIF_INODES (1 << QIF_INODES_B) +#define QIF_BTIME (1 << QIF_BTIME_B) +#define QIF_ITIME (1 << QIF_ITIME_B) #define QIF_LIMITS (QIF_BLIMITS | QIF_ILIMITS) #define QIF_USAGE (QIF_SPACE | QIF_INODES) #define QIF_TIMES (QIF_BTIME | QIF_ITIME) @@ -242,6 +251,11 @@ extern struct dqstats dqstats; #define DQ_FAKE_B 3 /* no limits only usage */ #define DQ_READ_B 4 /* dquot was read into memory */ #define DQ_ACTIVE_B 5 /* dquot is active (dquot_release not called) */ +#define DQ_LASTSET_B 6 /* Following 6 bits (see QIF_) are reserved\ + * for the mask of entries set via SETQUOTA\ + * quotactl. They are set under dq_data_lock\ + * and the quota format handling dquot can\ + * clear them when it sees fit. */ struct dquot { struct hlist_node dq_hash; /* Hash list in memory */ -- cgit v1.2.3 From 571b46e40bebb0d57130ca24c4a84dfd553adb91 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 30 Oct 2008 09:17:52 +0100 Subject: quota: Update version number 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 Signed-off-by: Mark Fasheh --- include/linux/quota.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/quota.h b/include/linux/quota.h index 6d98885c16da..ec82beb10424 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -36,8 +36,7 @@ #include #include -#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 */ -- cgit v1.2.3 From 3d9ea253a0e73dccaa869888ec2ceb17ea76c810 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 10 Oct 2008 16:12:23 +0200 Subject: 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 Signed-off-by: Mark Fasheh --- fs/dquot.c | 38 ++++++++++++++++++++++++++++++++------ include/linux/quotaops.h | 4 ++++ 2 files changed, 36 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 89226726daa8..ae8fd9e645cc 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; @@ -661,11 +659,29 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type) return dquot; } +/* + * 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 e840ca523175..e3a10272d471 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); -- cgit v1.2.3 From 12c77527e4138bc3b17d17b0e0c909e4fc84924f Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 20 Oct 2008 17:05:00 +0200 Subject: 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 Signed-off-by: Mark Fasheh --- fs/dquot.c | 36 ++++++++++++++++++++++++++++++++++++ include/linux/quotaops.h | 3 +++ 2 files changed, 39 insertions(+) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index ae8fd9e645cc..075dc76904e7 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 e3a10272d471..f4913948c305 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); -- cgit v1.2.3 From e97fcd95a4778a8caf1980c6c72fdf68185a0838 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 18 Nov 2008 17:15:24 -0800 Subject: jbd2: Add BH_JBDPrivateStart Add this so that file systems using JBD2 can safely allocate unused b_state bits. In this case, we add it so that Ocfs2 can define a single bit for tracking the validation state of a buffer. Acked-by: "Theodore Ts'o" Signed-off-by: Mark Fasheh --- include/linux/jbd2.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index c7d106ef22e2..f36645745489 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -329,6 +329,7 @@ enum jbd_state_bits { BH_State, /* Pins most journal_head state */ BH_JournalHead, /* Pins bh->b_private and jh->b_bh */ BH_Unshadow, /* Dummy bit, for BJ_Shadow wakeup filtering */ + BH_JBDPrivateStart, /* First bit available for private use by FS */ }; BUFFER_FNS(JBD, jbd) -- cgit v1.2.3 From 5cd9d5bb86daf632a40f90e2321ea9379e42f073 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 25 Nov 2008 15:31:31 +0100 Subject: quota: Unexport dqblk_v1.h and dqblk_v2.h Unexport header files dqblk_v[12].h since except for quota format ID they don't contain information userspace should be interested in. Move ID definitions to quota.h. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- include/linux/Kbuild | 2 -- include/linux/dqblk_v1.h | 3 --- include/linux/dqblk_v2.h | 3 --- include/linux/quota.h | 4 ++++ 4 files changed, 4 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/include/linux/Kbuild b/include/linux/Kbuild index 900a787cbae9..39da666067b9 100644 --- a/include/linux/Kbuild +++ b/include/linux/Kbuild @@ -56,8 +56,6 @@ header-y += dlm_device.h header-y += dlm_netlink.h header-y += dm-ioctl.h header-y += dn.h -header-y += dqblk_v1.h -header-y += dqblk_v2.h header-y += dqblk_xfs.h header-y += efs_fs_sb.h header-y += elf-fdpic.h diff --git a/include/linux/dqblk_v1.h b/include/linux/dqblk_v1.h index 9cea901f5bba..3713a7232dd8 100644 --- a/include/linux/dqblk_v1.h +++ b/include/linux/dqblk_v1.h @@ -5,9 +5,6 @@ #ifndef _LINUX_DQBLK_V1_H #define _LINUX_DQBLK_V1_H -/* Id of quota format */ -#define QFMT_VFS_OLD 1 - /* Root squash turned on */ #define V1_DQF_RSQUASH 1 diff --git a/include/linux/dqblk_v2.h b/include/linux/dqblk_v2.h index ff8af1b4bda7..18000a542677 100644 --- a/include/linux/dqblk_v2.h +++ b/include/linux/dqblk_v2.h @@ -7,9 +7,6 @@ #include -/* Id number of quota format */ -#define QFMT_VFS_V0 2 - /* Numbers of blocks needed for updates */ #define V2_INIT_ALLOC QTREE_INIT_ALLOC #define V2_INIT_REWRITE QTREE_INIT_REWRITE diff --git a/include/linux/quota.h b/include/linux/quota.h index ec82beb10424..d72d5d84fde5 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -70,6 +70,10 @@ #define Q_GETQUOTA 0x800007 /* get user quota structure */ #define Q_SETQUOTA 0x800008 /* set user quota structure */ +/* Quota format type IDs */ +#define QFMT_VFS_OLD 1 +#define QFMT_VFS_V0 2 + /* Size of block in which space limits are passed through the quota * interface */ #define QIF_DQBLKSIZE_BITS 10 -- cgit v1.2.3 From 7d9056ba20ebed6e3937a2e23183f6117919cb00 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 25 Nov 2008 15:31:32 +0100 Subject: quota: Export dquot_alloc() and dquot_destroy() functions These are default functions for creating and destroying quota structures and they should be used from filesystems. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/dquot.c | 6 ++++-- include/linux/quotaops.h | 2 ++ 2 files changed, 6 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/fs/dquot.c b/fs/dquot.c index 075dc76904e7..61bfff64e5af 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -413,10 +413,11 @@ out_dqlock: return ret; } -static void dquot_destroy(struct dquot *dquot) +void dquot_destroy(struct dquot *dquot) { kmem_cache_free(dquot_cachep, dquot); } +EXPORT_SYMBOL(dquot_destroy); static inline void do_destroy_dquot(struct dquot *dquot) { @@ -668,10 +669,11 @@ we_slept: spin_unlock(&dq_list_lock); } -static struct dquot *dquot_alloc(struct super_block *sb, int type) +struct dquot *dquot_alloc(struct super_block *sb, int type) { return kmem_cache_zalloc(dquot_cachep, GFP_NOFS); } +EXPORT_SYMBOL(dquot_alloc); static struct dquot *get_empty_dquot(struct super_block *sb, int type) { diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index f4913948c305..21b781a3350f 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -31,6 +31,8 @@ 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); +struct dquot *dquot_alloc(struct super_block *sb, int type); +void dquot_destroy(struct dquot *dquot); int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); int dquot_alloc_inode(const struct inode *inode, qsize_t number); -- cgit v1.2.3 From e06c8227fd94ec181849ba206bf032be31c4295c Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Thu, 11 Sep 2008 15:35:47 -0700 Subject: jbd2: Add buffer triggers Filesystems often to do compute intensive operation on some metadata. If this operation is repeated many times, it can be very expensive. It would be much nicer if the operation could be performed once before a buffer goes to disk. This adds triggers to jbd2 buffer heads. Just before writing a metadata buffer to the journal, jbd2 will optionally call a commit trigger associated with the buffer. If the journal is aborted, an abort trigger will be called on any dirty buffers as they are dropped from pending transactions. ocfs2 will use this feature. Initially I tried to come up with a more generic trigger that could be used for non-buffer-related events like transaction completion. It doesn't tie nicely, because the information a buffer trigger needs (specific to a journal_head) isn't the same as what a transaction trigger needs (specific to a tranaction_t or perhaps journal_t). So I implemented a buffer set, with the understanding that journal/transaction wide triggers should be implemented separately. There is only one trigger set allowed per buffer. I can't think of any reason to attach more than one set. Contrast this with a journal or transaction in which multiple places may want to watch the entire transaction separately. The trigger sets are considered static allocation from the jbd2 perspective. ocfs2 will just have one trigger set per block type, setting the same set on every bh of the same type. Signed-off-by: Joel Becker Cc: "Theodore Ts'o" Cc: Signed-off-by: Mark Fasheh --- fs/jbd2/commit.c | 9 +++++++++ fs/jbd2/journal.c | 19 ++++++++++++++++++ fs/jbd2/transaction.c | 47 ++++++++++++++++++++++++++++++++++++++++++++ include/linux/jbd2.h | 31 +++++++++++++++++++++++++++++ include/linux/journal-head.h | 8 ++++++++ 5 files changed, 114 insertions(+) (limited to 'include') diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index ebc667bc54a8..c8a1bace685a 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -509,6 +509,10 @@ void jbd2_journal_commit_transaction(journal_t *journal) if (is_journal_aborted(journal)) { clear_buffer_jbddirty(jh2bh(jh)); JBUFFER_TRACE(jh, "journal is aborting: refile"); + jbd2_buffer_abort_trigger(jh, + jh->b_frozen_data ? + jh->b_frozen_triggers : + jh->b_triggers); jbd2_journal_refile_buffer(journal, jh); /* If that was the last one, we need to clean up * any descriptor buffers which may have been @@ -844,6 +848,9 @@ restart_loop: * data. * * Otherwise, we can just throw away the frozen data now. + * + * We also know that the frozen data has already fired + * its triggers if they exist, so we can clear that too. */ if (jh->b_committed_data) { jbd2_free(jh->b_committed_data, bh->b_size); @@ -851,10 +858,12 @@ restart_loop: if (jh->b_frozen_data) { jh->b_committed_data = jh->b_frozen_data; jh->b_frozen_data = NULL; + jh->b_frozen_triggers = NULL; } } else if (jh->b_frozen_data) { jbd2_free(jh->b_frozen_data, bh->b_size); jh->b_frozen_data = NULL; + jh->b_frozen_triggers = NULL; } spin_lock(&journal->j_list_lock); diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index e70d657a19f8..f6bff9d6f8df 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -50,6 +50,7 @@ EXPORT_SYMBOL(jbd2_journal_unlock_updates); EXPORT_SYMBOL(jbd2_journal_get_write_access); EXPORT_SYMBOL(jbd2_journal_get_create_access); EXPORT_SYMBOL(jbd2_journal_get_undo_access); +EXPORT_SYMBOL(jbd2_journal_set_triggers); EXPORT_SYMBOL(jbd2_journal_dirty_metadata); EXPORT_SYMBOL(jbd2_journal_release_buffer); EXPORT_SYMBOL(jbd2_journal_forget); @@ -290,6 +291,7 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction, struct page *new_page; unsigned int new_offset; struct buffer_head *bh_in = jh2bh(jh_in); + struct jbd2_buffer_trigger_type *triggers; /* * The buffer really shouldn't be locked: only the current committing @@ -314,12 +316,22 @@ repeat: done_copy_out = 1; new_page = virt_to_page(jh_in->b_frozen_data); new_offset = offset_in_page(jh_in->b_frozen_data); + triggers = jh_in->b_frozen_triggers; } else { new_page = jh2bh(jh_in)->b_page; new_offset = offset_in_page(jh2bh(jh_in)->b_data); + triggers = jh_in->b_triggers; } mapped_data = kmap_atomic(new_page, KM_USER0); + /* + * Fire any commit trigger. Do this before checking for escaping, + * as the trigger may modify the magic offset. If a copy-out + * happens afterwards, it will have the correct data in the buffer. + */ + jbd2_buffer_commit_trigger(jh_in, mapped_data + new_offset, + triggers); + /* * Check for escaping */ @@ -352,6 +364,13 @@ repeat: new_page = virt_to_page(tmp); new_offset = offset_in_page(tmp); done_copy_out = 1; + + /* + * This isn't strictly necessary, as we're using frozen + * data for the escaping, but it keeps consistency with + * b_frozen_data usage. + */ + jh_in->b_frozen_triggers = jh_in->b_triggers; } /* diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 39b7805a599a..4f925a4f3d05 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -741,6 +741,12 @@ done: source = kmap_atomic(page, KM_USER0); memcpy(jh->b_frozen_data, source+offset, jh2bh(jh)->b_size); kunmap_atomic(source, KM_USER0); + + /* + * Now that the frozen data is saved off, we need to store + * any matching triggers. + */ + jh->b_frozen_triggers = jh->b_triggers; } jbd_unlock_bh_state(bh); @@ -943,6 +949,47 @@ out: return err; } +/** + * void jbd2_journal_set_triggers() - Add triggers for commit writeout + * @bh: buffer to trigger on + * @type: struct jbd2_buffer_trigger_type containing the trigger(s). + * + * Set any triggers on this journal_head. This is always safe, because + * triggers for a committing buffer will be saved off, and triggers for + * a running transaction will match the buffer in that transaction. + * + * Call with NULL to clear the triggers. + */ +void jbd2_journal_set_triggers(struct buffer_head *bh, + struct jbd2_buffer_trigger_type *type) +{ + struct journal_head *jh = bh2jh(bh); + + jh->b_triggers = type; +} + +void jbd2_buffer_commit_trigger(struct journal_head *jh, void *mapped_data, + struct jbd2_buffer_trigger_type *triggers) +{ + struct buffer_head *bh = jh2bh(jh); + + if (!triggers || !triggers->t_commit) + return; + + triggers->t_commit(triggers, bh, mapped_data, bh->b_size); +} + +void jbd2_buffer_abort_trigger(struct journal_head *jh, + struct jbd2_buffer_trigger_type *triggers) +{ + if (!triggers || !triggers->t_abort) + return; + + triggers->t_abort(triggers, jh2bh(jh)); +} + + + /** * int jbd2_journal_dirty_metadata() - mark a buffer as containing dirty metadata * @handle: transaction to add buffer to. diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index f36645745489..34456476e761 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1008,6 +1008,35 @@ int __jbd2_journal_clean_checkpoint_list(journal_t *journal); int __jbd2_journal_remove_checkpoint(struct journal_head *); void __jbd2_journal_insert_checkpoint(struct journal_head *, transaction_t *); + +/* + * Triggers + */ + +struct jbd2_buffer_trigger_type { + /* + * Fired just before a buffer is written to the journal. + * mapped_data is a mapped buffer that is the frozen data for + * commit. + */ + void (*t_commit)(struct jbd2_buffer_trigger_type *type, + struct buffer_head *bh, void *mapped_data, + size_t size); + + /* + * Fired during journal abort for dirty buffers that will not be + * committed. + */ + void (*t_abort)(struct jbd2_buffer_trigger_type *type, + struct buffer_head *bh); +}; + +extern void jbd2_buffer_commit_trigger(struct journal_head *jh, + void *mapped_data, + struct jbd2_buffer_trigger_type *triggers); +extern void jbd2_buffer_abort_trigger(struct journal_head *jh, + struct jbd2_buffer_trigger_type *triggers); + /* Buffer IO */ extern int jbd2_journal_write_metadata_buffer(transaction_t *transaction, @@ -1046,6 +1075,8 @@ extern int jbd2_journal_extend (handle_t *, int nblocks); extern int jbd2_journal_get_write_access(handle_t *, struct buffer_head *); extern int jbd2_journal_get_create_access (handle_t *, struct buffer_head *); extern int jbd2_journal_get_undo_access(handle_t *, struct buffer_head *); +void jbd2_journal_set_triggers(struct buffer_head *, + struct jbd2_buffer_trigger_type *type); extern int jbd2_journal_dirty_metadata (handle_t *, struct buffer_head *); extern void jbd2_journal_release_buffer (handle_t *, struct buffer_head *); extern int jbd2_journal_forget (handle_t *, struct buffer_head *); diff --git a/include/linux/journal-head.h b/include/linux/journal-head.h index bb70ebb6a2d5..525aac3c97df 100644 --- a/include/linux/journal-head.h +++ b/include/linux/journal-head.h @@ -12,6 +12,8 @@ typedef unsigned int tid_t; /* Unique transaction ID */ typedef struct transaction_s transaction_t; /* Compound transaction type */ + + struct buffer_head; struct journal_head { @@ -87,6 +89,12 @@ struct journal_head { * [j_list_lock] */ struct journal_head *b_cpnext, *b_cpprev; + + /* Trigger type */ + struct jbd2_buffer_trigger_type *b_triggers; + + /* Trigger type for the committing transaction's frozen data */ + struct jbd2_buffer_trigger_type *b_frozen_triggers; }; #endif /* JOURNAL_HEAD_H_INCLUDED */ -- cgit v1.2.3