From 28666d6dc3feca2b1983e6011df383299d8b6b64 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Fri, 13 Feb 2015 11:23:11 -0600 Subject: [PATCH 1/9] Add myself (Bob Peterson) as a maintainer of GFS2 This patch adds Bob Peterson as a maintainer of the GFS2 file system. It also changes the development repository to a shared location rather than Steve Whitehouse's private location. Signed-off-by: Bob Peterson --- MAINTAINERS | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index ddc5a8cf9a8a..f705c6c1b105 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4307,10 +4307,10 @@ F: scripts/get_maintainer.pl GFS2 FILE SYSTEM M: Steven Whitehouse +M: Bob Peterson L: cluster-devel@redhat.com W: http://sources.redhat.com/cluster/ -T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes.git -T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw.git +T: git git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git S: Supported F: Documentation/filesystems/gfs2*.txt F: fs/gfs2/ From 932e468a377f62569fe3818f160fadbaef3f26bd Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Fri, 13 Feb 2015 12:16:37 -0600 Subject: [PATCH 2/9] GFS2: gfs2_set_acl(): Cache "no acl" as well When removing a default acl or setting an access acl that is entirely represented in the file mode, we end up with acl == NULL in gfs2_set_acl(). In that case, bring gfs2 in line with other file systems and cache the NULL acl with set_cached_acl() instead of invalidating the cache with forget_cached_acl(). Signed-off-by: Andreas Gruenbacher Signed-off-by: Bob Peterson --- fs/gfs2/acl.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index 7b3143064af1..1be3b061c05c 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -110,11 +110,7 @@ int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) error = __gfs2_xattr_set(inode, name, data, len, 0, GFS2_EATYPE_SYS); if (error) goto out; - - if (acl) - set_cached_acl(inode, type, acl); - else - forget_cached_acl(inode, type); + set_cached_acl(inode, type, acl); out: kfree(data); return error; From f42a69fadc4b90b5d59c4672dcbe52dcd4b6a117 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Thu, 19 Feb 2015 08:02:16 -0600 Subject: [PATCH 3/9] GFS2: Allocate reservation during splice_write This patch adds a GFS2-specific function for splice_write which first calls function gfs2_rs_alloc to make sure a reservation structure has been allocated before attempting to reserve blocks. Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/file.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index 3e32bb8e2d7e..556d774f9aa9 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -1063,6 +1063,22 @@ static int gfs2_flock(struct file *file, int cmd, struct file_lock *fl) } } +static ssize_t gfs2_file_splice_write(struct pipe_inode_info *pipe, + struct file *out, loff_t *ppos, + size_t len, unsigned int flags) +{ + int error; + struct gfs2_inode *ip = GFS2_I(out->f_mapping->host); + + error = gfs2_rs_alloc(ip); + if (error) + return (ssize_t)error; + + gfs2_size_hint(out, *ppos, len); + + return iter_file_splice_write(pipe, out, ppos, len, flags); +} + const struct file_operations gfs2_file_fops = { .llseek = gfs2_llseek, .read = new_sync_read, @@ -1077,7 +1093,7 @@ const struct file_operations gfs2_file_fops = { .lock = gfs2_lock, .flock = gfs2_flock, .splice_read = generic_file_splice_read, - .splice_write = iter_file_splice_write, + .splice_write = gfs2_file_splice_write, .setlease = simple_nosetlease, .fallocate = gfs2_fallocate, }; @@ -1107,7 +1123,7 @@ const struct file_operations gfs2_file_fops_nolock = { .release = gfs2_release, .fsync = gfs2_fsync, .splice_read = generic_file_splice_read, - .splice_write = iter_file_splice_write, + .splice_write = gfs2_file_splice_write, .setlease = generic_setlease, .fallocate = gfs2_fallocate, }; From f1ea6f4ec0a48d7b6bbf4d380a0ac14d69fadb44 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Tue, 24 Feb 2015 07:22:28 -0600 Subject: [PATCH 4/9] GFS2: Move gfs2_file_splice_write outside of #ifdef This patch moves function gfs2_file_splice_write so it's not conditionally compiled. Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/file.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index 556d774f9aa9..7353c0a01a1e 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -931,6 +931,22 @@ out_uninit: return ret; } +static ssize_t gfs2_file_splice_write(struct pipe_inode_info *pipe, + struct file *out, loff_t *ppos, + size_t len, unsigned int flags) +{ + int error; + struct gfs2_inode *ip = GFS2_I(out->f_mapping->host); + + error = gfs2_rs_alloc(ip); + if (error) + return (ssize_t)error; + + gfs2_size_hint(out, *ppos, len); + + return iter_file_splice_write(pipe, out, ppos, len, flags); +} + #ifdef CONFIG_GFS2_FS_LOCKING_DLM /** @@ -1063,22 +1079,6 @@ static int gfs2_flock(struct file *file, int cmd, struct file_lock *fl) } } -static ssize_t gfs2_file_splice_write(struct pipe_inode_info *pipe, - struct file *out, loff_t *ppos, - size_t len, unsigned int flags) -{ - int error; - struct gfs2_inode *ip = GFS2_I(out->f_mapping->host); - - error = gfs2_rs_alloc(ip); - if (error) - return (ssize_t)error; - - gfs2_size_hint(out, *ppos, len); - - return iter_file_splice_write(pipe, out, ppos, len, flags); -} - const struct file_operations gfs2_file_fops = { .llseek = gfs2_llseek, .read = new_sync_read, From b8fbf471edb3dbf441716fd2a52a7ca76c381381 Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Wed, 18 Mar 2015 12:03:41 -0500 Subject: [PATCH 5/9] gfs2: perform quota checks against allocation parameters Use struct gfs2_alloc_parms as an argument to gfs2_quota_check() and gfs2_quota_lock_check() to check for quota violations while accounting for the new blocks requested by the current operation in ap->target. Previously, the number of new blocks requested during an operation were not accounted for during quota_check and would allow these operations to exceed quota. This was not very apparent since most operations allocated only 1 block at a time and quotas would get violated in the next operation. i.e. quota excess would only be by 1 block or so. With fallocate, (where we allocate a bunch of blocks at once) the quota excess is non-trivial and is addressed by this patch. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/aops.c | 6 +++--- fs/gfs2/bmap.c | 2 +- fs/gfs2/file.c | 15 ++++++++------- fs/gfs2/incore.h | 2 +- fs/gfs2/inode.c | 18 ++++++++++-------- fs/gfs2/quota.c | 6 +++--- fs/gfs2/quota.h | 8 +++++--- fs/gfs2/xattr.c | 2 +- 8 files changed, 32 insertions(+), 27 deletions(-) diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 4ad4f94edebe..7bc5c82423ea 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -671,12 +671,12 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping, if (alloc_required) { struct gfs2_alloc_parms ap = { .aflags = 0, }; - error = gfs2_quota_lock_check(ip); + requested = data_blocks + ind_blocks; + ap.target = requested; + error = gfs2_quota_lock_check(ip, &ap); if (error) goto out_unlock; - requested = data_blocks + ind_blocks; - ap.target = requested; error = gfs2_inplace_reserve(ip, &ap); if (error) goto out_qunlock; diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index f0b945ab853e..61296ecbd0e2 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -1224,7 +1224,7 @@ static int do_grow(struct inode *inode, u64 size) if (gfs2_is_stuffed(ip) && (size > (sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)))) { - error = gfs2_quota_lock_check(ip); + error = gfs2_quota_lock_check(ip, &ap); if (error) return error; diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index 7353c0a01a1e..c569adbc1431 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -429,11 +429,11 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) if (ret) goto out_unlock; - ret = gfs2_quota_lock_check(ip); - if (ret) - goto out_unlock; gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks); ap.target = data_blocks + ind_blocks; + ret = gfs2_quota_lock_check(ip, &ap); + if (ret) + goto out_unlock; ret = gfs2_inplace_reserve(ip, &ap); if (ret) goto out_quota_unlock; @@ -827,13 +827,13 @@ static long __gfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t offset += bytes; continue; } - error = gfs2_quota_lock_check(ip); - if (error) - return error; retry: gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks); - ap.target = data_blocks + ind_blocks; + + error = gfs2_quota_lock_check(ip, &ap); + if (error) + return error; error = gfs2_inplace_reserve(ip, &ap); if (error) { if (error == -ENOSPC && bytes > sdp->sd_sb.sb_bsize) { @@ -841,6 +841,7 @@ retry: bytes &= bsize_mask; if (bytes == 0) bytes = sdp->sd_sb.sb_bsize; + gfs2_quota_unlock(ip); goto retry; } goto out_qunlock; diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 7a2dbbc0d634..3a4ea50c9113 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -301,7 +301,7 @@ struct gfs2_blkreserv { * to the allocation code. */ struct gfs2_alloc_parms { - u32 target; + u64 target; u32 aflags; }; diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 73c72253faac..08bc84d7e768 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -382,7 +382,7 @@ static int alloc_dinode(struct gfs2_inode *ip, u32 flags, unsigned *dblocks) struct gfs2_alloc_parms ap = { .target = *dblocks, .aflags = flags, }; int error; - error = gfs2_quota_lock_check(ip); + error = gfs2_quota_lock_check(ip, &ap); if (error) goto out; @@ -525,7 +525,7 @@ static int link_dinode(struct gfs2_inode *dip, const struct qstr *name, int error; if (da->nr_blocks) { - error = gfs2_quota_lock_check(dip); + error = gfs2_quota_lock_check(dip, &ap); if (error) goto fail_quota_locks; @@ -953,7 +953,7 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, if (da.nr_blocks) { struct gfs2_alloc_parms ap = { .target = da.nr_blocks, }; - error = gfs2_quota_lock_check(dip); + error = gfs2_quota_lock_check(dip, &ap); if (error) goto out_gunlock; @@ -1470,7 +1470,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, if (da.nr_blocks) { struct gfs2_alloc_parms ap = { .target = da.nr_blocks, }; - error = gfs2_quota_lock_check(ndip); + error = gfs2_quota_lock_check(ndip, &ap); if (error) goto out_gunlock; @@ -1669,6 +1669,7 @@ static int setattr_chown(struct inode *inode, struct iattr *attr) kuid_t ouid, nuid; kgid_t ogid, ngid; int error; + struct gfs2_alloc_parms ap; ouid = inode->i_uid; ogid = inode->i_gid; @@ -1696,9 +1697,11 @@ static int setattr_chown(struct inode *inode, struct iattr *attr) if (error) goto out; + ap.target = gfs2_get_inode_blocks(&ip->i_inode); + if (!uid_eq(ouid, NO_UID_QUOTA_CHANGE) || !gid_eq(ogid, NO_GID_QUOTA_CHANGE)) { - error = gfs2_quota_check(ip, nuid, ngid); + error = gfs2_quota_check(ip, nuid, ngid, &ap); if (error) goto out_gunlock_q; } @@ -1713,9 +1716,8 @@ static int setattr_chown(struct inode *inode, struct iattr *attr) if (!uid_eq(ouid, NO_UID_QUOTA_CHANGE) || !gid_eq(ogid, NO_GID_QUOTA_CHANGE)) { - u64 blocks = gfs2_get_inode_blocks(&ip->i_inode); - gfs2_quota_change(ip, -blocks, ouid, ogid); - gfs2_quota_change(ip, blocks, nuid, ngid); + gfs2_quota_change(ip, -ap.target, ouid, ogid); + gfs2_quota_change(ip, ap.target, nuid, ngid); } out_end_trans: diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 3aa17d4d1cfc..964a769a7a86 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1094,7 +1094,8 @@ static int print_message(struct gfs2_quota_data *qd, char *type) return 0; } -int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) +int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid, + struct gfs2_alloc_parms *ap) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data *qd; @@ -1117,14 +1118,13 @@ int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) value = (s64)be64_to_cpu(qd->qd_qb.qb_value); spin_lock(&qd_lock); - value += qd->qd_change; + value += qd->qd_change + ap->target; spin_unlock(&qd_lock); if (be64_to_cpu(qd->qd_qb.qb_limit) && (s64)be64_to_cpu(qd->qd_qb.qb_limit) < value) { print_message(qd, "exceeded"); quota_send_warning(qd->qd_id, sdp->sd_vfs->s_dev, QUOTA_NL_BHARDWARN); - error = -EDQUOT; break; } else if (be64_to_cpu(qd->qd_qb.qb_warn) && diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 55d506eb3c4a..ad04b3acae2b 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -24,7 +24,8 @@ extern void gfs2_quota_unhold(struct gfs2_inode *ip); extern int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid); extern void gfs2_quota_unlock(struct gfs2_inode *ip); -extern int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid); +extern int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid, + struct gfs2_alloc_parms *ap); extern void gfs2_quota_change(struct gfs2_inode *ip, s64 change, kuid_t uid, kgid_t gid); @@ -37,7 +38,8 @@ extern int gfs2_quotad(void *data); extern void gfs2_wake_up_statfs(struct gfs2_sbd *sdp); -static inline int gfs2_quota_lock_check(struct gfs2_inode *ip) +static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, + struct gfs2_alloc_parms *ap) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); int ret; @@ -48,7 +50,7 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip) return ret; if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; - ret = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); + ret = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid, ap); if (ret) gfs2_quota_unlock(ip); return ret; diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c index 0b81f783f787..fd260ce8869a 100644 --- a/fs/gfs2/xattr.c +++ b/fs/gfs2/xattr.c @@ -732,7 +732,7 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er, if (error) return error; - error = gfs2_quota_lock_check(ip); + error = gfs2_quota_lock_check(ip, &ap); if (error) return error; From 25435e5ed611f310dda889940cbc4c45b3ecd9ec Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Wed, 18 Mar 2015 12:04:37 -0500 Subject: [PATCH 6/9] gfs2: allow quota_check and inplace_reserve to return available blocks struct gfs2_alloc_parms is passed to gfs2_quota_check() and gfs2_inplace_reserve() with ap->target containing the number of blocks being requested for allocation in the current operation. We add a new field to struct gfs2_alloc_parms called 'allowed'. gfs2_quota_check() and gfs2_inplace_reserve() return the max blocks allowed by quota and the max blocks allowed by the chosen rgrp respectively in 'allowed'. A new field 'min_target', when non-zero, tells gfs2_quota_check() and gfs2_inplace_reserve() to not return -EDQUOT/-ENOSPC when there are atleast 'min_target' blocks allowable/available. The assumption is that the caller is ok with just 'min_target' blocks and will likely proceed with allocating them. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/incore.h | 2 ++ fs/gfs2/quota.c | 52 ++++++++++++++++++++++++++++++++++++------------ fs/gfs2/rgrp.c | 20 ++++++++++++++----- fs/gfs2/rgrp.h | 3 ++- 4 files changed, 58 insertions(+), 19 deletions(-) diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 3a4ea50c9113..58b75abf6ab2 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -302,7 +302,9 @@ struct gfs2_blkreserv { */ struct gfs2_alloc_parms { u64 target; + u32 min_target; u32 aflags; + u64 allowed; }; enum { diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 964a769a7a86..55614685a31c 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1094,15 +1094,33 @@ static int print_message(struct gfs2_quota_data *qd, char *type) return 0; } +/** + * gfs2_quota_check - check if allocating new blocks will exceed quota + * @ip: The inode for which this check is being performed + * @uid: The uid to check against + * @gid: The gid to check against + * @ap: The allocation parameters. ap->target contains the requested + * blocks. ap->min_target, if set, contains the minimum blks + * requested. + * + * Returns: 0 on success. + * min_req = ap->min_target ? ap->min_target : ap->target; + * quota must allow atleast min_req blks for success and + * ap->allowed is set to the number of blocks allowed + * + * -EDQUOT otherwise, quota violation. ap->allowed is set to number + * of blocks available. + */ int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid, struct gfs2_alloc_parms *ap) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data *qd; - s64 value; + s64 value, warn, limit; unsigned int x; int error = 0; + ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ if (!test_bit(GIF_QD_LOCKED, &ip->i_flags)) return 0; @@ -1116,29 +1134,37 @@ int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid, qid_eq(qd->qd_id, make_kqid_gid(gid)))) continue; + warn = (s64)be64_to_cpu(qd->qd_qb.qb_warn); + limit = (s64)be64_to_cpu(qd->qd_qb.qb_limit); value = (s64)be64_to_cpu(qd->qd_qb.qb_value); spin_lock(&qd_lock); - value += qd->qd_change + ap->target; + value += qd->qd_change; spin_unlock(&qd_lock); - if (be64_to_cpu(qd->qd_qb.qb_limit) && (s64)be64_to_cpu(qd->qd_qb.qb_limit) < value) { - print_message(qd, "exceeded"); - quota_send_warning(qd->qd_id, - sdp->sd_vfs->s_dev, QUOTA_NL_BHARDWARN); - error = -EDQUOT; - break; - } else if (be64_to_cpu(qd->qd_qb.qb_warn) && - (s64)be64_to_cpu(qd->qd_qb.qb_warn) < value && + if (limit > 0 && (limit - value) < ap->allowed) + ap->allowed = limit - value; + /* If we can't meet the target */ + if (limit && limit < (value + (s64)ap->target)) { + /* If no min_target specified or we don't meet + * min_target, return -EDQUOT */ + if (!ap->min_target || ap->min_target > ap->allowed) { + print_message(qd, "exceeded"); + quota_send_warning(qd->qd_id, + sdp->sd_vfs->s_dev, + QUOTA_NL_BHARDWARN); + error = -EDQUOT; + break; + } + } else if (warn && warn < value && time_after_eq(jiffies, qd->qd_last_warn + - gfs2_tune_get(sdp, - gt_quota_warn_period) * HZ)) { + gfs2_tune_get(sdp, gt_quota_warn_period) + * HZ)) { quota_send_warning(qd->qd_id, sdp->sd_vfs->s_dev, QUOTA_NL_BSOFTWARN); error = print_message(qd, "warning"); qd->qd_last_warn = jiffies; } } - return error; } diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 9150207f365c..6af2396a317c 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1946,10 +1946,18 @@ static inline int fast_to_acquire(struct gfs2_rgrpd *rgd) * @ip: the inode to reserve space for * @ap: the allocation parameters * - * Returns: errno + * We try our best to find an rgrp that has at least ap->target blocks + * available. After a couple of passes (loops == 2), the prospects of finding + * such an rgrp diminish. At this stage, we return the first rgrp that has + * atleast ap->min_target blocks available. Either way, we set ap->allowed to + * the number of blocks available in the chosen rgrp. + * + * Returns: 0 on success, + * -ENOMEM if a suitable rgrp can't be found + * errno otherwise */ -int gfs2_inplace_reserve(struct gfs2_inode *ip, const struct gfs2_alloc_parms *ap) +int gfs2_inplace_reserve(struct gfs2_inode *ip, struct gfs2_alloc_parms *ap) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_rgrpd *begin = NULL; @@ -2012,7 +2020,7 @@ int gfs2_inplace_reserve(struct gfs2_inode *ip, const struct gfs2_alloc_parms *a /* Skip unuseable resource groups */ if ((rs->rs_rbm.rgd->rd_flags & (GFS2_RGF_NOALLOC | GFS2_RDF_ERROR)) || - (ap->target > rs->rs_rbm.rgd->rd_extfail_pt)) + (loops == 0 && ap->target > rs->rs_rbm.rgd->rd_extfail_pt)) goto skip_rgrp; if (sdp->sd_args.ar_rgrplvb) @@ -2027,11 +2035,13 @@ int gfs2_inplace_reserve(struct gfs2_inode *ip, const struct gfs2_alloc_parms *a goto check_rgrp; /* If rgrp has enough free space, use it */ - if (rs->rs_rbm.rgd->rd_free_clone >= ap->target) { + if (rs->rs_rbm.rgd->rd_free_clone >= ap->target || + (loops == 2 && ap->min_target && + rs->rs_rbm.rgd->rd_free_clone >= ap->min_target)) { ip->i_rgd = rs->rs_rbm.rgd; + ap->allowed = ip->i_rgd->rd_free_clone; return 0; } - check_rgrp: /* Check for unlinked inodes which can be reclaimed */ if (rs->rs_rbm.rgd->rd_flags & GFS2_RDF_CHECK) diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h index b104f4af3afd..68972ecfbb01 100644 --- a/fs/gfs2/rgrp.h +++ b/fs/gfs2/rgrp.h @@ -41,7 +41,8 @@ extern void gfs2_rgrp_go_unlock(struct gfs2_holder *gh); extern struct gfs2_alloc *gfs2_alloc_get(struct gfs2_inode *ip); #define GFS2_AF_ORLOV 1 -extern int gfs2_inplace_reserve(struct gfs2_inode *ip, const struct gfs2_alloc_parms *ap); +extern int gfs2_inplace_reserve(struct gfs2_inode *ip, + struct gfs2_alloc_parms *ap); extern void gfs2_inplace_release(struct gfs2_inode *ip); extern int gfs2_alloc_blocks(struct gfs2_inode *ip, u64 *bn, unsigned int *n, From d9be0cda774e7fe36651b04a89a42e9e354e43c3 Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Wed, 18 Mar 2015 12:05:15 -0500 Subject: [PATCH 7/9] gfs2: allow fallocate to max out quotas/fs efficiently We can quickly get an estimate of how many blocks are available for allocation restricted by quota and fs size respectively, using the ap->allowed field in the gfs2_alloc_parms structure. gfs2_quota_check() and gfs2_inplace_reserve() provide these values. Once we have the total number of blocks available to us, we can compute how many bytes of data can be written using those blocks instead of guessing inefficiently. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/file.c | 70 +++++++++++++++++++++++++++++++++----------------- 1 file changed, 47 insertions(+), 23 deletions(-) diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index c569adbc1431..4d3108792172 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -765,22 +765,30 @@ out: brelse(dibh); return error; } - -static void calc_max_reserv(struct gfs2_inode *ip, loff_t max, loff_t *len, - unsigned int *data_blocks, unsigned int *ind_blocks) +/** + * calc_max_reserv() - Reverse of write_calc_reserv. Given a number of + * blocks, determine how many bytes can be written. + * @ip: The inode in question. + * @len: Max cap of bytes. What we return in *len must be <= this. + * @data_blocks: Compute and return the number of data blocks needed + * @ind_blocks: Compute and return the number of indirect blocks needed + * @max_blocks: The total blocks available to work with. + * + * Returns: void, but @len, @data_blocks and @ind_blocks are filled in. + */ +static void calc_max_reserv(struct gfs2_inode *ip, loff_t *len, + unsigned int *data_blocks, unsigned int *ind_blocks, + unsigned int max_blocks) { + loff_t max = *len; const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - unsigned int max_blocks = ip->i_rgd->rd_free_clone; unsigned int tmp, max_data = max_blocks - 3 * (sdp->sd_max_height - 1); for (tmp = max_data; tmp > sdp->sd_diptrs;) { tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs); max_data -= tmp; } - /* This calculation isn't the exact reverse of gfs2_write_calc_reserve, - so it might end up with fewer data blocks */ - if (max_data <= *data_blocks) - return; + *data_blocks = max_data; *ind_blocks = max_blocks - max_data; *len = ((loff_t)max_data - 3) << sdp->sd_sb.sb_bsize_shift; @@ -797,7 +805,7 @@ static long __gfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_alloc_parms ap = { .aflags = 0, }; unsigned int data_blocks = 0, ind_blocks = 0, rblocks; - loff_t bytes, max_bytes; + loff_t bytes, max_bytes, max_blks = UINT_MAX; int error; const loff_t pos = offset; const loff_t count = len; @@ -819,6 +827,9 @@ static long __gfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t gfs2_size_hint(file, offset, len); + gfs2_write_calc_reserv(ip, PAGE_SIZE, &data_blocks, &ind_blocks); + ap.min_target = data_blocks + ind_blocks; + while (len > 0) { if (len < bytes) bytes = len; @@ -827,28 +838,41 @@ static long __gfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t offset += bytes; continue; } -retry: + + /* We need to determine how many bytes we can actually + * fallocate without exceeding quota or going over the + * end of the fs. We start off optimistically by assuming + * we can write max_bytes */ + max_bytes = (len > max_chunk_size) ? max_chunk_size : len; + + /* Since max_bytes is most likely a theoretical max, we + * calculate a more realistic 'bytes' to serve as a good + * starting point for the number of bytes we may be able + * to write */ gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks); ap.target = data_blocks + ind_blocks; error = gfs2_quota_lock_check(ip, &ap); if (error) return error; + /* ap.allowed tells us how many blocks quota will allow + * us to write. Check if this reduces max_blks */ + if (ap.allowed && ap.allowed < max_blks) + max_blks = ap.allowed; + error = gfs2_inplace_reserve(ip, &ap); - if (error) { - if (error == -ENOSPC && bytes > sdp->sd_sb.sb_bsize) { - bytes >>= 1; - bytes &= bsize_mask; - if (bytes == 0) - bytes = sdp->sd_sb.sb_bsize; - gfs2_quota_unlock(ip); - goto retry; - } + if (error) goto out_qunlock; - } - max_bytes = bytes; - calc_max_reserv(ip, (len > max_chunk_size)? max_chunk_size: len, - &max_bytes, &data_blocks, &ind_blocks); + + /* check if the selected rgrp limits our max_blks further */ + if (ap.allowed && ap.allowed < max_blks) + max_blks = ap.allowed; + + /* Almost done. Calculate bytes that can be written using + * max_blks. We also recompute max_bytes, data_blocks and + * ind_blocks */ + calc_max_reserv(ip, &max_bytes, &data_blocks, + &ind_blocks, max_blks); rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA + RES_RG_HDR + gfs2_rg_blocks(ip, data_blocks + ind_blocks); From 7b4ddfa7c9523a19fca30830f8b426964d359da2 Mon Sep 17 00:00:00 2001 From: Chengyu Song Date: Tue, 24 Mar 2015 09:37:53 -0500 Subject: [PATCH 8/9] gfs2: incorrect check for debugfs returns debugfs_create_dir and debugfs_create_file may return -ENODEV when debugfs is not configured, so the return value should be checked against ERROR_VALUE as well, otherwise the later dereference of the dentry pointer would crash the kernel. Signed-off-by: Chengyu Song Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/glock.c | 49 +++++++++++++++++++++++++++++-------------------- 1 file changed, 29 insertions(+), 20 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index f42dffba056a..0fa8062f85a7 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -2047,34 +2047,41 @@ static const struct file_operations gfs2_sbstats_fops = { int gfs2_create_debugfs_file(struct gfs2_sbd *sdp) { - sdp->debugfs_dir = debugfs_create_dir(sdp->sd_table_name, gfs2_root); - if (!sdp->debugfs_dir) - return -ENOMEM; - sdp->debugfs_dentry_glocks = debugfs_create_file("glocks", - S_IFREG | S_IRUGO, - sdp->debugfs_dir, sdp, - &gfs2_glocks_fops); - if (!sdp->debugfs_dentry_glocks) - goto fail; + struct dentry *dent; - sdp->debugfs_dentry_glstats = debugfs_create_file("glstats", - S_IFREG | S_IRUGO, - sdp->debugfs_dir, sdp, - &gfs2_glstats_fops); - if (!sdp->debugfs_dentry_glstats) + dent = debugfs_create_dir(sdp->sd_table_name, gfs2_root); + if (IS_ERR_OR_NULL(dent)) goto fail; + sdp->debugfs_dir = dent; - sdp->debugfs_dentry_sbstats = debugfs_create_file("sbstats", - S_IFREG | S_IRUGO, - sdp->debugfs_dir, sdp, - &gfs2_sbstats_fops); - if (!sdp->debugfs_dentry_sbstats) + dent = debugfs_create_file("glocks", + S_IFREG | S_IRUGO, + sdp->debugfs_dir, sdp, + &gfs2_glocks_fops); + if (IS_ERR_OR_NULL(dent)) goto fail; + sdp->debugfs_dentry_glocks = dent; + + dent = debugfs_create_file("glstats", + S_IFREG | S_IRUGO, + sdp->debugfs_dir, sdp, + &gfs2_glstats_fops); + if (IS_ERR_OR_NULL(dent)) + goto fail; + sdp->debugfs_dentry_glstats = dent; + + dent = debugfs_create_file("sbstats", + S_IFREG | S_IRUGO, + sdp->debugfs_dir, sdp, + &gfs2_sbstats_fops); + if (IS_ERR_OR_NULL(dent)) + goto fail; + sdp->debugfs_dentry_sbstats = dent; return 0; fail: gfs2_delete_debugfs_file(sdp); - return -ENOMEM; + return dent ? PTR_ERR(dent) : -ENOMEM; } void gfs2_delete_debugfs_file(struct gfs2_sbd *sdp) @@ -2100,6 +2107,8 @@ void gfs2_delete_debugfs_file(struct gfs2_sbd *sdp) int gfs2_register_debugfs(void) { gfs2_root = debugfs_create_dir("gfs2", NULL); + if (IS_ERR(gfs2_root)) + return PTR_ERR(gfs2_root); return gfs2_root ? 0 : -ENOMEM; } From 30133177957dca9a3e2a37b720f891d3225a92a1 Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Wed, 8 Apr 2015 09:03:56 -0500 Subject: [PATCH 9/9] gfs2: fix quota refresh race in do_glock() quotad periodically syncs in-memory quotas to the ondisk quota file and sets the QDF_REFRESH flag so that a subsequent read of a synced quota is re-read from disk. gfs2_quota_lock() checks for this flag and sets a 'force' bit to force re-read from disk if requested. However, there is a race condition here. It is possible for gfs2_quota_lock() to find the QDF_REFRESH flag unset (i.e force=0) and quotad comes in immediately after and syncs the relevant quota and sets the QDF_REFRESH flag. gfs2_quota_lock() resumes with force=0 and uses the stale in-memory quota usage values that result in miscalculations. This patch fixes this race by moving the check for the QDF_REFRESH flag check further out into the gfs2_quota_lock() process, i.e, in do_glock(), under the protection of the quota glock. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/quota.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 55614685a31c..5c27e48aa76f 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -923,6 +923,9 @@ restart: if (error) return error; + if (test_and_clear_bit(QDF_REFRESH, &qd->qd_flags)) + force_refresh = FORCE; + qd->qd_qb = *(struct gfs2_quota_lvb *)qd->qd_gl->gl_lksb.sb_lvbptr; if (force_refresh || qd->qd_qb.qb_magic != cpu_to_be32(GFS2_MAGIC)) { @@ -974,11 +977,8 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) sizeof(struct gfs2_quota_data *), sort_qd, NULL); for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) { - int force = NO_FORCE; qd = ip->i_res->rs_qa_qd[x]; - if (test_and_clear_bit(QDF_REFRESH, &qd->qd_flags)) - force = FORCE; - error = do_glock(qd, force, &ip->i_res->rs_qa_qd_ghs[x]); + error = do_glock(qd, NO_FORCE, &ip->i_res->rs_qa_qd_ghs[x]); if (error) break; }