/* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions * of the GNU General Public License version 2. */ /* * Quota change tags are associated with each transaction that allocates or * deallocates space. Those changes are accumulated locally to each node (in a * per-node file) and then are periodically synced to the quota file. This * avoids the bottleneck of constantly touching the quota file, but introduces * fuzziness in the current usage value of IDs that are being used on different * nodes in the cluster simultaneously. So, it is possible for a user on * multiple nodes to overrun their quota, but that overrun is controlable. * Since quota tags are part of transactions, there is no need for a quota check * program to be run on node crashes or anything like that. * * There are couple of knobs that let the administrator manage the quota * fuzziness. "quota_quantum" sets the maximum time a quota change can be * sitting on one node before being synced to the quota file. (The default is * 60 seconds.) Another knob, "quota_scale" controls how quickly the frequency * of quota file syncs increases as the user moves closer to their limit. The * more frequent the syncs, the more accurate the quota enforcement, but that * means that there is more contention between the nodes for the quota file. * The default value is one. This sets the maximum theoretical quota overrun * (with infinite node with infinite bandwidth) to twice the user's limit. (In * practice, the maximum overrun you see should be much less.) A "quota_scale" * number greater than one makes quota syncs more frequent and reduces the * maximum overrun. Numbers less than one (but greater than zero) make quota * syncs less frequent. * * GFS quotas also use per-ID Lock Value Blocks (LVBs) to cache the contents of * the quota file, so it is not being constantly read. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "gfs2.h" #include "incore.h" #include "bmap.h" #include "glock.h" #include "glops.h" #include "log.h" #include "meta_io.h" #include "quota.h" #include "rgrp.h" #include "super.h" #include "trans.h" #include "inode.h" #include "util.h" struct gfs2_quota_change_host { u64 qc_change; u32 qc_flags; /* GFS2_QCF_... */ struct kqid qc_id; }; static LIST_HEAD(qd_lru_list); static atomic_t qd_lru_count = ATOMIC_INIT(0); static DEFINE_SPINLOCK(qd_lock); unsigned long gfs2_qd_shrink_scan(struct shrinker *shrink, struct shrink_control *sc) { struct gfs2_quota_data *qd; struct gfs2_sbd *sdp; int nr_to_scan = sc->nr_to_scan; long freed = 0; if (!(sc->gfp_mask & __GFP_FS)) return SHRINK_STOP; spin_lock(&qd_lock); while (nr_to_scan && !list_empty(&qd_lru_list)) { qd = list_entry(qd_lru_list.next, struct gfs2_quota_data, qd_reclaim); sdp = qd->qd_gl->gl_sbd; /* Free from the filesystem-specific list */ list_del(&qd->qd_list); gfs2_assert_warn(sdp, !qd->qd_change); gfs2_assert_warn(sdp, !qd->qd_slot_count); gfs2_assert_warn(sdp, !qd->qd_bh_count); gfs2_glock_put(qd->qd_gl); atomic_dec(&sdp->sd_quota_count); /* Delete it from the common reclaim list */ list_del_init(&qd->qd_reclaim); atomic_dec(&qd_lru_count); spin_unlock(&qd_lock); kmem_cache_free(gfs2_quotad_cachep, qd); spin_lock(&qd_lock); nr_to_scan--; freed++; } spin_unlock(&qd_lock); return freed; } unsigned long gfs2_qd_shrink_count(struct shrinker *shrink, struct shrink_control *sc) { return vfs_pressure_ratio(atomic_read(&qd_lru_count)); } static u64 qd2index(struct gfs2_quota_data *qd) { struct kqid qid = qd->qd_id; return (2 * (u64)from_kqid(&init_user_ns, qid)) + ((qid.type == USRQUOTA) ? 0 : 1); } static u64 qd2offset(struct gfs2_quota_data *qd) { u64 offset; offset = qd2index(qd); offset *= sizeof(struct gfs2_quota); return offset; } static int qd_alloc(struct gfs2_sbd *sdp, struct kqid qid, struct gfs2_quota_data **qdp) { struct gfs2_quota_data *qd; int error; qd = kmem_cache_zalloc(gfs2_quotad_cachep, GFP_NOFS); if (!qd) return -ENOMEM; qd->qd_lockref.count = 1; spin_lock_init(&qd->qd_lockref.lock); qd->qd_id = qid; qd->qd_slot = -1; INIT_LIST_HEAD(&qd->qd_reclaim); error = gfs2_glock_get(sdp, qd2index(qd), &gfs2_quota_glops, CREATE, &qd->qd_gl); if (error) goto fail; *qdp = qd; return 0; fail: kmem_cache_free(gfs2_quotad_cachep, qd); return error; } static int qd_get(struct gfs2_sbd *sdp, struct kqid qid, struct gfs2_quota_data **qdp) { struct gfs2_quota_data *qd = NULL, *new_qd = NULL; int error, found; *qdp = NULL; for (;;) { found = 0; spin_lock(&qd_lock); list_for_each_entry(qd, &sdp->sd_quota_list, qd_list) { if (qid_eq(qd->qd_id, qid)) { lockref_get(&qd->qd_lockref); if (!list_empty(&qd->qd_reclaim)) { /* Remove it from reclaim list */ list_del_init(&qd->qd_reclaim); atomic_dec(&qd_lru_count); } found = 1; break; } } if (!found) qd = NULL; if (!qd && new_qd) { qd = new_qd; list_add(&qd->qd_list, &sdp->sd_quota_list); atomic_inc(&sdp->sd_quota_count); new_qd = NULL; } spin_unlock(&qd_lock); if (qd) { if (new_qd) { gfs2_glock_put(new_qd->qd_gl); kmem_cache_free(gfs2_quotad_cachep, new_qd); } *qdp = qd; return 0; } error = qd_alloc(sdp, qid, &new_qd); if (error) return error; } } static void qd_hold(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; gfs2_assert(sdp, !__lockref_is_dead(&qd->qd_lockref)); lockref_get(&qd->qd_lockref); } static void qd_put(struct gfs2_quota_data *qd) { spin_lock(&qd_lock); if (!lockref_put_or_lock(&qd->qd_lockref)) { /* Add to the reclaim list */ list_add_tail(&qd->qd_reclaim, &qd_lru_list); atomic_inc(&qd_lru_count); spin_unlock(&qd->qd_lockref.lock); } spin_unlock(&qd_lock); } static int slot_get(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; unsigned int c, o = 0, b; unsigned char byte = 0; spin_lock(&qd_lock); if (qd->qd_slot_count++) { spin_unlock(&qd_lock); return 0; } for (c = 0; c < sdp->sd_quota_chunks; c++) for (o = 0; o < PAGE_SIZE; o++) { byte = sdp->sd_quota_bitmap[c][o]; if (byte != 0xFF) goto found; } goto fail; found: for (b = 0; b < 8; b++) if (!(byte & (1 << b))) break; qd->qd_slot = c * (8 * PAGE_SIZE) + o * 8 + b; if (qd->qd_slot >= sdp->sd_quota_slots) goto fail; sdp->sd_quota_bitmap[c][o] |= 1 << b; spin_unlock(&qd_lock); return 0; fail: qd->qd_slot_count--; spin_unlock(&qd_lock); return -ENOSPC; } static void slot_hold(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; spin_lock(&qd_lock); gfs2_assert(sdp, qd->qd_slot_count); qd->qd_slot_count++; spin_unlock(&qd_lock); } static void gfs2_icbit_munge(struct gfs2_sbd *sdp, unsigned char **bitmap, unsigned int bit, int new_value) { unsigned int c, o, b = bit; int old_value; c = b / (8 * PAGE_SIZE); b %= 8 * PAGE_SIZE; o = b / 8; b %= 8; old_value = (bitmap[c][o] & (1 << b)); gfs2_assert_withdraw(sdp, !old_value != !new_value); if (new_value) bitmap[c][o] |= 1 << b; else bitmap[c][o] &= ~(1 << b); } static void slot_put(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; spin_lock(&qd_lock); gfs2_assert(sdp, qd->qd_slot_count); if (!--qd->qd_slot_count) { gfs2_icbit_munge(sdp, sdp->sd_quota_bitmap, qd->qd_slot, 0); qd->qd_slot = -1; } spin_unlock(&qd_lock); } static int bh_get(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; struct gfs2_inode *ip = GFS2_I(sdp->sd_qc_inode); unsigned int block, offset; struct buffer_head *bh; int error; struct buffer_head bh_map = { .b_state = 0, .b_blocknr = 0 }; mutex_lock(&sdp->sd_quota_mutex); if (qd->qd_bh_count++) { mutex_unlock(&sdp->sd_quota_mutex); return 0; } block = qd->qd_slot / sdp->sd_qc_per_block; offset = qd->qd_slot % sdp->sd_qc_per_block; bh_map.b_size = 1 << ip->i_inode.i_blkbits; error = gfs2_block_map(&ip->i_inode, block, &bh_map, 0); if (error) goto fail; error = gfs2_meta_read(ip->i_gl, bh_map.b_blocknr, DIO_WAIT, &bh); if (error) goto fail; error = -EIO; if (gfs2_metatype_check(sdp, bh, GFS2_METATYPE_QC)) goto fail_brelse; qd->qd_bh = bh; qd->qd_bh_qc = (struct gfs2_quota_change *) (bh->b_data + sizeof(struct gfs2_meta_header) + offset * sizeof(struct gfs2_quota_change)); mutex_unlock(&sdp->sd_quota_mutex); return 0; fail_brelse: brelse(bh); fail: qd->qd_bh_count--; mutex_unlock(&sdp->sd_quota_mutex); return error; } static void bh_put(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; mutex_lock(&sdp->sd_quota_mutex); gfs2_assert(sdp, qd->qd_bh_count); if (!--qd->qd_bh_count) { brelse(qd->qd_bh); qd->qd_bh = NULL; qd->qd_bh_qc = NULL; } mutex_unlock(&sdp->sd_quota_mutex); } static int qd_check_sync(struct gfs2_sbd *sdp, struct gfs2_quota_data *qd, u64 *sync_gen) { if (test_bit(QDF_LOCKED, &qd->qd_flags) || !test_bit(QDF_CHANGE, &qd->qd_flags) || (sync_gen && (qd->qd_sync_gen >= *sync_gen))) return 0; list_move_tail(&qd->qd_list, &sdp->sd_quota_list); set_bit(QDF_LOCKED, &qd->qd_flags); gfs2_assert_warn(sdp, !__lockref_is_dead(&qd->qd_lockref)); lockref_get(&qd->qd_lockref); qd->qd_change_sync = qd->qd_change; gfs2_assert_warn(sdp, qd->qd_slot_count); qd->qd_slot_count++; return 1; } static int qd_fish(struct gfs2_sbd *sdp, struct gfs2_quota_data **qdp) { struct gfs2_quota_data *qd = NULL; int error; int found = 0; *qdp = NULL; if (sdp->sd_vfs->s_flags & MS_RDONLY) return 0; spin_lock(&qd_lock); list_for_each_entry(qd, &sdp->sd_quota_list, qd_list) { found = qd_check_sync(sdp, qd, &sdp->sd_quota_sync_gen); if (found) break; } if (!found) qd = NULL; spin_unlock(&qd_lock); if (qd) { gfs2_assert_warn(sdp, qd->qd_change_sync); error = bh_get(qd); if (error) { clear_bit(QDF_LOCKED, &qd->qd_flags); slot_put(qd); qd_put(qd); return error; } } *qdp = qd; return 0; } static void qd_unlock(struct gfs2_quota_data *qd) { gfs2_assert_warn(qd->qd_gl->gl_sbd, test_bit(QDF_LOCKED, &qd->qd_flags)); clear_bit(QDF_LOCKED, &qd->qd_flags); bh_put(qd); slot_put(qd); qd_put(qd); } static int qdsb_get(struct gfs2_sbd *sdp, struct kqid qid, struct gfs2_quota_data **qdp) { int error; error = qd_get(sdp, qid, qdp); if (error) return error; error = slot_get(*qdp); if (error) goto fail; error = bh_get(*qdp); if (error) goto fail_slot; return 0; fail_slot: slot_put(*qdp); fail: qd_put(*qdp); return error; } static void qdsb_put(struct gfs2_quota_data *qd) { bh_put(qd); slot_put(qd); qd_put(qd); } int gfs2_quota_hold(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data **qd; int error; if (ip->i_res == NULL) { error = gfs2_rs_alloc(ip); if (error) return error; } qd = ip->i_res->rs_qa_qd; if (gfs2_assert_warn(sdp, !ip->i_res->rs_qa_qd_num) || gfs2_assert_warn(sdp, !test_bit(GIF_QD_LOCKED, &ip->i_flags))) return -EIO; if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; error = qdsb_get(sdp, make_kqid_uid(ip->i_inode.i_uid), qd); if (error) goto out; ip->i_res->rs_qa_qd_num++; qd++; error = qdsb_get(sdp, make_kqid_gid(ip->i_inode.i_gid), qd); if (error) goto out; ip->i_res->rs_qa_qd_num++; qd++; if (!uid_eq(uid, NO_UID_QUOTA_CHANGE) && !uid_eq(uid, ip->i_inode.i_uid)) { error = qdsb_get(sdp, make_kqid_uid(uid), qd); if (error) goto out; ip->i_res->rs_qa_qd_num++; qd++; } if (!gid_eq(gid, NO_GID_QUOTA_CHANGE) && !gid_eq(gid, ip->i_inode.i_gid)) { error = qdsb_get(sdp, make_kqid_gid(gid), qd); if (error) goto out; ip->i_res->rs_qa_qd_num++; qd++; } out: if (error) gfs2_quota_unhold(ip); return error; } void gfs2_quota_unhold(struct gfs2_inode *ip) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); unsigned int x; if (ip->i_res == NULL) return; gfs2_assert_warn(sdp, !test_bit(GIF_QD_LOCKED, &ip->i_flags)); for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) { qdsb_put(ip->i_res->rs_qa_qd[x]); ip->i_res->rs_qa_qd[x] = NULL; } ip->i_res->rs_qa_qd_num = 0; } static int sort_qd(const void *a, const void *b) { const struct gfs2_quota_data *qd_a = *(const struct gfs2_quota_data **)a; const struct gfs2_quota_data *qd_b = *(const struct gfs2_quota_data **)b; if (qid_lt(qd_a->qd_id, qd_b->qd_id)) return -1; if (qid_lt(qd_b->qd_id, qd_a->qd_id)) return 1; return 0; } static void do_qc(struct gfs2_quota_data *qd, s64 change) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; struct gfs2_inode *ip = GFS2_I(sdp->sd_qc_inode); struct gfs2_quota_change *qc = qd->qd_bh_qc; s64 x; mutex_lock(&sdp->sd_quota_mutex); gfs2_trans_add_meta(ip->i_gl, qd->qd_bh); if (!test_bit(QDF_CHANGE, &qd->qd_flags)) { qc->qc_change = 0; qc->qc_flags = 0; if (qd->qd_id.type == USRQUOTA) qc->qc_flags = cpu_to_be32(GFS2_QCF_USER); qc->qc_id = cpu_to_be32(from_kqid(&init_user_ns, qd->qd_id)); } x = be64_to_cpu(qc->qc_change) + change; qc->qc_change = cpu_to_be64(x); spin_lock(&qd_lock); qd->qd_change = x; spin_unlock(&qd_lock); if (!x) { gfs2_assert_warn(sdp, test_bit(QDF_CHANGE, &qd->qd_flags)); clear_bit(QDF_CHANGE, &qd->qd_flags); qc->qc_flags = 0; qc->qc_id = 0; slot_put(qd); qd_put(qd); } else if (!test_and_set_bit(QDF_CHANGE, &qd->qd_flags)) { qd_hold(qd); slot_hold(qd); } mutex_unlock(&sdp->sd_quota_mutex); } /** * gfs2_adjust_quota - adjust record of current block usage * @ip: The quota inode * @loc: Offset of the entry in the quota file * @change: The amount of usage change to record * @qd: The quota data * @fdq: The updated limits to record * * This function was mostly borrowed from gfs2_block_truncate_page which was * in turn mostly borrowed from ext3 * * Returns: 0 or -ve on error */ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc, s64 change, struct gfs2_quota_data *qd, struct fs_disk_quota *fdq) { struct inode *inode = &ip->i_inode; struct gfs2_sbd *sdp = GFS2_SB(inode); struct address_space *mapping = inode->i_mapping; unsigned long index = loc >> PAGE_CACHE_SHIFT; unsigned offset = loc & (PAGE_CACHE_SIZE - 1); unsigned blocksize, iblock, pos; struct buffer_head *bh; struct page *page; void *kaddr, *ptr; struct gfs2_quota q, *qp; int err, nbytes; u64 size; if (gfs2_is_stuffed(ip)) { err = gfs2_unstuff_dinode(ip, NULL); if (err) return err; } memset(&q, 0, sizeof(struct gfs2_quota)); err = gfs2_internal_read(ip, (char *)&q, &loc, sizeof(q)); if (err < 0) return err; err = -EIO; qp = &q; qp->qu_value = be64_to_cpu(qp->qu_value); qp->qu_value += change; qp->qu_value = cpu_to_be64(qp->qu_value); qd->qd_qb.qb_value = qp->qu_value; if (fdq) { if (fdq->d_fieldmask & FS_DQ_BSOFT) { qp->qu_warn = cpu_to_be64(fdq->d_blk_softlimit >> sdp->sd_fsb2bb_shift); qd->qd_qb.qb_warn = qp->qu_warn; } if (fdq->d_fieldmask & FS_DQ_BHARD) { qp->qu_limit = cpu_to_be64(fdq->d_blk_hardlimit >> sdp->sd_fsb2bb_shift); qd->qd_qb.qb_limit = qp->qu_limit; } if (fdq->d_fieldmask & FS_DQ_BCOUNT) { qp->qu_value = cpu_to_be64(fdq->d_bcount >> sdp->sd_fsb2bb_shift); qd->qd_qb.qb_value = qp->qu_value; } } /* Write the quota into the quota file on disk */ ptr = qp; nbytes = sizeof(struct gfs2_quota); get_a_page: page = find_or_create_page(mapping, index, GFP_NOFS); if (!page) return -ENOMEM; blocksize = inode->i_sb->s_blocksize; iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits); if (!page_has_buffers(page)) create_empty_buffers(page, blocksize, 0); bh = page_buffers(page); pos = blocksize; while (offset >= pos) { bh = bh->b_this_page; iblock++; pos += blocksize; } if (!buffer_mapped(bh)) { gfs2_block_map(inode, iblock, bh, 1); if (!buffer_mapped(bh)) goto unlock_out; /* If it's a newly allocated disk block for quota, zero it */ if (buffer_new(bh)) zero_user(page, pos - blocksize, bh->b_size); } if (PageUptodate(page)) set_buffer_uptodate(bh); if (!buffer_uptodate(bh)) { ll_rw_block(READ | REQ_META, 1, &bh); wait_on_buffer(bh); if (!buffer_uptodate(bh)) goto unlock_out; } gfs2_trans_add_data(ip->i_gl, bh); kaddr = kmap_atomic(page); if (offset + sizeof(struct gfs2_quota) > PAGE_CACHE_SIZE) nbytes = PAGE_CACHE_SIZE - offset; memcpy(kaddr + offset, ptr, nbytes); flush_dcache_page(page); kunmap_atomic(kaddr); unlock_page(page); page_cache_release(page); /* If quota straddles page boundary, we need to update the rest of the * quota at the beginning of the next page */ if ((offset + sizeof(struct gfs2_quota)) > PAGE_CACHE_SIZE) { ptr = ptr + nbytes; nbytes = sizeof(struct gfs2_quota) - nbytes; offset = 0; index++; goto get_a_page; } size = loc + sizeof(struct gfs2_quota); if (size > inode->i_size) i_size_write(inode, size); inode->i_mtime = inode->i_atime = CURRENT_TIME; mark_inode_dirty(inode); return 0; unlock_out: unlock_page(page); page_cache_release(page); return err; } static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) { struct gfs2_sbd *sdp = (*qda)->qd_gl->gl_sbd; struct gfs2_inode *ip = GFS2_I(sdp->sd_quota_inode); struct gfs2_alloc_parms ap = { .aflags = 0, }; unsigned int data_blocks, ind_blocks; struct gfs2_holder *ghs, i_gh; unsigned int qx, x; struct gfs2_quota_data *qd; unsigned reserved; loff_t offset; unsigned int nalloc = 0, blocks; int error; error = gfs2_rs_alloc(ip); if (error) return error; gfs2_write_calc_reserv(ip, sizeof(struct gfs2_quota), &data_blocks, &ind_blocks); ghs = kcalloc(num_qd, sizeof(struct gfs2_holder), GFP_NOFS); if (!ghs) return -ENOMEM; sort(qda, num_qd, sizeof(struct gfs2_quota_data *), sort_qd, NULL); mutex_lock(&ip->i_inode.i_mutex); for (qx = 0; qx < num_qd; qx++) { error = gfs2_glock_nq_init(qda[qx]->qd_gl, LM_ST_EXCLUSIVE, GL_NOCACHE, &ghs[qx]); if (error) goto out; } error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &i_gh); if (error) goto out; for (x = 0; x < num_qd; x++) { offset = qd2offset(qda[x]); if (gfs2_write_alloc_required(ip, offset, sizeof(struct gfs2_quota))) nalloc++; } /* * 1 blk for unstuffing inode if stuffed. We add this extra * block to the reservation unconditionally. If the inode * doesn't need unstuffing, the block will be released to the * rgrp since it won't be allocated during the transaction */ /* +3 in the end for unstuffing block, inode size update block * and another block in case quota straddles page boundary and * two blocks need to be updated instead of 1 */ blocks = num_qd * data_blocks + RES_DINODE + num_qd + 3; reserved = 1 + (nalloc * (data_blocks + ind_blocks)); ap.target = reserved; error = gfs2_inplace_reserve(ip, &ap); if (error) goto out_alloc; if (nalloc) blocks += gfs2_rg_blocks(ip, reserved) + nalloc * ind_blocks + RES_STATFS; error = gfs2_trans_begin(sdp, blocks, 0); if (error) goto out_ipres; for (x = 0; x < num_qd; x++) { qd = qda[x]; offset = qd2offset(qd); error = gfs2_adjust_quota(ip, offset, qd->qd_change_sync, qd, NULL); if (error) goto out_end_trans; do_qc(qd, -qd->qd_change_sync); set_bit(QDF_REFRESH, &qd->qd_flags); } error = 0; out_end_trans: gfs2_trans_end(sdp); out_ipres: gfs2_inplace_release(ip); out_alloc: gfs2_glock_dq_uninit(&i_gh); out: while (qx--) gfs2_glock_dq_uninit(&ghs[qx]); mutex_unlock(&ip->i_inode.i_mutex); kfree(ghs); gfs2_log_flush(ip->i_gl->gl_sbd, ip->i_gl); return error; } static int update_qd(struct gfs2_sbd *sdp, struct gfs2_quota_data *qd) { struct gfs2_inode *ip = GFS2_I(sdp->sd_quota_inode); struct gfs2_quota q; struct gfs2_quota_lvb *qlvb; loff_t pos; int error; memset(&q, 0, sizeof(struct gfs2_quota)); pos = qd2offset(qd); error = gfs2_internal_read(ip, (char *)&q, &pos, sizeof(q)); if (error < 0) return error; qlvb = (struct gfs2_quota_lvb *)qd->qd_gl->gl_lksb.sb_lvbptr; qlvb->qb_magic = cpu_to_be32(GFS2_MAGIC); qlvb->__pad = 0; qlvb->qb_limit = q.qu_limit; qlvb->qb_warn = q.qu_warn; qlvb->qb_value = q.qu_value; qd->qd_qb = *qlvb; return 0; } static int do_glock(struct gfs2_quota_data *qd, int force_refresh, struct gfs2_holder *q_gh) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; struct gfs2_inode *ip = GFS2_I(sdp->sd_quota_inode); struct gfs2_holder i_gh; int error; restart: error = gfs2_glock_nq_init(qd->qd_gl, LM_ST_SHARED, 0, q_gh); if (error) return error; 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)) { gfs2_glock_dq_uninit(q_gh); error = gfs2_glock_nq_init(qd->qd_gl, LM_ST_EXCLUSIVE, GL_NOCACHE, q_gh); if (error) return error; error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &i_gh); if (error) goto fail; error = update_qd(sdp, qd); if (error) goto fail_gunlock; gfs2_glock_dq_uninit(&i_gh); gfs2_glock_dq_uninit(q_gh); force_refresh = 0; goto restart; } return 0; fail_gunlock: gfs2_glock_dq_uninit(&i_gh); fail: gfs2_glock_dq_uninit(q_gh); return error; } int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data *qd; unsigned int x; int error = 0; error = gfs2_quota_hold(ip, uid, gid); if (error) return error; if (capable(CAP_SYS_RESOURCE) || sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; sort(ip->i_res->rs_qa_qd, ip->i_res->rs_qa_qd_num, 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]); if (error) break; } if (!error) set_bit(GIF_QD_LOCKED, &ip->i_flags); else { while (x--) gfs2_glock_dq_uninit(&ip->i_res->rs_qa_qd_ghs[x]); gfs2_quota_unhold(ip); } return error; } static int need_sync(struct gfs2_quota_data *qd) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; struct gfs2_tune *gt = &sdp->sd_tune; s64 value; unsigned int num, den; int do_sync = 1; if (!qd->qd_qb.qb_limit) return 0; spin_lock(&qd_lock); value = qd->qd_change; spin_unlock(&qd_lock); spin_lock(>->gt_spin); num = gt->gt_quota_scale_num; den = gt->gt_quota_scale_den; spin_unlock(>->gt_spin); if (value < 0) do_sync = 0; else if ((s64)be64_to_cpu(qd->qd_qb.qb_value) >= (s64)be64_to_cpu(qd->qd_qb.qb_limit)) do_sync = 0; else { value *= gfs2_jindex_size(sdp) * num; value = div_s64(value, den); value += (s64)be64_to_cpu(qd->qd_qb.qb_value); if (value < (s64)be64_to_cpu(qd->qd_qb.qb_limit)) do_sync = 0; } return do_sync; } void gfs2_quota_unlock(struct gfs2_inode *ip) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data *qda[4]; unsigned int count = 0; unsigned int x; int found; if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags)) goto out; for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) { struct gfs2_quota_data *qd; int sync; qd = ip->i_res->rs_qa_qd[x]; sync = need_sync(qd); gfs2_glock_dq_uninit(&ip->i_res->rs_qa_qd_ghs[x]); if (!sync) continue; spin_lock(&qd_lock); found = qd_check_sync(sdp, qd, NULL); spin_unlock(&qd_lock); if (!found) continue; gfs2_assert_warn(sdp, qd->qd_change_sync); if (bh_get(qd)) { clear_bit(QDF_LOCKED, &qd->qd_flags); slot_put(qd); qd_put(qd); continue; } qda[count++] = qd; } if (count) { do_sync(count, qda); for (x = 0; x < count; x++) qd_unlock(qda[x]); } out: gfs2_quota_unhold(ip); } #define MAX_LINE 256 static int print_message(struct gfs2_quota_data *qd, char *type) { struct gfs2_sbd *sdp = qd->qd_gl->gl_sbd; printk(KERN_INFO "GFS2: fsid=%s: quota %s for %s %u\n", sdp->sd_fsname, type, (qd->qd_id.type == USRQUOTA) ? "user" : "group", from_kqid(&init_user_ns, qd->qd_id)); return 0; } int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_quota_data *qd; s64 value; unsigned int x; int error = 0; if (!test_bit(GIF_QD_LOCKED, &ip->i_flags)) return 0; if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) { qd = ip->i_res->rs_qa_qd[x]; if (!(qid_eq(qd->qd_id, make_kqid_uid(uid)) || qid_eq(qd->qd_id, make_kqid_gid(gid)))) continue; value = (s64)be64_to_cpu(qd->qd_qb.qb_value); spin_lock(&qd_lock); 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 && time_after_eq(jiffies, qd->qd_last_warn + 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; } void gfs2_quota_change(struct gfs2_inode *ip, s64 change, kuid_t uid, kgid_t gid) { struct gfs2_quota_data *qd; unsigned int x; if (gfs2_assert_warn(GFS2_SB(&ip->i_inode), change)) return; if (ip->i_diskflags & GFS2_DIF_SYSTEM) return; for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) { qd = ip->i_res->rs_qa_qd[x]; if (qid_eq(qd->qd_id, make_kqid_uid(uid)) || qid_eq(qd->qd_id, make_kqid_gid(gid))) { do_qc(qd, change); } } } int gfs2_quota_sync(struct super_block *sb, int type) { struct gfs2_sbd *sdp = sb->s_fs_info; struct gfs2_quota_data **qda; unsigned int max_qd = PAGE_SIZE/sizeof(struct gfs2_holder); unsigned int num_qd; unsigned int x; int error = 0; qda = kcalloc(max_qd, sizeof(struct gfs2_quota_data *), GFP_KERNEL); if (!qda) return -ENOMEM; mutex_lock(&sdp->sd_quota_sync_mutex); sdp->sd_quota_sync_gen++; do { num_qd = 0; for (;;) { error = qd_fish(sdp, qda + num_qd); if (error || !qda[num_qd]) break; if (++num_qd == max_qd) break; } if (num_qd) { if (!error) error = do_sync(num_qd, qda); if (!error) for (x = 0; x < num_qd; x++) qda[x]->qd_sync_gen = sdp->sd_quota_sync_gen; for (x = 0; x < num_qd; x++) qd_unlock(qda[x]); } } while (!error && num_qd == max_qd); mutex_unlock(&sdp->sd_quota_sync_mutex); kfree(qda); return error; } int gfs2_quota_refresh(struct gfs2_sbd *sdp, struct kqid qid) { struct gfs2_quota_data *qd; struct gfs2_holder q_gh; int error; error = qd_get(sdp, qid, &qd); if (error) return error; error = do_glock(qd, FORCE, &q_gh); if (!error) gfs2_glock_dq_uninit(&q_gh); qd_put(qd); return error; } static void gfs2_quota_change_in(struct gfs2_quota_change_host *qc, const void *buf) { const struct gfs2_quota_change *str = buf; qc->qc_change = be64_to_cpu(str->qc_change); qc->qc_flags = be32_to_cpu(str->qc_flags); qc->qc_id = make_kqid(&init_user_ns, (qc->qc_flags & GFS2_QCF_USER)?USRQUOTA:GRPQUOTA, be32_to_cpu(str->qc_id)); } int gfs2_quota_init(struct gfs2_sbd *sdp) { struct gfs2_inode *ip = GFS2_I(sdp->sd_qc_inode); u64 size = i_size_read(sdp->sd_qc_inode); unsigned int blocks = size >> sdp->sd_sb.sb_bsize_shift; unsigned int x, slot = 0; unsigned int found = 0; u64 dblock; u32 extlen = 0; int error; if (gfs2_check_internal_file_size(sdp->sd_qc_inode, 1, 64 << 20)) return -EIO; sdp->sd_quota_slots = blocks * sdp->sd_qc_per_block; sdp->sd_quota_chunks = DIV_ROUND_UP(sdp->sd_quota_slots, 8 * PAGE_SIZE); error = -ENOMEM; sdp->sd_quota_bitmap = kcalloc(sdp->sd_quota_chunks, sizeof(unsigned char *), GFP_NOFS); if (!sdp->sd_quota_bitmap) return error; for (x = 0; x < sdp->sd_quota_chunks; x++) { sdp->sd_quota_bitmap[x] = kzalloc(PAGE_SIZE, GFP_NOFS); if (!sdp->sd_quota_bitmap[x]) goto fail; } for (x = 0; x < blocks; x++) { struct buffer_head *bh; unsigned int y; if (!extlen) { int new = 0; error = gfs2_extent_map(&ip->i_inode, x, &new, &dblock, &extlen); if (error) goto fail; } error = -EIO; bh = gfs2_meta_ra(ip->i_gl, dblock, extlen); if (!bh) goto fail; if (gfs2_metatype_check(sdp, bh, GFS2_METATYPE_QC)) { brelse(bh); goto fail; } for (y = 0; y < sdp->sd_qc_per_block && slot < sdp->sd_quota_slots; y++, slot++) { struct gfs2_quota_change_host qc; struct gfs2_quota_data *qd; gfs2_quota_change_in(&qc, bh->b_data + sizeof(struct gfs2_meta_header) + y * sizeof(struct gfs2_quota_change)); if (!qc.qc_change) continue; error = qd_alloc(sdp, qc.qc_id, &qd); if (error) { brelse(bh); goto fail; } set_bit(QDF_CHANGE, &qd->qd_flags); qd->qd_change = qc.qc_change; qd->qd_slot = slot; qd->qd_slot_count = 1; spin_lock(&qd_lock); gfs2_icbit_munge(sdp, sdp->sd_quota_bitmap, slot, 1); list_add(&qd->qd_list, &sdp->sd_quota_list); atomic_inc(&sdp->sd_quota_count); spin_unlock(&qd_lock); found++; } brelse(bh); dblock++; extlen--; } if (found) fs_info(sdp, "found %u quota changes\n", found); return 0; fail: gfs2_quota_cleanup(sdp); return error; } void gfs2_quota_cleanup(struct gfs2_sbd *sdp) { struct list_head *head = &sdp->sd_quota_list; struct gfs2_quota_data *qd; unsigned int x; spin_lock(&qd_lock); while (!list_empty(head)) { qd = list_entry(head->prev, struct gfs2_quota_data, qd_list); /* * To be removed in due course... we should be able to * ensure that all refs to the qd have done by this point * so that this rather odd test is not required */ spin_lock(&qd->qd_lockref.lock); if (qd->qd_lockref.count > 1 || (qd->qd_lockref.count && !test_bit(QDF_CHANGE, &qd->qd_flags))) { spin_unlock(&qd->qd_lockref.lock); list_move(&qd->qd_list, head); spin_unlock(&qd_lock); schedule(); spin_lock(&qd_lock); continue; } spin_unlock(&qd->qd_lockref.lock); list_del(&qd->qd_list); /* Also remove if this qd exists in the reclaim list */ if (!list_empty(&qd->qd_reclaim)) { list_del_init(&qd->qd_reclaim); atomic_dec(&qd_lru_count); } atomic_dec(&sdp->sd_quota_count); spin_unlock(&qd_lock); if (!qd->qd_lockref.count) { gfs2_assert_warn(sdp, !qd->qd_change); gfs2_assert_warn(sdp, !qd->qd_slot_count); } else gfs2_assert_warn(sdp, qd->qd_slot_count == 1); gfs2_assert_warn(sdp, !qd->qd_bh_count); gfs2_glock_put(qd->qd_gl); kmem_cache_free(gfs2_quotad_cachep, qd); spin_lock(&qd_lock); } spin_unlock(&qd_lock); gfs2_assert_warn(sdp, !atomic_read(&sdp->sd_quota_count)); if (sdp->sd_quota_bitmap) { for (x = 0; x < sdp->sd_quota_chunks; x++) kfree(sdp->sd_quota_bitmap[x]); kfree(sdp->sd_quota_bitmap); } } static void quotad_error(struct gfs2_sbd *sdp, const char *msg, int error) { if (error == 0 || error == -EROFS) return; if (!test_bit(SDF_SHUTDOWN, &sdp->sd_flags)) fs_err(sdp, "gfs2_quotad: %s error %d\n", msg, error); } static void quotad_check_timeo(struct gfs2_sbd *sdp, const char *msg, int (*fxn)(struct super_block *sb, int type), unsigned long t, unsigned long *timeo, unsigned int *new_timeo) { if (t >= *timeo) { int error = fxn(sdp->sd_vfs, 0); quotad_error(sdp, msg, error); *timeo = gfs2_tune_get_i(&sdp->sd_tune, new_timeo) * HZ; } else { *timeo -= t; } } static void quotad_check_trunc_list(struct gfs2_sbd *sdp) { struct gfs2_inode *ip; while(1) { ip = NULL; spin_lock(&sdp->sd_trunc_lock); if (!list_empty(&sdp->sd_trunc_list)) { ip = list_entry(sdp->sd_trunc_list.next, struct gfs2_inode, i_trunc_list); list_del_init(&ip->i_trunc_list); } spin_unlock(&sdp->sd_trunc_lock); if (ip == NULL) return; gfs2_glock_finish_truncate(ip); } } void gfs2_wake_up_statfs(struct gfs2_sbd *sdp) { if (!sdp->sd_statfs_force_sync) { sdp->sd_statfs_force_sync = 1; wake_up(&sdp->sd_quota_wait); } } /** * gfs2_quotad - Write cached quota changes into the quota file * @sdp: Pointer to GFS2 superblock * */ int gfs2_quotad(void *data) { struct gfs2_sbd *sdp = data; struct gfs2_tune *tune = &sdp->sd_tune; unsigned long statfs_timeo = 0; unsigned long quotad_timeo = 0; unsigned long t = 0; DEFINE_WAIT(wait); int empty; while (!kthread_should_stop()) { /* Update the master statfs file */ if (sdp->sd_statfs_force_sync) { int error = gfs2_statfs_sync(sdp->sd_vfs, 0); quotad_error(sdp, "statfs", error); statfs_timeo = gfs2_tune_get(sdp, gt_statfs_quantum) * HZ; } else quotad_check_timeo(sdp, "statfs", gfs2_statfs_sync, t, &statfs_timeo, &tune->gt_statfs_quantum); /* Update quota file */ quotad_check_timeo(sdp, "sync", gfs2_quota_sync, t, "ad_timeo, &tune->gt_quota_quantum); /* Check for & recover partially truncated inodes */ quotad_check_trunc_list(sdp); try_to_freeze(); t = min(quotad_timeo, statfs_timeo); prepare_to_wait(&sdp->sd_quota_wait, &wait, TASK_INTERRUPTIBLE); spin_lock(&sdp->sd_trunc_lock); empty = list_empty(&sdp->sd_trunc_list); spin_unlock(&sdp->sd_trunc_lock); if (empty && !sdp->sd_statfs_force_sync) t -= schedule_timeout(t); else t = 0; finish_wait(&sdp->sd_quota_wait, &wait); } return 0; } static int gfs2_quota_get_xstate(struct super_block *sb, struct fs_quota_stat *fqs) { struct gfs2_sbd *sdp = sb->s_fs_info; memset(fqs, 0, sizeof(struct fs_quota_stat)); fqs->qs_version = FS_QSTAT_VERSION; switch (sdp->sd_args.ar_quota) { case GFS2_QUOTA_ON: fqs->qs_flags |= (FS_QUOTA_UDQ_ENFD | FS_QUOTA_GDQ_ENFD); /*FALLTHRU*/ case GFS2_QUOTA_ACCOUNT: fqs->qs_flags |= (FS_QUOTA_UDQ_ACCT | FS_QUOTA_GDQ_ACCT); break; case GFS2_QUOTA_OFF: break; } if (sdp->sd_quota_inode) { fqs->qs_uquota.qfs_ino = GFS2_I(sdp->sd_quota_inode)->i_no_addr; fqs->qs_uquota.qfs_nblks = sdp->sd_quota_inode->i_blocks; } fqs->qs_uquota.qfs_nextents = 1; /* unsupported */ fqs->qs_gquota = fqs->qs_uquota; /* its the same inode in both cases */ fqs->qs_incoredqs = atomic_read(&qd_lru_count); return 0; } static int gfs2_get_dqblk(struct super_block *sb, struct kqid qid, struct fs_disk_quota *fdq) { struct gfs2_sbd *sdp = sb->s_fs_info; struct gfs2_quota_lvb *qlvb; struct gfs2_quota_data *qd; struct gfs2_holder q_gh; int error; memset(fdq, 0, sizeof(struct fs_disk_quota)); if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return -ESRCH; /* Crazy XFS error code */ if ((qid.type != USRQUOTA) && (qid.type != GRPQUOTA)) return -EINVAL; error = qd_get(sdp, qid, &qd); if (error) return error; error = do_glock(qd, FORCE, &q_gh); if (error) goto out; qlvb = (struct gfs2_quota_lvb *)qd->qd_gl->gl_lksb.sb_lvbptr; fdq->d_version = FS_DQUOT_VERSION; fdq->d_flags = (qid.type == USRQUOTA) ? FS_USER_QUOTA : FS_GROUP_QUOTA; fdq->d_id = from_kqid_munged(current_user_ns(), qid); fdq->d_blk_hardlimit = be64_to_cpu(qlvb->qb_limit) << sdp->sd_fsb2bb_shift; fdq->d_blk_softlimit = be64_to_cpu(qlvb->qb_warn) << sdp->sd_fsb2bb_shift; fdq->d_bcount = be64_to_cpu(qlvb->qb_value) << sdp->sd_fsb2bb_shift; gfs2_glock_dq_uninit(&q_gh); out: qd_put(qd); return error; } /* GFS2 only supports a subset of the XFS fields */ #define GFS2_FIELDMASK (FS_DQ_BSOFT|FS_DQ_BHARD|FS_DQ_BCOUNT) static int gfs2_set_dqblk(struct super_block *sb, struct kqid qid, struct fs_disk_quota *fdq) { struct gfs2_sbd *sdp = sb->s_fs_info; struct gfs2_inode *ip = GFS2_I(sdp->sd_quota_inode); struct gfs2_quota_data *qd; struct gfs2_holder q_gh, i_gh; unsigned int data_blocks, ind_blocks; unsigned int blocks = 0; int alloc_required; loff_t offset; int error; if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return -ESRCH; /* Crazy XFS error code */ if ((qid.type != USRQUOTA) && (qid.type != GRPQUOTA)) return -EINVAL; if (fdq->d_fieldmask & ~GFS2_FIELDMASK) return -EINVAL; error = qd_get(sdp, qid, &qd); if (error) return error; error = gfs2_rs_alloc(ip); if (error) goto out_put; mutex_lock(&ip->i_inode.i_mutex); error = gfs2_glock_nq_init(qd->qd_gl, LM_ST_EXCLUSIVE, 0, &q_gh); if (error) goto out_unlockput; error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &i_gh); if (error) goto out_q; /* Check for existing entry, if none then alloc new blocks */ error = update_qd(sdp, qd); if (error) goto out_i; /* If nothing has changed, this is a no-op */ if ((fdq->d_fieldmask & FS_DQ_BSOFT) && ((fdq->d_blk_softlimit >> sdp->sd_fsb2bb_shift) == be64_to_cpu(qd->qd_qb.qb_warn))) fdq->d_fieldmask ^= FS_DQ_BSOFT; if ((fdq->d_fieldmask & FS_DQ_BHARD) && ((fdq->d_blk_hardlimit >> sdp->sd_fsb2bb_shift) == be64_to_cpu(qd->qd_qb.qb_limit))) fdq->d_fieldmask ^= FS_DQ_BHARD; if ((fdq->d_fieldmask & FS_DQ_BCOUNT) && ((fdq->d_bcount >> sdp->sd_fsb2bb_shift) == be64_to_cpu(qd->qd_qb.qb_value))) fdq->d_fieldmask ^= FS_DQ_BCOUNT; if (fdq->d_fieldmask == 0) goto out_i; offset = qd2offset(qd); alloc_required = gfs2_write_alloc_required(ip, offset, sizeof(struct gfs2_quota)); if (gfs2_is_stuffed(ip)) alloc_required = 1; if (alloc_required) { struct gfs2_alloc_parms ap = { .aflags = 0, }; gfs2_write_calc_reserv(ip, sizeof(struct gfs2_quota), &data_blocks, &ind_blocks); blocks = 1 + data_blocks + ind_blocks; ap.target = blocks; error = gfs2_inplace_reserve(ip, &ap); if (error) goto out_i; blocks += gfs2_rg_blocks(ip, blocks); } /* Some quotas span block boundaries and can update two blocks, adding an extra block to the transaction to handle such quotas */ error = gfs2_trans_begin(sdp, blocks + RES_DINODE + 2, 0); if (error) goto out_release; /* Apply changes */ error = gfs2_adjust_quota(ip, offset, 0, qd, fdq); gfs2_trans_end(sdp); out_release: if (alloc_required) gfs2_inplace_release(ip); out_i: gfs2_glock_dq_uninit(&i_gh); out_q: gfs2_glock_dq_uninit(&q_gh); out_unlockput: mutex_unlock(&ip->i_inode.i_mutex); out_put: qd_put(qd); return error; } const struct quotactl_ops gfs2_quotactl_ops = { .quota_sync = gfs2_quota_sync, .get_xstate = gfs2_quota_get_xstate, .get_dqblk = gfs2_get_dqblk, .set_dqblk = gfs2_set_dqblk, };