This updates all relevant function definitions and callers to get rid of the global g_sbi, making it suitable for external use in liberofs.
Signed-off-by: Hongzhen Luo <hongz...@linux.alibaba.com> --- include/erofs/cache.h | 42 ++++++---- include/erofs/internal.h | 5 +- lib/blobchunk.c | 20 ++--- lib/cache.c | 162 +++++++++++++++++++++------------------ lib/compress.c | 40 +++++----- lib/inode.c | 43 ++++++----- lib/super.c | 21 ++--- lib/xattr.c | 10 +-- mkfs/main.c | 15 ++-- 9 files changed, 198 insertions(+), 160 deletions(-) diff --git a/include/erofs/cache.h b/include/erofs/cache.h index 288843e..16148ea 100644 --- a/include/erofs/cache.h +++ b/include/erofs/cache.h @@ -53,10 +53,20 @@ struct erofs_buffer_block { struct erofs_buffer_head buffers; }; -static inline const int get_alignsize(int type, int *type_ret) -{ - struct erofs_sb_info *sbi = &g_sbi; +struct erofs_buffer_manager { + struct erofs_buffer_block blkh; + erofs_blk_t tail_blkaddr, erofs_metablkcnt; + + /* buckets for all mapped buffer blocks to boost up allocation */ + struct list_head mapped_buckets[META + 1][EROFS_MAX_BLOCK_SIZE]; + /* last mapped buffer block to accelerate erofs_mapbh() */ + struct erofs_buffer_block *last_mapped_block; +}; + +static inline const int get_alignsize(struct erofs_sb_info *sbi, int type, + int *type_ret) +{ if (type == DATA) return erofs_blksiz(sbi); @@ -82,10 +92,10 @@ static inline const int get_alignsize(int type, int *type_ret) extern const struct erofs_bhops erofs_drop_directly_bhops; extern const struct erofs_bhops erofs_skip_write_bhops; -static inline erofs_off_t erofs_btell(struct erofs_buffer_head *bh, bool end) +static inline erofs_off_t erofs_btell(struct erofs_sb_info *sbi, + struct erofs_buffer_head *bh, bool end) { const struct erofs_buffer_block *bb = bh->block; - struct erofs_sb_info *sbi = &g_sbi; if (bb->blkaddr == NULL_ADDR) return NULL_ADDR_UL; @@ -101,20 +111,26 @@ static inline int erofs_bh_flush_generic_end(struct erofs_buffer_head *bh) return 0; } -void erofs_buffer_init(erofs_blk_t startblk); -int erofs_bh_balloon(struct erofs_buffer_head *bh, erofs_off_t incr); +void erofs_buffer_init(struct erofs_sb_info *sbi, erofs_blk_t startblk); +int erofs_bh_balloon(struct erofs_sb_info *sbi, struct erofs_buffer_head *bh, + erofs_off_t incr); -struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, +struct erofs_buffer_head *erofs_balloc(struct erofs_sb_info *sbi, + int type, erofs_off_t size, unsigned int required_ext, unsigned int inline_ext); -struct erofs_buffer_head *erofs_battach(struct erofs_buffer_head *bh, +struct erofs_buffer_head *erofs_battach(struct erofs_sb_info *sbi, + struct erofs_buffer_head *bh, int type, unsigned int size); -erofs_blk_t erofs_mapbh(struct erofs_buffer_block *bb); -int erofs_bflush(struct erofs_buffer_block *bb); +erofs_blk_t erofs_mapbh(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb); +int erofs_bflush(struct erofs_sb_info *sbi, struct erofs_buffer_block *bb); -void erofs_bdrop(struct erofs_buffer_head *bh, bool tryrevoke); -erofs_blk_t erofs_total_metablocks(void); +void erofs_bdrop(struct erofs_sb_info *sbi, struct erofs_buffer_head *bh, + bool tryrevoke); +erofs_blk_t erofs_total_metablocks(struct erofs_sb_info *sbi); +void erofs_buffer_exit(struct erofs_sb_info *sbi); #ifdef __cplusplus } diff --git a/include/erofs/internal.h b/include/erofs/internal.h index 5f70570..bad9f13 100644 --- a/include/erofs/internal.h +++ b/include/erofs/internal.h @@ -58,6 +58,7 @@ extern struct erofs_sb_info g_sbi; (roundup(addr, erofs_blksiz(sbi)) >> (sbi)->blkszbits) struct erofs_buffer_head; +struct erofs_buffer_manager; struct erofs_device_info { u8 tag[64]; @@ -132,6 +133,8 @@ struct erofs_sb_info { struct erofs_mkfs_dfops *mkfs_dfops; #endif bool useqpl; + + struct erofs_buffer_manager *bmngr; }; #define EROFS_SUPER_END (EROFS_SUPER_OFFSET + sizeof(struct erofs_super_block)) @@ -402,7 +405,7 @@ int erofs_read_superblock(struct erofs_sb_info *sbi); void erofs_put_super(struct erofs_sb_info *sbi); int erofs_writesb(struct erofs_sb_info *sbi, struct erofs_buffer_head *sb_bh, erofs_blk_t *blocks); -struct erofs_buffer_head *erofs_reserve_sb(void); +struct erofs_buffer_head *erofs_reserve_sb(struct erofs_sb_info *sbi); int erofs_enable_sb_chksum(struct erofs_sb_info *sbi, u32 *crc); /* namei.c */ diff --git a/lib/blobchunk.c b/lib/blobchunk.c index 9af223d..1169288 100644 --- a/lib/blobchunk.c +++ b/lib/blobchunk.c @@ -496,8 +496,8 @@ int erofs_mkfs_dump_blobs(struct erofs_sb_info *sbi) unsigned int i, ret; erofs_blk_t nblocks; - nblocks = erofs_mapbh(NULL); - pos_out = erofs_btell(bh_devt, false); + nblocks = erofs_mapbh(sbi, NULL); + pos_out = erofs_btell(sbi, bh_devt, false); i = 0; do { struct erofs_deviceslot dis = { @@ -513,17 +513,17 @@ int erofs_mkfs_dump_blobs(struct erofs_sb_info *sbi) nblocks += sbi->devs[i].blocks; } while (++i < sbi->extra_devices); bh_devt->op = &erofs_drop_directly_bhops; - erofs_bdrop(bh_devt, false); + erofs_bdrop(sbi, bh_devt, false); return 0; } - bh = erofs_balloc(DATA, datablob_size, 0, 0); + bh = erofs_balloc(sbi, DATA, datablob_size, 0, 0); if (IS_ERR(bh)) return PTR_ERR(bh); - erofs_mapbh(bh->block); + erofs_mapbh(sbi, bh->block); - pos_out = erofs_btell(bh, false); + pos_out = erofs_btell(sbi, bh, false); remapped_base = erofs_blknr(sbi, pos_out); pos_out += sbi->bdev.offset; if (blobfile) { @@ -535,7 +535,7 @@ int erofs_mkfs_dump_blobs(struct erofs_sb_info *sbi) ret = erofs_io_ftruncate(&sbi->bdev, pos_out + datablob_size); } bh->op = &erofs_drop_directly_bhops; - erofs_bdrop(bh, false); + erofs_bdrop(sbi, bh, false); return ret; } @@ -626,15 +626,15 @@ int erofs_mkfs_init_devices(struct erofs_sb_info *sbi, unsigned int devices) if (!sbi->devs) return -ENOMEM; - bh_devt = erofs_balloc(DEVT, + bh_devt = erofs_balloc(sbi, DEVT, sizeof(struct erofs_deviceslot) * devices, 0, 0); if (IS_ERR(bh_devt)) { free(sbi->devs); return PTR_ERR(bh_devt); } - erofs_mapbh(bh_devt->block); + erofs_mapbh(sbi, bh_devt->block); bh_devt->op = &erofs_skip_write_bhops; - sbi->devt_slotoff = erofs_btell(bh_devt, false) / EROFS_DEVT_SLOT_SIZE; + sbi->devt_slotoff = erofs_btell(sbi, bh_devt, false) / EROFS_DEVT_SLOT_SIZE; sbi->extra_devices = devices; erofs_sb_set_device_table(sbi); return 0; diff --git a/lib/cache.c b/lib/cache.c index 69a3055..d0afc92 100644 --- a/lib/cache.c +++ b/lib/cache.c @@ -9,17 +9,6 @@ #include <erofs/cache.h> #include "erofs/print.h" -static struct erofs_buffer_block blkh = { - .list = LIST_HEAD_INIT(blkh.list), - .blkaddr = NULL_ADDR, -}; -static erofs_blk_t tail_blkaddr, erofs_metablkcnt; - -/* buckets for all mapped buffer blocks to boost up allocation */ -static struct list_head mapped_buckets[META + 1][EROFS_MAX_BLOCK_SIZE]; -/* last mapped buffer block to accelerate erofs_mapbh() */ -static struct erofs_buffer_block *last_mapped_block = &blkh; - static int erofs_bh_flush_drop_directly(struct erofs_buffer_head *bh) { return erofs_bh_flush_generic_end(bh); @@ -38,39 +27,48 @@ const struct erofs_bhops erofs_skip_write_bhops = { .flush = erofs_bh_flush_skip_write, }; -void erofs_buffer_init(erofs_blk_t startblk) +void erofs_buffer_init(struct erofs_sb_info *sbi, erofs_blk_t startblk) { int i, j; - for (i = 0; i < ARRAY_SIZE(mapped_buckets); i++) - for (j = 0; j < ARRAY_SIZE(mapped_buckets[0]); j++) - init_list_head(&mapped_buckets[i][j]); - tail_blkaddr = startblk; + sbi->bmngr = malloc(sizeof(struct erofs_buffer_manager)); + if (!sbi->bmngr) { + erofs_err("fail to prepare buffer manager"); + return; + } + init_list_head(&sbi->bmngr->blkh.list); + sbi->bmngr->blkh.blkaddr = NULL_ADDR; + sbi->bmngr->last_mapped_block = &sbi->bmngr->blkh; + + for (i = 0; i < ARRAY_SIZE(sbi->bmngr->mapped_buckets); i++) + for (j = 0; j < ARRAY_SIZE(sbi->bmngr->mapped_buckets[0]); j++) + init_list_head(&sbi->bmngr->mapped_buckets[i][j]); + sbi->bmngr->tail_blkaddr = startblk; } -static void erofs_bupdate_mapped(struct erofs_buffer_block *bb) +static void erofs_bupdate_mapped(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb) { struct list_head *bkt; - struct erofs_sb_info *sbi = &g_sbi; if (bb->blkaddr == NULL_ADDR) return; - bkt = mapped_buckets[bb->type] + + bkt = sbi->bmngr->mapped_buckets[bb->type] + (bb->buffers.off & (erofs_blksiz(sbi) - 1)); list_del(&bb->mapped_list); list_add_tail(&bb->mapped_list, bkt); } /* return occupied bytes in specific buffer block if succeed */ -static int __erofs_battach(struct erofs_buffer_block *bb, +static int __erofs_battach(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb, struct erofs_buffer_head *bh, erofs_off_t incr, unsigned int alignsize, unsigned int extrasize, bool dryrun) { - struct erofs_sb_info *sbi = &g_sbi; const unsigned int blksiz = erofs_blksiz(sbi); const unsigned int blkmask = blksiz - 1; erofs_off_t boff = bb->buffers.off; @@ -87,7 +85,7 @@ static int __erofs_battach(struct erofs_buffer_block *bb, blkaddr = bb->blkaddr; if (blkaddr != NULL_ADDR) { - tailupdate = (tail_blkaddr == blkaddr + + tailupdate = (sbi->bmngr->tail_blkaddr == blkaddr + BLK_ROUND_UP(sbi, boff)); if (oob && !tailupdate) return -EINVAL; @@ -104,13 +102,15 @@ static int __erofs_battach(struct erofs_buffer_block *bb, bb->buffers.off = boff; /* need to update the tail_blkaddr */ if (tailupdate) - tail_blkaddr = blkaddr + BLK_ROUND_UP(sbi, boff); - erofs_bupdate_mapped(bb); + sbi->bmngr->tail_blkaddr = blkaddr + + BLK_ROUND_UP(sbi, boff); + erofs_bupdate_mapped(sbi, bb); } return ((alignedoffset + incr - 1) & blkmask) + 1; } -int erofs_bh_balloon(struct erofs_buffer_head *bh, erofs_off_t incr) +int erofs_bh_balloon(struct erofs_sb_info *sbi, struct erofs_buffer_head *bh, + erofs_off_t incr) { struct erofs_buffer_block *const bb = bh->block; @@ -118,16 +118,16 @@ int erofs_bh_balloon(struct erofs_buffer_head *bh, erofs_off_t incr) if (bh->list.next != &bb->buffers.list) return -EINVAL; - return __erofs_battach(bb, NULL, incr, 1, 0, false); + return __erofs_battach(sbi, bb, NULL, incr, 1, 0, false); } -static int erofs_bfind_for_attach(int type, erofs_off_t size, +static int erofs_bfind_for_attach(struct erofs_sb_info *sbi, + int type, erofs_off_t size, unsigned int required_ext, unsigned int inline_ext, unsigned int alignsize, struct erofs_buffer_block **bbp) { - struct erofs_sb_info *sbi = &g_sbi; const unsigned int blksiz = erofs_blksiz(sbi); struct erofs_buffer_block *cur, *bb; unsigned int used0, used_before, usedmax, used; @@ -153,7 +153,8 @@ static int erofs_bfind_for_attach(int type, erofs_off_t size, used_before = rounddown(blksiz - (size + required_ext + inline_ext), alignsize); for (; used_before; --used_before) { - struct list_head *bt = mapped_buckets[type] + used_before; + struct list_head *bt = sbi->bmngr->mapped_buckets[type] + + used_before; if (list_empty(bt)) continue; @@ -162,7 +163,7 @@ static int erofs_bfind_for_attach(int type, erofs_off_t size, /* last mapped block can be expended, don't handle it here */ if (list_next_entry(cur, list)->blkaddr == NULL_ADDR) { - DBG_BUGON(cur != last_mapped_block); + DBG_BUGON(cur != sbi->bmngr->last_mapped_block); continue; } @@ -170,7 +171,7 @@ static int erofs_bfind_for_attach(int type, erofs_off_t size, DBG_BUGON(cur->blkaddr == NULL_ADDR); DBG_BUGON(used_before != (cur->buffers.off & (blksiz - 1))); - ret = __erofs_battach(cur, NULL, size, alignsize, + ret = __erofs_battach(sbi, cur, NULL, size, alignsize, required_ext + inline_ext, true); if (ret < 0) { DBG_BUGON(1); @@ -188,10 +189,10 @@ static int erofs_bfind_for_attach(int type, erofs_off_t size, skip_mapped: /* try to start from the last mapped one, which can be expended */ - cur = last_mapped_block; - if (cur == &blkh) + cur = sbi->bmngr->last_mapped_block; + if (cur == &sbi->bmngr->blkh) cur = list_next_entry(cur, list); - for (; cur != &blkh; cur = list_next_entry(cur, list)) { + for (; cur != &sbi->bmngr->blkh; cur = list_next_entry(cur, list)) { used_before = cur->buffers.off & (blksiz - 1); /* skip if buffer block is just full */ @@ -202,7 +203,7 @@ skip_mapped: if (cur->type != type) continue; - ret = __erofs_battach(cur, NULL, size, alignsize, + ret = __erofs_battach(sbi, cur, NULL, size, alignsize, required_ext + inline_ext, true); if (ret < 0) continue; @@ -229,7 +230,8 @@ skip_mapped: return 0; } -struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, +struct erofs_buffer_head *erofs_balloc(struct erofs_sb_info *sbi, + int type, erofs_off_t size, unsigned int required_ext, unsigned int inline_ext) { @@ -237,7 +239,7 @@ struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, struct erofs_buffer_head *bh; unsigned int alignsize; - int ret = get_alignsize(type, &type); + int ret = get_alignsize(sbi, type, &type); if (ret < 0) return ERR_PTR(ret); @@ -246,7 +248,7 @@ struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, alignsize = ret; /* try to find if we could reuse an allocated buffer block */ - ret = erofs_bfind_for_attach(type, size, required_ext, inline_ext, + ret = erofs_bfind_for_attach(sbi, type, size, required_ext, inline_ext, alignsize, &bb); if (ret) return ERR_PTR(ret); @@ -266,9 +268,10 @@ struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, bb->buffers.off = 0; init_list_head(&bb->buffers.list); if (type == DATA) - list_add(&bb->list, &last_mapped_block->list); + list_add(&bb->list, + &sbi->bmngr->last_mapped_block->list); else - list_add_tail(&bb->list, &blkh.list); + list_add_tail(&bb->list, &sbi->bmngr->blkh.list); init_list_head(&bb->mapped_list); bh = malloc(sizeof(struct erofs_buffer_head)); @@ -278,7 +281,7 @@ struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, } } - ret = __erofs_battach(bb, bh, size, alignsize, + ret = __erofs_battach(sbi, bb, bh, size, alignsize, required_ext + inline_ext, false); if (ret < 0) { free(bh); @@ -287,13 +290,14 @@ struct erofs_buffer_head *erofs_balloc(int type, erofs_off_t size, return bh; } -struct erofs_buffer_head *erofs_battach(struct erofs_buffer_head *bh, +struct erofs_buffer_head *erofs_battach(struct erofs_sb_info *sbi, + struct erofs_buffer_head *bh, int type, unsigned int size) { struct erofs_buffer_block *const bb = bh->block; struct erofs_buffer_head *nbh; unsigned int alignsize; - int ret = get_alignsize(type, &type); + int ret = get_alignsize(sbi, type, &type); if (ret < 0) return ERR_PTR(ret); @@ -307,7 +311,7 @@ struct erofs_buffer_head *erofs_battach(struct erofs_buffer_head *bh, if (!nbh) return ERR_PTR(-ENOMEM); - ret = __erofs_battach(bb, nbh, size, alignsize, 0, false); + ret = __erofs_battach(sbi, bb, nbh, size, alignsize, 0, false); if (ret < 0) { free(nbh); return ERR_PTR(ret); @@ -315,61 +319,63 @@ struct erofs_buffer_head *erofs_battach(struct erofs_buffer_head *bh, return nbh; } -static erofs_blk_t __erofs_mapbh(struct erofs_buffer_block *bb) +static erofs_blk_t __erofs_mapbh(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb) { erofs_blk_t blkaddr; - struct erofs_sb_info *sbi = &g_sbi; if (bb->blkaddr == NULL_ADDR) { - bb->blkaddr = tail_blkaddr; - last_mapped_block = bb; - erofs_bupdate_mapped(bb); + bb->blkaddr = sbi->bmngr->tail_blkaddr; + sbi->bmngr->last_mapped_block = bb; + erofs_bupdate_mapped(sbi, bb); } blkaddr = bb->blkaddr + BLK_ROUND_UP(sbi, bb->buffers.off); - if (blkaddr > tail_blkaddr) - tail_blkaddr = blkaddr; + if (blkaddr > sbi->bmngr->tail_blkaddr) + sbi->bmngr->tail_blkaddr = blkaddr; return blkaddr; } -erofs_blk_t erofs_mapbh(struct erofs_buffer_block *bb) +erofs_blk_t erofs_mapbh(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb) { - struct erofs_buffer_block *t = last_mapped_block; + struct erofs_buffer_block *t = sbi->bmngr->last_mapped_block; if (bb && bb->blkaddr != NULL_ADDR) return bb->blkaddr; do { t = list_next_entry(t, list); - if (t == &blkh) + if (t == &sbi->bmngr->blkh) break; DBG_BUGON(t->blkaddr != NULL_ADDR); - (void)__erofs_mapbh(t); + (void)__erofs_mapbh(sbi, t); } while (t != bb); - return tail_blkaddr; + return sbi->bmngr->tail_blkaddr; } -static void erofs_bfree(struct erofs_buffer_block *bb) +static void erofs_bfree(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb) { DBG_BUGON(!list_empty(&bb->buffers.list)); - if (bb == last_mapped_block) - last_mapped_block = list_prev_entry(bb, list); + if (bb == sbi->bmngr->last_mapped_block) + sbi->bmngr->last_mapped_block = list_prev_entry(bb, list); list_del(&bb->mapped_list); list_del(&bb->list); free(bb); } -int erofs_bflush(struct erofs_buffer_block *bb) +int erofs_bflush(struct erofs_sb_info *sbi, + struct erofs_buffer_block *bb) { - struct erofs_sb_info *sbi = &g_sbi; const unsigned int blksiz = erofs_blksiz(sbi); struct erofs_buffer_block *p, *n; erofs_blk_t blkaddr; - list_for_each_entry_safe(p, n, &blkh.list, list) { + list_for_each_entry_safe(p, n, &sbi->bmngr->blkh.list, list) { struct erofs_buffer_head *bh, *nbh; unsigned int padding; bool skip = false; @@ -378,7 +384,7 @@ int erofs_bflush(struct erofs_buffer_block *bb) if (p == bb) break; - blkaddr = __erofs_mapbh(p); + blkaddr = __erofs_mapbh(sbi, p); list_for_each_entry_safe(bh, nbh, &p->buffers.list, list) { if (bh->op == &erofs_skip_write_bhops) { @@ -401,23 +407,25 @@ int erofs_bflush(struct erofs_buffer_block *bb) padding, true); if (p->type != DATA) - erofs_metablkcnt += BLK_ROUND_UP(sbi, p->buffers.off); + sbi->bmngr->erofs_metablkcnt += BLK_ROUND_UP(sbi, + p->buffers.off); erofs_dbg("block %u to %u flushed", p->blkaddr, blkaddr - 1); - erofs_bfree(p); + erofs_bfree(sbi, p); } return 0; } -void erofs_bdrop(struct erofs_buffer_head *bh, bool tryrevoke) +void erofs_bdrop(struct erofs_sb_info *sbi, struct erofs_buffer_head *bh, + bool tryrevoke) { struct erofs_buffer_block *const bb = bh->block; const erofs_blk_t blkaddr = bh->block->blkaddr; bool rollback = false; - struct erofs_sb_info *sbi = &g_sbi; /* tail_blkaddr could be rolled back after revoking all bhs */ if (tryrevoke && blkaddr != NULL_ADDR && - tail_blkaddr == blkaddr + BLK_ROUND_UP(sbi, bb->buffers.off)) + sbi->bmngr->tail_blkaddr == blkaddr + BLK_ROUND_UP(sbi, + bb->buffers.off)) rollback = true; bh->op = &erofs_drop_directly_bhops; @@ -427,13 +435,21 @@ void erofs_bdrop(struct erofs_buffer_head *bh, bool tryrevoke) return; if (!rollback && bb->type != DATA) - erofs_metablkcnt += BLK_ROUND_UP(sbi, bb->buffers.off); - erofs_bfree(bb); + sbi->bmngr->erofs_metablkcnt += BLK_ROUND_UP(sbi, bb->buffers.off); + erofs_bfree(sbi, bb); if (rollback) - tail_blkaddr = blkaddr; + sbi->bmngr->tail_blkaddr = blkaddr; } -erofs_blk_t erofs_total_metablocks(void) +erofs_blk_t erofs_total_metablocks(struct erofs_sb_info *sbi) { - return erofs_metablkcnt; + return sbi->bmngr->erofs_metablkcnt; } + +void erofs_buffer_exit(struct erofs_sb_info *sbi) +{ + if (sbi->bmngr) { + free(sbi->bmngr); + sbi->bmngr= NULL; + } +} \ No newline at end of file diff --git a/lib/compress.c b/lib/compress.c index b473587..e74e58c 100644 --- a/lib/compress.c +++ b/lib/compress.c @@ -1111,7 +1111,7 @@ int erofs_commit_compressed_file(struct z_erofs_compress_ictx *ictx, } if (compressed_blocks) { - ret = erofs_bh_balloon(bh, erofs_pos(sbi, compressed_blocks)); + ret = erofs_bh_balloon(sbi, bh, erofs_pos(sbi, compressed_blocks)); DBG_BUGON(ret != erofs_blksiz(sbi)); } else { if (!cfg.c_fragments && !cfg.c_dedupe) @@ -1126,7 +1126,7 @@ int erofs_commit_compressed_file(struct z_erofs_compress_ictx *ictx, bh->op = &erofs_skip_write_bhops; inode->bh_data = bh; } else { - erofs_bdrop(bh, false); + erofs_bdrop(sbi, bh, false); } inode->u.i_blocks = compressed_blocks; @@ -1148,7 +1148,7 @@ err_free_meta: free(compressmeta); inode->compressmeta = NULL; err_free_idata: - erofs_bdrop(bh, true); /* revoke buffer */ + erofs_bdrop(sbi, bh, true); /* revoke buffer */ if (inode->idata) { free(inode->idata); inode->idata = NULL; @@ -1366,14 +1366,14 @@ int erofs_mt_write_compressed_file(struct z_erofs_compress_ictx *ictx) pthread_cond_wait(&ictx->cond, &ictx->mutex); pthread_mutex_unlock(&ictx->mutex); - bh = erofs_balloc(DATA, 0, 0, 0); + bh = erofs_balloc(ictx->inode->sbi, DATA, 0, 0, 0); if (IS_ERR(bh)) { ret = PTR_ERR(bh); goto out; } DBG_BUGON(!head); - blkaddr = erofs_mapbh(bh->block); + blkaddr = erofs_mapbh(ictx->inode->sbi, bh->block); ret = 0; do { @@ -1531,12 +1531,12 @@ int erofs_write_compressed_file(struct z_erofs_compress_ictx *ictx) #endif /* allocate main data buffer */ - bh = erofs_balloc(DATA, 0, 0, 0); + bh = erofs_balloc(inode->sbi, DATA, 0, 0, 0); if (IS_ERR(bh)) { ret = PTR_ERR(bh); goto err_free_idata; } - blkaddr = erofs_mapbh(bh->block); /* start_blkaddr */ + blkaddr = erofs_mapbh(inode->sbi, bh->block); /* start_blkaddr */ ictx->seg_num = 1; sctx = (struct z_erofs_compress_sctx) { @@ -1560,7 +1560,7 @@ int erofs_write_compressed_file(struct z_erofs_compress_ictx *ictx) goto out; err_free_idata: - erofs_bdrop(bh, true); /* revoke buffer */ + erofs_bdrop(inode->sbi, bh, true); /* revoke buffer */ if (inode->idata) { free(inode->idata); inode->idata = NULL; @@ -1596,13 +1596,13 @@ static int z_erofs_build_compr_cfgs(struct erofs_sb_info *sbi, } }; - bh = erofs_battach(bh, META, sizeof(lz4alg)); + bh = erofs_battach(sbi, bh, META, sizeof(lz4alg)); if (IS_ERR(bh)) { DBG_BUGON(1); return PTR_ERR(bh); } - erofs_mapbh(bh->block); - ret = erofs_dev_write(sbi, &lz4alg, erofs_btell(bh, false), + erofs_mapbh(sbi, bh->block); + ret = erofs_dev_write(sbi, &lz4alg, erofs_btell(sbi, bh, false), sizeof(lz4alg)); bh->op = &erofs_drop_directly_bhops; } @@ -1620,13 +1620,13 @@ static int z_erofs_build_compr_cfgs(struct erofs_sb_info *sbi, } }; - bh = erofs_battach(bh, META, sizeof(lzmaalg)); + bh = erofs_battach(sbi, bh, META, sizeof(lzmaalg)); if (IS_ERR(bh)) { DBG_BUGON(1); return PTR_ERR(bh); } - erofs_mapbh(bh->block); - ret = erofs_dev_write(sbi, &lzmaalg, erofs_btell(bh, false), + erofs_mapbh(sbi, bh->block); + ret = erofs_dev_write(sbi, &lzmaalg, erofs_btell(sbi, bh, false), sizeof(lzmaalg)); bh->op = &erofs_drop_directly_bhops; } @@ -1644,13 +1644,13 @@ static int z_erofs_build_compr_cfgs(struct erofs_sb_info *sbi, } }; - bh = erofs_battach(bh, META, sizeof(zalg)); + bh = erofs_battach(sbi, bh, META, sizeof(zalg)); if (IS_ERR(bh)) { DBG_BUGON(1); return PTR_ERR(bh); } - erofs_mapbh(bh->block); - ret = erofs_dev_write(sbi, &zalg, erofs_btell(bh, false), + erofs_mapbh(sbi, bh->block); + ret = erofs_dev_write(sbi, &zalg, erofs_btell(sbi, bh, false), sizeof(zalg)); bh->op = &erofs_drop_directly_bhops; } @@ -1667,13 +1667,13 @@ static int z_erofs_build_compr_cfgs(struct erofs_sb_info *sbi, } }; - bh = erofs_battach(bh, META, sizeof(zalg)); + bh = erofs_battach(sbi, bh, META, sizeof(zalg)); if (IS_ERR(bh)) { DBG_BUGON(1); return PTR_ERR(bh); } - erofs_mapbh(bh->block); - ret = erofs_dev_write(sbi, &zalg, erofs_btell(bh, false), + erofs_mapbh(sbi, bh->block); + ret = erofs_dev_write(sbi, &zalg, erofs_btell(sbi, bh, false), sizeof(zalg)); bh->op = &erofs_drop_directly_bhops; } diff --git a/lib/inode.c b/lib/inode.c index 8d60088..2ca91ff 100644 --- a/lib/inode.c +++ b/lib/inode.c @@ -186,7 +186,8 @@ static int __allocate_inode_bh_data(struct erofs_inode *inode, } /* allocate main data buffer */ - bh = erofs_balloc(type, erofs_pos(inode->sbi, nblocks), 0, 0); + bh = erofs_balloc(inode->sbi, type, erofs_pos(inode->sbi, nblocks), + 0, 0); if (IS_ERR(bh)) return PTR_ERR(bh); @@ -194,7 +195,7 @@ static int __allocate_inode_bh_data(struct erofs_inode *inode, inode->bh_data = bh; /* get blkaddr of the bh */ - ret = erofs_mapbh(bh->block); + ret = erofs_mapbh(inode->sbi, bh->block); DBG_BUGON(ret < 0); /* write blocks except for the tail-end block */ @@ -314,8 +315,8 @@ erofs_nid_t erofs_lookupnid(struct erofs_inode *inode) erofs_off_t off, meta_offset; if (bh && (long long)inode->nid <= 0) { - erofs_mapbh(bh->block); - off = erofs_btell(bh, false); + erofs_mapbh(sbi, bh->block); + off = erofs_btell(sbi, bh, false); meta_offset = erofs_pos(sbi, sbi->meta_blkaddr); DBG_BUGON(off < meta_offset); @@ -576,7 +577,7 @@ int erofs_iflush(struct erofs_inode *inode) int ret; if (inode->bh) - off = erofs_btell(inode->bh, false); + off = erofs_btell(sbi, inode->bh, false); else off = erofs_iloc(inode); @@ -724,7 +725,7 @@ static int erofs_prepare_tail_block(struct erofs_inode *inode) bh = inode->bh_data; if (bh) { /* expend a block as the tail block (should be successful) */ - ret = erofs_bh_balloon(bh, erofs_blksiz(sbi)); + ret = erofs_bh_balloon(sbi, bh, erofs_blksiz(sbi)); if (ret != erofs_blksiz(sbi)) { DBG_BUGON(1); return -EIO; @@ -768,7 +769,7 @@ static int erofs_prepare_inode_buffer(struct erofs_inode *inode) inode->datalayout = EROFS_INODE_FLAT_PLAIN; } - bh = erofs_balloc(INODE, inodesize, 0, inode->idata_size); + bh = erofs_balloc(inode->sbi, INODE, inodesize, 0, inode->idata_size); if (bh == ERR_PTR(-ENOSPC)) { int ret; @@ -781,7 +782,7 @@ noinline: ret = erofs_prepare_tail_block(inode); if (ret) return ret; - bh = erofs_balloc(INODE, inodesize, 0, 0); + bh = erofs_balloc(inode->sbi, INODE, inodesize, 0, 0); if (IS_ERR(bh)) return PTR_ERR(bh); DBG_BUGON(inode->bh_inline); @@ -801,7 +802,7 @@ noinline: } /* allocate inline buffer */ - ibh = erofs_battach(bh, META, inode->idata_size); + ibh = erofs_battach(inode->sbi, bh, META, inode->idata_size); if (IS_ERR(ibh)) return PTR_ERR(ibh); @@ -818,7 +819,7 @@ noinline: static int erofs_bh_flush_write_inline(struct erofs_buffer_head *bh) { struct erofs_inode *const inode = bh->fsprivate; - const erofs_off_t off = erofs_btell(bh, false); + const erofs_off_t off = erofs_btell(inode->sbi, bh, false); int ret; ret = erofs_dev_write(inode->sbi, inode->idata, off, inode->idata_size); @@ -860,29 +861,29 @@ static int erofs_write_tail_end(struct erofs_inode *inode) erofs_off_t pos, zero_pos; if (!bh) { - bh = erofs_balloc(DATA, erofs_blksiz(sbi), 0, 0); + bh = erofs_balloc(sbi, DATA, erofs_blksiz(sbi), 0, 0); if (IS_ERR(bh)) return PTR_ERR(bh); bh->op = &erofs_skip_write_bhops; /* get blkaddr of bh */ - ret = erofs_mapbh(bh->block); + ret = erofs_mapbh(sbi, bh->block); inode->u.i_blkaddr = bh->block->blkaddr; inode->bh_data = bh; } else { if (inode->lazy_tailblock) { /* expend a tail block (should be successful) */ - ret = erofs_bh_balloon(bh, erofs_blksiz(sbi)); + ret = erofs_bh_balloon(sbi, bh, erofs_blksiz(sbi)); if (ret != erofs_blksiz(sbi)) { DBG_BUGON(1); return -EIO; } inode->lazy_tailblock = false; } - ret = erofs_mapbh(bh->block); + ret = erofs_mapbh(sbi, bh->block); } DBG_BUGON(ret < 0); - pos = erofs_btell(bh, true) - erofs_blksiz(sbi); + pos = erofs_btell(sbi, bh, true) - erofs_blksiz(sbi); /* 0'ed data should be padded at head for 0padding conversion */ if (erofs_sb_has_lz4_0padding(sbi) && inode->compressed_idata) { @@ -917,7 +918,7 @@ out: * Don't leave DATA buffers which were written in the global * buffer list. It will make balloc() slowly. */ - erofs_bdrop(bh, false); + erofs_bdrop(sbi, bh, false); inode->bh_data = NULL; } return 0; @@ -1157,8 +1158,8 @@ static void erofs_fixup_meta_blkaddr(struct erofs_inode *rootdir) struct erofs_sb_info *sbi = rootdir->sbi; erofs_off_t off, meta_offset; - erofs_mapbh(bh->block); - off = erofs_btell(bh, false); + erofs_mapbh(sbi, bh->block); + off = erofs_btell(sbi, bh, false); if (off > rootnid_maxoffset) meta_offset = round_up(off - rootnid_maxoffset, erofs_blksiz(sbi)); @@ -1176,16 +1177,16 @@ static int erofs_inode_reserve_data_blocks(struct erofs_inode *inode) struct erofs_buffer_head *bh; /* allocate data blocks */ - bh = erofs_balloc(DATA, alignedsz, 0, 0); + bh = erofs_balloc(sbi, DATA, alignedsz, 0, 0); if (IS_ERR(bh)) return PTR_ERR(bh); /* get blkaddr of the bh */ - (void)erofs_mapbh(bh->block); + (void)erofs_mapbh(sbi, bh->block); /* write blocks except for the tail-end block */ inode->u.i_blkaddr = bh->block->blkaddr; - erofs_bdrop(bh, false); + erofs_bdrop(sbi, bh, false); inode->datalayout = EROFS_INODE_FLAT_PLAIN; tarerofs_blocklist_write(inode->u.i_blkaddr, nblocks, inode->i_ino[1]); diff --git a/lib/super.c b/lib/super.c index 3fbaf66..56eec37 100644 --- a/lib/super.c +++ b/lib/super.c @@ -149,6 +149,7 @@ void erofs_put_super(struct erofs_sb_info *sbi) sbi->devs = NULL; } erofs_xattr_prefixes_cleanup(sbi); + erofs_buffer_exit(sbi); } int erofs_writesb(struct erofs_sb_info *sbi, struct erofs_buffer_head *sb_bh, @@ -176,7 +177,7 @@ int erofs_writesb(struct erofs_sb_info *sbi, struct erofs_buffer_head *sb_bh, char *buf; int ret; - *blocks = erofs_mapbh(NULL); + *blocks = erofs_mapbh(sbi, NULL); sb.blocks = cpu_to_le32(*blocks); memcpy(sb.uuid, sbi->uuid, sizeof(sb.uuid)); memcpy(sb.volume_name, sbi->volume_name, sizeof(sb.volume_name)); @@ -194,42 +195,42 @@ int erofs_writesb(struct erofs_sb_info *sbi, struct erofs_buffer_head *sb_bh, } memcpy(buf + EROFS_SUPER_OFFSET, &sb, sizeof(sb)); - ret = erofs_dev_write(sbi, buf, sb_bh ? erofs_btell(sb_bh, false) : 0, + ret = erofs_dev_write(sbi, buf, sb_bh ? erofs_btell(sbi, sb_bh, false) : 0, EROFS_SUPER_END); free(buf); if (sb_bh) - erofs_bdrop(sb_bh, false); + erofs_bdrop(sbi, sb_bh, false); return ret; } -struct erofs_buffer_head *erofs_reserve_sb(void) +struct erofs_buffer_head *erofs_reserve_sb(struct erofs_sb_info *sbi) { struct erofs_buffer_head *bh; int err; - erofs_buffer_init(0); - bh = erofs_balloc(META, 0, 0, 0); + erofs_buffer_init(sbi, 0); + bh = erofs_balloc(sbi, META, 0, 0, 0); if (IS_ERR(bh)) { erofs_err("failed to allocate super: %s", PTR_ERR(bh)); return bh; } bh->op = &erofs_skip_write_bhops; - err = erofs_bh_balloon(bh, EROFS_SUPER_END); + err = erofs_bh_balloon(sbi, bh, EROFS_SUPER_END); if (err < 0) { erofs_err("failed to balloon super: %s", erofs_strerror(err)); goto err_bdrop; } /* make sure that the super block should be the very first blocks */ - (void)erofs_mapbh(bh->block); - if (erofs_btell(bh, false) != 0) { + (void)erofs_mapbh(sbi, bh->block); + if (erofs_btell(sbi, bh, false) != 0) { erofs_err("failed to pin super block @ 0"); err = -EFAULT; goto err_bdrop; } return bh; err_bdrop: - erofs_bdrop(bh, true); + erofs_bdrop(sbi, bh, true); return ERR_PTR(err); } diff --git a/lib/xattr.c b/lib/xattr.c index b0f80e9..3356c3e 100644 --- a/lib/xattr.c +++ b/lib/xattr.c @@ -906,7 +906,7 @@ int erofs_build_shared_xattrs_from_path(struct erofs_sb_info *sbi, const char *p return -ENOMEM; } - bh = erofs_balloc(XATTR, shared_xattrs_size, 0, 0); + bh = erofs_balloc(sbi, XATTR, shared_xattrs_size, 0, 0); if (IS_ERR(bh)) { free(sorted_n); free(buf); @@ -914,8 +914,8 @@ int erofs_build_shared_xattrs_from_path(struct erofs_sb_info *sbi, const char *p } bh->op = &erofs_skip_write_bhops; - erofs_mapbh(bh->block); - off = erofs_btell(bh, false); + erofs_mapbh(sbi, bh->block); + off = erofs_btell(sbi, bh, false); sbi->xattr_blkaddr = off / erofs_blksiz(sbi); off %= erofs_blksiz(sbi); @@ -930,9 +930,9 @@ int erofs_build_shared_xattrs_from_path(struct erofs_sb_info *sbi, const char *p shared_xattrs_list = sorted_n[0]; free(sorted_n); bh->op = &erofs_drop_directly_bhops; - ret = erofs_dev_write(sbi, buf, erofs_btell(bh, false), shared_xattrs_size); + ret = erofs_dev_write(sbi, buf, erofs_btell(sbi, bh, false), shared_xattrs_size); free(buf); - erofs_bdrop(bh, false); + erofs_bdrop(sbi, bh, false); out: erofs_cleanxattrs(true); return ret; diff --git a/mkfs/main.c b/mkfs/main.c index 37f250b..1793fd8 100644 --- a/mkfs/main.c +++ b/mkfs/main.c @@ -1105,7 +1105,7 @@ static int erofs_mkfs_rebuild_load_trees(struct erofs_inode *root) return 0; } -static void erofs_mkfs_showsummaries(erofs_blk_t nblocks) +static void erofs_mkfs_showsummaries(struct erofs_sb_info *sbi, erofs_blk_t nblocks) { char uuid_str[37] = {}; char *incr = incremental_mode ? "new" : "total"; @@ -1121,7 +1121,7 @@ static void erofs_mkfs_showsummaries(erofs_blk_t nblocks) "Filesystem %s metadata blocks: %u\n" "Filesystem %s deduplicated bytes (of source files): %llu\n", uuid_str, nblocks, 1U << g_sbi.blkszbits, g_sbi.inos | 0ULL, - incr, erofs_total_metablocks(), + incr, erofs_total_metablocks(sbi), incr, g_sbi.saved_by_deduplication | 0ULL); } @@ -1245,7 +1245,7 @@ int main(int argc, char **argv) } if (!incremental_mode) { - sb_bh = erofs_reserve_sb(); + sb_bh = erofs_reserve_sb(&g_sbi); if (IS_ERR(sb_bh)) { err = PTR_ERR(sb_bh); goto exit; @@ -1270,7 +1270,7 @@ int main(int argc, char **argv) u.startblk = DIV_ROUND_UP(u.st.st_size, erofs_blksiz(&g_sbi)); else u.startblk = g_sbi.primarydevice_blocks; - erofs_buffer_init(u.startblk); + erofs_buffer_init(&g_sbi, u.startblk); sb_bh = NULL; } @@ -1392,7 +1392,7 @@ int main(int argc, char **argv) } /* flush all buffers except for the superblock */ - err = erofs_bflush(NULL); + err = erofs_bflush(&g_sbi, NULL); if (err) goto exit; @@ -1405,7 +1405,7 @@ int main(int argc, char **argv) goto exit; /* flush all remaining buffers */ - err = erofs_bflush(NULL); + err = erofs_bflush(&g_sbi, NULL); if (err) goto exit; @@ -1448,6 +1448,7 @@ exit: return 1; } erofs_update_progressinfo("Build completed.\n"); - erofs_mkfs_showsummaries(nblocks); + erofs_mkfs_showsummaries(&g_sbi, nblocks); + erofs_put_super(&g_sbi); return 0; } -- 2.43.5