Currently the sb block is parsed from within ext4_fill_super(), however since the new mount api separates the option parsing and super block setup into two distinct steps we need to move the parsing of sb block into handle_mount_opt(). In preparation for the next step we also need to refactor ext4_fill_super so that we can parse options separately. Unfortunately we still need to parse options specified in the super block itself and that needs to be done after we've read the super block from disk. Another complication is that we really want to apply the options from super block before we apply user specified mount options. So with this patch we're going through the following sequence: - parse user provided options (including sb block) - initialize sbi and store s_sb_block if provided - in __ext4_fill_super() - read the super block - parse and apply options specified in s_mount_opts - check and apply user provided options stored in ctx - continue with the regular ext4_fill_super operation It's ugly, but if we still want to support s_mount_opts we have to do it in this order. __ext4_fill_super would really benefit from some more refactoring, but that will have to wait until after the mount api conversion is done. Signed-off-by: Lukas Czerner <lczerner@xxxxxxxxxx> --- fs/ext4/super.c | 297 ++++++++++++++++++++++++++++++++---------------- 1 file changed, 196 insertions(+), 101 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index f3dd82fef453..417a929cb0ab 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1667,29 +1667,6 @@ static const match_table_t tokens = { {Opt_err, NULL}, }; -static ext4_fsblk_t get_sb_block(void **data) -{ - ext4_fsblk_t sb_block; - char *options = (char *) *data; - - if (!options || strncmp(options, "sb=", 3) != 0) - return 1; /* Default location */ - - options += 3; - /* TODO: use simple_strtoll with >32bit ext4 */ - sb_block = simple_strtoul(options, &options, 0); - if (*options && *options != ',') { - printk(KERN_ERR "EXT4-fs: Invalid sb specification: %s\n", - (char *) *data); - return 1; - } - if (*options == ',') - options++; - *data = (void *) options; - - return sb_block; -} - #define DEFAULT_JOURNAL_IOPRIO (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, 3)) static const char deprecated_msg[] = "Mount option \"%s\" will be removed by %s\n" @@ -1766,6 +1743,7 @@ static const struct mount_opts { {Opt_stripe, 0, MOPT_GTE0}, {Opt_resuid, 0, MOPT_GTE0}, {Opt_resgid, 0, MOPT_GTE0}, + {Opt_sb, 0, MOPT_GTE0}, {Opt_journal_dev, 0, MOPT_NO_EXT2 | MOPT_GTE0}, {Opt_journal_path, 0, MOPT_NO_EXT2 | MOPT_STRING}, {Opt_journal_ioprio, 0, MOPT_NO_EXT2 | MOPT_GTE0}, @@ -1854,6 +1832,7 @@ static int ext4_sb_read_encoding(const struct ext4_super_block *es, #define EXT4_SPEC_s_resuid (1 << 13) #define EXT4_SPEC_s_resgid (1 << 14) #define EXT4_SPEC_s_commit_interval (1 << 15) +#define EXT4_SPEC_s_sb_block (1 << 16) struct ext4_fs_context { char *s_qf_names[EXT4_MAXQUOTAS]; @@ -1881,6 +1860,7 @@ struct ext4_fs_context { u32 s_min_batch_time; kuid_t s_resuid; kgid_t s_resgid; + ext4_fsblk_t s_sb_block; }; #ifdef CONFIG_QUOTA @@ -1994,8 +1974,6 @@ static int handle_mount_opt(struct fs_context *fc, struct fs_parameter *param) case Opt_nouser_xattr: ext4_msg(NULL, KERN_WARNING, deprecated_msg, param->key, "3.5"); break; - case Opt_sb: - return 1; /* handled by get_sb_block() */ case Opt_removed: ext4_msg(NULL, KERN_WARNING, "Ignoring removed %s option", param->key); @@ -2108,6 +2086,14 @@ static int handle_mount_opt(struct fs_context *fc, struct fs_parameter *param) } ctx->s_resgid = gid; ctx->spec |= EXT4_SPEC_s_resgid; + } else if (token == Opt_sb) { + if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE) { + ext4_msg(NULL, KERN_WARNING, + "Ignoring %s option on remount", param->key); + } else { + ctx->s_sb_block = result.uint_32; + ctx->spec |= EXT4_SPEC_s_sb_block; + } } else if (token == Opt_journal_dev) { if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE) { ext4_msg(NULL, KERN_ERR, @@ -2202,28 +2188,15 @@ static int handle_mount_opt(struct fs_context *fc, struct fs_parameter *param) return 1; } -static int parse_options(char *options, struct super_block *sb, - unsigned long *journal_devnum, - unsigned int *journal_ioprio, - int is_remount) +static int parse_options(struct fs_context *fc, char *options) { - struct ext4_fs_context ctx; struct fs_parameter param; - struct fs_context fc; char *value; int ret; char *p; if (!options) - return 1; - - memset(&fc, 0, sizeof(fc)); - memset(&ctx, 0, sizeof(ctx)); - fc.fs_private = &ctx; - fc.s_fs_info = EXT4_SB(sb); - - if (is_remount) - fc.purpose = FS_CONTEXT_FOR_RECONFIGURE; + return 0; while ((p = strsep(&options, ",")) != NULL) { @@ -2250,27 +2223,63 @@ static int parse_options(char *options, struct super_block *sb, } } - ret = handle_mount_opt(&fc, ¶m); + ret = handle_mount_opt(fc, ¶m); kfree(param.string); if (ret < 0) - return 0; + return ret; } - ret = ext4_validate_options(&fc); + ret = ext4_validate_options(fc); if (ret < 0) + return ret; + + return 0; +} + +static int parse_apply_options(char *options, struct super_block *sb, + unsigned long *journal_devnum, + unsigned int *journal_ioprio, + int is_remount) +{ + struct ext4_fs_context *ctx; + struct fs_context *fc; + int ret = -ENOMEM; + + if (!options) return 0; - ret = ext4_check_opt_consistency(&fc, sb); + fc = kzalloc(sizeof(struct fs_context), GFP_KERNEL); + if (!fc) + return ret; + ctx = kzalloc(sizeof(struct ext4_fs_context), GFP_KERNEL); + if (!ctx) + goto out_free; + + fc->fs_private = ctx; + fc->s_fs_info = EXT4_SB(sb); + + if (is_remount) + fc->purpose = FS_CONTEXT_FOR_RECONFIGURE; + + ret = parse_options(fc, options); if (ret < 0) - return 0; + goto out_free; - if (ctx.spec & EXT4_SPEC_JOURNAL_DEV) - *journal_devnum = ctx.journal_devnum; - if (ctx.spec & EXT4_SPEC_JOURNAL_IOPRIO) - *journal_ioprio = ctx.journal_ioprio; + ret = ext4_check_opt_consistency(fc, sb); + if (ret < 0) + goto out_free; - ext4_apply_options(&fc, sb); - return 1; + if (ctx->spec & EXT4_SPEC_JOURNAL_DEV) + *journal_devnum = ctx->journal_devnum; + if (ctx->spec & EXT4_SPEC_JOURNAL_IOPRIO) + *journal_ioprio = ctx->journal_ioprio; + + ext4_apply_options(fc, sb); + +out_free: + kfree(ctx); + kfree(fc); + return 0; } static void ext4_apply_quota_options(struct fs_context *fc, @@ -4020,21 +4029,53 @@ static void ext4_set_resv_clusters(struct super_block *sb) atomic64_set(&sbi->s_resv_clusters, resv_clusters); } -static int ext4_fill_super(struct super_block *sb, void *data, int silent) +static void ext4_free_sbi(struct ext4_sb_info *sbi) +{ + if (!sbi) + return; + + kfree(sbi->s_blockgroup_lock); + fs_put_dax(sbi->s_daxdev); + kfree(sbi); +} + +static struct ext4_sb_info *ext4_alloc_sbi(struct super_block *sb) +{ + struct ext4_sb_info *sbi; + + sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); + if (!sbi) + return NULL; + + sbi->s_daxdev = fs_dax_get_by_bdev(sb->s_bdev); + + sbi->s_blockgroup_lock = + kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); + + if (!sbi->s_blockgroup_lock) + goto err_out; + + sb->s_fs_info = sbi; + sbi->s_sb = sb; + return sbi; +err_out: + fs_put_dax(sbi->s_daxdev); + kfree(sbi); + return NULL; +} + +static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb, + int silent) { - struct dax_device *dax_dev = fs_dax_get_by_bdev(sb->s_bdev); - char *orig_data = kstrdup(data, GFP_KERNEL); struct buffer_head *bh; struct ext4_super_block *es = NULL; - struct ext4_sb_info *sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); + struct ext4_sb_info *sbi = EXT4_SB(sb); ext4_fsblk_t block; - ext4_fsblk_t sb_block = get_sb_block(&data); ext4_fsblk_t logical_sb_block; unsigned long offset = 0; unsigned long journal_devnum = 0; unsigned long def_mount_opts; struct inode *root; - const char *descr; int ret = -ENOMEM; int blocksize, clustersize; unsigned int db_count; @@ -4044,26 +4085,13 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) int err = 0; unsigned int journal_ioprio = DEFAULT_JOURNAL_IOPRIO; ext4_group_t first_not_zeroed; + struct ext4_fs_context *ctx = fc->fs_private; - if ((data && !orig_data) || !sbi) - goto out_free_base; - - sbi->s_daxdev = dax_dev; - sbi->s_blockgroup_lock = - kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); - if (!sbi->s_blockgroup_lock) - goto out_free_base; - - sb->s_fs_info = sbi; - sbi->s_sb = sb; sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS; - sbi->s_sb_block = sb_block; if (sb->s_bdev->bd_part) sbi->s_sectors_written_start = part_stat_read(sb->s_bdev->bd_part, sectors[STAT_WRITE]); - /* Cleanup superblock name */ - strreplace(sb->s_id, '/', '!'); /* -EINVAL is default */ ret = -EINVAL; @@ -4078,10 +4106,10 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) * block sizes. We need to calculate the offset from buffer start. */ if (blocksize != EXT4_MIN_BLOCK_SIZE) { - logical_sb_block = sb_block * EXT4_MIN_BLOCK_SIZE; + logical_sb_block = sbi->s_sb_block * EXT4_MIN_BLOCK_SIZE; offset = do_div(logical_sb_block, blocksize); } else { - logical_sb_block = sb_block; + logical_sb_block = sbi->s_sb_block; } if (!(bh = sb_bread_unmovable(sb, logical_sb_block))) { @@ -4203,8 +4231,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) GFP_KERNEL); if (!s_mount_opts) goto failed_mount; - if (!parse_options(s_mount_opts, sb, &journal_devnum, - &journal_ioprio, 0)) { + if (!parse_apply_options(s_mount_opts, sb, &journal_devnum, + &journal_ioprio, 0)) { ext4_msg(sb, KERN_WARNING, "failed to parse options in superblock: %s", s_mount_opts); @@ -4212,10 +4240,19 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) kfree(s_mount_opts); } sbi->s_def_mount_opt = sbi->s_mount_opt; - if (!parse_options((char *) data, sb, &journal_devnum, - &journal_ioprio, 0)) + + /* Now check and apply options we've got in fs context */ + err = ext4_check_opt_consistency(fc, sb); + if (err < 0) goto failed_mount; + if (ctx->spec & EXT4_SPEC_JOURNAL_DEV) + journal_devnum = ctx->journal_devnum; + if (ctx->spec & EXT4_SPEC_JOURNAL_IOPRIO) + journal_ioprio = ctx->journal_ioprio; + + ext4_apply_options(fc, sb); + #ifdef CONFIG_UNICODE if (ext4_has_feature_casefold(sb) && !sbi->s_encoding) { const struct ext4_sb_encodings *encoding_info; @@ -4413,7 +4450,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) } brelse(bh); - logical_sb_block = sb_block * EXT4_MIN_BLOCK_SIZE; + logical_sb_block = sbi->s_sb_block * EXT4_MIN_BLOCK_SIZE; offset = do_div(logical_sb_block, blocksize); bh = sb_bread_unmovable(sb, logical_sb_block); if (!bh) { @@ -5004,15 +5041,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) ext4_msg(sb, KERN_INFO, "recovery complete"); ext4_mark_recovery_complete(sb, es); } - if (EXT4_SB(sb)->s_journal) { - if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) - descr = " journalled data mode"; - else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA) - descr = " ordered data mode"; - else - descr = " writeback data mode"; - } else - descr = "out journal"; if (test_opt(sb, DISCARD)) { struct request_queue *q = bdev_get_queue(sb->s_bdev); @@ -5022,13 +5050,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) "the device does not support discard"); } - if (___ratelimit(&ext4_mount_msg_ratelimit, "EXT4-fs mount")) - ext4_msg(sb, KERN_INFO, "mounted filesystem with%s. " - "Opts: %.*s%s%s", descr, - (int) sizeof(sbi->s_es->s_mount_opts), - sbi->s_es->s_mount_opts, - *sbi->s_es->s_mount_opts ? "; " : "", orig_data); - if (es->s_error_count) mod_timer(&sbi->s_err_report, jiffies + 300*HZ); /* 5 minutes */ @@ -5037,7 +5058,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) ratelimit_state_init(&sbi->s_warning_ratelimit_state, 5 * HZ, 10); ratelimit_state_init(&sbi->s_msg_ratelimit_state, 5 * HZ, 10); - kfree(orig_data); return 0; cantfind_ext4: @@ -5107,14 +5127,89 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) brelse(bh); out_fail: sb->s_fs_info = NULL; - kfree(sbi->s_blockgroup_lock); -out_free_base: - kfree(sbi); - kfree(orig_data); - fs_put_dax(dax_dev); return err ? err : ret; } +static void cleanup_ctx(struct ext4_fs_context *ctx) +{ + int i; + + if (!ctx) + return; + + for (i = 0; i < EXT4_MAXQUOTAS; i++) { + kfree(ctx->s_qf_names[i]); + } +} + +static int ext4_fill_super(struct super_block *sb, void *data, int silent) +{ + struct ext4_fs_context ctx; + struct ext4_sb_info *sbi; + struct fs_context fc; + const char *descr; + char *orig_data; + int ret = -ENOMEM; + + orig_data = kstrdup(data, GFP_KERNEL); + if (data && !orig_data) + return -ENOMEM; + + /* Cleanup superblock name */ + strreplace(sb->s_id, '/', '!'); + + memset(&fc, 0, sizeof(fc)); + memset(&ctx, 0, sizeof(ctx)); + fc.fs_private = &ctx; + + ret = parse_options(&fc, (char *) data); + if (ret < 0) + goto free_data; + + sbi = ext4_alloc_sbi(sb); + if (!sbi) { + ret = -ENOMEM; + goto free_data; + } + + fc.s_fs_info = sbi; + + sbi->s_sb_block = 1; /* Default super block location */ + if (ctx.spec & EXT4_SPEC_s_sb_block) + sbi->s_sb_block = ctx.s_sb_block; + + ret = __ext4_fill_super(&fc, sb, silent); + if (ret < 0) + goto free_sbi; + + if (EXT4_SB(sb)->s_journal) { + if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) + descr = " journalled data mode"; + else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA) + descr = " ordered data mode"; + else + descr = " writeback data mode"; + } else + descr = "out journal"; + + if (___ratelimit(&ext4_mount_msg_ratelimit, "EXT4-fs mount")) + ext4_msg(sb, KERN_INFO, "mounted filesystem with%s. " + "Opts: %.*s%s%s", descr, + (int) sizeof(sbi->s_es->s_mount_opts), + sbi->s_es->s_mount_opts, + *sbi->s_es->s_mount_opts ? "; " : "", (char *)orig_data); + + kfree(orig_data); + cleanup_ctx(&ctx); + return 0; +free_sbi: + ext4_free_sbi(sbi); +free_data: + kfree(orig_data); + cleanup_ctx(&ctx); + return ret; +} + /* * Setup any per-fs journal parameters now. We'll do this both on * initial mount, once the journal has been initialised but before we've @@ -5710,7 +5805,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) if (sbi->s_journal && sbi->s_journal->j_task->io_context) journal_ioprio = sbi->s_journal->j_task->io_context->ioprio; - if (!parse_options(data, sb, NULL, &journal_ioprio, 1)) { + if (!parse_apply_options(data, sb, NULL, &journal_ioprio, 1)) { err = -EINVAL; goto restore_opts; } -- 2.21.0