Plug buffer head leak on submit for the 7.2 kernel

Starting with the 7.2 kernel, buffer heads no longer lose a reference
during the submit-wait procedure. I didn't notice that change, so now I
would expect all buffer heads to be leaking. The fact that I still can't
tell the difference makes me believe, like I mentioned before, that
somehow the kernel is still disposing of the buffer heads. Either way,
fix this.

Signed-off-by: Ernesto A. Fernández <ernesto@corellium.com>

README.rst

@@ -57,10 +57,15 @@ Now you can just cd to the linux-apfs-rw directory and run::
 	make
 
 The resulting module is the apfs.ko file. Before you can use it you must insert
-it into the kernel, as well as its dependencies. Again as root::
+it into the kernel. Again as root::
+
+	insmod apfs.ko
+
+If you get an "unknown symbol in module" error, run::
 
 	modprobe libcrc32c
-	insmod apfs.ko
+
+and then try the insmod command again.
 
 Mount
 =====

apfs.h

@@ -44,10 +44,33 @@ static inline bool sb_rdonly(const struct super_block *sb) { return sb->s_flags
 #endif
 
 /* Compatibility wrapper around submit_bh() */
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(6, 0, 0) || RHEL_VERSION_GE(9, 2)
-#define apfs_submit_bh(op, op_flags, bh) submit_bh(op | op_flags, bh)
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(7, 2, 0)
+#define apfs_submit_bh(op, op_flags, bh)				\
+do {									\
+	put_bh(bh);							\
+	if (op == REQ_OP_READ)						\
+		bh_submit(bh, op | op_flags, bh_end_read);		\
+	else								\
+		bh_submit(bh, op | op_flags, bh_end_write);		\
+} while (0)
+#elif LINUX_VERSION_CODE >= KERNEL_VERSION(6, 0, 0) || RHEL_VERSION_GE(9, 2)
+#define apfs_submit_bh(op, op_flags, bh)				\
+do {									\
+	if (op == REQ_OP_READ)						\
+		bh->b_end_io = end_buffer_read_sync;			\
+	else								\
+		bh->b_end_io = end_buffer_write_sync;			\
+	submit_bh(op | op_flags, bh);					\
+} while (0)
 #else
-#define apfs_submit_bh(op, op_flags, bh) submit_bh(op, op_flags, bh)
+#define apfs_submit_bh(op, op_flags, bh)				\
+do {									\
+	if (op == REQ_OP_READ)						\
+		bh->b_end_io = end_buffer_read_sync;			\
+	else								\
+		bh->b_end_io = end_buffer_write_sync;			\
+	submit_bh(op, op_flags, bh);					\
+} while (0)
 #endif
 
 /*
@@ -1324,7 +1347,7 @@ __apfs_getblk(struct super_block *sb, sector_t block)
 		block -= nxi->nx_tier2_bno;
 	}
 
-#if LINUX_VERSION_CODE < KERNEL_VERSION(6, 7, 0)
+#if LINUX_VERSION_CODE < KERNEL_VERSION(6, 7, 0) && !RHEL_VERSION_GE(9, 8)
 	return __getblk_gfp(info->blki_bdev, block, sb->s_blocksize, __GFP_MOVABLE);
 #else
 	return bdev_getblk(info->blki_bdev, block, sb->s_blocksize, __GFP_MOVABLE);

dkms.conf

@@ -1,5 +1,5 @@
 PACKAGE_NAME="linux-apfs-rw"
-PACKAGE_VERSION="0.3.19"
+PACKAGE_VERSION="0.3.21"
 
 BUILT_MODULE_NAME[0]="apfs"
 DEST_MODULE_LOCATION[0]="/extra"

extents.c

@@ -2298,7 +2298,6 @@ int apfs_nonsparse_dstream_read(struct apfs_dstream_info *dstream, void *buf, si
 		if (!buffer_uptodate(bh)) {
 			get_bh(bh);
 			lock_buffer(bh);
-			bh->b_end_io = end_buffer_read_sync;
 			apfs_submit_bh(REQ_OP_READ, 0, bh);
 		}
 	}
@@ -2365,7 +2364,6 @@ void apfs_nonsparse_dstream_preread(struct apfs_dstream_info *dstream)
 		if (!buffer_uptodate(bh)) {
 			get_bh(bh);
 			lock_buffer(bh);
-			bh->b_end_io = end_buffer_read_sync;
 			apfs_submit_bh(REQ_OP_READ, 0, bh);
 		}
 		brelse(bh);

inode.c

@@ -556,7 +556,6 @@ int __apfs_write_begin(struct file *file, struct address_space *mapping, loff_t
 			if (buffer_mapped(bh) && !buffer_uptodate(bh)) {
 				get_bh(bh);
 				lock_buffer(bh);
-				bh->b_end_io = end_buffer_read_sync;
 				apfs_submit_bh(REQ_OP_READ, 0, bh);
 				wait_on_buffer(bh);
 				if (!buffer_uptodate(bh)) {

lzfse/lzfse_internal.h

@@ -70,7 +70,9 @@ typedef int64_t lzfse_offset;
 typedef int32_t lzfse_offset;
 #endif
 
+#if !defined(__APPLE__) || !defined(__MACH__)
 typedef uint64_t uintmax_t;
+#endif
 
 /*! @abstract Decoder state object for lzfse compressed blocks. */
 typedef struct {

spaceman.c

@@ -308,6 +308,7 @@ static int apfs_write_single_ip_bitmap(struct super_block *sb, char *bitmap, u32
 	err = apfs_transaction_join(sb, bh);
 	if (err)
 		goto fail;
+	put_bh(bh);
 	bh = NULL;
 
 	spaceman->sm_ip_bmaps[idx].dirty = false;

super.c

@@ -17,6 +17,7 @@
 #include "version.h"
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(7, 0, 0)
 #include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #endif
 
 #define APFS_MODULE_ID_STRING	"linux-apfs by eafer (" GIT_COMMIT ")"
@@ -342,8 +343,8 @@ static int apfs_read_main_super(struct super_block *sb)
 	xid = le64_to_cpu(msb_raw->nx_o.o_xid);
 	for (i = 0; i < desc_blocks; ++i) {
 		struct apfs_nx_superblock *desc_raw;
+		bool ignore;
 
-		brelse(desc_bh);
 		desc_bh = apfs_sb_bread(sb, desc_base + i);
 		if (!desc_bh) {
 			apfs_err(sb, "unable to read checkpoint descriptor");
@@ -351,18 +352,24 @@ static int apfs_read_main_super(struct super_block *sb)
 		}
 		desc_raw = (struct apfs_nx_superblock *)desc_bh->b_data;
 
+		ignore = false;
 		if (le32_to_cpu(desc_raw->nx_magic) != APFS_NX_MAGIC)
-			continue; /* Not a superblock */
+			ignore = true; /* Not a superblock */
 		if (le64_to_cpu(desc_raw->nx_o.o_xid) <= xid)
-			continue; /* Old */
+			ignore = true; /* Old */
 		if (!apfs_obj_verify_csum(sb, desc_bh))
-			continue; /* Corrupted */
+			ignore = true; /* Corrupted */
 
-		xid = le64_to_cpu(desc_raw->nx_o.o_xid);
-		msb_raw = desc_raw;
-		bno = desc_base + i;
-		brelse(bh);
-		bh = desc_bh;
+		if (ignore) {
+			brelse(desc_bh);
+		} else {
+			xid = le64_to_cpu(desc_raw->nx_o.o_xid);
+			msb_raw = desc_raw;
+			bno = desc_base + i;
+			brelse(bh);
+			bh = desc_bh;
+		}
+		desc_raw = NULL;
 		desc_bh = NULL;
 	}
 
@@ -1203,6 +1210,7 @@ enum {
 	Opt_readwrite, Opt_cknodes, Opt_uid, Opt_gid, Opt_vol, Opt_snap, Opt_tier2, Opt_err,
 };
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(7, 0, 0)
 static const match_table_t tokens = {
 	{Opt_readwrite, "readwrite"},
 	{Opt_cknodes, "cknodes"},
@@ -1213,6 +1221,18 @@ static const match_table_t tokens = {
 	{Opt_tier2, "tier2=%s"},
 	{Opt_err, NULL}
 };
+#else
+static const struct fs_parameter_spec apfs_param_spec[] = {
+	fsparam_flag	("readwrite",	Opt_readwrite),
+	fsparam_flag	("cknodes",	Opt_cknodes),
+	fsparam_uid	("uid",		Opt_uid),
+	fsparam_gid	("gid",		Opt_gid),
+	fsparam_u32	("vol",		Opt_vol),
+	fsparam_string	("snap",	Opt_snap),
+	fsparam_string	("tier2",	Opt_tier2),
+	{}
+};
+#endif
 
 /**
  * apfs_set_nx_flags - Set the mount flags for the container, if allowed
@@ -1252,6 +1272,26 @@ static void parse_options_set_flags(struct super_block *sb, struct apfs_sb_info
 	}
 }
 
+/**
+ * apfs_set_default_opts - Set the default mount options in the superblock
+ * @sbi: in-memory superblock info
+ */
+static void apfs_set_default_opts(struct apfs_sb_info *sbi)
+{
+	sbi->s_vol_nr = 0;
+	sbi->s_snap_name = NULL;
+	sbi->s_tier2_path = NULL;
+	sbi->s_mount_opt = 0;
+#ifdef CONFIG_APFS_RW_ALWAYS
+	/* Still risky, but some packagers want writable mounts by default */
+	sbi->s_mount_opt |= APFS_READWRITE;
+#endif
+	sbi->s_uid = INVALID_UID;
+	sbi->s_gid = INVALID_GID;
+}
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(7, 0, 0)
+
 /*
  * Many of the parse_options() functions in other file systems return 0
  * on error. This one returns an error code, and 0 on success.
@@ -1266,20 +1306,6 @@ static int parse_options(struct apfs_sb_info *sbi, char *options)
 	int option;
 	int err = 0;
 
-	/* Set default values before parsing */
-	sbi->s_vol_nr = 0;
-	sbi->s_snap_name = NULL;
-	sbi->s_tier2_path = NULL;
-	sbi->s_mount_opt = 0;
-
-#ifdef CONFIG_APFS_RW_ALWAYS
-	/* Still risky, but some packagers want writable mounts by default */
-	sbi->s_mount_opt |= APFS_READWRITE;
-#endif
-
-	sbi->s_uid = INVALID_UID;
-	sbi->s_gid = INVALID_GID;
-
 	if (!options)
 		return 0;
 
@@ -1351,6 +1377,62 @@ static int parse_options(struct apfs_sb_info *sbi, char *options)
 	return 0;
 }
 
+#else /* LINUX_VERSION_CODE < KERNEL_VERSION(7, 0, 0) */
+
+static int apfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	struct apfs_sb_info *sbi = fc->s_fs_info;
+	int opt;
+	struct fs_parse_result result;
+
+	opt = fs_parse(fc, apfs_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_readwrite:
+		/*
+		 * Write support is not safe yet, so keep it disabled
+		 * unless the user requests it explicitly.
+		 */
+		sbi->s_mount_opt |= APFS_READWRITE;
+		break;
+	case Opt_cknodes:
+		/*
+		 * Right now, node checksums are too costly to enable
+		 * by default.  TODO: try to improve this.
+		 */
+		sbi->s_mount_opt |= APFS_CHECK_NODES;
+		break;
+	case Opt_uid:
+		sbi->s_uid = result.uid;
+		break;
+	case Opt_gid:
+		sbi->s_gid = result.gid;
+		break;
+	case Opt_vol:
+		sbi->s_vol_nr = result.uint_32;
+		break;
+	case Opt_snap:
+		kfree(sbi->s_snap_name);
+		sbi->s_snap_name = kstrdup(param->string, GFP_KERNEL);
+		if (!sbi->s_snap_name)
+			return -ENOMEM;
+		break;
+	case Opt_tier2:
+		kfree(sbi->s_tier2_path);
+		sbi->s_tier2_path = kstrdup(param->string, GFP_KERNEL);
+		if (!sbi->s_tier2_path)
+			return -ENOMEM;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+#endif
+
 /**
  * apfs_check_nx_features - Check for unsupported features in the container
  * @sb: superblock structure
@@ -1924,6 +2006,7 @@ static struct dentry *apfs_mount(struct file_system_type *fs_type, int flags,
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
 		return ERR_PTR(-ENOMEM);
+	apfs_set_default_opts(sbi);
 	/* Set up the fields that sget() will need to id the superblock */
 	error = parse_options(sbi, data);
 	if (error)
@@ -2123,16 +2206,11 @@ static void apfs_free_fc(struct fs_context *fc)
 	}
 }
 
-static int apfs_parse_monolithic(struct fs_context *fc, void *data)
-{
-	return parse_options(fc->s_fs_info, data);
-}
-
 static const struct fs_context_operations apfs_context_ops = {
 	.get_tree		= apfs_get_tree,
 	.reconfigure		= apfs_reconfigure,
 	.free			= apfs_free_fc,
-	.parse_monolithic	= apfs_parse_monolithic,
+	.parse_param		= apfs_parse_param,
 };
 
 static int apfs_init_fs_context(struct fs_context *fc)
@@ -2143,6 +2221,7 @@ static int apfs_init_fs_context(struct fs_context *fc)
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
 		return -ENOMEM;
+	apfs_set_default_opts(sbi);
 
 	fc->s_fs_info = sbi;
 	fc->ops = &apfs_context_ops;

transaction.c

@@ -640,6 +640,9 @@ static int apfs_write_ephemeral_objects(struct super_block *sb)
 		}
 		data_len += obj_blkcnt;
 	}
+	brelse(cpm_bh);
+	cpm = NULL;
+	cpm_bh = NULL;
 
 	/*
 	 * The checkpoint superblock can't be set until the very end of the
@@ -721,7 +724,6 @@ static int apfs_transaction_commit_nx(struct super_block *sb)
 		clear_buffer_dirty(bh);
 		get_bh(bh);
 		lock_buffer(bh);
-		bh->b_end_io = end_buffer_write_sync;
 		apfs_submit_bh(REQ_OP_WRITE, REQ_SYNC, bh);
 	}
 	list_for_each_entry_safe(bhi, tmp, &nx_trans->t_buffers, list) {