Merge tag 'exfat-for-6.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat

Pull exfat updates from Namjae Jeon:

- Clean-up unnecessary codes as ->valid_size is supported

- buffered-IO fallback is no longer needed when using direct-IO

- Move ->valid_size extension from mmap to ->page_mkwrite. This
improves the overhead caused by unnecessary zero-out during mmap.

- Fix memleaks from exfat_load_bitmap() and exfat_create_upcase_table()

- Add sops->shutdown and ioctl

- Add Yuezhang Mo as a reviwer

* tag 'exfat-for-6.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat:
MAINTAINERS: exfat: add myself as reviewer
exfat: resolve memory leak from exfat_create_upcase_table()
exfat: move extend valid_size into ->page_mkwrite()
exfat: fix memory leak in exfat_load_bitmap()
exfat: Implement sops->shutdown and ioctl
exfat: do not fallback to buffered write
exfat: drop ->i_size_ondisk

MAINTAINERS

···
 EXFAT FILE SYSTEM
 M:	Namjae Jeon <[e-mail]>
 M:	Sungjong Seo <[e-mail]>
+R:	Yuezhang Mo <[e-mail]>
 L:	[e-mail]
 S:	Maintained
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat.git

fs/exfat/balloc.c

···
 				return -EIO;
 
 			type = exfat_get_entry_type(ep);
-			if (type == TYPE_UNUSED)
-				break;
-			if (type != TYPE_BITMAP)
-				continue;
-			if (ep->dentry.bitmap.flags == 0x0) {
+			if (type == TYPE_BITMAP &&
+			    ep->dentry.bitmap.flags == 0x0) {
 				int err;
 
 				err = exfat_allocate_bitmap(sb, ep);
···
 				return err;
 			}
 			brelse(bh);
+
+			if (type == TYPE_UNUSED)
+				return -EINVAL;
 		}
 
 		if (exfat_get_next_cluster(sb, &clu.dir))

fs/exfat/exfat_fs.h

···
 #include <linux/ratelimit.h>
 #include <linux/nls.h>
 #include <linux/blkdev.h>
+#include <uapi/linux/exfat.h>
 
 #define EXFAT_ROOT_INO		1
 
···
 #define DIR_CACHE_SIZE		\
 	(DIV_ROUND_UP(EXFAT_DEN_TO_B(ES_MAX_ENTRY_NUM), SECTOR_SIZE) + 1)
 
+/* Superblock flags */
+#define EXFAT_FLAGS_SHUTDOWN	1
+
 struct exfat_dentry_namebuf {
 	char *lfn;
 	int lfnbuf_len; /* usually MAX_UNINAME_BUF_SIZE */
···
 	unsigned int clu_srch_ptr; /* cluster search pointer */
 	unsigned int used_clusters; /* number of used clusters */
 
+	unsigned long s_exfat_flags; /* Exfat superblock flags */
+
 	struct mutex s_lock; /* superblock lock */
 	struct mutex bitmap_lock; /* bitmap lock */
 	struct exfat_mount_options options;
···
 	/* for avoiding the race between alloc and free */
 	unsigned int cache_valid_id;
 
-	/*
-	 * NOTE: i_size_ondisk is 64bits, so must hold ->inode_lock to access.
-	 * physically allocated size.
-	 */
-	loff_t i_size_ondisk;
-	/* block-aligned i_size (used in cont_write_begin) */
-	loff_t i_size_aligned;
 	/* on-disk position of directory entry or 0 */
 	loff_t i_pos;
 	loff_t valid_size;
···
 static inline struct exfat_inode_info *EXFAT_I(struct inode *inode)
 {
 	return container_of(inode, struct exfat_inode_info, vfs_inode);
+}
+
+static inline int exfat_forced_shutdown(struct super_block *sb)
+{
+	return test_bit(EXFAT_FLAGS_SHUTDOWN, &EXFAT_SB(sb)->s_exfat_flags);
 }
 
 /*
···
 	return clus >= EXFAT_FIRST_CLUSTER && clus < sbi->num_clusters;
 }
 
+static inline loff_t exfat_ondisk_size(const struct inode *inode)
+{
+	return ((loff_t)inode->i_blocks) << 9;
+}
+
 /* super.c */
 int exfat_set_volume_dirty(struct super_block *sb);
 int exfat_clear_volume_dirty(struct super_block *sb);
···
 long exfat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
 long exfat_compat_ioctl(struct file *filp, unsigned int cmd,
 				unsigned long arg);
+int exfat_force_shutdown(struct super_block *sb, u32 flags);
 
 /* namei.c */
 extern const struct dentry_operations exfat_dentry_ops;

fs/exfat/file.c

···
 	if (ret)
 		return ret;
 
-	num_clusters = EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi);
+	num_clusters = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi);
 	new_num_clusters = EXFAT_B_TO_CLU_ROUND_UP(size, sbi);
 
 	if (new_num_clusters == num_clusters)
···
 	/* Expanded range not zeroed, do not update valid_size */
 	i_size_write(inode, size);
 
-	ei->i_size_aligned = round_up(size, sb->s_blocksize);
-	ei->i_size_ondisk = ei->i_size_aligned;
 	inode->i_blocks = round_up(size, sbi->cluster_size) >> 9;
 	mark_inode_dirty(inode);
 
···
 	exfat_set_volume_dirty(sb);
 
 	num_clusters_new = EXFAT_B_TO_CLU_ROUND_UP(i_size_read(inode), sbi);
-	num_clusters_phys = EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi);
+	num_clusters_phys = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi);
 
 	exfat_chain_set(&clu, ei->start_clu, num_clusters_phys, ei->flags);
 
···
 	struct super_block *sb = inode->i_sb;
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	struct exfat_inode_info *ei = EXFAT_I(inode);
-	unsigned int blocksize = i_blocksize(inode);
-	loff_t aligned_size;
 	int err;
 
 	mutex_lock(&sbi->s_lock);
···
 
 	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9;
 write_size:
-	aligned_size = i_size_read(inode);
-	if (aligned_size & (blocksize - 1)) {
-		aligned_size |= (blocksize - 1);
-		aligned_size++;
-	}
-
-	if (ei->i_size_ondisk > i_size_read(inode))
-		ei->i_size_ondisk = aligned_size;
-
-	if (ei->i_size_aligned > i_size_read(inode))
-		ei->i_size_aligned = aligned_size;
 	mutex_unlock(&sbi->s_lock);
 }
 
···
 	struct inode *inode = dentry->d_inode;
 	unsigned int ia_valid;
 	int error;
+
+	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
+		return -EIO;
 
 	if ((attr->ia_valid & ATTR_SIZE) &&
 	    attr->ia_size > i_size_read(inode)) {
···
 	return 0;
 }
 
+static int exfat_ioctl_shutdown(struct super_block *sb, unsigned long arg)
+{
+	u32 flags;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (get_user(flags, (__u32 __user *)arg))
+		return -EFAULT;
+
+	return exfat_force_shutdown(sb, flags);
+}
+
 long exfat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
···
 		return exfat_ioctl_get_attributes(inode, user_attr);
 	case FAT_IOCTL_SET_ATTRIBUTES:
 		return exfat_ioctl_set_attributes(filp, user_attr);
+	case EXFAT_IOC_SHUTDOWN:
+		return exfat_ioctl_shutdown(inode->i_sb, arg);
 	case FITRIM:
 		return exfat_ioctl_fitrim(inode, arg);
 	default:
···
 	struct inode *inode = filp->f_mapping->host;
 	int err;
 
+	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
+		return -EIO;
+
 	err = __generic_file_fsync(filp, start, end, datasync);
 	if (err)
 		return err;
···
 	return blkdev_issue_flush(inode->i_sb->s_bdev);
 }
 
-static int exfat_file_zeroed_range(struct file *file, loff_t start, loff_t end)
+static int exfat_extend_valid_size(struct file *file, loff_t new_valid_size)
 {
 	int err;
+	loff_t pos;
 	struct inode *inode = file_inode(file);
+	struct exfat_inode_info *ei = EXFAT_I(inode);
 	struct address_space *mapping = inode->i_mapping;
 	const struct address_space_operations *ops = mapping->a_ops;
 
-	while (start < end) {
-		u32 zerofrom, len;
+	pos = ei->valid_size;
+	while (pos < new_valid_size) {
+		u32 len;
 		struct folio *folio;
 
-		zerofrom = start & (PAGE_SIZE - 1);
-		len = PAGE_SIZE - zerofrom;
-		if (start + len > end)
-			len = end - start;
+		len = PAGE_SIZE - (pos & (PAGE_SIZE - 1));
+		if (pos + len > new_valid_size)
+			len = new_valid_size - pos;
 
-		err = ops->write_begin(file, mapping, start, len, &folio, NULL);
+		err = ops->write_begin(file, mapping, pos, len, &folio, NULL);
 		if (err)
 			goto out;
 
-		folio_zero_range(folio, offset_in_folio(folio, start), len);
-
-		err = ops->write_end(file, mapping, start, len, len, folio, NULL);
+		err = ops->write_end(file, mapping, pos, len, len, folio, NULL);
 		if (err < 0)
 			goto out;
-		start += len;
+		pos += len;
 
 		balance_dirty_pages_ratelimited(mapping);
 		cond_resched();
···
 		goto unlock;
 
 	if (pos > valid_size) {
-		ret = exfat_file_zeroed_range(file, valid_size, pos);
+		ret = exfat_extend_valid_size(file, pos);
 		if (ret < 0 && ret != -ENOSPC) {
 			exfat_err(inode->i_sb,
 				"write: fail to zero from %llu to %llu(%zd)",
···
 	return ret;
 }
 
-static int exfat_file_mmap(struct file *file, struct vm_area_struct *vma)
+static vm_fault_t exfat_page_mkwrite(struct vm_fault *vmf)
 {
-	int ret;
+	int err;
+	struct vm_area_struct *vma = vmf->vma;
+	struct file *file = vma->vm_file;
 	struct inode *inode = file_inode(file);
 	struct exfat_inode_info *ei = EXFAT_I(inode);
-	loff_t start = ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
-	loff_t end = min_t(loff_t, i_size_read(inode),
+	loff_t start, end;
+
+	if (!inode_trylock(inode))
+		return VM_FAULT_RETRY;
+
+	start = ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+	end = min_t(loff_t, i_size_read(inode),
 			start + vma->vm_end - vma->vm_start);
 
-	if ((vma->vm_flags & VM_WRITE) && ei->valid_size < end) {
-		ret = exfat_file_zeroed_range(file, ei->valid_size, end);
-		if (ret < 0) {
-			exfat_err(inode->i_sb,
-				  "mmap: fail to zero from %llu to %llu(%d)",
-				  start, end, ret);
-			return ret;
+	if (ei->valid_size < end) {
+		err = exfat_extend_valid_size(file, end);
+		if (err < 0) {
+			inode_unlock(inode);
+			return vmf_fs_error(err);
 		}
 	}
 
-	return generic_file_mmap(file, vma);
+	inode_unlock(inode);
+
+	return filemap_page_mkwrite(vmf);
+}
+
+static const struct vm_operations_struct exfat_file_vm_ops = {
+	.fault		= filemap_fault,
+	.map_pages	= filemap_map_pages,
+	.page_mkwrite	= exfat_page_mkwrite,
+};
+
+static int exfat_file_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	file_accessed(file);
+	vma->vm_ops = &exfat_file_vm_ops;
+	return 0;
 }
 
 const struct file_operations exfat_file_operations = {

fs/exfat/inode.c

···
 {
 	int ret;
 
+	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
+		return -EIO;
+
 	mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock);
 	ret = __exfat_write_inode(inode, wbc->sync_mode == WB_SYNC_ALL);
 	mutex_unlock(&EXFAT_SB(inode->i_sb)->s_lock);
···
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	struct exfat_inode_info *ei = EXFAT_I(inode);
 	unsigned int local_clu_offset = clu_offset;
-	unsigned int num_to_be_allocated = 0, num_clusters = 0;
+	unsigned int num_to_be_allocated = 0, num_clusters;
 
-	if (ei->i_size_ondisk > 0)
-		num_clusters =
-			EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi);
+	num_clusters = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi);
 
 	if (clu_offset >= num_clusters)
 		num_to_be_allocated = clu_offset - num_clusters + 1;
···
 	return 0;
 }
 
-static int exfat_map_new_buffer(struct exfat_inode_info *ei,
-		struct buffer_head *bh, loff_t pos)
-{
-	if (buffer_delay(bh) && pos > ei->i_size_aligned)
-		return -EIO;
-	set_buffer_new(bh);
-
-	/*
-	 * Adjust i_size_aligned if i_size_ondisk is bigger than it.
-	 */
-	if (ei->i_size_ondisk > ei->i_size_aligned)
-		ei->i_size_aligned = ei->i_size_ondisk;
-	return 0;
-}
-
 static int exfat_get_block(struct inode *inode, sector_t iblock,
 		struct buffer_head *bh_result, int create)
 {
···
 	sector_t last_block;
 	sector_t phys = 0;
 	sector_t valid_blks;
-	loff_t pos;
 
 	mutex_lock(&sbi->s_lock);
 	last_block = EXFAT_B_TO_BLK_ROUND_UP(i_size_read(inode), sb);
···
 	mapped_blocks = sbi->sect_per_clus - sec_offset;
 	max_blocks = min(mapped_blocks, max_blocks);
 
-	pos = EXFAT_BLK_TO_B((iblock + 1), sb);
-	if ((create && iblock >= last_block) || buffer_delay(bh_result)) {
-		if (ei->i_size_ondisk < pos)
-			ei->i_size_ondisk = pos;
-	}
-
 	map_bh(bh_result, sb, phys);
 	if (buffer_delay(bh_result))
 		clear_buffer_delay(bh_result);
···
 		}
 
 		/* The area has not been written, map and mark as new. */
-		err = exfat_map_new_buffer(ei, bh_result, pos);
-		if (err) {
-			exfat_fs_error(sb,
-					"requested for bmap out of range(pos : (%llu) > i_size_aligned(%llu)\n",
-					pos, ei->i_size_aligned);
-			goto unlock_ret;
-		}
+		set_buffer_new(bh_result);
 
 		ei->valid_size = EXFAT_BLK_TO_B(iblock + max_blocks, sb);
 		mark_inode_dirty(inode);
···
 			 * The block has been partially written,
 			 * zero the unwritten part and map the block.
 			 */
-			loff_t size, off;
+			loff_t size, off, pos;
 
 			max_blocks = 1;
 
···
 			if (!bh_result->b_folio)
 				goto done;
 
-			pos -= sb->s_blocksize;
+			pos = EXFAT_BLK_TO_B(iblock, sb);
 			size = ei->valid_size - pos;
 			off = pos & (PAGE_SIZE - 1);
 
···
 static int exfat_writepages(struct address_space *mapping,
 		struct writeback_control *wbc)
 {
+	if (unlikely(exfat_forced_shutdown(mapping->host->i_sb)))
+		return -EIO;
+
 	return mpage_writepages(mapping, wbc, exfat_get_block);
 }
 
···
 {
 	int ret;
 
+	if (unlikely(exfat_forced_shutdown(mapping->host->i_sb)))
+		return -EIO;
+
 	ret = block_write_begin(mapping, pos, len, foliop, exfat_get_block);
 
 	if (ret < 0)
···
 	int err;
 
 	err = generic_write_end(file, mapping, pos, len, copied, folio, fsdata);
-
-	if (ei->i_size_aligned < i_size_read(inode)) {
-		exfat_fs_error(inode->i_sb,
-			"invalid size(size(%llu) > aligned(%llu)\n",
-			i_size_read(inode), ei->i_size_aligned);
-		return -EIO;
-	}
-
 	if (err < len)
 		exfat_write_failed(mapping, pos+len);
 
···
 	int rw = iov_iter_rw(iter);
 	ssize_t ret;
 
-	if (rw == WRITE) {
-		/*
-		 * FIXME: blockdev_direct_IO() doesn't use ->write_begin(),
-		 * so we need to update the ->i_size_aligned to block boundary.
-		 *
-		 * But we must fill the remaining area or hole by nul for
-		 * updating ->i_size_aligned
-		 *
-		 * Return 0, and fallback to normal buffered write.
-		 */
-		if (EXFAT_I(inode)->i_size_aligned < size)
-			return 0;
-	}
-
 	/*
 	 * Need to use the DIO_LOCKING for avoiding the race
 	 * condition of exfat_get_block() and ->truncate().
···
 	} else
 		size = pos + ret;
 
-	/* zero the unwritten part in the partially written block */
-	if (rw == READ && pos < ei->valid_size && ei->valid_size < size) {
+	if (rw == WRITE) {
+		/*
+		 * If the block had been partially written before this write,
+		 * ->valid_size will not be updated in exfat_get_block(),
+		 * update it here.
+		 */
+		if (ei->valid_size < size) {
+			ei->valid_size = size;
+			mark_inode_dirty(inode);
+		}
+	} else if (pos < ei->valid_size && ei->valid_size < size) {
+		/* zero the unwritten part in the partially written block */
 		iov_iter_revert(iter, size - ei->valid_size);
 		iov_iter_zero(size - ei->valid_size, iter);
 	}
···
 	}
 
 	i_size_write(inode, size);
-
-	/* ondisk and aligned size should be aligned with block size */
-	if (size & (inode->i_sb->s_blocksize - 1)) {
-		size |= (inode->i_sb->s_blocksize - 1);
-		size++;
-	}
-
-	ei->i_size_aligned = size;
-	ei->i_size_ondisk = size;
 
 	exfat_save_attr(inode, info->attr);
 

fs/exfat/namei.c

···
 
 		/* directory inode should be updated in here */
 		i_size_write(inode, size);
-		ei->i_size_ondisk += sbi->cluster_size;
-		ei->i_size_aligned += sbi->cluster_size;
 		ei->valid_size += sbi->cluster_size;
 		ei->flags = p_dir->flags;
 		inode->i_blocks += sbi->cluster_size >> 9;
···
 	loff_t i_pos;
 	int err;
 	loff_t size = i_size_read(dir);
+
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return -EIO;
 
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_set_volume_dirty(sb);
···
 	struct exfat_entry_set_cache es;
 	int entry, err = 0;
 
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return -EIO;
+
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_chain_dup(&cdir, &ei->dir);
 	entry = ei->entry;
···
 	loff_t i_pos;
 	int err;
 	loff_t size = i_size_read(dir);
+
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return -EIO;
 
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_set_volume_dirty(sb);
···
 	struct exfat_entry_set_cache es;
 	int entry, err;
 
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return -EIO;
+
 	mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock);
 
 	exfat_chain_dup(&cdir, &ei->dir);
···
 	struct super_block *sb = inode->i_sb;
 	struct exfat_entry_set_cache old_es, new_es;
 	int sync = IS_DIRSYNC(inode);
+
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return -EIO;
 
 	num_new_entries = exfat_calc_num_entries(p_uniname);
 	if (num_new_entries < 0)

fs/exfat/nls.c

···
 				le32_to_cpu(ep->dentry.upcase.checksum));
 
 			brelse(bh);
-			if (ret && ret != -EIO)
+			if (ret && ret != -EIO) {
+				/* free memory from exfat_load_upcase_table call */
+				exfat_free_upcase_table(sbi);
 				goto load_default;
+			}
 
 			/* load successfully */
 			return ret;

fs/exfat/super.c

···
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	int err = 0;
 
+	if (unlikely(exfat_forced_shutdown(sb)))
+		return 0;
+
 	if (!wait)
 		return 0;
 
···
 	return 0;
 }
 
+int exfat_force_shutdown(struct super_block *sb, u32 flags)
+{
+	int ret;
+	struct exfat_sb_info *sbi = sb->s_fs_info;
+	struct exfat_mount_options *opts = &sbi->options;
+
+	if (exfat_forced_shutdown(sb))
+		return 0;
+
+	switch (flags) {
+	case EXFAT_GOING_DOWN_DEFAULT:
+	case EXFAT_GOING_DOWN_FULLSYNC:
+		ret = bdev_freeze(sb->s_bdev);
+		if (ret)
+			return ret;
+		bdev_thaw(sb->s_bdev);
+		set_bit(EXFAT_FLAGS_SHUTDOWN, &sbi->s_exfat_flags);
+		break;
+	case EXFAT_GOING_DOWN_NOSYNC:
+		set_bit(EXFAT_FLAGS_SHUTDOWN, &sbi->s_exfat_flags);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (opts->discard)
+		opts->discard = 0;
+	return 0;
+}
+
+static void exfat_shutdown(struct super_block *sb)
+{
+	exfat_force_shutdown(sb, EXFAT_GOING_DOWN_NOSYNC);
+}
+
 static struct inode *exfat_alloc_inode(struct super_block *sb)
 {
 	struct exfat_inode_info *ei;
···
 	.sync_fs	= exfat_sync_fs,
 	.statfs		= exfat_statfs,
 	.show_options	= exfat_show_options,
+	.shutdown	= exfat_shutdown,
 };
 
 enum {
···
 
 	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9;
 	ei->i_pos = ((loff_t)sbi->root_dir << 32) | 0xffffffff;
-	ei->i_size_aligned = i_size_read(inode);
-	ei->i_size_ondisk = i_size_read(inode);
 
 	exfat_save_attr(inode, EXFAT_ATTR_SUBDIR);
 	ei->i_crtime = simple_inode_init_ts(inode);

include/uapi/linux/exfat.h

···
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+/*
+ * Copyright (C) 2024 Unisoc Technologies Co., Ltd.
+ */
+
+#ifndef _UAPI_LINUX_EXFAT_H
+#define _UAPI_LINUX_EXFAT_H
+#include <linux/types.h>
+#include <linux/ioctl.h>
+
+/*
+ * exfat-specific ioctl commands
+ */
+
+#define EXFAT_IOC_SHUTDOWN _IOR('X', 125, __u32)
+
+/*
+ * Flags used by EXFAT_IOC_SHUTDOWN
+ */
+
+#define EXFAT_GOING_DOWN_DEFAULT	0x0	/* default with full sync */
+#define EXFAT_GOING_DOWN_FULLSYNC	0x1     /* going down with full sync*/
+#define EXFAT_GOING_DOWN_NOSYNC         0x2     /* going down */
+
+#endif /* _UAPI_LINUX_EXFAT_H */