[PATCH v0 0/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

[PATCH v0 0/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

[David Timber]

This is part of my exFAT fragmentation optimisation project.

As exFAT is fundamentally based on linked-list structure, cluster
allocation is not a cheap operation - it involves walking the cluster
chain as well as bitmap so it's not particularly easy on flash devices.

The purpose of VDL is in part addressing the issue by allowing users
to clear the contents of files without changing the cluster layout by
only updating VDL which is O(1) op. I.e. reduced flash wear.

The fallocate support should see applications in recording devices like
dashcams, IP cameras and DVRs(digital video recorders). Such devices
typically implement FIFO-style file rotation. Also, the technique
could be potentially utilise to eliminate the chance of fragmentation
if the device preallocates clusters for large media files.

The `fallocate -d ...` and `cp --sparse=always ...`
commands(util-linux) can be used on the files in exFAT to detect zeros
leading up to EOF and update the VDL accordingly.

David Timber (1):
  exfat: add limited FALLOC_FL_ZERO_RANGE support

 fs/exfat/file.c | 75 +++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 64 insertions(+), 11 deletions(-)

-- 
2.53.0.1.ga224b40d3f.dirty

[PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

[David Timber]

As means of truncating the VDL of a regular file while maintaining the
layout of the allocated clusters, allow the use fallocate mode
FALLOC_FL_ZERO_RANGE with the range that covers EOF, with the support
of optional FALLOC_FL_KEEP_SIZE flag.

To reset the VDL to 0, userspace may use fallocate() like so:

	fallocate(fd, FALLOC_FL_ZERO_RANGE|FALLOC_FL_KEEP_SIZE,
                  lseek(fd, 0, SEEK_END));

FALLOC_FL_KEEP_SIZE flag is for multiple users to guard the file from
TOCTOU conditions. Without the flag, the behaviour is the same as
FALLOC_FL_ZERO_RANGE.

Signed-off-by: David Timber <dxdt@dev.snart.me>
---
 fs/exfat/file.c | 75 +++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 64 insertions(+), 11 deletions(-)

diff --git a/fs/exfat/file.c b/fs/exfat/file.c
index 2daf0dbabb24..cff61d1a9f68 100644
--- a/fs/exfat/file.c
+++ b/fs/exfat/file.c
@@ -36,7 +36,8 @@ static int exfat_cont_expand(struct inode *inode, loff_t size)
 	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)
+	WARN_ON(new_num_clusters < num_clusters);
+	if (new_num_clusters <= num_clusters)
 		goto out;
 
 	if (num_clusters) {
@@ -94,35 +95,87 @@ static int exfat_cont_expand(struct inode *inode, loff_t size)
 /*
  * Preallocate space for a file. This implements exfat's fallocate file
  * operation, which gets called from sys_fallocate system call. User space
- * requests len bytes at offset. In contrary to fat, we only support
- * FALLOC_FL_ALLOCATE_RANGE because by leaving the valid data length(VDL)
- * field, it is unnecessary to zero out the newly allocated clusters.
+ * requests len bytes at offset.
+ *
+ * In contrary to fat, FALLOC_FL_ALLOCATE_RANGE can be done without zeroing out
+ * the newly allocated clusters by leaving the valid data length(VDL) field
+ * unchanged.
+ *
+ * Due to the inherent limitation of the VDL scheme, FALLOC_FL_ZERO_RANGE is
+ * only possible when the requested range covers EOF.
  */
 static long exfat_fallocate(struct file *file, int mode,
 			  loff_t offset, loff_t len)
 {
 	struct inode *inode = file->f_mapping->host;
-	loff_t newsize = offset + len;
+	loff_t newsize, isize;
 	int err = 0;
 
 	/* No support for other modes */
-	if (mode != FALLOC_FL_ALLOCATE_RANGE)
+	switch (mode) {
+	case FALLOC_FL_ALLOCATE_RANGE:
+	case FALLOC_FL_ZERO_RANGE:
+	case FALLOC_FL_ZERO_RANGE|FALLOC_FL_KEEP_SIZE:
+		break;
+	default:
 		return -EOPNOTSUPP;
+	}
 
 	/* No support for dir */
 	if (!S_ISREG(inode->i_mode))
-		return -EOPNOTSUPP;
+		return mode & FALLOC_FL_ZERO_RANGE ? -EINVAL : -EOPNOTSUPP;
 
 	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
 		return -EIO;
 
 	inode_lock(inode);
 
-	if (newsize <= i_size_read(inode))
-		goto error;
+	newsize = offset + len;
+	isize = i_size_read(inode);
+
+	if (mode & FALLOC_FL_ZERO_RANGE) {
+		struct exfat_inode_info *ei = EXFAT_I(inode);
+		loff_t saved_validsize = ei->valid_size;
+
+		/* The requested range must span to or past EOF */
+		if (newsize < isize) {
+			err = -EOPNOTSUPP;
+			goto error;
+		}
+
+		/* valid_size can only be truncated */
+		if (offset < ei->valid_size)
+			ei->valid_size = offset;
+		/* If offset >= ei->valid_size, the range is already zeroed so that'd be no-op */
+
+		if (!(mode & FALLOC_FL_KEEP_SIZE) && isize < newsize) {
+			err = exfat_cont_expand(inode, newsize);
+			if (err) {
+				/* inode unchanged - revert valid_size */
+				ei->valid_size = saved_validsize;
+				goto error;
+			}
+			/* inode invalidated in exfat_cont_expand() */
+		} else {
+			/* update inode */
+			inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
 
-	/* This is just an expanding truncate */
-	err = exfat_cont_expand(inode, newsize);
+			mark_inode_dirty(inode);
+
+			if (IS_SYNC(inode))
+				return write_inode_now(inode, 1);
+		}
+
+		/* drop cache after the new valid_size */
+		if (ei->valid_size != saved_validsize)
+			truncate_pagecache(inode, ei->valid_size);
+	} else { /* mode == FALLOC_FL_ALLOCATE_RANGE */
+		if (newsize <= isize)
+			goto error;
+
+		/* This is just an expanding truncate */
+		err = exfat_cont_expand(inode, newsize);
+	}
 
 error:
 	inode_unlock(inode);
-- 
2.53.0.1.ga224b40d3f.dirty

[syzbot ci] Re: exfat: add limited FALLOC_FL_ZERO_RANGE support

[syzbot ci]

syzbot ci has tested the following series

[v0] exfat: add limited FALLOC_FL_ZERO_RANGE support
https://lore.kernel.org/all/20260319043553.301185-1-dxdt@dev.snart.me
* [PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

and found the following issue:
WARNING: lock held when returning to user space in exfat_fallocate

Full report is available here:
https://ci.syzbot.org/series/a58a79c9-a4e7-40b7-a5d3-8d4da244a00c

***

WARNING: lock held when returning to user space in exfat_fallocate

tree:      linux-next
URL:       https://kernel.googlesource.com/pub/scm/linux/kernel/git/next/linux-next
base:      b5d083a3ed1e2798396d5e491432e887da8d4a06
arch:      amd64
compiler:  Debian clang version 21.1.8 (++20251221033036+2078da43e25a-1~exp1~20251221153213.50), Debian LLD 21.1.8
config:    https://ci.syzbot.org/builds/6a43cfef-5b9a-4b39-a6b0-7d34bfb0f7b7/config
C repro:   https://ci.syzbot.org/findings/87d27fd2-e5b7-4174-8faf-3b746f7aa18d/c_repro
syz repro: https://ci.syzbot.org/findings/87d27fd2-e5b7-4174-8faf-3b746f7aa18d/syz_repro

exFAT-fs (loop0): failed to load upcase table (idx : 0x0000fd4f, chksum : 0x395e47cf, utbl_chksum : 0xe619d30d)
exFAT-fs (loop0): valid_size(150994954) is greater than size(10)
================================================
WARNING: lock held when returning to user space!
syzkaller #0 Not tainted
------------------------------------------------
syz.0.17/5982 is leaving the kernel with locks still held!
1 lock held by syz.0.17/5982:
 #0: ffff8881bc75c110 (&sb->s_type->i_mutex_key#24){+.+.}-{4:4}, at: inode_lock include/linux/fs.h:1028 [inline]
 #0: ffff8881bc75c110 (&sb->s_type->i_mutex_key#24){+.+.}-{4:4}, at: exfat_fallocate+0x163/0x4e0 fs/exfat/file.c:131


***

If these findings have caused you to resend the series or submit a
separate fix, please add the following tag to your commit message:
  Tested-by: syzbot@syzkaller.appspotmail.com

---
This report is generated by a bot. It may contain errors.
syzbot ci engineers can be reached at syzkaller@googlegroups.com.

RE: [syzbot ci] Re: exfat: add limited FALLOC_FL_ZERO_RANGE support

[David Timber]

#syz test

diff --git a/fs/exfat/file.c b/fs/exfat/file.c
index 2daf0dbabb24..dfa5fc89f77d 100644
--- a/fs/exfat/file.c
+++ b/fs/exfat/file.c
@@ -36,7 +36,8 @@ static int exfat_cont_expand(struct inode *inode, loff_t size)
 	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)
+	WARN_ON(new_num_clusters < num_clusters);
+	if (new_num_clusters <= num_clusters)
 		goto out;
 
 	if (num_clusters) {
@@ -94,35 +95,87 @@ static int exfat_cont_expand(struct inode *inode, loff_t size)
 /*
  * Preallocate space for a file. This implements exfat's fallocate file
  * operation, which gets called from sys_fallocate system call. User space
- * requests len bytes at offset. In contrary to fat, we only support
- * FALLOC_FL_ALLOCATE_RANGE because by leaving the valid data length(VDL)
- * field, it is unnecessary to zero out the newly allocated clusters.
+ * requests len bytes at offset.
+ *
+ * In contrary to fat, FALLOC_FL_ALLOCATE_RANGE can be done without zeroing out
+ * the newly allocated clusters by leaving the valid data length(VDL) field
+ * unchanged.
+ *
+ * Due to the inherent limitation of the VDL scheme, FALLOC_FL_ZERO_RANGE is
+ * only possible when the requested range covers EOF.
  */
 static long exfat_fallocate(struct file *file, int mode,
 			  loff_t offset, loff_t len)
 {
 	struct inode *inode = file->f_mapping->host;
-	loff_t newsize = offset + len;
+	loff_t newsize, isize;
 	int err = 0;
 
 	/* No support for other modes */
-	if (mode != FALLOC_FL_ALLOCATE_RANGE)
+	switch (mode) {
+	case FALLOC_FL_ALLOCATE_RANGE:
+	case FALLOC_FL_ZERO_RANGE:
+	case FALLOC_FL_ZERO_RANGE|FALLOC_FL_KEEP_SIZE:
+		break;
+	default:
 		return -EOPNOTSUPP;
+	}
 
 	/* No support for dir */
 	if (!S_ISREG(inode->i_mode))
-		return -EOPNOTSUPP;
+		return mode & FALLOC_FL_ZERO_RANGE ? -EINVAL : -EOPNOTSUPP;
 
 	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
 		return -EIO;
 
 	inode_lock(inode);
 
-	if (newsize <= i_size_read(inode))
-		goto error;
+	newsize = offset + len;
+	isize = i_size_read(inode);
+
+	if (mode & FALLOC_FL_ZERO_RANGE) {
+		struct exfat_inode_info *ei = EXFAT_I(inode);
+		loff_t saved_validsize = ei->valid_size;
+
+		/* The requested range must span to or past EOF */
+		if (newsize < isize) {
+			err = -EOPNOTSUPP;
+			goto error;
+		}
+
+		/* valid_size can only be truncated */
+		if (offset < ei->valid_size)
+			ei->valid_size = offset;
+		/* If offset >= ei->valid_size, the range is already zeroed so that'd be no-op */
+
+		if (!(mode & FALLOC_FL_KEEP_SIZE) && isize < newsize)
+			err = exfat_cont_expand(inode, newsize);
+			/* inode invalidated in exfat_cont_expand() */
+		else {
+			/* update inode */
+			inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
+			mark_inode_dirty(inode);
+
+			if (IS_SYNC(inode))
+				err = write_inode_now(inode, 1);
+		}
+
+		if (err) {
+			/* inode unchanged - revert valid_size */
+			ei->valid_size = saved_validsize;
+			goto error;
+		}
+
+		/* drop cache after the new valid_size */
+		if (ei->valid_size != saved_validsize)
+			truncate_pagecache(inode, ei->valid_size);
+	} else { /* mode == FALLOC_FL_ALLOCATE_RANGE */
+		if (newsize <= isize)
+			goto error;
 
-	/* This is just an expanding truncate */
-	err = exfat_cont_expand(inode, newsize);
+		/* This is just an expanding truncate */
+		err = exfat_cont_expand(inode, newsize);
+	}
 
 error:
 	inode_unlock(inode);

Re: RE: [syzbot ci] Re: exfat: add limited FALLOC_FL_ZERO_RANGE support

[syzbot ci]

syzbot-ci does not support `test` command

Re: [PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

[kernel test robot]

BCC: lkp@intel.com
CC: oe-kbuild-all@lists.linux.dev
In-Reply-To: <20260319043553.301185-2-dxdt@dev.snart.me>
References: <20260319043553.301185-2-dxdt@dev.snart.me>
TO: David Timber <dxdt@dev.snart.me>
TO: linkinjeon@kernel.org
TO: sj1557.seo@samsung.com
CC: yuezhang.mo@sony.com
CC: linux-fsdevel@vger.kernel.org
CC: linux-kernel@vger.kernel.org
CC: David Timber <dxdt@dev.snart.me>

Hi David,

kernel test robot noticed the following build warnings:

[auto build test WARNING on next-20260318]
[cannot apply to brauner-vfs/vfs.all v7.0-rc4 v7.0-rc3 v7.0-rc2 linus/master v7.0-rc4]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/David-Timber/exfat-add-limited-FALLOC_FL_ZERO_RANGE-support/20260319-185700
base:   next-20260318
patch link:    https://lore.kernel.org/r/20260319043553.301185-2-dxdt%40dev.snart.me
patch subject: [PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support
:::::: branch date: 3 days ago
:::::: commit date: 3 days ago
config: i386-randconfig-141-20260322 (https://download.01.org/0day-ci/archive/20260322/202603221252.dVVrrOrL-lkp@intel.com/config)
compiler: clang version 20.1.8 (https://github.com/llvm/llvm-project 87f0227cb60147a26a1eeb4fb06e3b505e9c7261)
smatch: v0.5.0-9004-gb810ac53

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Reported-by: Dan Carpenter <error27@gmail.com>
| Closes: https://lore.kernel.org/r/202603221252.dVVrrOrL-lkp@intel.com/

smatch warnings:
fs/exfat/file.c:183 exfat_fallocate() warn: inconsistent returns '&inode->i_rwsem'.

vim +183 fs/exfat/file.c

98d917047e8b7f4 Namjae Jeon  2020-03-02   94  
bf1797960c20f3d David Timber 2026-02-28   95  /*
bf1797960c20f3d David Timber 2026-02-28   96   * Preallocate space for a file. This implements exfat's fallocate file
bf1797960c20f3d David Timber 2026-02-28   97   * operation, which gets called from sys_fallocate system call. User space
1178dacb657facf David Timber 2026-03-19   98   * requests len bytes at offset.
1178dacb657facf David Timber 2026-03-19   99   *
1178dacb657facf David Timber 2026-03-19  100   * In contrary to fat, FALLOC_FL_ALLOCATE_RANGE can be done without zeroing out
1178dacb657facf David Timber 2026-03-19  101   * the newly allocated clusters by leaving the valid data length(VDL) field
1178dacb657facf David Timber 2026-03-19  102   * unchanged.
1178dacb657facf David Timber 2026-03-19  103   *
1178dacb657facf David Timber 2026-03-19  104   * Due to the inherent limitation of the VDL scheme, FALLOC_FL_ZERO_RANGE is
1178dacb657facf David Timber 2026-03-19  105   * only possible when the requested range covers EOF.
bf1797960c20f3d David Timber 2026-02-28  106   */
bf1797960c20f3d David Timber 2026-02-28  107  static long exfat_fallocate(struct file *file, int mode,
bf1797960c20f3d David Timber 2026-02-28  108  			  loff_t offset, loff_t len)
bf1797960c20f3d David Timber 2026-02-28  109  {
bf1797960c20f3d David Timber 2026-02-28  110  	struct inode *inode = file->f_mapping->host;
1178dacb657facf David Timber 2026-03-19  111  	loff_t newsize, isize;
bf1797960c20f3d David Timber 2026-02-28  112  	int err = 0;
bf1797960c20f3d David Timber 2026-02-28  113  
bf1797960c20f3d David Timber 2026-02-28  114  	/* No support for other modes */
1178dacb657facf David Timber 2026-03-19  115  	switch (mode) {
1178dacb657facf David Timber 2026-03-19  116  	case FALLOC_FL_ALLOCATE_RANGE:
1178dacb657facf David Timber 2026-03-19  117  	case FALLOC_FL_ZERO_RANGE:
1178dacb657facf David Timber 2026-03-19  118  	case FALLOC_FL_ZERO_RANGE|FALLOC_FL_KEEP_SIZE:
1178dacb657facf David Timber 2026-03-19  119  		break;
1178dacb657facf David Timber 2026-03-19  120  	default:
bf1797960c20f3d David Timber 2026-02-28  121  		return -EOPNOTSUPP;
1178dacb657facf David Timber 2026-03-19  122  	}
bf1797960c20f3d David Timber 2026-02-28  123  
bf1797960c20f3d David Timber 2026-02-28  124  	/* No support for dir */
bf1797960c20f3d David Timber 2026-02-28  125  	if (!S_ISREG(inode->i_mode))
1178dacb657facf David Timber 2026-03-19  126  		return mode & FALLOC_FL_ZERO_RANGE ? -EINVAL : -EOPNOTSUPP;
bf1797960c20f3d David Timber 2026-02-28  127  
bf1797960c20f3d David Timber 2026-02-28  128  	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
bf1797960c20f3d David Timber 2026-02-28  129  		return -EIO;
bf1797960c20f3d David Timber 2026-02-28  130  
bf1797960c20f3d David Timber 2026-02-28  131  	inode_lock(inode);
bf1797960c20f3d David Timber 2026-02-28  132  
1178dacb657facf David Timber 2026-03-19  133  	newsize = offset + len;
1178dacb657facf David Timber 2026-03-19  134  	isize = i_size_read(inode);
1178dacb657facf David Timber 2026-03-19  135  
1178dacb657facf David Timber 2026-03-19  136  	if (mode & FALLOC_FL_ZERO_RANGE) {
1178dacb657facf David Timber 2026-03-19  137  		struct exfat_inode_info *ei = EXFAT_I(inode);
1178dacb657facf David Timber 2026-03-19  138  		loff_t saved_validsize = ei->valid_size;
1178dacb657facf David Timber 2026-03-19  139  
1178dacb657facf David Timber 2026-03-19  140  		/* The requested range must span to or past EOF */
1178dacb657facf David Timber 2026-03-19  141  		if (newsize < isize) {
1178dacb657facf David Timber 2026-03-19  142  			err = -EOPNOTSUPP;
1178dacb657facf David Timber 2026-03-19  143  			goto error;
1178dacb657facf David Timber 2026-03-19  144  		}
1178dacb657facf David Timber 2026-03-19  145  
1178dacb657facf David Timber 2026-03-19  146  		/* valid_size can only be truncated */
1178dacb657facf David Timber 2026-03-19  147  		if (offset < ei->valid_size)
1178dacb657facf David Timber 2026-03-19  148  			ei->valid_size = offset;
1178dacb657facf David Timber 2026-03-19  149  		/* If offset >= ei->valid_size, the range is already zeroed so that'd be no-op */
1178dacb657facf David Timber 2026-03-19  150  
1178dacb657facf David Timber 2026-03-19  151  		if (!(mode & FALLOC_FL_KEEP_SIZE) && isize < newsize) {
1178dacb657facf David Timber 2026-03-19  152  			err = exfat_cont_expand(inode, newsize);
1178dacb657facf David Timber 2026-03-19  153  			if (err) {
1178dacb657facf David Timber 2026-03-19  154  				/* inode unchanged - revert valid_size */
1178dacb657facf David Timber 2026-03-19  155  				ei->valid_size = saved_validsize;
1178dacb657facf David Timber 2026-03-19  156  				goto error;
1178dacb657facf David Timber 2026-03-19  157  			}
1178dacb657facf David Timber 2026-03-19  158  			/* inode invalidated in exfat_cont_expand() */
1178dacb657facf David Timber 2026-03-19  159  		} else {
1178dacb657facf David Timber 2026-03-19  160  			/* update inode */
1178dacb657facf David Timber 2026-03-19  161  			inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
1178dacb657facf David Timber 2026-03-19  162  
1178dacb657facf David Timber 2026-03-19  163  			mark_inode_dirty(inode);
1178dacb657facf David Timber 2026-03-19  164  
1178dacb657facf David Timber 2026-03-19  165  			if (IS_SYNC(inode))
1178dacb657facf David Timber 2026-03-19  166  				return write_inode_now(inode, 1);
1178dacb657facf David Timber 2026-03-19  167  		}
1178dacb657facf David Timber 2026-03-19  168  
1178dacb657facf David Timber 2026-03-19  169  		/* drop cache after the new valid_size */
1178dacb657facf David Timber 2026-03-19  170  		if (ei->valid_size != saved_validsize)
1178dacb657facf David Timber 2026-03-19  171  			truncate_pagecache(inode, ei->valid_size);
1178dacb657facf David Timber 2026-03-19  172  	} else { /* mode == FALLOC_FL_ALLOCATE_RANGE */
1178dacb657facf David Timber 2026-03-19  173  		if (newsize <= isize)
bf1797960c20f3d David Timber 2026-02-28  174  			goto error;
bf1797960c20f3d David Timber 2026-02-28  175  
bf1797960c20f3d David Timber 2026-02-28  176  		/* This is just an expanding truncate */
bf1797960c20f3d David Timber 2026-02-28  177  		err = exfat_cont_expand(inode, newsize);
1178dacb657facf David Timber 2026-03-19  178  	}
bf1797960c20f3d David Timber 2026-02-28  179  
bf1797960c20f3d David Timber 2026-02-28  180  error:
bf1797960c20f3d David Timber 2026-02-28  181  	inode_unlock(inode);
bf1797960c20f3d David Timber 2026-02-28  182  
bf1797960c20f3d David Timber 2026-02-28 @183  	return err;
bf1797960c20f3d David Timber 2026-02-28  184  }
bf1797960c20f3d David Timber 2026-02-28  185  

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

Re: [PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support

[Dan Carpenter]

Hi David,

kernel test robot noticed the following build warnings:

https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/David-Timber/exfat-add-limited-FALLOC_FL_ZERO_RANGE-support/20260319-185700
base:   next-20260318
patch link:    https://lore.kernel.org/r/20260319043553.301185-2-dxdt%40dev.snart.me
patch subject: [PATCH v0 1/1] exfat: add limited FALLOC_FL_ZERO_RANGE support
config: i386-randconfig-141-20260322 (https://download.01.org/0day-ci/archive/20260322/202603221252.dVVrrOrL-lkp@intel.com/config)
compiler: clang version 20.1.8 (https://github.com/llvm/llvm-project 87f0227cb60147a26a1eeb4fb06e3b505e9c7261)
smatch: v0.5.0-9004-gb810ac53

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
| Closes: https://lore.kernel.org/r/202603221252.dVVrrOrL-lkp@intel.com/

smatch warnings:
fs/exfat/file.c:183 exfat_fallocate() warn: inconsistent returns '&inode->i_rwsem'.

vim +183 fs/exfat/file.c

bf1797960c20f3d David Timber 2026-02-28  107  static long exfat_fallocate(struct file *file, int mode,
bf1797960c20f3d David Timber 2026-02-28  108  			  loff_t offset, loff_t len)
bf1797960c20f3d David Timber 2026-02-28  109  {
bf1797960c20f3d David Timber 2026-02-28  110  	struct inode *inode = file->f_mapping->host;
1178dacb657facf David Timber 2026-03-19  111  	loff_t newsize, isize;
bf1797960c20f3d David Timber 2026-02-28  112  	int err = 0;
bf1797960c20f3d David Timber 2026-02-28  113  
bf1797960c20f3d David Timber 2026-02-28  114  	/* No support for other modes */
1178dacb657facf David Timber 2026-03-19  115  	switch (mode) {
1178dacb657facf David Timber 2026-03-19  116  	case FALLOC_FL_ALLOCATE_RANGE:
1178dacb657facf David Timber 2026-03-19  117  	case FALLOC_FL_ZERO_RANGE:
1178dacb657facf David Timber 2026-03-19  118  	case FALLOC_FL_ZERO_RANGE|FALLOC_FL_KEEP_SIZE:
1178dacb657facf David Timber 2026-03-19  119  		break;
1178dacb657facf David Timber 2026-03-19  120  	default:
bf1797960c20f3d David Timber 2026-02-28  121  		return -EOPNOTSUPP;
1178dacb657facf David Timber 2026-03-19  122  	}
bf1797960c20f3d David Timber 2026-02-28  123  
bf1797960c20f3d David Timber 2026-02-28  124  	/* No support for dir */
bf1797960c20f3d David Timber 2026-02-28  125  	if (!S_ISREG(inode->i_mode))
1178dacb657facf David Timber 2026-03-19  126  		return mode & FALLOC_FL_ZERO_RANGE ? -EINVAL : -EOPNOTSUPP;
bf1797960c20f3d David Timber 2026-02-28  127  
bf1797960c20f3d David Timber 2026-02-28  128  	if (unlikely(exfat_forced_shutdown(inode->i_sb)))
bf1797960c20f3d David Timber 2026-02-28  129  		return -EIO;
bf1797960c20f3d David Timber 2026-02-28  130  
bf1797960c20f3d David Timber 2026-02-28  131  	inode_lock(inode);
                                                ^^^^^^^^^^^^^^^^^^

bf1797960c20f3d David Timber 2026-02-28  132  
1178dacb657facf David Timber 2026-03-19  133  	newsize = offset + len;
1178dacb657facf David Timber 2026-03-19  134  	isize = i_size_read(inode);
1178dacb657facf David Timber 2026-03-19  135  
1178dacb657facf David Timber 2026-03-19  136  	if (mode & FALLOC_FL_ZERO_RANGE) {
1178dacb657facf David Timber 2026-03-19  137  		struct exfat_inode_info *ei = EXFAT_I(inode);
1178dacb657facf David Timber 2026-03-19  138  		loff_t saved_validsize = ei->valid_size;
1178dacb657facf David Timber 2026-03-19  139  
1178dacb657facf David Timber 2026-03-19  140  		/* The requested range must span to or past EOF */
1178dacb657facf David Timber 2026-03-19  141  		if (newsize < isize) {
1178dacb657facf David Timber 2026-03-19  142  			err = -EOPNOTSUPP;
1178dacb657facf David Timber 2026-03-19  143  			goto error;
1178dacb657facf David Timber 2026-03-19  144  		}
1178dacb657facf David Timber 2026-03-19  145  
1178dacb657facf David Timber 2026-03-19  146  		/* valid_size can only be truncated */
1178dacb657facf David Timber 2026-03-19  147  		if (offset < ei->valid_size)
1178dacb657facf David Timber 2026-03-19  148  			ei->valid_size = offset;
1178dacb657facf David Timber 2026-03-19  149  		/* If offset >= ei->valid_size, the range is already zeroed so that'd be no-op */
1178dacb657facf David Timber 2026-03-19  150  
1178dacb657facf David Timber 2026-03-19  151  		if (!(mode & FALLOC_FL_KEEP_SIZE) && isize < newsize) {
1178dacb657facf David Timber 2026-03-19  152  			err = exfat_cont_expand(inode, newsize);
1178dacb657facf David Timber 2026-03-19  153  			if (err) {
1178dacb657facf David Timber 2026-03-19  154  				/* inode unchanged - revert valid_size */
1178dacb657facf David Timber 2026-03-19  155  				ei->valid_size = saved_validsize;
1178dacb657facf David Timber 2026-03-19  156  				goto error;
1178dacb657facf David Timber 2026-03-19  157  			}
1178dacb657facf David Timber 2026-03-19  158  			/* inode invalidated in exfat_cont_expand() */
1178dacb657facf David Timber 2026-03-19  159  		} else {
1178dacb657facf David Timber 2026-03-19  160  			/* update inode */
1178dacb657facf David Timber 2026-03-19  161  			inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
1178dacb657facf David Timber 2026-03-19  162  
1178dacb657facf David Timber 2026-03-19  163  			mark_inode_dirty(inode);
1178dacb657facf David Timber 2026-03-19  164  
1178dacb657facf David Timber 2026-03-19  165  			if (IS_SYNC(inode))
1178dacb657facf David Timber 2026-03-19  166  				return write_inode_now(inode, 1);

I don't think this calls inode_unlock(inode);

1178dacb657facf David Timber 2026-03-19  167  		}
1178dacb657facf David Timber 2026-03-19  168  
1178dacb657facf David Timber 2026-03-19  169  		/* drop cache after the new valid_size */
1178dacb657facf David Timber 2026-03-19  170  		if (ei->valid_size != saved_validsize)
1178dacb657facf David Timber 2026-03-19  171  			truncate_pagecache(inode, ei->valid_size);
1178dacb657facf David Timber 2026-03-19  172  	} else { /* mode == FALLOC_FL_ALLOCATE_RANGE */
1178dacb657facf David Timber 2026-03-19  173  		if (newsize <= isize)
bf1797960c20f3d David Timber 2026-02-28  174  			goto error;
bf1797960c20f3d David Timber 2026-02-28  175  
bf1797960c20f3d David Timber 2026-02-28  176  		/* This is just an expanding truncate */
bf1797960c20f3d David Timber 2026-02-28  177  		err = exfat_cont_expand(inode, newsize);
1178dacb657facf David Timber 2026-03-19  178  	}
bf1797960c20f3d David Timber 2026-02-28  179  
bf1797960c20f3d David Timber 2026-02-28  180  error:
bf1797960c20f3d David Timber 2026-02-28  181  	inode_unlock(inode);
bf1797960c20f3d David Timber 2026-02-28  182  
bf1797960c20f3d David Timber 2026-02-28 @183  	return err;
bf1797960c20f3d David Timber 2026-02-28  184  }

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki