fs/exfat/namei.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
exfat_find_empty_entry() stores the return value of
exfat_alloc_cluster() in an unsigned int. When
exfat_alloc_cluster() returns a negative errno, it is
converted to a large positive value, which corrupts
error propagation to the caller.
Change the type of ret to int so that negative errno
values are preserved.
Fixes: 5f2aa075070c ("exfat: add inode operations")
Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
---
fs/exfat/namei.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c
index f5f1c4e8a29f..f2a87ecd79f9 100644
--- a/fs/exfat/namei.c
+++ b/fs/exfat/namei.c
@@ -304,8 +304,8 @@ static int exfat_find_empty_entry(struct inode *inode,
struct exfat_chain *p_dir, int num_entries,
struct exfat_entry_set_cache *es)
{
- int dentry;
- unsigned int ret, last_clu;
+ int dentry, ret;
+ unsigned int last_clu;
loff_t size = 0;
struct exfat_chain clu;
struct super_block *sb = inode->i_sb;
--
2.50.1.windows.1
Hi, Haotian,
On 25. 12. 3. 16:08, Haotian Zhang wrote:
> exfat_find_empty_entry() stores the return value of
> exfat_alloc_cluster() in an unsigned int. When
> exfat_alloc_cluster() returns a negative errno, it is
> converted to a large positive value, which corrupts
> error propagation to the caller.
Have you ever encountered an actual error?
IMO, due to implicit type conversion, it should work as follows,
so, I don't think there will be any real issues.
int -> unsigned int -> int
Anyway, it makes sense to modify the type of ret from unsigned int to int.
What about changing the title and comment?
Thanks.
SJ
>
> Change the type of ret to int so that negative errno
> values are preserved.
>
> Fixes: 5f2aa075070c ("exfat: add inode operations")
> Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
> ---
> fs/exfat/namei.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c
> index f5f1c4e8a29f..f2a87ecd79f9 100644
> --- a/fs/exfat/namei.c
> +++ b/fs/exfat/namei.c
> @@ -304,8 +304,8 @@ static int exfat_find_empty_entry(struct inode *inode,
> struct exfat_chain *p_dir, int num_entries,
> struct exfat_entry_set_cache *es)
> {
> - int dentry;
> - unsigned int ret, last_clu;
> + int dentry, ret;
> + unsigned int last_clu;
> loff_t size = 0;
> struct exfat_chain clu;
> struct super_block *sb = inode->i_sb;
Change the type of 'ret' from unsigned int to int in
exfat_find_empty_entry(). Although the implicit type conversion
(int -> unsigned int -> int) does not cause actual bugs in
practice, using int directly is more appropriate for storing
error codes returned by exfat_alloc_cluster().
This improves code clarity and consistency with standard error
handling practices.
Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
---
fs/exfat/namei.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c
index f5f1c4e8a29f..f2a87ecd79f9 100644
--- a/fs/exfat/namei.c
+++ b/fs/exfat/namei.c
@@ -304,8 +304,8 @@ static int exfat_find_empty_entry(struct inode *inode,
struct exfat_chain *p_dir, int num_entries,
struct exfat_entry_set_cache *es)
{
- int dentry;
- unsigned int ret, last_clu;
+ int dentry, ret;
+ unsigned int last_clu;
loff_t size = 0;
struct exfat_chain clu;
struct super_block *sb = inode->i_sb;
--
2.50.1.windows.1
On Fri, Dec 5, 2025 at 10:59 AM Haotian Zhang <vulab@iscas.ac.cn> wrote: > > Change the type of 'ret' from unsigned int to int in > exfat_find_empty_entry(). Although the implicit type conversion > (int -> unsigned int -> int) does not cause actual bugs in > practice, using int directly is more appropriate for storing > error codes returned by exfat_alloc_cluster(). > > This improves code clarity and consistency with standard error > handling practices. > > Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn> Applied it to #dev. Thanks!
On 25. 12. 5. 10:59, Haotian Zhang wrote:
> Change the type of 'ret' from unsigned int to int in
> exfat_find_empty_entry(). Although the implicit type conversion
> (int -> unsigned int -> int) does not cause actual bugs in
> practice, using int directly is more appropriate for storing
> error codes returned by exfat_alloc_cluster().
>
> This improves code clarity and consistency with standard error
> handling practices.
>
> Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
> ---
> fs/exfat/namei.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c
> index f5f1c4e8a29f..f2a87ecd79f9 100644
> --- a/fs/exfat/namei.c
> +++ b/fs/exfat/namei.c
> @@ -304,8 +304,8 @@ static int exfat_find_empty_entry(struct inode *inode,
> struct exfat_chain *p_dir, int num_entries,
> struct exfat_entry_set_cache *es)
> {
> - int dentry;
> - unsigned int ret, last_clu;
> + int dentry, ret;
> + unsigned int last_clu;
The patch looks good to me:
Reviewed-by: Sungjong Seo <sj1557.seo@samsung.com>
> loff_t size = 0;
> struct exfat_chain clu;
> struct super_block *sb = inode->i_sb;
© 2016 - 2025 Red Hat, Inc.