[PATCH next] ext4: Fix diagnostic printf formats

[PATCH next] ext4: Fix diagnostic printf formats

[david.laight.linux]

From: David Laight <david.laight.linux@gmail.com>

The formats for non-terminated names should be "%.*s" not "%*.s".
The kernel currently treats "%*.s" as equivalent to "%*s" whereas
userspace requires it be equivalent to "%*.0s".
Neither is correct here.

Signed-off-by: David Laight <david.laight.linux@gmail.com>
---
 fs/ext4/namei.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index c4b5e252af0e..7aaf5fbd4498 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -647,7 +647,7 @@ static struct stats dx_show_leaf(struct inode *dir,
 					/* Directory is not encrypted */
 					(void) ext4fs_dirhash(dir, de->name,
 						de->name_len, &h);
-					printk("%*.s:(U)%x.%u ", len,
+					printk("%.*s:(U)%x.%u ", len,
 					       name, h.hash,
 					       (unsigned) ((char *) de
 							   - base));
@@ -683,7 +683,7 @@ static struct stats dx_show_leaf(struct inode *dir,
 						(void) ext4fs_dirhash(dir,
 							de->name,
 							de->name_len, &h);
-					printk("%*.s:(E)%x.%u ", len, name,
+					printk("%.*s:(E)%x.%u ", len, name,
 					       h.hash, (unsigned) ((char *) de
 								   - base));
 					fscrypt_fname_free_buffer(
@@ -694,7 +694,7 @@ static struct stats dx_show_leaf(struct inode *dir,
 				char *name = de->name;
 				(void) ext4fs_dirhash(dir, de->name,
 						      de->name_len, &h);
-				printk("%*.s:%x.%u ", len, name, h.hash,
+				printk("%.*s:%x.%u ", len, name, h.hash,
 				       (unsigned) ((char *) de - base));
 #endif
 			}
-- 
2.39.5

Re: [PATCH next] ext4: Fix diagnostic printf formats

[Andy Shevchenko]

On Thu, Mar 26, 2026 at 08:18:04PM +0000, david.laight.linux@gmail.com wrote:

> The formats for non-terminated names should be "%.*s" not "%*.s".
> The kernel currently treats "%*.s" as equivalent to "%*s" whereas
> userspace requires it be equivalent to "%*.0s".
> Neither is correct here.

This entire code seems was never tested properly and it's a dead code
until one defines manually DX_DEBUG. It also has tons of plain printk()
calls that may behave differently if the first character is not printable
but maps to the level of printk().

I'm not sure how your patch helps with all that, but apparently the
printed data has to be NUL-terminated, otherwise I have no idea how
it was ever working without crashes.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH next] ext4: Fix diagnostic printf formats

[David Laight]

On Fri, 27 Mar 2026 12:48:56 +0200
Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:

> On Thu, Mar 26, 2026 at 08:18:04PM +0000, david.laight.linux@gmail.com wrote:
> 
> > The formats for non-terminated names should be "%.*s" not "%*.s".
> > The kernel currently treats "%*.s" as equivalent to "%*s" whereas
> > userspace requires it be equivalent to "%*.0s".
> > Neither is correct here.  
> 
> This entire code seems was never tested properly and it's a dead code
> until one defines manually DX_DEBUG. It also has tons of plain printk()
> calls that may behave differently if the first character is not printable
> but maps to the level of printk().
> 
> I'm not sure how your patch helps with all that, but apparently the
> printed data has to be NUL-terminated, otherwise I have no idea how
> it was ever working without crashes.
> 

I noticed that as well.
I suspect it way have worked for the person that wrote it because the
name strings all happened to be NUL terminated.
There is certainly likely to be a '\0' before you 'fall off' mapped
memory and crash - so maybe they just ignored the extra characters.

Clearly the other option is to delete it all.
But regardless the format string is wrong.

	David