[PATCH] Revert "ext4: apply umask if ACL support is disabled"

[PATCH] Revert "ext4: apply umask if ACL support is disabled"

[Max Kellermann]

This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
commit caused a regression because now the umask was applied to
symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
has been fixed somewhere else already.

Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
 fs/ext4/acl.h | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
index ef4c19e5f570..0c5a79c3b5d4 100644
--- a/fs/ext4/acl.h
+++ b/fs/ext4/acl.h
@@ -68,11 +68,6 @@ extern int ext4_init_acl(handle_t *, struct inode *, struct inode *);
 static inline int
 ext4_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
 {
-	/* usually, the umask is applied by posix_acl_create(), but if
-	   ext4 ACL support is disabled at compile time, we need to do
-	   it here, because posix_acl_create() will never be called */
-	inode->i_mode &= ~current_umask();
-
 	return 0;
 }
 #endif  /* CONFIG_EXT4_FS_POSIX_ACL */
-- 
2.39.2

Re: [PATCH] Revert "ext4: apply umask if ACL support is disabled"

[Christian Brauner]

On Fri, Mar 15, 2024 at 03:29:56PM +0100, Max Kellermann wrote:
> This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
> commit caused a regression because now the umask was applied to
> symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
> has been fixed somewhere else already.
> 
> Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> ---

Looks good to me,
Reviewed-by: Christian Brauner <brauner@kernel.org>

Re: [PATCH] Revert "ext4: apply umask if ACL support is disabled"

[Michael Forney]

Max Kellermann <max.kellermann@ionos.com> wrote:
> This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
> commit caused a regression because now the umask was applied to
> symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
> has been fixed somewhere else already.

Thanks, Max! I've verified that this fixes symlink modes for me,
as well as the flatpak corruption error I was getting.

> Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>

Tested-by: Michael Forney <mforney@mforney.org>

Re: [PATCH] Revert "ext4: apply umask if ACL support is disabled"

[Michael Forney]

Michael Forney <mforney@mforney.org> wrote:
> Max Kellermann <max.kellermann@ionos.com> wrote:
> > This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
> > commit caused a regression because now the umask was applied to
> > symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
> > has been fixed somewhere else already.
> 
> Thanks, Max! I've verified that this fixes symlink modes for me,
> as well as the flatpak corruption error I was getting.
> 
> > Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
> > Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> 
> Tested-by: Michael Forney <mforney@mforney.org>

Just checking in on this. I'd really like to see this regression
fixed. It currently affects versions 6.5 through 6.9-rc6 as well
as all longterm releases.

Re: [PATCH] Revert "ext4: apply umask if ACL support is disabled"

[Theodore Ts'o]

On Fri, 15 Mar 2024 15:29:56 +0100, Max Kellermann wrote:
> This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
> commit caused a regression because now the umask was applied to
> symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
> has been fixed somewhere else already.
> 
> 

Applied, thanks!

[1/1] Revert "ext4: apply umask if ACL support is disabled"
      commit: c77194965dd0dcc26f9c1671d2e74e4eb1248af5

Best regards,
-- 
Theodore Ts'o <tytso@mit.edu>

[PATCH 5.15.y] Revert "ext4: apply umask if ACL support is disabled"

[Jesper Nilsson]

From: Max Kellermann <max.kellermann@ionos.com>

This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
commit caused a regression because now the umask was applied to
symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
has been fixed somewhere else already.

Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Tested-by: Michael Forney <mforney@mforney.org>
Link: https://lore.kernel.org/r/20240315142956.2420360-1-max.kellermann@ionos.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
(cherry picked from commit c77194965dd0dcc26f9c1671d2e74e4eb1248af5)

---
This revert never reached linux-5.15.y, and caused the same regression
with symbolic links. The original problem is just as the revert states
fixed and the test program for O_TMPFILE works as expected.
---
 fs/ext4/acl.h | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
index 0db6ae0ca936..3219669732bf 100644
--- a/fs/ext4/acl.h
+++ b/fs/ext4/acl.h
@@ -68,11 +68,6 @@ extern int ext4_init_acl(handle_t *, struct inode *, struct inode *);
 static inline int
 ext4_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
 {
-	/* usually, the umask is applied by posix_acl_create(), but if
-	   ext4 ACL support is disabled at compile time, we need to do
-	   it here, because posix_acl_create() will never be called */
-	inode->i_mode &= ~current_umask();
-
 	return 0;
 }
 #endif  /* CONFIG_EXT4_FS_POSIX_ACL */
-- 
2.36.0

Re: [PATCH 5.15.y] Revert "ext4: apply umask if ACL support is disabled"

[Greg KH]

On Mon, Feb 10, 2025 at 09:28:01PM +0100, Jesper Nilsson wrote:
> From: Max Kellermann <max.kellermann@ionos.com>
> 
> This reverts commit 484fd6c1de13b336806a967908a927cc0356e312.  The
> commit caused a regression because now the umask was applied to
> symlinks and the fix is unnecessary because the umask/O_TMPFILE bug
> has been fixed somewhere else already.
> 
> Fixes: https://lore.kernel.org/lkml/28DSITL9912E1.2LSZUVTGTO52Q@mforney.org/
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> Reviewed-by: Christian Brauner <brauner@kernel.org>
> Tested-by: Michael Forney <mforney@mforney.org>
> Link: https://lore.kernel.org/r/20240315142956.2420360-1-max.kellermann@ionos.com
> Signed-off-by: Theodore Ts'o <tytso@mit.edu>
> (cherry picked from commit c77194965dd0dcc26f9c1671d2e74e4eb1248af5)
> 
> ---
> This revert never reached linux-5.15.y, and caused the same regression
> with symbolic links. The original problem is just as the revert states
> fixed and the test program for O_TMPFILE works as expected.

This revert never hit any of the other stable kernels either, should it
go to all of them?

Also, you didn't sign off on this :(

thanks,

greg k-h