[PATCH] udf: prevent memory leak in udf_new_inode

[PATCH] udf: prevent memory leak in udf_new_inode

[Navid Emamdoost]

In udf_new_inode if either udf_new_block or insert_inode_locked fials
the allocated memory for iinfo->i_ext.i_data should be released.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
---
 fs/udf/ialloc.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
index 0adb40718a5d..b8ab3acab6b6 100644
--- a/fs/udf/ialloc.c
+++ b/fs/udf/ialloc.c
@@ -86,6 +86,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
 			      dinfo->i_location.partitionReferenceNum,
 			      start, &err);
 	if (err) {
+		kfree(iinfo->i_ext.i_data);
 		iput(inode);
 		return ERR_PTR(err);
 	}
@@ -130,6 +131,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
 	inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
 	iinfo->i_crtime = inode->i_mtime;
 	if (unlikely(insert_inode_locked(inode) < 0)) {
+		kfree(iinfo->i_ext.i_data);
 		make_bad_inode(inode);
 		iput(inode);
 		return ERR_PTR(-EIO);
-- 
2.17.1

Re: [PATCH] udf: prevent memory leak in udf_new_inode

[Al Viro]

On Wed, Sep 25, 2019 at 04:39:03PM -0500, Navid Emamdoost wrote:
> In udf_new_inode if either udf_new_block or insert_inode_locked fials
> the allocated memory for iinfo->i_ext.i_data should be released.

"... because of such-and-such reasons" part appears to be missing.
Why should it be released there?

> Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
> ---
>  fs/udf/ialloc.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
> index 0adb40718a5d..b8ab3acab6b6 100644
> --- a/fs/udf/ialloc.c
> +++ b/fs/udf/ialloc.c
> @@ -86,6 +86,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
>  			      dinfo->i_location.partitionReferenceNum,
>  			      start, &err);
>  	if (err) {
> +		kfree(iinfo->i_ext.i_data);
>  		iput(inode);
>  		return ERR_PTR(err);
>  	}

Have you tested that?  Because it has all earmarks of double-free;
normal eviction pathway ought to free the damn thing.  <greps around
a bit>

Mind explaining what's to stop ->evict_inode (== udf_evict_inode) from
hitting
        kfree(iinfo->i_ext.i_data);
considering that this call of kfree() appears to be unconditional there?

> @@ -130,6 +131,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
>  	inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
>  	iinfo->i_crtime = inode->i_mtime;
>  	if (unlikely(insert_inode_locked(inode) < 0)) {
> +		kfree(iinfo->i_ext.i_data);
>  		make_bad_inode(inode);
>  		iput(inode);
>  		return ERR_PTR(-EIO);

And the same here.

Re: [PATCH] udf: prevent memory leak in udf_new_inode

[Jan Kara]

On Wed 25-09-19 23:24:08, Al Viro wrote:
> On Wed, Sep 25, 2019 at 04:39:03PM -0500, Navid Emamdoost wrote:
> > In udf_new_inode if either udf_new_block or insert_inode_locked fials
> > the allocated memory for iinfo->i_ext.i_data should be released.
> 
> "... because of such-and-such reasons" part appears to be missing.
> Why should it be released there?
> 
> > Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
> > ---
> >  fs/udf/ialloc.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
> > index 0adb40718a5d..b8ab3acab6b6 100644
> > --- a/fs/udf/ialloc.c
> > +++ b/fs/udf/ialloc.c
> > @@ -86,6 +86,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
> >  			      dinfo->i_location.partitionReferenceNum,
> >  			      start, &err);
> >  	if (err) {
> > +		kfree(iinfo->i_ext.i_data);
> >  		iput(inode);
> >  		return ERR_PTR(err);
> >  	}
> 
> Have you tested that?  Because it has all earmarks of double-free;
> normal eviction pathway ought to free the damn thing.  <greps around
> a bit>
> 
> Mind explaining what's to stop ->evict_inode (== udf_evict_inode) from
> hitting
>         kfree(iinfo->i_ext.i_data);
> considering that this call of kfree() appears to be unconditional there?

Exactly. udf_evict_inode() is responsible for freeing iinfo->i_ext.i_data
so the patch would result in double free.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH] udf: prevent memory leak in udf_new_inode

[Navid Emamdoost]

On Thu, Sep 26, 2019 at 10:00:31AM +0200, Jan Kara wrote:
> On Wed 25-09-19 23:24:08, Al Viro wrote:
> > On Wed, Sep 25, 2019 at 04:39:03PM -0500, Navid Emamdoost wrote:
> > > In udf_new_inode if either udf_new_block or insert_inode_locked fials
> > > the allocated memory for iinfo->i_ext.i_data should be released.
> > 
> > "... because of such-and-such reasons" part appears to be missing.
> > Why should it be released there?
> > 
> > > Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
> > > ---
> > >  fs/udf/ialloc.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
> > > index 0adb40718a5d..b8ab3acab6b6 100644
> > > --- a/fs/udf/ialloc.c
> > > +++ b/fs/udf/ialloc.c
> > > @@ -86,6 +86,7 @@ struct inode *udf_new_inode(struct inode *dir, umode_t mode)
> > >  			      dinfo->i_location.partitionReferenceNum,
> > >  			      start, &err);
> > >  	if (err) {
> > > +		kfree(iinfo->i_ext.i_data);
> > >  		iput(inode);
> > >  		return ERR_PTR(err);
> > >  	}
> > 
> > Have you tested that?  Because it has all earmarks of double-free;
> > normal eviction pathway ought to free the damn thing.  <greps around
> > a bit>
> > 
> > Mind explaining what's to stop ->evict_inode (== udf_evict_inode) from
> > hitting
> >         kfree(iinfo->i_ext.i_data);
> > considering that this call of kfree() appears to be unconditional there?
> 
> Exactly. udf_evict_inode() is responsible for freeing iinfo->i_ext.i_data
> so the patch would result in double free.
> 
> 								Honza
Thanks for clarification.
> -- 
> Jan Kara <jack@suse.com>
> SUSE Labs, CR