diff mbox series

ima: Fix memory leak in __ima_inode_hash()

Message ID 20221102163006.1039343-1-roberto.sassu@huaweicloud.com (mailing list archive)
State Not Applicable
Headers show
Series ima: Fix memory leak in __ima_inode_hash() | expand

Checks

Context Check Description
netdev/tree_selection success Not a local patch

Commit Message

Roberto Sassu Nov. 2, 2022, 4:30 p.m. UTC 
From: Roberto Sassu <roberto.sassu@huawei.com>

Commit f3cc6b25dcc5 ("ima: always measure and audit files in policy") lets
measurement or audit happen even if the file digest cannot be calculated.

As a result, iint->ima_hash could have been allocated despite
ima_collect_measurement() returning an error.

Since ima_hash belongs to a temporary inode metadata structure, declared
at the beginning of __ima_inode_hash(), just add a kfree() call if
ima_collect_measurement() returns an error different from -ENOMEM (in that
case, ima_hash should not have been allocated).

Cc: stable@vger.kernel.org
Fixes: 280fe8367b0d ("ima: Always return a file measurement in ima_file_hash()")
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
---
 security/integrity/ima/ima_main.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

Comments

Mimi Zohar Nov. 2, 2022, 10:04 p.m. UTC | #1 
Hi Roberto,

On Wed, 2022-11-02 at 17:30 +0100, Roberto Sassu wrote:
> From: Roberto Sassu <roberto.sassu@huawei.com>

Any chance you could fix your mailer?

> 
> Commit f3cc6b25dcc5 ("ima: always measure and audit files in policy") lets
> measurement or audit happen even if the file digest cannot be calculated.
> 
> As a result, iint->ima_hash could have been allocated despite
> ima_collect_measurement() returning an error.
> 
> Since ima_hash belongs to a temporary inode metadata structure, declared
> at the beginning of __ima_inode_hash(), just add a kfree() call if
> ima_collect_measurement() returns an error different from -ENOMEM (in that
> case, ima_hash should not have been allocated).
> 
> Cc: stable@vger.kernel.org
> Fixes: 280fe8367b0d ("ima: Always return a file measurement in ima_file_hash()")
> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>

Thanks,

Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
Roberto Sassu Nov. 3, 2022, 7:49 a.m. UTC | #2 
On Wed, 2022-11-02 at 18:04 -0400, Mimi Zohar wrote:
> Hi Roberto,
> 
> On Wed, 2022-11-02 at 17:30 +0100, Roberto Sassu wrote:
> > From: Roberto Sassu <roberto.sassu@huawei.com>
> 
> Any chance you could fix your mailer?

Hi Mimi

not sure how to fix this. I need to send from @huaweicloud.com because
some people didn't receive the patches from @huawei.com. But I still
prefer to have the original email in the patches.

Thanks

Roberto

> > Commit f3cc6b25dcc5 ("ima: always measure and audit files in
> > policy") lets
> > measurement or audit happen even if the file digest cannot be
> > calculated.
> > 
> > As a result, iint->ima_hash could have been allocated despite
> > ima_collect_measurement() returning an error.
> > 
> > Since ima_hash belongs to a temporary inode metadata structure,
> > declared
> > at the beginning of __ima_inode_hash(), just add a kfree() call if
> > ima_collect_measurement() returns an error different from -ENOMEM
> > (in that
> > case, ima_hash should not have been allocated).
> > 
> > Cc: stable@vger.kernel.org
> > Fixes: 280fe8367b0d ("ima: Always return a file measurement in
> > ima_file_hash()")
> > Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
> 
> Thanks,
> 
> Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
diff mbox series

Patch

diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
index 040b03ddc1c7..4a207a3ef7ef 100644
--- a/security/integrity/ima/ima_main.c
+++ b/security/integrity/ima/ima_main.c
@@ -542,8 +542,13 @@  static int __ima_inode_hash(struct inode *inode, struct file *file, char *buf,
 
 		rc = ima_collect_measurement(&tmp_iint, file, NULL, 0,
 					     ima_hash_algo, NULL);
-		if (rc < 0)
+		if (rc < 0) {
+			/* ima_hash could be allocated in case of failure. */
+			if (rc != -ENOMEM)
+				kfree(tmp_iint.ima_hash);
+
 			return -EOPNOTSUPP;
+		}
 
 		iint = &tmp_iint;
 		mutex_lock(&iint->mutex);