On IMA policy update, if a measure rule exists in the policy,
IMA_MEASURE is set for ima_policy_flags which makes the violation_check
variable always true. Coupled with a no-action on MAY_READ for a
FILE_CHECK call, we're always taking the inode_lock().

This becomes a performance problem for extremely heavy read-only workloads.
Therefore, prevent this only in the case there's no action to be taken.

Signed-off-by: Frederick Lawler <fred@cloudflare.com>
---
Changes since v1:
- Add MAY_WRITE && action check to violation_check to avoid MAY_READ
  only situations
---
 security/integrity/ima/ima_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
index 9b87556b03a7..11a91d9dae48 100644
--- a/security/integrity/ima/ima_main.c
+++ b/security/integrity/ima/ima_main.c
@@ -237,7 +237,8 @@ static int process_measurement(struct file *file, const struct cred *cred,
 				&allowed_algos);
 	violation_check = ((func == FILE_CHECK || func == MMAP_CHECK ||
 			    func == MMAP_CHECK_REQPROT) &&
-			   (ima_policy_flag & IMA_MEASURE));
+			   (ima_policy_flag & IMA_MEASURE) &&
+			   ((action & IMA_MEASURE) || (mask & MAY_WRITE)));
 	if (!action && !violation_check)
 		return 0;
 
-- 
2.43.0

My mistake, this is PATCH v2. I forgot to change the subject in git
send-email. I can resend if that's needed.

On 3/25/2025 7:18 PM, Frederick Lawler wrote:
> My mistake, this is PATCH v2. I forgot to change the subject in git
> send-email. I can resend if that's needed.

No problem. However, let's change:

mask & MAY_WRITE

with

file->f_mode & FMODE_WRITE

Thanks

Roberto