IMA: support for duplicate measurement records

IMA does not include duplicate file, buffer, or critical data measurement records since TPM extend is a very expensive operation. However, in some cases, the measurement of duplicate records is necessary to accurately determine the current state of the system. For instance - the file, buffer, or critical data measurement record may change from some value 'val#1', to 'val#2', and then back to 'val#1'. Currently, IMA will not measure the last change to 'val#1', since the hash of 'val#1' for the given record is already present in the measurement log. This limits the ability of the attestation service to accurately determine the current state of the system, because it would be interpreted as the system having 'val#2' for the given record.

Update ima_add_template_entry() to support measurement of duplicate records, driven by a Kconfig option - IMA_DISABLE_HTABLE.

---
Change Log v3:
- Incorporated feedback from Mimi on v2.
- Updated patch title and description to make it generic.
- Changed config description word 'data' to 'records'.
- Tested use cases for boot param "ima_policy=tcb".

Change Log v2:
- Incorporated feedback from Mimi on v1.
- The fix is not just applicable to measurement of critical data, it now applies to other buffers and file data as well.
- the fix is driven by a Kconfig option IMA_DISABLE_HTABLE, rather than a IMA policy condition - allow_dup.

 security/integrity/ima/Kconfig     | 7 +++++++
 security/integrity/ima/ima_queue.c | 5 +++--
 2 files changed, 10 insertions(+), 2 deletions(-)

index 12e9250c1bec..d0ceada99243 100644
@@ -334,3 +334,10 @@ config IMA_SECURE_AND_OR_TRUSTED_BOOT
        help
           This option is selected by architectures to enable secure and/or
           trusted boot based on IMA runtime policies.
+
+config IMA_DISABLE_HTABLE
+	bool "Disable htable to allow measurement of duplicate records"
+	depends on IMA
+	default n
+	help
+	   This option disables htable to allow measurement of duplicate records.

index c096ef8945c7..532da87ce519 100644
@@ -168,7 +168,7 @@ int ima_add_template_entry(struct ima_template_entry *entry, int violation,
 	int result = 0, tpmresult = 0;
 
 	mutex_lock(&ima_extend_list_mutex);
-	if (!violation) {
+	if (!violation && !IS_ENABLED(CONFIG_IMA_DISABLE_HTABLE)) {
 		if (ima_lookup_digest_entry(digest, entry->pcr)) {
 			audit_cause = "hash_exists";
 			result = -EEXIST;

@@ -176,7 +176,8 @@ int ima_add_template_entry(struct ima_template_entry *entry, int violation,
 		}
 	}
 
-	result = ima_add_digest_entry(entry, 1);
+	result = ima_add_digest_entry(entry,
+				      !IS_ENABLED(CONFIG_IMA_DISABLE_HTABLE));
 	if (result < 0) {
 		audit_cause = "ENOMEM";
 		audit_info = 0;

2.17.1

Hi Tushar,

LGTM.

Kind regards,
Petr

Hello Petr,

Thank you taking a look at the patch, and the 'Reviewed-by' tag. ~Tushar