Skip to contentSkip to conversation

LTP IMA fix bundle

PATCH In review Jia Zhang 3 participants 19 messages Jan 16, 2019 – Jan 24, 2019 Keys-trusted

lore
Share & tools
lore

Jia Zhang

author
#0 lore raw

v3 cover LTP IMA fix bundle

Changelog since V2:
- Rework patch 6. Here is a quick diff against V2:

testcases/kernel/security/integrity/ima/tests/ima_violations.sh +4 -3
index a1360b8..74223c2 100755
@@ -20,6 +20,7 @@
 # Test whether ToMToU and open_writer violations invalidatethe PCR and are logged.
 
 TST_SETUP="setup"
+TST_CLEANUP="cleanup"
 TST_CNT=3
 TST_NEEDS_DEVICE=1
 
@@ -46,10 +47,12 @@ setup()
        tst_res TINFO "using log $LOG"
 }
 
-reset_printk_ratelimit()
+cleanup()
 {
        [ "$PRINTK_RATE_LIMIT" != "0" ] && \
                sysctl -wq kernel.printk_ratelimit=$PRINTK_RATE_LIMIT
+
+       ima_cleanup
 }
 
 open_file_read()
@@ -163,8 +166,6 @@ test3()
 
        validate $num_violations $count $search
 
-       reset_printk_ratelimit
-
        # wait for ima_mmap to exit, so we can umount
        tst_sleep 2s
 }

The latest test results are attached:

Test Start Time: Wed Jan 16 10:44:10 2019
-----------------------------------------

Testcase                                           Result     Exit Value
--------                                           ------     ----------
ima_measurements                                   PASS       0    
ima_policy                                         PASS       0    
ima_tpm                                            CONF       32   
ima_violations                                     PASS       0    

-----------------------------------------------
Total Tests: 4
Total Skipped Tests: 1
Total Failures: 0
Kernel Version: 5.0.0-rc1+
Machine Architecture: x86_64
Hostname: test-machine

Jia

author
#1 lore raw

PATCH v3 1/6 ima/ima_boot_aggregate: Fix the definition of event log

According to [1], the structure of event log should be packed, and certain fields should be 32-bit unsigned integer. Fortunately, keeping natural alignment seems to make everything working as expected all the time.

[1] page 17,18 @https://trustedcomputinggroup.org/wp-content/uploads/TCG_EFI_Protocol_1_22_Final-v05.pdf

---
 testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c +4 -4
index f6e7be0..d85d222 100644
@@ -53,10 +53,10 @@ int main(int argc, char *argv[])
 	struct {
 		struct {
 			u_int32_t pcr;
-			int type;
-			unsigned char digest[SHA_DIGEST_LENGTH];
-			u_int16_t len;
-		} header;
+			u_int32_t type;
+			u_int8_t digest[SHA_DIGEST_LENGTH];
+			u_int32_t len;
+		} header __attribute__ ((packed));
 		char *data;
 	} event;
 	struct {
Signature · 1 line
1.8.3.1
maintainer Reviewed-by on v3 1/6
#2 lore raw

Re: [PATCH v3 1/6] ima/ima_boot_aggregate: Fix the definition of event log

Quoted 8 lines · Jia Zhang

According to [1], the structure of event log should be packed, and certain fields should be 32-bit unsigned integer. Fortunately, keeping natural alignment seems to make everything working as expected all the time.

[1] page 17,18 @https://trustedcomputinggroup.org/wp-content/uploads/TCG_EFI_Protocol_1_22_Final-v05.pdf

Quoted 23 lines
---
 testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c +4 -4
index f6e7be0..d85d222 100644
@@ -53,10 +53,10 @@ int main(int argc, char *argv[])
 	struct {
 		struct {
 			u_int32_t pcr;
-			int type;
-			unsigned char digest[SHA_DIGEST_LENGTH];
-			u_int16_t len;
-		} header;
+			u_int32_t type;
+			u_int8_t digest[SHA_DIGEST_LENGTH];
+			u_int32_t len;
+		} header __attribute__ ((packed));
 		char *data;
 	} event;
 	struct {
maintainer Reviewed-by on v3 1/6
#3 lore raw

Re: [PATCH v3 1/6] ima/ima_boot_aggregate: Fix the definition of event log

In reply to Jia Zhang
Quoted 8 lines · Jia Zhang

According to [1], the structure of event log should be packed, and certain fields should be 32-bit unsigned integer. Fortunately, keeping natural alignment seems to make everything working as expected all the time.

[1] page 17,18 @https://trustedcomputinggroup.org/wp-content/uploads/TCG_EFI_Protocol_1_22_Final-v05.pdf

Quoted 23 lines
---
 testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c +4 -4
index f6e7be0..d85d222 100644
@@ -53,10 +53,10 @@ int main(int argc, char *argv[])
 	struct {
 		struct {
 			u_int32_t pcr;
-			int type;
-			unsigned char digest[SHA_DIGEST_LENGTH];
-			u_int16_t len;
-		} header;
+			u_int32_t type;
+			u_int8_t digest[SHA_DIGEST_LENGTH];
+			u_int32_t len;
+		} header __attribute__ ((packed));
 		char *data;
 	} event;
 	struct {
author
#4 lore raw

PATCH v3 2/6 ima/ima_boot_aggregate: Don't hard code the length of sha1 hash

In reply to Jia Zhang

Instead, use SHA_DIGEST_LENGTH.

---
 testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c +5 -4
index d85d222..67be6a7 100644
@@ -42,7 +42,7 @@ static void display_sha1_digest(unsigned char *pcr)
 {
 	int i;
 
-	for (i = 0; i < 20; i++)
+	for (i = 0; i < SHA_DIGEST_LENGTH; i++)
 		printf("%02x", *(pcr + i) & 0xff);
 	printf("\n");
 }
@@ -94,8 +94,9 @@ int main(int argc, char *argv[])
 			display_sha1_digest(event.header.digest);
 		}
 		SHA1_Init(&c);
-		SHA1_Update(&c, pcr[event.header.pcr].digest, 20);
-		SHA1_Update(&c, event.header.digest, 20);
+		SHA1_Update(&c, pcr[event.header.pcr].digest,
+			    SHA_DIGEST_LENGTH);
+		SHA1_Update(&c, event.header.digest, SHA_DIGEST_LENGTH);
 		SHA1_Final(pcr[event.header.pcr].digest, &c);
 #if MAX_EVENT_DATA_SIZE < USHRT_MAX
 		if (event.header.len > MAX_EVENT_DATA_SIZE) {
@@ -116,7 +117,7 @@ int main(int argc, char *argv[])
 			printf("PCR-%2.2x: ", i);
 			display_sha1_digest(pcr[i].digest);
 		}
-		SHA1_Update(&c, pcr[i].digest, 20);
+		SHA1_Update(&c, pcr[i].digest, SHA_DIGEST_LENGTH);
 	}
 	SHA1_Final(boot_aggregate, &c);
 
Signature · 1 line
1.8.3.1
author
#5 lore raw

PATCH v3 3/6 ima/ima_boot_aggregate: Fix extending PCRs beyond PCR 0-7

In reply to Jia Zhang

The boot aggragate calculation should never touch PCRs beyond PCR 0-7, even a PCR extension really manipulates out-of-domain PCRs.

---
 .../security/integrity/ima/src/ima_boot_aggregate.c       | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
testcases/kernel/security/integrity/ima/src/ima_boot_aggregate.c +10 -5
index 67be6a7..98893b9 100644
@@ -93,11 +93,16 @@ int main(int argc, char *argv[])
 			printf("%03u ", event.header.pcr);
 			display_sha1_digest(event.header.digest);
 		}
-		SHA1_Init(&c);
-		SHA1_Update(&c, pcr[event.header.pcr].digest,
-			    SHA_DIGEST_LENGTH);
-		SHA1_Update(&c, event.header.digest, SHA_DIGEST_LENGTH);
-		SHA1_Final(pcr[event.header.pcr].digest, &c);
+
+		if (event.header.pcr < NUM_PCRS) {
+			SHA1_Init(&c);
+			SHA1_Update(&c, pcr[event.header.pcr].digest,
+				    SHA_DIGEST_LENGTH);
+			SHA1_Update(&c, event.header.digest,
+				    SHA_DIGEST_LENGTH);
+			SHA1_Final(pcr[event.header.pcr].digest, &c);
+		}
+
 #if MAX_EVENT_DATA_SIZE < USHRT_MAX
 		if (event.header.len > MAX_EVENT_DATA_SIZE) {
 			printf("Error event too long\n");
Signature · 1 line
1.8.3.1
author
#6 lore raw

PATCH v3 4/6 ima: Code cleanup

In reply to Jia Zhang

- Change the legacy function name from PATH_CHECK to FILE_CHECK.
- Use the variable IMA_POLICY instead of hard code path.

---
 testcases/kernel/security/integrity/ima/policy/measure.policy         | 2 +-
 testcases/kernel/security/integrity/ima/policy/measure.policy-invalid | 2 +-
 testcases/kernel/security/integrity/ima/tests/ima_policy.sh           | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
testcases/kernel/security/integrity/ima/policy/measure.policy +1 -1
index c68e722..9976ddf 100644
@@ -13,4 +13,4 @@ dont_measure fsmagic=0x01021994
 dont_measure fsmagic=0x73636673
 measure func=FILE_MMAP mask=MAY_EXEC
 measure func=BPRM_CHECK mask=MAY_EXEC
-measure func=PATH_CHECK mask=MAY_READ uid=0
+measure func=FILE_CHECK mask=MAY_READ uid=0
testcases/kernel/security/integrity/ima/policy/measure.policy-invalid +1 -1
index e406757..04dff89 100644
@@ -13,4 +13,4 @@ dont_measure fsmagic=0x01021994
 dnt_measure fsmagic=0x73636673
 measure func=FILE_MMAP mask=MAY_EXEC
 measure func=BPRM_CHECK mask=MAY_EXEC
-measure func=PATH_CHECK mask=MAY_READ uid=0
+measure func=FILE_CHECK mask=MAY_READ uid=0
testcases/kernel/security/integrity/ima/tests/ima_policy.sh +1 -1
index 64aa8cb..a0c7869 100755
@@ -28,7 +28,7 @@ check_policy_writable()
 {
 	local err="IMA policy already loaded and kernel not configured to enable multiple writes to it (need CONFIG_IMA_WRITE_POLICY=y)"
 
-	[ -f /sys/kernel/security/ima/policy ] || tst_brk TCONF "$err"
+	[ -f $IMA_POLICY ] || tst_brk TCONF "$err"
 	# CONFIG_IMA_READ_POLICY
 	echo "" 2> log > $IMA_POLICY
 	grep -q "Device or resource busy" log && tst_brk TCONF "$err"
Signature · 1 line
1.8.3.1
author
#7 lore raw

PATCH v3 5/6 ima: Rename the folder name for policy files to datafiles

In reply to Jia Zhang

If we choose to run ima_policy.sh locally without installation, a failure message is reported as following:

ima_policy 1 TCONF: missing <path>/ltp/testcases/kernel/security/integrity/ima/datafiles/measure.policy

TST_DATAROOT would be extended to datafiles but the policy files are actually placed under policy.

In order to make it easier, just rename the folder name to datafiles.

---
 testcases/kernel/security/integrity/ima/Makefile   |  2 +-
 .../security/integrity/ima/datafiles/Makefile      | 31 ++++++++++++++++++++++
 .../integrity/ima/datafiles/measure.policy         | 16 +++++++++++
 .../integrity/ima/datafiles/measure.policy-invalid | 16 +++++++++++
 .../kernel/security/integrity/ima/policy/Makefile  | 31 ----------------------
 .../security/integrity/ima/policy/measure.policy   | 16 -----------
 .../integrity/ima/policy/measure.policy-invalid    | 16 -----------
 7 files changed, 64 insertions(+), 64 deletions(-)
 create mode 100644 testcases/kernel/security/integrity/ima/datafiles/Makefile
 create mode 100644 testcases/kernel/security/integrity/ima/datafiles/measure.policy
 create mode 100644 testcases/kernel/security/integrity/ima/datafiles/measure.policy-invalid
 delete mode 100644 testcases/kernel/security/integrity/ima/policy/Makefile
 delete mode 100644 testcases/kernel/security/integrity/ima/policy/measure.policy
 delete mode 100644 testcases/kernel/security/integrity/ima/policy/measure.policy-invalid
testcases/kernel/security/integrity/ima/Makefile +1 -1
index 1290e6f..19b10ff 100644
@@ -24,6 +24,6 @@ top_srcdir		?= ../../../../..
 
 include $(top_srcdir)/include/mk/env_pre.mk
 
-SUBDIRS			:= policy src tests
+SUBDIRS			:= datafiles src tests
 
 include $(top_srcdir)/include/mk/generic_trunk_target.mk
testcases/kernel/security/integrity/ima/datafiles/Makefile new file +31 -0
new file mode 100644
index 0000000..a960f9d
@@ -0,0 +1,31 @@
+#
+#    testcases/kernel/security/integrity/ima/policy testcases Makefile.
+#
+#    Copyright (C) 2009, Cisco Systems Inc.
+#
+#    This program is free software; you can redistribute it and/or modify
+#    it under the terms of the GNU General Public License as published by
+#    the Free Software Foundation; either version 2 of the License, or
+#    (at your option) any later version.
+#
+#    This program is distributed in the hope that it will be useful,
+#    but WITHOUT ANY WARRANTY; without even the implied warranty of
+#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#    GNU General Public License for more details.
+#
+#    You should have received a copy of the GNU General Public License along
+#    with this program; if not, write to the Free Software Foundation, Inc.,
+#    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+# Ngie Cooper, July 2009
+#
+
+top_srcdir		?= ../../../../../..
+
+include	$(top_srcdir)/include/mk/env_pre.mk
+
+INSTALL_DIR		:= testcases/data/ima_policy
+
+INSTALL_TARGETS		:= measure*
+
+include $(top_srcdir)/include/mk/generic_leaf_target.mk
testcases/kernel/security/integrity/ima/datafiles/measure.policy new file +16 -0
new file mode 100644
index 0000000..9976ddf
@@ -0,0 +1,16 @@
+#
+# Integrity measure policy
+#
+# PROC_SUPER_MAGIC
+dont_measure fsmagic=0x9fa0
+# SYSFS_MAGIC
+dont_measure fsmagic=0x62656572
+# DEBUGFS_MAGIC
+dont_measure fsmagic=0x64626720
+# TMPFS_MAGIC
+dont_measure fsmagic=0x01021994
+# SECURITYFS_MAGIC
+dont_measure fsmagic=0x73636673
+measure func=FILE_MMAP mask=MAY_EXEC
+measure func=BPRM_CHECK mask=MAY_EXEC
+measure func=FILE_CHECK mask=MAY_READ uid=0
testcases/kernel/security/integrity/ima/datafiles/measure.policy-invalid new file +16 -0
new file mode 100644
index 0000000..04dff89
@@ -0,0 +1,16 @@
+#
+# Integrity measure policy
+#
+# PROC_SUPER_MAGIC
+dont_measure fsmagic=0x9fa0
+# SYSFS_MAGIC
+dont_measure fsmagic=0x62656572
+# DEBUGFS_MAGIC
+dont_measure fsmagic=0x64626720
+# TMPFS_MAGIC
+dont_measure fsmagic=0x01021994
+# SECURITYFS_MAGIC
+dnt_measure fsmagic=0x73636673
+measure func=FILE_MMAP mask=MAY_EXEC
+measure func=BPRM_CHECK mask=MAY_EXEC
+measure func=FILE_CHECK mask=MAY_READ uid=0
testcases/kernel/security/integrity/ima/policy/Makefile deleted +0 -31
deleted file mode 100644
index a960f9d..0000000
@@ -1,31 +0,0 @@
-#
-#    testcases/kernel/security/integrity/ima/policy testcases Makefile.
-#
-#    Copyright (C) 2009, Cisco Systems Inc.
-#
-#    This program is free software; you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation; either version 2 of the License, or
-#    (at your option) any later version.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License along
-#    with this program; if not, write to the Free Software Foundation, Inc.,
-#    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
-#
-# Ngie Cooper, July 2009
-#
-
-top_srcdir		?= ../../../../../..
-
-include	$(top_srcdir)/include/mk/env_pre.mk
-
-INSTALL_DIR		:= testcases/data/ima_policy
-
-INSTALL_TARGETS		:= measure*
-
-include $(top_srcdir)/include/mk/generic_leaf_target.mk
testcases/kernel/security/integrity/ima/policy/measure.policy deleted +0 -16
deleted file mode 100644
index 9976ddf..0000000
@@ -1,16 +0,0 @@
-#
-# Integrity measure policy
-#
-# PROC_SUPER_MAGIC
-dont_measure fsmagic=0x9fa0
-# SYSFS_MAGIC
-dont_measure fsmagic=0x62656572
-# DEBUGFS_MAGIC
-dont_measure fsmagic=0x64626720
-# TMPFS_MAGIC
-dont_measure fsmagic=0x01021994
-# SECURITYFS_MAGIC
-dont_measure fsmagic=0x73636673
-measure func=FILE_MMAP mask=MAY_EXEC
-measure func=BPRM_CHECK mask=MAY_EXEC
-measure func=FILE_CHECK mask=MAY_READ uid=0
testcases/kernel/security/integrity/ima/policy/measure.policy-invalid deleted +0 -16
deleted file mode 100644
index 04dff89..0000000
@@ -1,16 +0,0 @@
-#
-# Integrity measure policy
-#
-# PROC_SUPER_MAGIC
-dont_measure fsmagic=0x9fa0
-# SYSFS_MAGIC
-dont_measure fsmagic=0x62656572
-# DEBUGFS_MAGIC
-dont_measure fsmagic=0x64626720
-# TMPFS_MAGIC
-dont_measure fsmagic=0x01021994
-# SECURITYFS_MAGIC
-dnt_measure fsmagic=0x73636673
-measure func=FILE_MMAP mask=MAY_EXEC
-measure func=BPRM_CHECK mask=MAY_EXEC
-measure func=FILE_CHECK mask=MAY_READ uid=0
Signature · 1 line
1.8.3.1
#8 lore raw

Re: [PATCH v3 5/6] ima: Rename the folder name for policy files to datafiles

Hi Mimi, Jia,

Quoted 5 lines

If we choose to run ima_policy.sh locally without installation, a failure message is reported as following:

ima_policy 1 TCONF: missing <path>/ltp/testcases/kernel/security/integrity/ima/datafiles/measure.policy

Quoted 7 lines

TST_DATAROOT would be extended to datafiles but the policy files are actually placed under policy.

In order to make it easier, just rename the folder name to datafiles.

---

I wasn't sure about this one as make install isn't that hard to do even during debugging and policy is more descriptive. But other tests use datafiles as well, therefore taken as well.

Whole patchset merged, thank you both for your work.

Kind regards,
Petr

author
#9 lore raw

Re: [PATCH v3 5/6] ima: Rename the folder name for policy files to datafiles

Quoted 10 lines · Petr Vorel

Hi Mimi, Jia,

Quoted 7 lines

If we choose to run ima_policy.sh locally without installation, a failure message is reported as following:

ima_policy 1 TCONF: missing <path>/ltp/testcases/kernel/security/integrity/ima/datafiles/measure.policy

TST_DATAROOT would be extended to datafiles but the policy files are actually placed under policy.

Quoted 8 lines
Quoted 4 lines

In order to make it easier, just rename the folder name to datafiles.

---

I wasn't sure about this one as make install isn't that hard to do even during debugging and policy is more descriptive. But other tests use datafiles as well, therefore taken as well.

Actually, this patch resolves the issue if we just follow this step to run IMA-only test:

$ git clone https://github.com/linux-test-project/ltp.git
$ cd ltp
$ export LTP_ROOT="$PWD"
$ make autotools
$ ./configure
$ cd testcases/lib
$ make
$ cd ../kernel/security/integrity/ima
$ make
$ sudo PATH=$LTP_ROOT/testcases/lib:tests:$PATH ./tests/ima_policy.sh

A full LTP installation doesn't have this issue.

Jia

Quoted 6 lines

Whole patchset merged, thank you both for your work.

Kind regards,
Petr

#10 lore raw

PATCH v3 5/6 ima: Rename the folder name for policy files to datafiles

Hi Jia,

Quoted 18 lines
Quoted 3 lines

I wasn't sure about this one as make install isn't that hard to do even during debugging and policy is more descriptive. But other tests use datafiles as well, therefore taken as well.

Actually, this patch resolves the issue if we just follow this step to run IMA-only test:

$ git clone https://github.com/linux-test-project/ltp.git
$ cd ltp
$ export LTP_ROOT="$PWD"
$ make autotools
$ ./configure
$ cd testcases/lib
$ make
$ cd ../kernel/security/integrity/ima
$ make
$ sudo PATH=$LTP_ROOT/testcases/lib:tests:$PATH ./tests/ima_policy.sh
Quoted 1 line

A full LTP installation doesn't have this issue.

Just a side note: there is only one LTP installation: make install :). Your example is compilation and running test *without* installation.

The fact we require installation suggest runltp (still not replaced LTP runner):

$ ./runltp
FATAL: LTP not installed correctly
INFO:  Follow directions in INSTALL!

I usually do installation, also because I test patches on various VMs.

Kind regards,
Petr

author
#11 lore raw

PATCH v3 5/6 ima: Rename the folder name for policy files to datafiles

Quoted 24 lines · Petr Vorel

Hi Jia,

Quoted 18 lines
Quoted 3 lines

I wasn't sure about this one as make install isn't that hard to do even during debugging and policy is more descriptive. But other tests use datafiles as well, therefore taken as well.

Actually, this patch resolves the issue if we just follow this step to run IMA-only test:

$ git clone https://github.com/linux-test-project/ltp.git
$ cd ltp
$ export LTP_ROOT="$PWD"
$ make autotools
$ ./configure
$ cd testcases/lib
$ make
$ cd ../kernel/security/integrity/ima
$ make
$ sudo PATH=$LTP_ROOT/testcases/lib:tests:$PATH ./tests/ima_policy.sh
Quoted 1 line

A full LTP installation doesn't have this issue.

Just a side note: there is only one LTP installation: make install :). Your example is compilation and running test *without* installation.

Quoted 6 lines

The fact we require installation suggest runltp (still not replaced LTP runner):

$ ./runltp
FATAL: LTP not installed correctly
INFO:  Follow directions in INSTALL!

I usually do installation, also because I test patches on various VMs.

Yes. Actually I already switch to always do a full installation :)

Jia

Quoted 4 lines

Kind regards,
Petr

author
#12 lore raw

PATCH v3 6/6 ima/ima_violations: Temporarily remove the printk rate limit

In reply to Jia Zhang

The output frequency of audit log is limited by printk_ratelimit() in kernel if auditd not used. Thus, the test cases heavily depending on searching certain keywords in log file may fail if the matching patterns are exactly suppressed by printk_ratelimit().

In order to fix such a sort of failure, just temporarily remove the printk rate limit, and restore its original setting when doing cleanup.

---
 .../kernel/security/integrity/ima/tests/ima_setup.sh      |  2 +-
 .../kernel/security/integrity/ima/tests/ima_violations.sh | 15 +++++++++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)
testcases/kernel/security/integrity/ima/tests/ima_setup.sh +1 -1
index 6dfb4d2..fe60981 100644
@@ -20,7 +20,7 @@
 TST_TESTFUNC="test"
 TST_SETUP_CALLER="$TST_SETUP"
 TST_SETUP="ima_setup"
-TST_CLEANUP="ima_cleanup"
+TST_CLEANUP="${TST_CLEANUP:-ima_cleanup}"
 TST_NEEDS_TMPDIR=1
 TST_NEEDS_ROOT=1
 
testcases/kernel/security/integrity/ima/tests/ima_violations.sh +15 -0
index f3f40d4..74223c2 100755
@@ -20,6 +20,7 @@
 # Test whether ToMToU and open_writer violations invalidatethe PCR and are logged.
 
 TST_SETUP="setup"
+TST_CLEANUP="cleanup"
 TST_CNT=3
 TST_NEEDS_DEVICE=1
 
@@ -31,15 +32,29 @@ setup()
 	FILE="test.txt"
 	IMA_VIOLATIONS="$SECURITYFS/ima/violations"
 	LOG="/var/log/messages"
+	PRINTK_RATE_LIMIT="0"
 
 	if status_daemon auditd; then
 		LOG="/var/log/audit/audit.log"
+	else
+		tst_check_cmds sysctl
+
+		PRINTK_RATE_LIMIT=`sysctl -n kernel.printk_ratelimit`
+		sysctl -wq kernel.printk_ratelimit=0
 	fi
 	[ -f "$LOG" ] || \
 		tst_brk TBROK "log $LOG does not exist (bug in detection?)"
 	tst_res TINFO "using log $LOG"
 }
 
+cleanup()
+{
+	[ "$PRINTK_RATE_LIMIT" != "0" ] && \
+		sysctl -wq kernel.printk_ratelimit=$PRINTK_RATE_LIMIT
+
+	ima_cleanup
+}
+
 open_file_read()
 {
 	exec 3< $FILE || exit 1
Signature · 1 line
1.8.3.1
maintainer Reviewed-by on v3 6/6
#13 lore raw

Re: [PATCH v3 6/6] ima/ima_violations: Temporarily remove the printk rate limit

Quoted 10 lines · Jia Zhang

The output frequency of audit log is limited by printk_ratelimit() in kernel if auditd not used. Thus, the test cases heavily depending on searching certain keywords in log file may fail if the matching patterns are exactly suppressed by printk_ratelimit().

In order to fix such a sort of failure, just temporarily remove the printk rate limit, and restore its original setting when doing cleanup.

Thanks, I wasn't aware of the sysctl.  If the message isn't in /var/log/messages or /var/log/audit/audit.log, do we now need to also check journalctl?

Quoted 60 lines
---
 .../kernel/security/integrity/ima/tests/ima_setup.sh      |  2 +-
 .../kernel/security/integrity/ima/tests/ima_violations.sh | 15 +++++++++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)
testcases/kernel/security/integrity/ima/tests/ima_setup.sh +1 -1
index 6dfb4d2..fe60981 100644
@@ -20,7 +20,7 @@
 TST_TESTFUNC="test"
 TST_SETUP_CALLER="$TST_SETUP"
 TST_SETUP="ima_setup"
-TST_CLEANUP="ima_cleanup"
+TST_CLEANUP="${TST_CLEANUP:-ima_cleanup}"
 TST_NEEDS_TMPDIR=1
 TST_NEEDS_ROOT=1
 
testcases/kernel/security/integrity/ima/tests/ima_violations.sh +15 -0
index f3f40d4..74223c2 100755
@@ -20,6 +20,7 @@
 # Test whether ToMToU and open_writer violations invalidatethe PCR and are logged.
 
 TST_SETUP="setup"
+TST_CLEANUP="cleanup"
 TST_CNT=3
 TST_NEEDS_DEVICE=1
 
@@ -31,15 +32,29 @@ setup()
 	FILE="test.txt"
 	IMA_VIOLATIONS="$SECURITYFS/ima/violations"
 	LOG="/var/log/messages"
+	PRINTK_RATE_LIMIT="0"
 
 	if status_daemon auditd; then
 		LOG="/var/log/audit/audit.log"
+	else
+		tst_check_cmds sysctl
+
+		PRINTK_RATE_LIMIT=`sysctl -n kernel.printk_ratelimit`
+		sysctl -wq kernel.printk_ratelimit=0
 	fi
 	[ -f "$LOG" ] || \
 		tst_brk TBROK "log $LOG does not exist (bug in detection?)"
 	tst_res TINFO "using log $LOG"
 }
 
+cleanup()
+{
+	[ "$PRINTK_RATE_LIMIT" != "0" ] && \
+		sysctl -wq kernel.printk_ratelimit=$PRINTK_RATE_LIMIT
+
+	ima_cleanup
+}
+
 open_file_read()
 {
 	exec 3< $FILE || exit 1
author
#14 lore raw

PATCH v3 6/6 ima/ima_violations: Temporarily remove the printk rate limit

Quoted 10 lines · Mimi Zohar
Quoted 9 lines · Jia Zhang

The output frequency of audit log is limited by printk_ratelimit() in kernel if auditd not used. Thus, the test cases heavily depending on searching certain keywords in log file may fail if the matching patterns are exactly suppressed by printk_ratelimit().

In order to fix such a sort of failure, just temporarily remove the printk rate limit, and restore its original setting when doing cleanup.

Quoted 5 lines
Quoted 1 line

Thanks, I wasn't aware of the sysctl.  If the message isn't in /var/log/messages or /var/log/audit/audit.log, do we now need to also check journalctl?

Not necessary. If user land's auditd is not used, kauditd will send the log to /var/log/message with printk rate limit:

static void kauditd_printk_skb(struct sk_buff *skb)

{
        struct nlmsghdr *nlh = nlmsg_hdr(skb);
        char *data = nlmsg_data(nlh);

        if (nlh->nlmsg_type != AUDIT_EOE && printk_ratelimit())
                pr_notice("type=%d %s\n", nlh->nlmsg_type, data);
}

So just work around the potential loss with sysctl to temporarily disable the limit.

Jia

Quoted 63 lines
Quoted 60 lines
---
 .../kernel/security/integrity/ima/tests/ima_setup.sh      |  2 +-
 .../kernel/security/integrity/ima/tests/ima_violations.sh | 15 +++++++++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)
testcases/kernel/security/integrity/ima/tests/ima_setup.sh +1 -1
index 6dfb4d2..fe60981 100644
@@ -20,7 +20,7 @@
 TST_TESTFUNC="test"
 TST_SETUP_CALLER="$TST_SETUP"
 TST_SETUP="ima_setup"
-TST_CLEANUP="ima_cleanup"
+TST_CLEANUP="${TST_CLEANUP:-ima_cleanup}"
 TST_NEEDS_TMPDIR=1
 TST_NEEDS_ROOT=1
 
testcases/kernel/security/integrity/ima/tests/ima_violations.sh +15 -0
index f3f40d4..74223c2 100755
@@ -20,6 +20,7 @@
 # Test whether ToMToU and open_writer violations invalidatethe PCR and are logged.
 
 TST_SETUP="setup"
+TST_CLEANUP="cleanup"
 TST_CNT=3
 TST_NEEDS_DEVICE=1
 
@@ -31,15 +32,29 @@ setup()
 	FILE="test.txt"
 	IMA_VIOLATIONS="$SECURITYFS/ima/violations"
 	LOG="/var/log/messages"
+	PRINTK_RATE_LIMIT="0"
 
 	if status_daemon auditd; then
 		LOG="/var/log/audit/audit.log"
+	else
+		tst_check_cmds sysctl
+
+		PRINTK_RATE_LIMIT=`sysctl -n kernel.printk_ratelimit`
+		sysctl -wq kernel.printk_ratelimit=0
 	fi
 	[ -f "$LOG" ] || \
 		tst_brk TBROK "log $LOG does not exist (bug in detection?)"
 	tst_res TINFO "using log $LOG"
 }
 
+cleanup()
+{
+	[ "$PRINTK_RATE_LIMIT" != "0" ] && \
+		sysctl -wq kernel.printk_ratelimit=$PRINTK_RATE_LIMIT
+
+	ima_cleanup
+}
+
 open_file_read()
 {
 	exec 3< $FILE || exit 1
#15 lore raw
In reply to Jia Zhang

Hi Jia, Mimi,

Quoted 2 lines

Changelog since V2:
- Rework patch 6. Here is a quick diff against V2:

LGTM, just waiting if Mimi reviews 1st and 6th commit.

Kind regards,
Petr

author
#16 lore raw

Hi Mini,

Could you help to take a look at this V3 patch series?

Thanks,
Jia

Quoted 10 lines · Petr Vorel

Hi Jia, Mimi,

Quoted 2 lines

Changelog since V2:
- Rework patch 6. Here is a quick diff against V2:

LGTM, just waiting if Mimi reviews 1st and 6th commit.

Kind regards,
Petr

author
#17 lore raw
In reply to Petr Vorel

Hi Petr,

Is it good enough to merge this patch series?

Thanks,
Jia

Quoted 10 lines · Petr Vorel

Hi Jia, Mimi,

Quoted 2 lines

Changelog since V2:
- Rework patch 6. Here is a quick diff against V2:

LGTM, just waiting if Mimi reviews 1st and 6th commit.

Kind regards,
Petr