Skip to contentSkip to conversation

Repost of tpm driver suspend/resume test

PATCH In review Stefan Berger 4 participants 5 messages Mar 14, 2006 – Mar 15, 2006

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Repost of tpm driver suspend/resume test

This is a repost of a previously posted patch to test suspend and resume cycles of the TPM front- and backend.
Now that sysfs is mounted in the RAM disk of the test environment, the test code does not need to do this anymore. This patch also simplifies previous tests and the APIs used there.

xen-unstable.hg/tools/xm-test/tests/vtpm/03_vtpm-susp_res.py new file +120 -0
Index: xen/xen-unstable.hg/tools/xm-test/tests/vtpm/03_vtpm-susp_res.py
===================================================================
@@ -0,0 +1,120 @@
+#!/usr/bin/python
+
+# Copyright (C) International Business Machines Corp., 2006
+# Author: Stefan Berger <[email protected]>
+
+# Positive Test: create domain with virtual TPM attached at build time,
+#                check list of pcrs; suspend and resume the domain and
+#                check list of pcrs again
+
+from XmTestLib import *
+import commands
+import os
+import os.path
+
+def vtpm_cleanup(domName):
+    # Since this is only a temporary domain I clean up the domain from the
+    # virtual TPM directory
+    os.system("/etc/xen/scripts/vtpm-delete %s" % domName)
+
+if ENABLE_HVM_SUPPORT:
+    SKIP("vtpm-list not supported for HVM domains")
+
+if os.path.exists("/dev/tpm0") == False:
+    SKIP("This machine has no hardware TPM; cannot run this test")
+
+output = commands.getoutput("ps aux | grep vtpm_manager | grep -v grep")
+if output == "":
+    FAIL("virtual TPM manager must be started to run this test")
+
+# vtpm manager has been detected
+config = {"vtpm":"instance=1,backend=0"}
+domain = XmTestDomain(extraConfig=config)
+
+try:
+    domain.start()
+except DomainError, e:
+    if verbose:
+        print e.extra
+    vtpm_cleanup(domain.getName())
+    FAIL("Unable to create domain")
+
+domName = domain.getName()
+
+try:
+    console = XmConsole(domain.getName())
+except ConsoleError, e:
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+try:
+    console.sendInput("input")
+except ConsoleError, e:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+try:
+    run = console.runCmd("cat /sys/devices/platform/tpm_vtpm/pcrs")
+except ConsoleError, e:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+if re.search("No such file",run["output"]):
+    FAIL("TPM frontend support not compiled into (domU?) kernel")
+
+console.closeConsole()
+
+try:
+    status, ouptut = traceCommand("xm save %s %s.save" %
+                                  (domName, domName),
+                                  timeout=30)
+except TimeoutError, e:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+if status != 0:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL("xm save did not succeed")
+
+try:
+    status, ouptut = traceCommand("xm restore %s.save" %
+                                  (domName),
+                                  timeout=30)
+except TimeoutError, e:
+    os.remove("%s.save" % domName)
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+os.remove("%s.save" % domName)
+
+if status != 0:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL("xm restore did not succeed")
+
+try:
+    console = XmConsole(domain.getName())
+except ConsoleError, e:
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+try:
+    run = console.runCmd("cat /sys/devices/platform/tpm_vtpm/pcrs")
+except ConsoleError, e:
+    saveLog(console.getHistory())
+    vtpm_cleanup(domName)
+    FAIL(str(e))
+
+console.closeConsole()
+
+domain.stop()
+
+vtpm_cleanup(domName)
+
+if not re.search("PCR-00:",run["output"]):
+	FAIL("Virtual TPM is not working correctly on /dev/vtpm on backend side")
xen-unstable.hg/tools/xm-test/tests/vtpm/Makefile.am +2 -1
Index: xen/xen-unstable.hg/tools/xm-test/tests/vtpm/Makefile.am
===================================================================
@@ -2,7 +2,8 @@
 SUBDIRS =
 
 TESTS = 01_vtpm-list_pos.test \
-        02_vtpm-cat_pcrs.test
+        02_vtpm-cat_pcrs.test \
+        03_vtpm-susp_res.test
 
 XFAIL_TESTS =
 
xen-unstable.hg/tools/xm-test/tests/vtpm/01_vtpm-list_pos.py +6 -4
Index: xen/xen-unstable.hg/tools/xm-test/tests/vtpm/01_vtpm-list_pos.py
===================================================================
@@ -1,18 +1,20 @@
 #!/usr/bin/python
 
 # Copyright (C) International Business Machines Corp., 2006
-# Author: Stefan Berger <[email protected])
+# Author: Stefan Berger <[email protected]>
 
 # Positive Test: create domain with virtual TPM attached at build time,
 #                verify list
 
 
 from XmTestLib import *
+import commands
+import os
 
 def vtpm_cleanup(domName):
-	# Since this is only a temporary domain I clean up the domain from the
-	# virtual TPM directory
-	traceCommand("/etc/xen/scripts/vtpm-delete %s" % domName)
+    # Since this is only a temporary domain I clean up the domain from the
+    # virtual TPM directory
+    os.system("/etc/xen/scripts/vtpm-delete %s" % domName)
 
 if ENABLE_HVM_SUPPORT:
     SKIP("vtpm-list not supported for HVM domains")
xen-unstable.hg/tools/xm-test/tests/vtpm/02_vtpm-cat_pcrs.py +9 -17
Index: xen/xen-unstable.hg/tools/xm-test/tests/vtpm/02_vtpm-cat_pcrs.py
===================================================================
@@ -1,26 +1,28 @@
 #!/usr/bin/python
 
 # Copyright (C) International Business Machines Corp., 2006
-# Author: Stefan Berger <[email protected])
+# Author: Stefan Berger <[email protected]>
 
 # Positive Test: create domain with virtual TPM attached at build time,
 #                check list of pcrs
 
 from XmTestLib import *
+import commands
+import os
+import os.path
 
 def vtpm_cleanup(domName):
-	# Since this is only a temporary domain I clean up the domain from the
-	# virtual TPM directory
-	traceCommand("/etc/xen/scripts/vtpm-delete %s" % domName)
+    # Since this is only a temporary domain I clean up the domain from the
+    # virtual TPM directory
+    os.system("/etc/xen/scripts/vtpm-delete %s" % domName)
 
 if ENABLE_HVM_SUPPORT:
     SKIP("vtpm-list not supported for HVM domains")
 
-status, output = traceCommand("ls /dev/tpm0")
-if re.search("No such file or directory",output):
+if os.path.exists("/dev/tpm0") == False:
     SKIP("This machine has no hardware TPM; cannot run this test")
 
-status, output = traceCommand("ps aux | grep vtpm_manager | grep -v grep")
+output = commands.getoutput("ps aux | grep vtpm_manager | grep -v grep")
 if output == "":
     FAIL("virtual TPM manager must be started to run this test")
 
@@ -46,21 +48,11 @@ except ConsoleError, e:
 
 try:
     console.sendInput("input")
-    run = console.runCmd("ls /sys")
 except ConsoleError, e:
     saveLog(console.getHistory())
     vtpm_cleanup(domName)
     FAIL(str(e))
 
-if re.search("No such file",run["output"]):
-    try:
-        run = console.runCmd("mkdir /sys")
-        run = console.runCmd("mount -t sysfs /sys /sys")
-    except ConsoleError, e:
-        saveLog(console.getHistory())
-        vtpm_cleanup(domName)
-        FAIL(str(e))
-
 try:
     run = console.runCmd("cat /sys/devices/platform/tpm_vtpm/pcrs")
 except ConsoleError, e:

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#1 lore raw
Quoted 5 lines · Stefan Berger

This is a repost of a previously posted patch to test suspend and resume cycles of the TPM front- and backend.
Now that sysfs is mounted in the RAM disk of the test environment, the test code does not need to do this anymore. This patch also simplifies previous tests and the APIs used there.

Woody Marvel and Dan Stekloff still seem to be discussing the best way to detect sysfs capabilities. I presume your patch depends upon Dan's, and I've been holding that off until I get a sysfs-detection patch that everyone's happy with.

Ewan.

#2 lore raw
Quoted 8 lines · Ewan Mellor
Quoted 5 lines · Stefan Berger

This is a repost of a previously posted patch to test suspend and resume cycles of the TPM front- and backend.
Now that sysfs is mounted in the RAM disk of the test environment, the test code does not need to do this anymore. This patch also simplifies previous tests and the APIs used there.

Quoted 4 lines

Woody Marvel and Dan Stekloff still seem to be discussing the best way to detect sysfs capabilities. I presume your patch depends upon Dan's, and I've been holding that off until I get a sysfs-detection patch that everyone's happy with.

Hi Ewan,

Please put the sysfs patch in as is, Woody is fine with it.

Thanks,

Dan

#3 lore raw

Sorry Ewan if there was any mis-communications, the patch is fine. I was probing for greater understanding. Thanks.

wm
==============

Quoted 23 lines · Daniel Stekloff
Quoted 8 lines · Ewan Mellor
Quoted 5 lines · Stefan Berger

This is a repost of a previously posted patch to test suspend and resume cycles of the TPM front- and backend.
Now that sysfs is mounted in the RAM disk of the test environment, the test code does not need to do this anymore. This patch also simplifies previous tests and the APIs used there.

Quoted 4 lines

Woody Marvel and Dan Stekloff still seem to be discussing the best way to detect sysfs capabilities. I presume your patch depends upon Dan's, and I've been holding that off until I get a sysfs-detection patch that everyone's happy with.

Hi Ewan,

Please put the sysfs patch in as is, Woody is fine with it.

Thanks,

Dan

#4 lore raw
In reply to Ewan Mellor

[email protected] wrote on 03/15/2006 04:58:41 AM:

Quoted 3 lines · Stefan Berger
Quoted 2 lines

This is a repost of a previously posted patch to test suspend and

resume

Quoted 6 lines
Quoted 4 lines

cycles of the TPM front- and backend.
Now that sysfs is mounted in the RAM disk of the test environment, the test code does not need to do this anymore. This patch also simplifies previous tests and the APIs used there.

Woody Marvel and Dan Stekloff still seem to be discussing the best way

to

Quoted 1 line

detect sysfs capabilities. I presume your patch depends upon Dan's, and

I've

Quoted 1 line

been holding that off until I get a sysfs-detection patch that

everyone's

Quoted 1 line

happy with.

Actually, if you haven't put it in, yet, let me know and I will change the patch to result in a 'SKIP' if the vtpm_manager process as is not running.

Stefan

Quoted 7 lines

Ewan.

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel