Merge pull request #14 from arthurscchan/Development

Merge non-deterministic handling to master branch

Author: arthurscchan

benchmarkProgram/concurrentSyscall/grpReadWrite/cmdForkReadWrite/prepare

@@ -18,32 +18,25 @@ CODE=$(cat <<EOF
 #include <stdlib.h>
 #include <fcntl.h>
 #include <unistd.h>
-int main() { srand(time(NULL));
-if (fork()) {int id=open("`realpath $1`/test.txt", O_RDWR); char buf[1]; int i;
-	#ifdef READ
-	for (i=0;i<READ;i++) {
-		#ifdef RANDOM
-		sleep(rand()%20);
-		#endif
-		#ifdef PROGRAM
-		read(id, buf, 1);
-		#endif
-	}
+int main() {
+int id=open("`realpath $1`/test.txt", O_RDWR); char buf[1]; int i;
+fork();
+#ifdef READ
+for (i=0;i<READ;i++) {
+	#ifdef PROGRAM
+	read(id, buf, 1);
 	#endif
-	close(id);
-} else {int id=open("`realpath $1`/test.txt", O_RDWR); int i;
-	#ifdef WRITE
-	for (i=0;i<WRITE;i++) {
-		#ifdef RANDOM
-		sleep(rand()%20);
-		#endif
-		#ifdef PROGRAM
-		write(id, "TEST", 1);
-		#endif
-	}
+}
+#endif
+#ifdef WRITE
+for (i=0;i<WRITE;i++) {
+	#ifdef PROGRAM
+	write(id, "TEST", 1);
 	#endif
-	close(id);
-}}
+}
+#endif
+close(id);
+}
 EOF
 )
 

config/config.ini

@@ -1,19 +1,15 @@
 [spg]
 stage1tool = startSpade.py -d
 stage2handler = dot2Clingo.py
-template = output.dot-%%s
 
 [spn]
 stage1tool = startSpade.py -n
 stage2handler = neo4j2Clingo.py -d
-template = output.db-%%s
 
 [opu]
 stage1tool = startOpus.py
 stage2handler = neo4j2Clingo.py -d
-template = output.db-%%s
 
 [cam]
 stage1tool = startCamflow.py
 stage2handler = provJson2Clingo.py
-template = output.provjson-%%s

fullAutomation.py

@@ -35,6 +35,9 @@ def prepareDir(directory):
 	trial = 2
 baseDir = os.path.abspath(os.path.dirname(sys.argv[0]))
 tool = sys.argv[1]
+if tool != 'cam':
+	print ('Temporary disable spade and opus for new concurrency testing, please revert to git version a28cc39585c6e74dbeccfc3dccfeef537857ff22 for testing spade and opus in a stable version')
+	quit()
 toolBaseDir = os.path.abspath(sys.argv[2])
 benchmarkDir = os.path.abspath(sys.argv[3])
 stageDir = os.path.abspath('%s/stage/' % baseDir)
@@ -51,19 +54,22 @@ def prepareDir(directory):
 
 stage1Tool = config[tool]['stage1tool']
 stage2Handler = config[tool]['stage2handler']
-template = config[tool]['template']
 
 #Stage 1 - Start the tools and generate graph (neo4j / dot / provjson)
 start = time.time()
 print ('Starting stage 1...Generating provenance from native tools')
 
 os.system('sudo chmod +x %s/startTool/%s' % (baseDir, stage1Tool.split()[0]))
 stage1Command = 'sudo %s/startTool/%s %s %s %s %s %s %s %d' % (baseDir, stage1Tool, stageDir, workingDir, '%s' , '%s', toolBaseDir , '%s', trial)
+
 print ('Program')
-subprocess.call((stage1Command % (benchmarkDir, 'PROGRAM,RANDOM,READ=2,WRITE=2', 'program')).split())
+#programFingerprint = subprocess.check_output((stage1Command % (benchmarkDir, 'PROGRAM,RANDOM,READ=2,WRITE=2', 'program')).split())
+programFingerprint = subprocess.check_output((stage1Command % (benchmarkDir, 'PROGRAM,READ=2,WRITE=2', 'program')).split()).decode().split()
 print ('End Program')
+
 print ('Control')
-subprocess.call((stage1Command % (benchmarkDir, 'CONTROL,RANDOM,READ=2,WRITE=2', 'control')).split())
+#controlFingerprint = subprocess.check_output((stage1Command % (benchmarkDir, 'CONTROL,RANDOM,READ=2,WRITE=2', 'control')).split())
+controlFingerprint = subprocess.check_output((stage1Command % (benchmarkDir, 'CONTROL,READ=2,WRITE=2', 'control')).split()).decode().split()
 print ('End Control')
 
 print ('End of stage 1\n')
@@ -75,13 +81,26 @@ def prepareDir(directory):
 print ('Starting stage 2...Transforming provenance result to Clingo graph')
 
 os.system('sudo chmod +x %s/genClingoGraph/%s' % (baseDir, stage2Handler.split()[0]))
-stage2Command = 'sudo %s/genClingoGraph/%s %s %s %s' % (baseDir, stage2Handler, '%s', template, workingDir)
-for i in range(1,trial+1):
-	suffix = 'control-%d' % i
-	subprocess.call((stage2Command % (suffix,suffix)).split())
-for i in range(1,trial+1):
-	suffix = 'program-%d' % i
-	subprocess.call((stage2Command % (suffix,suffix)).split())
+stage2Command = 'sudo %s/genClingoGraph/%s %s %s %s' % (baseDir, stage2Handler, '%s', '%s', '%s')
+
+for fingerprint in controlFingerprint:
+	dir = os.path.abspath('%s/%s' % (workingDir, fingerprint))
+	if os.path.isdir(dir):
+		i = 1
+		for file in os.listdir(dir):
+			suffix = 'control-%d' % i
+			subprocess.call((stage2Command % (suffix,file,dir)).split())
+			i += 1
+
+for fingerprint in programFingerprint:
+	dir = os.path.abspath('%s/%s' % (workingDir, fingerprint))
+	if os.path.isdir(dir):
+		i = 1
+		for file in os.listdir(dir):
+			suffix = 'program-%d' % i
+			subprocess.call((stage2Command % (suffix,file,dir)).split())
+			i += 1
+
 print ('End of stage 2\n')
 end = time.time()
 t2 = end-start
@@ -93,32 +112,52 @@ def prepareDir(directory):
 os.system('sudo chmod +x %s/processGraph/generalizeGraph.py' % baseDir)
 stage3Command = 'sudo %s/processGraph/generalizeGraph.py %s %s %s' % (baseDir, workingDir, ('%s/processGraph/template.lp' % baseDir), '%s')
 
-command = stage3Command % ('control %s')
-for i in range(1,trial+1):
-	suffix = 'control-%d' % i
-	command = command % ('%s/clingo-%s %s' % (workingDir, suffix, '%s'))
-subprocess.call((command % ' ').split())
-
-command = stage3Command % ('program %s')
-for i in range(1,trial+1):
-	suffix = 'program-%d' % i
-	command = command % ('%s/clingo-%s %s' % (workingDir, suffix,'%s'))
-subprocess.call((command % '').split())
+for fingerprint in controlFingerprint:
+	dir = os.path.abspath('%s/%s' % (workingDir, fingerprint))
+	if os.path.isdir(dir):
+		command = stage3Command % ('control-%s %s' % (fingerprint, '%s'))
+		for file in ('%s/%s' % (dir,name) for name in os.listdir(dir) if name.startswith('clingo-control')):
+			command = command % ('%s %s' % (file, '%s'))
+		subprocess.call((command % '').split())
+
+for fingerprint in programFingerprint:
+	dir = os.path.abspath('%s/%s' % (workingDir, fingerprint))
+	if os.path.isdir(dir):
+		command = stage3Command % ('program-%s %s' % (fingerprint, '%s'))
+		for file in ('%s/%s' % (dir,name) for name in os.listdir(dir) if name.startswith('clingo-program')):
+			command = command % ('%s %s' % (file, '%s'))
+		subprocess.call((command % '').split())
 
 print ('End of stage 3\n')
 end = time.time()
 t3 = end-start
 
-shutil.copyfile('%s/general.clingo-program' % workingDir, '%s/general.clingo-program' % outDir)
-shutil.copyfile('%s/general.clingo-control' % workingDir, '%s/general.clingo-control' % outDir)
+for file in (file for file in os.listdir(workingDir) if file.startswith('general.clingo')):
+	shutil.copyfile('%s/%s' % (workingDir,file), '%s/%s' % (outDir,file))
 
 #Stage 4 - Compare and generate benchmark
 start = time.time()
 print ('Starting stage 4...Generating benchmark')
 
 os.system('sudo chmod +x %s/processGraph/findSubgraph.py' % baseDir)
-stage4Command = 'sudo %s/processGraph/findSubgraph.py %s %s 1 general.clingo-control general.clingo-program %s' % (baseDir, workingDir, ('%s/processGraph/template.lp' % baseDir), ('%s/result.clingo' % outDir))
-subprocess.call(stage4Command.split())
+for fingerprint in programFingerprint:
+	if len(controlFingerprint) > 1:
+		editDistance = sys.maxsize 
+		for backgroundFingerprint in controlFingerprint:
+			preStage4Command = '''sudo %s/processGraph/calEditDistance.py %s %s/general.clingo-control-%s 
+			%s/general.clingo-program-%s''' % (baseDir, ('''%s/processGraph/editdist.lp
+			''' % baseDir), workingDir, backgroundFingerprint, workingDir, fingerprint)
+			newEditDistance = int(subprocess.check_output(preStage4Command.split()))
+			if newEditDistance < editDistance:
+				editDistance = newEditDistance
+				background = backgroundFingerprint
+	else:
+		background = controlFingerprint[0]	
+	
+	stage4Command = '''sudo %s/processGraph/findSubgraph.py %s %s 1 general.clingo-control-%s general.clingo-program-%s %s
+        ''' % (baseDir, workingDir, ('%s/processGraph/template.lp' % baseDir), background, fingerprint, ('''%s/result-%s.clingo
+        ''' % (outDir,fingerprint)))
+	subprocess.call(stage4Command.split())
 
 print ('End of stage 4\n')
 end = time.time()

processGraph/__pycache__/clingoFunction.cpython-36.pyc

@@ -16,4 +16,4 @@
 
 result = dict()
 
-print (processGraph(sys.argv[2], sys.argv[3], clingoCode, False))
+print (processGraph(sys.argv[2], sys.argv[3], clingoCode, os.path.abspath('%s/../' % os.path.dirname(sys.argv[0])), False))

processGraph/calEditDistance.py

@@ -125,7 +125,7 @@ def compareProps(graph1Props, graph2Props, isGeneral):
 	return result
 
 #Graph Process
-def processGraph(graph1Path, graph2Path, clingoCode, isMapping):
+def processGraph(graph1Path, graph2Path, clingoCode, baseDir, isMapping):
 	#Read Graph
 	file = open(graph1Path, 'r')
 	graph1 = fixIdentifier(file.read(), 1)
@@ -142,7 +142,7 @@ def processGraph(graph1Path, graph2Path, clingoCode, isMapping):
 
 	#Clingo Operation
 	inputString = '%s\n%s\n%s'%(clingoCode, graph1, graph2)
-	pipe = subprocess.Popen(['../clingo/clingo', '--time-limit=30'], stdout=subprocess.PIPE, stdin=subprocess.PIPE, stderr=subprocess.STDOUT)
+	pipe = subprocess.Popen(['%s/clingo/clingo' % baseDir, '--time-limit=30'], stdout=subprocess.PIPE, stdin=subprocess.PIPE, stderr=subprocess.STDOUT)
 	mapResult = pipe.communicate(input=inputString.encode())[0]
 	if isMapping:
 		map = decodeClingoResult(mapResult.decode())

processGraph/clingoFunction.py

@@ -21,7 +21,7 @@
 os.chdir(workingDir)
 
 #Process Graph
-graph2Node, graph2Edge, graph1Props, graph2Props, map = processGraph(sys.argv[4], sys.argv[5], clingoCode, True)
+graph2Node, graph2Edge, graph1Props, graph2Props, map = processGraph(sys.argv[4], sys.argv[5], clingoCode, os.path.abspath('../'), True)
 
 # Properties are present if present in graph 2 and not in linked part of graph 1
 diffProps = dict()

processGraph/findSubgraph.py

@@ -28,7 +28,7 @@
 	quit()
 
 for i in range(4,len(sys.argv)-1):
-	graph2Node, graph2Edge, graph1Props, graph2Props, map = processGraph(sys.argv[i], sys.argv[i+1], clingoCode, True)
+	graph2Node, graph2Edge, graph1Props, graph2Props, map = processGraph(sys.argv[i], sys.argv[i+1], clingoCode, os.path.abspath('../'), True)
 
 	#Generalize properties
 	if map:

processGraph/generalizeGraph.py

@@ -1,11 +1,13 @@
 #!/usr/bin/env python3
 
 import os
+import re
 import sys
 import time
 import shutil
 import subprocess
 import json
+import hashlib
 
 #Merger Json
 def mergeJson(base, line):
@@ -92,14 +94,21 @@ def startCamflow(stagePath, workingPath, suffix, isModel):
 	subprocess.call('camflow --duplicate true'.split())
 	subprocess.call('camflow -e true'.split())
 #	subprocess.call('camflow -a true'.split())
-	os.system('%s/test' % stagePath)
+	subprocess.call(('trace-cmd record -e syscalls %s/test' % stagePath).split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 #	subprocess.call('camflow -a false'.split())
 	subprocess.call('camflow -e false'.split())
 #	subprocess.call('camflow --duplicate false'.split())
 	subprocess.call(('camflow --track-file %s/test false' % stagePath).split())
 	time.sleep(1)
 	subprocess.call('service camflowd stop'.split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
+	#Handle FTrace Fingerprint
+	ftraceResult = subprocess.check_output('trace-cmd report'.split())
+	if ftraceResult:
+		syscallList = [line.split(':')[1].strip() for line in ftraceResult.decode('ascii').split('\n') if re.match(r'^\s*test-((?!wait4).)*$',line)]
+	
+	fingerprint = hashlib.md5(''.join(syscallList).encode()).hexdigest()
+
 	#Process provenance result into 1 json
 	result={}
 	try:
@@ -124,9 +133,14 @@ def startCamflow(stagePath, workingPath, suffix, isModel):
 #	file.write(json.dumps(mergeNode(oldNode,result)))
 #	file.close()
 
+	#Handle fingerprint folder
+	if not os.path.exists('%s/%s' %(workingPath, fingerprint)):
+		os.makedirs('%s/%s' %(workingPath, fingerprint))
+		os.chown('%s/%s' %(workingPath, fingerprint), 1000, 1000)
+
 	if not isModel:
 		#Writing result to json
-		file = open('%s/output.provjson-%s' %(workingPath, suffix), 'w')
+		file = open('%s/%s/output.provjson-%s' %(workingPath, fingerprint, suffix), 'w')
 		file.write(json.dumps(result))
 		file.close()
 
@@ -135,6 +149,7 @@ def startCamflow(stagePath, workingPath, suffix, isModel):
 	except IOError:
 		pass
 
+	return fingerprint
 
 #Retrieve arguments
 trial = 0
@@ -160,12 +175,16 @@ def startCamflow(stagePath, workingPath, suffix, isModel):
 for item in macroOpt.split(','):
         gccMacro = "%s -D%s" %(gccMacro,item)
 
+fingerprintSet = set()
+
 #Create Model Data
 #subprocess.check_output(('%s/prepare %s %s --static' %(progPath, stagePath, gccMacro)).split())
 #startCamflow(stagePath, workingPath, '', True)
 
 for i in range(1, trial+1):
 	#Prepare the benchmark program
 	subprocess.check_output(('%s/prepare %s %s --static' %(progPath, stagePath, gccMacro)).split())
-	startCamflow(stagePath, workingPath, '%s-%d' % (suffix, i), False)
-		
+	fingerprintSet.add(startCamflow(stagePath, workingPath, '%s-%d' % (suffix, i), False))
+	
+for fingerprint in fingerprintSet:
+	print (fingerprint)

startTool/startCamflow.py

@@ -1,8 +1,10 @@
 #!/usr/bin/env python3
 
 import os
+import re
 import sys
 import time
+import hashlib
 import subprocess
 
 #Retrieve arguments
@@ -35,6 +37,8 @@
 
 os.chdir(workingPath)
 
+fingerprintSet = set()
+
 #Get Audit Log
 for i in range(1, trial+1):
 	#Prepare the benchmark program
@@ -69,10 +73,25 @@
 	subprocess.call((baseCommand % 'server start').split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
 	#Capture Provenance
-	subprocess.call((baseCommand % ('process launch %s/test' % stagePath)).split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+	subprocess.call((baseCommand % ('process launch trace-cmd record -e syscalls %s/test' % stagePath)).split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
 	#Stop Opus Server
 	subprocess.call((baseCommand % 'server stop').split(), stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
-	#Wait for Data Flushing
-	time.sleep(2)
+	#Handle FTrace Fingerprint
+	ftraceResult = subprocess.check_output('trace-cmd report'.split())
+	if ftraceResult:
+		syscallList = [line.split(':')[1].strip() for line in ftraceResult.decode('ascii').split('\n') if re.match(r'^\s*test-((?!wait4).)*$',line)]
+
+	fingerprint = hashlib.md5(''.join(syscallList).encode()).hexdigest()
+	fingerprintSet.add(fingerprint)
+
+	#Handle fingerprint folder
+	if not os.path.exists('%s/%s' %(workingPath, fingerprint)):
+		os.makedirs('%s/%s' %(workingPath, fingerprint))
+		os.chown('%s/%s' %(workingPath, fingerprint), 1000, 1000)
+
+	os.rename('%s/output.db-%s-%d' % (workingPath,suffix,i),'%s/%s/output.db-%s-%d' % (workingPath,fingerprint,suffix,i))
+
+for fingerprint in fingerprintSet:
+	print(fingerprint)