: eclone-1: Test basic functionality

From: Sukadev Bhattiprolu <[email protected]>
Date: Sat, 30 Jan 2010 12:49:30 -0800
Subject: [PATCH 1/6] eclone-1: Test basic functionality

Verify that a child process gets the expected pid and arguments on stack when it is created with eclone() system call.

Changelog[v2]:
- Use libeclone.a from user-cr git tree so tests are portable across architectures.
- Fix a couple of nits identified by Serge Hallyn

---
 Makefile              |    2 +-
 eclone/Makefile       |   21 +++++++++
 eclone/eclone-1.c     |  119 +++++++++++++++++++++++++++++++++++++++++++++++++
 eclone/eclone-tests.h |   66 +++++++++++++++++++++++++++
 4 files changed, 207 insertions(+), 1 deletions(-)
 create mode 100644 eclone/Makefile
 create mode 100644 eclone/eclone-1.c
 create mode 100644 eclone/eclone-tests.h

index 95f8b6e..1d412b9 100644
@@ -1,5 +1,5 @@
 SUBDIRS = libcrtest counterloop fileio simple userns ipc sleep \
-	  process-tree futex epoll taskfs
+	  process-tree futex epoll taskfs eclone
 
 targets = ns_exec mysu
 

new file mode 100644
index 0000000..10b63d3
@@ -0,0 +1,21 @@
+#
+# libeclone.a is built/installed from user-cr git-tree. Point USER_CR_DIR
+# to the directory containing libeclone.a and genstack.h
+#
+USER_CR_DIR ?= ../../user-cr
+
+CFLAGS = -Wall -I $(USER_CR_DIR)
+
+LDFLAGS =
+
+LIB_ECLONE = $(USER_CR_DIR)/libeclone.a
+
+PROGS = eclone-1
+
+all: $(PROGS)
+
+$(PROGS): %: %.c
+	$(CC) $(CFLAGS) -o $@ $< $(LIB_ECLONE) $(LDFLAGS)
+
+clean:
+	rm -f *.o $(PROGS)

1.6.6.1

From: Sukadev Bhattiprolu <[email protected]>
Date: Sat, 30 Jan 2010 12:51:23 -0800
Subject: [PATCH 2/6] eclone-2: Ensure eclone() fails if selected pid is in use

Ensure that eclone() system call fails with -EBUSY if selected pid is in use.

Changelog[v2]:
- Use libeclone.a from user-cr git tree so tests are portable across architectures.
- Fix a few nits identified by Serge Hallyn

---
 eclone/Makefile   |    2 +-
 eclone/eclone-2.c |  116 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 117 insertions(+), 1 deletions(-)
 create mode 100644 eclone/eclone-2.c

index 10b63d3..91aaffc 100644
@@ -10,7 +10,7 @@ LDFLAGS =
 
 LIB_ECLONE = $(USER_CR_DIR)/libeclone.a
 
-PROGS = eclone-1
+PROGS = eclone-1 eclone-2
 
 all: $(PROGS)
 

1.6.6.1

From: Sukadev Bhattiprolu <[email protected]>
Date: Sat, 30 Jan 2010 12:53:04 -0800
Subject: [PATCH 3/6] eclone-3: Verify eclone() fails if reserved fields are not 0.

To ensure future extendability, we want the unused fields to be 0. Ensure eclone() fails if ->reserved0 field is non-0.

Changelog[v2]:
- Use libeclone.a from user-cr git tree so tests are portable across architectures.
- Fix some nits identified by Serge Hallyn

---
 eclone/Makefile   |    2 +-
 eclone/eclone-3.c |  116 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 117 insertions(+), 1 deletions(-)
 create mode 100644 eclone/eclone-3.c

index 91aaffc..ee0471f 100644
@@ -10,7 +10,7 @@ LDFLAGS =
 
 LIB_ECLONE = $(USER_CR_DIR)/libeclone.a
 
-PROGS = eclone-1 eclone-2
+PROGS = eclone-1 eclone-2 eclone-3
 
 all: $(PROGS)
 

1.6.6.1

From: Sukadev Bhattiprolu <[email protected]>
Date: Mon, 1 Feb 2010 18:10:32 -0800
Subject: [PATCH 4/6] eclone-4: Verify eclone() fails if clone_flags_high is non-zero

To ensure future extendability, we want the unused fields to be 0. Ensure eclone() fails if ->clone_flags_high field is non-0.

Changelog[v2]:
- Use libeclone.a from user-cr git tree so tests are portable across architectures.
- Fix some nits identified by Serge Hallyn

---
 eclone/Makefile   |    2 +-
 eclone/eclone-4.c |  115 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+), 1 deletions(-)
 create mode 100644 eclone/eclone-4.c

index ee0471f..ba3d6ac 100644
@@ -10,7 +10,7 @@ LDFLAGS =
 
 LIB_ECLONE = $(USER_CR_DIR)/libeclone.a
 
-PROGS = eclone-1 eclone-2 eclone-3
+PROGS = eclone-1 eclone-2 eclone-3 eclone-4
 
 all: $(PROGS)
 

1.6.6.1

From: Sukadev Bhattiprolu <[email protected]>
Date: Mon, 1 Feb 2010 18:13:51 -0800
Subject: [PATCH 5/6] eclone-5: nr_pids must not exceed nesting level

Verify that eclone() fails if nr_pids exceeds the current nesting level of pid namespaces. Also verify that eclone() succeeds in choosing a pid for a process in a descendant pid namespace.

Changelog[v2]:
- Use libeclone.a from user-cr git tree so tests are portable across architectures.
- Fix some nits identified by Serge Hallyn

---
 eclone/Makefile   |    2 +-
 eclone/eclone-5.c |  175 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 176 insertions(+), 1 deletions(-)
 create mode 100644 eclone/eclone-5.c

index ba3d6ac..7bd5585 100644
@@ -10,7 +10,7 @@ LDFLAGS =
 
 LIB_ECLONE = $(USER_CR_DIR)/libeclone.a
 
-PROGS = eclone-1 eclone-2 eclone-3 eclone-4
+PROGS = eclone-1 eclone-2 eclone-3 eclone-4 eclone-5
 
 all: $(PROGS)
 

1.6.6.1

This is a bit of a shame - it failed for me at first because I was running it inside a container. But there is no way for a task to know it's own nsdepth right?

So I went ahead and pushed these tests (plus a few changes).

thanks,
-serge

Serge E. Hallyn [serue-r/[email protected]] wrote:

| Quoting Sukadev Bhattiprolu ([email protected]):
| > 
| > From: Sukadev Bhattiprolu <[email protected]>
| > Date: Mon, 1 Feb 2010 18:13:51 -0800
| > Subject: [PATCH 5/6] eclone-5: nr_pids must not exceed nesting level
| > 
| > Verify that eclone() fails if nr_pids exceeds the current nesting level
| > of pid namespaces. Also verify that eclone() succeeds in choosing a pid
| > for a process in a descendant pid namespace.
| 
| This is a bit of a shame - it failed for me at first because I was
| running it inside a container.  But there is no way for a task to
| know it's own nsdepth right?

Nope - or at least I don't know of any way.

We could modify the test to set nr_pids to a ridiculously large value just to test that eclone() fails when nr_pids exceeds nesting level.

|
| So I went ahead and pushed these tests (plus a few changes).

Thanks,

Sukadev

Yeah though it's nice to be able to test the boundary conditions.

I suppose it could take an optional nesting level argument - if not specified use 20. Then in documentation recommend that people run

	eclone-5 2

from init_pid_ns. The motivation would be to have the automated testing not return false positives!

From: Sukadev Bhattiprolu <[email protected]>
Date: Tue, 2 Feb 2010 11:21:07 -0800
Subject: [PATCH 6/6] eclone/runtests.sh: Wrapper script for eclone tests

A simple wrapper script to run all eclone tests in this directory.

---
 eclone/runtests.sh |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)
 create mode 100755 eclone/runtests.sh

new file mode 100755
index 0000000..de3f586
@@ -0,0 +1,23 @@
+#!/bin/bash
+
+TEST_LOG=`mktemp -p . runtests-XXXXXXX`.log
+
+echo Logfile: $TEST_LOG
+
+> $TEST_LOG
+
+TESTS="eclone-1 eclone-2 eclone-3 eclone-4 eclone-5"
+#TESTS="eclone-5"
+
+for t in $TESTS
+do
+	echo "===== Test: $t" >> $TEST_LOG
+	./$t >> $TEST_LOG 2>&1
+	if [ $? -eq 0 ]; then
+		echo "Test '$t' PASSED" >> $TEST_LOG
+	else
+		echo "Test '$t' FAILED" >> $TEST_LOG
+	fi
+done
+
+cat $TEST_LOG

1.6.6.1