From: Sukadev Bhattiprolu <sukadev@xxxxxxxxxxxxxxxxxx>
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.
NOTE: The myclone() function in eclone-lib.c supports just x86 for now.
Needs to be ported to other architectures.
Signed-off-by: Sukadev Bhattiprolu <sukadev@xxxxxxxxxxxxxxxxxx>
---
Makefile | 2 +-
eclone/Makefile | 16 +++++++
eclone/clone_args.h | 37 ++++++++++++++++
eclone/eclone-1.c | 121 +++++++++++++++++++++++++++++++++++++++++++++++++++
eclone/eclone-lib.c | 85 ++++++++++++++++++++++++++++++++++++
5 files changed, 260 insertions(+), 1 deletions(-)
create mode 100644 eclone/Makefile
create mode 100644 eclone/clone_args.h
create mode 100644 eclone/eclone-1.c
create mode 100644 eclone/eclone-lib.c
diff --git a/Makefile b/Makefile
index 95f8b6e..1d412b9 100644
--- a/Makefile
+++ b/Makefile
@@ -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
diff --git a/eclone/Makefile b/eclone/Makefile
new file mode 100644
index 0000000..43cef65
--- /dev/null
+++ b/eclone/Makefile
@@ -0,0 +1,16 @@
+
+CFLAGS = -Wall
+
+LDFLAGS =
+
+PROGS = eclone-1
+
+all: $(PROGS)
+
+eclone-lib.o: eclone-lib.c clone_args.h
+
+$(PROGS): %: %.c eclone-lib.o
+ $(CC) $(CFLAGS) -o $@ $< eclone-lib.o $(LDFLAGS)
+
+clean:
+ rm -f $(PROGS) eclone-lib.o
diff --git a/eclone/clone_args.h b/eclone/clone_args.h
new file mode 100644
index 0000000..393f139
--- /dev/null
+++ b/eclone/clone_args.h
@@ -0,0 +1,37 @@
+#include <errno.h>
+
+#define __NR_clone_with_pids 337
+#define __NR_clone3 337
+#define __NR_eclone 337
+#define __NR_clone 120
+#define __NR_exit 1
+#define __NR_getpid 20
+
+#define CLONE_NEWPID 0x20000000
+#define CLONE_CHILD_SETTID 0x01000000
+#define CLONE_PARENT_SETTID 0x00100000
+#define CLONE_UNUSED 0x00001000
+
+#define STACKSIZE 8192
+
+typedef unsigned long long u64;
+typedef unsigned int u32;
+typedef int pid_t;
+struct clone_args {
+ u64 clone_flags_high;
+
+ u64 child_stack_base;
+ u64 child_stack_size;
+
+ u64 parent_tid_ptr;
+ u64 child_tid_ptr;
+
+ u32 nr_pids;
+
+ u32 reserved0;
+};
+
+void *setup_stack(int (*child_fn)(void *), void *child_arg, int stack_size);
+int eclone(int flags_low, struct clone_args *clone_args, int args_size,
+ int *pids);
+int gettid();
diff --git a/eclone/eclone-1.c b/eclone/eclone-1.c
new file mode 100644
index 0000000..06a10f9
--- /dev/null
+++ b/eclone/eclone-1.c
@@ -0,0 +1,121 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <errno.h>
+#include <unistd.h>
+#include <signal.h>
+#include <string.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <sys/syscall.h>
+#include "clone_args.h"
+
+int verbose = 0;
+int child_tid, parent_tid;
+
+#define CHILD_TID1 377
+#define CHILD_TID2 399
+#define CHILD_ARG (void *)0x979797
+
+pid_t pids_list[] = { CHILD_TID1, CHILD_TID2 };
+
+int do_child(void *arg)
+{
+ sleep(3);
+
+ if (arg != CHILD_ARG) {
+ printf("ERROR: Expected arg %p, actual %p\n", CHILD_ARG, arg);
+ exit(1);
+ }
+
+ if (gettid() != CHILD_TID1) {
+ printf("FAIL: Child expected pid %d, actual %d\n", CHILD_TID1,
+ getpid());
+ exit(2);
+ } else {
+ printf("PASS: Child got expected pid %d\n", CHILD_TID1);
+ exit(0);
+ }
+
+}
+
+static int myclone(int (*child_fn)(void *), void *child_arg,
+ unsigned int flags_low, int nr_pids, pid_t *pids_list)
+{
+ int rc;
+ void *stack;
+ struct clone_args ca;
+ int args_size;
+
+ stack = setup_stack(child_fn, child_arg, STACKSIZE);
+ if (!stack) {
+ printf("ERROR: setup_stack returns NULL for size %d\n",
+ STACKSIZE);
+ exit(1);
+ }
+
+ memset(&ca, 0, sizeof(ca));
+ ca.child_stack_base = (u64)(int)stack;
+ ca.child_stack_size = (u64)0;
+ ca.parent_tid_ptr = (u64)((int)&parent_tid);
+ ca.child_tid_ptr = (u64)((int)&child_tid);
+ ca.nr_pids = nr_pids;
+
+ if (verbose) {
+ printf("[%d, %d]: Parent:\n\t child_stack 0x%p, ptidp %llx, "
+ "ctidp %llx, pids %p\n", getpid(), gettid(),
+ stack, ca.parent_tid_ptr, ca.child_tid_ptr,
+ pids_list);
+ }
+
+ args_size = sizeof(struct clone_args);
+ rc = eclone(flags_low, &ca, args_size, pids_list);
+
+ if (verbose) {
+ printf("[%d, %d]: eclone() returned %d, error %d\n", getpid(),
+ gettid(), rc, errno);
+ fflush(stdout);
+ }
+
+ if (rc < 0) {
+ printf("ERROR: ");
+ exit(1);
+ }
+
+ return rc;
+}
+
+int main()
+{
+ int rc, pid, status;
+ unsigned long flags;
+ int nr_pids = 1;
+
+ flags = SIGCHLD|CLONE_PARENT_SETTID|CLONE_CHILD_SETTID;
+
+ pid = myclone(do_child, (void *)CHILD_ARG, flags, nr_pids, pids_list);
+
+ if (verbose) {
+ printf("[%d, %d]: Parent waiting for %d\n", getpid(),
+ gettid(), pid);
+ }
+
+ rc = waitpid(pid, &status, __WALL);
+ if (rc < 0) {
+ printf("ERROR: ");
+ verbose = 1;
+ }
+
+ if (verbose) {
+ printf("\twaitpid(): child %d, rc %d, error %d, status 0x%x\n",
+ getpid(), rc, errno, status);
+ if (rc >=0) {
+ if (WIFEXITED(status)) {
+ printf("\t EXITED, %d\n", WEXITSTATUS(status));
+ } else if (WIFSIGNALED(status)) {
+ printf("\t SIGNALED, %d\n", WTERMSIG(status));
+ }
+ }
+ }
+
+ return 0;
+}
diff --git a/eclone/eclone-lib.c b/eclone/eclone-lib.c
new file mode 100644
index 0000000..f547945
--- /dev/null
+++ b/eclone/eclone-lib.c
@@ -0,0 +1,85 @@
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include "clone_args.h"
+
+/* From Oren Laadan */
+
+#if defined(__i386__) && defined(__NR_eclone)
+/*
+ * libc doesn't support eclone() yet...
+ * (see: http://lkml.indiana.edu/hypermail/linux/kernel/9604.3/0204.html)
+ */
+
+int eclone(int flags_low, struct clone_args *clone_args, int args_size,
+ int *pids)
+{
+ long retval;
+
+ __asm__ __volatile__(
+ "movl %3, %%ebx\n\t" /* flags_low -> 1st (ebx) */
+ "movl %4, %%ecx\n\t" /* clone_args -> 2nd (ecx) */
+ "movl %5, %%edx\n\t" /* args_size -> 3rd (edx) */
+ "movl %6, %%esi\n\t" /* pids -> 4th (esi) */
+
+ "pushl %%ebp\n\t" /* save value of ebp */
+ "int $0x80\n\t" /* Linux/i386 system call */
+ "testl %0,%0\n\t" /* check return value */
+ "jne 1f\n\t" /* jump if parent */
+ "popl %%esi\n\t" /* get subthread function */
+ "call *%%esi\n\t" /* start subthread function */
+ "movl %2,%0\n\t"
+ "int $0x80\n" /* exit system call: exit subthread */
+ "1:\n\t"
+ "popl %%ebp\t" /* restore parent's ebp */
+
+ :"=a" (retval)
+
+ :"0" (__NR_clone3),
+ "i" (__NR_exit),
+ "b" (flags_low),
+ "c" (clone_args),
+ "d" (args_size),
+ "D" (pids)
+ );
+
+ if (retval < 0) {
+ errno = -retval;
+ retval = -1;
+ }
+ return retval;
+}
+
+void *setup_stack(int (*child_fn)(void *), void *child_arg, int size)
+{
+ void *stack_base;
+ void **stack_top;
+
+ stack_base = malloc(size);
+ if (!stack_base)
+ return NULL;
+
+ stack_top = (void **)((char *)stack_base + (size - 4));
+ *--stack_top = child_arg;
+ *--stack_top = child_fn;
+
+ return stack_top;
+}
+
+#endif
+
+/* gettid() is a bit more useful than getpid() when messing with clone() */
+int gettid()
+{
+ int rc;
+
+ rc = syscall(__NR_gettid, 0, 0, 0);
+ if (rc < 0) {
+ printf("rc %d, errno %d\n", rc, errno);
+ exit(1);
+ }
+ return rc;
+}
--
1.6.6.1
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/containers
