[WIP] user_mode_thread
Signed-off-by: Andy Lutomirski <luto@kernel.org>
diff --git a/kernel/Makefile b/kernel/Makefile
index eb26e12c..319df88 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -9,7 +9,7 @@
extable.o params.o \
kthread.o sys_ni.o nsproxy.o \
notifier.o ksysfs.o cred.o reboot.o \
- async.o range.o smpboot.o ucount.o
+ async.o range.o smpboot.o ucount.o user_mode_thread.o
obj-$(CONFIG_MULTIUSER) += groups.o
diff --git a/kernel/user_mode_thread.c b/kernel/user_mode_thread.c
new file mode 100644
index 0000000..90023fe
--- /dev/null
+++ b/kernel/user_mode_thread.c
@@ -0,0 +1,254 @@
+/*
+ * Kernel threads that run user code.
+ */
+
+#include <linux/user_mode_thread.h>
+
+/* TODO: fix this crap */
+#include <linux/slab.h>
+#include <linux/init.h>
+#include <linux/unistd.h>
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/completion.h>
+#include <linux/personality.h>
+#include <linux/mempolicy.h>
+#include <linux/sem.h>
+#include <linux/file.h>
+#include <linux/fdtable.h>
+#include <linux/iocontext.h>
+#include <linux/key.h>
+#include <linux/binfmts.h>
+#include <linux/mman.h>
+#include <linux/mmu_notifier.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/vmacache.h>
+#include <linux/nsproxy.h>
+#include <linux/capability.h>
+#include <linux/cpu.h>
+#include <linux/cgroup.h>
+#include <linux/security.h>
+#include <linux/hugetlb.h>
+#include <linux/seccomp.h>
+#include <linux/swap.h>
+#include <linux/syscalls.h>
+#include <linux/jiffies.h>
+#include <linux/futex.h>
+#include <linux/compat.h>
+#include <linux/kthread.h>
+#include <linux/task_io_accounting_ops.h>
+#include <linux/rcupdate.h>
+#include <linux/ptrace.h>
+#include <linux/mount.h>
+#include <linux/audit.h>
+#include <linux/memcontrol.h>
+#include <linux/ftrace.h>
+#include <linux/proc_fs.h>
+#include <linux/profile.h>
+#include <linux/rmap.h>
+#include <linux/ksm.h>
+#include <linux/acct.h>
+#include <linux/tsacct_kern.h>
+#include <linux/cn_proc.h>
+#include <linux/freezer.h>
+#include <linux/delayacct.h>
+#include <linux/taskstats_kern.h>
+#include <linux/random.h>
+#include <linux/tty.h>
+#include <linux/blkdev.h>
+#include <linux/fs_struct.h>
+#include <linux/magic.h>
+#include <linux/perf_event.h>
+#include <linux/posix-timers.h>
+#include <linux/user-return-notifier.h>
+#include <linux/oom.h>
+#include <linux/khugepaged.h>
+#include <linux/signalfd.h>
+#include <linux/uprobes.h>
+#include <linux/aio.h>
+#include <linux/compiler.h>
+#include <linux/sysctl.h>
+#include <linux/kcov.h>
+
+static void userthread_entry(void *arg);
+
+struct userthread_internal_info {
+ int retval; /* 0 means successfully started. */
+ struct completion started;
+ const char *name;
+
+ struct userthread_info *info;
+};
+
+int create_user_mode_thread(struct userthread_info *info,
+ const char *name)
+{
+ pid_t child;
+ unsigned int flags;
+
+ struct userthread_internal_info iinfo = {
+ .info = info,
+ .name = name,
+ .retval = 0,
+ };
+
+ init_completion(&iinfo.started);
+
+ /*
+ * TODO: we shouldn't fork directly. Instead we should
+ * fork off of kthreadd, a workqueue, or some similar
+ * clean context, IMO.
+ *
+ * TODO: We actually do want CLONE_VM becuase we should
+ * probably go through the exec() infrastructure to give
+ * us a real context, even though we don't want to run a
+ * real program and we probably don't want a vdso.
+ */
+ flags = current->flags;
+ current->flags = flags | PF_KTHREAD;
+ child = _do_fork(CLONE_UNTRACED,
+ (unsigned long)userthread_entry,
+ (unsigned long)&iinfo, NULL, NULL, 0);
+ current->flags = flags;
+ if (child < 0)
+ return (int)child;
+
+ wait_for_completion(&iinfo.started);
+ if (iinfo.retval < 0)
+ return iinfo.retval;
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(create_user_mode_thread);
+
+static void userthread_entry(void *arg)
+{
+ struct userthread_internal_info *iinfo = arg;
+ struct userthread_info *info = iinfo->info;
+ struct cred *cred;
+
+ current->flags &= ~PF_KTHREAD;
+
+ /* TODO: set_task_comm? */
+ strncpy(current->comm, iinfo->name, sizeof(current->comm));
+
+ /* Reset signal state */
+ spin_lock_irq(¤t->sighand->siglock);
+ flush_signal_handlers(current, 1);
+ spin_unlock_irq(¤t->sighand->siglock);
+
+ /* Reset priority */
+ set_user_nice(current, 0);
+
+ /* Reset creds */
+ cred = prepare_kernel_cred(current);
+ if (!cred) {
+ iinfo->retval = -ENOMEM;
+ goto fail;
+ }
+ commit_creds(cred);
+
+ iinfo->retval = info->init(info);
+ if (iinfo->retval < 0)
+ goto fail;
+
+ complete(&iinfo->started);
+
+ /*
+ * We expect main to return so it lands in user mode. main is
+ * responsible for all cleanup once it starts executing.
+ */
+ info->main(info);
+ return;
+
+fail:
+ complete(&iinfo->started);
+ do_exit(0);
+
+}
+
+#ifdef CONFIG_DEBUG_FS
+
+#include <linux/debugfs.h>
+
+struct hello_info {
+ struct userthread_info info;
+ char buf[32];
+};
+
+static int hello_init(struct userthread_info *info)
+{
+ return 0;
+}
+
+static void hello_main(struct userthread_info *info)
+{
+ struct hello_info *h =
+ container_of(info, struct hello_info, info);
+
+ struct hello_info mine = *h;
+ kfree(h);
+
+ pr_err("MAIN: %s\n", mine.buf);
+
+ start_thread(current_pt_regs(), 0xbaadc0de, 0xdeadbeef);
+ return; /* segfault! */
+}
+
+static ssize_t hello_write(struct file *file,
+ const char __user *user_buf,
+ size_t count, loff_t *ppos)
+{
+ char buf[32];
+ size_t buf_size;
+ struct hello_info *h;
+ int ret;
+
+ if (count == 0)
+ return 0;
+
+ buf_size = min(count, (sizeof(buf)-1));
+ if (copy_from_user(buf, user_buf, buf_size))
+ return -EFAULT;
+
+ buf[buf_size] = '\0';
+
+ h = kmalloc(sizeof(*h), GFP_KERNEL);
+ if (!h) {
+ return -ENOMEM;
+ }
+
+ h->info.init = hello_init;
+ h->info.main = hello_main;
+ memcpy(h->buf, buf, sizeof(buf));
+
+ ret = create_user_mode_thread(&h->info, "khelloworld");
+ if (ret != 0) {
+ kfree(h);
+ return ret;
+ }
+
+ return count;
+}
+
+static const struct file_operations hello_fops = {
+ .open = simple_open,
+ .write = hello_write,
+ .llseek = default_llseek,
+};
+
+static int init_test(void)
+{
+ struct dentry *dir;
+
+ dir = debugfs_create_dir("user_mode_thread_test", NULL);
+ if (!dir)
+ return -ENOMEM;
+
+ debugfs_create_file("say_hello", 0600, dir, NULL, &hello_fops);
+ return 0;
+}
+device_initcall(init_test);
+
+#endif
