Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/signal

Pull big execve/kernel_thread/fork unification series from Al Viro:
 "All architectures are converted to new model.  Quite a bit of that
  stuff is actually shared with architecture trees; in such cases it's
  literally shared branch pulled by both, not a cherry-pick.

  A lot of ugliness and black magic is gone (-3KLoC total in this one):

   - kernel_thread()/kernel_execve()/sys_execve() redesign.

     We don't do syscalls from kernel anymore for either kernel_thread()
     or kernel_execve():

     kernel_thread() is essentially clone(2) with callback run before we
     return to userland, the callbacks either never return or do
     successful do_execve() before returning.

     kernel_execve() is a wrapper for do_execve() - it doesn't need to
     do transition to user mode anymore.

     As a result kernel_thread() and kernel_execve() are
     arch-independent now - they live in kernel/fork.c and fs/exec.c
     resp.  sys_execve() is also in fs/exec.c and it's completely
     architecture-independent.

   - daemonize() is gone, along with its parts in fs/*.c

   - struct pt_regs * is no longer passed to do_fork/copy_process/
     copy_thread/do_execve/search_binary_handler/->load_binary/do_coredump.

   - sys_fork()/sys_vfork()/sys_clone() unified; some architectures
     still need wrappers (ones with callee-saved registers not saved in
     pt_regs on syscall entry), but the main part of those suckers is in
     kernel/fork.c now."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/signal: (113 commits)
  do_coredump(): get rid of pt_regs argument
  print_fatal_signal(): get rid of pt_regs argument
  ptrace_signal(): get rid of unused arguments
  get rid of ptrace_signal_deliver() arguments
  new helper: signal_pt_regs()
  unify default ptrace_signal_deliver
  flagday: kill pt_regs argument of do_fork()
  death to idle_regs()
  don't pass regs to copy_process()
  flagday: don't pass regs to copy_thread()
  bfin: switch to generic vfork, get rid of pointless wrappers
  xtensa: switch to generic clone()
  openrisc: switch to use of generic fork and clone
  unicore32: switch to generic clone(2)
  score: switch to generic fork/vfork/clone
  c6x: sanitize copy_thread(), get rid of clone(2) wrapper, switch to generic clone()
  take sys_fork/sys_vfork/sys_clone prototypes to linux/syscalls.h
  mn10300: switch to generic fork/vfork/clone
  h8300: switch to generic fork/vfork/clone
  tile: switch to generic clone()
  ...

Conflicts:
	arch/microblaze/include/asm/Kbuild

8 files changed, 94 insertions, 199 deletions

diff --git a/arch/openrisc/Kconfig b/arch/openrisc/Kconfig
index 05f2ba4..e7f1a29 100644
--- a/arch/openrisc/Kconfig
+++ b/arch/openrisc/Kconfig
@@ -22,6 +22,8 @@ config OPENRISC
 	select GENERIC_STRNCPY_FROM_USER
 	select GENERIC_STRNLEN_USER
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config MMU
 	def_bool y
diff --git a/arch/openrisc/include/asm/processor.h b/arch/openrisc/include/asm/processor.h
index 43decdb..3369138 100644
--- a/arch/openrisc/include/asm/processor.h
+++ b/arch/openrisc/include/asm/processor.h
@@ -81,8 +81,6 @@ struct thread_struct {
 #define KSTK_ESP(tsk)   (task_pt_regs(tsk)->sp)
 
 
-extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
 void start_thread(struct pt_regs *regs, unsigned long nip, unsigned long sp);
 void release_thread(struct task_struct *);
 unsigned long get_wchan(struct task_struct *p);
diff --git a/arch/openrisc/include/asm/syscalls.h b/arch/openrisc/include/asm/syscalls.h
index 84a978a..8ee8168 100644
--- a/arch/openrisc/include/asm/syscalls.h
+++ b/arch/openrisc/include/asm/syscalls.h
@@ -24,4 +24,11 @@ asmlinkage long sys_or1k_atomic(unsigned long type, unsigned long *v1,
 
 #include <asm-generic/syscalls.h>
 
+asmlinkage long __sys_clone(unsigned long clone_flags, unsigned long newsp,
+			void __user *parent_tid, void __user *child_tid, int tls);
+asmlinkage long __sys_fork(void);
+
+#define sys_clone __sys_clone
+#define sys_fork __sys_fork
+
 #endif /* __ASM_OPENRISC_SYSCALLS_H */
diff --git a/arch/openrisc/include/uapi/asm/unistd.h b/arch/openrisc/include/uapi/asm/unistd.h
index 437bdbb..5082b80 100644
--- a/arch/openrisc/include/uapi/asm/unistd.h
+++ b/arch/openrisc/include/uapi/asm/unistd.h
@@ -20,6 +20,10 @@
 
 #define sys_mmap2 sys_mmap_pgoff
 
+#define __ARCH_WANT_SYS_EXECVE
+#define __ARCH_WANT_SYS_FORK
+#define __ARCH_WANT_SYS_CLONE
+
 #include <asm-generic/unistd.h>
 
 #define __NR_or1k_atomic __NR_arch_specific_syscall
diff --git a/arch/openrisc/kernel/Makefile b/arch/openrisc/kernel/Makefile
index e1ee0fa..35f92ce 100644
--- a/arch/openrisc/kernel/Makefile
+++ b/arch/openrisc/kernel/Makefile
@@ -5,7 +5,7 @@
 extra-y	:= head.o vmlinux.lds
 
 obj-y	:= setup.o idle.o or32_ksyms.o process.o dma.o \
-	   traps.o time.o irq.o entry.o ptrace.o signal.o sys_or32.o \
+	   traps.o time.o irq.o entry.o ptrace.o signal.o \
 	   sys_call_table.o
 
 obj-$(CONFIG_MODULES)		+= module.o
diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index ddfcaa8..5e5b306 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
@@ -894,6 +894,16 @@ ENTRY(ret_from_fork)
 	l.jal	schedule_tail
 	 l.nop
 
+	/* Check if we are a kernel thread */
+	l.sfeqi	r20,0
+	l.bf	1f
+	 l.nop
+
+	/* ...we are a kernel thread so invoke the requested callback */
+	l.jalr	r20
+	 l.or	r3,r22,r0
+
+1:
 	/* _syscall_returns expect r11 to contain return value */
 	l.lwz	r11,PT_GPR11(r1)
 
@@ -915,26 +925,6 @@ ENTRY(ret_from_fork)
 	l.j	_syscall_return
 	 l.nop
 
-/* Since syscalls don't save call-clobbered registers, the args to
- * kernel_thread_helper will need to be passed through callee-saved
- * registers and copied to the parameter registers when the thread
- * begins running.
- *
- * See arch/openrisc/kernel/process.c:
- * The args are passed as follows:
- *   arg1 (r3) : passed in r20
- *   arg2 (r4) : passed in r22
- */
-
-ENTRY(_kernel_thread_helper)
-	l.or	r3,r20,r0
-	l.or	r4,r22,r0
-	l.movhi	r31,hi(kernel_thread_helper)
-	l.ori	r31,r31,lo(kernel_thread_helper)
-	l.jr	r31
-	 l.nop
-
-
 /* ========================================================[ switch ] === */
 
 /*
@@ -1044,8 +1034,13 @@ ENTRY(_switch)
 	/* Unwind stack to pre-switch state */
 	l.addi  r1,r1,(INT_FRAME_SIZE)
 
-	/* Return via the link-register back to where we 'came from', where that can be
-	 * either schedule() or return_from_fork()... */
+	/* Return via the link-register back to where we 'came from', where
+	 * that may be either schedule(), ret_from_fork(), or
+	 * ret_from_kernel_thread().  If we are returning to a new thread,
+	 * we are expected to have set up the arg to schedule_tail already,
+	 * hence we do so here unconditionally:
+	 */
+	l.lwz   r3,TI_STACK(r3)		/* Load 'prev' as schedule_tail arg */
 	l.jr	r9
 	 l.nop
 
@@ -1076,22 +1071,18 @@ _fork_save_extra_regs_and_call:
 	l.jr	r29
 	 l.sw    PT_GPR28(r1),r28
 
-ENTRY(sys_clone)
-	l.movhi	r29,hi(_sys_clone)
-	l.ori	r29,r29,lo(_sys_clone)
+ENTRY(__sys_clone)
+	l.movhi	r29,hi(sys_clone)
+	l.ori	r29,r29,lo(sys_clone)
 	l.j	_fork_save_extra_regs_and_call
 	 l.addi	r7,r1,0
 
-ENTRY(sys_fork)
-	l.movhi	r29,hi(_sys_fork)
-	l.ori	r29,r29,lo(_sys_fork)
+ENTRY(__sys_fork)
+	l.movhi	r29,hi(sys_fork)
+	l.ori	r29,r29,lo(sys_fork)
 	l.j	_fork_save_extra_regs_and_call
 	 l.addi	r3,r1,0
 
-ENTRY(sys_execve)
-	l.j	_sys_execve
-	 l.addi	r6,r1,0
-
 ENTRY(sys_sigaltstack)
 	l.j	_sys_sigaltstack
 	 l.addi	r5,r1,0
diff --git a/arch/openrisc/kernel/process.c b/arch/openrisc/kernel/process.c
index c35f3ab..00c233b 100644
--- a/arch/openrisc/kernel/process.c
+++ b/arch/openrisc/kernel/process.c
@@ -109,66 +109,83 @@ void release_thread(struct task_struct *dead_task)
  */
 extern asmlinkage void ret_from_fork(void);
 
+/*
+ * copy_thread
+ * @clone_flags: flags
+ * @usp: user stack pointer or fn for kernel thread
+ * @arg: arg to fn for kernel thread; always NULL for userspace thread
+ * @p: the newly created task
+ * @regs: CPU context to copy for userspace thread; always NULL for kthread
+ *
+ * At the top of a newly initialized kernel stack are two stacked pt_reg
+ * structures.  The first (topmost) is the userspace context of the thread.
+ * The second is the kernelspace context of the thread.
+ *
+ * A kernel thread will not be returning to userspace, so the topmost pt_regs
+ * struct can be uninitialized; it _does_ need to exist, though, because
+ * a kernel thread can become a userspace thread by doing a kernel_execve, in
+ * which case the topmost context will be initialized and used for 'returning'
+ * to userspace.
+ *
+ * The second pt_reg struct needs to be initialized to 'return' to
+ * ret_from_fork.  A kernel thread will need to set r20 to the address of
+ * a function to call into (with arg in r22); userspace threads need to set
+ * r20 to NULL in which case ret_from_fork will just continue a return to
+ * userspace.
+ *
+ * A kernel thread 'fn' may return; this is effectively what happens when
+ * kernel_execve is called.  In that case, the userspace pt_regs must have
+ * been initialized (which kernel_execve takes care of, see start_thread
+ * below); ret_from_fork will then continue its execution causing the
+ * 'kernel thread' to return to userspace as a userspace thread.
+ */
+
 int
 copy_thread(unsigned long clone_flags, unsigned long usp,
-	    unsigned long unused, struct task_struct *p, struct pt_regs *regs)
+	    unsigned long arg, struct task_struct *p)
 {
-	struct pt_regs *childregs;
+	struct pt_regs *userregs;
 	struct pt_regs *kregs;
 	unsigned long sp = (unsigned long)task_stack_page(p) + THREAD_SIZE;
-	struct thread_info *ti;
 	unsigned long top_of_kernel_stack;
 
 	top_of_kernel_stack = sp;
 
 	p->set_child_tid = p->clear_child_tid = NULL;
 
-	/* Copy registers */
-	/* redzone */
-	sp -= STACK_FRAME_OVERHEAD;
+	/* Locate userspace context on stack... */
+	sp -= STACK_FRAME_OVERHEAD;	/* redzone */
 	sp -= sizeof(struct pt_regs);
-	childregs = (struct pt_regs *)sp;
+	userregs = (struct pt_regs *) sp;
 
-	/* Copy parent registers */
-	*childregs = *regs;
+	/* ...and kernel context */
+	sp -= STACK_FRAME_OVERHEAD;	/* redzone */
+	sp -= sizeof(struct pt_regs);
+	kregs = (struct pt_regs *)sp;
 
-	if ((childregs->sr & SPR_SR_SM) == 1) {
-		/* for kernel thread, set `current_thread_info'
-		 * and stackptr in new task
-		 */
-		childregs->sp = (unsigned long)task_stack_page(p) + THREAD_SIZE;
-		childregs->gpr[10] = (unsigned long)task_thread_info(p);
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(kregs, 0, sizeof(struct pt_regs));
+		kregs->gpr[20] = usp; /* fn, kernel thread */
+		kregs->gpr[22] = arg;
 	} else {
-		childregs->sp = usp;
-	}
-
-	childregs->gpr[11] = 0;	/* Result from fork() */
+		*userregs = *current_pt_regs();
 
-	/*
-	 * The way this works is that at some point in the future
-	 * some task will call _switch to switch to the new task.
-	 * That will pop off the stack frame created below and start
-	 * the new task running at ret_from_fork.  The new task will
-	 * do some house keeping and then return from the fork or clone
-	 * system call, using the stack frame created above.
-	 */
-	/* redzone */
-	sp -= STACK_FRAME_OVERHEAD;
-	sp -= sizeof(struct pt_regs);
-	kregs = (struct pt_regs *)sp;
+		if (usp)
+			userregs->sp = usp;
+		userregs->gpr[11] = 0;	/* Result from fork() */
 
-	ti = task_thread_info(p);
-	ti->ksp = sp;
+		kregs->gpr[20] = 0;	/* Userspace thread */
+	}
 
-	/* kregs->sp must store the location of the 'pre-switch' kernel stack
-	 * pointer... for a newly forked process, this is simply the top of
-	 * the kernel stack.
+	/*
+	 * _switch wants the kernel stack page in pt_regs->sp so that it
+	 * can restore it to thread_info->ksp... see _switch for details.
 	 */
 	kregs->sp = top_of_kernel_stack;
-	kregs->gpr[3] = (unsigned long)current;	/* arg to schedule_tail */
-	kregs->gpr[10] = (unsigned long)task_thread_info(p);
 	kregs->gpr[9] = (unsigned long)ret_from_fork;
 
+	task_thread_info(p)->ksp = (unsigned long)kregs;
+
 	return 0;
 }
 
@@ -177,16 +194,14 @@ copy_thread(unsigned long clone_flags, unsigned long usp,
  */
 void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long sp)
 {
-	unsigned long sr = regs->sr & ~SPR_SR_SM;
+	unsigned long sr = mfspr(SPR_SR) & ~SPR_SR_SM;
 
 	set_fs(USER_DS);
-	memset(regs->gpr, 0, sizeof(regs->gpr));
+	memset(regs, 0, sizeof(struct pt_regs));
 
 	regs->pc = pc;
 	regs->sr = sr;
 	regs->sp = sp;
-
-/*	printk("start thread, ksp = %lx\n", current_thread_info()->ksp);*/
 }
 
 /* Fill in the fpu structure for a core dump.  */
@@ -237,74 +252,9 @@ void dump_elf_thread(elf_greg_t *dest, struct pt_regs* regs)
 	dest[35] = 0;
 }
 
-extern void _kernel_thread_helper(void);
-
-void __noreturn kernel_thread_helper(int (*fn) (void *), void *arg)
-{
-	do_exit(fn(arg));
-}
-
-/*
- * Create a kernel thread.
- */
-int kernel_thread(int (*fn) (void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-	regs.gpr[20] = (unsigned long)fn;
-	regs.gpr[22] = (unsigned long)arg;
-	regs.sr = mfspr(SPR_SR);
-	regs.pc = (unsigned long)_kernel_thread_helper;
-
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED,
-		       0, &regs, 0, NULL, NULL);
-}
-
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage long _sys_execve(const char __user *name,
-			    const char __user * const __user *argv,
-			    const char __user * const __user *envp,
-			    struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(name);
-	error = PTR_ERR(filename);
-
-	if (IS_ERR(filename))
-		goto out;
-
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-
-out:
-	return error;
-}
-
 unsigned long get_wchan(struct task_struct *p)
 {
 	/* TODO */
 
 	return 0;
 }
-
-int kernel_execve(const char *filename, char *const argv[], char *const envp[])
-{
-	register long __res asm("r11") = __NR_execve;
-	register long __a asm("r3") = (long)(filename);
-	register long __b asm("r4") = (long)(argv);
-	register long __c asm("r5") = (long)(envp);
-	__asm__ volatile ("l.sys 1"
-			  : "=r" (__res), "=r"(__a), "=r"(__b), "=r"(__c)
-			  : "0"(__res), "1"(__a), "2"(__b), "3"(__c)
-			  : "r6", "r7", "r8", "r12", "r13", "r15",
-			    "r17", "r19", "r21", "r23", "r25", "r27",
-			    "r29", "r31");
-	__asm__ volatile ("l.nop");
-	return __res;
-}
diff --git a/arch/openrisc/kernel/sys_or32.c b/arch/openrisc/kernel/sys_or32.c
deleted file mode 100644
index 5706008..0000000
--- a/arch/openrisc/kernel/sys_or32.c
+++ /dev/null
@@ -1,57 +0,0 @@
-/*
- * OpenRISC sys_or32.c
- *
- * Linux architectural port borrowing liberally from similar works of
- * others.  All original copyrights apply as per the original source
- * declaration.
- *
- * Modifications for the OpenRISC architecture:
- * Copyright (C) 2003 Matjaz Breskvar <phoenix@bsemi.com>
- * Copyright (C) 2010-2011 Jonas Bonn <jonas@southpole.se>
- *
- *      This program is free software; you can redistribute it and/or
- *      modify it under the terms of the GNU General Public License
- *      as published by the Free Software Foundation; either version
- *      2 of the License, or (at your option) any later version.
- *
- * This file contains various random system calls that
- * have a non-standard calling sequence on some platforms.
- * Since we don't have to do any backwards compatibility, our
- * versions are done in the most "normal" way possible.
- */
-
-#include <linux/errno.h>
-#include <linux/syscalls.h>
-#include <linux/mm.h>
-
-#include <asm/syscalls.h>
-
-/* These are secondary entry points as the primary entry points are defined in
- * entry.S where we add the 'regs' parameter value
- */
-
-asmlinkage long _sys_clone(unsigned long clone_flags, unsigned long newsp,
-			   int __user *parent_tid, int __user *child_tid,
-			   struct pt_regs *regs)
-{
-	long ret;
-
-	/* FIXME: Is alignment necessary? */
-	/* newsp = ALIGN(newsp, 4); */
-
-	if (!newsp)
-		newsp = regs->sp;
-
-	ret = do_fork(clone_flags, newsp, regs, 0, parent_tid, child_tid);
-
-	return ret;
-}
-
-asmlinkage int _sys_fork(struct pt_regs *regs)
-{
-#ifdef CONFIG_MMU
-	return do_fork(SIGCHLD, regs->sp, regs, 0, NULL, NULL);
-#else
-	return -EINVAL;
-#endif
-}