Merge branches 'no-rebases', 'arch-avr32', 'arch-blackfin', 'arch-cris', 'arch-h8300', 'arch-m32r', 'arch-mn10300', 'arch-score', 'arch-sh' and 'arch-powerpc' into for-next

arch/arm64/Kconfig

@@ -7,6 +7,8 @@ config ARM64
 	select GENERIC_IOMAP
 	select GENERIC_IRQ_PROBE
 	select GENERIC_IRQ_SHOW
+	select GENERIC_KERNEL_EXECVE
+	select GENERIC_KERNEL_THREAD
 	select GENERIC_SMP_IDLE_THREAD
 	select GENERIC_TIME_VSYSCALL
 	select HARDIRQS_SW_RESEND

arch/arm64/include/asm/processor.h

@@ -128,11 +128,6 @@ unsigned long get_wchan(struct task_struct *p);
 extern struct task_struct *cpu_switch_to(struct task_struct *prev,
 					 struct task_struct *next);
 
-/*
- * Create a new kernel thread
- */
-extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
-
 #define task_pt_regs(p) \
 	((struct pt_regs *)(THREAD_START_SP + task_stack_page(p)) - 1)
 

arch/arm64/include/asm/syscalls.h

@@ -23,18 +23,16 @@
 /*
  * System call wrappers implemented in kernel/entry.S.
  */
-asmlinkage long sys_execve_wrapper(const char __user *filename,
-				   const char __user *const __user *argv,
-				   const char __user *const __user *envp);
-asmlinkage long sys_clone_wrapper(unsigned long clone_flags,
-				  unsigned long newsp,
-				  void __user *parent_tid,
-				  unsigned long tls_val,
-				  void __user *child_tid);
 asmlinkage long sys_rt_sigreturn_wrapper(void);
 asmlinkage long sys_sigaltstack_wrapper(const stack_t __user *uss,
 					stack_t __user *uoss);
 
+/*
+ * AArch64 sys_clone implementation has a different prototype than the generic
+ * one (additional TLS value argument).
+ */
+#define sys_clone	sys_clone
+
 #include <asm-generic/syscalls.h>
 
 #endif	/* __ASM_SYSCALLS_H */

arch/arm64/include/asm/unistd.h

@@ -25,4 +25,5 @@
 #define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_COMPAT_SYS_SENDFILE
 #endif
+#define __ARCH_WANT_SYS_EXECVE
 #include <uapi/asm/unistd.h>

arch/arm64/include/asm/unistd32.h

@@ -23,7 +23,7 @@
 
 __SYSCALL(0,   sys_restart_syscall)
 __SYSCALL(1,   sys_exit)
-__SYSCALL(2,   compat_sys_fork_wrapper)
+__SYSCALL(2,   compat_sys_fork)
 __SYSCALL(3,   sys_read)
 __SYSCALL(4,   sys_write)
 __SYSCALL(5,   compat_sys_open)
@@ -32,7 +32,7 @@ __SYSCALL(7,   sys_ni_syscall)			/* 7 was sys_waitpid */
 __SYSCALL(8,   sys_creat)
 __SYSCALL(9,   sys_link)
 __SYSCALL(10,  sys_unlink)
-__SYSCALL(11,  compat_sys_execve_wrapper)
+__SYSCALL(11,  compat_sys_execve)
 __SYSCALL(12,  sys_chdir)
 __SYSCALL(13,  sys_ni_syscall)			/* 13 was sys_time */
 __SYSCALL(14,  sys_mknod)
@@ -141,7 +141,7 @@ __SYSCALL(116, compat_sys_sysinfo)
 __SYSCALL(117, sys_ni_syscall)			/* 117 was sys_ipc */
 __SYSCALL(118, sys_fsync)
 __SYSCALL(119, compat_sys_sigreturn_wrapper)
-__SYSCALL(120, compat_sys_clone_wrapper)
+__SYSCALL(120, sys_clone)
 __SYSCALL(121, sys_setdomainname)
 __SYSCALL(122, sys_newuname)
 __SYSCALL(123, sys_ni_syscall)			/* 123 was sys_modify_ldt */
@@ -211,7 +211,7 @@ __SYSCALL(186, compat_sys_sigaltstack_wrapper)
 __SYSCALL(187, compat_sys_sendfile)
 __SYSCALL(188, sys_ni_syscall)			/* 188 reserved */
 __SYSCALL(189, sys_ni_syscall)			/* 189 reserved */
-__SYSCALL(190, compat_sys_vfork_wrapper)
+__SYSCALL(190, compat_sys_vfork)
 __SYSCALL(191, compat_sys_getrlimit)		/* SuS compliant getrlimit */
 __SYSCALL(192, sys_mmap_pgoff)
 __SYSCALL(193, compat_sys_truncate64_wrapper)

arch/arm64/kernel/entry.S

@@ -594,7 +594,7 @@ work_resched:
 /*
  * "slow" syscall return path.
  */
-ENTRY(ret_to_user)
+ret_to_user:
 	disable_irq				// disable interrupts
 	ldr	x1, [tsk, #TI_FLAGS]
 	and	x2, x1, #_TIF_WORK_MASK
@@ -611,7 +611,10 @@ ENDPROC(ret_to_user)
  */
 ENTRY(ret_from_fork)
 	bl	schedule_tail
-	get_thread_info tsk
+	cbz	x19, 1f				// not a kernel thread
+	mov	x0, x20
+	blr	x19
+1:	get_thread_info tsk
 	b	ret_to_user
 ENDPROC(ret_from_fork)
 
@@ -673,16 +676,6 @@ __sys_trace_return:
 /*
  * Special system call wrappers.
  */
-ENTRY(sys_execve_wrapper)
-	mov	x3, sp
-	b	sys_execve
-ENDPROC(sys_execve_wrapper)
-
-ENTRY(sys_clone_wrapper)
-	mov	x5, sp
-	b	sys_clone
-ENDPROC(sys_clone_wrapper)
-
 ENTRY(sys_rt_sigreturn_wrapper)
 	mov	x0, sp
 	b	sys_rt_sigreturn

arch/arm64/kernel/process.c

@@ -240,27 +240,41 @@ int copy_thread(unsigned long clone_flags, unsigned long stack_start,
 	struct pt_regs *childregs = task_pt_regs(p);
 	unsigned long tls = p->thread.tp_value;
 
-	*childregs = *regs;
-	childregs->regs[0] = 0;
+	memset(&p->thread.cpu_context, 0, sizeof(struct cpu_context));
 
-	if (is_compat_thread(task_thread_info(p)))
-		childregs->compat_sp = stack_start;
-	else {
+	if (likely(regs)) {
+		*childregs = *regs;
+		childregs->regs[0] = 0;
+		if (is_compat_thread(task_thread_info(p))) {
+			if (stack_start)
+				childregs->compat_sp = stack_start;
+		} else {
+			/*
+			 * Read the current TLS pointer from tpidr_el0 as it may be
+			 * out-of-sync with the saved value.
+			 */
+			asm("mrs %0, tpidr_el0" : "=r" (tls));
+			if (stack_start) {
+				/* 16-byte aligned stack mandatory on AArch64 */
+				if (stack_start & 15)
+					return -EINVAL;
+				childregs->sp = stack_start;
+			}
+		}
 		/*
-		 * Read the current TLS pointer from tpidr_el0 as it may be
-		 * out-of-sync with the saved value.
+		 * If a TLS pointer was passed to clone (4th argument), use it
+		 * for the new thread.
 		 */
-		asm("mrs %0, tpidr_el0" : "=r" (tls));
-		childregs->sp = stack_start;
+		if (clone_flags & CLONE_SETTLS)
+			tls = regs->regs[3];
+	} else {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		childregs->pstate = PSR_MODE_EL1h;
+		p->thread.cpu_context.x19 = stack_start;
+		p->thread.cpu_context.x20 = stk_sz;
 	}
-
-	memset(&p->thread.cpu_context, 0, sizeof(struct cpu_context));
-	p->thread.cpu_context.sp = (unsigned long)childregs;
 	p->thread.cpu_context.pc = (unsigned long)ret_from_fork;
-
-	/* If a TLS pointer was passed to clone, use that for the new thread. */
-	if (clone_flags & CLONE_SETTLS)
-		tls = regs->regs[3];
+	p->thread.cpu_context.sp = (unsigned long)childregs;
 	p->thread.tp_value = tls;
 
 	ptrace_hw_copy_thread(p);
@@ -309,43 +323,6 @@ struct task_struct *__switch_to(struct task_struct *prev,
 	return last;
 }
 
-/*
- * Shuffle the argument into the correct register before calling the
- * thread function.  x1 is the thread argument, x2 is the pointer to
- * the thread function, and x3 points to the exit function.
- */
-extern void kernel_thread_helper(void);
-asm(	".section .text\n"
-"	.align\n"
-"	.type	kernel_thread_helper, #function\n"
-"kernel_thread_helper:\n"
-"	mov	x0, x1\n"
-"	mov	x30, x3\n"
-"	br	x2\n"
-"	.size	kernel_thread_helper, . - kernel_thread_helper\n"
-"	.previous");
-
-#define kernel_thread_exit	do_exit
-
-/*
- * Create a kernel thread.
- */
-pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-	regs.regs[1] = (unsigned long)arg;
-	regs.regs[2] = (unsigned long)fn;
-	regs.regs[3] = (unsigned long)kernel_thread_exit;
-	regs.pc = (unsigned long)kernel_thread_helper;
-	regs.pstate = PSR_MODE_EL1h;
-
-	return do_fork(flags|CLONE_VM|CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 unsigned long get_wchan(struct task_struct *p)
 {
 	struct stackframe frame;

arch/arm64/kernel/sys.c

@@ -31,79 +31,11 @@
  */
 asmlinkage long sys_clone(unsigned long clone_flags, unsigned long newsp,
 			  int __user *parent_tidptr, unsigned long tls_val,
-			  int __user *child_tidptr, struct pt_regs *regs)
+			  int __user *child_tidptr)
 {
-	if (!newsp)
-		newsp = regs->sp;
-	/* 16-byte aligned stack mandatory on AArch64 */
-	if (newsp & 15)
-		return -EINVAL;
-	return do_fork(clone_flags, newsp, regs, 0, parent_tidptr, child_tidptr);
-}
-
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage long sys_execve(const char __user *filenamei,
-			   const char __user *const __user *argv,
-			   const char __user *const __user *envp,
-			   struct pt_regs *regs)
-{
-	long error;
-	struct filename *filename;
-
-	filename = getname(filenamei);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-out:
-	return error;
-}
-
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	struct pt_regs regs;
-	int ret;
-
-	memset(&regs, 0, sizeof(struct pt_regs));
-	ret = do_execve(filename,
-			(const char __user *const __user *)argv,
-			(const char __user *const __user *)envp, &regs);
-	if (ret < 0)
-		goto out;
-
-	/*
-	 * Save argc to the register structure for userspace.
-	 */
-	regs.regs[0] = ret;
-
-	/*
-	 * We were successful.  We won't be returning to our caller, but
-	 * instead to user space by manipulating the kernel stack.
-	 */
-	asm(	"add	x0, %0, %1\n\t"
-		"mov	x1, %2\n\t"
-		"mov	x2, %3\n\t"
-		"bl	memmove\n\t"	/* copy regs to top of stack */
-		"mov	x27, #0\n\t"	/* not a syscall */
-		"mov	x28, %0\n\t"	/* thread structure */
-		"mov	sp, x0\n\t"	/* reposition stack pointer */
-		"b	ret_to_user"
-		:
-		: "r" (current_thread_info()),
-		  "Ir" (THREAD_START_SP - sizeof(regs)),
-		  "r" (&regs),
-		  "Ir" (sizeof(regs))
-		: "x0", "x1", "x2", "x27", "x28", "x30", "memory");
-
- out:
-	return ret;
+	return do_fork(clone_flags, newsp, current_pt_regs(), 0,
+			parent_tidptr, child_tidptr);
 }
-EXPORT_SYMBOL(kernel_execve);
 
 asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
 			 unsigned long prot, unsigned long flags,
@@ -118,8 +50,6 @@ asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
 /*
  * Wrappers to pass the pt_regs argument.
  */
-#define sys_execve		sys_execve_wrapper
-#define sys_clone		sys_clone_wrapper
 #define sys_rt_sigreturn	sys_rt_sigreturn_wrapper
 #define sys_sigaltstack		sys_sigaltstack_wrapper
 

arch/arm64/kernel/sys32.S

@@ -26,25 +26,6 @@
 /*
  * System call wrappers for the AArch32 compatibility layer.
  */
-compat_sys_fork_wrapper:
-	mov	x0, sp
-	b	compat_sys_fork
-ENDPROC(compat_sys_fork_wrapper)
-
-compat_sys_vfork_wrapper:
-	mov	x0, sp
-	b	compat_sys_vfork
-ENDPROC(compat_sys_vfork_wrapper)
-
-compat_sys_execve_wrapper:
-	mov	x3, sp
-	b	compat_sys_execve
-ENDPROC(compat_sys_execve_wrapper)
-
-compat_sys_clone_wrapper:
-	mov	x5, sp
-	b	compat_sys_clone
-ENDPROC(compat_sys_clone_wrapper)
 
 compat_sys_sigreturn_wrapper:
 	mov	x0, sp

arch/arm64/kernel/sys_compat.c

@@ -28,43 +28,15 @@
 #include <asm/cacheflush.h>
 #include <asm/unistd32.h>
 
-asmlinkage int compat_sys_fork(struct pt_regs *regs)
+asmlinkage int compat_sys_fork(void)
 {
-	return do_fork(SIGCHLD, regs->compat_sp, regs, 0, NULL, NULL);
+	return do_fork(SIGCHLD, 0, current_pt_regs(), 0, NULL, NULL);
 }
 
-asmlinkage int compat_sys_clone(unsigned long clone_flags, unsigned long newsp,
-			  int __user *parent_tidptr, int tls_val,
-			  int __user *child_tidptr, struct pt_regs *regs)
+asmlinkage int compat_sys_vfork(void)
 {
-	if (!newsp)
-		newsp = regs->compat_sp;
-
-	return do_fork(clone_flags, newsp, regs, 0, parent_tidptr, child_tidptr);
-}
-
-asmlinkage int compat_sys_vfork(struct pt_regs *regs)
-{
-	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs->compat_sp,
-		       regs, 0, NULL, NULL);
-}
-
-asmlinkage int compat_sys_execve(const char __user *filenamei,
-				 compat_uptr_t argv, compat_uptr_t envp,
-				 struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(filenamei);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = compat_do_execve(filename->name, compat_ptr(argv),
-					compat_ptr(envp), regs);
-	putname(filename);
-out:
-	return error;
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, 0,
+		       current_pt_regs(), 0, NULL, NULL);
 }
 
 asmlinkage int compat_sys_sched_rr_get_interval(compat_pid_t pid,

arch/avr32/Kconfig

@@ -17,6 +17,8 @@ config AVR32
 	select GENERIC_CLOCKEVENTS
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	help
 	  AVR32 is a high-performance 32-bit RISC microprocessor core,
 	  designed for cost-sensitive embedded applications, with particular

arch/avr32/include/asm/processor.h

@@ -142,9 +142,6 @@ struct task_struct;
 /* Free all resources held by a thread */
 extern void release_thread(struct task_struct *);
 
-/* Create a kernel thread without removing it from tasklists */
-extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
-
 /* Return saved PC of a blocked thread */
 #define thread_saved_pc(tsk)    ((tsk)->thread.cpu_context.pc)
 

arch/avr32/include/asm/unistd.h

@@ -39,6 +39,7 @@
 #define __ARCH_WANT_SYS_GETPGRP
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 /*
  * "Conditional" syscalls

arch/avr32/kernel/Makefile

@@ -7,7 +7,7 @@ extra-y				:= head.o vmlinux.lds
 obj-$(CONFIG_SUBARCH_AVR32B)	+= entry-avr32b.o
 obj-y				+= syscall_table.o syscall-stubs.o irq.o
 obj-y				+= setup.o traps.o ocd.o ptrace.o
-obj-y				+= signal.o sys_avr32.o process.o time.o
+obj-y				+= signal.o process.o time.o
 obj-y				+= switch_to.o cpu.o
 obj-$(CONFIG_MODULES)		+= module.o avr32_ksyms.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o

arch/avr32/kernel/entry-avr32b.S

@@ -251,13 +251,15 @@ syscall_badsys:
 	.global ret_from_fork
 ret_from_fork:
 	call   schedule_tail
+	mov	r12, 0
+	rjmp	syscall_return
 
-	/* check for syscall tracing */
-	get_thread_info r0
-	ld.w	r1, r0[TI_flags]
-	andl	r1, _TIF_ALLWORK_MASK, COH
-	brne	syscall_exit_work
-	rjmp    syscall_exit_cont
+	.global ret_from_kernel_thread
+ret_from_kernel_thread:
+	call   schedule_tail
+	mov	r12, r0
+	mov	lr, r2	/* syscall_return */
+	mov	pc, r1
 
 syscall_trace_enter:
 	pushm	r8-r12

arch/avr32/kernel/process.c

@@ -68,44 +68,6 @@ void machine_restart(char *cmd)
 	while (1) ;
 }
 
-/*
- * PC is actually discarded when returning from a system call -- the
- * return address must be stored in LR. This function will make sure
- * LR points to do_exit before starting the thread.
- *
- * Also, when returning from fork(), r12 is 0, so we must copy the
- * argument as well.
- *
- *  r0 : The argument to the main thread function
- *  r1 : The address of do_exit
- *  r2 : The address of the main thread function
- */
-asmlinkage extern void kernel_thread_helper(void);
-__asm__("	.type	kernel_thread_helper, @function\n"
-	"kernel_thread_helper:\n"
-	"	mov	r12, r0\n"
-	"	mov	lr, r2\n"
-	"	mov	pc, r1\n"
-	"	.size	kernel_thread_helper, . - kernel_thread_helper");
-
-int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-	regs.r0 = (unsigned long)arg;
-	regs.r1 = (unsigned long)fn;
-	regs.r2 = (unsigned long)do_exit;
-	regs.lr = (unsigned long)kernel_thread_helper;
-	regs.pc = (unsigned long)kernel_thread_helper;
-	regs.sr = MODE_SUPERVISOR;
-
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED,
-		       0, &regs, 0, NULL, NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 /*
  * Free current thread data structures etc
  */
@@ -332,26 +294,31 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 }
 
 asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
+asmlinkage void syscall_return(void);
 
 int copy_thread(unsigned long clone_flags, unsigned long usp,
-		unsigned long unused,
+		unsigned long arg,
 		struct task_struct *p, struct pt_regs *regs)
 {
-	struct pt_regs *childregs;
-
-	childregs = ((struct pt_regs *)(THREAD_SIZE + (unsigned long)task_stack_page(p))) - 1;
-	*childregs = *regs;
-
-	if (user_mode(regs))
+	struct pt_regs *childregs = task_pt_regs(p);
+
+	if (unlikely(!regs)) {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		p->thread.cpu_context.r0 = arg;
+		p->thread.cpu_context.r1 = usp; /* fn */
+		p->thread.cpu_context.r2 = syscall_return;
+		p->thread.cpu_context.pc = (unsigned long)ret_from_kernel_thread;
+		childregs->sr = MODE_SUPERVISOR;
+	} else {
+		*childregs = *regs;
 		childregs->sp = usp;
-	else
-		childregs->sp = (unsigned long)task_stack_page(p) + THREAD_SIZE;
-
-	childregs->r12 = 0; /* Set return value for child */
+		childregs->r12 = 0; /* Set return value for child */
+		p->thread.cpu_context.pc = (unsigned long)ret_from_fork;
+	}
 
 	p->thread.cpu_context.sr = MODE_SUPERVISOR | SR_GM;
 	p->thread.cpu_context.ksp = (unsigned long)childregs;
-	p->thread.cpu_context.pc = (unsigned long)ret_from_fork;
 
 	clear_tsk_thread_flag(p, TIF_DEBUG);
 	if ((clone_flags & CLONE_PTRACE) && test_thread_flag(TIF_DEBUG))
@@ -382,27 +349,6 @@ asmlinkage int sys_vfork(struct pt_regs *regs)
 		       0, NULL, NULL);
 }
 
-asmlinkage int sys_execve(const char __user *ufilename,
-			  const char __user *const __user *uargv,
-			  const char __user *const __user *uenvp,
-			  struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(ufilename);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-
-	error = do_execve(filename->name, uargv, uenvp, regs);
-	putname(filename);
-
-out:
-	return error;
-}
-
-
 /*
  * This function is supposed to answer the question "who called
  * schedule()?"

arch/avr32/kernel/sys_avr32.c

@@ -1,24 +0,0 @@
-/*
- * Copyright (C) 2004-2006 Atmel Corporation
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-#include <linux/unistd.h>
-
-int kernel_execve(const char *file,
-		  const char *const *argv,
-		  const char *const *envp)
-{
-	register long scno asm("r8") = __NR_execve;
-	register long sc1 asm("r12") = (long)file;
-	register long sc2 asm("r11") = (long)argv;
-	register long sc3 asm("r10") = (long)envp;
-
-	asm volatile("scall"
-		     : "=r"(sc1)
-		     : "r"(scno), "0"(sc1), "r"(sc2), "r"(sc3)
-		     : "cc", "memory");
-	return sc1;
-}

arch/avr32/kernel/syscall-stubs.S

@@ -50,12 +50,6 @@ __sys_vfork:
 	mov	r12, sp
 	rjmp	sys_vfork
 
-	.global	__sys_execve
-	.type	__sys_execve,@function
-__sys_execve:
-	mov	r9, sp
-	rjmp	sys_execve
-
 	.global	__sys_mmap2
 	.type	__sys_mmap2,@function
 __sys_mmap2:

arch/avr32/kernel/syscall_table.S

@@ -24,7 +24,7 @@ sys_call_table:
 	.long	sys_creat
 	.long	sys_link
 	.long	sys_unlink		/* 10 */
-	.long	__sys_execve
+	.long	sys_execve
 	.long	sys_chdir
 	.long	sys_time
 	.long	sys_mknod

arch/blackfin/Kconfig

@@ -45,6 +45,8 @@ config BLACKFIN
 	select ARCH_USES_GETTIMEOFFSET if !GENERIC_CLOCKEVENTS
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config GENERIC_CSUM
 	def_bool y

arch/blackfin/include/asm/processor.h

@@ -75,8 +75,6 @@ static inline void release_thread(struct task_struct *dead_task)
 {
 }
 
-extern int kernel_thread(int (*fn) (void *), void *arg, unsigned long flags);
-
 /*
  * Free current thread data structures etc..
  */

arch/blackfin/include/asm/unistd.h

@@ -446,6 +446,7 @@
 #define __ARCH_WANT_SYS_NICE
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 /*
  * "Conditional" syscalls

arch/blackfin/kernel/entry.S

@@ -46,22 +46,16 @@ ENTRY(_ret_from_fork)
 	SP += -12;
 	pseudo_long_call _schedule_tail, p5;
 	SP += 12;
-	r0 = [sp + PT_IPEND];
-	cc = bittst(r0,1);
-	if cc jump .Lin_kernel;
+	p1 = [sp++];
+	r0 = [sp++];
+	cc = p1 == 0;
+	if cc jump .Lfork;
+	sp += -12;
+	call (p1);
+	sp += 12;
+.Lfork:
 	RESTORE_CONTEXT
 	rti;
-.Lin_kernel:
-	bitclr(r0,1);
-	[sp + PT_IPEND] = r0;
-	/* do a 'fake' RTI by jumping to [RETI]
-	 * to avoid clearing supervisor mode in child
-	 */
-	r0 = [sp + PT_PC];
-	[sp + PT_P0] = r0;
-
-	RESTORE_ALL_SYS
-	jump (p0);
 ENDPROC(_ret_from_fork)
 
 ENTRY(_sys_vfork)

arch/blackfin/kernel/process.c

@@ -101,40 +101,6 @@ void cpu_idle(void)
 	}
 }
 
-/*
- * This gets run with P1 containing the
- * function to call, and R1 containing
- * the "args".  Note P0 is clobbered on the way here.
- */
-void kernel_thread_helper(void);
-__asm__(".section .text\n"
-	".align 4\n"
-	"_kernel_thread_helper:\n\t"
-	"\tsp += -12;\n\t"
-	"\tr0 = r1;\n\t" "\tcall (p1);\n\t" "\tcall _do_exit;\n" ".previous");
-
-/*
- * Create a kernel thread.
- */
-pid_t kernel_thread(int (*fn) (void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-	regs.r1 = (unsigned long)arg;
-	regs.p1 = (unsigned long)fn;
-	regs.pc = (unsigned long)kernel_thread_helper;
-	regs.orig_p0 = -1;
-	/* Set bit 2 to tell ret_from_fork we should be returning to kernel
-	   mode.  */
-	regs.ipend = 0x8002;
-	__asm__ __volatile__("%0 = syscfg;":"=da"(regs.syscfg):);
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL,
-		       NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 /*
  * Do necessary setup to start up a newly executed thread.
  *
@@ -193,38 +159,31 @@ copy_thread(unsigned long clone_flags,
 	    struct task_struct *p, struct pt_regs *regs)
 {
 	struct pt_regs *childregs;
+	unsigned long *v;
 
 	childregs = (struct pt_regs *) (task_stack_page(p) + THREAD_SIZE) - 1;
-	*childregs = *regs;
-	childregs->r0 = 0;
+	v = ((unsigned long *)childregs) - 2;
+	if (unlikely(!regs)) {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		v[0] = usp;
+		v[1] = topstk;
+		childregs->orig_p0 = -1;
+		childregs->ipend = 0x8000;
+		__asm__ __volatile__("%0 = syscfg;":"=da"(childregs->syscfg):);
+		p->thread.usp = 0;
+	} else {
+		*childregs = *regs;
+		childregs->r0 = 0;
+		p->thread.usp = usp;
+		v[0] = v[1] = 0;
+	}
 
-	p->thread.usp = usp;
-	p->thread.ksp = (unsigned long)childregs;
+	p->thread.ksp = (unsigned long)v;
 	p->thread.pc = (unsigned long)ret_from_fork;
 
 	return 0;
 }
 
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys_execve(const char __user *name,
-			  const char __user *const __user *argv,
-			  const char __user *const __user *envp)
-{
-	int error;
-	struct filename *filename;
-	struct pt_regs *regs = (struct pt_regs *)((&name) + 6);
-
-	filename = getname(name);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		return error;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-	return error;
-}
-
 unsigned long get_wchan(struct task_struct *p)
 {
 	unsigned long fp, pc;

arch/blackfin/mach-common/entry.S

@@ -530,61 +530,6 @@ ENTRY(_trap) /* Exception: 4th entry into system event table(supervisor mode)*/
 	jump .Lsyscall_really_exit;
 ENDPROC(_trap)
 
-ENTRY(_kernel_execve)
-	link SIZEOF_PTREGS;
-	p0 = sp;
-	r3 = SIZEOF_PTREGS / 4;
-	r4 = 0(x);
-.Lclear_regs:
-	[p0++] = r4;
-	r3 += -1;
-	cc = r3 == 0;
-	if !cc jump .Lclear_regs (bp);
-
-	p0 = sp;
-	sp += -16;
-	[sp + 12] = p0;
-	pseudo_long_call _do_execve, p5;
-	SP += 16;
-	cc = r0 == 0;
-	if ! cc jump .Lexecve_failed;
-	/* Success.  Copy our temporary pt_regs to the top of the kernel
-	 * stack and do a normal exception return.
-	 */
-	r1 = sp;
-	r0 = (-KERNEL_STACK_SIZE) (x);
-	r1 = r1 & r0;
-	p2 = r1;
-	p3 = [p2];
-	r0 = KERNEL_STACK_SIZE - 4 (z);
-	p1 = r0;
-	p1 = p1 + p2;
-
-	p0 = fp;
-	r4 = [p0--];
-	r3 = SIZEOF_PTREGS / 4;
-.Lcopy_regs:
-	r4 = [p0--];
-	[p1--] = r4;
-	r3 += -1;
-	cc = r3 == 0;
-	if ! cc jump .Lcopy_regs (bp);
-
-	r0 = (KERNEL_STACK_SIZE - SIZEOF_PTREGS) (z);
-	p1 = r0;
-	p1 = p1 + p2;
-	sp = p1;
-	r0 = syscfg;
-	[SP + PT_SYSCFG] = r0;
-	[p3 + (TASK_THREAD + THREAD_KSP)] = sp;
-
-	RESTORE_CONTEXT;
-	rti;
-.Lexecve_failed:
-	unlink;
-	rts;
-ENDPROC(_kernel_execve)
-
 ENTRY(_system_call)
 	/* Store IPEND */
 	p2.l = lo(IPEND);

arch/c6x/Kconfig

@@ -18,6 +18,7 @@ config C6X
 	select OF_EARLY_FLATTREE
 	select GENERIC_CLOCKEVENTS
 	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	select MODULES_USE_ELF_RELA
 
 config MMU

arch/c6x/include/uapi/asm/unistd.h

@@ -14,7 +14,6 @@
  *   more details.
  */
 
-#define __ARCH_WANT_KERNEL_EXECVE
 #define __ARCH_WANT_SYS_EXECVE
 
 /* Use the standard ABI for syscalls. */

arch/c6x/kernel/entry.S

@@ -413,19 +413,9 @@ ENTRY(ret_from_kernel_thread)
 0:
 	B	.S2	B10		   /* call fn */
 	LDW	.D2T1	*+SP(REGS_A1+8),A4 /* get arg */
-	MVKL	.S2	sys_exit,B11
-	MVKH	.S2	sys_exit,B11
-	ADDKPC	.S2	0f,B3,1
-0:
-	BNOP	.S2	B11,5	/* jump to sys_exit */
+	ADDKPC	.S2	ret_from_fork_2,B3,3
 ENDPROC(ret_from_kernel_thread)
 
-ENTRY(ret_from_kernel_execve)
-	GET_THREAD_INFO A12
-	BNOP	.S2	syscall_exit,4
-	ADD	.D2X	A4,-8,SP
-ENDPROC(ret_from_kernel_execve)
-
 	;;
 	;; These are the interrupt handlers, responsible for calling __do_IRQ()
 	;; int6 is used for syscalls (see _system_call entry)

arch/cris/Kconfig

@@ -49,6 +49,8 @@ config CRIS
 	select GENERIC_SMP_IDLE_THREAD if ETRAX_ARCH_V32
 	select GENERIC_CMOS_UPDATE
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config HZ
 	int

arch/cris/arch-v10/kernel/entry.S

@@ -35,6 +35,7 @@
 	.globl system_call
 	.globl ret_from_intr
 	.globl ret_from_fork
+	.globl ret_from_kernel_thread
 	.globl resume
 	.globl multiple_interrupt
 	.globl hwbreakpoint
@@ -81,7 +82,14 @@ ret_from_fork:
 	jsr schedule_tail
 	ba  ret_from_sys_call
 	nop
-		
+
+ret_from_kernel_thread:
+	jsr schedule_tail
+	move.d	$r2, $r10	; argument is here
+	jsr	$r1		; call the payload
+	moveq	0, $r9		; no syscall restarts, TYVM...
+	ba  ret_from_sys_call
+
 ret_from_intr:
 	;; check for resched if preemptive kernel or if we're going back to user-mode 
 	;; this test matches the user_regs(regs) macro
@@ -586,13 +594,6 @@ _ugdb_handle_breakpoint:
 	ba	do_sigtrap		; SIGTRAP the offending process. 
 	pop	$dccr			; Restore dccr in delay slot.
 
-	.global kernel_execve
-kernel_execve:
-	move.d __NR_execve, $r9
-	break 13
-	ret
-	nop
-
 	.data
 
 hw_bp_trigs:

arch/cris/arch-v10/kernel/process.c

@@ -17,6 +17,7 @@
 #include <arch/svinto.h>
 #include <linux/init.h>
 #include <arch/system.h>
+#include <asm/ptrace.h>
 
 #ifdef CONFIG_ETRAX_GPIO
 void etrax_gpio_wake_up_check(void); /* drivers/gpio.c */
@@ -81,31 +82,6 @@ unsigned long thread_saved_pc(struct task_struct *t)
 	return task_pt_regs(t)->irp;
 }
 
-static void kernel_thread_helper(void* dummy, int (*fn)(void *), void * arg)
-{
-  fn(arg);
-  do_exit(-1); /* Should never be called, return bad exit value */
-}
-
-/*
- * Create a kernel thread
- */
-int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-        /* Don't use r10 since that is set to 0 in copy_thread */
-	regs.r11 = (unsigned long)fn;
-	regs.r12 = (unsigned long)arg;
-	regs.irp = (unsigned long)kernel_thread_helper;
-	regs.dccr = 1 << I_DCCR_BITNR;
-
-	/* Ok, create the new process.. */
-        return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
-}
-
 /* setup the child's kernel stack with a pt_regs and switch_stack on it.
  * it will be un-nested during _resume and _ret_from_sys_call when the
  * new thread is scheduled.
@@ -115,29 +91,35 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
  *
  */
 asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
 
 int copy_thread(unsigned long clone_flags, unsigned long usp,
-		unsigned long unused,
+		unsigned long arg,
 		struct task_struct *p, struct pt_regs *regs)
 {
-	struct pt_regs * childregs;
-	struct switch_stack *swstack;
+	struct pt_regs *childregs = task_pt_regs(p);
+	struct switch_stack *swstack = ((struct switch_stack *)childregs) - 1;
 	
 	/* put the pt_regs structure at the end of the new kernel stack page and fix it up
 	 * remember that the task_struct doubles as the kernel stack for the task
 	 */
 
-	childregs = task_pt_regs(p);
-        
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(swstack, 0,
+			sizeof(struct switch_stack) + sizeof(struct pt_regs));
+		swstack->r1 = usp;
+		swstack->r2 = arg;
+		childregs->dccr = 1 << I_DCCR_BITNR;
+		swstack->return_ip = (unsigned long) ret_from_kernel_thread;
+		p->thread.ksp = (unsigned long) swstack;
+		p->thread.usp = 0;
+		return 0;
+	}
 	*childregs = *regs;  /* struct copy of pt_regs */
-        
-        p->set_child_tid = p->clear_child_tid = NULL;
 
         childregs->r10 = 0;  /* child returns 0 after a fork/clone */
-	
-	/* put the switch stack right below the pt_regs */
 
-	swstack = ((struct switch_stack *)childregs) - 1;
+	/* put the switch stack right below the pt_regs */
 
 	swstack->r9 = 0; /* parameter to ret_from_sys_call, 0 == dont restart the syscall */
 
@@ -147,7 +129,7 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	
 	/* fix the user-mode stackpointer */
 
-	p->thread.usp = usp;	
+	p->thread.usp = usp;
 
 	/* and the kernel-mode one */
 
@@ -161,68 +143,28 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	return 0;
 }
 
-/* 
- * Be aware of the "magic" 7th argument in the four system-calls below.
- * They need the latest stackframe, which is put as the 7th argument by
- * entry.S. The previous arguments are dummies or actually used, but need
- * to be defined to reach the 7th argument.
- *
- * N.B.: Another method to get the stackframe is to use current_regs(). But
- * it returns the latest stack-frame stacked when going from _user mode_ and
- * some of these (at least sys_clone) are called from kernel-mode sometimes
- * (for example during kernel_thread, above) and thus cannot use it. Thus,
- * to be sure not to get any surprises, we use the method for the other calls
- * as well.
- */
-
-asmlinkage int sys_fork(long r10, long r11, long r12, long r13, long mof, long srp,
-			struct pt_regs *regs)
+asmlinkage int sys_fork(void)
 {
-	return do_fork(SIGCHLD, rdusp(), regs, 0, NULL, NULL);
+	return do_fork(SIGCHLD, rdusp(), current_pt_regs(), 0, NULL, NULL);
 }
 
 /* if newusp is 0, we just grab the old usp */
 /* FIXME: Is parent_tid/child_tid really third/fourth argument? Update lib? */
 asmlinkage int sys_clone(unsigned long newusp, unsigned long flags,
-			 int* parent_tid, int* child_tid, long mof, long srp,
-			 struct pt_regs *regs)
+			 int* parent_tid, int* child_tid)
 {
 	if (!newusp)
 		newusp = rdusp();
-	return do_fork(flags, newusp, regs, 0, parent_tid, child_tid);
+	return do_fork(flags, newusp, current_pt_regs(), 0, parent_tid, child_tid);
 }
 
 /* vfork is a system call in i386 because of register-pressure - maybe
  * we can remove it and handle it in libc but we put it here until then.
  */
 
-asmlinkage int sys_vfork(long r10, long r11, long r12, long r13, long mof, long srp,
-			 struct pt_regs *regs)
-{
-        return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, rdusp(), regs, 0, NULL, NULL);
-}
-
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys_execve(const char *fname,
-			  const char *const *argv,
-			  const char *const *envp,
-			  long r13, long mof, long srp, 
-			  struct pt_regs *regs)
+asmlinkage int sys_vfork(void)
 {
-	int error;
-	struct filename *filename;
-
-	filename = getname(fname);
-	error = PTR_ERR(filename);
-
-	if (IS_ERR(filename))
-	        goto out;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
- out:
-	return error;
+        return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, rdusp(), current_pt_regs(), 0, NULL, NULL);
 }
 
 unsigned long get_wchan(struct task_struct *p)

arch/cris/arch-v32/kernel/entry.S

@@ -31,6 +31,7 @@
 	.globl system_call
 	.globl ret_from_intr
 	.globl ret_from_fork
+	.globl ret_from_kernel_thread
 	.globl resume
 	.globl multiple_interrupt
 	.globl nmi_interrupt
@@ -84,6 +85,18 @@ ret_from_fork:
 	nop
 	.size	ret_from_fork, . - ret_from_fork
 
+	.type	ret_from_kernel_thread,@function
+ret_from_kernel_thread:
+	jsr schedule_tail
+	nop
+	move.d	$r2, $r10
+	jsr	$r1
+	nop
+	moveq	0, $r9			; no syscall restarts, TYVM...
+	ba  ret_from_sys_call
+	nop
+	.size	ret_from_kernel_thread, . - ret_from_kernel_thread
+
 	.type	ret_from_intr,@function
 ret_from_intr:
 	;; Check for resched if preemptive kernel, or if we're going back to
@@ -531,15 +544,6 @@ _ugdb_handle_exception:
 	ba	do_sigtrap		; SIGTRAP the offending process.
 	move.d	[$sp+], $r0		; Restore R0 in delay slot.
 
-	.global kernel_execve
-	.type	kernel_execve,@function
-kernel_execve:
-	move.d __NR_execve, $r9
-	break 13
-	ret
-	nop
-	.size	kernel_execve, . - kernel_execve
-
 	.data
 
 	.section .rodata,"a"

arch/cris/arch-v32/kernel/process.c

@@ -16,6 +16,7 @@
 #include <hwregs/reg_map.h>
 #include <hwregs/timer_defs.h>
 #include <hwregs/intr_vect_defs.h>
+#include <asm/ptrace.h>
 
 extern void stop_watchdog(void);
 
@@ -94,31 +95,6 @@ unsigned long thread_saved_pc(struct task_struct *t)
 	return task_pt_regs(t)->erp;
 }
 
-static void
-kernel_thread_helper(void* dummy, int (*fn)(void *), void * arg)
-{
-	fn(arg);
-	do_exit(-1); /* Should never be called, return bad exit value. */
-}
-
-/* Create a kernel thread. */
-int
-kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-        /* Don't use r10 since that is set to 0 in copy_thread. */
-	regs.r11 = (unsigned long) fn;
-	regs.r12 = (unsigned long) arg;
-	regs.erp = (unsigned long) kernel_thread_helper;
-	regs.ccs = 1 << (I_CCS_BITNR + CCS_SHIFT);
-
-	/* Create the new process. */
-        return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
-}
-
 /*
  * Setup the child's kernel stack with a pt_regs and call switch_stack() on it.
  * It will be unnested during _resume and _ret_from_sys_call when the new thread
@@ -129,23 +105,33 @@ kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
  */
 
 extern asmlinkage void ret_from_fork(void);
+extern asmlinkage void ret_from_kernel_thread(void);
 
 int
 copy_thread(unsigned long clone_flags, unsigned long usp,
-	unsigned long unused,
+	unsigned long arg,
 	struct task_struct *p, struct pt_regs *regs)
 {
-	struct pt_regs *childregs;
-	struct switch_stack *swstack;
+	struct pt_regs *childregs = task_pt_regs(p);
+	struct switch_stack *swstack = ((struct switch_stack *) childregs) - 1;
 
 	/*
 	 * Put the pt_regs structure at the end of the new kernel stack page and
 	 * fix it up. Note: the task_struct doubles as the kernel stack for the
 	 * task.
 	 */
-	childregs = task_pt_regs(p);
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(swstack, 0,
+			sizeof(struct switch_stack) + sizeof(struct pt_regs));
+		swstack->r1 = usp;
+		swstack->r2 = arg;
+		childregs->ccs = 1 << (I_CCS_BITNR + CCS_SHIFT);
+		swstack->return_ip = (unsigned long) ret_from_kernel_thread;
+		p->thread.ksp = (unsigned long) swstack;
+		p->thread.usp = 0;
+		return 0;
+	}
 	*childregs = *regs;	/* Struct copy of pt_regs. */
-        p->set_child_tid = p->clear_child_tid = NULL;
         childregs->r10 = 0;	/* Child returns 0 after a fork/clone. */
 
 	/* Set a new TLS ?
@@ -156,7 +142,6 @@ copy_thread(unsigned long clone_flags, unsigned long usp,
 	}
 
 	/* Put the switch stack right below the pt_regs. */
-	swstack = ((struct switch_stack *) childregs) - 1;
 
 	/* Parameter to ret_from_sys_call. 0 is don't restart the syscall. */
 	swstack->r9 = 0;
@@ -174,35 +159,21 @@ copy_thread(unsigned long clone_flags, unsigned long usp,
 	return 0;
 }
 
-/*
- * Be aware of the "magic" 7th argument in the four system-calls below.
- * They need the latest stackframe, which is put as the 7th argument by
- * entry.S. The previous arguments are dummies or actually used, but need
- * to be defined to reach the 7th argument.
- *
- * N.B.: Another method to get the stackframe is to use current_regs(). But
- * it returns the latest stack-frame stacked when going from _user mode_ and
- * some of these (at least sys_clone) are called from kernel-mode sometimes
- * (for example during kernel_thread, above) and thus cannot use it. Thus,
- * to be sure not to get any surprises, we use the method for the other calls
- * as well.
- */
 asmlinkage int
-sys_fork(long r10, long r11, long r12, long r13, long mof, long srp,
-	struct pt_regs *regs)
+sys_fork(void)
 {
-	return do_fork(SIGCHLD, rdusp(), regs, 0, NULL, NULL);
+	return do_fork(SIGCHLD, rdusp(), current_pt_regs(), 0, NULL, NULL);
 }
 
 /* FIXME: Is parent_tid/child_tid really third/fourth argument? Update lib? */
 asmlinkage int
 sys_clone(unsigned long newusp, unsigned long flags, int *parent_tid, int *child_tid,
-	unsigned long tls, long srp, struct pt_regs *regs)
+	unsigned long tls)
 {
 	if (!newusp)
 		newusp = rdusp();
 
-	return do_fork(flags, newusp, regs, 0, parent_tid, child_tid);
+	return do_fork(flags, newusp, current_pt_regs(), 0, parent_tid, child_tid);
 }
 
 /*
@@ -210,32 +181,9 @@ sys_clone(unsigned long newusp, unsigned long flags, int *parent_tid, int *child
  * we can remove it and handle it in libc but we put it here until then.
  */
 asmlinkage int
-sys_vfork(long r10, long r11, long r12, long r13, long mof, long srp,
-	struct pt_regs *regs)
-{
-	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, rdusp(), regs, 0, NULL, NULL);
-}
-
-/* sys_execve() executes a new program. */
-asmlinkage int
-sys_execve(const char *fname,
-	   const char *const *argv,
-	   const char *const *envp, long r13, long mof, long srp,
-	   struct pt_regs *regs)
+sys_vfork(void)
 {
-	int error;
-	struct filename *filename;
-
-	filename = getname(fname);
-	error = PTR_ERR(filename);
-
-	if (IS_ERR(filename))
-	        goto out;
-
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
- out:
-	return error;
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, rdusp(), current_pt_regs(), 0, NULL, NULL);
 }
 
 unsigned long

arch/cris/include/asm/processor.h

@@ -49,8 +49,6 @@ struct task_struct;
 #define task_pt_regs(task) user_regs(task_thread_info(task))
 #define current_regs() task_pt_regs(current)
 
-extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
 unsigned long get_wchan(struct task_struct *p);
 
 #define KSTK_ESP(tsk)   ((tsk) == current ? rdusp() : (tsk)->thread.usp)

arch/cris/include/asm/unistd.h

@@ -371,6 +371,7 @@
 #define __ARCH_WANT_SYS_SIGPROCMASK
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 /*
  * "Conditional" syscalls

arch/cris/kernel/crisksyms.c

@@ -30,7 +30,6 @@ extern void __negdi2(void);
 extern void iounmap(volatile void * __iomem);
 
 /* Platform dependent support */
-EXPORT_SYMBOL(kernel_thread);
 EXPORT_SYMBOL(get_cmos_time);
 EXPORT_SYMBOL(loops_per_usec);
 

arch/h8300/Kconfig

@@ -8,6 +8,8 @@ config H8300
 	select GENERIC_IRQ_SHOW
 	select GENERIC_CPU_DEVICES
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config SYMBOL_PREFIX
 	string

arch/h8300/include/asm/processor.h

@@ -107,8 +107,6 @@ static inline void release_thread(struct task_struct *dead_task)
 {
 }
 
-extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
 /*
  * Free current thread data structures etc..
  */

arch/h8300/include/asm/ptrace.h

@@ -60,6 +60,8 @@ struct pt_regs {
 #define user_mode(regs) (!((regs)->ccr & PS_S))
 #define instruction_pointer(regs) ((regs)->pc)
 #define profile_pc(regs) instruction_pointer(regs)
+#define current_pt_regs() ((struct pt_regs *) \
+	(THREAD_SIZE + (unsigned long)current_thread_info()) - 1)
 #endif /* __KERNEL__ */
 #endif /* __ASSEMBLY__ */
 #endif /* _H8300_PTRACE_H */

arch/h8300/include/asm/unistd.h

@@ -356,6 +356,7 @@
 #define __ARCH_WANT_SYS_SIGPROCMASK
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 /*
  * "Conditional" syscalls

arch/h8300/kernel/entry.S

@@ -158,6 +158,7 @@ INTERRUPTS = 128
 .globl SYMBOL_NAME(system_call)
 .globl SYMBOL_NAME(ret_from_exception)
 .globl SYMBOL_NAME(ret_from_fork)
+.globl SYMBOL_NAME(ret_from_kernel_thread)
 .globl SYMBOL_NAME(ret_from_interrupt)
 .globl SYMBOL_NAME(interrupt_redirect_table)
 .globl SYMBOL_NAME(sw_ksp),SYMBOL_NAME(sw_usp)
@@ -330,6 +331,14 @@ SYMBOL_NAME_LABEL(ret_from_fork)
 	jsr	@SYMBOL_NAME(schedule_tail)
 	jmp	@SYMBOL_NAME(ret_from_exception)
 
+SYMBOL_NAME_LABEL(ret_from_kernel_thread)
+	mov.l	er2,er0
+	jsr	@SYMBOL_NAME(schedule_tail)
+	mov.l	@(LER4:16,sp),er0
+	mov.l	@(LER5:16,sp),er1
+	jsr	@er1
+	jmp	@SYMBOL_NAME(ret_from_exception)
+
 SYMBOL_NAME_LABEL(resume)
 	/*
 	 * Beware - when entering resume, offset of tss is in d1,

arch/h8300/kernel/h8300_ksyms.c

@@ -33,7 +33,6 @@ EXPORT_SYMBOL(strncmp);
 
 EXPORT_SYMBOL(ip_fast_csum);
 
-EXPORT_SYMBOL(kernel_thread);
 EXPORT_SYMBOL(enable_irq);
 EXPORT_SYMBOL(disable_irq);
 

arch/h8300/kernel/process.c

@@ -47,6 +47,7 @@ void (*pm_power_off)(void) = NULL;
 EXPORT_SYMBOL(pm_power_off);
 
 asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
 
 /*
  * The idle loop on an H8/300..
@@ -122,39 +123,6 @@ void show_regs(struct pt_regs * regs)
 		printk("\n");
 }
 
-/*
- * Create a kernel thread
- */
-int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
-{
-	long retval;
-	long clone_arg;
-	mm_segment_t fs;
-
-	fs = get_fs();
-	set_fs (KERNEL_DS);
-	clone_arg = flags | CLONE_VM;
-	__asm__("mov.l sp,er3\n\t"
-		"sub.l er2,er2\n\t"
-		"mov.l %2,er1\n\t"
-		"mov.l %1,er0\n\t"
-		"trapa #0\n\t"
-		"cmp.l sp,er3\n\t"
-		"beq 1f\n\t"
-		"mov.l %4,er0\n\t"
-		"mov.l %3,er1\n\t"
-		"jsr @er1\n\t"
-		"mov.l %5,er0\n\t"
-		"trapa #0\n"
-		"1:\n\t"
-		"mov.l er0,%0"
-		:"=r"(retval)
-		:"i"(__NR_clone),"g"(clone_arg),"g"(fn),"g"(arg),"i"(__NR_exit)
-		:"er0","er1","er2","er3");
-	set_fs (fs);
-	return retval;
-}
-
 void flush_thread(void)
 {
 }
@@ -198,6 +166,13 @@ int copy_thread(unsigned long clone_flags,
 
 	childregs = (struct pt_regs *) (THREAD_SIZE + task_stack_page(p)) - 1;
 
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		childregs->retpc = (unsigned long) ret_from_kernel_thread;
+		childregs->er4 = topstk; /* arg */
+		childregs->er5 = usp; /* fn */
+		p->thread.ksp = (unsigned long)childregs;
+	}
 	*childregs = *regs;
 	childregs->retpc = (unsigned long) ret_from_fork;
 	childregs->er0 = 0;
@@ -208,27 +183,6 @@ int copy_thread(unsigned long clone_flags,
 	return 0;
 }
 
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys_execve(const char *name,
-			  const char *const *argv,
-			  const char *const *envp,
-			  int dummy, ...)
-{
-	int error;
-	struct filename *filename;
-	struct pt_regs *regs = (struct pt_regs *) ((unsigned char *)&dummy-4);
-
-	filename = getname(name);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		return error;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-	return error;
-}
-
 unsigned long thread_saved_pc(struct task_struct *tsk)
 {
 	return ((struct pt_regs *)tsk->thread.esp0)->pc;

arch/h8300/kernel/sys_h8300.c

@@ -46,29 +46,3 @@ asmlinkage void syscall_print(void *dummy,...)
                ((regs->pc)&0xffffff)-2,regs->orig_er0,regs->er1,regs->er2,regs->er3,regs->er0);
 }
 #endif
-
-/*
- * Do a system call from kernel instead of calling sys_execve so we
- * end up with proper pt_regs.
- */
-asmlinkage
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	register long res __asm__("er0");
-	register const char *const *_c __asm__("er3") = envp;
-	register const char *const *_b __asm__("er2") = argv;
-	register const char * _a __asm__("er1") = filename;
-	__asm__ __volatile__ ("mov.l %1,er0\n\t"
-			"trapa	#0\n\t"
-			: "=r" (res)
-			: "g" (__NR_execve),
-			  "g" (_a),
-			  "g" (_b),
-			  "g" (_c)
-			: "cc", "memory");
-	return res;
-}
-
-

arch/hexagon/Kconfig

@@ -31,6 +31,8 @@ config HEXAGON
 	select GENERIC_CLOCKEVENTS
 	select GENERIC_CLOCKEVENTS_BROADCAST
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	---help---
 	  Qualcomm Hexagon is a processor architecture designed for high
 	  performance and low power across a wide variety of applications.

arch/hexagon/include/asm/processor.h

@@ -34,7 +34,6 @@
 struct task_struct;
 
 /*  this is defined in arch/process.c  */
-extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
 extern unsigned long thread_saved_pc(struct task_struct *tsk);
 
 extern void start_thread(struct pt_regs *, unsigned long, unsigned long);

arch/hexagon/include/uapi/asm/ptrace.h

@@ -32,4 +32,8 @@
 extern int regs_query_register_offset(const char *name);
 extern const char *regs_query_register_name(unsigned int offset);
 
+#define current_pt_regs() \
+	((struct pt_regs *) \
+	 ((unsigned long)current_thread_info() + THREAD_SIZE) - 1)
+
 #endif

arch/hexagon/include/uapi/asm/unistd.h

@@ -27,5 +27,6 @@
  */
 
 #define sys_mmap2 sys_mmap_pgoff
+#define __ARCH_WANT_SYS_EXECVE
 
 #include <asm-generic/unistd.h>

arch/hexagon/kernel/process.c

@@ -25,33 +25,6 @@
 #include <linux/uaccess.h>
 #include <linux/slab.h>
 
-/*
- * Kernel thread creation.  The desired kernel function is "wrapped"
- * in the kernel_thread_helper function, which does cleanup
- * afterwards.
- */
-static void __noreturn kernel_thread_helper(void *arg, int (*fn)(void *))
-{
-	do_exit(fn(arg));
-}
-
-int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-	/*
-	 * Yes, we're exploting illicit knowledge of the ABI here.
-	 */
-	regs.r00 = (unsigned long) arg;
-	regs.r01 = (unsigned long) fn;
-	pt_set_elr(&regs, (unsigned long)kernel_thread_helper);
-	pt_set_kmode(&regs);
-
-	return do_fork(flags|CLONE_VM|CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 /*
  * Program thread launch.  Often defined as a macro in processor.h,
  * but we're shooting for a small footprint and it's not an inner-loop
@@ -114,7 +87,7 @@ unsigned long thread_saved_pc(struct task_struct *tsk)
  * Copy architecture-specific thread state
  */
 int copy_thread(unsigned long clone_flags, unsigned long usp,
-		unsigned long unused, struct task_struct *p,
+		unsigned long arg, struct task_struct *p,
 		struct pt_regs *regs)
 {
 	struct thread_info *ti = task_thread_info(p);
@@ -125,61 +98,50 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	childregs = (struct pt_regs *) (((unsigned long) ti + THREAD_SIZE) -
 					sizeof(*childregs));
 
-	memcpy(childregs, regs, sizeof(*childregs));
 	ti->regs = childregs;
 
 	/*
 	 * Establish kernel stack pointer and initial PC for new thread
+	 * Note that unlike the usual situation, we do not copy the
+	 * parent's callee-saved here; those are in pt_regs and whatever
+	 * we leave here will be overridden on return to userland.
 	 */
 	ss = (struct hexagon_switch_stack *) ((unsigned long) childregs -
 						    sizeof(*ss));
 	ss->lr = (unsigned long)ret_from_fork;
 	p->thread.switch_sp = ss;
-
-	/* If User mode thread, set pt_reg stack pointer as per parameter */
-	if (user_mode(childregs)) {
-		pt_set_rte_sp(childregs, usp);
-
-		/* Child sees zero return value */
-		childregs->r00 = 0;
-
-		/*
-		 * The clone syscall has the C signature:
-		 * int [r0] clone(int flags [r0],
-		 *           void *child_frame [r1],
-		 *           void *parent_tid [r2],
-		 *           void *child_tid [r3],
-		 *           void *thread_control_block [r4]);
-		 * ugp is used to provide TLS support.
-		 */
-		if (clone_flags & CLONE_SETTLS)
-			childregs->ugp = childregs->r04;
-
-		/*
-		 * Parent sees new pid -- not necessary, not even possible at
-		 * this point in the fork process
-		 * Might also want to set things like ti->addr_limit
-		 */
-	} else {
-		/*
-		 * If kernel thread, resume stack is kernel stack base.
-		 * Note that this is pointer arithmetic on pt_regs *
-		 */
-		pt_set_rte_sp(childregs, (unsigned long)(childregs + 1));
-		/*
-		 * We need the current thread_info fast path pointer
-		 * set up in pt_regs.  The register to be used is
-		 * parametric for assembler code, but the mechanism
-		 * doesn't drop neatly into C.  Needs to be fixed.
-		 */
-		childregs->THREADINFO_REG = (unsigned long) ti;
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		/* r24 <- fn, r25 <- arg */
+		ss->r2524 = usp | ((u64)arg << 32);
+		pt_set_kmode(childregs);
+		return 0;
 	}
+	memcpy(childregs, regs, sizeof(*childregs));
+	ss->r2524 = 0;
+
+	pt_set_rte_sp(childregs, usp);
+
+	/* Child sees zero return value */
+	childregs->r00 = 0;
+
+	/*
+	 * The clone syscall has the C signature:
+	 * int [r0] clone(int flags [r0],
+	 *           void *child_frame [r1],
+	 *           void *parent_tid [r2],
+	 *           void *child_tid [r3],
+	 *           void *thread_control_block [r4]);
+	 * ugp is used to provide TLS support.
+	 */
+	if (clone_flags & CLONE_SETTLS)
+		childregs->ugp = childregs->r04;
 
 	/*
-	 * thread_info pointer is pulled out of task_struct "stack"
-	 * field on switch_to.
+	 * Parent sees new pid -- not necessary, not even possible at
+	 * this point in the fork process
+	 * Might also want to set things like ti->addr_limit
 	 */
-	p->stack = (void *)ti;
 
 	return 0;
 }

arch/hexagon/kernel/signal.c

@@ -249,14 +249,14 @@ void do_notify_resume(struct pt_regs *regs, unsigned long thread_info_flags)
  */
 asmlinkage int sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss)
 {
-	struct pt_regs *regs = current_thread_info()->regs;
+	struct pt_regs *regs = current_pt_regs();
 
 	return do_sigaltstack(uss, uoss, regs->r29);
 }
 
 asmlinkage int sys_rt_sigreturn(void)
 {
-	struct pt_regs *regs = current_thread_info()->regs;
+	struct pt_regs *regs = current_pt_regs();
 	struct rt_sigframe __user *frame;
 	sigset_t blocked;
 

arch/hexagon/kernel/syscall.c

@@ -35,55 +35,13 @@
  * See signal.c for signal-related system call wrappers.
  */
 
-asmlinkage int sys_execve(char __user *ufilename,
-			  const char __user *const __user *argv,
-			  const char __user *const __user *envp)
-{
-	struct pt_regs *pregs = current_thread_info()->regs;
-	struct filename *filename;
-	int retval;
-
-	filename = getname(ufilename);
-	retval = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		return retval;
-
-	retval = do_execve(filename->name, argv, envp, pregs);
-	putname(filename);
-
-	return retval;
-}
-
 asmlinkage int sys_clone(unsigned long clone_flags, unsigned long newsp,
 			 unsigned long parent_tidp, unsigned long child_tidp)
 {
-	struct pt_regs *pregs = current_thread_info()->regs;
+	struct pt_regs *pregs = current_pt_regs();
 
 	if (!newsp)
 		newsp = pregs->SP;
 	return do_fork(clone_flags, newsp, pregs, 0, (int __user *)parent_tidp,
 		       (int __user *)child_tidp);
 }
-
-/*
- * Do a system call from the kernel, so as to have a proper pt_regs
- * and recycle the sys_execvpe infrustructure.
- */
-int kernel_execve(const char *filename,
-		  const char *const argv[], const char *const envp[])
-{
-	register unsigned long __a0 asm("r0") = (unsigned long) filename;
-	register unsigned long __a1 asm("r1") = (unsigned long) argv;
-	register unsigned long __a2 asm("r2") = (unsigned long) envp;
-	int retval;
-
-	__asm__ volatile(
-		"	R6 = #%4;\n"
-		"	trap0(#1);\n"
-		"	%0 = R0;\n"
-		: "=r" (retval)
-		: "r" (__a0), "r" (__a1), "r" (__a2), "i" (__NR_execve)
-	);
-
-	return retval;
-}

arch/hexagon/kernel/vm_entry.S

@@ -266,4 +266,8 @@ _K_enter_machcheck:
 	.globl ret_from_fork
 ret_from_fork:
 	call schedule_tail
+	P0 = cmp.eq(R24, #0);
+	if P0 jump return_from_syscall
+	R0 = R25;
+	callr R24
 	jump return_from_syscall

arch/ia64/Kconfig

@@ -42,6 +42,8 @@ config IA64
 	select GENERIC_TIME_VSYSCALL_OLD
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	default y
 	help
 	  The Itanium Processor Family is Intel's 64-bit successor to

arch/ia64/include/asm/processor.h

@@ -340,22 +340,6 @@ struct task_struct;
  */
 #define release_thread(dead_task)
 
-/*
- * This is the mechanism for creating a new kernel thread.
- *
- * NOTE 1: Only a kernel-only process (ie the swapper or direct
- * descendants who haven't done an "execve()") should use this: it
- * will work within a system call from a "real" process, but the
- * process memory space will not be free'd until both the parent and
- * the child have exited.
- *
- * NOTE 2: This MUST NOT be an inlined function.  Otherwise, we get
- * into trouble in init/main.c when the child thread returns to
- * do_basic_setup() and the timing is such that free_initmem() has
- * been called already.
- */
-extern pid_t kernel_thread (int (*fn)(void *), void *arg, unsigned long flags);
-
 /* Get wait channel for task P.  */
 extern unsigned long get_wchan (struct task_struct *p);
 

arch/ia64/include/asm/unistd.h

@@ -29,6 +29,7 @@
 
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 #if !defined(__ASSEMBLY__) && !defined(ASSEMBLER)
 

arch/ia64/kernel/entry.S

@@ -61,14 +61,13 @@ ENTRY(ia64_execve)
 	 * Allocate 8 input registers since ptrace() may clobber them
 	 */
 	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
-	alloc loc1=ar.pfs,8,2,4,0
+	alloc loc1=ar.pfs,8,2,3,0
 	mov loc0=rp
 	.body
 	mov out0=in0			// filename
 	;;				// stop bit between alloc and call
 	mov out1=in1			// argv
 	mov out2=in2			// envp
-	add out3=16,sp			// regs
 	br.call.sptk.many rp=sys_execve
 .ret0:
 	cmp4.ge p6,p7=r8,r0
@@ -76,7 +75,6 @@ ENTRY(ia64_execve)
 	sxt4 r8=r8			// return 64-bit result
 	;;
 	stf.spill [sp]=f0
-(p6)	cmp.ne pKStk,pUStk=r0,r0	// a successful execve() lands us in user-mode...
 	mov rp=loc0
 (p6)	mov ar.pfs=r0			// clear ar.pfs on success
 (p7)	br.ret.sptk.many rp
@@ -484,19 +482,6 @@ GLOBAL_ENTRY(prefetch_stack)
 	br.ret.sptk.many rp
 END(prefetch_stack)
 
-GLOBAL_ENTRY(kernel_execve)
-	rum psr.ac
-	mov r15=__NR_execve			// put syscall number in place
-	break __BREAK_SYSCALL
-	br.ret.sptk.many rp
-END(kernel_execve)
-
-GLOBAL_ENTRY(clone)
-	mov r15=__NR_clone			// put syscall number in place
-	break __BREAK_SYSCALL
-	br.ret.sptk.many rp
-END(clone)
-
 	/*
 	 * Invoke a system call, but do some tracing before and after the call.
 	 * We MUST preserve the current register frame throughout this routine
@@ -600,6 +585,27 @@ GLOBAL_ENTRY(ia64_strace_leave_kernel)
 .ret4:	br.cond.sptk ia64_leave_kernel
 END(ia64_strace_leave_kernel)
 
+ENTRY(call_payload)
+	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(0)
+	/* call the kernel_thread payload; fn is in r4, arg - in r5 */
+	alloc loc1=ar.pfs,0,3,1,0
+	mov loc0=rp
+	mov loc2=gp
+	mov out0=r5		// arg
+	ld8 r14 = [r4], 8	// fn.address
+	;;
+	mov b6 = r14
+	ld8 gp = [r4]		// fn.gp
+	;;
+	br.call.sptk.many rp=b6	// fn(arg)
+.ret12:	mov gp=loc2
+	mov rp=loc0
+	mov ar.pfs=loc1
+	/* ... and if it has returned, we are going to userland */
+	cmp.ne pKStk,pUStk=r0,r0
+	br.ret.sptk.many rp
+END(call_payload)
+
 GLOBAL_ENTRY(ia64_ret_from_clone)
 	PT_REGS_UNWIND_INFO(0)
 {	/*
@@ -616,6 +622,7 @@ GLOBAL_ENTRY(ia64_ret_from_clone)
 	br.call.sptk.many rp=ia64_invoke_schedule_tail
 }
 .ret8:
+(pKStk)	br.call.sptk.many rp=call_payload
 	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
 	;;
 	ld4 r2=[r2]

arch/ia64/kernel/head.S

@@ -1093,19 +1093,6 @@ GLOBAL_ENTRY(cycle_to_cputime)
 END(cycle_to_cputime)
 #endif /* CONFIG_VIRT_CPU_ACCOUNTING */
 
-GLOBAL_ENTRY(start_kernel_thread)
-	.prologue
-	.save rp, r0				// this is the end of the call-chain
-	.body
-	alloc r2 = ar.pfs, 0, 0, 2, 0
-	mov out0 = r9
-	mov out1 = r11;;
-	br.call.sptk.many rp = kernel_thread_helper;;
-	mov out0 = r8
-	br.call.sptk.many rp = sys_exit;;
-1:	br.sptk.few 1b				// not reached
-END(start_kernel_thread)
-
 #ifdef CONFIG_IA64_BRL_EMU
 
 /*

arch/ia64/kernel/process.c

@@ -401,64 +401,15 @@ copy_thread(unsigned long clone_flags,
 	struct pt_regs *child_ptregs;
 	int retval = 0;
 
-#ifdef CONFIG_SMP
-	/*
-	 * For SMP idle threads, fork_by_hand() calls do_fork with
-	 * NULL regs.
-	 */
-	if (!regs)
-		return 0;
-#endif
-
-	stack = ((struct switch_stack *) regs) - 1;
-
 	child_ptregs = (struct pt_regs *) ((unsigned long) p + IA64_STK_OFFSET) - 1;
 	child_stack = (struct switch_stack *) child_ptregs - 1;
 
-	/* copy parent's switch_stack & pt_regs to child: */
-	memcpy(child_stack, stack, sizeof(*child_ptregs) + sizeof(*child_stack));
-
 	rbs = (unsigned long) current + IA64_RBS_OFFSET;
 	child_rbs = (unsigned long) p + IA64_RBS_OFFSET;
-	rbs_size = stack->ar_bspstore - rbs;
-
-	/* copy the parent's register backing store to the child: */
-	memcpy((void *) child_rbs, (void *) rbs, rbs_size);
-
-	if (likely(user_mode(child_ptregs))) {
-		if (clone_flags & CLONE_SETTLS)
-			child_ptregs->r13 = regs->r16;	/* see sys_clone2() in entry.S */
-		if (user_stack_base) {
-			child_ptregs->r12 = user_stack_base + user_stack_size - 16;
-			child_ptregs->ar_bspstore = user_stack_base;
-			child_ptregs->ar_rnat = 0;
-			child_ptregs->loadrs = 0;
-		}
-	} else {
-		/*
-		 * Note: we simply preserve the relative position of
-		 * the stack pointer here.  There is no need to
-		 * allocate a scratch area here, since that will have
-		 * been taken care of by the caller of sys_clone()
-		 * already.
-		 */
-		child_ptregs->r12 = (unsigned long) child_ptregs - 16; /* kernel sp */
-		child_ptregs->r13 = (unsigned long) p;		/* set `current' pointer */
-	}
-	child_stack->ar_bspstore = child_rbs + rbs_size;
-	child_stack->b0 = (unsigned long) &ia64_ret_from_clone;
 
 	/* copy parts of thread_struct: */
 	p->thread.ksp = (unsigned long) child_stack - 16;
 
-	/* stop some PSR bits from being inherited.
-	 * the psr.up/psr.pp bits must be cleared on fork but inherited on execve()
-	 * therefore we must specify them explicitly here and not include them in
-	 * IA64_PSR_BITS_TO_CLEAR.
-	 */
-	child_ptregs->cr_ipsr = ((child_ptregs->cr_ipsr | IA64_PSR_BITS_TO_SET)
-				 & ~(IA64_PSR_BITS_TO_CLEAR | IA64_PSR_PP | IA64_PSR_UP));
-
 	/*
 	 * NOTE: The calling convention considers all floating point
 	 * registers in the high partition (fph) to be scratch.  Since
@@ -480,8 +431,66 @@ copy_thread(unsigned long clone_flags,
 #	define THREAD_FLAGS_TO_SET	0
 	p->thread.flags = ((current->thread.flags & ~THREAD_FLAGS_TO_CLEAR)
 			   | THREAD_FLAGS_TO_SET);
+
 	ia64_drop_fpu(p);	/* don't pick up stale state from a CPU's fph */
 
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		if (unlikely(!user_stack_base)) {
+			/* fork_idle() called us */
+			return 0;
+		}
+		memset(child_stack, 0, sizeof(*child_ptregs) + sizeof(*child_stack));
+		child_stack->r4 = user_stack_base;	/* payload */
+		child_stack->r5 = user_stack_size;	/* argument */
+		/*
+		 * Preserve PSR bits, except for bits 32-34 and 37-45,
+		 * which we can't read.
+		 */
+		child_ptregs->cr_ipsr = ia64_getreg(_IA64_REG_PSR) | IA64_PSR_BN;
+		/* mark as valid, empty frame */
+		child_ptregs->cr_ifs = 1UL << 63;
+		child_stack->ar_fpsr = child_ptregs->ar_fpsr
+			= ia64_getreg(_IA64_REG_AR_FPSR);
+		child_stack->pr = (1 << PRED_KERNEL_STACK);
+		child_stack->ar_bspstore = child_rbs;
+		child_stack->b0 = (unsigned long) &ia64_ret_from_clone;
+
+		/* stop some PSR bits from being inherited.
+		 * the psr.up/psr.pp bits must be cleared on fork but inherited on execve()
+		 * therefore we must specify them explicitly here and not include them in
+		 * IA64_PSR_BITS_TO_CLEAR.
+		 */
+		child_ptregs->cr_ipsr = ((child_ptregs->cr_ipsr | IA64_PSR_BITS_TO_SET)
+				 & ~(IA64_PSR_BITS_TO_CLEAR | IA64_PSR_PP | IA64_PSR_UP));
+
+		return 0;
+	}
+	stack = ((struct switch_stack *) regs) - 1;
+	/* copy parent's switch_stack & pt_regs to child: */
+	memcpy(child_stack, stack, sizeof(*child_ptregs) + sizeof(*child_stack));
+
+	/* copy the parent's register backing store to the child: */
+	rbs_size = stack->ar_bspstore - rbs;
+	memcpy((void *) child_rbs, (void *) rbs, rbs_size);
+	if (clone_flags & CLONE_SETTLS)
+		child_ptregs->r13 = regs->r16;	/* see sys_clone2() in entry.S */
+	if (user_stack_base) {
+		child_ptregs->r12 = user_stack_base + user_stack_size - 16;
+		child_ptregs->ar_bspstore = user_stack_base;
+		child_ptregs->ar_rnat = 0;
+		child_ptregs->loadrs = 0;
+	}
+	child_stack->ar_bspstore = child_rbs + rbs_size;
+	child_stack->b0 = (unsigned long) &ia64_ret_from_clone;
+
+	/* stop some PSR bits from being inherited.
+	 * the psr.up/psr.pp bits must be cleared on fork but inherited on execve()
+	 * therefore we must specify them explicitly here and not include them in
+	 * IA64_PSR_BITS_TO_CLEAR.
+	 */
+	child_ptregs->cr_ipsr = ((child_ptregs->cr_ipsr | IA64_PSR_BITS_TO_SET)
+				 & ~(IA64_PSR_BITS_TO_CLEAR | IA64_PSR_PP | IA64_PSR_UP));
+
 #ifdef CONFIG_PERFMON
 	if (current->thread.pfm_context)
 		pfm_inherit(p, child_ptregs);
@@ -608,57 +617,6 @@ dump_fpu (struct pt_regs *pt, elf_fpregset_t dst)
 	return 1;	/* f0-f31 are always valid so we always return 1 */
 }
 
-long
-sys_execve (const char __user *filename,
-	    const char __user *const __user *argv,
-	    const char __user *const __user *envp,
-	    struct pt_regs *regs)
-{
-	struct filename *fname;
-	int error;
-
-	fname = getname(filename);
-	error = PTR_ERR(fname);
-	if (IS_ERR(fname))
-		goto out;
-	error = do_execve(fname->name, argv, envp, regs);
-	putname(fname);
-out:
-	return error;
-}
-
-pid_t
-kernel_thread (int (*fn)(void *), void *arg, unsigned long flags)
-{
-	extern void start_kernel_thread (void);
-	unsigned long *helper_fptr = (unsigned long *) &start_kernel_thread;
-	struct {
-		struct switch_stack sw;
-		struct pt_regs pt;
-	} regs;
-
-	memset(&regs, 0, sizeof(regs));
-	regs.pt.cr_iip = helper_fptr[0];	/* set entry point (IP) */
-	regs.pt.r1 = helper_fptr[1];		/* set GP */
-	regs.pt.r9 = (unsigned long) fn;	/* 1st argument */
-	regs.pt.r11 = (unsigned long) arg;	/* 2nd argument */
-	/* Preserve PSR bits, except for bits 32-34 and 37-45, which we can't read.  */
-	regs.pt.cr_ipsr = ia64_getreg(_IA64_REG_PSR) | IA64_PSR_BN;
-	regs.pt.cr_ifs = 1UL << 63;		/* mark as valid, empty frame */
-	regs.sw.ar_fpsr = regs.pt.ar_fpsr = ia64_getreg(_IA64_REG_AR_FPSR);
-	regs.sw.ar_bspstore = (unsigned long) current + IA64_RBS_OFFSET;
-	regs.sw.pr = (1 << PRED_KERNEL_STACK);
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs.pt, 0, NULL, NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
-/* This gets called from kernel_thread() via ia64_invoke_thread_helper().  */
-int
-kernel_thread_helper (int (*fn)(void *), void *arg)
-{
-	return (*fn)(arg);
-}
-
 /*
  * Flush thread state.  This is called when a thread does an execve().
  */

arch/m32r/Kconfig

@@ -15,6 +15,8 @@ config M32R
 	select GENERIC_ATOMIC64
 	select ARCH_USES_GETTIMEOFFSET
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config SBUS
 	bool

arch/m32r/include/asm/processor.h

@@ -118,11 +118,6 @@ struct mm_struct;
 /* Free all resources held by a thread. */
 extern void release_thread(struct task_struct *);
 
-/*
- * create a kernel thread without removing it from tasklists
- */
-extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
 /* Copy and release all segment info associated with a VM */
 extern void copy_segments(struct task_struct *p, struct mm_struct * mm);
 extern void release_segments(struct mm_struct * mm);

arch/m32r/include/asm/ptrace.h

@@ -139,6 +139,8 @@ extern void withdraw_debug_trap(struct pt_regs *regs);
 
 #define task_pt_regs(task) \
         ((struct pt_regs *)(task_stack_page(task) + THREAD_SIZE) - 1)
+#define current_pt_regs() ((struct pt_regs *) \
+	((unsigned long)current_thread_info() + THREAD_SIZE) - 1)
 
 #endif /* __KERNEL */
 

arch/m32r/include/asm/unistd.h

@@ -352,6 +352,7 @@
 #define __ARCH_WANT_SYS_OLDUMOUNT
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 #define __IGNORE_lchown
 #define __IGNORE_setuid

arch/m32r/kernel/entry.S

@@ -125,6 +125,15 @@
 	and	\reg, sp
 	.endm
 
+ENTRY(ret_from_kernel_thread)
+	pop	r0
+	bl	schedule_tail
+	GET_THREAD_INFO(r8)
+	ld	r0, R0(r8)
+	ld	r1, R1(r8)
+	jl	r1
+	bra	syscall_exit
+
 ENTRY(ret_from_fork)
 	pop	r0
 	bl	schedule_tail

arch/m32r/kernel/process.c

@@ -164,41 +164,6 @@ void show_regs(struct pt_regs * regs)
 #endif
 }
 
-/*
- * Create a kernel thread
- */
-
-/*
- * This is the mechanism for creating a new kernel thread.
- *
- * NOTE! Only a kernel-only process(ie the swapper or direct descendants
- * who haven't done an "execve()") should use this: it will work within
- * a system call from a "real" process, but the process memory space will
- * not be free'd until both the parent and the child have exited.
- */
-static void kernel_thread_helper(void *nouse, int (*fn)(void *), void *arg)
-{
-	fn(arg);
-	do_exit(-1);
-}
-
-int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof (regs));
-	regs.r1 = (unsigned long)fn;
-	regs.r2 = (unsigned long)arg;
-
-	regs.bpc = (unsigned long)kernel_thread_helper;
-
-	regs.psw = M32R_PSW_BIE;
-
-	/* Ok, create the new process. */
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL,
-		NULL);
-}
-
 /*
  * Free current thread data structures etc..
  */
@@ -227,29 +192,35 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 }
 
 int copy_thread(unsigned long clone_flags, unsigned long spu,
-	unsigned long unused, struct task_struct *tsk, struct pt_regs *regs)
+	unsigned long arg, struct task_struct *tsk, struct pt_regs *regs)
 {
 	struct pt_regs *childregs = task_pt_regs(tsk);
 	extern void ret_from_fork(void);
-
-	/* Copy registers */
-	*childregs = *regs;
-
-	childregs->spu = spu;
-	childregs->r0 = 0;	/* Child gets zero as return value */
-	regs->r0 = tsk->pid;
+	extern void ret_from_kernel_thread(void);
+
+	if (unlikely(tsk->flags & PF_KTHREAD)) {
+		memset(childregs, 0, sizeof(struct pt_regs));
+		childregs->psw = M32R_PSW_BIE;
+		childregs->r1 = spu;	/* fn */
+		childregs->r0 = arg;
+		tsk->thread.lr = (unsigned long)ret_from_kernel_thread;
+	} else {
+		/* Copy registers */
+		*childregs = *regs;
+		childregs->spu = spu;
+		childregs->r0 = 0;	/* Child gets zero as return value */
+		tsk->thread.lr = (unsigned long)ret_from_fork;
+	}
 	tsk->thread.sp = (unsigned long)childregs;
-	tsk->thread.lr = (unsigned long)ret_from_fork;
 
 	return 0;
 }
 
-asmlinkage int sys_fork(unsigned long r0, unsigned long r1, unsigned long r2,
-	unsigned long r3, unsigned long r4, unsigned long r5, unsigned long r6,
-	struct pt_regs regs)
+asmlinkage int sys_fork(void)
 {
 #ifdef CONFIG_MMU
-	return do_fork(SIGCHLD, regs.spu, &regs, 0, NULL, NULL);
+	struct pt_regs *regs = current_pt_regs();
+	return do_fork(SIGCHLD, regs->spu, regs, 0, NULL, NULL);
 #else
 	return -EINVAL;
 #endif /* CONFIG_MMU */
@@ -257,14 +228,13 @@ asmlinkage int sys_fork(unsigned long r0, unsigned long r1, unsigned long r2,
 
 asmlinkage int sys_clone(unsigned long clone_flags, unsigned long newsp,
 			 unsigned long parent_tidptr,
-			 unsigned long child_tidptr,
-			 unsigned long r4, unsigned long r5, unsigned long r6,
-			 struct pt_regs regs)
+			 unsigned long child_tidptr)
 {
+	struct pt_regs *regs = current_pt_regs();
 	if (!newsp)
-		newsp = regs.spu;
+		newsp = regs->spu;
 
-	return do_fork(clone_flags, newsp, &regs, 0,
+	return do_fork(clone_flags, newsp, regs, 0,
 		       (int __user *)parent_tidptr, (int __user *)child_tidptr);
 }
 
@@ -278,37 +248,13 @@ asmlinkage int sys_clone(unsigned long clone_flags, unsigned long newsp,
  * do not have enough call-clobbered registers to hold all
  * the information you need.
  */
-asmlinkage int sys_vfork(unsigned long r0, unsigned long r1, unsigned long r2,
-	unsigned long r3, unsigned long r4, unsigned long r5, unsigned long r6,
-	struct pt_regs regs)
+asmlinkage int sys_vfork(void)
 {
-	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs.spu, &regs, 0,
+	struct pt_regs *regs = current_pt_regs();
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs->spu, regs, 0,
 			NULL, NULL);
 }
 
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys_execve(const char __user *ufilename,
-			  const char __user *const __user *uargv,
-			  const char __user *const __user *uenvp,
-			  unsigned long r3, unsigned long r4, unsigned long r5,
-			  unsigned long r6, struct pt_regs regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(ufilename);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-
-	error = do_execve(filename->name, uargv, uenvp, &regs);
-	putname(filename);
-out:
-	return error;
-}
-
 /*
  * These bracket the sleeping functions..
  */

arch/m32r/kernel/sys_m32r.c

@@ -88,24 +88,3 @@ asmlinkage int sys_cachectl(char *addr, int nbytes, int op)
 	/* Not implemented yet. */
 	return -ENOSYS;
 }
-
-/*
- * Do a system call from kernel instead of calling sys_execve so we
- * end up with proper pt_regs.
- */
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	register long __scno __asm__ ("r7") = __NR_execve;
-	register long __arg3 __asm__ ("r2") = (long)(envp);
-	register long __arg2 __asm__ ("r1") = (long)(argv);
-	register long __res __asm__ ("r0") = (long)(filename);
-	__asm__ __volatile__ (
-		"trap #" SYSCALL_VECTOR "|| nop"
-		: "=r" (__res)
-		: "r" (__scno), "0" (__res), "r" (__arg2),
-			"r" (__arg3)
-		: "memory");
-	return __res;
-}

arch/m68k/Kconfig

@@ -16,6 +16,7 @@ config M68K
 	select ARCH_WANT_IPC_PARSE_VERSION
 	select ARCH_USES_GETTIMEOFFSET if MMU && !COLDFIRE
 	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_REL
 	select MODULES_USE_ELF_RELA

arch/m68k/include/asm/unistd.h

@@ -32,7 +32,6 @@
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_SYS_EXECVE
-#define __ARCH_WANT_KERNEL_EXECVE
 
 /*
  * "Conditional" syscalls

arch/m68k/kernel/entry.S

@@ -115,16 +115,9 @@ ENTRY(ret_from_kernel_thread)
 	| a3 contains the kernel thread payload, d7 - its argument
 	movel	%d1,%sp@-
 	jsr	schedule_tail
-	GET_CURRENT(%d0)
 	movel	%d7,(%sp)
 	jsr	%a3@
 	addql	#4,%sp
-	movel	%d0,(%sp)
-	jra	sys_exit
-
-ENTRY(ret_from_kernel_execve)
-	movel	4(%sp), %sp
-	GET_CURRENT(%d0)
 	jra	ret_from_exception
 
 #if defined(CONFIG_COLDFIRE) || !defined(CONFIG_MMU)

arch/microblaze/Kconfig

@@ -26,6 +26,8 @@ config MICROBLAZE
 	select GENERIC_ATOMIC64
 	select GENERIC_CLOCKEVENTS
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 config SWAP
 	def_bool n

arch/microblaze/include/asm/processor.h

@@ -31,6 +31,7 @@ extern const struct seq_operations cpuinfo_op;
 void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long usp);
 
 extern void ret_from_fork(void);
+extern void ret_from_kernel_thread(void);
 
 # endif /* __ASSEMBLY__ */
 
@@ -78,11 +79,6 @@ extern unsigned long thread_saved_pc(struct task_struct *t);
 
 extern unsigned long get_wchan(struct task_struct *p);
 
-/*
- * create a kernel thread without removing it from tasklists
- */
-extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
-
 # define KSTK_EIP(tsk)	(0)
 # define KSTK_ESP(tsk)	(0)
 
@@ -131,8 +127,6 @@ extern inline void release_thread(struct task_struct *dead_task)
 {
 }
 
-extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
-
 /* Free current thread data structures etc.  */
 static inline void exit_thread(void)
 {

arch/microblaze/include/asm/unistd.h

@@ -422,6 +422,7 @@
 #define __ARCH_WANT_SYS_SIGPROCMASK
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 /*
  * "Conditional" syscalls

arch/microblaze/kernel/entry-nommu.S

@@ -474,6 +474,14 @@ ENTRY(ret_from_fork)
 	brid	ret_to_user
 	nop
 
+ENTRY(ret_from_kernel_thread)
+	brlid	r15, schedule_tail
+	addk	r5, r0, r3
+	brald	r15, r20
+	addk	r5, r0, r19
+	brid	ret_to_user
+	addk	r3, r0, r0
+
 work_pending:
 	enable_irq
 
@@ -559,10 +567,6 @@ sys_clone:
 	brid	microblaze_clone
 	addk	r7, r1, r0
 
-sys_execve:
-	brid	microblaze_execve
-	addk	r8, r1, r0
-
 sys_rt_sigreturn_wrapper:
 	brid	sys_rt_sigreturn
 	addk	r5, r1, r0

arch/microblaze/kernel/entry.S

@@ -293,24 +293,6 @@ C_ENTRY(_user_exception):
 	swi	r1, r0, TOPHYS(PER_CPU(ENTRY_SP)) /* save stack */
 	addi	r14, r14, 4	/* return address is 4 byte after call */
 
-	mfs	r1, rmsr
-	nop
-	andi	r1, r1, MSR_UMS
-	bnei	r1, 1f
-
-/* Kernel-mode state save - kernel execve */
-	lwi	r1, r0, TOPHYS(PER_CPU(ENTRY_SP)); /* Reload kernel stack-ptr*/
-	tophys(r1,r1);
-
-	addik	r1, r1, -PT_SIZE; /* Make room on the stack. */
-	SAVE_REGS
-
-	swi	r1, r1, PT_MODE; /* pt_regs -> kernel mode */
-	brid	2f;
-	nop;				/* Fill delay slot */
-
-/* User-mode state save.  */
-1:
 	lwi	r1, r0, TOPHYS(PER_CPU(CURRENT_SAVE)); /* get saved current */
 	tophys(r1,r1);
 	lwi	r1, r1, TS_THREAD_INFO;	/* get stack from task_struct */
@@ -479,11 +461,20 @@ C_ENTRY(sys_fork_wrapper):
    saved context).  */
 C_ENTRY(ret_from_fork):
 	bralid	r15, schedule_tail; /* ...which is schedule_tail's arg */
-	add	r3, r5, r0;	/* switch_thread returns the prev task */
+	add	r5, r3, r0;	/* switch_thread returns the prev task */
 				/* ( in the delay slot ) */
 	brid	ret_from_trap;	/* Do normal trap return */
 	add	r3, r0, r0;	/* Child's fork call should return 0. */
 
+C_ENTRY(ret_from_kernel_thread):
+	bralid	r15, schedule_tail; /* ...which is schedule_tail's arg */
+	add	r5, r3, r0;	/* switch_thread returns the prev task */
+				/* ( in the delay slot ) */
+	brald	r15, r20	/* fn was left in r20 */
+	addk	r5, r0, r19	/* ... and argument - in r19 */
+	brid	ret_from_trap
+	add	r3, r0, r0
+
 C_ENTRY(sys_vfork):
 	brid	microblaze_vfork	/* Do real work (tail-call) */
 	addik	r5, r1, 0
@@ -498,10 +489,6 @@ C_ENTRY(sys_clone):
 	brid	do_fork		/* Do real work (tail-call) */
 	add     r8, r0, r0;             /* Arg 3: (unused) */
 
-C_ENTRY(sys_execve):
-	brid	microblaze_execve;	/* Do real work (tail-call).*/
-	addik	r8, r1, 0;		/* add user context as 4th arg */
-
 C_ENTRY(sys_rt_sigreturn_wrapper):
 	brid	sys_rt_sigreturn	/* Do real work */
 	addik	r5, r1, 0;		/* add user context as 1st arg */

arch/microblaze/kernel/process.c

@@ -119,46 +119,38 @@ void flush_thread(void)
 }
 
 int copy_thread(unsigned long clone_flags, unsigned long usp,
-		unsigned long unused,
+		unsigned long arg,
 		struct task_struct *p, struct pt_regs *regs)
 {
 	struct pt_regs *childregs = task_pt_regs(p);
 	struct thread_info *ti = task_thread_info(p);
 
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		/* if we're creating a new kernel thread then just zeroing all
+		 * the registers. That's OK for a brand new thread.*/
+		memset(childregs, 0, sizeof(struct pt_regs));
+		memset(&ti->cpu_context, 0, sizeof(struct cpu_context));
+		ti->cpu_context.r1  = (unsigned long)childregs;
+		ti->cpu_context.r20 = (unsigned long)usp; /* fn */
+		ti->cpu_context.r19 = (unsigned long)arg;
+		childregs->pt_mode = 1;
+		local_save_flags(childregs->msr);
+#ifdef CONFIG_MMU
+		ti->cpu_context.msr = childregs->msr & ~MSR_IE;
+#endif
+		ti->cpu_context.r15 = (unsigned long)ret_from_kernel_thread - 8;
+		return 0;
+	}
 	*childregs = *regs;
-	if (user_mode(regs))
-		childregs->r1 = usp;
-	else
-		childregs->r1 = ((unsigned long) ti) + THREAD_SIZE;
+	childregs->r1 = usp;
 
-#ifndef CONFIG_MMU
 	memset(&ti->cpu_context, 0, sizeof(struct cpu_context));
 	ti->cpu_context.r1 = (unsigned long)childregs;
+#ifndef CONFIG_MMU
 	ti->cpu_context.msr = (unsigned long)childregs->msr;
 #else
+	childregs->msr |= MSR_UMS;
 
-	/* if creating a kernel thread then update the current reg (we don't
-	 * want to use the parent's value when restoring by POP_STATE) */
-	if (kernel_mode(regs))
-		/* save new current on stack to use POP_STATE */
-		childregs->CURRENT_TASK = (unsigned long)p;
-	/* if returning to user then use the parent's value of this register */
-
-	/* if we're creating a new kernel thread then just zeroing all
-	 * the registers. That's OK for a brand new thread.*/
-	/* Pls. note that some of them will be restored in POP_STATE */
-	if (kernel_mode(regs))
-		memset(&ti->cpu_context, 0, sizeof(struct cpu_context));
-	/* if this thread is created for fork/vfork/clone, then we want to
-	 * restore all the parent's context */
-	/* in addition to the registers which will be restored by POP_STATE */
-	else {
-		ti->cpu_context = *(struct cpu_context *)regs;
-		childregs->msr |= MSR_UMS;
-	}
-
-	/* FIXME STATE_SAVE_PT_OFFSET; */
-	ti->cpu_context.r1  = (unsigned long)childregs;
 	/* we should consider the fact that childregs is a copy of the parent
 	 * regs which were saved immediately after entering the kernel state
 	 * before enabling VM. This MSR will be restored in switch_to and
@@ -209,29 +201,6 @@ unsigned long thread_saved_pc(struct task_struct *tsk)
 }
 #endif
 
-static void kernel_thread_helper(int (*fn)(void *), void *arg)
-{
-	fn(arg);
-	do_exit(-1);
-}
-
-int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-	/* store them in non-volatile registers */
-	regs.r5 = (unsigned long)fn;
-	regs.r6 = (unsigned long)arg;
-	local_save_flags(regs.msr);
-	regs.pc = (unsigned long)kernel_thread_helper;
-	regs.pt_mode = 1;
-
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
-			&regs, 0, NULL, NULL);
-}
-EXPORT_SYMBOL_GPL(kernel_thread);
-
 unsigned long get_wchan(struct task_struct *p)
 {
 /* TBD (used by procfs) */
@@ -246,6 +215,7 @@ void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long usp)
 	regs->pt_mode = 0;
 #ifdef CONFIG_MMU
 	regs->msr |= MSR_UMS;
+	regs->msr &= ~MSR_VM;
 #endif
 }
 

arch/microblaze/kernel/sys_microblaze.c

@@ -48,24 +48,6 @@ asmlinkage long microblaze_clone(int flags, unsigned long stack,
 	return do_fork(flags, stack, regs, 0, NULL, NULL);
 }
 
-asmlinkage long microblaze_execve(const char __user *filenamei,
-				  const char __user *const __user *argv,
-				  const char __user *const __user *envp,
-				  struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(filenamei);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-out:
-	return error;
-}
-
 asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
 			unsigned long prot, unsigned long flags,
 			unsigned long fd, off_t pgoff)
@@ -75,24 +57,3 @@ asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
 
 	return sys_mmap_pgoff(addr, len, prot, flags, fd, pgoff >> PAGE_SHIFT);
 }
-
-/*
- * Do a system call from kernel instead of calling sys_execve so we
- * end up with proper pt_regs.
- */
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	register const char *__a __asm__("r5") = filename;
-	register const void *__b __asm__("r6") = argv;
-	register const void *__c __asm__("r7") = envp;
-	register unsigned long __syscall __asm__("r12") = __NR_execve;
-	register unsigned long __ret __asm__("r3");
-	__asm__ __volatile__ ("brki r14, 0x8"
-			: "=r" (__ret), "=r" (__syscall)
-			: "1" (__syscall), "r" (__a), "r" (__b), "r" (__c)
-			: "r4", "r8", "r9",
-			"r10", "r11", "r14", "cc", "memory");
-	return __ret;
-}

arch/mips/Kconfig

@@ -40,6 +40,8 @@ config MIPS
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_REL
 	select MODULES_USE_ELF_RELA if 64BIT
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 menu "Machine selection"
 

arch/mips/include/asm/processor.h

@@ -310,8 +310,6 @@ struct task_struct;
 /* Free all resources held by a thread. */
 #define release_thread(thread) do { } while(0)
 
-extern long kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
 extern unsigned long thread_saved_pc(struct task_struct *tsk);
 
 /*

arch/mips/include/asm/ptrace.h

@@ -61,4 +61,10 @@ static inline void die_if_kernel(const char *str, struct pt_regs *regs)
 		die(str, regs);
 }
 
+#define current_pt_regs()						\
+({									\
+	unsigned long sp = (unsigned long)__builtin_frame_address(0);	\
+	(struct pt_regs *)((sp | (THREAD_SIZE - 1)) + 1 - 32) - 1;	\
+})
+
 #endif /* _ASM_PTRACE_H */

arch/mips/include/asm/unistd.h

@@ -20,6 +20,7 @@
 #define __ARCH_OMIT_COMPAT_SYS_GETDENTS64
 #define __ARCH_WANT_OLD_READDIR
 #define __ARCH_WANT_SYS_ALARM
+#define __ARCH_WANT_SYS_EXECVE
 #define __ARCH_WANT_SYS_GETHOSTNAME
 #define __ARCH_WANT_SYS_IPC
 #define __ARCH_WANT_SYS_PAUSE

arch/mips/kernel/entry.S

@@ -65,6 +65,12 @@ need_resched:
 	b	need_resched
 #endif
 
+FEXPORT(ret_from_kernel_thread)
+	jal	schedule_tail		# a0 = struct task_struct *prev
+	move	a0, s1
+	jal	s0
+	j	syscall_exit
+
 FEXPORT(ret_from_fork)
 	jal	schedule_tail		# a0 = struct task_struct *prev
 

arch/mips/kernel/linux32.c

@@ -3,7 +3,6 @@
  *
  * Copyright (C) 2000 Silicon Graphics, Inc.
  * Written by Ulf Carlsson (ulfc@engr.sgi.com)
- * sys32_execve from ia64/ia32 code, Feb 2000, Kanoj Sarcar (kanoj@sgi.com)
  */
 #include <linux/compiler.h>
 #include <linux/mm.h>
@@ -77,26 +76,6 @@ out:
 	return error;
 }
 
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys32_execve(nabi_no_regargs struct pt_regs regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname(compat_ptr(regs.regs[4]));
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = compat_do_execve(filename->name, compat_ptr(regs.regs[5]),
-				 compat_ptr(regs.regs[6]), &regs);
-	putname(filename);
-
-out:
-	return error;
-}
-
 #define RLIM_INFINITY32	0x7fffffff
 #define RESOURCE32(x) ((x > RLIM_INFINITY32) ? RLIM_INFINITY32 : x)
 

arch/mips/kernel/mips_ksyms.c

@@ -32,8 +32,6 @@ EXPORT_SYMBOL(memset);
 EXPORT_SYMBOL(memcpy);
 EXPORT_SYMBOL(memmove);
 
-EXPORT_SYMBOL(kernel_thread);
-
 /*
  * Functions that operate on entire pages.  Mostly used by memory management.
  */

arch/mips/kernel/process.c

@@ -84,6 +84,7 @@ void __noreturn cpu_idle(void)
 }
 
 asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
 
 void start_thread(struct pt_regs * regs, unsigned long pc, unsigned long sp)
 {
@@ -113,7 +114,7 @@ void flush_thread(void)
 }
 
 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 *regs)
 {
 	struct thread_info *ti = task_thread_info(p);
 	struct pt_regs *childregs;
@@ -136,19 +137,30 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	childregs = (struct pt_regs *) childksp - 1;
 	/*  Put the stack after the struct pt_regs.  */
 	childksp = (unsigned long) childregs;
+	p->thread.cp0_status = read_c0_status() & ~(ST0_CU2|ST0_CU1);
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		unsigned long status = p->thread.cp0_status;
+		memset(childregs, 0, sizeof(struct pt_regs));
+		ti->addr_limit = KERNEL_DS;
+		p->thread.reg16 = usp; /* fn */
+		p->thread.reg17 = arg;
+		p->thread.reg29 = childksp;
+		p->thread.reg31 = (unsigned long) ret_from_kernel_thread;
+#if defined(CONFIG_CPU_R3000) || defined(CONFIG_CPU_TX39XX)
+		status = (status & ~(ST0_KUP | ST0_IEP | ST0_IEC)) |
+			 ((status & (ST0_KUC | ST0_IEC)) << 2);
+#else
+		status |= ST0_EXL;
+#endif
+		childregs->cp0_status = status;
+		return 0;
+	}
 	*childregs = *regs;
 	childregs->regs[7] = 0;	/* Clear error flag */
-
 	childregs->regs[2] = 0;	/* Child gets zero as return value */
+	childregs->regs[29] = usp;
+	ti->addr_limit = USER_DS;
 
-	if (childregs->cp0_status & ST0_CU0) {
-		childregs->regs[28] = (unsigned long) ti;
-		childregs->regs[29] = childksp;
-		ti->addr_limit = KERNEL_DS;
-	} else {
-		childregs->regs[29] = usp;
-		ti->addr_limit = USER_DS;
-	}
 	p->thread.reg29 = (unsigned long) childregs;
 	p->thread.reg31 = (unsigned long) ret_from_fork;
 
@@ -156,7 +168,6 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	 * New tasks lose permission to use the fpu. This accelerates context
 	 * switching for most programs since they don't use the fpu.
 	 */
-	p->thread.cp0_status = read_c0_status() & ~(ST0_CU2|ST0_CU1);
 	childregs->cp0_status &= ~(ST0_CU2|ST0_CU1);
 
 #ifdef CONFIG_MIPS_MT_SMTC
@@ -221,35 +232,6 @@ int dump_task_fpu(struct task_struct *t, elf_fpregset_t *fpr)
 	return 1;
 }
 
-/*
- * Create a kernel thread
- */
-static void __noreturn kernel_thread_helper(void *arg, int (*fn)(void *))
-{
-	do_exit(fn(arg));
-}
-
-long kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-
-	regs.regs[4] = (unsigned long) arg;
-	regs.regs[5] = (unsigned long) fn;
-	regs.cp0_epc = (unsigned long) kernel_thread_helper;
-	regs.cp0_status = read_c0_status();
-#if defined(CONFIG_CPU_R3000) || defined(CONFIG_CPU_TX39XX)
-	regs.cp0_status = (regs.cp0_status & ~(ST0_KUP | ST0_IEP | ST0_IEC)) |
-			  ((regs.cp0_status & (ST0_KUC | ST0_IEC)) << 2);
-#else
-	regs.cp0_status |= ST0_EXL;
-#endif
-
-	/* Ok, create the new process.. */
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
-}
-
 /*
  *
  */

arch/mips/kernel/scall64-n32.S

@@ -167,7 +167,7 @@ EXPORT(sysn32_call_table)
 	PTR	sys_getsockopt
 	PTR	sys_clone			/* 6055 */
 	PTR	sys_fork
-	PTR	sys32_execve
+	PTR	compat_sys_execve
 	PTR	sys_exit
 	PTR	compat_sys_wait4
 	PTR	sys_kill			/* 6060 */

arch/mips/kernel/scall64-o32.S

@@ -203,7 +203,7 @@ sys_call_table:
 	PTR	sys_creat
 	PTR	sys_link
 	PTR	sys_unlink			/* 4010 */
-	PTR	sys32_execve
+	PTR	compat_sys_execve
 	PTR	sys_chdir
 	PTR	compat_sys_time
 	PTR	sys_mknod

arch/mips/kernel/syscall.c

@@ -127,28 +127,6 @@ _sys_clone(nabi_no_regargs struct pt_regs regs)
 	               parent_tidptr, child_tidptr);
 }
 
-/*
- * sys_execve() executes a new program.
- */
-asmlinkage int sys_execve(nabi_no_regargs struct pt_regs regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname((const char __user *) (long)regs.regs[4]);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = do_execve(filename->name,
-			  (const char __user *const __user *) (long)regs.regs[5],
-	                  (const char __user *const __user *) (long)regs.regs[6],
-			  &regs);
-	putname(filename);
-
-out:
-	return error;
-}
-
 SYSCALL_DEFINE1(set_thread_area, unsigned long, addr)
 {
 	struct thread_info *ti = task_thread_info(current);
@@ -313,34 +291,3 @@ asmlinkage void bad_stack(void)
 {
 	do_exit(SIGSEGV);
 }
-
-/*
- * Do a system call from kernel instead of calling sys_execve so we
- * end up with proper pt_regs.
- */
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	register unsigned long __a0 asm("$4") = (unsigned long) filename;
-	register unsigned long __a1 asm("$5") = (unsigned long) argv;
-	register unsigned long __a2 asm("$6") = (unsigned long) envp;
-	register unsigned long __a3 asm("$7");
-	unsigned long __v0;
-
-	__asm__ volatile ("					\n"
-	"	.set	noreorder				\n"
-	"	li	$2, %5		# __NR_execve		\n"
-	"	syscall						\n"
-	"	move	%0, $2					\n"
-	"	.set	reorder					\n"
-	: "=&r" (__v0), "=r" (__a3)
-	: "r" (__a0), "r" (__a1), "r" (__a2), "i" (__NR_execve)
-	: "$2", "$8", "$9", "$10", "$11", "$12", "$13", "$14", "$15", "$24",
-	  "memory");
-
-	if (__a3 == 0)
-		return __v0;
-
-	return -__v0;
-}

arch/mn10300/Kconfig

@@ -9,6 +9,7 @@ config MN10300
 	select HAVE_NMI_WATCHDOG if MN10300_WD_TIMER
 	select GENERIC_CLOCKEVENTS
 	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 	select MODULES_USE_ELF_RELA
 
 config AM33_2

arch/mn10300/include/asm/unistd.h

@@ -44,7 +44,6 @@
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_SYS_EXECVE
-#define __ARCH_WANT_KERNEL_EXECVE
 
 /*
  * "Conditional" syscalls

arch/mn10300/kernel/entry.S

@@ -60,13 +60,8 @@ ENTRY(ret_from_kernel_thread)
 	mov	(REG_D0,fp),d0
 	mov	(REG_A0,fp),a0
 	calls	(a0)
-	jmp	sys_exit
-
-ENTRY(ret_from_kernel_execve)
-	add	-12,d0	/* pt_regs -> frame */
-	mov	d0,sp
-	GET_THREAD_INFO a2
 	clr	d0
+	mov	d0,(REG_D0,fp)
 	jmp	syscall_exit
 
 ###############################################################################

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

arch/openrisc/include/uapi/asm/unistd.h

@@ -20,6 +20,8 @@
 
 #define sys_mmap2 sys_mmap_pgoff
 
+#define __ARCH_WANT_SYS_EXECVE
+
 #include <asm-generic/unistd.h>
 
 #define __NR_or1k_atomic __NR_arch_specific_syscall

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
 
@@ -1088,10 +1083,6 @@ ENTRY(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

arch/openrisc/kernel/process.c

@@ -109,66 +109,82 @@ 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 *regs)
 {
-	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 = *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;
+		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 +193,14 @@ 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 +251,9 @@ 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;
-}

arch/parisc/Kconfig

@@ -22,6 +22,8 @@ config PARISC
 	select GENERIC_STRNCPY_FROM_USER
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_THREAD
+	select GENERIC_KERNEL_EXECVE
 
 	help
 	  The PA-RISC microprocessor is designed by Hewlett-Packard and used

arch/parisc/include/asm/unistd.h

@@ -163,6 +163,7 @@ type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4, type5 arg5)	\
 #define __ARCH_WANT_SYS_RT_SIGACTION
 #define __ARCH_WANT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_SYS_EXECVE
 
 #endif /* __ASSEMBLY__ */
 

arch/parisc/kernel/entry.S

@@ -707,60 +707,10 @@ ENTRY(end_fault_vector)
 	.import		handle_interruption,code
 	.import		do_cpu_irq_mask,code
 
-	/*
-	 * r26 = function to be called
-	 * r25 = argument to pass in
-	 * r24 = flags for do_fork()
-	 *
-	 * Kernel threads don't ever return, so they don't need
-	 * a true register context. We just save away the arguments
-	 * for copy_thread/ret_ to properly set up the child.
-	 */
-
-#define CLONE_VM 0x100	/* Must agree with <linux/sched.h> */
-#define CLONE_UNTRACED 0x00800000
-
-	.import do_fork
-ENTRY(__kernel_thread)
-	STREG	%r2, -RP_OFFSET(%r30)
-
-	copy	%r30, %r1
-	ldo	PT_SZ_ALGN(%r30),%r30
-#ifdef CONFIG_64BIT
-	/* Yo, function pointers in wide mode are little structs... -PB */
-	ldd	24(%r26), %r2
-	STREG	%r2, PT_GR27(%r1)	/* Store childs %dp */
-	ldd	16(%r26), %r26
-
-	STREG	%r22, PT_GR22(%r1)	/* save r22 (arg5) */
-	copy	%r0, %r22		/* user_tid */
-#endif
-	STREG	%r26, PT_GR26(%r1)  /* Store function & argument for child */
-	STREG	%r25, PT_GR25(%r1)
-	ldil	L%CLONE_UNTRACED, %r26
-	ldo	CLONE_VM(%r26), %r26   /* Force CLONE_VM since only init_mm */
-	or	%r26, %r24, %r26      /* will have kernel mappings.	 */
-	ldi	1, %r25			/* stack_start, signals kernel thread */
-	stw	%r0, -52(%r30)	     	/* user_tid */
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-	BL	do_fork, %r2
-	copy	%r1, %r24		/* pt_regs */
-
-	/* Parent Returns here */
-
-	LDREG	-PT_SZ_ALGN-RP_OFFSET(%r30), %r2
-	ldo	-PT_SZ_ALGN(%r30), %r30
-	bv	%r0(%r2)
-	nop
-ENDPROC(__kernel_thread)
-
 	/*
 	 * Child Returns here
 	 *
-	 * copy_thread moved args from temp save area set up above
-	 * into task save area.
+	 * copy_thread moved args into task save area.
 	 */
 
 ENTRY(ret_from_kernel_thread)
@@ -769,51 +719,17 @@ ENTRY(ret_from_kernel_thread)
 	BL	schedule_tail, %r2
 	nop
 
-	LDREG	TI_TASK-THREAD_SZ_ALGN(%r30), %r1
+	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30), %r1
 	LDREG	TASK_PT_GR25(%r1), %r26
 #ifdef CONFIG_64BIT
 	LDREG	TASK_PT_GR27(%r1), %r27
-	LDREG	TASK_PT_GR22(%r1), %r22
 #endif
 	LDREG	TASK_PT_GR26(%r1), %r1
 	ble	0(%sr7, %r1)
 	copy	%r31, %r2
-
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-	loadgp				/* Thread could have been in a module */
-#endif
-#ifndef CONFIG_64BIT
-	b	sys_exit
-#else
-	load32	sys_exit, %r1
-	bv	%r0(%r1)
-#endif
-	ldi	0, %r26
-ENDPROC(ret_from_kernel_thread)
-
-	.import	sys_execve, code
-ENTRY(__execve)
-	copy	%r2, %r15
-	copy	%r30, %r16
-	ldo	PT_SZ_ALGN(%r30), %r30
-	STREG	%r26, PT_GR26(%r16)
-	STREG	%r25, PT_GR25(%r16)
-	STREG	%r24, PT_GR24(%r16)
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-	BL	sys_execve, %r2
-	copy	%r16, %r26
-
-	cmpib,=,n 0,%r28,intr_return    /* forward */
-
-	/* yes, this will trap and die. */
-	copy	%r15, %r2
-	copy	%r16, %r30
-	bv	%r0(%r2)
+	b	finish_child_return
 	nop
-ENDPROC(__execve)
+ENDPROC(ret_from_kernel_thread)
 
 
 	/*
@@ -1776,49 +1692,27 @@ ENTRY(sys_fork_wrapper)
 	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30), %r1
 	ldo	TASK_REGS(%r1),%r1
 	reg_save %r1
-	mfctl	%cr27, %r3
-	STREG	%r3, PT_CR27(%r1)
-
-	STREG	%r2,-RP_OFFSET(%r30)
-	ldo	FRAME_SIZE(%r30),%r30
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-
-	/* These are call-clobbered registers and therefore
-	   also syscall-clobbered (we hope). */
-	STREG	%r2,PT_GR19(%r1)	/* save for child */
-	STREG	%r30,PT_GR21(%r1)
+	mfctl	%cr27, %r28
+	STREG	%r28, PT_CR27(%r1)
 
 	LDREG	PT_GR30(%r1),%r25
 	copy	%r1,%r24
-	BL	sys_clone,%r2
+	b	sys_clone
 	ldi	SIGCHLD,%r26
-
-	LDREG	-RP_OFFSET-FRAME_SIZE(%r30),%r2
-wrapper_exit:
-	ldo	-FRAME_SIZE(%r30),%r30		/* get the stackframe */
-	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30),%r1
-	ldo	TASK_REGS(%r1),%r1	 /* get pt regs */
-
-	LDREG	PT_CR27(%r1), %r3
-	mtctl	%r3, %cr27
-	reg_restore %r1
-
-	/* strace expects syscall # to be preserved in r20 */
-	ldi	__NR_fork,%r20
-	bv %r0(%r2)
-	STREG	%r20,PT_GR20(%r1)
 ENDPROC(sys_fork_wrapper)
 
 	/* Set the return value for the child */
 ENTRY(child_return)
 	BL	schedule_tail, %r2
 	nop
+finish_child_return:
+	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30), %r1
+	ldo	TASK_REGS(%r1),%r1	 /* get pt regs */
 
-	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE-FRAME_SIZE(%r30), %r1
-	LDREG	TASK_PT_GR19(%r1),%r2
-	b	wrapper_exit
+	LDREG	PT_CR27(%r1), %r3
+	mtctl	%r3, %cr27
+	reg_restore %r1
+	b	syscall_exit
 	copy	%r0,%r28
 ENDPROC(child_return)
 
@@ -1827,23 +1721,10 @@ ENTRY(sys_clone_wrapper)
 	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30),%r1
 	ldo	TASK_REGS(%r1),%r1	/* get pt regs */
 	reg_save %r1
-	mfctl	%cr27, %r3
-	STREG	%r3, PT_CR27(%r1)
-
-	STREG	%r2,-RP_OFFSET(%r30)
-	ldo	FRAME_SIZE(%r30),%r30
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-
-	/* WARNING - Clobbers r19 and r21, userspace must save these! */
-	STREG	%r2,PT_GR19(%r1)	/* save for child */
-	STREG	%r30,PT_GR21(%r1)
-	BL	sys_clone,%r2
+	mfctl	%cr27, %r28
+	STREG	%r28, PT_CR27(%r1)
+	b	sys_clone
 	copy	%r1,%r24
-
-	b	wrapper_exit
-	LDREG	-RP_OFFSET-FRAME_SIZE(%r30),%r2
 ENDPROC(sys_clone_wrapper)
 
 
@@ -1851,72 +1732,14 @@ ENTRY(sys_vfork_wrapper)
 	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30),%r1
 	ldo	TASK_REGS(%r1),%r1	/* get pt regs */
 	reg_save %r1
-	mfctl	%cr27, %r3
-	STREG	%r3, PT_CR27(%r1)
+	mfctl	%cr27, %r28
+	STREG	%r28, PT_CR27(%r1)
 
-	STREG	%r2,-RP_OFFSET(%r30)
-	ldo	FRAME_SIZE(%r30),%r30
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-
-	STREG	%r2,PT_GR19(%r1)	/* save for child */
-	STREG	%r30,PT_GR21(%r1)
-
-	BL	sys_vfork,%r2
+	b	sys_vfork
 	copy	%r1,%r26
-
-	b	wrapper_exit
-	LDREG	-RP_OFFSET-FRAME_SIZE(%r30),%r2
 ENDPROC(sys_vfork_wrapper)
 
 	
-	.macro  execve_wrapper execve
-	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30),%r1
-	ldo	TASK_REGS(%r1),%r1	/* get pt regs */
-
-	/*
-	 * Do we need to save/restore r3-r18 here?
-	 * I don't think so. why would new thread need old
-	 * threads registers?
-	 */
-
-	/* %arg0 - %arg3 are already saved for us. */
-
-	STREG %r2,-RP_OFFSET(%r30)
-	ldo FRAME_SIZE(%r30),%r30
-#ifdef CONFIG_64BIT
-	ldo	-16(%r30),%r29		/* Reference param save area */
-#endif
-	BL \execve,%r2
-	copy %r1,%arg0
-
-	ldo -FRAME_SIZE(%r30),%r30
-	LDREG -RP_OFFSET(%r30),%r2
-
-	/* If exec succeeded we need to load the args */
-
-	ldo -1024(%r0),%r1
-	cmpb,>>= %r28,%r1,error_\execve
-	copy %r2,%r19
-
-error_\execve:
-	bv %r0(%r19)
-	nop
-	.endm
-
-	.import sys_execve
-ENTRY(sys_execve_wrapper)
-	execve_wrapper sys_execve
-ENDPROC(sys_execve_wrapper)
-
-#ifdef CONFIG_64BIT
-	.import sys32_execve
-ENTRY(sys32_execve_wrapper)
-	execve_wrapper sys32_execve
-ENDPROC(sys32_execve_wrapper)
-#endif
-
 ENTRY(sys_rt_sigreturn_wrapper)
 	LDREG	TI_TASK-THREAD_SZ_ALGN-FRAME_SIZE(%r30),%r26
 	ldo	TASK_REGS(%r26),%r26	/* get pt regs */

arch/parisc/kernel/process.c

@@ -52,6 +52,7 @@
 
 #include <asm/io.h>
 #include <asm/asm-offsets.h>
+#include <asm/assembly.h>
 #include <asm/pdc.h>
 #include <asm/pdc_chassis.h>
 #include <asm/pgalloc.h>
@@ -164,23 +165,6 @@ void machine_power_off(void)
 void (*pm_power_off)(void) = machine_power_off;
 EXPORT_SYMBOL(pm_power_off);
 
-/*
- * Create a kernel thread
- */
-
-extern pid_t __kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
-pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
-{
-
-	/*
-	 * FIXME: Once we are sure we don't need any debug here,
-	 *	  kernel_thread can become a #define.
-	 */
-
-	return __kernel_thread(fn, arg, flags);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 /*
  * Free current thread data structures etc..
  */
@@ -256,8 +240,8 @@ sys_vfork(struct pt_regs *regs)
 
 int
 copy_thread(unsigned long clone_flags, unsigned long usp,
-	    unsigned long unused,	/* in ia64 this is "user_stack_size" */
-	    struct task_struct * p, struct pt_regs * pregs)
+	    unsigned long arg,
+	    struct task_struct *p, struct pt_regs *pregs)
 {
 	struct pt_regs * cregs = &(p->thread.regs);
 	void *stack = task_stack_page(p);
@@ -270,48 +254,32 @@ copy_thread(unsigned long clone_flags, unsigned long usp,
 #ifdef CONFIG_HPUX
 	extern void * const hpux_child_return;
 #endif
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		memset(cregs, 0, sizeof(struct pt_regs));
+		if (!usp) /* idle thread */
+			return 0;
 
-	*cregs = *pregs;
-
-	/* Set the return value for the child.  Note that this is not
-           actually restored by the syscall exit path, but we put it
-           here for consistency in case of signals. */
-	cregs->gr[28] = 0; /* child */
-
-	/*
-	 * We need to differentiate between a user fork and a
-	 * kernel fork. We can't use user_mode, because the
-	 * the syscall path doesn't save iaoq. Right now
-	 * We rely on the fact that kernel_thread passes
-	 * in zero for usp.
-	 */
-	if (usp == 1) {
 		/* kernel thread */
-		cregs->ksp = (unsigned long)stack + THREAD_SZ_ALGN;
 		/* Must exit via ret_from_kernel_thread in order
 		 * to call schedule_tail()
 		 */
+		cregs->ksp = (unsigned long)stack + THREAD_SZ_ALGN + FRAME_SIZE;
 		cregs->kpc = (unsigned long) &ret_from_kernel_thread;
 		/*
 		 * Copy function and argument to be called from
 		 * ret_from_kernel_thread.
 		 */
 #ifdef CONFIG_64BIT
-		cregs->gr[27] = pregs->gr[27];
+		cregs->gr[27] = ((unsigned long *)usp)[3];
+		cregs->gr[26] = ((unsigned long *)usp)[2];
+#else
+		cregs->gr[26] = usp;
 #endif
-		cregs->gr[26] = pregs->gr[26];
-		cregs->gr[25] = pregs->gr[25];
+		cregs->gr[25] = arg;
 	} else {
 		/* user thread */
-		/*
-		 * Note that the fork wrappers are responsible
-		 * for setting gr[21].
-		 */
-
-		/* Use same stack depth as parent */
-		cregs->ksp = (unsigned long)stack
-			+ (pregs->gr[21] & (THREAD_SIZE - 1));
 		cregs->gr[30] = usp;
+		cregs->ksp = (unsigned long)stack + THREAD_SZ_ALGN + FRAME_SIZE;
 		if (personality(p->personality) == PER_HPUX) {
 #ifdef CONFIG_HPUX
 			cregs->kpc = (unsigned long) &hpux_child_return;
@@ -323,8 +291,7 @@ copy_thread(unsigned long clone_flags, unsigned long usp,
 		}
 		/* Setup thread TLS area from the 4th parameter in clone */
 		if (clone_flags & CLONE_SETTLS)
-		  cregs->cr27 = pregs->gr[23];
-	
+			cregs->cr27 = pregs->gr[23];
 	}
 
 	return 0;
@@ -335,39 +302,6 @@ unsigned long thread_saved_pc(struct task_struct *t)
 	return t->thread.regs.kpc;
 }
 
-/*
- * sys_execve() executes a new program.
- */
-
-asmlinkage int sys_execve(struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	filename = getname((const char __user *) regs->gr[26]);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = do_execve(filename->name,
-			  (const char __user *const __user *) regs->gr[25],
-			  (const char __user *const __user *) regs->gr[24],
-			  regs);
-	putname(filename);
-out:
-
-	return error;
-}
-
-extern int __execve(const char *filename,
-		    const char *const argv[],
-		    const char *const envp[], struct task_struct *task);
-int kernel_execve(const char *filename,
-		  const char *const argv[],
-		  const char *const envp[])
-{
-	return __execve(filename, argv, envp, current);
-}
-
 unsigned long
 get_wchan(struct task_struct *p)
 {

arch/parisc/kernel/sys_parisc32.c

@@ -53,28 +53,6 @@
 #define DBG(x)
 #endif
 
-/*
- * sys32_execve() executes a new program.
- */
-
-asmlinkage int sys32_execve(struct pt_regs *regs)
-{
-	int error;
-	struct filename *filename;
-
-	DBG(("sys32_execve(%p) r26 = 0x%lx\n", regs, regs->gr[26]));
-	filename = getname((const char __user *) regs->gr[26]);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = compat_do_execve(filename->name, compat_ptr(regs->gr[25]),
-				 compat_ptr(regs->gr[24]), regs);
-	putname(filename);
-out:
-
-	return error;
-}
-
 asmlinkage long sys32_unimplemented(int r26, int r25, int r24, int r23,
 	int r22, int r21, int r20)
 {

arch/parisc/kernel/syscall_table.S

@@ -66,7 +66,7 @@
 	ENTRY_SAME(creat)
 	ENTRY_SAME(link)
 	ENTRY_SAME(unlink)		/* 10 */
-	ENTRY_DIFF(execve_wrapper)
+	ENTRY_COMP(execve)
 	ENTRY_SAME(chdir)
 	/* See comments in kernel/time.c!!! Maybe we don't need this? */
 	ENTRY_COMP(time)

arch/powerpc/Kconfig

@@ -144,6 +144,7 @@ config PPC
 	select GENERIC_KERNEL_THREAD
 	select HAVE_MOD_ARCH_SPECIFIC
 	select MODULES_USE_ELF_RELA
+	select GENERIC_KERNEL_EXECVE
 
 config EARLY_PRINTK
 	bool