sh: Make syscall tracer use tracehook notifiers, add TIF_NOTIFY_RESUME.

This follows the changes in commits:

7d6d637dac2050f30a1b57b0a3dc5de4a10616ba
4f72c4279eab1e5f3ed1ac4e55d4527617582392

on powerpc. Adding in TIF_NOTIFY_RESUME, and cleaning up the syscall
tracing to be more generic. This is an incremental step to turning
on tracehook, as well as unifying more of the ptrace and signal code
across the 32/64 split.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

arch/sh/include/asm/thread_info.h

@@ -119,10 +119,11 @@ static inline struct thread_info *current_thread_info(void)
 #define TIF_SINGLESTEP		4	/* singlestepping active */
 #define TIF_SYSCALL_AUDIT	5	/* syscall auditing active */
 #define TIF_SECCOMP		6	/* secure computing */
+#define TIF_NOTIFY_RESUME	7	/* callback before returning to user */
 #define TIF_USEDFPU		16	/* FPU was used by this task this quantum (SMP) */
 #define TIF_POLLING_NRFLAG	17	/* true if poll_idle() is polling TIF_NEED_RESCHED */
 #define TIF_MEMDIE		18
-#define TIF_FREEZE		19
+#define TIF_FREEZE		19	/* Freezing for suspend */
 
 #define _TIF_SYSCALL_TRACE	(1 << TIF_SYSCALL_TRACE)
 #define _TIF_SIGPENDING		(1 << TIF_SIGPENDING)
@@ -131,6 +132,7 @@ static inline struct thread_info *current_thread_info(void)
 #define _TIF_SINGLESTEP		(1 << TIF_SINGLESTEP)
 #define _TIF_SYSCALL_AUDIT	(1 << TIF_SYSCALL_AUDIT)
 #define _TIF_SECCOMP		(1 << TIF_SECCOMP)
+#define _TIF_NOTIFY_RESUME	(1 << TIF_NOTIFY_RESUME)
 #define _TIF_USEDFPU		(1 << TIF_USEDFPU)
 #define _TIF_POLLING_NRFLAG	(1 << TIF_POLLING_NRFLAG)
 #define _TIF_FREEZE		(1 << TIF_FREEZE)
@@ -146,9 +148,10 @@ static inline struct thread_info *current_thread_info(void)
 				 _TIF_SYSCALL_AUDIT | _TIF_SECCOMP)
 
 /* work to do on any return to u-space */
-#define _TIF_ALLWORK_MASK	(_TIF_SYSCALL_TRACE | _TIF_SIGPENDING    | \
-				 _TIF_NEED_RESCHED  | _TIF_SYSCALL_AUDIT | \
-				 _TIF_SINGLESTEP    | _TIF_RESTORE_SIGMASK)
+#define _TIF_ALLWORK_MASK	(_TIF_SYSCALL_TRACE | _TIF_SIGPENDING      | \
+				 _TIF_NEED_RESCHED  | _TIF_SYSCALL_AUDIT   | \
+				 _TIF_SINGLESTEP    | _TIF_RESTORE_SIGMASK | \
+				 _TIF_NOTIFY_RESUME)
 
 /* work to do on interrupt/exception return */
 #define _TIF_WORK_MASK		(_TIF_ALLWORK_MASK & ~(_TIF_SYSCALL_TRACE | \

arch/sh/kernel/cpu/sh5/entry.S

@@ -987,11 +987,11 @@ work_resched:
 work_notifysig:
 	gettr	tr1, LINK
 
-	movi	do_signal, r6
+	movi	do_notify_resume, r6
 	ptabs	r6, tr0
 	or	SP, ZERO, r2
-	or	ZERO, ZERO, r3
-	blink	tr0, LINK	    /* Call do_signal(regs, 0), return here */
+	or	r7, ZERO, r3
+	blink	tr0, LINK	    /* Call do_notify_resume(regs, current_thread_info->flags), return here */
 
 restore_all:
 	/* Do prefetches */
@@ -1305,13 +1305,15 @@ syscall_allowed:
 	beq/l	r6, ZERO, tr0
 
 	/* Trace it by calling syscall_trace before and after */
-	movi	syscall_trace, r4
+	movi	do_syscall_trace_enter, r4
 	or	SP, ZERO, r2
-	or	ZERO, ZERO, r3
 	ptabs	r4, tr0
 	blink	tr0, LINK
 
-	/* Reload syscall number as r5 is trashed by syscall_trace */
+	/* Save the retval */
+	st.q	SP, FRAME_R(2), r2
+
+	/* Reload syscall number as r5 is trashed by do_syscall_trace_enter */
 	ld.q	SP, FRAME_S(FSYSCALL_ID), r5
 	andi	r5, 0x1ff, r5
 
@@ -1343,9 +1345,8 @@ syscall_ret_trace:
 	/* We get back here only if under trace */
 	st.q	SP, FRAME_R(9), r2	/* Save return value */
 
-	movi	syscall_trace, LINK
+	movi	do_syscall_trace_leave, LINK
 	or	SP, ZERO, r2
-	movi	1, r3
 	ptabs	LINK, tr0
 	blink	tr0, LINK
 

arch/sh/kernel/entry-common.S

@@ -211,10 +211,8 @@ syscall_exit_work:
 	 nop
 #endif
 	sti
-	! XXX setup arguments...
 	mov	r15, r4
-	mov	#1, r5
-	mov.l	4f, r0			! do_syscall_trace
+	mov.l	8f, r0			! do_syscall_trace_leave
 	jsr	@r0
 	 nop
 	bra	resume_userspace
@@ -223,12 +221,11 @@ syscall_exit_work:
 	.align	2
 syscall_trace_entry:
 	!                     	Yes it is traced.
-	! XXX setup arguments...
 	mov     r15, r4
-	mov     #0, r5
-	mov.l	4f, r11		! Call do_syscall_trace which notifies
+	mov.l	7f, r11		! Call do_syscall_trace_enter which notifies
 	jsr	@r11	    	! superior (will chomp R[0-7])
 	 nop
+	mov.l	r0, @(OFF_R0,r15)	! Save return value
 	!			Reload R0-R4 from kernel stack, where the
 	!   	    	    	parent may have modified them using
 	!   	    	    	ptrace(POKEUSR).  (Note that R0-R2 are
@@ -389,8 +386,9 @@ syscall_exit:
 #endif
 2:	.long	NR_syscalls
 3:	.long	sys_call_table
-4:	.long	do_syscall_trace
 #ifdef CONFIG_TRACE_IRQFLAGS
 5:	.long	trace_hardirqs_on
 6:	.long	trace_hardirqs_off
 #endif
+7:	.long	do_syscall_trace_enter
+8:	.long	do_syscall_trace_leave

arch/sh/kernel/ptrace_32.c

@@ -21,6 +21,7 @@
 #include <linux/io.h>
 #include <linux/audit.h>
 #include <linux/seccomp.h>
+#include <linux/tracehook.h>
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/system.h>
@@ -216,41 +217,38 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
 	return ret;
 }
 
-asmlinkage void do_syscall_trace(struct pt_regs *regs, int entryexit)
+asmlinkage long do_syscall_trace_enter(struct pt_regs *regs)
 {
-	struct task_struct *tsk = current;
+	long ret = 0;
 
 	secure_computing(regs->regs[0]);
 
-	if (unlikely(current->audit_context) && entryexit)
-		audit_syscall_exit(AUDITSC_RESULT(regs->regs[0]),
-				   regs->regs[0]);
-
-	if (!test_thread_flag(TIF_SYSCALL_TRACE) &&
-	    !test_thread_flag(TIF_SINGLESTEP))
-		goto out;
-	if (!(tsk->ptrace & PT_PTRACED))
-		goto out;
-
-	/* the 0x80 provides a way for the tracing parent to distinguish
-	   between a syscall stop and SIGTRAP delivery */
-	ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD) &&
-				 !test_thread_flag(TIF_SINGLESTEP) ? 0x80 : 0));
-
-	/*
-	 * this isn't the same as continuing with a signal, but it will do
-	 * for normal use.  strace only continues with a signal if the
-	 * stopping signal is not SIGTRAP.  -brl
-	 */
-	if (tsk->exit_code) {
-		send_sig(tsk->exit_code, tsk, 1);
-		tsk->exit_code = 0;
-	}
+	if (test_thread_flag(TIF_SYSCALL_TRACE) &&
+	    tracehook_report_syscall_entry(regs))
+		/*
+		 * Tracing decided this syscall should not happen.
+		 * We'll return a bogus call number to get an ENOSYS
+		 * error, but leave the original number in regs->regs[0].
+		 */
+		ret = -1L;
 
-out:
-	if (unlikely(current->audit_context) && !entryexit)
+	if (unlikely(current->audit_context))
 		audit_syscall_entry(AUDIT_ARCH_SH, regs->regs[3],
 				    regs->regs[4], regs->regs[5],
 				    regs->regs[6], regs->regs[7]);
 
+	return ret ?: regs->regs[0];
+}
+
+asmlinkage void do_syscall_trace_leave(struct pt_regs *regs)
+{
+	int step;
+
+	if (unlikely(current->audit_context))
+		audit_syscall_exit(AUDITSC_RESULT(regs->regs[0]),
+				   regs->regs[0]);
+
+	step = test_thread_flag(TIF_SINGLESTEP);
+	if (step || test_thread_flag(TIF_SYSCALL_TRACE))
+		tracehook_report_syscall_exit(regs, step);
 }

arch/sh/kernel/ptrace_64.c

@@ -28,6 +28,7 @@
 #include <linux/syscalls.h>
 #include <linux/audit.h>
 #include <linux/seccomp.h>
+#include <linux/tracehook.h>
 #include <asm/io.h>
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
@@ -221,40 +222,37 @@ asmlinkage int sh64_ptrace(long request, long pid, long addr, long data)
 	return sys_ptrace(request, pid, addr, data);
 }
 
-asmlinkage void syscall_trace(struct pt_regs *regs, int entryexit)
+asmlinkage long long do_syscall_trace_enter(struct pt_regs *regs)
 {
-	struct task_struct *tsk = current;
+	long long ret = 0;
 
 	secure_computing(regs->regs[9]);
 
-	if (unlikely(current->audit_context) && entryexit)
-		audit_syscall_exit(AUDITSC_RESULT(regs->regs[9]),
-				   regs->regs[9]);
-
-	if (!test_thread_flag(TIF_SYSCALL_TRACE) &&
-	    !test_thread_flag(TIF_SINGLESTEP))
-		goto out;
-	if (!(tsk->ptrace & PT_PTRACED))
-		goto out;
-
-	ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD) &&
-				!test_thread_flag(TIF_SINGLESTEP) ? 0x80 : 0));
-
-	/*
-	 * this isn't the same as continuing with a signal, but it will do
-	 * for normal use.  strace only continues with a signal if the
-	 * stopping signal is not SIGTRAP.  -brl
-	 */
-	if (tsk->exit_code) {
-		send_sig(tsk->exit_code, tsk, 1);
-		tsk->exit_code = 0;
-	}
+	if (test_thread_flag(TIF_SYSCALL_TRACE) &&
+	    tracehook_report_syscall_entry(regs))
+		/*
+		 * Tracing decided this syscall should not happen.
+		 * We'll return a bogus call number to get an ENOSYS
+		 * error, but leave the original number in regs->regs[0].
+		 */
+		ret = -1LL;
 
-out:
-	if (unlikely(current->audit_context) && !entryexit)
+	if (unlikely(current->audit_context))
 		audit_syscall_entry(AUDIT_ARCH_SH, regs->regs[1],
 				    regs->regs[2], regs->regs[3],
 				    regs->regs[4], regs->regs[5]);
+
+	return ret ?: regs->regs[9];
+}
+
+asmlinkage void do_syscall_trace_leave(struct pt_regs *regs)
+{
+	if (unlikely(current->audit_context))
+		audit_syscall_exit(AUDITSC_RESULT(regs->regs[9]),
+				   regs->regs[9]);
+
+	if (test_thread_flag(TIF_SYSCALL_TRACE))
+		tracehook_report_syscall_exit(regs, 0);
 }
 
 /* Called with interrupts disabled */

arch/sh/kernel/signal_32.c

@@ -24,6 +24,7 @@
 #include <linux/binfmts.h>
 #include <linux/freezer.h>
 #include <linux/io.h>
+#include <linux/tracehook.h>
 #include <asm/system.h>
 #include <asm/ucontext.h>
 #include <asm/uaccess.h>
@@ -507,14 +508,13 @@ handle_signal(unsigned long sig, struct k_sigaction *ka, siginfo_t *info,
 		switch (regs->regs[0]) {
 			case -ERESTART_RESTARTBLOCK:
 			case -ERESTARTNOHAND:
+			no_system_call_restart:
 				regs->regs[0] = -EINTR;
 				break;
 
 			case -ERESTARTSYS:
-				if (!(ka->sa.sa_flags & SA_RESTART)) {
-					regs->regs[0] = -EINTR;
-					break;
-				}
+				if (!(ka->sa.sa_flags & SA_RESTART))
+					goto no_system_call_restart;
 			/* fallthrough */
 			case -ERESTARTNOINTR:
 				regs->regs[0] = save_r0;
@@ -589,12 +589,15 @@ static void do_signal(struct pt_regs *regs, unsigned int save_r0)
 			 * clear the TIF_RESTORE_SIGMASK flag */
 			if (test_thread_flag(TIF_RESTORE_SIGMASK))
 				clear_thread_flag(TIF_RESTORE_SIGMASK);
+
+			tracehook_signal_handler(signr, &info, &ka, regs,
+					test_thread_flag(TIF_SINGLESTEP));
 		}
 
 		return;
 	}
 
- no_signal:
+no_signal:
 	/* Did we come from a system call? */
 	if (regs->tra >= 0) {
 		/* Restart the system call - no handlers present */
@@ -618,9 +621,14 @@ static void do_signal(struct pt_regs *regs, unsigned int save_r0)
 }
 
 asmlinkage void do_notify_resume(struct pt_regs *regs, unsigned int save_r0,
-				 __u32 thread_info_flags)
+				 unsigned long thread_info_flags)
 {
 	/* deal with pending signal delivery */
-	if (thread_info_flags & (_TIF_SIGPENDING | _TIF_RESTORE_SIGMASK))
+	if (thread_info_flags & _TIF_SIGPENDING)
 		do_signal(regs, save_r0);
+
+	if (thread_info_flags & _TIF_NOTIFY_RESUME) {
+		clear_thread_flag(TIF_NOTIFY_RESUME);
+		tracehook_notify_resume(regs);
+	}
 }

arch/sh/kernel/signal_64.c

@@ -22,6 +22,7 @@
 #include <linux/ptrace.h>
 #include <linux/unistd.h>
 #include <linux/stddef.h>
+#include <linux/tracehook.h>
 #include <asm/ucontext.h>
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
@@ -42,7 +43,84 @@
 
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
-asmlinkage int do_signal(struct pt_regs *regs, sigset_t *oldset);
+/*
+ * Note that 'init' is a special process: it doesn't get signals it doesn't
+ * want to handle. Thus you cannot kill init even with a SIGKILL even by
+ * mistake.
+ *
+ * Note that we go through the signals twice: once to check the signals that
+ * the kernel can handle, and then we build all the user-level signal handling
+ * stack-frames in one go after that.
+ */
+static int do_signal(struct pt_regs *regs, sigset_t *oldset)
+{
+	siginfo_t info;
+	int signr;
+	struct k_sigaction ka;
+
+	/*
+	 * We want the common case to go fast, which
+	 * is why we may in certain cases get here from
+	 * kernel mode. Just return without doing anything
+	 * if so.
+	 */
+	if (!user_mode(regs))
+		return 1;
+
+	if (try_to_freeze())
+		goto no_signal;
+
+	if (test_thread_flag(TIF_RESTORE_SIGMASK))
+		oldset = &current->saved_sigmask;
+	else if (!oldset)
+		oldset = &current->blocked;
+
+	signr = get_signal_to_deliver(&info, &ka, regs, 0);
+
+	if (signr > 0) {
+		/* Whee!  Actually deliver the signal.  */
+		handle_signal(signr, &info, &ka, oldset, regs);
+
+		/*
+		 * If a signal was successfully delivered, the saved sigmask
+		 * is in its frame, and we can clear the TIF_RESTORE_SIGMASK
+		 * flag.
+		 */
+		if (test_thread_flag(TIF_RESTORE_SIGMASK))
+			clear_thread_flag(TIF_RESTORE_SIGMASK);
+
+		tracehook_signal_handler(signr, &info, &ka, regs, 0);
+		return 1;
+	}
+
+no_signal:
+	/* Did we come from a system call? */
+	if (regs->syscall_nr >= 0) {
+		/* Restart the system call - no handlers present */
+		switch (regs->regs[REG_RET]) {
+		case -ERESTARTNOHAND:
+		case -ERESTARTSYS:
+		case -ERESTARTNOINTR:
+			/* Decode Syscall # */
+			regs->regs[REG_RET] = regs->syscall_nr;
+			regs->pc -= 4;
+			break;
+
+		case -ERESTART_RESTARTBLOCK:
+			regs->regs[REG_RET] = __NR_restart_syscall;
+			regs->pc -= 4;
+			break;
+		}
+	}
+
+	/* No signal to deliver -- put the saved sigmask back */
+	if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
+		clear_thread_flag(TIF_RESTORE_SIGMASK);
+		sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
+	}
+
+	return 0;
+}
 
 /*
  * Atomically swap in the new signal mask, and wait for a signal.
@@ -643,14 +721,13 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka,
 		switch (regs->regs[REG_RET]) {
 			case -ERESTART_RESTARTBLOCK:
 			case -ERESTARTNOHAND:
+			no_system_call_restart:
 				regs->regs[REG_RET] = -EINTR;
 				break;
 
 			case -ERESTARTSYS:
-				if (!(ka->sa.sa_flags & SA_RESTART)) {
-					regs->regs[REG_RET] = -EINTR;
-					break;
-				}
+				if (!(ka->sa.sa_flags & SA_RESTART))
+					goto no_system_call_restart;
 			/* fallthrough */
 			case -ERESTARTNOINTR:
 				/* Decode syscall # */
@@ -673,80 +750,13 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka,
 	spin_unlock_irq(&current->sighand->siglock);
 }
 
-/*
- * Note that 'init' is a special process: it doesn't get signals it doesn't
- * want to handle. Thus you cannot kill init even with a SIGKILL even by
- * mistake.
- *
- * Note that we go through the signals twice: once to check the signals that
- * the kernel can handle, and then we build all the user-level signal handling
- * stack-frames in one go after that.
- */
-int do_signal(struct pt_regs *regs, sigset_t *oldset)
+asmlinkage void do_notify_resume(struct pt_regs *regs, unsigned long thread_info_flags)
 {
-	siginfo_t info;
-	int signr;
-	struct k_sigaction ka;
-
-	/*
-	 * We want the common case to go fast, which
-	 * is why we may in certain cases get here from
-	 * kernel mode. Just return without doing anything
-	 * if so.
-	 */
-	if (!user_mode(regs))
-		return 1;
-
-	if (try_to_freeze())
-		goto no_signal;
-
-	if (test_thread_flag(TIF_RESTORE_SIGMASK))
-		oldset = &current->saved_sigmask;
-	else if (!oldset)
-		oldset = &current->blocked;
-
-	signr = get_signal_to_deliver(&info, &ka, regs, 0);
-
-	if (signr > 0) {
-		/* Whee!  Actually deliver the signal.  */
-		handle_signal(signr, &info, &ka, oldset, regs);
+	if (thread_info_flags & _TIF_SIGPENDING)
+		do_signal(regs, 0);
 
-		/*
-		 * If a signal was successfully delivered, the saved sigmask
-		 * is in its frame, and we can clear the TIF_RESTORE_SIGMASK
-		 * flag.
-		 */
-		if (test_thread_flag(TIF_RESTORE_SIGMASK))
-			clear_thread_flag(TIF_RESTORE_SIGMASK);
-
-		return 1;
+	if (thread_info_flags & _TIF_NOTIFY_RESUME) {
+		clear_thread_flag(TIF_NOTIFY_RESUME);
+		tracehook_notify_resume(regs);
 	}
-
-no_signal:
-	/* Did we come from a system call? */
-	if (regs->syscall_nr >= 0) {
-		/* Restart the system call - no handlers present */
-		switch (regs->regs[REG_RET]) {
-		case -ERESTARTNOHAND:
-		case -ERESTARTSYS:
-		case -ERESTARTNOINTR:
-			/* Decode Syscall # */
-			regs->regs[REG_RET] = regs->syscall_nr;
-			regs->pc -= 4;
-			break;
-
-		case -ERESTART_RESTARTBLOCK:
-			regs->regs[REG_RET] = __NR_restart_syscall;
-			regs->pc -= 4;
-			break;
-		}
-	}
-
-	/* No signal to deliver -- put the saved sigmask back */
-	if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
-		clear_thread_flag(TIF_RESTORE_SIGMASK);
-		sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
-	}
-
-	return 0;
 }