Inline the syscallenter() and syscallret(). This reduces the time measured

by the syscall entry speed microbenchmarks by ~10% on amd64.

Submitted by:	jhb
Approved by:	re (bz)
MFC after:	2 weeks

11 files changed, 229 insertions, 165 deletions

diff --git a/sys/amd64/amd64/trap.c b/sys/amd64/amd64/trap.c
index 4e5f8b8aa9f5..a36a6be13ce3 100644
--- a/sys/amd64/amd64/trap.c
+++ b/sys/amd64/amd64/trap.c
@@ -883,6 +883,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 /*
  *	syscall -	system call request C handler
  *
diff --git a/sys/amd64/ia32/ia32_syscall.c b/sys/amd64/ia32/ia32_syscall.c
index 0b46dae7a5e3..d79272a01dcf 100644
--- a/sys/amd64/ia32/ia32_syscall.c
+++ b/sys/amd64/ia32/ia32_syscall.c
@@ -163,6 +163,8 @@ ia32_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 void
 ia32_syscall(struct trapframe *frame)
 {
diff --git a/sys/i386/i386/trap.c b/sys/i386/i386/trap.c
index 5a8016c9c69d..dddecaa807ff 100644
--- a/sys/i386/i386/trap.c
+++ b/sys/i386/i386/trap.c
@@ -1054,6 +1054,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 /*
  *	syscall -	system call request C handler
  *
diff --git a/sys/ia64/ia32/ia32_trap.c b/sys/ia64/ia32/ia32_trap.c
index 57f479169560..6755358189a1 100644
--- a/sys/ia64/ia32/ia32_trap.c
+++ b/sys/ia64/ia32/ia32_trap.c
@@ -144,6 +144,8 @@ ia32_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 static void
 ia32_syscall(struct trapframe *tf)
 {
diff --git a/sys/ia64/ia64/trap.c b/sys/ia64/ia64/trap.c
index 4b84f769340e..d587b05868fd 100644
--- a/sys/ia64/ia64/trap.c
+++ b/sys/ia64/ia64/trap.c
@@ -929,6 +929,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (0);
 }
 
+#include "../../kern/subr_syscall.c"
+
 /*
  * Process a system call.
  *
diff --git a/sys/kern/subr_syscall.c b/sys/kern/subr_syscall.c
new file mode 100644
index 000000000000..cb0d9295ed2b
--- /dev/null
+++ b/sys/kern/subr_syscall.c
@@ -0,0 +1,213 @@
+/*-
+ * Copyright (C) 1994, David Greenman
+ * Copyright (c) 1990, 1993
+ *	The Regents of the University of California.  All rights reserved.
+ * Copyright (C) 2010 Konstantin Belousov <kib@freebsd.org>
+ *
+ * This code is derived from software contributed to Berkeley by
+ * the University of Utah, and William Jolitz.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *	This product includes software developed by the University of
+ *	California, Berkeley and its contributors.
+ * 4. Neither the name of the University nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ *	from: @(#)trap.c	7.4 (Berkeley) 5/13/91
+ */
+
+#include "opt_capsicum.h"
+#include "opt_ktrace.h"
+#include "opt_kdtrace.h"
+
+__FBSDID("$FreeBSD$");
+
+#include <sys/capability.h>
+#include <sys/ktr.h>
+#ifdef KTRACE
+#include <sys/uio.h>
+#include <sys/ktrace.h>
+#endif
+#include <security/audit/audit.h>
+
+static inline int
+syscallenter(struct thread *td, struct syscall_args *sa)
+{
+	struct proc *p;
+	int error, traced;
+
+	PCPU_INC(cnt.v_syscall);
+	p = td->td_proc;
+
+	td->td_pticks = 0;
+	if (td->td_ucred != p->p_ucred)
+		cred_update_thread(td);
+	if (p->p_flag & P_TRACED) {
+		traced = 1;
+		PROC_LOCK(p);
+		td->td_dbgflags &= ~TDB_USERWR;
+		td->td_dbgflags |= TDB_SCE;
+		PROC_UNLOCK(p);
+	} else
+		traced = 0;
+	error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
+#ifdef KTRACE
+	if (KTRPOINT(td, KTR_SYSCALL))
+		ktrsyscall(sa->code, sa->narg, sa->args);
+#endif
+
+	CTR6(KTR_SYSC,
+"syscall: td=%p pid %d %s (%#lx, %#lx, %#lx)",
+	    td, td->td_proc->p_pid, syscallname(p, sa->code),
+	    sa->args[0], sa->args[1], sa->args[2]);
+
+	if (error == 0) {
+		STOPEVENT(p, S_SCE, sa->narg);
+		PTRACESTOP_SC(p, td, S_PT_SCE);
+		if (td->td_dbgflags & TDB_USERWR) {
+			/*
+			 * Reread syscall number and arguments if
+			 * debugger modified registers or memory.
+			 */
+			error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
+#ifdef KTRACE
+			if (KTRPOINT(td, KTR_SYSCALL))
+				ktrsyscall(sa->code, sa->narg, sa->args);
+#endif
+			if (error != 0)
+				goto retval;
+		}
+
+#ifdef CAPABILITY_MODE
+		/*
+		 * In capability mode, we only allow access to system calls
+		 * flagged with SYF_CAPENABLED.
+		 */
+		if (IN_CAPABILITY_MODE(td) &&
+		    !(sa->callp->sy_flags & SYF_CAPENABLED)) {
+			error = ECAPMODE;
+			goto retval;
+		}
+#endif
+
+		error = syscall_thread_enter(td, sa->callp);
+		if (error != 0)
+			goto retval;
+
+#ifdef KDTRACE_HOOKS
+		/*
+		 * If the systrace module has registered it's probe
+		 * callback and if there is a probe active for the
+		 * syscall 'entry', process the probe.
+		 */
+		if (systrace_probe_func != NULL && sa->callp->sy_entry != 0)
+			(*systrace_probe_func)(sa->callp->sy_entry, sa->code,
+			    sa->callp, sa->args, 0);
+#endif
+
+		AUDIT_SYSCALL_ENTER(sa->code, td);
+		error = (sa->callp->sy_call)(td, sa->args);
+		AUDIT_SYSCALL_EXIT(error, td);
+
+		/* Save the latest error return value. */
+		td->td_errno = error;
+
+#ifdef KDTRACE_HOOKS
+		/*
+		 * If the systrace module has registered it's probe
+		 * callback and if there is a probe active for the
+		 * syscall 'return', process the probe.
+		 */
+		if (systrace_probe_func != NULL && sa->callp->sy_return != 0)
+			(*systrace_probe_func)(sa->callp->sy_return, sa->code,
+			    sa->callp, NULL, (error) ? -1 : td->td_retval[0]);
+#endif
+		syscall_thread_exit(td, sa->callp);
+		CTR4(KTR_SYSC, "syscall: p=%p error=%d return %#lx %#lx",
+		    p, error, td->td_retval[0], td->td_retval[1]);
+	}
+ retval:
+	if (traced) {
+		PROC_LOCK(p);
+		td->td_dbgflags &= ~TDB_SCE;
+		PROC_UNLOCK(p);
+	}
+	(p->p_sysent->sv_set_syscall_retval)(td, error);
+	return (error);
+}
+
+static inline void
+syscallret(struct thread *td, int error, struct syscall_args *sa __unused)
+{
+	struct proc *p;
+	int traced;
+
+	p = td->td_proc;
+
+	/*
+	 * Check for misbehavior.
+	 */
+	WITNESS_WARN(WARN_PANIC, NULL, "System call %s returning",
+	    syscallname(p, sa->code));
+	KASSERT(td->td_critnest == 0,
+	    ("System call %s returning in a critical section",
+	    syscallname(p, sa->code)));
+	KASSERT(td->td_locks == 0,
+	    ("System call %s returning with %d locks held",
+	     syscallname(p, sa->code), td->td_locks));
+
+	/*
+	 * Handle reschedule and other end-of-syscall issues
+	 */
+	userret(td, td->td_frame);
+
+	CTR4(KTR_SYSC, "syscall %s exit thread %p pid %d proc %s",
+	    syscallname(p, sa->code), td, td->td_proc->p_pid, td->td_name);
+
+#ifdef KTRACE
+	if (KTRPOINT(td, KTR_SYSRET))
+		ktrsysret(sa->code, error, td->td_retval[0]);
+#endif
+
+	if (p->p_flag & P_TRACED) {
+		traced = 1;
+		PROC_LOCK(p);
+		td->td_dbgflags |= TDB_SCX;
+		PROC_UNLOCK(p);
+	} else
+		traced = 0;
+	/*
+	 * This works because errno is findable through the
+	 * register set.  If we ever support an emulation where this
+	 * is not the case, this code will need to be revisited.
+	 */
+	STOPEVENT(p, S_SCX, sa->code);
+	PTRACESTOP_SC(p, td, S_PT_SCX);
+	if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) {
+		PROC_LOCK(p);
+		td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK);
+		PROC_UNLOCK(p);
+	}
+}
diff --git a/sys/kern/subr_trap.c b/sys/kern/subr_trap.c
index 3527ed1fbedf..42adf80f04fe 100644
--- a/sys/kern/subr_trap.c
+++ b/sys/kern/subr_trap.c
@@ -252,7 +252,6 @@ ast(struct trapframe *framep)
 	mtx_assert(&Giant, MA_NOTOWNED);
 }
 
-#ifdef HAVE_SYSCALL_ARGS_DEF
 const char *
 syscallname(struct proc *p, u_int code)
 {
@@ -264,164 +263,3 @@ syscallname(struct proc *p, u_int code)
 		return (unknown);
 	return (sv->sv_syscallnames[code]);
 }
-
-int
-syscallenter(struct thread *td, struct syscall_args *sa)
-{
-	struct proc *p;
-	int error, traced;
-
-	PCPU_INC(cnt.v_syscall);
-	p = td->td_proc;
-
-	td->td_pticks = 0;
-	if (td->td_ucred != p->p_ucred)
-		cred_update_thread(td);
-	if (p->p_flag & P_TRACED) {
-		traced = 1;
-		PROC_LOCK(p);
-		td->td_dbgflags &= ~TDB_USERWR;
-		td->td_dbgflags |= TDB_SCE;
-		PROC_UNLOCK(p);
-	} else
-		traced = 0;
-	error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
-#ifdef KTRACE
-	if (KTRPOINT(td, KTR_SYSCALL))
-		ktrsyscall(sa->code, sa->narg, sa->args);
-#endif
-
-	CTR6(KTR_SYSC,
-"syscall: td=%p pid %d %s (%#lx, %#lx, %#lx)",
-	    td, td->td_proc->p_pid, syscallname(p, sa->code),
-	    sa->args[0], sa->args[1], sa->args[2]);
-
-	if (error == 0) {
-		STOPEVENT(p, S_SCE, sa->narg);
-		PTRACESTOP_SC(p, td, S_PT_SCE);
-		if (td->td_dbgflags & TDB_USERWR) {
-			/*
-			 * Reread syscall number and arguments if
-			 * debugger modified registers or memory.
-			 */
-			error = (p->p_sysent->sv_fetch_syscall_args)(td, sa);
-#ifdef KTRACE
-			if (KTRPOINT(td, KTR_SYSCALL))
-				ktrsyscall(sa->code, sa->narg, sa->args);
-#endif
-			if (error != 0)
-				goto retval;
-		}
-
-#ifdef CAPABILITY_MODE
-		/*
-		 * In capability mode, we only allow access to system calls
-		 * flagged with SYF_CAPENABLED.
-		 */
-		if (IN_CAPABILITY_MODE(td) &&
-		    !(sa->callp->sy_flags & SYF_CAPENABLED)) {
-			error = ECAPMODE;
-			goto retval;
-		}
-#endif
-
-		error = syscall_thread_enter(td, sa->callp);
-		if (error != 0)
-			goto retval;
-
-#ifdef KDTRACE_HOOKS
-		/*
-		 * If the systrace module has registered it's probe
-		 * callback and if there is a probe active for the
-		 * syscall 'entry', process the probe.
-		 */
-		if (systrace_probe_func != NULL && sa->callp->sy_entry != 0)
-			(*systrace_probe_func)(sa->callp->sy_entry, sa->code,
-			    sa->callp, sa->args, 0);
-#endif
-
-		AUDIT_SYSCALL_ENTER(sa->code, td);
-		error = (sa->callp->sy_call)(td, sa->args);
-		AUDIT_SYSCALL_EXIT(error, td);
-
-		/* Save the latest error return value. */
-		td->td_errno = error;
-
-#ifdef KDTRACE_HOOKS
-		/*
-		 * If the systrace module has registered it's probe
-		 * callback and if there is a probe active for the
-		 * syscall 'return', process the probe.
-		 */
-		if (systrace_probe_func != NULL && sa->callp->sy_return != 0)
-			(*systrace_probe_func)(sa->callp->sy_return, sa->code,
-			    sa->callp, NULL, (error) ? -1 : td->td_retval[0]);
-#endif
-		syscall_thread_exit(td, sa->callp);
-		CTR4(KTR_SYSC, "syscall: p=%p error=%d return %#lx %#lx",
-		    p, error, td->td_retval[0], td->td_retval[1]);
-	}
- retval:
-	if (traced) {
-		PROC_LOCK(p);
-		td->td_dbgflags &= ~TDB_SCE;
-		PROC_UNLOCK(p);
-	}
-	(p->p_sysent->sv_set_syscall_retval)(td, error);
-	return (error);
-}
-
-void
-syscallret(struct thread *td, int error, struct syscall_args *sa __unused)
-{
-	struct proc *p;
-	int traced;
-
-	p = td->td_proc;
-
-	/*
-	 * Check for misbehavior.
-	 */
-	WITNESS_WARN(WARN_PANIC, NULL, "System call %s returning",
-	    syscallname(p, sa->code));
-	KASSERT(td->td_critnest == 0,
-	    ("System call %s returning in a critical section",
-	    syscallname(p, sa->code)));
-	KASSERT(td->td_locks == 0,
-	    ("System call %s returning with %d locks held",
-	     syscallname(p, sa->code), td->td_locks));
-
-	/*
-	 * Handle reschedule and other end-of-syscall issues
-	 */
-	userret(td, td->td_frame);
-
-	CTR4(KTR_SYSC, "syscall %s exit thread %p pid %d proc %s",
-	    syscallname(p, sa->code), td, td->td_proc->p_pid, td->td_name);
-
-#ifdef KTRACE
-	if (KTRPOINT(td, KTR_SYSRET))
-		ktrsysret(sa->code, error, td->td_retval[0]);
-#endif
-
-	if (p->p_flag & P_TRACED) {
-		traced = 1;
-		PROC_LOCK(p);
-		td->td_dbgflags |= TDB_SCX;
-		PROC_UNLOCK(p);
-	} else
-		traced = 0;
-	/*
-	 * This works because errno is findable through the
-	 * register set.  If we ever support an emulation where this
-	 * is not the case, this code will need to be revisited.
-	 */
-	STOPEVENT(p, S_SCX, sa->code);
-	PTRACESTOP_SC(p, td, S_PT_SCX);
-	if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) {
-		PROC_LOCK(p);
-		td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK);
-		PROC_UNLOCK(p);
-	}
-}
-#endif /* HAVE_SYSCALL_ARGS_DEF */
diff --git a/sys/powerpc/aim/trap.c b/sys/powerpc/aim/trap.c
index 51dee0eabdb5..93feb515728d 100644
--- a/sys/powerpc/aim/trap.c
+++ b/sys/powerpc/aim/trap.c
@@ -445,6 +445,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 void
 syscall(struct trapframe *frame)
 {
diff --git a/sys/powerpc/booke/trap.c b/sys/powerpc/booke/trap.c
index fc506bd361ce..091fce08f1fb 100644
--- a/sys/powerpc/booke/trap.c
+++ b/sys/powerpc/booke/trap.c
@@ -375,6 +375,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 void
 syscall(struct trapframe *frame)
 {
diff --git a/sys/sparc64/sparc64/trap.c b/sys/sparc64/sparc64/trap.c
index db05f06ba926..7037e462ad0e 100644
--- a/sys/sparc64/sparc64/trap.c
+++ b/sys/sparc64/sparc64/trap.c
@@ -606,6 +606,8 @@ cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
 	return (error);
 }
 
+#include "../../kern/subr_syscall.c"
+
 /*
  * Syscall handler
  * The arguments to the syscall are passed in the out registers by the caller,
diff --git a/sys/sys/proc.h b/sys/sys/proc.h
index 67adbe5b5899..fb9791303ddc 100644
--- a/sys/sys/proc.h
+++ b/sys/sys/proc.h
@@ -873,9 +873,6 @@ void	cpu_switch(struct thread *, struct thread *, struct mtx *);
 void	cpu_throw(struct thread *, struct thread *) __dead2;
 void	unsleep(struct thread *);
 void	userret(struct thread *, struct trapframe *);
-struct syscall_args;
-int	syscallenter(struct thread *, struct syscall_args *);
-void	syscallret(struct thread *, int, struct syscall_args *);
 
 void	cpu_exit(struct thread *);
 void	exit1(struct thread *, int) __dead2;