Refactor page fault/error functions, normalizing their operation relative

to other architectures.  Kernel lock only the minimum (in preparation for
further unlocking down the line)
ok kettenis

1 files changed, 154 insertions, 200 deletions

diff --git a/sys/arch/sparc64/sparc64/trap.c b/sys/arch/sparc64/sparc64/trap.c
index 7cc15c50792..2b9e05d0d2f 100644
--- a/sys/arch/sparc64/sparc64/trap.c
+++ b/sys/arch/sparc64/sparc64/trap.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: trap.c,v 1.106 2020/10/21 17:54:33 deraadt Exp $	*/
+/*	$OpenBSD: trap.c,v 1.107 2020/10/23 16:54:35 deraadt Exp $	*/
 /*	$NetBSD: trap.c,v 1.73 2001/08/09 01:03:01 eeh Exp $ */
 
 /*
@@ -718,6 +718,21 @@ pmap_unuse_final(struct proc *p)
 	p->p_addr->u_pcb.pcb_nsaved = 0;
 }
 
+static inline int
+accesstype(unsigned int type, u_long sfsr)
+{
+	/*
+	 * If it was a FAST_DATA_ACCESS_MMU_MISS we have no idea what the
+	 * access was since the SFSR is not set.  But we should never get
+	 * here from there.
+	 */
+	if (type == T_FDMMU_MISS || (sfsr & SFSR_FV) == 0)
+		return PROT_READ;
+	else if (sfsr & SFSR_W)
+		return PROT_READ | PROT_WRITE;
+	return PROT_READ;
+}
+
 /*
  * This routine handles MMU generated faults.  About half
  * of them could be recoverable through uvm_fault.
@@ -726,40 +741,18 @@ void
 data_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
     vaddr_t addr, vaddr_t sfva, u_long sfsr)
 {
-	u_int64_t tstate;
-	struct proc *p;
-	struct vmspace *vm;
-	vaddr_t va;
-	int rv;
-	vm_prot_t access_type;
+	struct proc *p = curproc;
+	vaddr_t va = trunc_page(addr);
+	vm_prot_t access_type = accesstype(type, sfsr);
 	vaddr_t onfault;
 	union sigval sv;
+	int signal, sicode, error;
 
 	uvmexp.traps++;
-	if ((p = curproc) == NULL)	/* safety check */
+	if (p == NULL)		/* safety check */
 		p = &proc0;
 
-	tstate = tf->tf_tstate;
-
-	/* Find the faulting va to give to uvm_fault */
-	va = trunc_page(addr);
-
-	/*
-	 * Now munch on protections.
-	 *
-	 * If it was a FAST_DATA_ACCESS_MMU_MISS we have no idea what the
-	 * access was since the SFSR is not set.  But we should never get
-	 * here from there.
-	 */
-	if (type == T_FDMMU_MISS || (sfsr & SFSR_FV) == 0) {
-		/* Punt */
-		access_type = PROT_READ;
-	} else {
-		access_type = (sfsr & SFSR_W) ? PROT_READ | PROT_WRITE
-			: PROT_READ;
-	}
-	if (tstate & TSTATE_PRIV) {
-		KERNEL_LOCK();
+	if (tf->tf_tstate & TSTATE_PRIV) {
 #ifdef DDB
 		extern char Lfsprobe[];
 		/*
@@ -779,11 +772,13 @@ data_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 			goto kfault;
 		if (!(addr & TLB_TAG_ACCESS_CTX)) {
 			/* CTXT == NUCLEUS */
-			rv = uvm_fault(kernel_map, va, 0, access_type);
-			if (rv == 0) {
-				KERNEL_UNLOCK();
+
+			KERNEL_LOCK();
+			error = uvm_fault(kernel_map, va, 0, access_type);
+			KERNEL_UNLOCK();
+
+			if (error == 0)
 				return;
-			}
 			goto kfault;
 		}
 	} else {
@@ -793,15 +788,13 @@ data_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 		    "[%s]%d/%d sp=%lx inside %lx-%lx: not MAP_STACK\n",
 		    uvm_map_inentry_sp, p->p_vmspace->vm_map.sserial))
 			goto out;
-
-		KERNEL_LOCK();
 	}
 
-	vm = p->p_vmspace;
-	/* alas! must call the horrible vm code */
 	onfault = (vaddr_t)p->p_addr->u_pcb.pcb_onfault;
 	p->p_addr->u_pcb.pcb_onfault = NULL;
-	rv = uvm_fault(&vm->vm_map, (vaddr_t)va, 0, access_type);
+	KERNEL_LOCK();
+	error = uvm_fault(&p->p_vmspace->vm_map, (vaddr_t)va, 0, access_type);
+	KERNEL_UNLOCK();
 	p->p_addr->u_pcb.pcb_onfault = (void *)onfault;
 
 	/*
@@ -811,58 +804,52 @@ data_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 	 * the current limit and we need to reflect that as an access
 	 * error.
 	 */
-	if (rv == 0)
+	if (error == 0) {
 		uvm_grow(p, va);
+		goto out;
+	}
 
-	if (rv != 0) {
-		/*
-		 * Pagein failed.  If doing copyin/out, return to onfault
-		 * address.  Any other page fault in kernel, die; if user
-		 * fault, deliver SIGSEGV.
-		 */
-		int signal, sicode;
-
-		if (tstate & TSTATE_PRIV) {
+	/*
+	 * Pagein failed.  If doing copyin/out, return to onfault
+	 * address.  Any other page fault in kernel, die; if user
+	 * fault, deliver SIGSEGV.
+	 */
+	if (tf->tf_tstate & TSTATE_PRIV) {
 kfault:
-			onfault = (long)p->p_addr->u_pcb.pcb_onfault;
-			if (!onfault) {
-				(void) splhigh();
-				panic("kernel data fault: pc=%lx addr=%lx",
-				    pc, addr);
-				/* NOTREACHED */
-			}
-			tf->tf_pc = onfault;
-			tf->tf_npc = onfault + 4;
-			KERNEL_UNLOCK();
-			return;
-		}
-
-		if (type == T_FDMMU_MISS || (sfsr & SFSR_FV) == 0)
-			sv.sival_ptr = (void *)va;
-		else
-			sv.sival_ptr = (void *)sfva;
-
-		signal = SIGSEGV;
-		sicode = SEGV_MAPERR;
-		if (rv == ENOMEM) {
-			printf("UVM: pid %d (%s), uid %d killed: out of swap\n",
-			    p->p_p->ps_pid, p->p_p->ps_comm,
-			    p->p_ucred ? (int)p->p_ucred->cr_uid : -1);
-			signal = SIGKILL;
-		}
-		if (rv == EACCES)
-			sicode = SEGV_ACCERR;
-		if (rv == EIO) {
-			signal = SIGBUS;
-			sicode = BUS_OBJERR;
+		onfault = (long)p->p_addr->u_pcb.pcb_onfault;
+		if (!onfault) {
+			(void) splhigh();
+			panic("kernel data fault: pc=%lx addr=%lx",
+			    pc, addr);
+			/* NOTREACHED */
 		}
-		trapsignal(p, signal, access_type, sicode, sv);
+		tf->tf_pc = onfault;
+		tf->tf_npc = onfault + 4;
+		return;
 	}
 
-	KERNEL_UNLOCK();
+	if (type == T_FDMMU_MISS || (sfsr & SFSR_FV) == 0)
+		sv.sival_ptr = (void *)va;
+	else
+		sv.sival_ptr = (void *)sfva;
+
+	signal = SIGSEGV;
+	sicode = SEGV_MAPERR;
+	if (error == ENOMEM) {
+		printf("UVM: pid %d (%s), uid %d killed: out of swap\n",
+		    p->p_p->ps_pid, p->p_p->ps_comm,
+		    p->p_ucred ? (int)p->p_ucred->cr_uid : -1);
+		signal = SIGKILL;
+	} else if (error == EACCES)
+		sicode = SEGV_ACCERR;
+	else if (error == EIO) {
+		signal = SIGBUS;
+		sicode = BUS_OBJERR;
+	}
+	trapsignal(p, signal, access_type, sicode, sv);
 
 out:
-	if ((tstate & TSTATE_PRIV) == 0) {
+	if ((tf->tf_tstate & TSTATE_PRIV) == 0) {
 		userret(p);
 		share_fpu(p, tf);
 	}
@@ -879,18 +866,15 @@ void
 data_access_error(struct trapframe64 *tf, unsigned type, vaddr_t afva,
     u_long afsr, vaddr_t sfva, u_long sfsr)
 {
+	struct proc *p = curproc;
 	u_long pc;
-	u_int64_t tstate;
-	struct proc *p;
 	vaddr_t onfault;
 	union sigval sv;
 
 	uvmexp.traps++;
-	if ((p = curproc) == NULL)	/* safety check */
+	if (p == NULL)		/* safety check */
 		p = &proc0;
 
-	tstate = tf->tf_tstate;
-
 	/*
 	 * Catch PCI config space reads.
 	 */
@@ -901,19 +885,16 @@ data_access_error(struct trapframe64 *tf, unsigned type, vaddr_t afva,
 
 	pc = tf->tf_pc;
 
-	sv.sival_ptr = (void *)pc;
-
 	onfault = (long)p->p_addr->u_pcb.pcb_onfault;
 	printf("data error type %x sfsr=%lx sfva=%lx afsr=%lx afva=%lx tf=%p\n",
-		type, sfsr, sfva, afsr, afva, tf);
+	    type, sfsr, sfva, afsr, afva, tf);
 
 	if (afsr == 0 && sfsr == 0) {
 		printf("data_access_error: no fault\n");
 		goto out;	/* No fault. Why were we called? */
 	}
 
-	if (tstate & TSTATE_PRIV) {
-
+	if (tf->tf_tstate & TSTATE_PRIV) {
 		if (!onfault) {
 			(void) splhigh();
 			panic("data fault: pc=%lx addr=%lx sfsr=%lb",
@@ -935,10 +916,11 @@ data_access_error(struct trapframe64 *tf, unsigned type, vaddr_t afva,
 		return;
 	}
 
+	sv.sival_ptr = (void *)pc;
 	trapsignal(p, SIGSEGV, PROT_READ | PROT_WRITE, SEGV_MAPERR, sv);
 
 out:
-	if ((tstate & TSTATE_PRIV) == 0) {
+	if ((tf->tf_tstate & TSTATE_PRIV) == 0) {
 		userret(p);
 		share_fpu(p, tf);
 	}
@@ -952,45 +934,34 @@ void
 text_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
     u_long sfsr)
 {
-	u_int64_t tstate;
-	struct proc *p;
-	struct vmspace *vm;
-	vaddr_t va;
-	int rv;
-	vm_prot_t access_type;
+	struct proc *p = curproc;
+	vaddr_t va = trunc_page(pc);
+	vm_prot_t access_type = PROT_EXEC;
 	union sigval sv;
-
-	sv.sival_ptr = (void *)pc;
+	int signal, sicode, error;
 
 	uvmexp.traps++;
-	if ((p = curproc) == NULL)	/* safety check */
+	if (p == NULL)		/* safety check */
 		panic("text_access_fault: no curproc");
 
-	tstate = tf->tf_tstate;
-
-	va = trunc_page(pc);
-
-	/* Now munch on protections... */
+	sv.sival_ptr = (void *)pc;
 
-	access_type = PROT_EXEC;
-	if (tstate & TSTATE_PRIV) {
+	if (tf->tf_tstate & TSTATE_PRIV) {
 		(void) splhigh();
 		panic("kernel text_access_fault: pc=%lx va=%lx", pc, va);
 		/* NOTREACHED */
-	} else {
-		p->p_md.md_tf = tf;
-		refreshcreds(p);
-		if (!uvm_map_inentry(p, &p->p_spinentry, PROC_STACK(p),
-		    "[%s]%d/%d sp=%lx inside %lx-%lx: not MAP_STACK\n",
-		    uvm_map_inentry_sp, p->p_vmspace->vm_map.sserial))
-			goto out;
 	}
 
-	KERNEL_LOCK();
+	p->p_md.md_tf = tf;
+	refreshcreds(p);
+	if (!uvm_map_inentry(p, &p->p_spinentry, PROC_STACK(p),
+	    "[%s]%d/%d sp=%lx inside %lx-%lx: not MAP_STACK\n",
+	    uvm_map_inentry_sp, p->p_vmspace->vm_map.sserial))
+		goto out;
 
-	vm = p->p_vmspace;
-	/* alas! must call the horrible vm code */
-	rv = uvm_fault(&vm->vm_map, va, 0, access_type);
+	KERNEL_LOCK();
+	error = uvm_fault(&p->p_vmspace->vm_map, va, 0, access_type);
+	KERNEL_UNLOCK();
 
 	/*
 	 * If this was a stack access we keep track of the maximum
@@ -999,37 +970,33 @@ text_access_fault(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 	 * the current limit and we need to reflect that as an access
 	 * error.
 	 */
-	if (rv == 0)
+	if (error == 0) {
 		uvm_grow(p, va);
+		goto out;
+	}
 
-	if (rv != 0) {
-		/*
-		 * Pagein failed. Any other page fault in kernel, die; if user
-		 * fault, deliver SIGSEGV.
-		 */
-		int signal, sicode;
-
-		if (tstate & TSTATE_PRIV) {
-			(void) splhigh();
-			panic("kernel text fault: pc=%llx", (unsigned long long)pc);
-			/* NOTREACHED */
-		}
-
-		signal = SIGSEGV;
-		sicode = SEGV_MAPERR;
-		if (rv == EACCES)
-			sicode = SEGV_ACCERR;
-		if (rv == EIO) {
-			signal = SIGBUS;
-			sicode = BUS_OBJERR;
-		}
-		trapsignal(p, signal, access_type, sicode, sv);
+	/*
+	 * Pagein failed. Any other page fault in kernel, die; if user
+	 * fault, deliver SIGSEGV.
+	 */
+	if (tf->tf_tstate & TSTATE_PRIV) {
+		(void) splhigh();
+		panic("kernel text fault: pc=%llx", (unsigned long long)pc);
+		/* NOTREACHED */
 	}
 
-	KERNEL_UNLOCK();
+	signal = SIGSEGV;
+	sicode = SEGV_MAPERR;
+	if (error == EACCES)
+		sicode = SEGV_ACCERR;
+	else if (error == EIO) {
+		signal = SIGBUS;
+		sicode = BUS_OBJERR;
+	}
+	trapsignal(p, signal, access_type, sicode, sv);
 
 out:
-	if ((tstate & TSTATE_PRIV) == 0) {
+	if ((tf->tf_tstate & TSTATE_PRIV) == 0) {
 		userret(p);
 		share_fpu(p, tf);
 	}
@@ -1047,27 +1014,24 @@ void
 text_access_error(struct trapframe64 *tf, unsigned type, vaddr_t pc,
     u_long sfsr, vaddr_t afva, u_long afsr)
 {
-	int64_t tstate;
-	struct proc *p;
-	struct vmspace *vm;
-	vaddr_t va;
-	int rv;
-	vm_prot_t access_type;
+	struct proc *p = curproc;
+	vaddr_t va = trunc_page(pc);
+	vm_prot_t access_type = PROT_EXEC;
 	union sigval sv;
+	int signal, sicode, error;
 
-	sv.sival_ptr = (void *)pc;
 	uvmexp.traps++;
-	if ((p = curproc) == NULL)	/* safety check */
+	if (p == NULL)		/* safety check */
 		p = &proc0;
 
-	tstate = tf->tf_tstate;
+	sv.sival_ptr = (void *)pc;
 
 	if ((afsr) != 0) {
 		printf("text_access_error: memory error...\n");
-		printf("text memory error type %d sfsr=%lx sfva=%lx afsr=%lx afva=%lx tf=%p\n",
-		       type, sfsr, pc, afsr, afva, tf);
+		printf("type %d sfsr=%lx sfva=%lx afsr=%lx afva=%lx tf=%p\n",
+		    type, sfsr, pc, afsr, afva, tf);
 
-		if (tstate & TSTATE_PRIV)
+		if (tf->tf_tstate & TSTATE_PRIV)
 			panic("text_access_error: kernel memory error");
 
 		/* User fault -- Berr */
@@ -1077,29 +1041,23 @@ text_access_error(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 	if ((sfsr & SFSR_FV) == 0 || (sfsr & SFSR_FT) == 0)
 		goto out;	/* No fault. Why were we called? */
 
-	va = trunc_page(pc);
-
-	/* Now munch on protections... */
-	access_type = PROT_EXEC;
-	if (tstate & TSTATE_PRIV) {
+	if (tf->tf_tstate & TSTATE_PRIV) {
 		(void) splhigh();
 		panic("kernel text error: pc=%lx sfsr=%lb", pc,
 		    sfsr, SFSR_BITS);
 		/* NOTREACHED */
-	} else {
-		p->p_md.md_tf = tf;
-		refreshcreds(p);
-		if (!uvm_map_inentry(p, &p->p_spinentry, PROC_STACK(p),
-		    "[%s]%d/%d sp=%lx inside %lx-%lx: not MAP_STACK\n",
-		    uvm_map_inentry_sp, p->p_vmspace->vm_map.sserial))
-			goto out;
 	}
 
-	KERNEL_LOCK();
+	p->p_md.md_tf = tf;
+	refreshcreds(p);
+	if (!uvm_map_inentry(p, &p->p_spinentry, PROC_STACK(p),
+	    "[%s]%d/%d sp=%lx inside %lx-%lx: not MAP_STACK\n",
+	    uvm_map_inentry_sp, p->p_vmspace->vm_map.sserial))
+		goto out;
 
-	vm = p->p_vmspace;
-	/* alas! must call the horrible vm code */
-	rv = uvm_fault(&vm->vm_map, va, 0, access_type);
+	KERNEL_LOCK();
+	error = uvm_fault(&p->p_vmspace->vm_map, va, 0, access_type);
+	KERNEL_UNLOCK();
 
 	/*
 	 * If this was a stack access we keep track of the maximum
@@ -1108,39 +1066,35 @@ text_access_error(struct trapframe64 *tf, unsigned type, vaddr_t pc,
 	 * the current limit and we need to reflect that as an access
 	 * error.
 	 */
-	if (rv == 0)
+	if (error == 0) {
 		uvm_grow(p, va);
+		goto out;
+	}
 
-	if (rv != 0) {
-		/*
-		 * Pagein failed.  If doing copyin/out, return to onfault
-		 * address.  Any other page fault in kernel, die; if user
-		 * fault, deliver SIGSEGV.
-		 */
-		int signal, sicode;
-
-		if (tstate & TSTATE_PRIV) {
-			(void) splhigh();
-			panic("kernel text error: pc=%lx sfsr=%lb", pc,
-			    sfsr, SFSR_BITS);
-			/* NOTREACHED */
-		}
-
-		signal = SIGSEGV;
-		sicode = SEGV_MAPERR;
-		if (rv == EACCES)
-			sicode = SEGV_ACCERR;
-		if (rv == EIO) {
-			signal = SIGBUS;
-			sicode = BUS_OBJERR;
-		}
-		trapsignal(p, signal, access_type, sicode, sv);
+	/*
+	 * Pagein failed.  If doing copyin/out, return to onfault
+	 * address.  Any other page fault in kernel, die; if user
+	 * fault, deliver SIGSEGV.
+	 */
+	if (tf->tf_tstate & TSTATE_PRIV) {
+		(void) splhigh();
+		panic("kernel text error: pc=%lx sfsr=%lb", pc,
+		    sfsr, SFSR_BITS);
+		/* NOTREACHED */
 	}
 
-	KERNEL_UNLOCK();
+	signal = SIGSEGV;
+	sicode = SEGV_MAPERR;
+	if (error == EACCES)
+		sicode = SEGV_ACCERR;
+	else if (error == EIO) {
+		signal = SIGBUS;
+		sicode = BUS_OBJERR;
+	}
+	trapsignal(p, signal, access_type, sicode, sv);
 
 out:
-	if ((tstate & TSTATE_PRIV) == 0) {
+	if ((tf->tf_tstate & TSTATE_PRIV) == 0) {
 		userret(p);
 		share_fpu(p, tf);
 	}