[PATCH] x86-64 updates

Lots of x86-64 updates. Merge with 2.4 and NUMA works now. Also reenabled
the preemptive kernel. And some other bug fixes.
IOMMU disabled by default now because it has problems.

 - Add more CONFIG options for device driver debugging and iommu
   force/debug.  (don't enable iommu force currently)
 - Some S3/ACPI fixes/cleanups from Pavel.
 - Set MSG_COMPAT_* in msg_flags for networking 32bit emulation.
   This unfortunately still doesn't fix the fd passing problems.
 - Sync PCI IOMMU code with 2.4 (minor fixes, flush less often)
 - Really fix UP compilation (Pavel)
 - Reenable preempt
 - Fix CONFIG_DISCONTIGMEM bootup and enable.  Still needs more tuning.
 - Fix some bugs in topology discovery and clean code up.
 - Don't put unwind tables into object files
 - Some kernel debugging hooks
 - Move CPU detection into early real mode code to better interact with
   vesafb consoles
 - Initialize mode in real mode character output
 - New 32bit FPU signal save/restore
 - Various fixes in FPU handling in ptrace
 - Fix security holes in ptrace (32bit and 64bit)
 - Fix serial ioctl (including security hole)
 - Add bluetooth ioctls to 32bit emu (from sparc64)
 - Correctly enable si_val in queued signals in 32bit emulation
 - Rework SEM_STAT emulation.  LTP still fails unfortunately.
 - Fix error case in msg* emulation
 - Fix debug register access from ptrace (Michal Ludvig, me)
 - Fix handling of NULL arguments in 32bit execve
 - Fix some error cases for 32bit readv/writev (LTP still complains)
 - Remove rate control from unimplemented syscall warnings
 - Fix error message for missing aperture
 - Turn some APIC printks into Dprintk to make the bootup more quiet
 - Some fixes for no APIC (probably still broken), add disableapic
   option (untested)
 - Sync K8 MCE handler with 2.4.  Should work a lot better now.
 - Remove never used KDB hooks
 - Fix buffer overflow in command line copying
 - Merge from i386: use separate status word for lazy FPU state
 - Don't force the IOMMU for dma masks < 4GB.
 - Print backtrace in Sysrq-T (from Andrea)
 - Merge from i386: fix FPU race in fork.
 - Disable NX mode by default for now
 - Rewrite dump_pagetable
 - Fix off by one bug in ioremap (i386 merge)
 - Merge from i386: handle VIA pci bridge bugs
 - Disable NUMA ACPI support (no SRAT support yet)
 - Fix aio 32bit emulation
 - Increase 32bit address space to nearly 4GB
 - Add exit_group syscall
 - Fix TLS setting in clone (Ulrich Drepper)

arch/x86_64/Kconfig

@@ -179,9 +179,7 @@ config SMP
 
 	  If you don't know what to do here, say N.
 
-# broken currently
 config PREEMPT
-	depends on NOT_WORKING
 	bool "Preemptible Kernel"
 	---help---
 	  This option reduces the latency of the kernel when reacting to
@@ -200,7 +198,7 @@ config PREEMPT
 # someone write a better help text please.
 config K8_NUMA
        bool "K8 NUMA support"
-       depends on SMP && NOT_WORKING
+       depends on SMP
        help
 	  Enable NUMA (Non Unified Memory Architecture) support for
 	  AMD Opteron Multiprocessor systems. The kernel will try to allocate
@@ -590,10 +588,8 @@ config DEBUG_SLAB
 	  allocation as well as poisoning memory on free to catch use of freed
 	  memory.
 
-#   bool '  Memory mapped I/O debugging' CONFIG_DEBUG_IOVIRT
 config MAGIC_SYSRQ
 	bool "Magic SysRq key"
-	depends on DEBUG_KERNEL
 	help
 	  If you say Y here, you will have some control over the system even
 	  if the system crashes for example during kernel debugging (e.g., you
@@ -639,14 +635,37 @@ config KALLSYMS
 
 config FRAME_POINTER
        bool "Compile the kernel with frame pointers"
-       depends on DEBUG_KERNEL	
        help
 	 Compile the kernel with frame pointers. This may help for some 
 	 debugging with external debuggers. Note the standard oops backtracer 
-	 doesn't make use of it and the x86-64 kernel doesn't ensure an consistent
+	 doesn't make use of this  and the x86-64 kernel doesn't ensure an consistent
 	 frame pointer through inline assembly (semaphores etc.)
 	 Normally you should say N.
 
+config IOMMU_DEBUG
+       bool "Force IOMMU to on" 
+       help
+         Force the IOMMU to on even when you have less than 4GB of memory and add 
+	 debugging code. 
+	 Can be disabled at boot time with iommu=noforce.
+
+config IOMMU_LEAK
+       bool "IOMMU leak tracing"
+       depends on DEBUG_KERNEL
+       help
+         Add a simple leak tracer to the IOMMU code. This is useful when you
+	 are debugging a buggy device driver that leaks IOMMU mappings.
+       
+config MCE_DEBUG
+       bool "K8 Machine check debugging mode"
+       default y 
+       help
+         Turn on all Machine Check debugging for device driver problems.	  
+	 This can cause panics, but is useful to find device driver problems.
+
+#config X86_REMOTE_DEBUG
+#       bool "kgdb debugging stub"
+
 endmenu
 
 source "security/Kconfig"

arch/x86_64/Makefile

@@ -47,6 +47,10 @@ CFLAGS += -fno-reorder-blocks
 # should lower this a lot and see how much .text is saves	
 CFLAGS += -finline-limit=2000
 #CFLAGS += -g
+# don't enable this when you use kgdb:
+ifneq ($(CONFIG_X86_REMOTE_DEBUG),y)
+CFLAGS += -fno-asynchronous-unwind-tables
+endif
 
 head-y := arch/x86_64/kernel/head.o arch/x86_64/kernel/head64.o arch/x86_64/kernel/init_task.o
 

arch/x86_64/boot/compressed/misc.c

@@ -274,7 +274,7 @@ static void error(char *x)
 	puts(x);
 	puts("\n\n -- System halted");
 
-	while(1);	/* Halt */
+	while(1);
 }
 
 void setup_normal_output_buffer(void)
@@ -429,8 +429,6 @@ int decompress_kernel(struct moveparams *mv, void *rmode)
 	else setup_output_buffer_if_we_run_high(mv);
 
 	makecrc();
-	puts("Checking CPU type...");
-	check_cpu();
 	puts(".\nDecompressing Linux...");
 	gunzip();
 	puts("done.\nBooting the kernel.\n");

arch/x86_64/boot/setup.S

@@ -42,6 +42,7 @@
  * if CX/DX have been changed in the e801 call and if so use AX/BX .
  * Michael Miller, April 2001 <michaelm@mjmm.org>
  *
+ * Added long mode checking and SSE force. March 2003, Andi Kleen.		
  */
 
 #include <linux/config.h>
@@ -200,10 +201,10 @@ fin:	ret
 prtsp2:	call	prtspc		# Print double space
 prtspc:	movb	$0x20, %al	# Print single space (note: fall-thru)
 
-# Part of above routine, this one just prints ascii al
-prtchr:	pushw	%ax
+prtchr:	
+	pushw	%ax
 	pushw	%cx
-	xorb	%bh, %bh
+	movw	$0007,%bx
 	movw	$0x01, %cx
 	movb	$0x0e, %ah
 	int	$0x10
@@ -280,6 +281,75 @@ good_sig:
 loader_panic_mess: .string "Wrong loader, giving up..."
 
 loader_ok:
+	/* check for long mode. */
+	/* we have to do this before the VESA setup, otherwise the user
+	   can't see the error message. */
+	
+	pushw	%ds
+	movw	%cs,%ax
+	movw	%ax,%ds
+	
+	/* minimum CPUID flags for x86-64 */
+	/* see http://www.x86-64.org/lists/discuss/msg02971.html */		
+#define SSE_MASK ((1<<25)|(1<<26))
+#define REQUIRED_MASK1 ((1<<0)|(1<<3)|(1<<4)|(1<<5)|(1<<6)|(1<<8)|(1<<11)| \
+					   (1<<13)|(1<<15)|(1<<24)|(1<<29))
+
+	pushfl				/* standard way to check for cpuid */
+	popl	%eax
+	movl	%eax,%ebx
+	xorl	$0x200000,%eax
+	pushl	%eax
+	popfl
+	pushfl
+	popl	%eax
+	cmpl	%eax,%ebx
+	jz	no_longmode		/* cpu has no cpuid */
+	movl	$0x80000000,%eax
+	cpuid
+	cmpl	$0x80000001,%eax
+	jb	no_longmode		/* no extended cpuid */
+	xor	%di,%di
+	cmpl	$0x68747541,%ebx	/* AuthenticAMD */
+	jnz	noamd
+	cmpl	$0x69746e65,%edx
+	jnz	noamd
+	cmpl	$0x444d4163,%ecx
+	jnz	noamd
+	mov	$1,%di			/* cpu is from AMD */
+noamd:		
+	movl	$0x80000001,%eax		
+	cpuid
+	andl	$REQUIRED_MASK1,%edx
+	xorl	$REQUIRED_MASK1,%edx
+	jnz	no_longmode
+sse_test:		
+	movl	$1,%eax
+	cpuid
+	andl	$SSE_MASK,%edx
+	cmpl	$SSE_MASK,%edx
+	je	sse_ok
+	test	%di,%di
+	jz	no_longmode	/* only try to force SSE on AMD */ 
+	movl	$0xc0010015,%ecx	/* HWCR */
+	rdmsr
+	btr	$15,%eax	/* enable SSE */
+	wrmsr
+	xor	%di,%di		/* don't loop */
+	jmp	sse_test	/* try again */	
+no_longmode:
+	call	beep
+	lea	long_mode_panic,%si
+	call	prtstr
+no_longmode_loop:		
+	jmp	no_longmode_loop
+long_mode_panic:
+	.string "Your CPU does not support long mode. Use a 32bit distribution."
+	.byte 0
+	
+sse_ok:
+	popw	%ds
+	
 # Get memory size (extended mem, kB)
 
 	xorl	%eax, %eax

arch/x86_64/ia32/fpu32.c

@@ -77,17 +77,20 @@ static inline int convert_fxsr_from_user(struct i387_fxsave_struct *fxsave,
 	struct _fpxreg *to;
 	struct _fpreg *from;
 	int i;
-	int err; 
-	__u32 v;
+	u32 v;
+	int err = 0;
 
-	err = __get_user(fxsave->cwd, &buf->cw); 
-	err |= __get_user(fxsave->swd, &buf->sw);
-	err |= __get_user(fxsave->twd, &buf->tag);
+#define G(num,val) err |= __get_user(val, num + (u32 *)buf)
+	G(0, fxsave->cwd);
+	G(1, fxsave->swd);
+	G(2, fxsave->twd);
 	fxsave->twd = twd_i387_to_fxsr(fxsave->twd);
-	err |= __get_user(fxsave->rip, &buf->ipoff); 
-	err |= __get_user(fxsave->rdp, &buf->dataoff); 
-	err |= __get_user(v, &buf->cssel); 
-	fxsave->fop = v >> 16;
+	G(3, fxsave->rip);
+	G(4, v);
+	fxsave->fop = v>>16;	/* cs ignored */
+	G(5, fxsave->rdp);
+	/* 6: ds ignored */
+#undef G
 	if (err) 
 		return -1; 
 
@@ -109,21 +112,29 @@ static inline int convert_fxsr_to_user(struct _fpstate_ia32 *buf,
 	struct _fpreg *to;
 	struct _fpxreg *from;
 	int i;
-	u32 ds; 
-	int err; 
-
-	err = __put_user((unsigned long)fxsave->cwd | 0xffff0000, &buf->cw);
-	err |= __put_user((unsigned long)fxsave->swd | 0xffff0000, &buf->sw);
-	err |= __put_user((u32)fxsave->rip, &buf->ipoff); 
-	err |= __put_user((u32)(regs->cs | ((u32)fxsave->fop << 16)), 
-			  &buf->cssel); 
-	err |= __put_user((u32)twd_fxsr_to_i387(fxsave), &buf->tag); 
-	err |= __put_user((u32)fxsave->rdp, &buf->dataoff); 
-	if (tsk == current) 
-		asm("movl %%ds,%0 " : "=r" (ds)); 
-	else /* ptrace. task has stopped. */
+	u16 cs,ds; 
+	int err = 0; 
+
+	if (tsk == current) {
+		/* should be actually ds/cs at fpu exception time,
+		   but that information is not available in 64bit mode. */
+		asm("movw %%ds,%0 " : "=r" (ds)); 
+		asm("movw %%cs,%0 " : "=r" (cs)); 		
+	} else { /* ptrace. task has stopped. */
 		ds = tsk->thread.ds;
-	err |= __put_user(ds, &buf->datasel); 
+		cs = regs->cs;
+	} 
+
+#define P(num,val) err |= __put_user(val, num + (u32 *)buf)
+	P(0, (u32)fxsave->cwd | 0xffff0000);
+	P(1, (u32)fxsave->swd | 0xffff0000);
+	P(2, twd_fxsr_to_i387(fxsave));
+	P(3, (u32)fxsave->rip);
+	P(4,  cs | ((u32)fxsave->fop) << 16); 
+	P(5, fxsave->rdp);
+	P(6, 0xffff0000 | ds);
+#undef P
+
 	if (err) 
 		return -1; 
 
@@ -144,9 +155,9 @@ int restore_i387_ia32(struct task_struct *tsk, struct _fpstate_ia32 *buf, int fs
 				     &buf->_fxsr_env[0],
 				     sizeof(struct i387_fxsave_struct)))
 			return -1;
-	} 
 	tsk->thread.i387.fxsave.mxcsr &= 0xffbf;
-	current->used_math = 1;
+		tsk->used_math = 1;
+	} 
 	return convert_fxsr_from_user(&tsk->thread.i387.fxsave, buf);
 }  
 
@@ -157,12 +168,11 @@ int save_i387_ia32(struct task_struct *tsk,
 {
 	int err = 0;
 
-	if (!tsk->used_math) 
-		return 0;
-	tsk->used_math = 0; 
-	unlazy_fpu(tsk);
+	init_fpu(tsk);
 	if (convert_fxsr_to_user(buf, &tsk->thread.i387.fxsave, regs, tsk))
 		return -1;
+	if (fsave)
+		return 0;
 	err |= __put_user(tsk->thread.i387.fxsave.swd, &buf->status);
 	if (fsave) 
 		return err ? -1 : 1; 	

arch/x86_64/ia32/ia32_ioctl.c

@@ -39,6 +39,7 @@
 #include <linux/cdrom.h>
 #include <linux/loop.h>
 #include <linux/auto_fs.h>
+#include <linux/auto_fs4.h>
 #include <linux/devfs_fs.h>
 #include <linux/tty.h>
 #include <linux/vt_kern.h>
@@ -60,6 +61,8 @@
 #include <linux/if_tun.h>
 #include <linux/dirent.h>
 #include <linux/ctype.h>
+#include <net/bluetooth/bluetooth.h>
+#include <net/bluetooth/rfcomm.h>
 #if defined(CONFIG_BLK_DEV_LVM) || defined(CONFIG_BLK_DEV_LVM_MODULE)
 /* Ugh. This header really is not clean */
 #define min min
@@ -2906,35 +2909,28 @@ static int serial_struct_ioctl(unsigned fd, unsigned cmd,  void *ptr)
 {
 	typedef struct serial_struct SS;
 	struct serial_struct32 *ss32 = ptr; 
-	int err = 0;
+	int err;
 	struct serial_struct ss; 
 	mm_segment_t oldseg = get_fs(); 
-	set_fs(KERNEL_DS);
 	if (cmd == TIOCSSERIAL) { 
-		err = -EFAULT;
 		if (copy_from_user(&ss, ss32, sizeof(struct serial_struct32)))
-			goto out;
+			return -EFAULT;
 		memmove(&ss.iomem_reg_shift, ((char*)&ss.iomem_base)+4, 
 			sizeof(SS)-offsetof(SS,iomem_reg_shift)); 
 		ss.iomem_base = (void *)((unsigned long)ss.iomem_base & 0xffffffff);
 	}
-	if (!err)
+	set_fs(KERNEL_DS);
 		err = sys_ioctl(fd,cmd,(unsigned long)(&ss)); 
+	set_fs(oldseg);
 	if (cmd == TIOCGSERIAL && err >= 0) { 
-		__u32 base;
 		if (__copy_to_user(ss32,&ss,offsetof(SS,iomem_base)) ||
-		    __copy_to_user(&ss32->iomem_reg_shift,
-				   &ss.iomem_reg_shift,
-				   sizeof(SS) - offsetof(SS, iomem_reg_shift)))
-			err = -EFAULT;
-		if (ss.iomem_base > (unsigned char *)0xffffffff)
-			base = -1; 
-		else
-			base = (unsigned long)ss.iomem_base;
-		err |= __put_user(base, &ss32->iomem_base); 		
+		    __put_user((unsigned long)ss.iomem_base  >> 32 ? 
+			       0xffffffff : (unsigned)(unsigned long)ss.iomem_base,
+			       &ss32->iomem_base) ||
+		    __put_user(ss.iomem_reg_shift, &ss32->iomem_reg_shift) ||
+		    __put_user(ss.port_high, &ss32->port_high))
+			return -EFAULT;
 	} 
- out:
-	set_fs(oldseg);
 	return err;	
 }
 
@@ -3045,7 +3041,14 @@ static int do_blkgetsize64(unsigned int fd, unsigned int cmd,
        return sys_ioctl(fd, BLKGETSIZE64, arg);
 }
 
+/* Bluetooth ioctls */
+#define HCIUARTSETPROTO        _IOW('U', 200, int)
+#define HCIUARTGETPROTO        _IOR('U', 201, int)
 
+#define BNEPCONNADD    _IOW('B', 200, int)
+#define BNEPCONNDEL    _IOW('B', 201, int)
+#define BNEPGETCONNLIST        _IOR('B', 210, int)
+#define BNEPGETCONNINFO        _IOR('B', 211, int)
 
 struct usbdevfs_ctrltransfer32 {
 	__u8 bRequestType;
@@ -4093,6 +4096,7 @@ COMPATIBLE_IOCTL(AUTOFS_IOC_FAIL)
 COMPATIBLE_IOCTL(AUTOFS_IOC_CATATONIC)
 COMPATIBLE_IOCTL(AUTOFS_IOC_PROTOVER)
 COMPATIBLE_IOCTL(AUTOFS_IOC_EXPIRE)
+COMPATIBLE_IOCTL(AUTOFS_IOC_EXPIRE_MULTI)
 /* DEVFS */
 COMPATIBLE_IOCTL(DEVFSDIOC_GET_PROTO_REV)
 COMPATIBLE_IOCTL(DEVFSDIOC_SET_EVENT_MASK)
@@ -4200,6 +4204,17 @@ COMPATIBLE_IOCTL(HCISETLINKMODE)
 COMPATIBLE_IOCTL(HCISETACLMTU)
 COMPATIBLE_IOCTL(HCISETSCOMTU)
 COMPATIBLE_IOCTL(HCIINQUIRY)
+COMPATIBLE_IOCTL(HCIUARTSETPROTO)
+COMPATIBLE_IOCTL(HCIUARTGETPROTO)
+COMPATIBLE_IOCTL(RFCOMMCREATEDEV)
+COMPATIBLE_IOCTL(RFCOMMRELEASEDEV)
+COMPATIBLE_IOCTL(RFCOMMGETDEVLIST)
+COMPATIBLE_IOCTL(RFCOMMGETDEVINFO)
+COMPATIBLE_IOCTL(RFCOMMSTEALDLC)
+COMPATIBLE_IOCTL(BNEPCONNADD)
+COMPATIBLE_IOCTL(BNEPCONNDEL)
+COMPATIBLE_IOCTL(BNEPGETCONNLIST)
+COMPATIBLE_IOCTL(BNEPGETCONNINFO)
 /* Misc. */
 COMPATIBLE_IOCTL(0x41545900)		/* ATYIO_CLKR */
 COMPATIBLE_IOCTL(0x41545901)		/* ATYIO_CLKW */

arch/x86_64/ia32/ia32_signal.c

@@ -47,9 +47,16 @@ static int ia32_copy_siginfo_to_user(siginfo_t32 *to, siginfo_t *from)
 {
 	if (!access_ok (VERIFY_WRITE, to, sizeof(siginfo_t)))
 		return -EFAULT;
-	if (from->si_code < 0)
-		return __copy_to_user(to, from, sizeof(siginfo_t));
-	else {
+	if (from->si_code < 0) { 
+		/* the only field that's different is the alignment
+		   of the pointer in sigval_t. Move that 4 bytes down including
+		   padding. */
+		memmove(&((siginfo_t32 *)&from)->si_int,
+			&from->si_int, 
+			sizeof(siginfo_t) - offsetof(siginfo_t, si_int));
+		/* last 4 bytes stay the same */
+		return __copy_to_user(to, from, sizeof(siginfo_t32));
+	} else {
 		int err;
 
 		/* If you change siginfo_t structure, please be sure
@@ -59,7 +66,7 @@ static int ia32_copy_siginfo_to_user(siginfo_t32 *to, siginfo_t *from)
 		   3 ints plus the relevant union member.  */
 		err = __put_user(from->si_signo, &to->si_signo);
 		err |= __put_user(from->si_errno, &to->si_errno);
-		err |= __put_user((short)from->si_code, &to->si_code);
+		err |= __put_user(from->si_code, &to->si_code);
 		/* First 32bits of unions are always present.  */
 		err |= __put_user(from->si_pid, &to->si_pid);
 		switch (from->si_code >> 16) {
@@ -108,6 +115,7 @@ sys32_sigaltstack(const stack_ia32_t *uss_ptr, stack_ia32_t *uoss_ptr,
 	mm_segment_t seg; 
 	if (uss_ptr) { 
 		u32 ptr;
+		memset(&uss,0,sizeof(stack_t));
 	if (!access_ok(VERIFY_READ,uss_ptr,sizeof(stack_ia32_t)) ||
 		    __get_user(ptr, &uss_ptr->ss_sp) ||
 		    __get_user(uss.ss_flags, &uss_ptr->ss_flags) ||
@@ -340,8 +348,11 @@ ia32_setup_sigcontext(struct sigcontext_ia32 *sc, struct _fpstate_ia32 *fpstate,
 	tmp = save_i387_ia32(current, fpstate, regs, 0);
 	if (tmp < 0)
 	  err = -EFAULT;
-	else
+	else { 
+		current->used_math = 0;
+		stts();
 	  err |= __put_user((u32)(u64)(tmp ? fpstate : NULL), &sc->fpstate);
+	}
 
 	/* non-iBCS2 extensions.. */
 	err |= __put_user(mask, &sc->oldmask);

arch/x86_64/ia32/ipc32.c

@@ -187,12 +187,58 @@ ipc_parse_version32 (int *cmd)
 	}
 }
 
+static int put_semid(void *user_semid, struct semid64_ds *s, int version)
+{
+	int err2;
+	switch (version) { 
+	case IPC_64: { 
+		struct semid64_ds32 *usp64 = (struct semid64_ds32 *) user_semid;
+		
+		if (!access_ok(VERIFY_WRITE, usp64, sizeof(*usp64))) {
+			err2 = -EFAULT;
+			break;
+		} 
+		err2 = __put_user(s->sem_perm.key, &usp64->sem_perm.key);
+		err2 |= __put_user(s->sem_perm.uid, &usp64->sem_perm.uid);
+		err2 |= __put_user(s->sem_perm.gid, &usp64->sem_perm.gid);
+		err2 |= __put_user(s->sem_perm.cuid, &usp64->sem_perm.cuid);
+		err2 |= __put_user(s->sem_perm.cgid, &usp64->sem_perm.cgid);
+		err2 |= __put_user(s->sem_perm.mode, &usp64->sem_perm.mode);
+		err2 |= __put_user(s->sem_perm.seq, &usp64->sem_perm.seq);
+		err2 |= __put_user(s->sem_otime, &usp64->sem_otime);
+		err2 |= __put_user(s->sem_ctime, &usp64->sem_ctime);
+		err2 |= __put_user(s->sem_nsems, &usp64->sem_nsems);
+		break;
+	}
+	default: {
+		struct semid_ds32 *usp32 = (struct semid_ds32 *) user_semid;
+		
+		if (!access_ok(VERIFY_WRITE, usp32, sizeof(*usp32))) {
+			err2 = -EFAULT;
+			break;
+		} 
+		err2 = __put_user(s->sem_perm.key, &usp32->sem_perm.key);
+		err2 |= __put_user(s->sem_perm.uid, &usp32->sem_perm.uid);
+		err2 |= __put_user(s->sem_perm.gid, &usp32->sem_perm.gid);
+		err2 |= __put_user(s->sem_perm.cuid, &usp32->sem_perm.cuid);
+		err2 |= __put_user(s->sem_perm.cgid, &usp32->sem_perm.cgid);
+		err2 |= __put_user(s->sem_perm.mode, &usp32->sem_perm.mode);
+		err2 |= __put_user(s->sem_perm.seq, &usp32->sem_perm.seq);
+		err2 |= __put_user(s->sem_otime, &usp32->sem_otime);
+		err2 |= __put_user(s->sem_ctime, &usp32->sem_ctime);
+		err2 |= __put_user(s->sem_nsems, &usp32->sem_nsems);
+		break;
+	}
+	}
+	return err2;
+}
+
 static int
 semctl32 (int first, int second, int third, void *uptr)
 {
 	union semun fourth;
 	u32 pad;
-	int err = 0, err2;
+	int err;
 	struct semid64_ds s;
 	mm_segment_t old_fs;
 	int version = ipc_parse_version32(&third);
@@ -225,46 +271,10 @@ semctl32 (int first, int second, int third, void *uptr)
 		fourth.__pad = &s;
 		old_fs = get_fs();
 		set_fs(KERNEL_DS);
-		err = sys_semctl(first, second|IPC_64, third, fourth);
+		err = sys_semctl(first, second, third|IPC_64, fourth);
 		set_fs(old_fs);
-
-		if (version == IPC_64) {
-			struct semid64_ds32 *usp64 = (struct semid64_ds32 *) A(pad);
-
-			if (!access_ok(VERIFY_WRITE, usp64, sizeof(*usp64))) {
-				err = -EFAULT;
-				break;
-			}
-			err2 = __put_user(s.sem_perm.key, &usp64->sem_perm.key);
-			err2 |= __put_user(s.sem_perm.uid, &usp64->sem_perm.uid);
-			err2 |= __put_user(s.sem_perm.gid, &usp64->sem_perm.gid);
-			err2 |= __put_user(s.sem_perm.cuid, &usp64->sem_perm.cuid);
-			err2 |= __put_user(s.sem_perm.cgid, &usp64->sem_perm.cgid);
-			err2 |= __put_user(s.sem_perm.mode, &usp64->sem_perm.mode);
-			err2 |= __put_user(s.sem_perm.seq, &usp64->sem_perm.seq);
-			err2 |= __put_user(s.sem_otime, &usp64->sem_otime);
-			err2 |= __put_user(s.sem_ctime, &usp64->sem_ctime);
-			err2 |= __put_user(s.sem_nsems, &usp64->sem_nsems);
-		} else {
-			struct semid_ds32 *usp32 = (struct semid_ds32 *) A(pad);
-
-			if (!access_ok(VERIFY_WRITE, usp32, sizeof(*usp32))) {
-				err = -EFAULT;
-				break;
-			}
-			err2 = __put_user(s.sem_perm.key, &usp32->sem_perm.key);
-			err2 |= __put_user(s.sem_perm.uid, &usp32->sem_perm.uid);
-			err2 |= __put_user(s.sem_perm.gid, &usp32->sem_perm.gid);
-			err2 |= __put_user(s.sem_perm.cuid, &usp32->sem_perm.cuid);
-			err2 |= __put_user(s.sem_perm.cgid, &usp32->sem_perm.cgid);
-			err2 |= __put_user(s.sem_perm.mode, &usp32->sem_perm.mode);
-			err2 |= __put_user(s.sem_perm.seq, &usp32->sem_perm.seq);
-			err2 |= __put_user(s.sem_otime, &usp32->sem_otime);
-			err2 |= __put_user(s.sem_ctime, &usp32->sem_ctime);
-			err2 |= __put_user(s.sem_nsems, &usp32->sem_nsems);
-		}
-		if (err2)
-		    err = -EFAULT;
+		if (!err)
+			err = put_semid((void *)A(pad), &s, version);
 		break;
 	default:
 		err = -EINVAL;
@@ -343,6 +353,7 @@ do_sys32_msgrcv (int first, int second, int msgtyp, int third, int version, void
 	return err;
 }
 
+
 static int
 msgctl32 (int first, int second, void *uptr)
 {
@@ -387,7 +398,6 @@ msgctl32 (int first, int second, void *uptr)
 		set_fs(KERNEL_DS);
 		err = sys_msgctl(first, second|IPC_64, (void *) &m64);
 		set_fs(old_fs);
-
 		if (version == IPC_64) {
 			if (!access_ok(VERIFY_WRITE, up64, sizeof(*up64))) {
 				err = -EFAULT;
@@ -608,7 +618,9 @@ shmctl32 (int first, int second, void *uptr)
 		if (err2)
 			err = -EFAULT;
 		break;
-
+	default:
+		err = -EINVAL;
+		break;
 	}
 	return err;
 }

arch/x86_64/ia32/ptrace32.c

@@ -8,7 +8,7 @@
  * This allows to access 64bit processes too; but there is no way to see the extended 
  * register contents.
  *
- * $Id: ptrace32.c,v 1.12 2002/03/24 13:02:02 ak Exp $
+ * $Id: ptrace32.c,v 1.16 2003/03/14 16:06:35 ak Exp $
  */ 
 
 #include <linux/kernel.h>
@@ -22,11 +22,9 @@
 #include <asm/errno.h>
 #include <asm/debugreg.h>
 #include <asm/i387.h>
-#include <asm/desc.h>
-#include <asm/ldt.h>
 #include <asm/fpu32.h>
-#include <linux/mm.h>
 #include <linux/ptrace.h>
+#include <linux/mm.h>
 
 #define R32(l,q) \
 	case offsetof(struct user32, regs.l): stack[offsetof(struct pt_regs, q)/8] = val; break
@@ -39,29 +37,26 @@ static int putreg32(struct task_struct *child, unsigned regno, u32 val)
 	switch (regno) {
 	case offsetof(struct user32, regs.fs):
 		if (val && (val & 3) != 3) return -EIO; 
-	        child->thread.fs = val; 
+		child->thread.fs = val & 0xffff; 
 		break;
 	case offsetof(struct user32, regs.gs):
 		if (val && (val & 3) != 3) return -EIO; 
-		child->thread.gs = val;
+		child->thread.gs = val & 0xffff;
 		break;
 	case offsetof(struct user32, regs.ds):
 		if (val && (val & 3) != 3) return -EIO; 
-		child->thread.ds = val;
+		child->thread.ds = val & 0xffff;
 		break;
 	case offsetof(struct user32, regs.es):
-		if (val && (val & 3) != 3) return -EIO; 
-		child->thread.es = val;
+		child->thread.es = val & 0xffff;
 		break;
-
 	case offsetof(struct user32, regs.ss): 
 		if ((val & 3) != 3) return -EIO;
-		stack[offsetof(struct pt_regs, ss)/8] = val; 
+        	stack[offsetof(struct pt_regs, ss)/8] = val & 0xffff;
 		break;
-
 	case offsetof(struct user32, regs.cs): 
 		if ((val & 3) != 3) return -EIO;
-		stack[offsetof(struct pt_regs, cs)/8] = val; 
+		stack[offsetof(struct pt_regs, cs)/8] = val & 0xffff;
 		break;
 
 	R32(ebx, rbx); 
@@ -79,8 +74,16 @@ static int putreg32(struct task_struct *child, unsigned regno, u32 val)
 		stack[offsetof(struct pt_regs, eflags)/8] = val & 0x44dd5; 
 		break;
 
-	case offsetof(struct user32, u_debugreg[0]) ... offsetof(struct user32, u_debugreg[6]):
-		child->thread.debugreg[(regno-offsetof(struct user32, u_debugreg[0]))/4] = val; 
+	case offsetof(struct user32, u_debugreg[4]): 
+	case offsetof(struct user32, u_debugreg[5]):
+		return -EIO;
+
+	case offsetof(struct user32, u_debugreg[0]) ...
+	     offsetof(struct user32, u_debugreg[3]):
+	case offsetof(struct user32, u_debugreg[6]):
+		child->thread.debugreg
+			[(regno-offsetof(struct user32, u_debugreg[0]))/4] 
+			= val; 
 		break; 
 
 	case offsetof(struct user32, u_debugreg[7]):
@@ -170,11 +173,19 @@ static struct task_struct *find_target(int request, int pid, int *err)
 	if (child)
 		get_task_struct(child);
 	read_unlock(&tasklist_lock);
-	*err = ptrace_check_attach(child,0);
-	if (*err == 0)
+	if (child) { 
+		*err = -EPERM;
+		if (child->pid == 1) 
+			goto out;
+		*err = ptrace_check_attach(child, request == PTRACE_KILL); 
+		if (*err < 0) 
+			goto out;
 		return child; 
+	} 
+ out:
 	put_task_struct(child);
 	return NULL; 
+	
 } 
 
 extern asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, unsigned long data);
@@ -187,6 +198,9 @@ asmlinkage long sys32_ptrace(long request, u32 pid, u32 addr, u32 data)
 	__u32 val;
 
 	switch (request) { 
+	default:
+		return sys_ptrace(request, pid, addr, data); 
+
 	case PTRACE_PEEKTEXT:
 	case PTRACE_PEEKDATA:
 	case PTRACE_POKEDATA:
@@ -201,9 +215,6 @@ asmlinkage long sys32_ptrace(long request, u32 pid, u32 addr, u32 data)
 	case PTRACE_GETFPXREGS:
 		break;
 		
-	default:
-		ret = sys_ptrace(request, pid, addr, data); 
-		return ret;
 	} 
 
 	child = find_target(request, pid, &ret);
@@ -261,7 +272,6 @@ asmlinkage long sys32_ptrace(long request, u32 pid, u32 addr, u32 data)
 			ret = -EIO;
 			break;
 		}
-		empty_fpu(child); 
 		ret = 0; 
 		for ( i = 0; i <= 16*4; i += sizeof(u32) ) {
 			ret |= __get_user(tmp, (u32 *) (unsigned long) data);
@@ -271,33 +281,47 @@ asmlinkage long sys32_ptrace(long request, u32 pid, u32 addr, u32 data)
 		break;
 	}
 
-	case PTRACE_SETFPREGS:
-		empty_fpu(child); 
+	case PTRACE_GETFPREGS:
+		ret = -EIO; 
+		if (!access_ok(VERIFY_READ, (void *)(u64)data, 
+			       sizeof(struct user_i387_struct)))
+			break;
 		save_i387_ia32(child, (void *)(u64)data, childregs, 1);
 		ret = 0; 
 			break;
 
-	case PTRACE_GETFPREGS:
-		empty_fpu(child); 
-		restore_i387_ia32(child, (void *)(u64)data, 1);
+	case PTRACE_SETFPREGS:
+		ret = -EIO;
+		if (!access_ok(VERIFY_WRITE, (void *)(u64)data, 
+			       sizeof(struct user_i387_struct)))
+			break;
 		ret = 0;
+		/* don't check EFAULT to be bug-to-bug compatible to i386 */
+		restore_i387_ia32(child, (void *)(u64)data, 1);
 		break;
 
 	case PTRACE_GETFPXREGS: { 
 		struct user32_fxsr_struct *u = (void *)(u64)data; 
-		empty_fpu(child); 
-		ret = copy_to_user(u, &child->thread.i387.fxsave, sizeof(*u));
-		ret |= __put_user(childregs->cs, &u->fcs);
-		ret |= __put_user(child->thread.ds, &u->fos); 
-		if (ret) 
+		init_fpu(child); 
+		ret = -EIO;
+		if (!access_ok(VERIFY_WRITE, u, sizeof(*u)))
+			break;
 			ret = -EFAULT;
+		if (__copy_to_user(u, &child->thread.i387.fxsave, sizeof(*u)))
+			break;
+		ret = __put_user(childregs->cs, &u->fcs);
+		ret |= __put_user(child->thread.ds, &u->fos); 
 		break; 
 	} 
 	case PTRACE_SETFPXREGS: { 
 		struct user32_fxsr_struct *u = (void *)(u64)data; 
-		empty_fpu(child); 
-		/* no error checking to be bug to bug compatible with i386 */ 
-		copy_from_user(&child->thread.i387.fxsave, u, sizeof(*u));
+		unlazy_fpu(child);
+		ret = -EIO;
+		if (!access_ok(VERIFY_READ, u, sizeof(*u)))
+			break;
+		/* no checking to be bug-to-bug compatible with i386 */
+		__copy_from_user(&child->thread.i387.fxsave, u, sizeof(*u));
+		child->used_math = 1;
 	        child->thread.i387.fxsave.mxcsr &= 0xffbf;
 		ret = 0; 
 			break;

arch/x86_64/ia32/sys_ia32.c

@@ -57,6 +57,7 @@
 #include <linux/binfmts.h>
 #include <linux/init.h>
 #include <linux/aio_abi.h>
+#include <linux/aio.h>
 #include <linux/compat.h>
 #include <linux/vfs.h>
 #include <linux/ptrace.h>
@@ -74,6 +75,7 @@
 
 #define A(__x)		((unsigned long)(__x))
 #define AA(__x)		((unsigned long)(__x))
+#define u32_to_ptr(x)	((void *)(u64)(x))
 #define ROUND_UP(x,a)	((__typeof__(x))(((unsigned long)(x) + ((a) - 1)) & ~((a) - 1)))
 #define NAME_OFFSET(de) ((int) ((de)->d_name - (char *) (de)))
 
@@ -738,7 +740,7 @@ asmlinkage ssize_t sys_readv(unsigned long,const struct iovec *,unsigned long);
 asmlinkage ssize_t sys_writev(unsigned long,const struct iovec *,unsigned long);
 
 static struct iovec *
-get_compat_iovec(struct compat_iovec *iov32, struct iovec *iov_buf, u32 count, int type, int *errp)
+get_compat_iovec(struct compat_iovec *iov32, struct iovec *iov_buf, u32 *count, int type, int *errp)
 {
 	int i;
 	u32 buf, len;
@@ -747,15 +749,18 @@ get_compat_iovec(struct compat_iovec *iov32, struct iovec *iov_buf, u32 count, i
 
 	/* Get the "struct iovec" from user memory */
 
-	if (!count)
+	*errp = 0;
+	if (!*count)
 		return 0;
-	if (count > UIO_MAXIOV)
+	*errp = -EINVAL;
+	if (*count > UIO_MAXIOV)
 		return(struct iovec *)0;
-	if(verify_area(VERIFY_READ, iov32, sizeof(struct compat_iovec)*count))
+	*errp = -EFAULT;
+	if(verify_area(VERIFY_READ, iov32, sizeof(struct compat_iovec)*(*count)))
 		return(struct iovec *)0;
-	if (count > UIO_FASTIOV) {
+	if (*count > UIO_FASTIOV) {
 		*errp = -ENOMEM; 
-		iov = kmalloc(count*sizeof(struct iovec), GFP_KERNEL);
+		iov = kmalloc(*count*sizeof(struct iovec), GFP_KERNEL);
 		if (!iov)
 			return((struct iovec *)0);
 	} else
@@ -763,14 +768,19 @@ get_compat_iovec(struct compat_iovec *iov32, struct iovec *iov_buf, u32 count, i
 
 	ivp = iov;
 	totlen = 0;
-	for (i = 0; i < count; i++) {
+	for (i = 0; i < *count; i++) {
 		*errp = __get_user(len, &iov32->iov_len) |
 		  	__get_user(buf, &iov32->iov_base);	
 		if (*errp)
 			goto error;
 		*errp = verify_area(type, (void *)A(buf), len);
-		if (*errp) 
+		if (*errp) {
+			if (i > 0) { 
+				*count = i;
+				break;
+			} 
 			goto error;
+		}
 		/* SuS checks: */
 		*errp = -EINVAL; 
 		if ((int)len < 0)
@@ -799,7 +809,7 @@ sys32_readv(int fd, struct compat_iovec *vector, u32 count)
 	int ret;
 	mm_segment_t old_fs = get_fs();
 
-	if ((iov = get_compat_iovec(vector, iovstack, count, VERIFY_WRITE, &ret)) == NULL)
+	if ((iov = get_compat_iovec(vector, iovstack, &count, VERIFY_WRITE, &ret)) == NULL)
 		return ret;
 	set_fs(KERNEL_DS);
 	ret = sys_readv(fd, iov, count);
@@ -817,7 +827,7 @@ sys32_writev(int fd, struct compat_iovec *vector, u32 count)
 	int ret;
 	mm_segment_t old_fs = get_fs();
 
-	if ((iov = get_compat_iovec(vector, iovstack, count, VERIFY_READ, &ret)) == NULL)
+	if ((iov = get_compat_iovec(vector, iovstack, &count, VERIFY_READ, &ret)) == NULL)
 		return ret;
 	set_fs(KERNEL_DS);
 	ret = sys_writev(fd, iov, count);
@@ -1672,21 +1682,26 @@ static int nargs(u32 src, char **dst)
 	return cnt; 
 } 
 
-long sys32_execve(char *name, u32 argv, u32 envp, struct pt_regs regs)
+asmlinkage long sys32_execve(char *name, u32 argv, u32 envp, struct pt_regs regs)
 { 
 	mm_segment_t oldseg; 
-	char **buf; 
-	int na,ne;
+	char **buf = NULL; 
+	int na = 0,ne = 0;
 	int ret;
-	unsigned sz; 
+	unsigned sz = 0; 
 
+	if (argv) {
 	na = nargs(argv, NULL); 
 	if (na < 0) 
 		return -EFAULT; 
+	} 	
+	if (envp) { 
 	ne = nargs(envp, NULL); 
 	if (ne < 0) 
 		return -EFAULT; 
+	}
 
+	if (argv || envp) { 
 	sz = (na+ne)*sizeof(void *); 
 	if (sz > PAGE_SIZE) 
 		buf = vmalloc(sz); 
@@ -1694,14 +1709,19 @@ long sys32_execve(char *name, u32 argv, u32 envp, struct pt_regs regs)
 		buf = kmalloc(sz, GFP_KERNEL); 
 	if (!buf)
 		return -ENOMEM; 
+	} 
 	
+	if (argv) { 
 	ret = nargs(argv, buf);
 	if (ret < 0)
 		goto free;
+	}
 
+	if (envp) { 
 	ret = nargs(envp, buf + na); 
 	if (ret < 0)
 		goto free; 
+	}
 
 	name = getname(name); 
 	ret = PTR_ERR(name); 
@@ -1710,7 +1730,7 @@ long sys32_execve(char *name, u32 argv, u32 envp, struct pt_regs regs)
 
 	oldseg = get_fs(); 
 	set_fs(KERNEL_DS);
-	ret = do_execve(name, buf, buf+na, &regs);  
+	ret = do_execve(name, argv ? buf : NULL, envp ? buf+na : NULL, &regs);  
 	set_fs(oldseg); 
 
 	if (ret == 0)
@@ -1719,10 +1739,12 @@ long sys32_execve(char *name, u32 argv, u32 envp, struct pt_regs regs)
 	putname(name);
  
 free:
+	if (argv || envp) { 
 	if (sz > PAGE_SIZE)
 		vfree(buf); 
 	else
 	kfree(buf);
+	}
 	return ret; 
 } 
 
@@ -2012,12 +2034,8 @@ long asmlinkage sys32_nfsservctl(int cmd, void *notused, void *notused2)
 
 long sys32_module_warning(void)
 { 
-	static long warn_time = -(60*HZ); 
-	if (time_before(warn_time + 60*HZ,jiffies) && strcmp(current->comm,"klogd")) { 
 		printk(KERN_INFO "%s: 32bit 2.4.x modutils not supported on 64bit kernel\n",
 		       current->comm);
-		warn_time = jiffies;
-	} 
 	return -ENOSYS ;
 } 
 
@@ -2055,6 +2073,7 @@ long sys32_sched_getaffinity(pid_t pid, unsigned int len,
 	return err;
 }
 
+
 extern long sys_io_setup(unsigned nr_reqs, aio_context_t *ctx);
 
 long sys32_io_setup(unsigned nr_reqs, u32 *ctx32p)
@@ -2071,48 +2090,47 @@ long sys32_io_setup(unsigned nr_reqs, u32 *ctx32p)
 	return ret;
 } 
 
-extern asmlinkage long sys_io_submit(aio_context_t ctx_id, long nr,
-				     struct iocb **iocbpp);
-
-long sys32_io_submit(aio_context_t ctx_id, unsigned long nr,
+asmlinkage long sys32_io_submit(aio_context_t ctx_id, int nr,
 		   u32 *iocbpp)
 {
-	mm_segment_t oldfs = get_fs(); 
-	int k, err = 0;
-	struct iocb **iocb64; 
-	if (nr > 128) 
+	struct kioctx *ctx;
+	long ret = 0;
+	int i;
+	
+	if (unlikely(nr < 0))
+		return -EINVAL;
+
+	if (unlikely(!access_ok(VERIFY_READ, iocbpp, (nr*sizeof(*iocbpp)))))
+		return -EFAULT;
+
+	ctx = lookup_ioctx(ctx_id);
+	if (unlikely(!ctx)) {
+		pr_debug("EINVAL: io_submit: invalid context id\n");
 		return -EINVAL; 
-	iocb64 = kmalloc(sizeof(struct iocb *) * nr, GFP_KERNEL);
-	if (!iocb64)
-		return -ENOMEM;
-	for (k = 0; k < nr && !err; k++) { 
-		u64 val1, val2;
-		u32 iocb32;
-		struct iocb *iocb; 	
-		err = get_user(iocb32, (u32 *)(u64)iocbpp[k]); 
-		iocb64[k] = iocb = (void *)(u64)iocb32; 
-		
-		if (get_user(val1, &iocb->aio_buf) ||
-		    get_user(val2, &iocb->aio_nbytes)) 
-			err = -EFAULT; 
-		else if (!val1) /* should check cmd */ 
-			;
-		else if (verify_area(VERIFY_WRITE, (void*)val1, val2))
-			err = -EFAULT; 
-
-		/* paranoia check - remove it when you are sure they
-		   are not pointers */
-		if (get_user(val1, &iocb->aio_reserved2) || val1 ||
-		    get_user(val2, &iocb->aio_reserved2) || val2)
-			err = -EFAULT; 		    		   
 	} 
-	if (!err) {
-		set_fs(KERNEL_DS);
-		err = sys_io_submit(ctx_id, nr, iocb64);
-		set_fs(oldfs); 
+
+	for (i=0; i<nr; i++) {
+		u32 p32;
+		struct iocb *user_iocb, tmp;
+
+		if (unlikely(__get_user(p32, iocbpp + i))) {
+			ret = -EFAULT;
+			break;
 	} 
-	kfree(iocb64);
-	return err;		
+		user_iocb = u32_to_ptr(p32);
+
+		if (unlikely(copy_from_user(&tmp, user_iocb, sizeof(tmp)))) {
+			ret = -EFAULT;
+			break;
+		}
+
+		ret = io_submit_one(ctx, user_iocb, &tmp);
+		if (ret)
+			break;
+	}
+
+	put_ioctx(ctx);
+	return i ? i : ret;
 }
 
 extern asmlinkage long sys_io_getevents(aio_context_t ctx_id,
@@ -2140,7 +2158,7 @@ asmlinkage long sys32_io_getevents(aio_context_t ctx_id,
 	set_fs(KERNEL_DS); 
 	ret = sys_io_getevents(ctx_id,min_nr,nr,events,timeout ? &t : NULL); 
 	set_fs(oldfs); 
-	if (timeout && put_compat_timespec(&t, timeout))
+	if (!ret && timeout && put_compat_timespec(&t, timeout))
 		return -EFAULT; 		
 	return ret;
 } 
@@ -2172,12 +2190,8 @@ asmlinkage long sys32_open(const char * filename, int flags, int mode)
 
 long sys32_vm86_warning(void)
 { 
-	static long warn_time = -(60*HZ); 
-	if (time_before(warn_time + 60*HZ,jiffies)) { 
 		printk(KERN_INFO "%s: vm86 mode not supported on 64 bit kernel\n",
 		       current->comm);
-		warn_time = jiffies;
-	} 
 	return -ENOSYS ;
 } 
 

arch/x86_64/kernel/Makefile

@@ -10,14 +10,13 @@ obj-y	:= process.o semaphore.o signal.o entry.o traps.o irq.o \
 		setup64.o bluesmoke.o bootflag.o e820.o reboot.o
 
 obj-$(CONFIG_MTRR)	+= mtrr/
+obj-$(CONFIG_ACPI)	+= acpi/
 obj-$(CONFIG_X86_MSR)	+= msr.o
 obj-$(CONFIG_X86_CPUID)	+= cpuid.o
 obj-$(CONFIG_SMP)	+= smp.o smpboot.o trampoline.o
 obj-$(CONFIG_X86_LOCAL_APIC)	+= apic.o  nmi.o
 obj-$(CONFIG_X86_IO_APIC)	+= io_apic.o mpparse.o
 obj-$(CONFIG_SOFTWARE_SUSPEND)	+= suspend.o suspend_asm.o
-obj-$(CONFIG_ACPI)             += acpi.o
-obj-$(CONFIG_ACPI_SLEEP)       += wakeup.o
 obj-$(CONFIG_EARLY_PRINTK)    += early_printk.o
 obj-$(CONFIG_GART_IOMMU) += pci-gart.o aperture.o
 obj-$(CONFIG_DUMMY_IOMMU) += pci-nommu.o

arch/x86_64/kernel/aperture.c

@@ -105,7 +105,8 @@ void __init iommu_hole_init(void)
 	if (!fix && !fallback_aper_force) 
 		return; 
 
-	printk("Your BIOS is broken and doesn't leave a aperture memory hole\n");
+	printk("Your BIOS doesn't leave a aperture memory hole\n");
+	printk("Please enable the IOMMU option in the BIOS setup\n"); 
 	aper_alloc = allocate_aperture(); 
 	if (!aper_alloc) 
 		return; 

arch/x86_64/kernel/apic.c

@@ -408,7 +408,7 @@ void __init setup_local_APIC (void)
 		if (maxlvt > 3)		/* Due to the Pentium erratum 3AP. */
 			apic_write(APIC_ESR, 0);
 		value = apic_read(APIC_ESR);
-		printk("ESR value before enabling vector: %08x\n", value);
+		Dprintk("ESR value before enabling vector: %08x\n", value);
 
 		value = ERROR_APIC_VECTOR;      // enables sending errors
 		apic_write_around(APIC_LVTERR, value);
@@ -418,7 +418,7 @@ void __init setup_local_APIC (void)
 		if (maxlvt > 3)
 			apic_write(APIC_ESR, 0);
 		value = apic_read(APIC_ESR);
-		printk("ESR value after enabling vector: %08x\n", value);
+		Dprintk("ESR value after enabling vector: %08x\n", value);
 	} else {
 		if (esr_disable)	
 			/* 
@@ -1080,9 +1080,10 @@ int __init APIC_init_uniprocessor (void)
 	if (nmi_watchdog == NMI_LOCAL_APIC)
 		check_nmi_watchdog();
 #ifdef CONFIG_X86_IO_APIC
-	if (smp_found_config)
-		if (!skip_ioapic_setup && nr_ioapics)
+	if (smp_found_config && !skip_ioapic_setup && nr_ioapics)
 			setup_IO_APIC();
+	else
+		nr_ioapics = 0;
 #endif
 	setup_boot_APIC_clock();
 

arch/x86_64/kernel/entry.S

@@ -359,8 +359,13 @@ ENTRY(stub_rt_sigreturn)
 /* 0(%rsp): interrupt number */ 
 	.macro interrupt func
 	cld
+#ifdef CONFIG_X86_REMOTE_DEBUG
+	SAVE_ALL	
+	movq %rsp,%rdi
+#else		
 	SAVE_ARGS
 	leaq -ARGOFFSET(%rsp),%rdi	# arg1 for handler
+#endif	
 	testl $3,CS(%rdi)
 	je 1f
 	swapgs	
@@ -667,18 +672,6 @@ ENTRY(execve)
 	ret
 
 ENTRY(page_fault)
-#ifdef CONFIG_KDB
-        pushq %rcx
-        pushq %rdx
-        pushq %rax
-        movl  $473,%ecx
-        rdmsr
-        andl  $0xfffffffe,%eax          /* Disable last branch recording */
-        wrmsr
-        popq  %rax
-        popq  %rdx
-        popq  %rcx
-#endif 
 	errorentry do_page_fault
 
 ENTRY(coprocessor_error)

arch/x86_64/kernel/head.S

@@ -194,7 +194,7 @@ ENTRY(no_long_mode)
 	jmp	1b
 
 .org 0xf00
-pGDT32:
+ENTRY(pGDT32):
 	.word	gdt32_end-gdt_table32
 	.long	gdt_table32-__START_KERNEL_map
 
@@ -307,6 +307,15 @@ ENTRY(level3_physmem_pgt)
 	.quad	0x0000000000105007		/* -> level2_kernel_pgt (so that __va works even before pagetable_init) */
 
 .org 0xb000
+ENTRY(wakeup_level4_pgt)
+	.quad	0x0000000000102007		/* -> level3_ident_pgt */
+	.fill	255,8,0
+	.quad	0x000000000010a007
+	.fill	254,8,0
+	/* (2^48-(2*1024*1024*1024))/(2^39) = 511 */
+	.quad	0x0000000000103007		/* -> level3_kernel_pgt */
+
+.org 0xc000
 .data
 
 	.align 16
@@ -371,3 +380,4 @@ ENTRY(idt_table)
 	.quad   0
 	.quad 	0
 	.endr
+

arch/x86_64/kernel/head64.c

@@ -15,6 +15,7 @@
 #include <asm/processor.h>
 #include <asm/proto.h>
 #include <asm/smp.h>
+#include <asm/bootsetup.h>
 
 /* Don't add a printk in there. printk relies on the PDA which is not initialized 
    yet. */
@@ -51,7 +52,7 @@ static void __init copy_bootdata(char *real_mode_data)
 		printk("old bootloader convention, maybe loadlin?\n");
 	}
 	command_line = (char *) ((u64)(new_data));
-	memcpy(saved_command_line, command_line, 2048);
+	memcpy(saved_command_line, command_line, COMMAND_LINE_SIZE);
 	printk("Bootdata ok (command line is %s)\n", saved_command_line);	
 }
 

arch/x86_64/kernel/i387.c

@@ -42,7 +42,7 @@ void __init fpu_init(void)
 
 	/* clean state in init */
 	stts();
-	clear_thread_flag(TIF_USEDFPU);
+	current_thread_info()->status = 0;
 	current->used_math = 0;
 }
 
@@ -51,13 +51,12 @@ void __init fpu_init(void)
  * so initialize it and set the mxcsr to its default.
  * remeber the current task has used the FPU.
  */
-void init_fpu(void)
+void init_fpu(struct task_struct *child)
 {
-	struct task_struct *me = current;
-	memset(&me->thread.i387.fxsave, 0, sizeof(struct i387_fxsave_struct));
-	me->thread.i387.fxsave.cwd = 0x37f;
-	me->thread.i387.fxsave.mxcsr = 0x1f80;
-	me->used_math = 1;
+	memset(&child->thread.i387.fxsave, 0, sizeof(struct i387_fxsave_struct));
+	child->thread.i387.fxsave.cwd = 0x37f;
+	child->thread.i387.fxsave.mxcsr = 0x1f80;
+	child->used_math = 1;
 }
 
 /*
@@ -81,7 +80,7 @@ int save_i387(struct _fpstate *buf)
 	if (!tsk->used_math) 
 		return 0;
 	tsk->used_math = 0; /* trigger finit */ 
-	if (test_thread_flag(TIF_USEDFPU)) { 
+	if (tsk->thread_info->status & TS_USEDFPU) {
 		err = save_i387_checking((struct i387_fxsave_struct *)buf);
 		if (err) return err;
 		stts();
@@ -99,7 +98,7 @@ int save_i387(struct _fpstate *buf)
 
 int get_fpregs(struct user_i387_struct *buf, struct task_struct *tsk)
 {
-	empty_fpu(tsk);
+	init_fpu(tsk);
 	return __copy_to_user((void *)buf, &tsk->thread.i387.fxsave,
 			       sizeof(struct user_i387_struct)) ? -EFAULT : 0;
 }

arch/x86_64/kernel/nmi.c

@@ -25,13 +25,15 @@
 #include <asm/mpspec.h>
 #include <asm/nmi.h>
 #include <asm/msr.h>
+#include <asm/proto.h>
+#include <asm/kdebug.h>
 
 extern void default_do_nmi(struct pt_regs *);
 
 unsigned int nmi_watchdog = NMI_LOCAL_APIC;
 static unsigned int nmi_hz = HZ;
 unsigned int nmi_perfctr_msr;	/* the MSR to reset in NMI handler */
-extern void show_registers(struct pt_regs *regs);
+int nmi_watchdog_disabled;
 
 #define K7_EVNTSEL_ENABLE	(1 << 22)
 #define K7_EVNTSEL_INT		(1 << 20)
@@ -251,15 +253,13 @@ void touch_nmi_watchdog (void)
 		alert_counter[i] = 0;
 }
 
-void nmi_watchdog_tick (struct pt_regs * regs)
+void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason)
 {
+	if (nmi_watchdog_disabled)
+		return;
+
+	int sum, cpu = safe_smp_processor_id();
 
-	/*
-	 * Since current_thread_info()-> is always on the stack, and we
-	 * always switch the stack NMI-atomically, it's safe to use
-	 * smp_processor_id().
-	 */
-	int sum, cpu = smp_processor_id();
 	sum = read_pda(apic_timer_irqs);
 
 	if (last_irq_sums[cpu] == sum) {
@@ -269,6 +269,10 @@ void nmi_watchdog_tick (struct pt_regs * regs)
 		 */
 		alert_counter[cpu]++;
 		if (alert_counter[cpu] == 5*nmi_hz) {
+			if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_BAD) { 
+				alert_counter[cpu] = 0; 
+				return;
+			} 
 			spin_lock(&nmi_print_lock);
 			/*
 			 * We are in trouble anyway, lets at least try

arch/x86_64/kernel/pci-gart.c

@@ -8,7 +8,7 @@
  * See Documentation/DMA-mapping.txt for the interface specification.
  * 
  * Copyright 2002 Andi Kleen, SuSE Labs.
- * $Id: pci-gart.c,v 1.12 2002/09/19 19:25:32 ak Exp $
+ * $Id: pci-gart.c,v 1.20 2003/03/12 08:23:29 ak Exp $
  */
 
 /* 
@@ -19,9 +19,12 @@ agpgart_be
 
 possible future tuning: 
  fast path for sg streaming mappings 
- more intelligent flush strategy - flush only a single NB?
+ more intelligent flush strategy - flush only a single NB? flush only when
+ gart area fills up and alloc_iommu wraps. 
+ don't flush on allocation - need to unmap the gart area first to avoid prefetches
+ by the CPU
  move boundary between IOMMU and AGP in GART dynamically
- could use exact fit in the gart in alloc_consistent, not order of two.
+  
 */ 
 
 #include <linux/config.h>
@@ -49,7 +52,11 @@ u32 *iommu_gatt_base; 		/* Remapping table */
 
 int no_iommu; 
 static int no_agp; 
+#ifdef CONFIG_IOMMU_DEBUG
 int force_mmu = 1;
+#else
+int force_mmu = 0;
+#endif
 
 extern int fallback_aper_order;
 extern int fallback_aper_force;
@@ -58,10 +65,9 @@ extern int fallback_aper_force;
 static spinlock_t iommu_bitmap_lock = SPIN_LOCK_UNLOCKED;
 static unsigned long *iommu_gart_bitmap; /* guarded by iommu_bitmap_lock */
 
-#define GPTE_MASK 0xfffffff000
 #define GPTE_VALID    1
 #define GPTE_COHERENT 2
-#define GPTE_ENCODE(x,flag) (((x) & 0xfffffff0) | ((x) >> 28) | GPTE_VALID | (flag))
+#define GPTE_ENCODE(x) (((x) & 0xfffff000) | (((x) >> 32) << 4) | GPTE_VALID | GPTE_COHERENT)
 #define GPTE_DECODE(x) (((x) & 0xfffff000) | (((u64)(x) & 0xff0) << 28))
 
 #define for_all_nb(dev) \
@@ -72,7 +78,6 @@ static unsigned long *iommu_gart_bitmap; /* guarded by iommu_bitmap_lock */
 #define EMERGENCY_PAGES 32 /* = 128KB */ 
 
 #ifdef CONFIG_AGP
-extern int agp_amdk8_init(void);
 extern int agp_init(void);
 #define AGPEXTERN extern
 #else
@@ -130,7 +135,7 @@ void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size,
 {
 	void *memory;
 	int gfp = GFP_ATOMIC;
-	int order, i;
+	int i;
 	unsigned long iommu_page;
 
 	if (hwdev == NULL || hwdev->dma_mask < 0xffffffff || no_iommu)
@@ -140,15 +145,15 @@ void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size,
 	 * First try to allocate continuous and use directly if already 
 	 * in lowmem. 
 	 */ 
-	order = get_order(size);
-	memory = (void *)__get_free_pages(gfp, order);
+	size = round_up(size, PAGE_SIZE); 
+	memory = (void *)__get_free_pages(gfp, get_order(size));
 	if (memory == NULL) {
 		return NULL; 
 	} else {
 		int high = (unsigned long)virt_to_bus(memory) + size
 			>= 0xffffffff;
 		int mmu = high;
-		if (force_mmu) 
+		if (force_mmu && !(gfp & GFP_DMA)) 
 			mmu = 1;
 		if (no_iommu) { 
 			if (high) goto error;
@@ -161,19 +166,21 @@ void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size,
 		}
 	} 
 
-	iommu_page = alloc_iommu(1<<order);
+	size >>= PAGE_SHIFT;
+
+	iommu_page = alloc_iommu(size);
 	if (iommu_page == -1)
 		goto error; 
 
    	/* Fill in the GATT, allocating pages as needed. */
-	for (i = 0; i < 1<<order; i++) { 
+	for (i = 0; i < size; i++) { 
 		unsigned long phys_mem; 
 		void *mem = memory + i*PAGE_SIZE;
 		if (i > 0) 
 			atomic_inc(&virt_to_page(mem)->count); 
 		phys_mem = virt_to_phys(mem); 
-		BUG_ON(phys_mem & ~PTE_MASK); 
-		iommu_gatt_base[iommu_page + i] = GPTE_ENCODE(phys_mem,GPTE_COHERENT); 
+		BUG_ON(phys_mem & ~PHYSICAL_PAGE_MASK); 
+		iommu_gatt_base[iommu_page + i] = GPTE_ENCODE(phys_mem); 
 	} 
 
 	flush_gart();
@@ -181,7 +188,7 @@ void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size,
 	return memory; 
 	
  error:
-	free_pages((unsigned long)memory, order); 
+	free_pages((unsigned long)memory, get_order(size)); 
 	return NULL; 
 }
 
@@ -193,30 +200,32 @@ void pci_free_consistent(struct pci_dev *hwdev, size_t size,
 			 void *vaddr, dma_addr_t bus)
 {
 	u64 pte;
-	int order = get_order(size);
 	unsigned long iommu_page;
 	int i;
 
+	size = round_up(size, PAGE_SIZE); 
 	if (bus < iommu_bus_base || bus > iommu_bus_base + iommu_size) { 
-		free_pages((unsigned long)vaddr, order); 		
+		free_pages((unsigned long)vaddr, get_order(size)); 		
 		return;
 	} 
+	size >>= PAGE_SHIFT;
 	iommu_page = (bus - iommu_bus_base) / PAGE_SIZE;
-	for (i = 0; i < 1<<order; i++) {
+	for (i = 0; i < size; i++) {
 		pte = iommu_gatt_base[iommu_page + i];
 		BUG_ON((pte & GPTE_VALID) == 0); 
 		iommu_gatt_base[iommu_page + i] = 0; 		
 		free_page((unsigned long) __va(GPTE_DECODE(pte)));
 	} 
 	flush_gart(); 
-	free_iommu(iommu_page, 1<<order);
+	free_iommu(iommu_page, size);
 }
 
 #ifdef CONFIG_IOMMU_LEAK
 /* Debugging aid for drivers that don't free their IOMMU tables */
 static void **iommu_leak_tab; 
 static int leak_trace;
-int iommu_leak_dumppages = 20; 
+int iommu_leak_pages = 20; 
+extern unsigned long printk_address(unsigned long);
 void dump_leak(void)
 {
 	int i;
@@ -224,10 +233,13 @@ void dump_leak(void)
 	if (dump || !iommu_leak_tab) return;
 	dump = 1;
 	show_stack(NULL);
-	printk("Dumping %d pages from end of IOMMU:\n", iommu_leak_dumppages); 
-	for (i = 0; i < iommu_leak_dumppages; i++) 
-		printk("[%lu: %lx] ",
-		       iommu_pages-i,(unsigned long) iommu_leak_tab[iommu_pages-i]); 
+	/* Very crude. dump some from the end of the table too */ 
+	printk("Dumping %d pages from end of IOMMU:\n", iommu_leak_pages); 
+	for (i = 0; i < iommu_leak_pages; i+=2) {
+		printk("%lu: ", iommu_pages-i);
+		printk_address((unsigned long) iommu_leak_tab[iommu_pages-i]);
+		printk("%c", (i+1)%2 == 0 ? '\n' : ' '); 
+	} 
 	printk("\n");
 }
 #endif
@@ -275,7 +287,8 @@ static inline int need_iommu(struct pci_dev *dev, unsigned long addr, size_t siz
 	return mmu; 
 }
 
-dma_addr_t pci_map_single(struct pci_dev *dev, void *addr, size_t size,int dir)
+dma_addr_t __pci_map_single(struct pci_dev *dev, void *addr, size_t size,
+			    int dir, int flush)
 { 
 	unsigned long iommu_page;
 	unsigned long phys_mem, bus;
@@ -297,13 +310,18 @@ dma_addr_t pci_map_single(struct pci_dev *dev, void *addr, size_t size,int dir)
 
 	phys_mem &= PAGE_MASK;
 	for (i = 0; i < npages; i++, phys_mem += PAGE_SIZE) {
-		BUG_ON(phys_mem & ~PTE_MASK); 
+		BUG_ON(phys_mem & ~PHYSICAL_PAGE_MASK); 
 		
 		/* 
 		 * Set coherent mapping here to avoid needing to flush
 		 * the caches on mapping.
 		 */
-		iommu_gatt_base[iommu_page + i] = GPTE_ENCODE(phys_mem, GPTE_COHERENT);
+		iommu_gatt_base[iommu_page + i] = GPTE_ENCODE(phys_mem);
+
+#ifdef CONFIG_IOMMU_DEBUG
+		/* paranoia check */
+		BUG_ON(GPTE_DECODE(iommu_gatt_base[iommu_page+i]) != phys_mem); 
+#endif
 
 #ifdef CONFIG_IOMMU_LEAK
 		/* XXX need eventually caller of pci_map_sg */
@@ -311,6 +329,7 @@ dma_addr_t pci_map_single(struct pci_dev *dev, void *addr, size_t size,int dir)
 			iommu_leak_tab[iommu_page + i] = __builtin_return_address(0); 
 #endif
 	}
+	if (flush)
 	flush_gart(); 
 
 	bus = iommu_bus_base + iommu_page*PAGE_SIZE; 
@@ -341,7 +360,7 @@ void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t dma_addr,
 	free_iommu(iommu_page, npages);
 }
 
-EXPORT_SYMBOL(pci_map_single);
+EXPORT_SYMBOL(__pci_map_single);
 EXPORT_SYMBOL(pci_unmap_single);
 
 static __init unsigned long check_iommu_size(unsigned long aper, u64 aper_size)
@@ -438,8 +457,6 @@ static __init int init_k8_gatt(agp_kern_info *info)
 	}
 	flush_gart(); 
 	
-	global_flush_tlb();
-		
 	printk("PCI-DMA: aperture base @ %x size %u KB\n", aper_base, aper_size>>10); 
 	return 0;
 
@@ -530,8 +547,10 @@ void __init pci_iommu_init(void)
    off   don't use the IOMMU
    leak  turn on simple iommu leak tracing (only when CONFIG_IOMMU_LEAK is on)
    memaper[=order] allocate an own aperture over RAM with size 32MB^order.  
+   noforce don't force IOMMU usage. Should be fastest.
+   force  Force IOMMU and turn on unmap debugging.
 */
-__init int iommu_setup(char *opt, char **end) 
+__init int iommu_setup(char *opt) 
 { 
     int arg;
     char *p = opt;
@@ -552,17 +571,21 @@ __init int iommu_setup(char *opt, char **end)
 			    fallback_aper_order = arg;
 	    } 
 #ifdef CONFIG_IOMMU_LEAK
-	    if (!memcmp(p,"leak", 4))
+	    if (!memcmp(p,"leak", 4)) { 
 		    leak_trace = 1;
+		    p += 4; 
+		    if (*p == '=') ++p;
+		    if (isdigit(*p) && get_option(&p, &arg))
+			    iommu_leak_pages = arg;
+	    } else
 #endif
 	    if (isdigit(*p) && get_option(&p, &arg)) 
 		    iommu_size = arg;
 	    do {
-		    if (*p == ' ' || *p == 0) { 
-			    *end = p; 
+		    if (*p == ' ' || *p == 0) 
 			    return 0; 
-		    }
 	    } while (*p++ != ','); 
     }
+    return 1;
 } 
 

arch/x86_64/kernel/process.c

@@ -192,6 +192,8 @@ void show_regs(struct pt_regs * regs)
 	       fs,fsindex,gs,gsindex,shadowgs); 
 	printk("CS:  %04x DS: %04x ES: %04x CR0: %016lx\n", cs, ds, es, cr0); 
 	printk("CR2: %016lx CR3: %016lx CR4: %016lx\n", cr2, cr3, cr4);
+
+	show_trace(&regs->rsp);
 }
 
 extern void load_gs_index(unsigned);
@@ -260,6 +262,14 @@ static inline u32 read_32bit_tls(struct task_struct *t, int tls)
 		(((u32)desc->base2) << 24);
 }
 
+/*
+ * This gets called before we allocate a new thread and copy
+ * the current task into it.
+ */
+void prepare_to_copy(struct task_struct *tsk)
+{
+	unlazy_fpu(tsk);
+}
 
 int copy_thread(int nr, unsigned long clone_flags, unsigned long rsp, 
 		unsigned long unused,
@@ -294,9 +304,6 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long rsp,
 	asm("movl %%es,%0" : "=m" (p->thread.es));
 	asm("movl %%ds,%0" : "=m" (p->thread.ds));
 
-	unlazy_fpu(me);	
-	p->thread.i387 = me->thread.i387;
-
 	if (unlikely(me->thread.io_bitmap_ptr != NULL)) { 
 		p->thread.io_bitmap_ptr = kmalloc((IO_BITMAP_SIZE+1)*4, GFP_KERNEL);
 		if (!p->thread.io_bitmap_ptr) 
@@ -314,7 +321,7 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long rsp,
 			err = ia32_child_tls(p, childregs); 
 		else 			
 #endif	 
-			err = do_arch_prctl(p, ARCH_SET_FS, childregs->r10); 
+			err = do_arch_prctl(p, ARCH_SET_FS, childregs->r8); 
 		if (err) 
 			goto out;
 	}

arch/x86_64/kernel/ptrace.c

@@ -240,8 +240,8 @@ asmlinkage long sys_ptrace(long request, long pid, long addr, long data)
 		unsigned long tmp;
 
 		ret = -EIO;
-		if ((addr & 3) || addr < 0 || 
-		    addr > sizeof(struct user) - 3)
+		if ((addr & 7) || addr < 0 || 
+		    addr > sizeof(struct user) - 7)
 			break;
 
 		tmp = 0;  /* Default return condition */
@@ -250,7 +250,7 @@ asmlinkage long sys_ptrace(long request, long pid, long addr, long data)
 		if(addr >= (long) &dummy->u_debugreg[0] &&
 		   addr <= (long) &dummy->u_debugreg[7]){
 			addr -= (long) &dummy->u_debugreg[0];
-			addr = addr >> 2;
+			addr = addr >> 3;
 			tmp = child->thread.debugreg[addr];
 		}
 		ret = put_user(tmp,(unsigned long *) data);
@@ -268,8 +268,8 @@ asmlinkage long sys_ptrace(long request, long pid, long addr, long data)
 
 	case PTRACE_POKEUSR: /* write the word at location addr in the USER area */
 		ret = -EIO;
-		if ((addr & 3) || addr < 0 || 
-		    addr > sizeof(struct user) - 3)
+		if ((addr & 7) || addr < 0 || 
+		    addr > sizeof(struct user) - 7)
 			break;
 
 		if (addr < sizeof(struct user_regs_struct)) {
@@ -290,6 +290,11 @@ asmlinkage long sys_ptrace(long request, long pid, long addr, long data)
 			  if(addr < (long) &dummy->u_debugreg[4] &&
 			     ((unsigned long) data) >= TASK_SIZE-3) break;
 			  
+			  if (addr == (long) &dummy->u_debugreg[6]) {
+				  if (data >> 32)
+					  goto out_tsk;
+			  }
+
 			  if(addr == (long) &dummy->u_debugreg[7]) {
 				  data &= ~DR_CONTROL_RESERVED;
 				  for(i=0; i<4; i++)
@@ -298,7 +303,7 @@ asmlinkage long sys_ptrace(long request, long pid, long addr, long data)
 			  }
 
 			  addr -= (long) &dummy->u_debugreg;
-			  addr = addr >> 2;
+			  addr = addr >> 3;
 			  child->thread.debugreg[addr] = data;
 			  ret = 0;
 		  }

arch/x86_64/kernel/reboot.c

@@ -101,7 +101,6 @@ void machine_restart(char * __unused)
 	 * Stop all CPUs and turn off local APICs and the IO-APIC, so
 	 * other OSs see a clean IRQ state.
 	 */
-	if (notify_die(DIE_STOP,"cpustop",0,0) != NOTIFY_BAD)
 	smp_send_stop();
 	disable_IO_APIC();
 #endif

arch/x86_64/kernel/setup.c

@@ -197,12 +197,15 @@ static __init void parse_cmdline_early (char ** cmdline_p)
 		if (!memcmp(from, "acpi=off", 8))
 			acpi_disabled = 1;
 
+		if (!memcmp(from, "disableapic", 11))
+			disable_apic = 1;
+
 		if (!memcmp(from, "mem=", 4))
 			parse_memopt(from+4, &from); 
 
 #ifdef CONFIG_GART_IOMMU 
 		if (!memcmp(from,"iommu=",6)) { 
-			iommu_setup(from+6, &from); 
+			iommu_setup(from+6); 
 		}
 #endif
 

arch/x86_64/kernel/setup64.c

@@ -40,7 +40,7 @@ struct desc_ptr idt_descr = { 256 * 16, (unsigned long) idt_table };
 char boot_cpu_stack[IRQSTACKSIZE] __cacheline_aligned;
 
 unsigned long __supported_pte_mask = ~0UL;
-static int do_not_nx = 0;
+static int do_not_nx = 1;
 
 static int __init nonx_setup(char *str)
 {
@@ -98,6 +98,8 @@ void pda_init(int cpu)
 	pda->cpudata_offset = 0;
 	pda->kernelstack = 
 		(unsigned long)stack_thread_info() - PDA_STACKOFFSET + THREAD_SIZE; 
+	pda->active_mm = &init_mm;
+	pda->mmu_state = 0;
 
 	if (cpu == 0) {
 		/* others are initialized in smpboot.c */
@@ -121,8 +123,6 @@ void pda_init(int cpu)
 	asm volatile("movq %0,%%cr3" :: "r" (__pa(level4))); 
 
 	pda->irqstackptr += IRQSTACKSIZE-64;
-	pda->active_mm = &init_mm;
-	pda->mmu_state = 0;
 } 
 
 #define EXCEPTION_STK_ORDER 0 /* >= N_EXCEPTION_STACKS*EXCEPTION_STKSZ */

arch/x86_64/kernel/signal.c

@@ -225,6 +225,7 @@ get_stack(struct k_sigaction *ka, struct pt_regs *regs, unsigned long size)
 	rsp = regs->rsp - 128;
 
 	/* This is the X/Open sanctioned signal stack switching.  */
+	/* RED-PEN: redzone on that stack? */
 	if (ka->sa.sa_flags & SA_ONSTACK) {
 		if (sas_ss_flags(rsp) == 0)
 			rsp = current->sas_ss_sp + current->sas_ss_size;
@@ -433,7 +434,8 @@ int do_signal(struct pt_regs *regs, sigset_t *oldset)
 		 * have been cleared if the watchpoint triggered
 		 * inside the kernel.
 		 */
-		__asm__("movq %0,%%db7"	: : "r" (current->thread.debugreg[7]));
+		if (current->thread.debugreg[7])
+			asm volatile("movq %0,%%db7"	: : "r" (current->thread.debugreg[7]));
 
 		/* Whee!  Actually deliver the signal.  */
 		handle_signal(signr, &info, oldset, regs);

arch/x86_64/kernel/smp.c

@@ -485,16 +485,16 @@ asmlinkage void smp_call_function_interrupt(void)
 	}
 }
 
-
 /* Slow. Should be only used for debugging. */
 int slow_smp_processor_id(void)
 { 
 	int stack_location;
 	unsigned long sp = (unsigned long)&stack_location; 
-	int cpu;
-	unsigned long mask;
+	int offset = 0, cpu;
+
+	for (offset = 0; (cpu_online_map >> offset); offset = cpu + 1) { 
+		cpu = ffz(~(cpu_online_map >> offset));
 
-	for_each_cpu(cpu, mask) { 
 		if (sp >= (u64)cpu_pda[cpu].irqstackptr - IRQSTACKSIZE && 
 		    sp <= (u64)cpu_pda[cpu].irqstackptr)
 			return cpu;

arch/x86_64/kernel/smpboot.c

@@ -318,8 +318,6 @@ void __init smp_callin(void)
 	 */
  	smp_store_cpu_info(cpuid);
 
-	notify_die(DIE_CPUINIT, "cpuinit", NULL, 0);
-
 	local_irq_disable();
 
 	/*
@@ -898,6 +896,8 @@ static void __init smp_boot_cpus(unsigned int max_cpus)
 	 */
 	if (!skip_ioapic_setup && nr_ioapics)
 		setup_IO_APIC();
+	else
+		nr_ioapics = 0;
 
 	setup_boot_APIC_clock();
 

arch/x86_64/kernel/suspend.c

@@ -28,7 +28,7 @@
 #include <asm/io.h>
 #include <asm/proto.h>
 
-static struct saved_context saved_context;
+struct saved_context saved_context;
 
 unsigned long saved_context_eax, saved_context_ebx, saved_context_ecx, saved_context_edx;
 unsigned long saved_context_esp, saved_context_ebp, saved_context_esi, saved_context_edi;

arch/x86_64/kernel/traps.c

@@ -110,7 +110,7 @@ int printk_address(unsigned long address)
 } 
 #endif
 
-static inline unsigned long *in_exception_stack(int cpu, unsigned long stack) 
+unsigned long *in_exception_stack(int cpu, unsigned long stack) 
 { 
 	int k;
 	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
@@ -249,7 +249,7 @@ void show_stack(unsigned long * rsp)
 void dump_stack(void)
 {
 	unsigned long dummy;
-	show_stack(&dummy);
+	show_trace(&dummy);
 }
 
 void show_registers(struct pt_regs *regs)
@@ -344,7 +344,6 @@ void die(const char * str, struct pt_regs * regs, long err)
 	show_registers(regs);
 	bust_spinlocks(0);
 	spin_unlock_irq(&die_lock);
-	notify_die(DIE_OOPS, (char *)str, regs, err);
 	do_exit(SIGSEGV);
 }
 
@@ -419,6 +418,8 @@ static void do_trap(int trapnr, int signr, char *str,
 #define DO_ERROR(trapnr, signr, str, name) \
 asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
 { \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) == NOTIFY_BAD) \
+		return; \
 	do_trap(trapnr, signr, str, regs, error_code, NULL); \
 }
 
@@ -430,10 +431,13 @@ asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
 	info.si_errno = 0; \
 	info.si_code = sicode; \
 	info.si_addr = (void *)siaddr; \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) == NOTIFY_BAD) \
+		return; \
 	do_trap(trapnr, signr, str, regs, error_code, &info); \
 }
 
 DO_ERROR_INFO( 0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->rip)
+DO_ERROR( 3, SIGTRAP, "int3", int3);
 DO_ERROR( 4, SIGSEGV, "overflow", overflow)
 DO_ERROR( 5, SIGSEGV, "bounds", bounds)
 DO_ERROR_INFO( 6, SIGILL,  "invalid operand", invalid_op, ILL_ILLOPN, regs->rip)
@@ -446,13 +450,6 @@ DO_ERROR(12, SIGBUS,  "stack segment", stack_segment)
 DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, get_cr2())
 DO_ERROR(18, SIGSEGV, "reserved", reserved)
 
-asmlinkage void do_int3(struct pt_regs * regs, long error_code)
-{
-	if (notify_die(DIE_INT3, "int3", regs, error_code) == NOTIFY_BAD)
-		return;
-	do_trap(3, SIGTRAP, "int3", regs, error_code, NULL);
-}
-
 extern void dump_pagetable(unsigned long);
 
 asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
@@ -493,6 +490,8 @@ asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
 			regs->rip = fixup->fixup;
 			return;
 		}
+		notify_die(DIE_GPF, "general protection fault", regs, error_code,
+			   13, SIGSEGV); 
 		die("general protection fault", regs, error_code);
 	}
 }
@@ -537,14 +536,14 @@ asmlinkage void default_do_nmi(struct pt_regs * regs)
 		 * so it must be the NMI watchdog.
 		 */
 		if (nmi_watchdog) {
-			nmi_watchdog_tick(regs);
+			nmi_watchdog_tick(regs,reason);
 			return;
 		}
 #endif
 		unknown_nmi_error(reason, regs);
 		return;
 	}
-	if (notify_die(DIE_NMI, "nmi", regs, reason) == NOTIFY_BAD)
+	if (notify_die(DIE_NMI, "nmi", regs, reason, 0, SIGINT) == NOTIFY_BAD)
 		return; 
 	if (reason & 0x80)
 		mem_parity_error(reason, regs);
@@ -569,6 +568,7 @@ asmlinkage void do_debug(struct pt_regs * regs, long error_code)
 
 #ifdef CONFIG_CHECKING
        { 
+	       /* RED-PEN interaction with debugger - could destroy gs */
                unsigned long gs; 
                struct x8664_pda *pda = cpu_pda + safe_smp_processor_id(); 
                rdmsrl(MSR_GS_BASE, gs); 
@@ -583,9 +583,6 @@ asmlinkage void do_debug(struct pt_regs * regs, long error_code)
 
 	conditional_sti(regs);
 
-	if (notify_die(DIE_DEBUG, "debug", regs, error_code) == NOTIFY_BAD)
-		return; 
-
 	/* Mask out spurious debug traps due to lazy DR7 setting */
 	if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) {
 		if (!tsk->thread.debugreg[7]) { 
@@ -618,17 +615,22 @@ asmlinkage void do_debug(struct pt_regs * regs, long error_code)
 	info.si_signo = SIGTRAP;
 	info.si_errno = 0;
 	info.si_code = TRAP_BRKPT;
-	info.si_addr = ((regs->cs & 3) == 0) ? (void *)tsk->thread.rip : 
-	                                        (void *)regs->rip;
+	if ((regs->cs & 3) == 0) 
+		goto clear_dr7; 
+
+	info.si_addr = (void *)regs->rip;
 	force_sig_info(SIGTRAP, &info, tsk);	
 clear_dr7:
-	asm("movq %0,%%db7"::"r"(0UL));
+	asm volatile("movq %0,%%db7"::"r"(0UL));
+	notify_die(DIE_DEBUG, "debug", regs, error_code, 1, SIGTRAP);
 	return;
 
 clear_TF_reenable:
 	set_tsk_thread_flag(tsk, TIF_SINGLESTEP);
 
 clear_TF:
+	/* RED-PEN could cause spurious errors */
+	if (notify_die(DIE_DEBUG, "debug2", regs, error_code, 1, SIGTRAP) != NOTIFY_BAD)
 	regs->eflags &= ~TF_MASK;
 	return;
 }
@@ -775,9 +777,9 @@ asmlinkage void math_state_restore(void)
 	clts();			/* Allow maths ops (or we recurse) */
 
 	if (!me->used_math)
-		init_fpu();
+		init_fpu(me);
 	restore_fpu_checking(&me->thread.i387.fxsave);
-	set_thread_flag(TIF_USEDFPU); 
+	me->thread_info->status |= TS_USEDFPU;
 }
 
 asmlinkage void math_emulate(void)
@@ -787,7 +789,7 @@ asmlinkage void math_emulate(void)
 
 void do_call_debug(struct pt_regs *regs) 
 { 
-	notify_die(DIE_CALL, "debug call", regs, 0); 
+	notify_die(DIE_CALL, "debug call", regs, 0, 255, SIGINT); 
 }
 
 void __init trap_init(void)
@@ -819,8 +821,6 @@ void __init trap_init(void)
        
 	set_intr_gate(KDB_VECTOR, call_debug);
        
-	notify_die(DIE_TRAPINIT, "traps initialized", 0, 0); 
-	
 	/*
 	 * Should be a barrier for any external CPU state.
 	 */

arch/x86_64/mm/fault.c

@@ -57,29 +57,41 @@ void bust_spinlocks(int yes)
 	}
 }
 
+static int bad_address(void *p) 
+{ 
+	unsigned long dummy;
+	return __get_user(dummy, (unsigned long *)p);
+} 
+
 void dump_pagetable(unsigned long address)
 {
-	static char *name[] = { "PML4", "PGD", "PDE", "PTE" }; 
-	int i, shift;
-	unsigned long page;
-
-	shift = 9+9+9+12;
-	address &= ~0xFFFF000000000000UL;
-	asm("movq %%cr3,%0" : "=r" (page)); 
-	for (i = 0; i < 4; i++) { 
-		unsigned long *padr = (unsigned long *) __va(page); 
-		padr += (address >> shift) & 0x1FFU;
-		if (__get_user(page, padr)) { 
-			printk("%s: bad %p\n", name[i], padr); 
-			break;
-		}
-		printk("%s: %016lx ", name[i], page); 
-		if ((page & (1 | (1<<7))) != 1) /* Not present or 2MB page */
-			break;
-		page &= ~0xFFFUL;
-		shift -= (i == 0) ? 12 : 9;
-	} 
+	pml4_t *pml4;
+	asm("movq %%cr3,%0" : "=r" (pml4));
+
+	pml4 = __va((unsigned long)pml4 & PHYSICAL_PAGE_MASK); 
+	pml4 += pml4_index(address);
+	printk("PML4 %lx ", pml4_val(*pml4));
+	if (bad_address(pml4)) goto bad;
+	if (!pml4_present(*pml4)) goto ret; 
+
+	pgd_t *pgd = __pgd_offset_k((pgd_t *)pml4_page(*pml4), address); 
+	if (bad_address(pgd)) goto bad;
+	printk("PGD %lx ", pgd_val(*pgd)); 
+	if (!pgd_present(*pgd))	goto ret;
+
+	pmd_t *pmd = pmd_offset(pgd, address); 
+	if (bad_address(pmd)) goto bad;
+	printk("PMD %lx ", pmd_val(*pmd));
+	if (!pmd_present(*pmd))	goto ret;	 
+
+	pte_t *pte = pte_offset_kernel(pmd, address);
+	if (bad_address(pte)) goto bad;
+	printk("PTE %lx", pte_val(*pte)); 
+ret:
 	printk("\n");
+	return;
+bad:
+	printk("BAD\n");
 }
 
 int page_fault_trace; 

arch/x86_64/mm/ioremap.c

@@ -150,7 +150,7 @@ void * __ioremap(unsigned long phys_addr, unsigned long size, unsigned long flag
 	 */
 	offset = phys_addr & ~PAGE_MASK;
 	phys_addr &= PAGE_MASK;
-	size = PAGE_ALIGN(last_addr) - phys_addr;
+	size = PAGE_ALIGN(last_addr+1) - phys_addr;
 
 	/*
 	 * Ok, go for it..

arch/x86_64/mm/k8topology.c

@@ -21,7 +21,7 @@
 #include <asm/pci-direct.h>
 #include <asm/numa.h>
 
-static int find_northbridge(void)
+static __init int find_northbridge(void)
 {
 	int num; 
 
@@ -45,7 +45,8 @@ int __init k8_scan_nodes(unsigned long start, unsigned long end)
 { 
 	unsigned long prevbase;
 	struct node nodes[MAXNODE];
-	int nodeid, numnodes, maxnode, i, nb; 
+	int nodeid, i, nb; 
+	int found = 0;
 
 	nb = find_northbridge(); 
 	if (nb < 0) 
@@ -53,12 +54,13 @@ int __init k8_scan_nodes(unsigned long start, unsigned long end)
 
 	printk(KERN_INFO "Scanning NUMA topology in Northbridge %d\n", nb); 
 
-	numnodes = (read_pci_config(0, nb, 0, 0x60 ) >> 4) & 3; 
+	numnodes = (1 << ((read_pci_config(0, nb, 0, 0x60 ) >> 4) & 3)); 
+
+	printk(KERN_INFO "Assuming %d nodes\n", numnodes - 1); 
 
 	memset(&nodes,0,sizeof(nodes)); 
 	prevbase = 0;
-	maxnode = -1; 
-	for (i = 0; i < MAXNODE; i++) { 
+	for (i = 0; i < numnodes; i++) { 
 		unsigned long base,limit; 
 
 		base = read_pci_config(0, nb, 1, 0x40 + i*8);
@@ -66,18 +68,16 @@ int __init k8_scan_nodes(unsigned long start, unsigned long end)
 
 		nodeid = limit & 3; 
 		if (!limit) { 
-			printk(KERN_INFO "Skipping node entry %d (base %lx)\n", i,			       base);
-			continue;
+			printk(KERN_ERR "Skipping node entry %d (base %lx)\n", i,			       base);
+			return -1;
 		}
 		if ((base >> 8) & 3 || (limit >> 8) & 3) {
 			printk(KERN_ERR "Node %d using interleaving mode %lx/%lx\n", 
 			       nodeid, (base>>8)&3, (limit>>8) & 3); 
 			return -1; 
 		}	
-		if (nodeid > maxnode) 
-			maxnode = nodeid; 
 		if ((1UL << nodeid) & nodes_present) { 
-			printk("Node %d already present. Skipping\n", nodeid);
+			printk(KERN_INFO "Node %d already present. Skipping\n", nodeid);
 			continue;
 		}
 
@@ -98,17 +98,19 @@ int __init k8_scan_nodes(unsigned long start, unsigned long end)
 			base = start; 
 		if (limit > end) 
 			limit = end; 
-		if (limit == base) 
+		if (limit == base) { 
+			printk(KERN_ERR "Empty node %d\n", nodeid); 
 			continue; 
+		}
 		if (limit < base) { 
-			printk(KERN_INFO"Node %d bogus settings %lx-%lx. Ignored.\n",
+			printk(KERN_ERR "Node %d bogus settings %lx-%lx.\n",
 			       nodeid, base, limit); 			       
-			continue; 
+			return -1;
 		} 
 		
 		/* Could sort here, but pun for now. Should not happen anyroads. */
 		if (prevbase > base) { 
-			printk(KERN_INFO "Node map not sorted %lx,%lx\n",
+			printk(KERN_ERR "Node map not sorted %lx,%lx\n",
 			       prevbase,base);
 			return -1;
 		}
@@ -116,23 +118,26 @@ int __init k8_scan_nodes(unsigned long start, unsigned long end)
 		printk(KERN_INFO "Node %d MemBase %016lx Limit %016lx\n", 
 		       nodeid, base, limit); 
 		
+		found++;
+		
 		nodes[nodeid].start = base; 
 		nodes[nodeid].end = limit;
 
 		prevbase = base;
 	} 
 
-	if (maxnode <= 0)
+	if (!found)
 		return -1; 
 
-	memnode_shift = compute_hash_shift(nodes,maxnode,end);
+	memnode_shift = compute_hash_shift(nodes);
 	if (memnode_shift < 0) { 
 		printk(KERN_ERR "No NUMA node hash function found. Contact maintainer\n"); 
 		return -1; 
 	} 
 	printk(KERN_INFO "Using node hash shift of %d\n", memnode_shift); 
 
-	early_for_all_nodes(i) { 
+	for (i = 0; i < numnodes; i++) { 
+		if (nodes[i].start != nodes[i].end)
 		setup_node_bootmem(i, nodes[i].start, nodes[i].end); 
 	} 
 

arch/x86_64/mm/numa.c

@@ -26,11 +26,10 @@ u8  memnodemap[NODEMAPSIZE];
 static int numa_off __initdata; 
 
 unsigned long nodes_present; 
-int maxnode;
 
 static int emunodes __initdata;
 
-int compute_hash_shift(struct node *nodes, int numnodes, u64 maxmem)
+int __init compute_hash_shift(struct node *nodes)
 {
 	int i; 
 	int shift = 24;
@@ -39,12 +38,16 @@ int compute_hash_shift(struct node *nodes, int numnodes, u64 maxmem)
 	/* When in doubt use brute force. */
 	while (shift < 48) { 
 		memset(memnodemap,0xff,sizeof(*memnodemap) * NODEMAPSIZE); 
-		early_for_all_nodes (i) { 
+		for (i = 0; i < numnodes; i++) { 
+			if (nodes[i].start == nodes[i].end) 
+				continue;
 			for (addr = nodes[i].start; 
 			     addr < nodes[i].end; 
 			     addr += (1UL << shift)) {
-				if (memnodemap[addr >> shift] != 0xff) { 
-					printk("node %d shift %d addr %Lx conflict %d\n", 
+				if (memnodemap[addr >> shift] != 0xff && 
+				    memnodemap[addr >> shift] != i) { 
+					printk(KERN_INFO 
+					    "node %d shift %d addr %Lx conflict %d\n", 
 					       i, shift, addr, memnodemap[addr>>shift]);
 					goto next; 
 				} 
@@ -101,9 +104,8 @@ void __init setup_node_bootmem(int nodeid, unsigned long start, unsigned long en
 
 	reserve_bootmem_node(NODE_DATA(nodeid), nodedata_phys, pgdat_size); 
 	reserve_bootmem_node(NODE_DATA(nodeid), bootmap_start, bootmap_pages<<PAGE_SHIFT);
-
-	if (nodeid > maxnode) 
-		maxnode = nodeid;
+	if (nodeid + 1 > numnodes) 
+		numnodes = nodeid + 1;
 	nodes_present |= (1UL << nodeid); 
 } 
 
@@ -151,6 +153,7 @@ int __init numa_initmem_init(unsigned long start_pfn, unsigned long end_pfn)
 		int i;
 		if (emunodes > MAXNODE) 
 			emunodes = MAXNODE; 
+		memset(&nodes, 0, sizeof(nodes)); 
 		printk(KERN_INFO "Faking %d nodes of size %ld MB\n", emunodes, nodesize>>20); 
 		for (i = 0; i < emunodes; i++) { 
 			unsigned long end = (i+1)*nodesize; 
@@ -160,7 +163,7 @@ int __init numa_initmem_init(unsigned long start_pfn, unsigned long end_pfn)
 			nodes[i].end = end; 
 			setup_node_bootmem(i, nodes[i].start, nodes[i].end);
 		}
-		memnode_shift = compute_hash_shift(nodes, emunodes, nodes[i-1].end); 
+		memnode_shift = compute_hash_shift(nodes); 
 		return 0;
 	} 
 

arch/x86_64/pci/irq.c

@@ -618,11 +618,20 @@ void pcibios_penalize_isa_irq(int irq)
 int pirq_enable_irq(struct pci_dev *dev)
 {
 	u8 pin;
+	extern int interrupt_line_quirk;
 	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin);
 	if (pin && !pcibios_lookup_irq(dev, 1) && !dev->irq) {
+		/* With IDE legacy devices the IRQ lookup failure is not a problem.. */
+		if (dev->class >> 8 == PCI_CLASS_STORAGE_IDE && !(dev->class & 0x5))
+			return 0;
+		
 		printk(KERN_WARNING "PCI: No IRQ known for interrupt pin %c of device %s.\n",
 		       'A' + pin - 1, dev->slot_name);
 	}
+	/* VIA bridges use interrupt line for apic/pci steering across
+	   the V-Link */
+	else if (interrupt_line_quirk)
+		pci_write_config_byte(dev, PCI_INTERRUPT_LINE, dev->irq);
 
 	return 0;
 }

include/asm-x86_64/apic.h

@@ -75,7 +75,7 @@ extern void smp_local_timer_interrupt (struct pt_regs * regs);
 extern void setup_boot_APIC_clock (void);
 extern void setup_secondary_APIC_clock (void);
 extern void setup_apic_nmi_watchdog (void);
-extern inline void nmi_watchdog_tick (struct pt_regs * regs);
+extern void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason);
 extern int APIC_init_uniprocessor (void);
 extern void disable_APIC_timer(void);
 extern void enable_APIC_timer(void);

include/asm-x86_64/debugreg.h

@@ -58,7 +58,7 @@
    We can slow the instruction pipeline for instructions coming via the
    gdt or the ldt if we want to.  I am not sure why this is an advantage */
 
-#define DR_CONTROL_RESERVED (0xFC00) /* Reserved by Intel */
+#define DR_CONTROL_RESERVED (0xFFFFFFFFFC00) /* Reserved by Intel */
 #define DR_LOCAL_SLOWDOWN (0x100)   /* Local slow the pipeline */
 #define DR_GLOBAL_SLOWDOWN (0x200)  /* Global slow the pipeline */
 

include/asm-x86_64/desc.h

@@ -8,6 +8,7 @@
 #ifndef __ASSEMBLY__
 
 #include <asm/segment.h>
+#include <asm/mmu.h>
 
 // 8 byte segment descriptor
 struct desc_struct { 

include/asm-x86_64/hdreg.h

@@ -7,6 +7,4 @@
 #ifndef __ASMx86_64_HDREG_H
 #define __ASMx86_64_HDREG_H
 
-typedef unsigned long ide_ioreg_t;
-
 #endif /* __ASMx86_64_HDREG_H */

include/asm-x86_64/i387.h

@@ -19,15 +19,15 @@
 #include <asm/thread_info.h>
 
 extern void fpu_init(void);
-extern void init_fpu(void);
-int save_i387(struct _fpstate *buf);
+extern void init_fpu(struct task_struct *child);
+extern int save_i387(struct _fpstate *buf);
 
 static inline int need_signal_i387(struct task_struct *me) 
 { 
 	if (!me->used_math)
 		return 0;
 	me->used_math = 0; 
-	if (me->thread_info->flags & _TIF_USEDFPU)
+	if (me->thread_info->status & TS_USEDFPU)
 		return 0;
 	return 1;
 } 
@@ -39,14 +39,14 @@ static inline int need_signal_i387(struct task_struct *me)
 #define kernel_fpu_end() stts()
 
 #define unlazy_fpu(tsk) do { \
-	if ((tsk)->thread_info->flags & _TIF_USEDFPU) \
+	if ((tsk)->thread_info->status & TS_USEDFPU) \
 		save_init_fpu(tsk); \
 } while (0)
 
 #define clear_fpu(tsk) do { \
-	if ((tsk)->thread_info->flags & _TIF_USEDFPU) {		\
+	if ((tsk)->thread_info->status & TS_USEDFPU) {		\
 		asm volatile("fwait");				\
-		(tsk)->thread_info->flags &= ~_TIF_USEDFPU;	\
+		(tsk)->thread_info->status &= ~TS_USEDFPU;	\
 		stts();						\
 	}							\
 } while (0)
@@ -114,11 +114,11 @@ static inline int save_i387_checking(struct i387_fxsave_struct *fx)
 
 static inline void kernel_fpu_begin(void)
 {
-	struct task_struct *me = current;
-	if (test_tsk_thread_flag(me,TIF_USEDFPU)) {
-		asm volatile("fxsave %0 ; fnclex"
-			      : "=m" (me->thread.i387.fxsave));
-		clear_tsk_thread_flag(me, TIF_USEDFPU);
+	struct thread_info *me = current_thread_info();
+	if (me->status & TS_USEDFPU) { 
+		asm volatile("rex64 ; fxsave %0 ; fnclex"
+			      : "=m" (me->task->thread.i387.fxsave));
+		me->status &= ~TS_USEDFPU;
 		return;
 	}
 	clts();
@@ -128,7 +128,7 @@ static inline void save_init_fpu( struct task_struct *tsk )
 {
 	asm volatile( "fxsave %0 ; fnclex"
 		      : "=m" (tsk->thread.i387.fxsave));
-	tsk->thread_info->flags &= ~TIF_USEDFPU;
+	tsk->thread_info->status &= ~TS_USEDFPU;
 	stts();
 }
 
@@ -140,18 +140,4 @@ static inline int restore_i387(struct _fpstate *buf)
 	return restore_fpu_checking((struct i387_fxsave_struct *)buf);
 }
 
-
-static inline void empty_fpu(struct task_struct *child)
-{
-	if (!child->used_math) {
-		/* Simulate an empty FPU. */
-		memset(&child->thread.i387.fxsave,0,sizeof(struct i387_fxsave_struct));
-		child->thread.i387.fxsave.cwd = 0x037f; 
-		child->thread.i387.fxsave.swd = 0;
-		child->thread.i387.fxsave.twd = 0; 
-		child->thread.i387.fxsave.mxcsr = 0x1f80;
-	}
-	child->used_math = 1; 
-}		
-
 #endif /* __ASM_X86_64_I387_H */

include/asm-x86_64/kdebug.h

@@ -9,8 +9,13 @@ struct die_args {
 	struct pt_regs *regs;
 	const char *str;
 	long err; 
+	int trapnr;
+	int signr;
 }; 
 
+/* Note - you should never unregister because that can race with NMIs.
+   If you really want to do it first unregister - then synchronize_kernel - then free. 
+  */
 extern struct notifier_block *die_chain;
 
 /* Grossly misnamed. */
@@ -21,15 +26,16 @@ enum die_val {
 	DIE_PANIC,
 	DIE_NMI,
 	DIE_DIE,
+	DIE_NMIWATCHDOG,
+	DIE_KERNELDEBUG,
+	DIE_TRAP,
+	DIE_GPF,
 	DIE_CALL,
-	DIE_CPUINIT,	/* not really a die, but .. */
-	DIE_TRAPINIT,	/* not really a die, but .. */
-	DIE_STOP, 
 }; 
 	
-static inline int notify_die(enum die_val val,char *str,struct pt_regs *regs,long err)
+static inline int notify_die(enum die_val val,char *str,struct pt_regs *regs,long err,int trap, int sig)
 { 
-	struct die_args args = { regs: regs, str: str, err: err }; 
+	struct die_args args = { .regs=regs, .str=str, .err=err, .trapnr=trap,.signr=sig }; 
 	return notifier_call_chain(&die_chain, val, &args); 
 } 
 

include/asm-x86_64/numa.h

@@ -8,13 +8,11 @@ struct node {
 	u64 start,end; 
 };
 
-#define for_all_nodes(x) for ((x) = 0; (x) <= maxnode; (x)++) \
+#define for_all_nodes(x) for ((x) = 0; (x) < numnodes; (x)++) \
 				if ((1UL << (x)) & nodes_present)
 
-#define early_for_all_nodes(n) \
-	for (n=0; n<MAXNODE;n++) if (nodes[n].start!=nodes[n].end)
 
-extern int compute_hash_shift(struct node *nodes, int numnodes, u64 maxmem);
+extern int compute_hash_shift(struct node *nodes);
 extern unsigned long nodes_present;
 
 #define ZONE_ALIGN (1UL << (MAX_ORDER+PAGE_SHIFT))

include/asm-x86_64/pci.h

@@ -44,8 +44,7 @@ int pcibios_set_irq_routing(struct pci_dev *dev, int pin, int irq);
 
 struct pci_dev;
 
-extern int iommu_setup(char *opt, char **end);
-
+extern int iommu_setup(char *opt);
 extern void pci_iommu_init(void);
 
 /* Allocate and map kernel buffer using consistent mode DMA for a device.
@@ -77,10 +76,11 @@ extern void pci_free_consistent(struct pci_dev *hwdev, size_t size,
  * Once the device is given the dma address, the device owns this memory
  * until either pci_unmap_single or pci_dma_sync_single is performed.
  */
-extern dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr,
-			  size_t size, int direction);
+extern dma_addr_t __pci_map_single(struct pci_dev *hwdev, void *ptr,
+			  size_t size, int direction, int flush);
+
 
-extern void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t addr,
+void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t addr,
 				   size_t size, int direction);
 
 /*
@@ -118,12 +118,16 @@ static inline void pci_dma_sync_sg(struct pci_dev *hwdev,
 	BUG_ON(direction == PCI_DMA_NONE); 
 } 
 
-#define PCI_DMA_BUS_IS_PHYS	0
+/* The PCI address space does equal the physical memory
+ * address space.  The networking and block device layers use
+ * this boolean for bounce buffer decisions.
+ */
+#define PCI_DMA_BUS_IS_PHYS	(0)
 
 
 #else
-static inline dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr,
-					size_t size, int direction)
+static inline dma_addr_t __pci_map_single(struct pci_dev *hwdev, void *ptr,
+					size_t size, int direction, int flush)
 {
 	dma_addr_t addr; 
 
@@ -210,6 +214,11 @@ extern int pci_map_sg(struct pci_dev *hwdev, struct scatterlist *sg,
 extern void pci_unmap_sg(struct pci_dev *hwdev, struct scatterlist *sg,
 			 int nents, int direction);
 
+static inline dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr,
+			  size_t size, int direction)
+{
+	return __pci_map_single(hwdev,ptr,size,direction,1); 
+}
 
 #define pci_unmap_page pci_unmap_single
 

include/asm-x86_64/pgtable.h

@@ -283,6 +283,7 @@ static inline int pmd_large(pmd_t pte) {
 #define pml4_page(pml4) ((unsigned long) __va(pml4_val(pml4) & PTE_MASK))
 #define pml4_index(address) ((address >> PML4_SHIFT) & (PTRS_PER_PML4-1))
 #define pml4_offset_k(address) (init_level4_pgt + pml4_index(address))
+#define pml4_present(pml4) (pml4_val(pml4) & _PAGE_PRESENT)
 #define mk_kernel_pml4(address) ((pml4_t){ (address) | _KERNPG_TABLE })
 #define level3_offset_k(dir, address) ((pgd_t *) pml4_page(*(dir)) + pgd_index(address))
 

include/asm-x86_64/processor.h

@@ -269,7 +269,7 @@ struct mm_struct;
 extern void release_thread(struct task_struct *);
 
 /* Prepare to copy thread state - unlazy all lazy status */
-#define prepare_to_copy(tsk)	do { } while (0)
+extern void prepare_to_copy(struct task_struct *tsk);
 
 /*
  * create a kernel thread without removing it from tasklists
@@ -308,8 +308,8 @@ extern inline void sync_core(void)
 #define ARCH_HAS_PREFETCHW
 #define ARCH_HAS_SPINLOCK_PREFETCH
 
-#define prefetch(x) __builtin_prefetch((x),0)
-#define prefetchw(x) __builtin_prefetch((x),1)
+#define prefetch(x) __builtin_prefetch((x),0,1)
+#define prefetchw(x) __builtin_prefetch((x),1,1)
 #define spin_lock_prefetch(x)  prefetchw(x)
 #define cpu_relax()   rep_nop()
 

include/asm-x86_64/proto.h

@@ -6,6 +6,7 @@
 /* misc architecture specific prototypes */
 
 struct cpuinfo_x86; 
+struct pt_regs;
 
 extern void get_cpu_vendor(struct cpuinfo_x86*);
 extern void start_kernel(void);
@@ -41,6 +42,8 @@ extern void free_bootmem_generic(unsigned long phys, unsigned len);
 extern unsigned long end_pfn_map; 
 
 extern void show_stack(unsigned long * rsp);
+extern void show_trace(unsigned long * rsp);
+extern void show_registers(struct pt_regs *regs);
 
 extern void exception_table_check(void);
 

include/asm-x86_64/suspend.h

@@ -11,7 +11,7 @@ arch_prepare_suspend(void)
 {
 }
 
-/* image of the saved processor state */
+/* Image of the saved processor state. If you touch this, fix acpi_wakeup.S. */
 struct saved_context {
   	u16 ds, es, fs, gs, ss;
 	unsigned long gs_base, gs_kernel_base, fs_base;

include/asm-x86_64/system.h

@@ -22,18 +22,18 @@
 struct save_context_frame { 
 	unsigned long rbp; 
 	unsigned long rbx;
+	unsigned long r11;
+	unsigned long r10;
+	unsigned long r9;
+	unsigned long r8;
 	unsigned long rcx;
 	unsigned long rdx;	
-	unsigned long rsi;
-	unsigned long rdi; 
 	unsigned long r15;
 	unsigned long r14;
 	unsigned long r13;
 	unsigned long r12;
-	unsigned long r11;
-	unsigned long r10;
-	unsigned long r9;
-	unsigned long r8; 
+	unsigned long rdi;
+	unsigned long rsi;
 }; 
 
 /* frame pointer must be last for get_wchan */
@@ -43,19 +43,20 @@ struct save_context_frame {
    rbp needs to be always explicitely saved because gcc cannot clobber the
    frame pointer and the scheduler is compiled with frame pointers. -AK */
 #define SAVE_CONTEXT \
-	__PUSH(r8) __PUSH(r9) __PUSH(r10) __PUSH(r11) __PUSH(r12) __PUSH(r13) \
-	__PUSH(r14) __PUSH(r15) \
-	__PUSH(rdi) __PUSH(rsi) \
-	__PUSH(rdx) __PUSH(rcx) \
+	__PUSH(rsi) __PUSH(rdi) \
+    __PUSH(r12) __PUSH(r13) __PUSH(r14) __PUSH(r15)  \
+	__PUSH(rdx) __PUSH(rcx) __PUSH(r8) __PUSH(r9) __PUSH(r10) __PUSH(r11)  \
 	__PUSH(rbx) __PUSH(rbp) 
 #define RESTORE_CONTEXT \
 	__POP(rbp) __POP(rbx) \
-	__POP(rcx) __POP(rdx) \
-	__POP(rsi) __POP(rdi) \
-	__POP(r15) __POP(r14) __POP(r13) __POP(r12) __POP(r11) __POP(r10) \
-	__POP(r9) __POP(r8)
+	__POP(r11) __POP(r10) __POP(r9) __POP(r8) __POP(rcx) __POP(rdx) \
+	__POP(r15) __POP(r14) __POP(r13) __POP(r12) \
+	__POP(rdi) __POP(rsi)
 
 /* RED-PEN: pipeline stall on ret because it is not predicted */
+/* RED-PEN: the register saving could be optimized */
+/* frame pointer must be last for get_wchan */
+
 #define switch_to(prev,next,last) \
 	asm volatile(SAVE_CONTEXT						    \
 		     "movq %%rsp,%[prevrsp]\n\t"				    \

include/asm-x86_64/thread_info.h

@@ -27,6 +27,7 @@ struct thread_info {
 	struct task_struct	*task;		/* main task structure */
 	struct exec_domain	*exec_domain;	/* execution domain */
 	__u32			flags;		/* low level flags */
+	__u32			status;		/* thread synchronous flags */
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;
 
@@ -100,16 +101,14 @@ static inline struct thread_info *stack_thread_info(void)
 #define TIF_SIGPENDING		2	/* signal pending */
 #define TIF_NEED_RESCHED	3	/* rescheduling necessary */
 #define TIF_SINGLESTEP		4	/* reenable singlestep on user return*/
-#define TIF_USEDFPU		16	/* FPU was used by this task this quantum */
-#define TIF_POLLING_NRFLAG	17	/* true if poll_idle() is polling TIF_NEED_RESCHED */
-#define TIF_IA32		18	/* 32bit process */ 
+#define TIF_POLLING_NRFLAG	16	/* true if poll_idle() is polling TIF_NEED_RESCHED */
+#define TIF_IA32		17	/* 32bit process */ 
 
 #define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
 #define _TIF_NOTIFY_RESUME	(1<<TIF_NOTIFY_RESUME)
 #define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
 #define _TIF_SINGLESTEP		(1<<TIF_SINGLESTEP)
 #define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
-#define _TIF_USEDFPU		(1<<TIF_USEDFPU)
 #define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
 #define _TIF_IA32		(1<<TIF_IA32)
 
@@ -118,6 +117,15 @@ static inline struct thread_info *stack_thread_info(void)
 
 #define PREEMPT_ACTIVE     0x4000000
 
+/*
+ * Thread-synchronous status.
+ *
+ * This is different from the flags in that nobody else
+ * ever touches our thread-synchronous status, so we don't
+ * have to worry about atomic accesses.
+ */
+#define TS_USEDFPU		0x0001	/* FPU was used by this task this quantum (SMP) */
+
 #endif /* __KERNEL__ */
 
 #endif /* _ASM_THREAD_INFO_H */

include/asm-x86_64/unistd.h

@@ -520,8 +520,10 @@ __SYSCALL(__NR_clock_gettime, sys_clock_gettime)
 __SYSCALL(__NR_clock_getres, sys_clock_getres)
 #define __NR_clock_nanosleep	230
 __SYSCALL(__NR_clock_nanosleep, sys_clock_nanosleep)
+#define __NR_exit_group		231
+__SYSCALL(__NR_exit_group, sys_exit_group)
 
-#define __NR_syscall_max __NR_clock_nanosleep
+#define __NR_syscall_max __NR_exit_group
 #ifndef __NO_STUBS
 
 /* user-visible error numbers are in the range -1 - -4095 */