[PATCH] x86-64 updates for 2.5.54

More x86-64 updates for 2.5.54.  Most noticeable change is that the
64bit X server works again.

This only changes x86-64 specific files.  It requires some AGP driver
changes I'm sending separately.

 - Some Makefile cleanups from Sam Ravnborg
 - Make sure extended registers in 32bit processes are zeroed and not
   accessible/changeable from ptrace.  This is to avoid potential
   security bugs with non 64bit clean 32bit emulation functions (they
   often are overflow prone etc.)
 - Some 32bit emulation cleanups from Stephen Rothwell
 - Make copy_*_user source const to fix warnings.
 - Set fs/gs to dummy values when the 64bit segment base is set to not
   confuse the context switch (Karsten Keil, me)
	* still one mysterious bug in this area unfortunately.
 - Make MAP_32BIT for 64bit processes only map in the first 31bit,
   because it is usually used to map small model code.  This fixes the X
   server crashes.  Some cleanups in this area.
 - Don't set O_LARGEFILE for 32bit open
 - Handle ptregs calls from 32bit syscall correctly.
 - Implement aio io_getevents for 32bit.
 - Remove buggy unused command handler in nfsd 32bit emulation.
 - Convert timespec in semtimedop (thanks to Anton for telling me about
   this)
 - Ignore long mode flag from 32bit modify_ldt.  This fixes Wine, which
   left it uninitialized (bug found by Karsten Keil)
 - Merge with i386
 - Handle new kallsyms
 - Remove some superfluous bootup printks

arch/x86_64/Makefile

@@ -57,30 +57,32 @@ drivers-$(CONFIG_PCI)			+= arch/x86_64/pci/
 # FIXME: is drivers- right ?
 drivers-$(CONFIG_OPROFILE)		+= arch/x86_64/oprofile/
 
-boot := arch/x86_64/boot
+makeboot =$(Q)$(MAKE) -f scripts/Makefile.build obj=arch/x86_64/boot $(1)
 
 .PHONY: zImage bzImage compressed zlilo bzlilo zdisk bzdisk install \
-	archclean
+		clean archclean archmrproper
 
 BOOTIMAGE=arch/x86_64/boot/bzImage
 zImage zlilo zdisk: BOOTIMAGE=arch/x86_64/boot/zImage
 
 zImage bzImage: vmlinux
-	$(Q)$(MAKE) $(build)=$(boot) $(BOOTIMAGE)
+	$(call makeboot,$(BOOTIMAGE))
 
 compressed: zImage
 
 zlilo bzlilo: vmlinux
-	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) zlilo
+	$(call makeboot,BOOTIMAGE=$(BOOTIMAGE) zlilo)
 
 zdisk bzdisk: vmlinux
-	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) zdisk
+	$(call makeboot,BOOTIMAGE=$(BOOTIMAGE) zdisk)
 
 install: vmlinux
-	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) install
+	$(call makeboot,BOOTIMAGE=$(BOOTIMAGE) install)
 
 archclean:
-	$(Q)$(MAKE) $(clean)=$(boot)
+	$(Q)$(MAKE) $(clean)=arch/i386/boot
+
+archmrproper:
 
 prepare: include/asm-$(ARCH)/offset.h
 
@@ -104,3 +106,4 @@ define archhelp
   echo  '                  (distribution) /sbin/installkernel or'
   echo  '        	  install to $$(INSTALL_PATH) and run lilo'
 endef
+

arch/x86_64/ia32/ia32_binfmt.c

@@ -261,6 +261,8 @@ static void elf32_init(struct pt_regs *regs)
 	regs->rax = 0;
 	regs->rbx = 0; 
 	regs->rbp = 0; 
+	regs->r8 = regs->r9 = regs->r10 = regs->r11 = regs->r12 =
+		regs->r13 = regs->r14 = regs->r15 = 0; 
     me->thread.fs = 0; 
 	me->thread.gs = 0;
 	me->thread.fsindex = 0; 

arch/x86_64/ia32/ia32_ioctl.c

@@ -450,13 +450,13 @@ struct ifreq32 {
                 struct  ifmap32 ifru_map;
                 char    ifru_slave[IFNAMSIZ];   /* Just fits the size */
 		char	ifru_newname[IFNAMSIZ];
-                __kernel_caddr_t32 ifru_data;
+                compat_caddr_t ifru_data;
         } ifr_ifru;
 };
 
 struct ifconf32 {
         int     ifc_len;                        /* size of buffer       */
-        __kernel_caddr_t32  ifcbuf;
+        compat_caddr_t  ifcbuf;
 };
 
 #ifdef CONFIG_NET
@@ -876,7 +876,7 @@ static int hdio_getgeo(unsigned int fd, unsigned int cmd, unsigned long arg)
 
 struct fb_fix_screeninfo32 {
 	char			id[16];
-        __kernel_caddr_t32	smem_start;
+        compat_caddr_t	smem_start;
 	__u32			smem_len;
 	__u32			type;
 	__u32			type_aux;
@@ -885,7 +885,7 @@ struct fb_fix_screeninfo32 {
 	__u16			ypanstep;
 	__u16			ywrapstep;
 	__u32			line_length;
-        __kernel_caddr_t32	mmio_start;
+        compat_caddr_t	mmio_start;
 	__u32			mmio_len;
 	__u32			accel;
 	__u16			reserved[3];
@@ -894,10 +894,10 @@ struct fb_fix_screeninfo32 {
 struct fb_cmap32 {
 	__u32			start;
 	__u32			len;
-	__kernel_caddr_t32	red;
-	__kernel_caddr_t32	green;
-	__kernel_caddr_t32	blue;
-	__kernel_caddr_t32	transp;
+	compat_caddr_t	red;
+	compat_caddr_t	green;
+	compat_caddr_t	blue;
+	compat_caddr_t	transp;
 };
 
 static int fb_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
@@ -1036,7 +1036,7 @@ struct floppy_struct32 {
 	unsigned char	rate;
 	unsigned char	spec1;
 	unsigned char	fmt_gap;
-	const __kernel_caddr_t32 name;
+	const compat_caddr_t name;
 };
 
 struct floppy_drive_params32 {
@@ -1075,7 +1075,7 @@ struct floppy_drive_struct32 {
 	int		fd_ref;
 	int		fd_device;
 	int		last_checked;
-	__kernel_caddr_t32 dmabuf;
+	compat_caddr_t dmabuf;
 	int		bufblocks;
 };
 
@@ -1600,7 +1600,7 @@ static int ppp_sock_fprog_ioctl_trans(unsigned int fd, unsigned int cmd, unsigne
 }
 
 struct ppp_option_data32 {
-	__kernel_caddr_t32	ptr;
+	compat_caddr_t	ptr;
 	__u32			length;
 	int			transmit;
 };
@@ -1681,8 +1681,8 @@ struct mtget32 {
 	__u32	mt_dsreg;
 	__u32	mt_gstat;
 	__u32	mt_erreg;
-	__kernel_daddr_t32	mt_fileno;
-	__kernel_daddr_t32	mt_blkno;
+	compat_daddr_t	mt_fileno;
+	compat_daddr_t	mt_blkno;
 };
 #define MTIOCGET32	_IOR('m', 2, struct mtget32)
 
@@ -1800,7 +1800,7 @@ static int mt_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
 
 struct cdrom_read32 {
 	int			cdread_lba;
-	__kernel_caddr_t32	cdread_bufaddr;
+	compat_caddr_t	cdread_bufaddr;
 	int			cdread_buflen;
 };
 
@@ -1808,16 +1808,16 @@ struct cdrom_read_audio32 {
 	union cdrom_addr	addr;
 	u_char			addr_format;
 	int			nframes;
-	__kernel_caddr_t32	buf;
+	compat_caddr_t	buf;
 };
 
 struct cdrom_generic_command32 {
 	unsigned char		cmd[CDROM_PACKET_SIZE];
-	__kernel_caddr_t32	buffer;
+	compat_caddr_t	buffer;
 	unsigned int		buflen;
 	int			stat;
-	__kernel_caddr_t32	sense;
-	__kernel_caddr_t32	reserved[3];
+	compat_caddr_t	sense;
+	compat_caddr_t	reserved[3];
 };
 
 static int cdrom_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
@@ -1826,7 +1826,7 @@ static int cdrom_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long ar
 	struct cdrom_read cdread;
 	struct cdrom_read_audio cdreadaudio;
 	struct cdrom_generic_command cgc;
-	__kernel_caddr_t32 addr;
+	compat_caddr_t addr;
 	char *data = 0;
 	void *karg;
 	int err = 0;
@@ -2124,12 +2124,12 @@ static int do_smb_getmountuid(unsigned int fd, unsigned int cmd, unsigned long a
 struct atmif_sioc32 {
         int                number;
         int                length;
-        __kernel_caddr_t32 arg;
+        compat_caddr_t arg;
 };
 
 struct atm_iobuf32 {
 	int                length;
-	__kernel_caddr_t32 buffer;
+	compat_caddr_t buffer;
 };
 
 #define ATM_GETLINKRATE32 _IOW('a', ATMIOC_ITF+1, struct atmif_sioc32)
@@ -2190,7 +2190,7 @@ static int do_atm_iobuf(unsigned int fd, unsigned int cmd, unsigned long arg)
 
 	iobuf.length = iobuf32.length;
 
-	if (iobuf32.buffer == (__kernel_caddr_t32) NULL || iobuf32.length == 0) {
+	if (iobuf32.buffer == (compat_caddr_t) NULL || iobuf32.length == 0) {
 		iobuf.buffer = (void*)(unsigned long)iobuf32.buffer;
 	} else {
 		iobuf.buffer = kmalloc(iobuf.length, GFP_KERNEL);
@@ -2244,7 +2244,7 @@ static int do_atmif_sioc(unsigned int fd, unsigned int cmd, unsigned long arg)
         sioc.number = sioc32.number;
         sioc.length = sioc32.length;
         
-	if (sioc32.arg == (__kernel_caddr_t32) NULL || sioc32.length == 0) {
+	if (sioc32.arg == (compat_caddr_t) NULL || sioc32.length == 0) {
 		sioc.arg = (void*)(unsigned long)sioc32.arg;
         } else {
                 sioc.arg = kmalloc(sioc.length, GFP_KERNEL);

arch/x86_64/ia32/ia32entry.S

@@ -76,6 +76,7 @@ cstar_do_call:
 	IA32_ARG_FIXUP 1
 	call *ia32_sys_call_table(,%rax,8)
 	.globl cstar_sysret
+	/* label must directly follow call */
 cstar_sysret:	
 	movq %rax,RAX-ARGOFFSET(%rsp)
 	GET_THREAD_INFO(%r10)
@@ -130,7 +131,7 @@ ENTRY(ia32_syscall)
 	pushq %rax
 	cld
 	/* note the registers are not zero extended to the sf.
-	   this could be a problem */
+	   this could be a problem. */
 	SAVE_ARGS
 	GET_THREAD_INFO(%r10)
 	bt $TIF_SYSCALL_TRACE,threadinfo_flags(%r10)
@@ -190,6 +191,8 @@ ENTRY(ia32_ptregs_common)
 	call *%rax
 	movq %r15, %r11
 	RESTORE_REST
+	cmpq $cstar_sysret,%r11
+	je   int_ret_from_sys_call /* misbalances the call/ret stack. sorry */
 	pushq %r11
 	ret
 
@@ -202,7 +205,7 @@ ia32_sys_call_table:
 	.quad stub32_fork
 	.quad sys_read
 	.quad sys_write
-	.quad sys_open		/* 5 */
+	.quad sys32_open		/* 5 */
 	.quad sys_close
 	.quad sys32_waitpid
 	.quad sys_creat
@@ -320,13 +323,13 @@ ia32_sys_call_table:
 	.quad stub32_clone		/* 120 */
 	.quad sys_setdomainname
 	.quad sys_uname
-	.quad sys_modify_ldt
+	.quad sys32_modify_ldt
 	.quad sys32_adjtimex
 	.quad sys32_mprotect		/* 125 */
 	.quad sys32_sigprocmask
 	.quad sys32_module_warning	/* create_module */
-	.quad sys32_module_warning	/* init_module */
-	.quad sys32_module_warning	/* delete module */
+	.quad sys_init_module
+	.quad sys_delete_module
 	.quad sys32_module_warning	/* 130  get_kernel_syms */
 	.quad ni_syscall	/* quotactl */ 
 	.quad sys_getpgid
@@ -444,8 +447,8 @@ ia32_sys_call_table:
 	.quad sys_get_thread_area
 	.quad sys32_io_setup
 	.quad sys_io_destroy
-	.quad sys_io_getevents
-	.quad sys_io_submit
+	.quad sys32_io_getevents
+	.quad sys32_io_submit
 	.quad sys_io_cancel
 	.quad sys_ni_syscall /* 250 alloc_huge_pages */
 	.quad sys_ni_syscall /* free_huge_pages */

arch/x86_64/ia32/ipc32.c

@@ -40,10 +40,10 @@ struct ipc_perm32 {
 
 struct ipc64_perm32 {
         unsigned key;
-	__kernel_uid32_t32 uid;
-	__kernel_gid32_t32 gid;
-	__kernel_uid32_t32 cuid;
-	__kernel_gid32_t32 cgid;
+	compat_uid32_t uid;
+	compat_gid32_t gid;
+	compat_uid32_t cuid;
+	compat_gid32_t cgid;
 	unsigned short mode;
 	unsigned short __pad1;
 	unsigned short seq;
@@ -86,8 +86,8 @@ struct msqid_ds32 {
 	unsigned short msg_cbytes;
 	unsigned short msg_qnum;
 	unsigned short msg_qbytes;
-	__kernel_ipc_pid_t32 msg_lspid;
-	__kernel_ipc_pid_t32 msg_lrpid;
+	compat_ipc_pid_t msg_lspid;
+	compat_ipc_pid_t msg_lrpid;
 };
 
 struct msqid64_ds32 {
@@ -113,8 +113,8 @@ struct shmid_ds32 {
 	compat_time_t shm_atime;
 	compat_time_t shm_dtime;
 	compat_time_t shm_ctime;
-	__kernel_ipc_pid_t32 shm_cpid;
-	__kernel_ipc_pid_t32 shm_lpid;
+	compat_ipc_pid_t shm_cpid;
+	compat_ipc_pid_t shm_lpid;
 	unsigned short shm_nattch;
 };
 
@@ -613,6 +613,31 @@ shmctl32 (int first, int second, void *uptr)
 	return err;
 }
 
+extern int sem_ctls[];
+
+static long semtimedop32(int semid, struct sembuf *sb, 
+			 unsigned nsops, struct compat_timespec *ts32)
+{ 
+	struct timespec ts;
+	mm_segment_t oldfs = get_fs(); 
+	long ret;
+
+	if (nsops > sem_ctls[2]) 
+		return -E2BIG;
+	if (!access_ok(VERIFY_READ, sb, nsops * sizeof(struct sembuf)))
+		return -EFAULT; 
+	if (ts32 && 
+	    (get_user(ts.tv_sec, &ts32->tv_sec)  || 
+	     __get_user(ts.tv_nsec, &ts32->tv_nsec)))
+		return -EFAULT;
+
+	set_fs(KERNEL_DS);  
+	ret = sys_semtimedop(semid, sb, nsops, ts32 ? &ts : NULL);
+	set_fs(oldfs); 
+	return ret;
+}
+
+
 asmlinkage long
 sys32_ipc (u32 call, int first, int second, int third, u32 ptr, u32 fifth)
 {
@@ -627,9 +652,8 @@ sys32_ipc (u32 call, int first, int second, int third, u32 ptr, u32 fifth)
 		return sys_semtimedop(first, (struct sembuf *)AA(ptr), second,
 				      NULL);
 	      case TIMEDSEMOP:
-		/* struct sembuf is the same on 32 and 64bit :)) */
-		return sys_semtimedop(first, (struct sembuf *)AA(ptr), second,
-				      (const struct timespec *)AA(fifth));
+		return semtimedop32(first, (struct sembuf *)AA(ptr), second,
+				  (struct compat_timespec *)AA(fifth)); 
 	      case SEMGET:
 		return sys_semget(first, second, third);
 	      case SEMCTL:

arch/x86_64/ia32/sys_ia32.c

@@ -65,6 +65,7 @@
 #include <asm/semaphore.h>
 #include <asm/ipc.h>
 #include <asm/atomic.h>
+#include <asm/ldt.h>
 
 #include <net/scm.h>
 #include <net/sock.h>
@@ -92,6 +93,11 @@ extern asmlinkage long sys_newlstat(char * filename, struct stat * statbuf);
 extern asmlinkage long sys_newfstat(unsigned int fd, struct stat * statbuf);
 
 
+extern asmlinkage long sys_newstat(char * filename, struct stat * statbuf);
+extern asmlinkage long sys_newlstat(char * filename, struct stat * statbuf);
+extern asmlinkage long sys_newfstat(unsigned int fd, struct stat * statbuf);
+
+
 int cp_compat_stat(struct kstat *kbuf, struct compat_stat *ubuf)
 {
 	if (verify_area(VERIFY_WRITE, ubuf, sizeof(struct compat_stat)) ||
@@ -228,19 +234,12 @@ sys32_mmap(struct mmap_arg_struct *arg)
 	if (a.prot & PROT_READ) 
 		a.prot |= PROT_EXEC; 
 
-	a.flags |= MAP_32BIT;
-
 	mm = current->mm; 
 	down_write(&mm->mmap_sem); 
 	retval = do_mmap_pgoff(file, a.addr, a.len, a.prot, a.flags, a.offset>>PAGE_SHIFT);
 	if (file)
 		fput(file);
 
-	/* Cannot wrap */
-	if (retval+a.len >= IA32_PAGE_OFFSET && (long)retval > 0) { 
-		do_munmap(mm, retval, a.len); 
-		retval = -ENOMEM; 
-	} 
 	up_write(&mm->mmap_sem); 
 
 	return retval;
@@ -479,6 +478,8 @@ put_tv32(struct compat_timeval *o, struct timeval *i)
 	return err; 
 }
 
+extern int do_setitimer(int which, struct itimerval *, struct itimerval *);
+
 asmlinkage long
 sys32_alarm(unsigned int seconds)
 {
@@ -1662,6 +1663,31 @@ sys32_sendfile(int out_fd, int in_fd, compat_off_t *offset, s32 count)
 	return ret;
 }
 
+extern asmlinkage long sys_modify_ldt(int func, void *ptr, 
+				      unsigned long bytecount);
+
+asmlinkage long sys32_modify_ldt(int func, void *ptr, unsigned long bytecount)
+{
+	long ret;
+	if (func == 0x1 || func == 0x11) { 
+		struct user_desc info;
+		mm_segment_t old_fs = get_fs();
+		if (bytecount != sizeof(struct user_desc))
+			return -EINVAL;
+		if (copy_from_user(&info, ptr, sizeof(struct user_desc)))
+			return -EFAULT;
+		/* lm bit was undefined in the 32bit ABI and programs
+		   give it random values. Force it to zero here. */
+		info.lm = 0; 
+		set_fs(KERNEL_DS);
+		ret = sys_modify_ldt(func, &info, bytecount);
+		set_fs(old_fs);
+	}  else { 
+		ret = sys_modify_ldt(func, ptr, bytecount); 
+	}
+	return ret;
+}
+
 /* Handle adjtimex compatability. */
 
 struct timex32 {
@@ -1757,15 +1783,9 @@ asmlinkage long sys32_mmap2(unsigned long addr, unsigned long len,
 		prot |= PROT_EXEC;
 
 	down_write(&mm->mmap_sem);
-	error = do_mmap_pgoff(file, addr, len, prot, flags|MAP_32BIT, pgoff);
+	error = do_mmap_pgoff(file, addr, len, prot, flags, pgoff);
 	up_write(&mm->mmap_sem);
 
-	/* cannot wrap */
-	if (error+len >= IA32_PAGE_OFFSET && (long)error >= 0) { 
-		do_munmap(mm, error, len); 
-		error = -ENOMEM; 
-	} 
-
 	if (file)
 		fput(file);
 	return error;
@@ -2110,54 +2130,6 @@ static int nfs_exp32_trans(struct nfsctl_arg *karg, struct nfsctl_arg32 *arg32)
 	return err;
 }
 
-static int nfs_uud32_trans(struct nfsctl_arg *karg, struct nfsctl_arg32 *arg32)
-{
-	u32 uaddr;
-	int i;
-	int err;
-
-	memset(karg, 0, sizeof(*karg));
-	if(get_user(karg->ca_version, &arg32->ca32_version))
-		return -EFAULT;
-	karg->ca_umap.ug_ident = (char *)get_zeroed_page(GFP_USER);
-	if(!karg->ca_umap.ug_ident)
-		return -ENOMEM;
-	err = get_user(uaddr, &arg32->ca32_umap.ug32_ident);
-	if(strncpy_from_user(karg->ca_umap.ug_ident,
-			     (char *)A(uaddr), PAGE_SIZE) <= 0)
-		return -EFAULT;
-	err |= __get_user(karg->ca_umap.ug_uidbase,
-		      &arg32->ca32_umap.ug32_uidbase);
-	err |= __get_user(karg->ca_umap.ug_uidlen,
-		      &arg32->ca32_umap.ug32_uidlen);
-	err |= __get_user(uaddr, &arg32->ca32_umap.ug32_udimap);
-	if (err)
-		return -EFAULT;
-	karg->ca_umap.ug_udimap = kmalloc((sizeof(uid_t) * karg->ca_umap.ug_uidlen),
-					  GFP_USER);
-	if(!karg->ca_umap.ug_udimap)
-		return -ENOMEM;
-	for(i = 0; i < karg->ca_umap.ug_uidlen; i++)
-		err |= __get_user(karg->ca_umap.ug_udimap[i],
-			      &(((compat_pid_t *)A(uaddr))[i]));
-	err |= __get_user(karg->ca_umap.ug_gidbase,
-		      &arg32->ca32_umap.ug32_gidbase);
-	err |= __get_user(karg->ca_umap.ug_uidlen,
-		      &arg32->ca32_umap.ug32_gidlen);
-	err |= __get_user(uaddr, &arg32->ca32_umap.ug32_gdimap);
-	if (err)
-		return -EFAULT;
-	karg->ca_umap.ug_gdimap = kmalloc((sizeof(gid_t) * karg->ca_umap.ug_uidlen),
-					  GFP_USER);
-	if(!karg->ca_umap.ug_gdimap)
-		return -ENOMEM;
-	for(i = 0; i < karg->ca_umap.ug_gidlen; i++)
-		err |= __get_user(karg->ca_umap.ug_gdimap[i],
-			      &(((compat_gid_t *)A(uaddr))[i]));
-
-	return err;
-}
-
 static int nfs_getfh32_trans(struct nfsctl_arg *karg, struct nfsctl_arg32 *arg32)
 {
 	int err;
@@ -2246,10 +2218,6 @@ long asmlinkage sys32_nfsservctl(int cmd, struct nfsctl_arg32 *arg32, union nfsc
 	case NFSCTL_UNEXPORT:
 		err = nfs_exp32_trans(karg, arg32);
 		break;
-	/* This one is unimplemented, be we're ready for it. */
-	case NFSCTL_UGIDUPDATE:
-		err = nfs_uud32_trans(karg, arg32);
-		break;
 	case NFSCTL_GETFH:
 		err = nfs_getfh32_trans(karg, arg32);
 		break;
@@ -2306,7 +2274,7 @@ 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 modutils not supported on 64bit kernel\n",
+		printk(KERN_INFO "%s: 32bit 2.4.x modutils not supported on 64bit kernel\n",
 		       current->comm);
 		warn_time = jiffies;
 	} 
@@ -2434,7 +2402,64 @@ long sys32_io_submit(aio_context_t ctx_id, unsigned long nr,
 	return err;		
 }
 
-/* XXX aio getevents */
+extern asmlinkage long sys_io_getevents(aio_context_t ctx_id,
+					  long min_nr,
+					  long nr,
+					  struct io_event *events,
+					  struct timespec *timeout);
+
+asmlinkage long sys32_io_getevents(aio_context_t ctx_id,
+				 unsigned long min_nr,
+				 unsigned long nr,
+				 struct io_event *events,
+				 struct compat_timespec *timeout)
+{ 	
+	long ret;
+	mm_segment_t oldfs; 
+	struct timespec t32;
+	/* Harden against bogus ptrace */
+	if (nr >= 0xffffffff || 
+	    !access_ok(VERIFY_WRITE, events, nr * sizeof(struct io_event)))
+		return -EFAULT;
+	if (timeout && 
+	    (get_user(t32.tv_sec, &timeout->tv_sec)  || 
+	     __get_user(t32.tv_nsec, &timeout->tv_nsec)))
+		return -EFAULT; 
+	oldfs = get_fs();
+	set_fs(KERNEL_DS); 
+	ret = sys_io_getevents(ctx_id,min_nr,nr,events,timeout ? &t32 : NULL); 
+	set_fs(oldfs); 
+	if (timeout && 
+	    (__put_user(t32.tv_sec, &timeout->tv_sec) || 
+	     __put_user(t32.tv_nsec, &timeout->tv_nsec)))
+		return -EFAULT; 		
+	return ret;
+} 
+
+asmlinkage long sys32_open(const char * filename, int flags, int mode)
+{
+	char * tmp;
+	int fd, error;
+
+	/* don't force O_LARGEFILE */
+	tmp = getname(filename);
+	fd = PTR_ERR(tmp);
+	if (!IS_ERR(tmp)) {
+		fd = get_unused_fd();
+		if (fd >= 0) {
+			struct file *f = filp_open(tmp, flags, mode);
+			error = PTR_ERR(f);
+			if (unlikely(IS_ERR(f))) {
+				put_unused_fd(fd); 
+				fd = error;
+			} else
+				fd_install(fd, f);
+		}
+		putname(tmp);
+	}
+	return fd;
+}
+
 
 long sys32_vm86_warning(void)
 { 

arch/x86_64/kernel/head.S

@@ -365,7 +365,8 @@ ENTRY(cpu_gdt_table)
 	.quad	0,0			/* TSS */
 	.quad	0			/* LDT */
 	.quad   0,0,0			/* three TLS descriptors */ 
-	.quad	0,0			/* pad to cache line boundary */
+	.quad	0x00cff2000000ffff	/* dummy descriptor for long base */
+	.quad	0			/* pad to cache line boundary */
 gdt_end:	
 	.globl gdt_end
 

arch/x86_64/kernel/msr.c

@@ -79,7 +79,7 @@ static inline int rdmsr_eio(u32 reg, u32 *eax, u32 *edx)
 	       "	jmp 2b\n"
 	       ".previous\n"
 	       ".section __ex_table,\"a\"\n"
-	       "	.align 4\n"
+	       "	.align 8\n"
 	       "	.quad 1b,3b\n"
 	       ".previous"
 	       : "=&bDS" (err), "=a" (*eax), "=d" (*edx)

arch/x86_64/kernel/pci-gart.c

@@ -72,6 +72,7 @@ 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
@@ -453,10 +454,14 @@ void __init pci_iommu_init(void)
 	unsigned long aper_size;
 	unsigned long iommu_start;
 		
-#ifndef CONFIG_AGP
+#ifndef CONFIG_AGP_AMD_8151
 	no_agp = 1; 
 #else
-	no_agp = no_agp || (agp_init() < 0) || (agp_copy_info(&info) < 0); 
+	/* Add other K8 AGP bridge drivers here */
+	no_agp = no_agp || 
+		(agp_init() < 0) || 
+		(agp_amdk8_init() < 0) || 
+		(agp_copy_info(&info) < 0); 
 #endif	
 
 	if (no_iommu || (!force_mmu && end_pfn < 0xffffffff>>PAGE_SHIFT)) { 

arch/x86_64/kernel/process.c

@@ -534,43 +534,45 @@ unsigned long get_wchan(struct task_struct *p)
 int sys_arch_prctl(int code, unsigned long addr)
 { 
 	int ret = 0; 
-	unsigned long tmp; 
 
 	switch (code) { 
 	case ARCH_SET_GS:
 		if (addr >= TASK_SIZE) 
 			return -EPERM; 
-		load_gs_index(0);
-		current->thread.gsindex = 0;
+		get_cpu();
+		load_gs_index(__USER_LONGBASE);
+		current->thread.gsindex = __USER_LONGBASE;
 		current->thread.gs = addr;
 		ret = checking_wrmsrl(MSR_KERNEL_GS_BASE, addr); 
+		put_cpu();
 		break;
 	case ARCH_SET_FS:
 		/* Not strictly needed for fs, but do it for symmetry
 		   with gs */
 		if (addr >= TASK_SIZE)
 			return -EPERM; 
-		asm volatile("movl %0,%%fs" :: "r" (0));
-		current->thread.fsindex = 0;
+		get_cpu();
+		asm volatile("movl %0,%%fs" :: "r" (__USER_LONGBASE));
+		current->thread.fsindex = __USER_LONGBASE;
 		current->thread.fs = addr;
 		ret = checking_wrmsrl(MSR_FS_BASE, addr); 
+		put_cpu();
 		break;
 
 		/* Returned value may not be correct when the user changed fs/gs */ 
 	case ARCH_GET_FS:
-		rdmsrl(MSR_FS_BASE, tmp);
-		ret = put_user(tmp, (unsigned long *)addr); 
+		ret = put_user(current->thread.fs, (unsigned long *)addr); 
 		break; 
 
 	case ARCH_GET_GS: 
-		rdmsrl(MSR_KERNEL_GS_BASE, tmp); 
-		ret = put_user(tmp, (unsigned long *)addr); 
+		ret = put_user(current->thread.gs, (unsigned long *)addr); 
 		break;
 
 	default:
 		ret = -EINVAL;
 		break;
 	} 
+
 	return ret;	
 } 
 
@@ -657,6 +659,7 @@ asmlinkage int sys_set_thread_area(struct user_desc *u_info)
 #define GET_LIMIT_PAGES(desc)	(((desc)->b >> 23) & 1)
 #define GET_PRESENT(desc)	(((desc)->b >> 15) & 1)
 #define GET_USEABLE(desc)	(((desc)->b >> 20) & 1)
+#define GET_LONGMODE(desc)	(((desc)->b >> 21) & 1)
 
 asmlinkage int sys_get_thread_area(struct user_desc *u_info)
 {
@@ -681,6 +684,7 @@ asmlinkage int sys_get_thread_area(struct user_desc *u_info)
 	info.limit_in_pages = GET_LIMIT_PAGES(desc);
 	info.seg_not_present = !GET_PRESENT(desc);
 	info.useable = GET_USEABLE(desc);
+	info.lm = GET_LONGMODE(desc);
 
 	if (copy_to_user(u_info, &info, sizeof(info)))
 		return -EFAULT;

arch/x86_64/kernel/ptrace.c

@@ -91,6 +91,11 @@ static int putreg(struct task_struct *child,
 	unsigned long regno, unsigned long value)
 {
 	unsigned long tmp; 
+	
+	/* Some code in the 64bit emulation may not be 64bit clean.
+	   Don't take any chances. */
+	if (test_tsk_thread_flag(child, TIF_IA32))
+		value &= 0xffffffff;
 	switch (regno) {
 		case offsetof(struct user_regs_struct,fs):
 			if (value && (value & 3) != 3)
@@ -145,6 +150,7 @@ static int putreg(struct task_struct *child,
 
 static unsigned long getreg(struct task_struct *child, unsigned long regno)
 {
+	unsigned long val;
 	switch (regno) {
 		case offsetof(struct user_regs_struct, fs):
 			return child->thread.fsindex;
@@ -160,7 +166,10 @@ static unsigned long getreg(struct task_struct *child, unsigned long regno)
 			return child->thread.gs;
 		default:
 			regno = regno - sizeof(struct pt_regs);
-			return get_stack_long(child, regno);
+			val = get_stack_long(child, regno);
+			if (test_tsk_thread_flag(child, TIF_IA32))
+				val &= 0xffffffff;
+			return val;
 	}
 
 }

arch/x86_64/kernel/smpboot.c

@@ -799,11 +799,6 @@ static void __init smp_boot_cpus(unsigned int max_cpus)
 	printk("CPU%d: ", 0);
 	print_cpu_info(&cpu_data[0]);
 
-	/*
-	 * We have the boot CPU online for sure.
-	 */
-	set_bit(0, &cpu_online_map);
-	set_bit(0, &cpu_callout_map);
 	x86_apicid_to_cpu[boot_cpu_id] = 0;
 	x86_cpu_to_apicid[0] = boot_cpu_id;
 	current_thread_info()->cpu = 0;

arch/x86_64/kernel/sys_x86_64.c

@@ -18,6 +18,7 @@
 
 #include <asm/uaccess.h>
 #include <asm/ipc.h>
+#include <asm/ia32.h>
 
 /*
  * sys_pipe() is the normal C calling standard for creating
@@ -70,18 +71,27 @@ unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr, unsi
 	struct vm_area_struct *vma;
 	unsigned long end = TASK_SIZE;
 
-	if (test_thread_flag(TIF_IA32))
-		flags |= MAP_32BIT; 
-	if (flags & MAP_32BIT) 
-		end = 0xffffffff-1;
-	if (len > end)
-		return -ENOMEM;
-	if (!addr) { 
-		addr = TASK_UNMAPPED_64;
-		if (flags & MAP_32BIT) {
+	if (test_thread_flag(TIF_IA32)) { 
+		if (!addr) 
 			addr = TASK_UNMAPPED_32;
+		end = IA32_PAGE_OFFSET; 
+	} else if (flags & MAP_32BIT) { 
+		/* This is usually used needed to map code in small model, so it needs to 
+		   be in the first 31bit. Limit it to that.
+		   This means we need to move the unmapped base down for this case. This can 
+		   give conflicts with the heap, but we assume that glibc malloc knows how 
+		   to fall back to mmap. Give it 1GB of playground for now. -AK */ 
+		if (!addr) 
+			addr = 0x40000000; 
+		end = 0x80000000;		
+	} else { 
+		if (!addr) 
+			addr = TASK_UNMAPPED_64; 
+		end = TASK_SIZE; 
 		}
-	} 
+
+	if (len > end)
+		return -ENOMEM;
 	addr = PAGE_ALIGN(addr);
 
 	for (vma = find_vma(current->mm, addr); ; vma = vma->vm_next) {

arch/x86_64/kernel/traps.c

@@ -87,8 +87,9 @@ int printk_address(unsigned long address)
 	const char *symname;
 	char *modname;
 	char *delim = ":"; 
+	char namebuf[128];
 
-	symname = kallsyms_lookup(address, &symsize, &offset, &modname); 
+	symname = kallsyms_lookup(address, &symsize, &offset, &modname, namebuf); 
 	if (!symname) 
 		return printk("[<%016lx>]", address);
 	if (!modname) 

arch/x86_64/kernel/vsyscall.c

@@ -151,7 +151,6 @@ static void __init map_vsyscall(void)
 
 static int __init vsyscall_init(void)
 {
-	printk("VSYSCALL: consistency checks...");
 	if ((unsigned long) &vgettimeofday != VSYSCALL_ADDR(__NR_vgettimeofday))
 		panic("vgettimeofday link addr broken");
 	if ((unsigned long) &vtime != VSYSCALL_ADDR(__NR_vtime))
@@ -159,9 +158,7 @@ static int __init vsyscall_init(void)
 	if (VSYSCALL_ADDR(0) != __fix_to_virt(VSYSCALL_FIRST_PAGE))
 		panic("fixmap first vsyscall %lx should be %lx", __fix_to_virt(VSYSCALL_FIRST_PAGE),
 		      VSYSCALL_ADDR(0));
-	printk("passed...mapping...");
 	map_vsyscall();
-	printk("done.\n");
 
 	return 0;
 }

include/asm-x86_64/compat.h

@@ -10,16 +10,22 @@
 typedef u32		compat_size_t;
 typedef s32		compat_ssize_t;
 typedef s32		compat_time_t;
-typedef s32		compat_suseconds_t;
 typedef s32		compat_clock_t;
 typedef s32		compat_pid_t;
 typedef u16		compat_uid_t;
 typedef u16		compat_gid_t;
+typedef u32		compat_uid32_t;
+typedef u32		compat_gid32_t;
 typedef u16		compat_mode_t;
 typedef u32		compat_ino_t;
 typedef u16		compat_dev_t;
 typedef s32		compat_off_t;
+typedef s64		compat_loff_t;
 typedef u16		compat_nlink_t;
+typedef u16		compat_ipc_pid_t;
+typedef s32		compat_daddr_t;
+typedef u32		compat_caddr_t;
+typedef __kernel_fsid_t	compat_fsid_t;
 
 struct compat_timespec {
 	compat_time_t	tv_sec;

include/asm-x86_64/ia32.h

@@ -11,17 +11,6 @@
  * 32 bit structures for IA32 support.
  */
 
-/* 32bit compatibility types */
-typedef unsigned short	       __kernel_ipc_pid_t32;
-typedef unsigned 				__kernel_uid32_t32;
-typedef unsigned 				__kernel_gid32_t32;
-typedef unsigned short	       __kernel_umode_t32;
-typedef int		       __kernel_daddr_t32;
-typedef unsigned int	       __kernel_caddr_t32;
-typedef long		       __kernel_loff_t32;
-typedef __kernel_fsid_t	       __kernel_fsid_t32;
-
-
 /* fcntl.h */
 struct flock32 {
        short l_type;
@@ -130,7 +119,7 @@ struct statfs32 {
        int f_bavail;
        int f_files;
        int f_ffree;
-       __kernel_fsid_t32 f_fsid;
+       compat_fsid_t f_fsid;
        int f_namelen;  /* SunOS ignores this field. */
        int f_spare[6];
 };

include/asm-x86_64/segment.h

@@ -6,6 +6,8 @@
 
 #define __KERNEL32_CS   0x38
 
+#define __USER_LONGBASE	((GDT_ENTRY_LONGBASE * 8)  | 3)
+
 /* 
  * we cannot use the same code segment descriptor for user and kernel
  * even not in the long flat model, because of different DPL /kkeil 
@@ -23,6 +25,7 @@
 #define GDT_ENTRY_LDT 10
 #define GDT_ENTRY_TLS_MIN 11
 #define GDT_ENTRY_TLS_MAX 13
+#define GDT_ENTRY_LONGBASE 14
 
 #define GDT_ENTRY_TLS_ENTRIES 3
 

include/asm-x86_64/smp.h

@@ -83,16 +83,17 @@ extern inline int find_first_cpu(void)
 	return ffz(~cpu_online_map); 	
 } 
 
+/* RED-PEN different from i386 */
 #define for_each_cpu(i) \
 	for((i) = find_first_cpu(); (i)>=0; (i)=find_next_cpu(i))
 
-extern volatile unsigned long cpu_callout_map;
-/* We don't mark CPUs online until __cpu_up(), so we need another measure */
 static inline int num_booting_cpus(void)
 {
 	return hweight32(cpu_callout_map);
 }
 
+extern volatile unsigned long cpu_callout_map;
+
 /*
  * Some lowlevel functions might want to know about
  * the real APIC ID <-> CPU # mapping.

include/asm-x86_64/uaccess.h

@@ -240,7 +240,7 @@ extern unsigned long copy_user_generic(void *to, const void *from, unsigned len)
 extern unsigned long copy_to_user(void *to, const void *from, unsigned len); 
 extern unsigned long copy_from_user(void *to, const void *from, unsigned len); 
 
-static inline int __copy_from_user(void *dst, void *src, unsigned size) 
+static inline int __copy_from_user(void *dst, const void *src, unsigned size) 
 { 
 	if (!__builtin_constant_p(size))
 		return copy_user_generic(dst,src,size);
@@ -269,7 +269,7 @@ static inline int __copy_from_user(void *dst, void *src, unsigned size)
 	}
 }	
 
-static inline int __copy_to_user(void *dst, void *src, unsigned size) 
+static inline int __copy_to_user(void *dst, const void *src, unsigned size) 
 { 
 	if (!__builtin_constant_p(size))
 		return copy_user_generic(dst,src,size);