Merge with rsync://rsync.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git

arch/arm/boot/compressed/head-xscale.S

@@ -47,3 +47,10 @@ __XScale_start:
                orr     r7, r7, #(MACH_TYPE_GTWX5715 & 0xff00)
 #endif
 
+#ifdef CONFIG_ARCH_IXP2000
+		mov	r1, #-1
+		mov	r0, #0xd6000000
+		str	r1, [r0, #0x14]
+		str	r1, [r0, #0x18]
+#endif
+

arch/arm/kernel/entry-armv.S

@@ -269,7 +269,7 @@ __pabt_svc:
 	add	r5, sp, #S_PC
 	ldmia	r7, {r2 - r4}			@ Get USR pc, cpsr
 
-#if __LINUX_ARM_ARCH__ < 6
+#if __LINUX_ARM_ARCH__ < 6 && !defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
 	@ make sure our user space atomic helper is aborted
 	cmp	r2, #VIRT_OFFSET
 	bichs	r3, r3, #PSR_Z_BIT
@@ -616,11 +616,17 @@ __kuser_helper_start:
 
 __kuser_cmpxchg:				@ 0xffff0fc0
 
-#if __LINUX_ARM_ARCH__ < 6
+#if defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
 
-#ifdef CONFIG_SMP  /* sanity check */
-#error "CONFIG_SMP on a machine supporting pre-ARMv6 processors?"
-#endif
+	/*
+	 * Poor you.  No fast solution possible...
+	 * The kernel itself must perform the operation.
+	 * A special ghost syscall is used for that (see traps.c).
+	 */
+	swi	#0x9ffff0
+	mov	pc, lr
+
+#elif __LINUX_ARM_ARCH__ < 6
 
 	/*
 	 * Theory of operation:

arch/arm/kernel/traps.c

@@ -464,6 +464,55 @@ asmlinkage int arm_syscall(int no, struct pt_regs *regs)
 #endif
 		return 0;
 
+#ifdef CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG
+	/*
+	 * Atomically store r1 in *r2 if *r2 is equal to r0 for user space.
+	 * Return zero in r0 if *MEM was changed or non-zero if no exchange
+	 * happened.  Also set the user C flag accordingly.
+	 * If access permissions have to be fixed up then non-zero is
+	 * returned and the operation has to be re-attempted.
+	 *
+	 * *NOTE*: This is a ghost syscall private to the kernel.  Only the
+	 * __kuser_cmpxchg code in entry-armv.S should be aware of its
+	 * existence.  Don't ever use this from user code.
+	 */
+	case 0xfff0:
+	{
+		extern void do_DataAbort(unsigned long addr, unsigned int fsr,
+					 struct pt_regs *regs);
+		unsigned long val;
+		unsigned long addr = regs->ARM_r2;
+		struct mm_struct *mm = current->mm;
+		pgd_t *pgd; pmd_t *pmd; pte_t *pte;
+
+		regs->ARM_cpsr &= ~PSR_C_BIT;
+		spin_lock(&mm->page_table_lock);
+		pgd = pgd_offset(mm, addr);
+		if (!pgd_present(*pgd))
+			goto bad_access;
+		pmd = pmd_offset(pgd, addr);
+		if (!pmd_present(*pmd))
+			goto bad_access;
+		pte = pte_offset_map(pmd, addr);
+		if (!pte_present(*pte) || !pte_write(*pte))
+			goto bad_access;
+		val = *(unsigned long *)addr;
+		val -= regs->ARM_r0;
+		if (val == 0) {
+			*(unsigned long *)addr = regs->ARM_r1;
+			regs->ARM_cpsr |= PSR_C_BIT;
+		}
+		spin_unlock(&mm->page_table_lock);
+		return val;
+
+		bad_access:
+		spin_unlock(&mm->page_table_lock);
+		/* simulate a read access fault */
+		do_DataAbort(addr, 15 + (1 << 11), regs);
+		return -1;
+	}
+#endif
+
 	default:
 		/* Calls 9f00xx..9f07ff are defined to return -ENOSYS
 		   if not implemented, rather than raising SIGILL.  This

arch/arm/lib/io-writesw-armv4.S

@@ -87,9 +87,9 @@ ENTRY(__raw_writesw)
 		subs	r2, r2, #2
 		orr	ip, ip, r3, push_hbyte1
 		strh	ip, [r0]
-		bpl	2b
+		bpl	1b
 
-3:		tst	r2, #1
-2:		movne	ip, r3, lsr #8
+		tst	r2, #1
+3:		movne	ip, r3, lsr #8
 		strneh	ip, [r0]
 		mov	pc, lr

arch/arm/mach-pxa/mainstone.c

@@ -304,6 +304,15 @@ static void __init mainstone_map_io(void)
 	PWER  = 0xC0000002;
 	PRER  = 0x00000002;
 	PFER  = 0x00000002;
+ 	/*	for use I SRAM as framebuffer.	*/
+ 	PSLR |= 0xF04;
+ 	PCFR = 0x66;
+ 	/*	For Keypad wakeup.	*/
+ 	KPC &=~KPC_ASACT;
+ 	KPC |=KPC_AS;
+ 	PKWR  = 0x000FD000;
+ 	/*	Need read PKWR back after set it.	*/
+ 	PKWR;
 }
 
 MACHINE_START(MAINSTONE, "Intel HCDDBBVA0 Development Platform (aka Mainstone)")

arch/arm/mach-pxa/pm.c

@@ -29,9 +29,6 @@
  */
 #undef DEBUG
 
-extern void pxa_cpu_suspend(void);
-extern void pxa_cpu_resume(void);
-
 #define SAVE(x)		sleep_save[SLEEP_SAVE_##x] = x
 #define RESTORE(x)	x = sleep_save[SLEEP_SAVE_##x]
 
@@ -63,6 +60,12 @@ enum {	SLEEP_SAVE_START = 0,
 	SLEEP_SAVE_ICMR,
 	SLEEP_SAVE_CKEN,
 
+#ifdef CONFIG_PXA27x
+ 	SLEEP_SAVE_MDREFR,
+ 	SLEEP_SAVE_PWER, SLEEP_SAVE_PCFR, SLEEP_SAVE_PRER,
+ 	SLEEP_SAVE_PFER, SLEEP_SAVE_PKWR,
+#endif
+
 	SLEEP_SAVE_CKSUM,
 
 	SLEEP_SAVE_SIZE
@@ -75,9 +78,7 @@ static int pxa_pm_enter(suspend_state_t state)
 	unsigned long checksum = 0;
 	struct timespec delta, rtc;
 	int i;
-
-	if (state != PM_SUSPEND_MEM)
-		return -EINVAL;
+	extern void pxa_cpu_pm_enter(suspend_state_t state);
 
 #ifdef CONFIG_IWMMXT
 	/* force any iWMMXt context to ram **/
@@ -100,16 +101,17 @@ static int pxa_pm_enter(suspend_state_t state)
 	SAVE(GAFR2_L); SAVE(GAFR2_U);
 
 #ifdef CONFIG_PXA27x
+	SAVE(MDREFR);
 	SAVE(GPLR3); SAVE(GPDR3); SAVE(GRER3); SAVE(GFER3); SAVE(PGSR3);
 	SAVE(GAFR3_L); SAVE(GAFR3_U);
+	SAVE(PWER); SAVE(PCFR); SAVE(PRER);
+	SAVE(PFER); SAVE(PKWR);
 #endif
 
 	SAVE(ICMR);
 	ICMR = 0;
 
 	SAVE(CKEN);
-	CKEN = 0;
-
 	SAVE(PSTR);
 
 	/* Note: wake up source are set up in each machine specific files */
@@ -123,16 +125,13 @@ static int pxa_pm_enter(suspend_state_t state)
 	/* Clear sleep reset status */
 	RCSR = RCSR_SMR;
 
-	/* set resume return address */
-	PSPR = virt_to_phys(pxa_cpu_resume);
-
 	/* before sleeping, calculate and save a checksum */
 	for (i = 0; i < SLEEP_SAVE_SIZE - 1; i++)
 		checksum += sleep_save[i];
 	sleep_save[SLEEP_SAVE_CKSUM] = checksum;
 
 	/* *** go zzz *** */
-	pxa_cpu_suspend();
+	pxa_cpu_pm_enter(state);
 
 	/* after sleeping, validate the checksum */
 	checksum = 0;
@@ -145,7 +144,7 @@ static int pxa_pm_enter(suspend_state_t state)
 		LUB_HEXLED = 0xbadbadc5;
 #endif
 		while (1)
-			pxa_cpu_suspend();
+			pxa_cpu_pm_enter(state);
 	}
 
 	/* ensure not to come back here if it wasn't intended */
@@ -162,8 +161,11 @@ static int pxa_pm_enter(suspend_state_t state)
 	RESTORE(PGSR0); RESTORE(PGSR1); RESTORE(PGSR2);
 
 #ifdef CONFIG_PXA27x
+	RESTORE(MDREFR);
 	RESTORE(GAFR3_L); RESTORE(GAFR3_U); RESTORE_GPLEVEL(3);
 	RESTORE(GPDR3); RESTORE(GRER3); RESTORE(GFER3); RESTORE(PGSR3);
+	RESTORE(PWER); RESTORE(PCFR); RESTORE(PRER);
+	RESTORE(PFER); RESTORE(PKWR);
 #endif
 
 	PSSR = PSSR_RDH | PSSR_PH;
@@ -197,7 +199,9 @@ unsigned long sleep_phys_sp(void *sp)
  */
 static int pxa_pm_prepare(suspend_state_t state)
 {
-	return 0;
+	extern int pxa_cpu_pm_prepare(suspend_state_t state);
+
+	return pxa_cpu_pm_prepare(state);
 }
 
 /*

arch/arm/mach-pxa/pxa25x.c

@@ -102,3 +102,32 @@ unsigned int get_lcdclk_frequency_10khz(void)
 }
 
 EXPORT_SYMBOL(get_lcdclk_frequency_10khz);
+
+
+int pxa_cpu_pm_prepare(suspend_state_t state)
+{
+	switch (state) {
+	case PM_SUSPEND_MEM:
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void pxa_cpu_pm_enter(suspend_state_t state)
+{
+	extern void pxa_cpu_suspend(unsigned int);
+	extern void pxa_cpu_resume(void);
+
+	CKEN = 0;
+
+	switch (state) {
+	case PM_SUSPEND_MEM:
+		/* set resume return address */
+		PSPR = virt_to_phys(pxa_cpu_resume);
+		pxa_cpu_suspend(3);
+		break;
+	}
+}

arch/arm/mach-pxa/pxa27x.c

@@ -120,6 +120,38 @@ EXPORT_SYMBOL(get_clk_frequency_khz);
 EXPORT_SYMBOL(get_memclk_frequency_10khz);
 EXPORT_SYMBOL(get_lcdclk_frequency_10khz);
 
+int pxa_cpu_pm_prepare(suspend_state_t state)
+{
+	switch (state) {
+	case PM_SUSPEND_MEM:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+void pxa_cpu_pm_enter(suspend_state_t state)
+{
+	extern void pxa_cpu_standby(void);
+	extern void pxa_cpu_suspend(unsigned int);
+	extern void pxa_cpu_resume(void);
+
+	CKEN = CKEN22_MEMC | CKEN9_OSTIMER;
+
+	/* ensure voltage-change sequencer not initiated, which hangs */
+	PCFR &= ~PCFR_FVC;
+
+	/* Clear edge-detect status register. */
+	PEDR = 0xDF12FE1B;
+
+	switch (state) {
+	case PM_SUSPEND_MEM:
+		/* set resume return address */
+		PSPR = virt_to_phys(pxa_cpu_resume);
+		pxa_cpu_suspend(3);
+		break;
+	}
+}
 
 /*
  * device registration specific to PXA27x.

arch/arm/mach-s3c2410/dma.c

@@ -785,6 +785,10 @@ int s3c2410_dma_free(dmach_t channel, s3c2410_dma_client_t *client)
 	chan->client = NULL;
 	chan->in_use = 0;
 
+	if (chan->irq_claimed)
+		free_irq(chan->irq, (void *)chan);
+	chan->irq_claimed = 0;
+
 	local_irq_restore(flags);
 
 	return 0;

arch/arm/mm/Kconfig

@@ -228,7 +228,6 @@ config CPU_SA1100
 	select CPU_CACHE_V4WB
 	select CPU_CACHE_VIVT
 	select CPU_TLB_V4WB
-	select CPU_MINICACHE
 
 # XScale
 config CPU_XSCALE
@@ -239,7 +238,6 @@ config CPU_XSCALE
 	select CPU_ABRT_EV5T
 	select CPU_CACHE_VIVT
 	select CPU_TLB_V4WBI
-	select CPU_MINICACHE
 
 # ARMv6
 config CPU_V6
@@ -345,11 +343,6 @@ config CPU_TLB_V4WBI
 config CPU_TLB_V6
 	bool
 
-config CPU_MINICACHE
-	bool
-	help
-	  Processor has a minicache.
-
 comment "Processor Features"
 
 config ARM_THUMB
@@ -429,3 +422,11 @@ config HAS_TLS_REG
 	  assume directly accessing that register and always obtain the
 	  expected value only on ARMv7 and above.
 
+config NEEDS_SYSCALL_FOR_CMPXCHG
+	bool
+	default y if SMP && (CPU_32v5 || CPU_32v4 || CPU_32v3)
+	help
+	  SMP on a pre-ARMv6 processor?  Well OK then.
+	  Forget about fast user space cmpxchg support.
+	  It is just not possible.
+

arch/arm/mm/Makefile

@@ -31,8 +31,6 @@ obj-$(CONFIG_CPU_COPY_V6)	+= copypage-v6.o mmu.o
 obj-$(CONFIG_CPU_SA1100)	+= copypage-v4mc.o
 obj-$(CONFIG_CPU_XSCALE)	+= copypage-xscale.o
 
-obj-$(CONFIG_CPU_MINICACHE)	+= minicache.o
-
 obj-$(CONFIG_CPU_TLB_V3)	+= tlb-v3.o
 obj-$(CONFIG_CPU_TLB_V4WT)	+= tlb-v4.o
 obj-$(CONFIG_CPU_TLB_V4WB)	+= tlb-v4wb.o

arch/arm/mm/copypage-xscale.S

-/*
- *  linux/arch/arm/lib/copypage-xscale.S
- *
- *  Copyright (C) 2001 Russell King
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/constants.h>
-
-/*
- * General note:
- *  We don't really want write-allocate cache behaviour for these functions
- *  since that will just eat through 8K of the cache.
- */
-
-	.text
-	.align	5
-/*
- * XScale optimised copy_user_page
- *  r0 = destination
- *  r1 = source
- *  r2 = virtual user address of ultimate destination page
- *
- * The source page may have some clean entries in the cache already, but we
- * can safely ignore them - break_cow() will flush them out of the cache
- * if we eventually end up using our copied page.
- *
- * What we could do is use the mini-cache to buffer reads from the source
- * page.  We rely on the mini-cache being smaller than one page, so we'll
- * cycle through the complete cache anyway.
- */
-ENTRY(xscale_mc_copy_user_page)
-	stmfd	sp!, {r4, r5, lr}
-	mov	r5, r0
-	mov	r0, r1
-	bl	map_page_minicache
-	mov	r1, r5
-	mov	lr, #PAGE_SZ/64-1
-
-	/*
-	 * Strangely enough, best performance is achieved
-	 * when prefetching destination as well.  (NP)
-	 */
-	pld	[r0, #0]
-	pld	[r0, #32]
-	pld	[r1, #0]
-	pld	[r1, #32]
-
-1:	pld	[r0, #64]
-	pld	[r0, #96]
-	pld	[r1, #64]
-	pld	[r1, #96]
-
-2:	ldrd	r2, [r0], #8
-	ldrd	r4, [r0], #8
-	mov	ip, r1
-	strd	r2, [r1], #8
-	ldrd	r2, [r0], #8
-	strd	r4, [r1], #8
-	ldrd	r4, [r0], #8
-	strd	r2, [r1], #8
-	strd	r4, [r1], #8
-	mcr	p15, 0, ip, c7, c10, 1		@ clean D line
-	ldrd	r2, [r0], #8
-	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line
-	ldrd	r4, [r0], #8
-	mov	ip, r1
-	strd	r2, [r1], #8
-	ldrd	r2, [r0], #8
-	strd	r4, [r1], #8
-	ldrd	r4, [r0], #8
-	strd	r2, [r1], #8
-	strd	r4, [r1], #8
-	mcr	p15, 0, ip, c7, c10, 1		@ clean D line
-	subs	lr, lr, #1
-	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line
-	bgt	1b
-	beq	2b
-
-	ldmfd	sp!, {r4, r5, pc}
-
-	.align	5
-/*
- * XScale optimised clear_user_page
- *  r0 = destination
- *  r1 = virtual user address of ultimate destination page
- */
-ENTRY(xscale_mc_clear_user_page)
-	mov	r1, #PAGE_SZ/32
-	mov	r2, #0
-	mov	r3, #0
-1:	mov	ip, r0
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	mcr	p15, 0, ip, c7, c10, 1		@ clean D line
-	subs	r1, r1, #1
-	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line
-	bne	1b
-	mov	pc, lr
-
-	__INITDATA
-
-	.type	xscale_mc_user_fns, #object
-ENTRY(xscale_mc_user_fns)
-	.long	xscale_mc_clear_user_page
-	.long	xscale_mc_copy_user_page
-	.size	xscale_mc_user_fns, . - xscale_mc_user_fns

arch/arm/mm/copypage-xscale.c

+/*
+ *  linux/arch/arm/lib/copypage-xscale.S
+ *
+ *  Copyright (C) 1995-2005 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This handles the mini data cache, as found on SA11x0 and XScale
+ * processors.  When we copy a user page page, we map it in such a way
+ * that accesses to this page will not touch the main data cache, but
+ * will be cached in the mini data cache.  This prevents us thrashing
+ * the main data cache on page faults.
+ */
+#include <linux/init.h>
+#include <linux/mm.h>
+
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/tlbflush.h>
+
+/*
+ * 0xffff8000 to 0xffffffff is reserved for any ARM architecture
+ * specific hacks for copying pages efficiently.
+ */
+#define COPYPAGE_MINICACHE	0xffff8000
+
+#define minicache_pgprot __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | \
+				  L_PTE_CACHEABLE)
+
+#define TOP_PTE(x)	pte_offset_kernel(top_pmd, x)
+
+static DEFINE_SPINLOCK(minicache_lock);
+
+/*
+ * XScale mini-dcache optimised copy_user_page
+ *
+ * We flush the destination cache lines just before we write the data into the
+ * corresponding address.  Since the Dcache is read-allocate, this removes the
+ * Dcache aliasing issue.  The writes will be forwarded to the write buffer,
+ * and merged as appropriate.
+ */
+static void __attribute__((naked))
+mc_copy_user_page(void *from, void *to)
+{
+	/*
+	 * Strangely enough, best performance is achieved
+	 * when prefetching destination as well.  (NP)
+	 */
+	asm volatile(
+	"stmfd	sp!, {r4, r5, lr}		\n\
+	mov	lr, %2				\n\
+	pld	[r0, #0]			\n\
+	pld	[r0, #32]			\n\
+	pld	[r1, #0]			\n\
+	pld	[r1, #32]			\n\
+1:	pld	[r0, #64]			\n\
+	pld	[r0, #96]			\n\
+	pld	[r1, #64]			\n\
+	pld	[r1, #96]			\n\
+2:	ldrd	r2, [r0], #8			\n\
+	ldrd	r4, [r0], #8			\n\
+	mov	ip, r1				\n\
+	strd	r2, [r1], #8			\n\
+	ldrd	r2, [r0], #8			\n\
+	strd	r4, [r1], #8			\n\
+	ldrd	r4, [r0], #8			\n\
+	strd	r2, [r1], #8			\n\
+	strd	r4, [r1], #8			\n\
+	mcr	p15, 0, ip, c7, c10, 1		@ clean D line\n\
+	ldrd	r2, [r0], #8			\n\
+	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line\n\
+	ldrd	r4, [r0], #8			\n\
+	mov	ip, r1				\n\
+	strd	r2, [r1], #8			\n\
+	ldrd	r2, [r0], #8			\n\
+	strd	r4, [r1], #8			\n\
+	ldrd	r4, [r0], #8			\n\
+	strd	r2, [r1], #8			\n\
+	strd	r4, [r1], #8			\n\
+	mcr	p15, 0, ip, c7, c10, 1		@ clean D line\n\
+	subs	lr, lr, #1			\n\
+	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line\n\
+	bgt	1b				\n\
+	beq	2b				\n\
+	ldmfd	sp!, {r4, r5, pc}		"
+	:
+	: "r" (from), "r" (to), "I" (PAGE_SIZE / 64 - 1));
+}
+
+void xscale_mc_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	spin_lock(&minicache_lock);
+
+	set_pte(TOP_PTE(COPYPAGE_MINICACHE), pfn_pte(__pa(kfrom) >> PAGE_SHIFT, minicache_pgprot));
+	flush_tlb_kernel_page(COPYPAGE_MINICACHE);
+
+	mc_copy_user_page((void *)COPYPAGE_MINICACHE, kto);
+
+	spin_unlock(&minicache_lock);
+}
+
+/*
+ * XScale optimised clear_user_page
+ */
+void __attribute__((naked))
+xscale_mc_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm volatile(
+	"mov	r1, %0				\n\
+	mov	r2, #0				\n\
+	mov	r3, #0				\n\
+1:	mov	ip, r0				\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	mcr	p15, 0, ip, c7, c10, 1		@ clean D line\n\
+	subs	r1, r1, #1			\n\
+	mcr	p15, 0, ip, c7, c6, 1		@ invalidate D line\n\
+	bne	1b				\n\
+	mov	pc, lr"
+	:
+	: "I" (PAGE_SIZE / 32));
+}
+
+struct cpu_user_fns xscale_mc_user_fns __initdata = {
+	.cpu_clear_user_page	= xscale_mc_clear_user_page, 
+	.cpu_copy_user_page	= xscale_mc_copy_user_page,
+};

arch/arm/mm/minicache.c

-/*
- *  linux/arch/arm/mm/minicache.c
- *
- *  Copyright (C) 2001 Russell King
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- *
- * This handles the mini data cache, as found on SA11x0 and XScale
- * processors.  When we copy a user page page, we map it in such a way
- * that accesses to this page will not touch the main data cache, but
- * will be cached in the mini data cache.  This prevents us thrashing
- * the main data cache on page faults.
- */
-#include <linux/init.h>
-#include <linux/mm.h>
-
-#include <asm/page.h>
-#include <asm/pgtable.h>
-#include <asm/tlbflush.h>
-
-/*
- * 0xffff8000 to 0xffffffff is reserved for any ARM architecture
- * specific hacks for copying pages efficiently.
- */
-#define minicache_address (0xffff8000)
-#define minicache_pgprot __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | \
-				  L_PTE_CACHEABLE)
-
-static pte_t *minicache_pte;
-
-/*
- * Note that this is intended to be called only from the copy_user_page
- * asm code; anything else will require special locking to prevent the
- * mini-cache space being re-used.  (Note: probably preempt unsafe).
- *
- * We rely on the fact that the minicache is 2K, and we'll be pushing
- * 4K of data through it, so we don't actually have to specifically
- * flush the minicache when we change the mapping.
- *
- * Note also: assert(PAGE_OFFSET <= virt < high_memory).
- * Unsafe: preempt, kmap.
- */
-unsigned long map_page_minicache(unsigned long virt)
-{
-	set_pte(minicache_pte, pfn_pte(__pa(virt) >> PAGE_SHIFT, minicache_pgprot));
-	flush_tlb_kernel_page(minicache_address);
-
-	return minicache_address;
-}
-
-static int __init minicache_init(void)
-{
-	pgd_t *pgd;
-	pmd_t *pmd;
-
-	spin_lock(&init_mm.page_table_lock);
-
-	pgd = pgd_offset_k(minicache_address);
-	pmd = pmd_alloc(&init_mm, pgd, minicache_address);
-	if (!pmd)
-		BUG();
-	minicache_pte = pte_alloc_kernel(&init_mm, pmd, minicache_address);
-	if (!minicache_pte)
-		BUG();
-
-	spin_unlock(&init_mm.page_table_lock);
-
-	return 0;
-}
-
-core_initcall(minicache_init);

arch/i386/kernel/Makefile

@@ -43,7 +43,7 @@ obj-$(CONFIG_SCx200)		+= scx200.o
 # Note: kbuild does not track this dependency due to usage of .incbin
 $(obj)/vsyscall.o: $(obj)/vsyscall-int80.so $(obj)/vsyscall-sysenter.so
 targets += $(foreach F,int80 sysenter,vsyscall-$F.o vsyscall-$F.so)
-targets += vsyscall.lds
+targets += vsyscall-note.o vsyscall.lds
 
 # The DSO images are built using a special linker script.
 quiet_cmd_syscall = SYSCALL $@

arch/ia64/kernel/module.c

@@ -825,14 +825,16 @@ apply_relocate_add (Elf64_Shdr *sechdrs, const char *strtab, unsigned int symind
 		 * XXX Should have an arch-hook for running this after final section
 		 *     addresses have been selected...
 		 */
-		/* See if gp can cover the entire core module:  */
-		uint64_t gp = (uint64_t) mod->module_core + MAX_LTOFF / 2;
-		if (mod->core_size >= MAX_LTOFF)
+		uint64_t gp;
+		if (mod->core_size > MAX_LTOFF)
 			/*
 			 * This takes advantage of fact that SHF_ARCH_SMALL gets allocated
 			 * at the end of the module.
 			 */
-			gp = (uint64_t) mod->module_core + mod->core_size - MAX_LTOFF / 2;
+			gp = mod->core_size - MAX_LTOFF / 2;
+		else
+			gp = mod->core_size / 2;
+		gp = (uint64_t) mod->module_core + ((gp + 7) & -8);
 		mod->arch.gp = gp;
 		DEBUGP("%s: placing gp at 0x%lx\n", __FUNCTION__, gp);
 	}

arch/ia64/kernel/ptrace.c

@@ -635,11 +635,17 @@ ia64_flush_fph (struct task_struct *task)
 {
 	struct ia64_psr *psr = ia64_psr(ia64_task_regs(task));
 
+	/*
+	 * Prevent migrating this task while
+	 * we're fiddling with the FPU state
+	 */
+	preempt_disable();
 	if (ia64_is_local_fpu_owner(task) && psr->mfh) {
 		psr->mfh = 0;
 		task->thread.flags |= IA64_THREAD_FPH_VALID;
 		ia64_save_fpu(&task->thread.fph[0]);
 	}
+	preempt_enable();
 }
 
 /*

arch/ia64/kernel/setup.c

@@ -720,7 +720,8 @@ cpu_init (void)
 	ia64_set_kr(IA64_KR_PT_BASE, __pa(ia64_imva(empty_zero_page)));
 
 	/*
-	 * Initialize default control register to defer all speculative faults.  The
+	 * Initialize default control register to defer speculative faults except
+	 * for those arising from TLB misses, which are not deferred.  The
 	 * kernel MUST NOT depend on a particular setting of these bits (in other words,
 	 * the kernel must have recovery code for all speculative accesses).  Turn on
 	 * dcr.lc as per recommendation by the architecture team.  Most IA-32 apps

arch/ia64/kernel/traps.c

@@ -111,6 +111,24 @@ ia64_bad_break (unsigned long break_num, struct pt_regs *regs)
 	siginfo_t siginfo;
 	int sig, code;
 
+	/* break.b always sets cr.iim to 0, which causes problems for
+	 * debuggers.  Get the real break number from the original instruction,
+	 * but only for kernel code.  User space break.b is left alone, to
+	 * preserve the existing behaviour.  All break codings have the same
+	 * format, so there is no need to check the slot type.
+	 */
+	if (break_num == 0 && !user_mode(regs)) {
+		struct ia64_psr *ipsr = ia64_psr(regs);
+		unsigned long *bundle = (unsigned long *)regs->cr_iip;
+		unsigned long slot;
+		switch (ipsr->ri) {
+		      case 0:  slot = (bundle[0] >>  5); break;
+		      case 1:  slot = (bundle[0] >> 46) | (bundle[1] << 18); break;
+		      default: slot = (bundle[1] >> 23); break;
+		}
+		break_num = ((slot >> 36 & 1) << 20) | (slot >> 6 & 0xfffff);
+	}
+
 	/* SIGILL, SIGFPE, SIGSEGV, and SIGBUS want these field initialized: */
 	siginfo.si_addr = (void __user *) (regs->cr_iip + ia64_psr(regs)->ri);
 	siginfo.si_imm = break_num;
@@ -202,13 +220,21 @@ disabled_fph_fault (struct pt_regs *regs)
 
 	/* first, grant user-level access to fph partition: */
 	psr->dfh = 0;
+
+	/*
+	 * Make sure that no other task gets in on this processor
+	 * while we're claiming the FPU
+	 */
+	preempt_disable();
 #ifndef CONFIG_SMP
 	{
 		struct task_struct *fpu_owner
 			= (struct task_struct *)ia64_get_kr(IA64_KR_FPU_OWNER);
 
-		if (ia64_is_local_fpu_owner(current))
+		if (ia64_is_local_fpu_owner(current)) {
+			preempt_enable_no_resched();
 			return;
+		}
 
 		if (fpu_owner)
 			ia64_flush_fph(fpu_owner);
@@ -226,6 +252,7 @@ disabled_fph_fault (struct pt_regs *regs)
 		 */
 		psr->mfh = 1;
 	}
+	preempt_enable_no_resched();
 }
 
 static inline int

arch/ia64/mm/init.c

@@ -305,8 +305,9 @@ setup_gate (void)
 	struct page *page;
 
 	/*
-	 * Map the gate page twice: once read-only to export the ELF headers etc. and once
-	 * execute-only page to enable privilege-promotion via "epc":
+	 * Map the gate page twice: once read-only to export the ELF
+	 * headers etc. and once execute-only page to enable
+	 * privilege-promotion via "epc":
 	 */
 	page = virt_to_page(ia64_imva(__start_gate_section));
 	put_kernel_page(page, GATE_ADDR, PAGE_READONLY);
@@ -315,6 +316,20 @@ setup_gate (void)
 	put_kernel_page(page, GATE_ADDR + PAGE_SIZE, PAGE_GATE);
 #else
 	put_kernel_page(page, GATE_ADDR + PERCPU_PAGE_SIZE, PAGE_GATE);
+	/* Fill in the holes (if any) with read-only zero pages: */
+	{
+		unsigned long addr;
+
+		for (addr = GATE_ADDR + PAGE_SIZE;
+		     addr < GATE_ADDR + PERCPU_PAGE_SIZE;
+		     addr += PAGE_SIZE)
+		{
+			put_kernel_page(ZERO_PAGE(0), addr,
+					PAGE_READONLY);
+			put_kernel_page(ZERO_PAGE(0), addr + PERCPU_PAGE_SIZE,
+					PAGE_READONLY);
+		}
+	}
 #endif
 	ia64_patch_gate();
 }