Merge branch 'master' into upstream

Documentation/memory-barriers.txt

@@ -19,6 +19,7 @@ Contents:
      - Control dependencies.
      - SMP barrier pairing.
      - Examples of memory barrier sequences.
+     - Read memory barriers vs load speculation.
 
  (*) Explicit kernel barriers.
 
@@ -248,7 +249,7 @@ And there are a number of things that _must_ or _must_not_ be assumed:
      we may get either of:
 
 	STORE *A = X; Y = LOAD *A;
-	STORE *A = Y;
+	STORE *A = Y = X;
 
 
 =========================
@@ -344,9 +345,12 @@ Memory barriers come in four basic varieties:
 
  (4) General memory barriers.
 
-     A general memory barrier is a combination of both a read memory barrier
-     and a write memory barrier.  It is a partial ordering over both loads and
-     stores.
+     A general memory barrier gives a guarantee that all the LOAD and STORE
+     operations specified before the barrier will appear to happen before all
+     the LOAD and STORE operations specified after the barrier with respect to
+     the other components of the system.
+
+     A general memory barrier is a partial ordering over both loads and stores.
 
      General memory barriers imply both read and write memory barriers, and so
      can substitute for either.
@@ -546,9 +550,9 @@ write barrier, though, again, a general barrier is viable:
 	===============	===============
 	a = 1;
 	<write barrier>
-	b = 2;		x = a;
+	b = 2;		x = b;
 			<read barrier>
-			y = b;
+			y = a;
 
 Or:
 
@@ -563,6 +567,18 @@ Or:
 Basically, the read barrier always has to be there, even though it can be of
 the "weaker" type.
 
+[!] Note that the stores before the write barrier would normally be expected to
+match the loads after the read barrier or data dependency barrier, and vice
+versa:
+
+	CPU 1                           CPU 2
+	===============                 ===============
+	a = 1;           }----   --->{  v = c
+	b = 2;           }    \ /    {  w = d
+	<write barrier>        \        <read barrier>
+	c = 3;           }    / \    {  x = a;
+	d = 4;           }----   --->{  y = b;
+
 
 EXAMPLES OF MEMORY BARRIER SEQUENCES
 ------------------------------------
@@ -600,8 +616,8 @@ STORE B, STORE C } all occuring before the unordered set of { STORE D, STORE E
 	|       |       +------+
 	+-------+       :      :
 	                   |
-	                   | Sequence in which stores committed to memory system
-	                   | by CPU 1
+	                   | Sequence in which stores are committed to the
+	                   | memory system by CPU 1
 	                   V
 
 
@@ -683,14 +699,12 @@ then the following will occur:
 	                               |        :       :       |       |
 	                               |        :       :       | CPU 2 |
 	                               |        +-------+       |       |
-	                                \       | X->9  |------>|       |
-	                                 \      +-------+       |       |
-	                                  ----->| B->2  |       |       |
-	                                        +-------+       |       |
-	     Makes sure all effects --->    ddddddddddddddddd   |       |
-	     prior to the store of C            +-------+       |       |
-	     are perceptible to                 | B->2  |------>|       |
-	     successive loads                   +-------+       |       |
+	                               |        | X->9  |------>|       |
+	                               |        +-------+       |       |
+	  Makes sure all effects --->   \   ddddddddddddddddd   |       |
+	  prior to the store of C        \      +-------+       |       |
+	  are perceptible to              ----->| B->2  |------>|       |
+	  subsequent loads                      +-------+       |       |
 	                                        :       :       +-------+
 
 
@@ -699,73 +713,239 @@ following sequence of events:
 
 	CPU 1			CPU 2
 	=======================	=======================
+		{ A = 0, B = 9 }
 	STORE A=1
-	STORE B=2
-	STORE C=3
 	<write barrier>
-	STORE D=4
-	STORE E=5
-				LOAD A
+	STORE B=2
 				LOAD B
-				LOAD C
-				LOAD D
-				LOAD E
+				LOAD A
 
 Without intervention, CPU 2 may then choose to perceive the events on CPU 1 in
 some effectively random order, despite the write barrier issued by CPU 1:
 
-	+-------+       :      :
-	|       |       +------+
-	|       |------>| C=3  | }
-	|       |  :    +------+ }
-	|       |  :    | A=1  | }
-	|       |  :    +------+ }
-	| CPU 1 |  :    | B=2  | }---
-	|       |       +------+ }   \
-	|       |   wwwwwwwwwwwww}    \
-	|       |       +------+ }     \          :       :       +-------+
-	|       |  :    | E=5  | }      \         +-------+       |       |
-	|       |  :    +------+ }       \      { | C->3  |------>|       |
-	|       |------>| D=4  | }        \     { +-------+    :  |       |
-	|       |       +------+           \    { | E->5  |    :  |       |
-	+-------+       :      :            \   { +-------+    :  |       |
-	                           Transfer  -->{ | A->1  |    :  | CPU 2 |
-	                          from CPU 1    { +-------+    :  |       |
-	                           to CPU 2     { | D->4  |    :  |       |
-	                                        { +-------+    :  |       |
-	                                        { | B->2  |------>|       |
-	                                          +-------+       |       |
-	                                          :       :       +-------+
-
-
-If, however, a read barrier were to be placed between the load of C and the
-load of D on CPU 2, then the partial ordering imposed by CPU 1 will be
-perceived correctly by CPU 2.
+	+-------+       :      :                :       :
+	|       |       +------+                +-------+
+	|       |------>| A=1  |------      --->| A->0  |
+	|       |       +------+      \         +-------+
+	| CPU 1 |   wwwwwwwwwwwwwwww   \    --->| B->9  |
+	|       |       +------+        |       +-------+
+	|       |------>| B=2  |---     |       :       :
+	|       |       +------+   \    |       :       :       +-------+
+	+-------+       :      :    \   |       +-------+       |       |
+	                             ---------->| B->2  |------>|       |
+	                                |       +-------+       | CPU 2 |
+	                                |       | A->0  |------>|       |
+	                                |       +-------+       |       |
+	                                |       :       :       +-------+
+	                                 \      :       :
+	                                  \     +-------+
+	                                   ---->| A->1  |
+	                                        +-------+
+	                                        :       :
 
-	+-------+       :      :
-	|       |       +------+
-	|       |------>| C=3  | }
-	|       |  :    +------+ }
-	|       |  :    | A=1  | }---
-	|       |  :    +------+ }   \
-	| CPU 1 |  :    | B=2  | }    \
-	|       |       +------+       \
-	|       |   wwwwwwwwwwwwwwww    \
-	|       |       +------+         \        :       :       +-------+
-	|       |  :    | E=5  | }        \       +-------+       |       |
-	|       |  :    +------+ }---      \    { | C->3  |------>|       |
-	|       |------>| D=4  | }   \      \   { +-------+    :  |       |
-	|       |       +------+      \      -->{ | B->2  |    :  |       |
-	+-------+       :      :       \        { +-------+    :  |       |
-	                                \       { | A->1  |    :  | CPU 2 |
-	                                 \        +-------+       |       |
-	   At this point the read ---->   \   rrrrrrrrrrrrrrrrr   |       |
-	   barrier causes all effects      \      +-------+       |       |
-	   prior to the storage of C        \   { | E->5  |    :  |       |
-	   to be perceptible to CPU 2        -->{ +-------+    :  |       |
-	                                        { | D->4  |------>|       |
-	                                          +-------+       |       |
-	                                          :       :       +-------+
+
+If, however, a read barrier were to be placed between the load of E and the
+load of A on CPU 2:
+
+	CPU 1			CPU 2
+	=======================	=======================
+		{ A = 0, B = 9 }
+	STORE A=1
+	<write barrier>
+	STORE B=2
+				LOAD B
+				<read barrier>
+				LOAD A
+
+then the partial ordering imposed by CPU 1 will be perceived correctly by CPU
+2:
+
+	+-------+       :      :                :       :
+	|       |       +------+                +-------+
+	|       |------>| A=1  |------      --->| A->0  |
+	|       |       +------+      \         +-------+
+	| CPU 1 |   wwwwwwwwwwwwwwww   \    --->| B->9  |
+	|       |       +------+        |       +-------+
+	|       |------>| B=2  |---     |       :       :
+	|       |       +------+   \    |       :       :       +-------+
+	+-------+       :      :    \   |       +-------+       |       |
+	                             ---------->| B->2  |------>|       |
+	                                |       +-------+       | CPU 2 |
+	                                |       :       :       |       |
+	                                |       :       :       |       |
+	  At this point the read ---->   \  rrrrrrrrrrrrrrrrr   |       |
+	  barrier causes all effects      \     +-------+       |       |
+	  prior to the storage of B        ---->| A->1  |------>|       |
+	  to be perceptible to CPU 2            +-------+       |       |
+	                                        :       :       +-------+
+
+
+To illustrate this more completely, consider what could happen if the code
+contained a load of A either side of the read barrier:
+
+	CPU 1			CPU 2
+	=======================	=======================
+		{ A = 0, B = 9 }
+	STORE A=1
+	<write barrier>
+	STORE B=2
+				LOAD B
+				LOAD A [first load of A]
+				<read barrier>
+				LOAD A [second load of A]
+
+Even though the two loads of A both occur after the load of B, they may both
+come up with different values:
+
+	+-------+       :      :                :       :
+	|       |       +------+                +-------+
+	|       |------>| A=1  |------      --->| A->0  |
+	|       |       +------+      \         +-------+
+	| CPU 1 |   wwwwwwwwwwwwwwww   \    --->| B->9  |
+	|       |       +------+        |       +-------+
+	|       |------>| B=2  |---     |       :       :
+	|       |       +------+   \    |       :       :       +-------+
+	+-------+       :      :    \   |       +-------+       |       |
+	                             ---------->| B->2  |------>|       |
+	                                |       +-------+       | CPU 2 |
+	                                |       :       :       |       |
+	                                |       :       :       |       |
+	                                |       +-------+       |       |
+	                                |       | A->0  |------>| 1st   |
+	                                |       +-------+       |       |
+	  At this point the read ---->   \  rrrrrrrrrrrrrrrrr   |       |
+	  barrier causes all effects      \     +-------+       |       |
+	  prior to the storage of B        ---->| A->1  |------>| 2nd   |
+	  to be perceptible to CPU 2            +-------+       |       |
+	                                        :       :       +-------+
+
+
+But it may be that the update to A from CPU 1 becomes perceptible to CPU 2
+before the read barrier completes anyway:
+
+	+-------+       :      :                :       :
+	|       |       +------+                +-------+
+	|       |------>| A=1  |------      --->| A->0  |
+	|       |       +------+      \         +-------+
+	| CPU 1 |   wwwwwwwwwwwwwwww   \    --->| B->9  |
+	|       |       +------+        |       +-------+
+	|       |------>| B=2  |---     |       :       :
+	|       |       +------+   \    |       :       :       +-------+
+	+-------+       :      :    \   |       +-------+       |       |
+	                             ---------->| B->2  |------>|       |
+	                                |       +-------+       | CPU 2 |
+	                                |       :       :       |       |
+	                                 \      :       :       |       |
+	                                  \     +-------+       |       |
+	                                   ---->| A->1  |------>| 1st   |
+	                                        +-------+       |       |
+	                                    rrrrrrrrrrrrrrrrr   |       |
+	                                        +-------+       |       |
+	                                        | A->1  |------>| 2nd   |
+	                                        +-------+       |       |
+	                                        :       :       +-------+
+
+
+The guarantee is that the second load will always come up with A == 1 if the
+load of B came up with B == 2.  No such guarantee exists for the first load of
+A; that may come up with either A == 0 or A == 1.
+
+
+READ MEMORY BARRIERS VS LOAD SPECULATION
+----------------------------------------
+
+Many CPUs speculate with loads: that is they see that they will need to load an
+item from memory, and they find a time where they're not using the bus for any
+other loads, and so do the load in advance - even though they haven't actually
+got to that point in the instruction execution flow yet.  This permits the
+actual load instruction to potentially complete immediately because the CPU
+already has the value to hand.
+
+It may turn out that the CPU didn't actually need the value - perhaps because a
+branch circumvented the load - in which case it can discard the value or just
+cache it for later use.
+
+Consider:
+
+	CPU 1	   		CPU 2
+	=======================	=======================
+	 	   		LOAD B
+	 	   		DIVIDE		} Divide instructions generally
+	 	   		DIVIDE		} take a long time to perform
+	 	   		LOAD A
+
+Which might appear as this:
+
+	                                        :       :       +-------+
+	                                        +-------+       |       |
+	                                    --->| B->2  |------>|       |
+	                                        +-------+       | CPU 2 |
+	                                        :       :DIVIDE |       |
+	                                        +-------+       |       |
+	The CPU being busy doing a --->     --->| A->0  |~~~~   |       |
+	division speculates on the              +-------+   ~   |       |
+	LOAD of A                               :       :   ~   |       |
+	                                        :       :DIVIDE |       |
+	                                        :       :   ~   |       |
+	Once the divisions are complete -->     :       :   ~-->|       |
+	the CPU can then perform the            :       :       |       |
+	LOAD with immediate effect              :       :       +-------+
+
+
+Placing a read barrier or a data dependency barrier just before the second
+load:
+
+	CPU 1	   		CPU 2
+	=======================	=======================
+	 	   		LOAD B
+	 	   		DIVIDE
+	 	   		DIVIDE
+				<read barrier>
+	 	   		LOAD A
+
+will force any value speculatively obtained to be reconsidered to an extent
+dependent on the type of barrier used.  If there was no change made to the
+speculated memory location, then the speculated value will just be used:
+
+	                                        :       :       +-------+
+	                                        +-------+       |       |
+	                                    --->| B->2  |------>|       |
+	                                        +-------+       | CPU 2 |
+	                                        :       :DIVIDE |       |
+	                                        +-------+       |       |
+	The CPU being busy doing a --->     --->| A->0  |~~~~   |       |
+	division speculates on the              +-------+   ~   |       |
+	LOAD of A                               :       :   ~   |       |
+	                                        :       :DIVIDE |       |
+	                                        :       :   ~   |       |
+	                                        :       :   ~   |       |
+	                                    rrrrrrrrrrrrrrrr~   |       |
+	                                        :       :   ~   |       |
+	                                        :       :   ~-->|       |
+	                                        :       :       |       |
+	                                        :       :       +-------+
+
+
+but if there was an update or an invalidation from another CPU pending, then
+the speculation will be cancelled and the value reloaded:
+
+	                                        :       :       +-------+
+	                                        +-------+       |       |
+	                                    --->| B->2  |------>|       |
+	                                        +-------+       | CPU 2 |
+	                                        :       :DIVIDE |       |
+	                                        +-------+       |       |
+	The CPU being busy doing a --->     --->| A->0  |~~~~   |       |
+	division speculates on the              +-------+   ~   |       |
+	LOAD of A                               :       :   ~   |       |
+	                                        :       :DIVIDE |       |
+	                                        :       :   ~   |       |
+	                                        :       :   ~   |       |
+	                                    rrrrrrrrrrrrrrrrr   |       |
+	                                        +-------+       |       |
+	The speculation is discarded --->   --->| A->1  |------>|       |
+	and an updated value is                 +-------+       |       |
+	retrieved                               :       :       +-------+
 
 
 ========================
@@ -901,7 +1081,7 @@ IMPLICIT KERNEL MEMORY BARRIERS
 ===============================
 
 Some of the other functions in the linux kernel imply memory barriers, amongst
-which are locking, scheduling and memory allocation functions.
+which are locking and scheduling functions.
 
 This specification is a _minimum_ guarantee; any particular architecture may
 provide more substantial guarantees, but these may not be relied upon outside
@@ -966,6 +1146,20 @@ equivalent to a full barrier, but a LOCK followed by an UNLOCK is not.
     barriers is that the effects instructions outside of a critical section may
     seep into the inside of the critical section.
 
+A LOCK followed by an UNLOCK may not be assumed to be full memory barrier
+because it is possible for an access preceding the LOCK to happen after the
+LOCK, and an access following the UNLOCK to happen before the UNLOCK, and the
+two accesses can themselves then cross:
+
+	*A = a;
+	LOCK
+	UNLOCK
+	*B = b;
+
+may occur as:
+
+	LOCK, STORE *B, STORE *A, UNLOCK
+
 Locks and semaphores may not provide any guarantee of ordering on UP compiled
 systems, and so cannot be counted on in such a situation to actually achieve
 anything at all - especially with respect to I/O accesses - unless combined
@@ -1016,8 +1210,6 @@ Other functions that imply barriers:
 
  (*) schedule() and similar imply full memory barriers.
 
- (*) Memory allocation and release functions imply full memory barriers.
-
 
 =================================
 INTER-CPU LOCKING BARRIER EFFECTS

arch/alpha/Kconfig

@@ -453,7 +453,7 @@ config ALPHA_IRONGATE
 
 config GENERIC_HWEIGHT
 	bool
-	default y if !ALPHA_EV6 && !ALPHA_EV67
+	default y if !ALPHA_EV67
 
 config ALPHA_AVANTI
 	bool

arch/arm/mach-ep93xx/ts72xx.c

@@ -111,21 +111,21 @@ static void __init ts72xx_map_io(void)
 	}
 }
 
-static unsigned char ts72xx_rtc_readb(unsigned long addr)
+static unsigned char ts72xx_rtc_readbyte(unsigned long addr)
 {
 	__raw_writeb(addr, TS72XX_RTC_INDEX_VIRT_BASE);
 	return __raw_readb(TS72XX_RTC_DATA_VIRT_BASE);
 }
 
-static void ts72xx_rtc_writeb(unsigned char value, unsigned long addr)
+static void ts72xx_rtc_writebyte(unsigned char value, unsigned long addr)
 {
 	__raw_writeb(addr, TS72XX_RTC_INDEX_VIRT_BASE);
 	__raw_writeb(value, TS72XX_RTC_DATA_VIRT_BASE);
 }
 
 static struct m48t86_ops ts72xx_rtc_ops = {
-	.readb			= ts72xx_rtc_readb,
-	.writeb			= ts72xx_rtc_writeb,
+	.readbyte		= ts72xx_rtc_readbyte,
+	.writebyte		= ts72xx_rtc_writebyte,
 };
 
 static struct platform_device ts72xx_rtc_device = {

arch/arm/mach-imx/irq.c

@@ -127,7 +127,7 @@ static void
 imx_gpio_ack_irq(unsigned int irq)
 {
 	DEBUG_IRQ("%s: irq %d\n", __FUNCTION__, irq);
-	ISR(IRQ_TO_REG(irq)) |= 1 << ((irq - IRQ_GPIOA(0)) % 32);
+	ISR(IRQ_TO_REG(irq)) = 1 << ((irq - IRQ_GPIOA(0)) % 32);
 }
 
 static void

arch/arm/mach-integrator/integrator_cp.c

@@ -232,8 +232,6 @@ static void __init intcp_init_irq(void)
 	for (i = IRQ_PIC_START; i <= IRQ_PIC_END; i++) {
 		if (i == 11)
 			i = 22;
-		if (i == IRQ_CP_CPPLDINT)
-			i++;
 		if (i == 29)
 			break;
 		set_irq_chip(i, &pic_chip);
@@ -259,8 +257,7 @@ static void __init intcp_init_irq(void)
 		set_irq_flags(i, IRQF_VALID | IRQF_PROBE);
 	}
 
-	set_irq_handler(IRQ_CP_CPPLDINT, sic_handle_irq);
-	pic_unmask_irq(IRQ_CP_CPPLDINT);
+	set_irq_chained_handler(IRQ_CP_CPPLDINT, sic_handle_irq);
 }
 
 /*

arch/arm/mach-pxa/spitz.c

@@ -371,6 +371,7 @@ static int spitz_ohci_init(struct device *dev)
 static struct pxaohci_platform_data spitz_ohci_platform_data = {
 	.port_mode	= PMM_NPS_MODE,
 	.init		= spitz_ohci_init,
+	.power_budget	= 150,
 };
 
 

arch/arm/mach-sa1100/neponset.c

@@ -59,6 +59,14 @@ neponset_irq_handler(unsigned int irq, struct irqdesc *desc, struct pt_regs *reg
 		if (irr & (IRR_ETHERNET | IRR_USAR)) {
 			desc->chip->mask(irq);
 
+			/*
+			 * Ack the interrupt now to prevent re-entering
+			 * this neponset handler.  Again, this is safe
+			 * since we'll check the IRR register prior to
+			 * leaving.
+			 */
+			desc->chip->ack(irq);
+
 			if (irr & IRR_ETHERNET) {
 				d = irq_desc + IRQ_NEPONSET_SMC9196;
 				desc_handle_irq(IRQ_NEPONSET_SMC9196, d, regs);

arch/arm/mach-versatile/core.c

@@ -112,10 +112,9 @@ void __init versatile_init_irq(void)
 {
 	unsigned int i;
 
-	vic_init(VA_VIC_BASE, IRQ_VIC_START, ~(1 << 31));
+	vic_init(VA_VIC_BASE, IRQ_VIC_START, ~0);
 
-	set_irq_handler(IRQ_VICSOURCE31, sic_handle_irq);
-	enable_irq(IRQ_VICSOURCE31);
+	set_irq_chained_handler(IRQ_VICSOURCE31, sic_handle_irq);
 
 	/* Do second interrupt controller */
 	writel(~0, VA_SIC_BASE + SIC_IRQ_ENABLE_CLEAR);

arch/i386/kernel/acpi/earlyquirk.c

@@ -5,17 +5,34 @@
 #include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/pci.h>
+#include <linux/acpi.h>
+
 #include <asm/pci-direct.h>
 #include <asm/acpi.h>
 #include <asm/apic.h>
 
+#ifdef CONFIG_ACPI
+
+static int nvidia_hpet_detected __initdata;
+
+static int __init nvidia_hpet_check(unsigned long phys, unsigned long size)
+{
+	nvidia_hpet_detected = 1;
+	return 0;
+}
+#endif
+
 static int __init check_bridge(int vendor, int device)
 {
 #ifdef CONFIG_ACPI
-	/* According to Nvidia all timer overrides are bogus. Just ignore
-	   them all. */
+	/* According to Nvidia all timer overrides are bogus unless HPET
+	   is enabled. */
 	if (vendor == PCI_VENDOR_ID_NVIDIA) {
-		acpi_skip_timer_override = 1;
+		nvidia_hpet_detected = 0;
+		acpi_table_parse(ACPI_HPET, nvidia_hpet_check);
+		if (nvidia_hpet_detected == 0) {
+			acpi_skip_timer_override = 1;
+		}
 	}
 #endif
 	if (vendor == PCI_VENDOR_ID_ATI && timer_over_8254 == 1) {

arch/i386/kernel/setup.c

@@ -1547,15 +1547,18 @@ void __init setup_arch(char **cmdline_p)
 	if (efi_enabled)
 		efi_map_memmap();
 
-#ifdef CONFIG_X86_IO_APIC
-	check_acpi_pci();	/* Checks more than just ACPI actually */
-#endif
-
 #ifdef CONFIG_ACPI
 	/*
 	 * Parse the ACPI tables for possible boot-time SMP configuration.
 	 */
 	acpi_boot_table_init();
+#endif
+
+#ifdef CONFIG_X86_IO_APIC
+	check_acpi_pci();	/* Checks more than just ACPI actually */
+#endif
+
+#ifdef CONFIG_ACPI
 	acpi_boot_init();
 
 #if defined(CONFIG_SMP) && defined(CONFIG_X86_PC)

arch/powerpc/kernel/prom_init.c

@@ -822,6 +822,7 @@ static void __init prom_send_capabilities(void)
 		/* try calling the ibm,client-architecture-support method */
 		if (call_prom_ret("call-method", 3, 2, &ret,
 				  ADDR("ibm,client-architecture-support"),
+				  root,
 				  ADDR(ibm_architecture_vec)) == 0) {
 			/* the call exists... */
 			if (ret)
@@ -1622,6 +1623,15 @@ static int __init prom_find_machine_type(void)
 			if (strstr(p, RELOC("Power Macintosh")) ||
 			    strstr(p, RELOC("MacRISC")))
 				return PLATFORM_POWERMAC;
+#ifdef CONFIG_PPC64
+			/* We must make sure we don't detect the IBM Cell
+			 * blades as pSeries due to some firmware issues,
+			 * so we do it here.
+			 */
+			if (strstr(p, RELOC("IBM,CBEA")) ||
+			    strstr(p, RELOC("IBM,CPBW-1.0")))
+				return PLATFORM_GENERIC;
+#endif /* CONFIG_PPC64 */
 			i += sl + 1;
 		}
 	}

arch/powerpc/kernel/signal_32.c

@@ -803,10 +803,13 @@ static int do_setcontext(struct ucontext __user *ucp, struct pt_regs *regs, int
 		if (__get_user(cmcp, &ucp->uc_regs))
 			return -EFAULT;
 		mcp = (struct mcontext __user *)(u64)cmcp;
+		/* no need to check access_ok(mcp), since mcp < 4GB */
 	}
 #else
 	if (__get_user(mcp, &ucp->uc_regs))
 		return -EFAULT;
+	if (!access_ok(VERIFY_READ, mcp, sizeof(*mcp)))
+		return -EFAULT;
 #endif
 	restore_sigmask(&set);
 	if (restore_user_regs(regs, mcp, sig))
@@ -908,13 +911,14 @@ int sys_debug_setcontext(struct ucontext __user *ctx,
 {
 	struct sig_dbg_op op;
 	int i;
+	unsigned char tmp;
 	unsigned long new_msr = regs->msr;
 #if defined(CONFIG_4xx) || defined(CONFIG_BOOKE)
 	unsigned long new_dbcr0 = current->thread.dbcr0;
 #endif
 
 	for (i=0; i<ndbg; i++) {
-		if (__copy_from_user(&op, dbg, sizeof(op)))
+		if (copy_from_user(&op, dbg + i, sizeof(op)))
 			return -EFAULT;
 		switch (op.dbg_type) {
 		case SIG_DBG_SINGLE_STEPPING:
@@ -959,6 +963,11 @@ int sys_debug_setcontext(struct ucontext __user *ctx,
 	current->thread.dbcr0 = new_dbcr0;
 #endif
 
+	if (!access_ok(VERIFY_READ, ctx, sizeof(*ctx))
+	    || __get_user(tmp, (u8 __user *) ctx)
+	    || __get_user(tmp, (u8 __user *) (ctx + 1) - 1))
+		return -EFAULT;
+
 	/*
 	 * If we get a fault copying the context into the kernel's
 	 * image of the user's registers, we can't just return -EFAULT

arch/powerpc/kernel/signal_64.c

@@ -182,6 +182,8 @@ static long restore_sigcontext(struct pt_regs *regs, sigset_t *set, int sig,
 	err |= __get_user(msr, &sc->gp_regs[PT_MSR]);
 	if (err)
 		return err;
+	if (v_regs && !access_ok(VERIFY_READ, v_regs, 34 * sizeof(vector128)))
+		return -EFAULT;
 	/* Copy 33 vec registers (vr0..31 and vscr) from the stack */
 	if (v_regs != 0 && (msr & MSR_VEC) != 0)
 		err |= __copy_from_user(current->thread.vr, v_regs,

arch/powerpc/platforms/cell/setup.c

@@ -125,14 +125,13 @@ static void __init cell_init_early(void)
 
 static int __init cell_probe(void)
 {
-	/* XXX This is temporary, the Cell maintainer will come up with
-	 * more appropriate detection logic
-	 */
 	unsigned long root = of_get_flat_dt_root();
-	if (!of_flat_dt_is_compatible(root, "IBM,CPBW-1.0"))
-		return 0;
 
-	return 1;
+	if (of_flat_dt_is_compatible(root, "IBM,CBEA") ||
+	    of_flat_dt_is_compatible(root, "IBM,CPBW-1.0"))
+		return 1;
+
+	return 0;
 }
 
 /*

arch/powerpc/platforms/pseries/setup.c

@@ -389,6 +389,7 @@ static int __init pSeries_probe_hypertas(unsigned long node,
 
 static int __init pSeries_probe(void)
 {
+	unsigned long root = of_get_flat_dt_root();
  	char *dtype = of_get_flat_dt_prop(of_get_flat_dt_root(),
  					  "device_type", NULL);
  	if (dtype == NULL)
@@ -396,6 +397,13 @@ static int __init pSeries_probe(void)
  	if (strcmp(dtype, "chrp"))
 		return 0;
 
+	/* Cell blades firmware claims to be chrp while it's not. Until this
+	 * is fixed, we need to avoid those here.
+	 */
+	if (of_flat_dt_is_compatible(root, "IBM,CPBW-1.0") ||
+	    of_flat_dt_is_compatible(root, "IBM,CBEA"))
+		return 0;
+
 	DBG("pSeries detected, looking for LPAR capability...\n");
 
 	/* Now try to figure out if we are running on LPAR */

arch/sparc/kernel/smp.c

@@ -69,6 +69,17 @@ void __init smp_store_cpu_info(int id)
 						     "clock-frequency", 0);
 	cpu_data(id).prom_node = cpu_node;
 	cpu_data(id).mid = cpu_get_hwmid(cpu_node);
+
+	/* this is required to tune the scheduler correctly */
+	/* is it possible to have CPUs with different cache sizes? */
+	if (id == boot_cpu_id) {
+		int cache_line,cache_nlines;
+		cache_line = 0x20;
+		cache_line = prom_getintdefault(cpu_node, "ecache-line-size", cache_line);
+		cache_nlines = 0x8000;
+		cache_nlines = prom_getintdefault(cpu_node, "ecache-nlines", cache_nlines);
+		max_cache_size = cache_line * cache_nlines;
+	}
 	if (cpu_data(id).mid < 0)
 		panic("No MID found for CPU%d at node 0x%08d", id, cpu_node);
 }

arch/sparc64/kernel/pci_sun4v.c

@@ -599,18 +599,128 @@ struct pci_iommu_ops pci_sun4v_iommu_ops = {
 
 /* SUN4V PCI configuration space accessors. */
 
-static inline int pci_sun4v_out_of_range(struct pci_pbm_info *pbm, unsigned int bus, unsigned int device, unsigned int func)
+struct pdev_entry {
+	struct pdev_entry	*next;
+	u32			devhandle;
+	unsigned int		bus;
+	unsigned int		device;
+	unsigned int		func;
+};
+
+#define PDEV_HTAB_SIZE	16
+#define PDEV_HTAB_MASK	(PDEV_HTAB_SIZE - 1)
+static struct pdev_entry *pdev_htab[PDEV_HTAB_SIZE];
+
+static inline unsigned int pdev_hashfn(u32 devhandle, unsigned int bus, unsigned int device, unsigned int func)
 {
-	if (bus == pbm->pci_first_busno) {
-		if (device == 0 && func == 0)
-			return 0;
-		return 1;
+	unsigned int val;
+
+	val = (devhandle ^ (devhandle >> 4));
+	val ^= bus;
+	val ^= device;
+	val ^= func;
+
+	return val & PDEV_HTAB_MASK;
+}
+
+static int pdev_htab_add(u32 devhandle, unsigned int bus, unsigned int device, unsigned int func)
+{
+	struct pdev_entry *p = kmalloc(sizeof(*p), GFP_KERNEL);
+	struct pdev_entry **slot;
+
+	if (!p)
+		return -ENOMEM;
+
+	slot = &pdev_htab[pdev_hashfn(devhandle, bus, device, func)];
+	p->next = *slot;
+	*slot = p;
+
+	p->devhandle = devhandle;
+	p->bus = bus;
+	p->device = device;
+	p->func = func;
+
+	return 0;
+}
+
+/* Recursively descend into the OBP device tree, rooted at toplevel_node,
+ * looking for a PCI device matching bus and devfn.
+ */
+static int obp_find(struct linux_prom_pci_registers *pregs, int toplevel_node, unsigned int bus, unsigned int devfn)
+{
+	toplevel_node = prom_getchild(toplevel_node);
+
+	while (toplevel_node != 0) {
+		int ret = obp_find(pregs, toplevel_node, bus, devfn);
+
+		if (ret != 0)
+			return ret;
+
+		ret = prom_getproperty(toplevel_node, "reg", (char *) pregs,
+				       sizeof(*pregs) * PROMREG_MAX);
+		if (ret == 0 || ret == -1)
+			goto next_sibling;
+
+		if (((pregs[0].phys_hi >> 16) & 0xff) == bus &&
+		    ((pregs[0].phys_hi >> 8) & 0xff) == devfn)
+			break;
+
+	next_sibling:
+		toplevel_node = prom_getsibling(toplevel_node);
+	}
+
+	return toplevel_node;
+}
+
+static int pdev_htab_populate(struct pci_pbm_info *pbm)
+{
+	struct linux_prom_pci_registers pr[PROMREG_MAX];
+	u32 devhandle = pbm->devhandle;
+	unsigned int bus;
+
+	for (bus = pbm->pci_first_busno; bus <= pbm->pci_last_busno; bus++) {
+		unsigned int devfn;
+
+		for (devfn = 0; devfn < 256; devfn++) {
+			unsigned int device = PCI_SLOT(devfn);
+			unsigned int func = PCI_FUNC(devfn);
+
+			if (obp_find(pr, pbm->prom_node, bus, devfn)) {
+				int err = pdev_htab_add(devhandle, bus,
+							device, func);
+				if (err)
+					return err;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static struct pdev_entry *pdev_find(u32 devhandle, unsigned int bus, unsigned int device, unsigned int func)
+{
+	struct pdev_entry *p;
+
+	p = pdev_htab[pdev_hashfn(devhandle, bus, device, func)];
+	while (p) {
+		if (p->devhandle == devhandle &&
+		    p->bus == bus &&
+		    p->device == device &&
+		    p->func == func)
+			break;
+
+		p = p->next;
 	}
 
+	return p;
+}
+
+static inline int pci_sun4v_out_of_range(struct pci_pbm_info *pbm, unsigned int bus, unsigned int device, unsigned int func)
+{
 	if (bus < pbm->pci_first_busno ||
 	    bus > pbm->pci_last_busno)
 		return 1;
-	return 0;
+	return pdev_find(pbm->devhandle, bus, device, func) == NULL;
 }
 
 static int pci_sun4v_read_pci_cfg(struct pci_bus *bus_dev, unsigned int devfn,
@@ -1063,6 +1173,8 @@ static void pci_sun4v_pbm_init(struct pci_controller_info *p, int prom_node, u32
 
 	pci_sun4v_get_bus_range(pbm);
 	pci_sun4v_iommu_init(pbm);
+
+	pdev_htab_populate(pbm);
 }
 
 void sun4v_pci_init(int node, char *model_name)

arch/sparc64/kernel/smp.c

@@ -1287,6 +1287,40 @@ int setup_profiling_timer(unsigned int multiplier)
 	return 0;
 }
 
+static void __init smp_tune_scheduling(void)
+{
+	int instance, node;
+	unsigned int def, smallest = ~0U;
+
+	def = ((tlb_type == hypervisor) ?
+	       (3 * 1024 * 1024) :
+	       (4 * 1024 * 1024));
+
+	instance = 0;
+	while (!cpu_find_by_instance(instance, &node, NULL)) {
+		unsigned int val;
+
+		val = prom_getintdefault(node, "ecache-size", def);
+		if (val < smallest)
+			smallest = val;
+
+		instance++;
+	}
+
+	/* Any value less than 256K is nonsense.  */
+	if (smallest < (256U * 1024U))
+		smallest = 256 * 1024;
+
+	max_cache_size = smallest;
+
+	if (smallest < 1U * 1024U * 1024U)
+		printk(KERN_INFO "Using max_cache_size of %uKB\n",
+		       smallest / 1024U);
+	else
+		printk(KERN_INFO "Using max_cache_size of %uMB\n",
+		       smallest / 1024U / 1024U);
+}
+
 /* Constrain the number of cpus to max_cpus.  */
 void __init smp_prepare_cpus(unsigned int max_cpus)
 {
@@ -1322,6 +1356,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 	}
 
 	smp_store_cpu_info(boot_cpu_id);
+	smp_tune_scheduling();
 }
 
 /* Set this up early so that things like the scheduler can init

arch/sparc64/kernel/sparc64_ksyms.c

@@ -297,7 +297,6 @@ EXPORT_SYMBOL(svr4_getcontext);
 EXPORT_SYMBOL(svr4_setcontext);
 EXPORT_SYMBOL(compat_sys_ioctl);
 EXPORT_SYMBOL(sparc32_open);
-EXPORT_SYMBOL(sys_close);
 #endif
 
 /* Special internal versions of library functions. */

arch/sparc64/kernel/traps.c

@@ -1797,7 +1797,9 @@ static const char *sun4v_err_type_to_str(u32 type)
 	};
 }
 
-static void sun4v_log_error(struct sun4v_error_entry *ent, int cpu, const char *pfx, atomic_t *ocnt)
+extern void __show_regs(struct pt_regs * regs);
+
+static void sun4v_log_error(struct pt_regs *regs, struct sun4v_error_entry *ent, int cpu, const char *pfx, atomic_t *ocnt)
 {
 	int cnt;
 
@@ -1830,6 +1832,8 @@ static void sun4v_log_error(struct sun4v_error_entry *ent, int cpu, const char *
 	       pfx,
 	       ent->err_raddr, ent->err_size, ent->err_cpu);
 
+	__show_regs(regs);
+
 	if ((cnt = atomic_read(ocnt)) != 0) {
 		atomic_set(ocnt, 0);
 		wmb();
@@ -1862,7 +1866,7 @@ void sun4v_resum_error(struct pt_regs *regs, unsigned long offset)
 
 	put_cpu();
 
-	sun4v_log_error(&local_copy, cpu,
+	sun4v_log_error(regs, &local_copy, cpu,
 			KERN_ERR "RESUMABLE ERROR",
 			&sun4v_resum_oflow_cnt);
 }
@@ -1910,7 +1914,7 @@ void sun4v_nonresum_error(struct pt_regs *regs, unsigned long offset)
 	}
 #endif
 
-	sun4v_log_error(&local_copy, cpu,
+	sun4v_log_error(regs, &local_copy, cpu,
 			KERN_EMERG "NON-RESUMABLE ERROR",
 			&sun4v_nonresum_oflow_cnt);
 
@@ -2200,7 +2204,6 @@ static inline struct reg_window *kernel_stack_up(struct reg_window *rw)
 void die_if_kernel(char *str, struct pt_regs *regs)
 {
 	static int die_counter;
-	extern void __show_regs(struct pt_regs * regs);
 	extern void smp_report_regs(void);
 	int count = 0;
 	

arch/x86_64/kernel/io_apic.c

@@ -271,6 +271,18 @@ __setup("enable_8254_timer", setup_enable_8254_timer);
 #include <linux/pci_ids.h>
 #include <linux/pci.h>
 
+
+#ifdef CONFIG_ACPI
+
+static int nvidia_hpet_detected __initdata;
+
+static int __init nvidia_hpet_check(unsigned long phys, unsigned long size)
+{
+	nvidia_hpet_detected = 1;
+	return 0;
+}
+#endif
+
 /* Temporary Hack. Nvidia and VIA boards currently only work with IO-APIC
    off. Check for an Nvidia or VIA PCI bridge and turn it off.
    Use pci direct infrastructure because this runs before the PCI subsystem. 
@@ -317,11 +329,19 @@ void __init check_ioapic(void)
 					return;
 				case PCI_VENDOR_ID_NVIDIA:
 #ifdef CONFIG_ACPI
-					/* All timer overrides on Nvidia
-				           seem to be wrong. Skip them. */
-					acpi_skip_timer_override = 1;
-					printk(KERN_INFO 
-	     "Nvidia board detected. Ignoring ACPI timer override.\n");
+					/*
+					 * All timer overrides on Nvidia are
+					 * wrong unless HPET is enabled.
+					 */
+					nvidia_hpet_detected = 0;
+					acpi_table_parse(ACPI_HPET,
+							nvidia_hpet_check);
+					if (nvidia_hpet_detected == 0) {
+						acpi_skip_timer_override = 1;
+						printk(KERN_INFO "Nvidia board "
+						    "detected. Ignoring ACPI "
+						    "timer override.\n");
+					}
 #endif
 					/* RED-PEN skip them on mptables too? */
 					return;

block/as-iosched.c

@@ -1648,17 +1648,17 @@ static void as_exit_queue(elevator_t *e)
  * initialize elevator private data (as_data), and alloc a arq for
  * each request on the free lists
  */
-static int as_init_queue(request_queue_t *q, elevator_t *e)
+static void *as_init_queue(request_queue_t *q, elevator_t *e)
 {
 	struct as_data *ad;
 	int i;
 
 	if (!arq_pool)
-		return -ENOMEM;
+		return NULL;
 
 	ad = kmalloc_node(sizeof(*ad), GFP_KERNEL, q->node);
 	if (!ad)
-		return -ENOMEM;
+		return NULL;
 	memset(ad, 0, sizeof(*ad));
 
 	ad->q = q; /* Identify what queue the data belongs to */
@@ -1667,7 +1667,7 @@ static int as_init_queue(request_queue_t *q, elevator_t *e)
 				GFP_KERNEL, q->node);
 	if (!ad->hash) {
 		kfree(ad);
-		return -ENOMEM;
+		return NULL;
 	}
 
 	ad->arq_pool = mempool_create_node(BLKDEV_MIN_RQ, mempool_alloc_slab,
@@ -1675,7 +1675,7 @@ static int as_init_queue(request_queue_t *q, elevator_t *e)
 	if (!ad->arq_pool) {
 		kfree(ad->hash);
 		kfree(ad);
-		return -ENOMEM;
+		return NULL;
 	}
 
 	/* anticipatory scheduling helpers */
@@ -1696,14 +1696,13 @@ static int as_init_queue(request_queue_t *q, elevator_t *e)
 	ad->antic_expire = default_antic_expire;
 	ad->batch_expire[REQ_SYNC] = default_read_batch_expire;
 	ad->batch_expire[REQ_ASYNC] = default_write_batch_expire;
-	e->elevator_data = ad;
 
 	ad->current_batch_expires = jiffies + ad->batch_expire[REQ_SYNC];
 	ad->write_batch_count = ad->batch_expire[REQ_ASYNC] / 10;
 	if (ad->write_batch_count < 2)
 		ad->write_batch_count = 2;
 
-	return 0;
+	return ad;
 }
 
 /*

block/cfq-iosched.c

@@ -2251,14 +2251,14 @@ static void cfq_exit_queue(elevator_t *e)
 	kfree(cfqd);
 }
 
-static int cfq_init_queue(request_queue_t *q, elevator_t *e)
+static void *cfq_init_queue(request_queue_t *q, elevator_t *e)
 {
 	struct cfq_data *cfqd;
 	int i;
 
 	cfqd = kmalloc(sizeof(*cfqd), GFP_KERNEL);
 	if (!cfqd)
-		return -ENOMEM;
+		return NULL;
 
 	memset(cfqd, 0, sizeof(*cfqd));
 
@@ -2288,8 +2288,6 @@ static int cfq_init_queue(request_queue_t *q, elevator_t *e)
 	for (i = 0; i < CFQ_QHASH_ENTRIES; i++)
 		INIT_HLIST_HEAD(&cfqd->cfq_hash[i]);
 
-	e->elevator_data = cfqd;
-
 	cfqd->queue = q;
 
 	cfqd->max_queued = q->nr_requests / 4;
@@ -2316,14 +2314,14 @@ static int cfq_init_queue(request_queue_t *q, elevator_t *e)
 	cfqd->cfq_slice_async_rq = cfq_slice_async_rq;
 	cfqd->cfq_slice_idle = cfq_slice_idle;
 
-	return 0;
+	return cfqd;
 out_crqpool:
 	kfree(cfqd->cfq_hash);
 out_cfqhash:
 	kfree(cfqd->crq_hash);
 out_crqhash:
 	kfree(cfqd);
-	return -ENOMEM;
+	return NULL;
 }
 
 static void cfq_slab_kill(void)

block/deadline-iosched.c

@@ -613,24 +613,24 @@ static void deadline_exit_queue(elevator_t *e)
  * initialize elevator private data (deadline_data), and alloc a drq for
  * each request on the free lists
  */
-static int deadline_init_queue(request_queue_t *q, elevator_t *e)
+static void *deadline_init_queue(request_queue_t *q, elevator_t *e)
 {
 	struct deadline_data *dd;
 	int i;
 
 	if (!drq_pool)
-		return -ENOMEM;
+		return NULL;
 
 	dd = kmalloc_node(sizeof(*dd), GFP_KERNEL, q->node);
 	if (!dd)
-		return -ENOMEM;
+		return NULL;
 	memset(dd, 0, sizeof(*dd));
 
 	dd->hash = kmalloc_node(sizeof(struct list_head)*DL_HASH_ENTRIES,
 				GFP_KERNEL, q->node);
 	if (!dd->hash) {
 		kfree(dd);
-		return -ENOMEM;
+		return NULL;
 	}
 
 	dd->drq_pool = mempool_create_node(BLKDEV_MIN_RQ, mempool_alloc_slab,
@@ -638,7 +638,7 @@ static int deadline_init_queue(request_queue_t *q, elevator_t *e)
 	if (!dd->drq_pool) {
 		kfree(dd->hash);
 		kfree(dd);
-		return -ENOMEM;
+		return NULL;
 	}
 
 	for (i = 0; i < DL_HASH_ENTRIES; i++)
@@ -653,8 +653,7 @@ static int deadline_init_queue(request_queue_t *q, elevator_t *e)
 	dd->writes_starved = writes_starved;
 	dd->front_merges = 1;
 	dd->fifo_batch = fifo_batch;
-	e->elevator_data = dd;
-	return 0;
+	return dd;
 }
 
 static void deadline_put_request(request_queue_t *q, struct request *rq)

block/elevator.c

@@ -121,16 +121,16 @@ static struct elevator_type *elevator_get(const char *name)
 	return e;
 }
 
-static int elevator_attach(request_queue_t *q, struct elevator_queue *eq)
+static void *elevator_init_queue(request_queue_t *q, struct elevator_queue *eq)
 {
-	int ret = 0;
+	return eq->ops->elevator_init_fn(q, eq);
+}
 
+static void elevator_attach(request_queue_t *q, struct elevator_queue *eq,
+			   void *data)
+{
 	q->elevator = eq;
-
-	if (eq->ops->elevator_init_fn)
-		ret = eq->ops->elevator_init_fn(q, eq);
-
-	return ret;
+	eq->elevator_data = data;
 }
 
 static char chosen_elevator[16];
@@ -181,6 +181,7 @@ int elevator_init(request_queue_t *q, char *name)
 	struct elevator_type *e = NULL;
 	struct elevator_queue *eq;
 	int ret = 0;
+	void *data;
 
 	INIT_LIST_HEAD(&q->queue_head);
 	q->last_merge = NULL;
@@ -202,10 +203,13 @@ int elevator_init(request_queue_t *q, char *name)
 	if (!eq)
 		return -ENOMEM;
 
-	ret = elevator_attach(q, eq);
-	if (ret)
+	data = elevator_init_queue(q, eq);
+	if (!data) {
 		kobject_put(&eq->kobj);
+		return -ENOMEM;
+	}
 
+	elevator_attach(q, eq, data);
 	return ret;
 }
 
@@ -722,13 +726,16 @@ int elv_register_queue(struct request_queue *q)
 	return error;
 }
 
+static void __elv_unregister_queue(elevator_t *e)
+{
+	kobject_uevent(&e->kobj, KOBJ_REMOVE);
+	kobject_del(&e->kobj);
+}
+
 void elv_unregister_queue(struct request_queue *q)
 {
-	if (q) {
-		elevator_t *e = q->elevator;
-		kobject_uevent(&e->kobj, KOBJ_REMOVE);
-		kobject_del(&e->kobj);
-	}
+	if (q)
+		__elv_unregister_queue(q->elevator);
 }
 
 int elv_register(struct elevator_type *e)
@@ -780,6 +787,7 @@ EXPORT_SYMBOL_GPL(elv_unregister);
 static int elevator_switch(request_queue_t *q, struct elevator_type *new_e)
 {
 	elevator_t *old_elevator, *e;
+	void *data;
 
 	/*
 	 * Allocate new elevator
@@ -788,6 +796,12 @@ static int elevator_switch(request_queue_t *q, struct elevator_type *new_e)
 	if (!e)
 		return 0;
 
+	data = elevator_init_queue(q, e);
+	if (!data) {
+		kobject_put(&e->kobj);
+		return 0;
+	}
+
 	/*
 	 * Turn on BYPASS and drain all requests w/ elevator private data
 	 */
@@ -806,19 +820,19 @@ static int elevator_switch(request_queue_t *q, struct elevator_type *new_e)
 		elv_drain_elevator(q);
 	}
 
-	spin_unlock_irq(q->queue_lock);
-
 	/*
-	 * unregister old elevator data
+	 * Remember old elevator.
 	 */
-	elv_unregister_queue(q);
 	old_elevator = q->elevator;
 
 	/*
 	 * attach and start new elevator
 	 */
-	if (elevator_attach(q, e))
-		goto fail;
+	elevator_attach(q, e, data);
+
+	spin_unlock_irq(q->queue_lock);
+
+	__elv_unregister_queue(old_elevator);
 
 	if (elv_register_queue(q))
 		goto fail_register;
@@ -837,7 +851,6 @@ static int elevator_switch(request_queue_t *q, struct elevator_type *new_e)
 	 */
 	elevator_exit(e);
 	e = NULL;
-fail:
 	q->elevator = old_elevator;
 	elv_register_queue(q);
 	clear_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags);

block/noop-iosched.c

@@ -65,16 +65,15 @@ noop_latter_request(request_queue_t *q, struct request *rq)
 	return list_entry(rq->queuelist.next, struct request, queuelist);
 }
 
-static int noop_init_queue(request_queue_t *q, elevator_t *e)
+static void *noop_init_queue(request_queue_t *q, elevator_t *e)
 {
 	struct noop_data *nd;
 
 	nd = kmalloc(sizeof(*nd), GFP_KERNEL);
 	if (!nd)
-		return -ENOMEM;
+		return NULL;
 	INIT_LIST_HEAD(&nd->queue);
-	e->elevator_data = nd;
-	return 0;
+	return nd;
 }
 
 static void noop_exit_queue(elevator_t *e)

drivers/acpi/processor_perflib.c

@@ -577,6 +577,8 @@ acpi_processor_register_performance(struct acpi_processor_performance
 		return_VALUE(-EBUSY);
 	}
 
+	WARN_ON(!performance);
+
 	pr->performance = performance;
 
 	if (acpi_processor_get_performance_info(pr)) {
@@ -609,7 +611,8 @@ acpi_processor_unregister_performance(struct acpi_processor_performance
 		return_VOID;
 	}
 
-	kfree(pr->performance->states);
+	if (pr->performance)
+		kfree(pr->performance->states);
 	pr->performance = NULL;
 
 	acpi_cpufreq_remove_file(pr);

drivers/char/Makefile

@@ -41,9 +41,9 @@ obj-$(CONFIG_N_HDLC)		+= n_hdlc.o
 obj-$(CONFIG_AMIGA_BUILTIN_SERIAL) += amiserial.o
 obj-$(CONFIG_SX)		+= sx.o generic_serial.o
 obj-$(CONFIG_RIO)		+= rio/ generic_serial.o
-obj-$(CONFIG_HVC_DRIVER)	+= hvc_console.o
 obj-$(CONFIG_HVC_CONSOLE)	+= hvc_vio.o hvsi.o
 obj-$(CONFIG_HVC_RTAS)		+= hvc_rtas.o
+obj-$(CONFIG_HVC_DRIVER)	+= hvc_console.o
 obj-$(CONFIG_RAW_DRIVER)	+= raw.o
 obj-$(CONFIG_SGI_SNSC)		+= snsc.o snsc_event.o
 obj-$(CONFIG_MMTIMER)		+= mmtimer.o

drivers/char/n_tty.c

@@ -1384,8 +1384,10 @@ static ssize_t read_chan(struct tty_struct *tty, struct file *file,
 		 * longer than TTY_THRESHOLD_UNTHROTTLE in canonical mode,
 		 * we won't get any more characters.
 		 */
-		if (n_tty_chars_in_buffer(tty) <= TTY_THRESHOLD_UNTHROTTLE)
+		if (n_tty_chars_in_buffer(tty) <= TTY_THRESHOLD_UNTHROTTLE) {
+			n_tty_set_room(tty);
 			check_unthrottle(tty);
+		}
 
 		if (b - buf >= minimum)
 			break;

drivers/message/fusion/mptspi.c

@@ -831,6 +831,7 @@ mptspi_ioc_reset(MPT_ADAPTER *ioc, int reset_phase)
 	return rc;
 }
 
+#ifdef CONFIG_PM
 /*
  * spi module resume handler
  */
@@ -846,6 +847,7 @@ mptspi_resume(struct pci_dev *pdev)
 
 	return rc;
 }
+#endif
 
 /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
 /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/

drivers/message/i2o/exec-osm.c

@@ -55,6 +55,7 @@ struct i2o_exec_wait {
 	u32 m;			/* message id */
 	struct i2o_message *msg;	/* pointer to the reply message */
 	struct list_head list;	/* node in global wait list */
+	spinlock_t lock;	/* lock before modifying */
 };
 
 /* Work struct needed to handle LCT NOTIFY replies */
@@ -87,6 +88,7 @@ static struct i2o_exec_wait *i2o_exec_wait_alloc(void)
 		return NULL;
 
 	INIT_LIST_HEAD(&wait->list);
+	spin_lock_init(&wait->lock);
 
 	return wait;
 };
@@ -125,6 +127,7 @@ int i2o_msg_post_wait_mem(struct i2o_controller *c, struct i2o_message *msg,
 	DECLARE_WAIT_QUEUE_HEAD(wq);
 	struct i2o_exec_wait *wait;
 	static u32 tcntxt = 0x80000000;
+	long flags;
 	int rc = 0;
 
 	wait = i2o_exec_wait_alloc();
@@ -146,33 +149,28 @@ int i2o_msg_post_wait_mem(struct i2o_controller *c, struct i2o_message *msg,
 	wait->tcntxt = tcntxt++;
 	msg->u.s.tcntxt = cpu_to_le32(wait->tcntxt);
 
+	wait->wq = &wq;
+	/*
+	 * we add elements to the head, because if a entry in the list will
+	 * never be removed, we have to iterate over it every time
+	 */
+	list_add(&wait->list, &i2o_exec_wait_list);
+
 	/*
 	 * Post the message to the controller. At some point later it will
 	 * return. If we time out before it returns then complete will be zero.
 	 */
 	i2o_msg_post(c, msg);
 
-	if (!wait->complete) {
-		wait->wq = &wq;
-		/*
-		 * we add elements add the head, because if a entry in the list
-		 * will never be removed, we have to iterate over it every time
-		 */
-		list_add(&wait->list, &i2o_exec_wait_list);
-
-		wait_event_interruptible_timeout(wq, wait->complete,
-						 timeout * HZ);
+	wait_event_interruptible_timeout(wq, wait->complete, timeout * HZ);
 
-		wait->wq = NULL;
-	}
+	spin_lock_irqsave(&wait->lock, flags);
 
-	barrier();
+	wait->wq = NULL;
 
-	if (wait->complete) {
+	if (wait->complete)
 		rc = le32_to_cpu(wait->msg->body[0]) >> 24;
-		i2o_flush_reply(c, wait->m);
-		i2o_exec_wait_free(wait);
-	} else {
+	else {
 		/*
 		 * We cannot remove it now. This is important. When it does
 		 * terminate (which it must do if the controller has not
@@ -186,6 +184,13 @@ int i2o_msg_post_wait_mem(struct i2o_controller *c, struct i2o_message *msg,
 		rc = -ETIMEDOUT;
 	}
 
+	spin_unlock_irqrestore(&wait->lock, flags);
+
+	if (rc != -ETIMEDOUT) {
+		i2o_flush_reply(c, wait->m);
+		i2o_exec_wait_free(wait);
+	}
+
 	return rc;
 };
 
@@ -213,7 +218,6 @@ static int i2o_msg_post_wait_complete(struct i2o_controller *c, u32 m,
 {
 	struct i2o_exec_wait *wait, *tmp;
 	unsigned long flags;
-	static spinlock_t lock = SPIN_LOCK_UNLOCKED;
 	int rc = 1;
 
 	/*
@@ -223,23 +227,24 @@ static int i2o_msg_post_wait_complete(struct i2o_controller *c, u32 m,
 	 * already expired. Not much we can do about that except log it for
 	 * debug purposes, increase timeout, and recompile.
 	 */
-	spin_lock_irqsave(&lock, flags);
 	list_for_each_entry_safe(wait, tmp, &i2o_exec_wait_list, list) {
 		if (wait->tcntxt == context) {
-			list_del(&wait->list);
+			spin_lock_irqsave(&wait->lock, flags);
 
-			spin_unlock_irqrestore(&lock, flags);
+			list_del(&wait->list);
 
 			wait->m = m;
 			wait->msg = msg;
 			wait->complete = 1;
 
-			barrier();
-
-			if (wait->wq) {
-				wake_up_interruptible(wait->wq);
+			if (wait->wq)
 				rc = 0;
-			} else {
+			else
+				rc = -1;
+
+			spin_unlock_irqrestore(&wait->lock, flags);
+
+			if (rc) {
 				struct device *dev;
 
 				dev = &c->pdev->dev;
@@ -248,15 +253,13 @@ static int i2o_msg_post_wait_complete(struct i2o_controller *c, u32 m,
 					 c->name);
 				i2o_dma_free(dev, &wait->dma);
 				i2o_exec_wait_free(wait);
-				rc = -1;
-			}
+			} else
+				wake_up_interruptible(wait->wq);
 
 			return rc;
 		}
 	}
 
-	spin_unlock_irqrestore(&lock, flags);
-
 	osm_warn("%s: Bogus reply in POST WAIT (tr-context: %08x)!\n", c->name,
 		 context);
 
@@ -322,14 +325,9 @@ static DEVICE_ATTR(product_id, S_IRUGO, i2o_exec_show_product_id, NULL);
 static int i2o_exec_probe(struct device *dev)
 {
 	struct i2o_device *i2o_dev = to_i2o_device(dev);
-	struct i2o_controller *c = i2o_dev->iop;
 
 	i2o_event_register(i2o_dev, &i2o_exec_driver, 0, 0xffffffff);
 
-	c->exec = i2o_dev;
-
-	i2o_exec_lct_notify(c, c->lct->change_ind + 1);
-
 	device_create_file(dev, &dev_attr_vendor_id);
 	device_create_file(dev, &dev_attr_product_id);
 
@@ -523,6 +521,8 @@ static int i2o_exec_lct_notify(struct i2o_controller *c, u32 change_ind)
 	struct device *dev;
 	struct i2o_message *msg;
 
+	down(&c->lct_lock);
+
 	dev = &c->pdev->dev;
 
 	if (i2o_dma_realloc
@@ -545,6 +545,8 @@ static int i2o_exec_lct_notify(struct i2o_controller *c, u32 change_ind)
 
 	i2o_msg_post(c, msg);
 
+	up(&c->lct_lock);
+
 	return 0;
 };
 

drivers/message/i2o/iop.c

@@ -804,8 +804,6 @@ void i2o_iop_remove(struct i2o_controller *c)
 
 	/* Ask the IOP to switch to RESET state */
 	i2o_iop_reset(c);
-
-	put_device(&c->device);
 }
 
 /**
@@ -1059,7 +1057,7 @@ struct i2o_controller *i2o_iop_alloc(void)
 
 	snprintf(poolname, sizeof(poolname), "i2o_%s_msg_inpool", c->name);
 	if (i2o_pool_alloc
-	    (&c->in_msg, poolname, I2O_INBOUND_MSG_FRAME_SIZE * 4,
+	    (&c->in_msg, poolname, I2O_INBOUND_MSG_FRAME_SIZE * 4 + sizeof(u32),
 	     I2O_MSG_INPOOL_MIN)) {
 		kfree(c);
 		return ERR_PTR(-ENOMEM);

drivers/net/sky2.c

@@ -187,12 +187,11 @@ static u16 gm_phy_read(struct sky2_hw *hw, unsigned port, u16 reg)
 	return v;
 }
 
-static int sky2_set_power_state(struct sky2_hw *hw, pci_power_t state)
+static void sky2_set_power_state(struct sky2_hw *hw, pci_power_t state)
 {
 	u16 power_control;
 	u32 reg1;
 	int vaux;
-	int ret = 0;
 
 	pr_debug("sky2_set_power_state %d\n", state);
 	sky2_write8(hw, B2_TST_CTRL1, TST_CFG_WRITE_ON);
@@ -275,12 +274,10 @@ static int sky2_set_power_state(struct sky2_hw *hw, pci_power_t state)
 		break;
 	default:
 		printk(KERN_ERR PFX "Unknown power state %d\n", state);
-		ret = -1;
 	}
 
 	sky2_pci_write16(hw, hw->pm_cap + PCI_PM_CTRL, power_control);
 	sky2_write8(hw, B2_TST_CTRL1, TST_CFG_WRITE_OFF);
-	return ret;
 }
 
 static void sky2_phy_reset(struct sky2_hw *hw, unsigned port)
@@ -2164,6 +2161,13 @@ static void sky2_descriptor_error(struct sky2_hw *hw, unsigned port,
 /* If idle then force a fake soft NAPI poll once a second
  * to work around cases where sharing an edge triggered interrupt.
  */
+static inline void sky2_idle_start(struct sky2_hw *hw)
+{
+	if (idle_timeout > 0)
+		mod_timer(&hw->idle_timer,
+			  jiffies + msecs_to_jiffies(idle_timeout));
+}
+
 static void sky2_idle(unsigned long arg)
 {
 	struct sky2_hw *hw = (struct sky2_hw *) arg;
@@ -2183,6 +2187,9 @@ static int sky2_poll(struct net_device *dev0, int *budget)
 	int work_done = 0;
 	u32 status = sky2_read32(hw, B0_Y2_SP_EISR);
 
+	if (!~status)
+		goto out;
+
 	if (status & Y2_IS_HW_ERR)
 		sky2_hw_intr(hw);
 
@@ -2219,7 +2226,7 @@ static int sky2_poll(struct net_device *dev0, int *budget)
 
 	if (sky2_more_work(hw))
 		return 1;
-
+out:
 	netif_rx_complete(dev0);
 
 	sky2_read32(hw, B0_Y2_SP_LISR);
@@ -3350,9 +3357,7 @@ static int __devinit sky2_probe(struct pci_dev *pdev,
 	sky2_write32(hw, B0_IMSK, Y2_IS_BASE);
 
 	setup_timer(&hw->idle_timer, sky2_idle, (unsigned long) hw);
-	if (idle_timeout > 0)
-		mod_timer(&hw->idle_timer,
-			  jiffies + msecs_to_jiffies(idle_timeout));
+	sky2_idle_start(hw);
 
 	pci_set_drvdata(pdev, hw);
 
@@ -3425,8 +3430,14 @@ static int sky2_suspend(struct pci_dev *pdev, pm_message_t state)
 {
 	struct sky2_hw *hw = pci_get_drvdata(pdev);
 	int i;
+	pci_power_t pstate = pci_choose_state(pdev, state);
+
+	if (!(pstate == PCI_D3hot || pstate == PCI_D3cold))
+		return -EINVAL;
+
+	del_timer_sync(&hw->idle_timer);
 
-	for (i = 0; i < 2; i++) {
+	for (i = 0; i < hw->ports; i++) {
 		struct net_device *dev = hw->dev[i];
 
 		if (dev) {
@@ -3438,7 +3449,10 @@ static int sky2_suspend(struct pci_dev *pdev, pm_message_t state)
 		}
 	}
 
-	return sky2_set_power_state(hw, pci_choose_state(pdev, state));
+	sky2_write32(hw, B0_IMSK, 0);
+	pci_save_state(pdev);
+	sky2_set_power_state(hw, pstate);
+	return 0;
 }
 
 static int sky2_resume(struct pci_dev *pdev)
@@ -3448,15 +3462,15 @@ static int sky2_resume(struct pci_dev *pdev)
 
 	pci_restore_state(pdev);
 	pci_enable_wake(pdev, PCI_D0, 0);
-	err = sky2_set_power_state(hw, PCI_D0);
-	if (err)
-		goto out;
+	sky2_set_power_state(hw, PCI_D0);
 
 	err = sky2_reset(hw);
 	if (err)
 		goto out;
 
-	for (i = 0; i < 2; i++) {
+	sky2_write32(hw, B0_IMSK, Y2_IS_BASE);
+
+	for (i = 0; i < hw->ports; i++) {
 		struct net_device *dev = hw->dev[i];
 		if (dev && netif_running(dev)) {
 			netif_device_attach(dev);
@@ -3465,10 +3479,12 @@ static int sky2_resume(struct pci_dev *pdev)
 				printk(KERN_ERR PFX "%s: could not up: %d\n",
 				       dev->name, err);
 				dev_close(dev);
-				break;
+				goto out;
 			}
 		}
 	}
+
+	sky2_idle_start(hw);
 out:
 	return err;
 }

drivers/net/tg3.c

@@ -69,8 +69,8 @@
 
 #define DRV_MODULE_NAME		"tg3"
 #define PFX DRV_MODULE_NAME	": "
-#define DRV_MODULE_VERSION	"3.58"
-#define DRV_MODULE_RELDATE	"May 22, 2006"
+#define DRV_MODULE_VERSION	"3.59"
+#define DRV_MODULE_RELDATE	"June 8, 2006"
 
 #define TG3_DEF_MAC_MODE	0
 #define TG3_DEF_RX_MODE		0
@@ -4485,9 +4485,8 @@ static void tg3_disable_nvram_access(struct tg3 *tp)
 /* tp->lock is held. */
 static void tg3_write_sig_pre_reset(struct tg3 *tp, int kind)
 {
-	if (!(tp->tg3_flags2 & TG3_FLG2_SUN_570X))
-		tg3_write_mem(tp, NIC_SRAM_FIRMWARE_MBOX,
-			      NIC_SRAM_FIRMWARE_MBOX_MAGIC1);
+	tg3_write_mem(tp, NIC_SRAM_FIRMWARE_MBOX,
+		      NIC_SRAM_FIRMWARE_MBOX_MAGIC1);
 
 	if (tp->tg3_flags2 & TG3_FLG2_ASF_NEW_HANDSHAKE) {
 		switch (kind) {
@@ -4568,13 +4567,12 @@ static int tg3_chip_reset(struct tg3 *tp)
 	void (*write_op)(struct tg3 *, u32, u32);
 	int i;
 
-	if (!(tp->tg3_flags2 & TG3_FLG2_SUN_570X)) {
-		tg3_nvram_lock(tp);
-		/* No matching tg3_nvram_unlock() after this because
-		 * chip reset below will undo the nvram lock.
-		 */
-		tp->nvram_lock_cnt = 0;
-	}
+	tg3_nvram_lock(tp);
+
+	/* No matching tg3_nvram_unlock() after this because
+	 * chip reset below will undo the nvram lock.
+	 */
+	tp->nvram_lock_cnt = 0;
 
 	if (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5752 ||
 	    GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5755 ||
@@ -4727,20 +4725,25 @@ static int tg3_chip_reset(struct tg3 *tp)
 		tw32_f(MAC_MODE, 0);
 	udelay(40);
 
-	if (!(tp->tg3_flags2 & TG3_FLG2_SUN_570X)) {
-		/* Wait for firmware initialization to complete. */
-		for (i = 0; i < 100000; i++) {
-			tg3_read_mem(tp, NIC_SRAM_FIRMWARE_MBOX, &val);
-			if (val == ~NIC_SRAM_FIRMWARE_MBOX_MAGIC1)
-				break;
-			udelay(10);
-		}
-		if (i >= 100000) {
-			printk(KERN_ERR PFX "tg3_reset_hw timed out for %s, "
-			       "firmware will not restart magic=%08x\n",
-			       tp->dev->name, val);
-			return -ENODEV;
-		}
+	/* Wait for firmware initialization to complete. */
+	for (i = 0; i < 100000; i++) {
+		tg3_read_mem(tp, NIC_SRAM_FIRMWARE_MBOX, &val);
+		if (val == ~NIC_SRAM_FIRMWARE_MBOX_MAGIC1)
+			break;
+		udelay(10);
+	}
+
+	/* Chip might not be fitted with firmare.  Some Sun onboard
+	 * parts are configured like that.  So don't signal the timeout
+	 * of the above loop as an error, but do report the lack of
+	 * running firmware once.
+	 */
+	if (i >= 100000 &&
+	    !(tp->tg3_flags2 & TG3_FLG2_NO_FWARE_REPORTED)) {
+		tp->tg3_flags2 |= TG3_FLG2_NO_FWARE_REPORTED;
+
+		printk(KERN_INFO PFX "%s: No firmware running.\n",
+		       tp->dev->name);
 	}
 
 	if ((tp->tg3_flags2 & TG3_FLG2_PCI_EXPRESS) &&
@@ -9075,9 +9078,6 @@ static void __devinit tg3_nvram_init(struct tg3 *tp)
 {
 	int j;
 
-	if (tp->tg3_flags2 & TG3_FLG2_SUN_570X)
-		return;
-
 	tw32_f(GRC_EEPROM_ADDR,
 	     (EEPROM_ADDR_FSM_RESET |
 	      (EEPROM_DEFAULT_CLOCK_PERIOD <<
@@ -9210,11 +9210,6 @@ static int tg3_nvram_read(struct tg3 *tp, u32 offset, u32 *val)
 {
 	int ret;
 
-	if (tp->tg3_flags2 & TG3_FLG2_SUN_570X) {
-		printk(KERN_ERR PFX "Attempt to do nvram_read on Sun 570X\n");
-		return -EINVAL;
-	}
-
 	if (!(tp->tg3_flags & TG3_FLAG_NVRAM))
 		return tg3_nvram_read_using_eeprom(tp, offset, val);
 
@@ -9447,11 +9442,6 @@ static int tg3_nvram_write_block(struct tg3 *tp, u32 offset, u32 len, u8 *buf)
 {
 	int ret;
 
-	if (tp->tg3_flags2 & TG3_FLG2_SUN_570X) {
-		printk(KERN_ERR PFX "Attempt to do nvram_write on Sun 570X\n");
-		return -EINVAL;
-	}
-
 	if (tp->tg3_flags & TG3_FLAG_EEPROM_WRITE_PROT) {
 		tw32_f(GRC_LOCAL_CTRL, tp->grc_local_ctrl &
 		       ~GRC_LCLCTRL_GPIO_OUTPUT1);
@@ -9578,15 +9568,19 @@ static void __devinit tg3_get_eeprom_hw_cfg(struct tg3 *tp)
 	pci_write_config_dword(tp->pdev, TG3PCI_MISC_HOST_CTRL,
 			       tp->misc_host_ctrl);
 
+	/* The memory arbiter has to be enabled in order for SRAM accesses
+	 * to succeed.  Normally on powerup the tg3 chip firmware will make
+	 * sure it is enabled, but other entities such as system netboot
+	 * code might disable it.
+	 */
+	val = tr32(MEMARB_MODE);
+	tw32(MEMARB_MODE, val | MEMARB_MODE_ENABLE);
+
 	tp->phy_id = PHY_ID_INVALID;
 	tp->led_ctrl = LED_CTRL_MODE_PHY_1;
 
-	/* Do not even try poking around in here on Sun parts.  */
-	if (tp->tg3_flags2 & TG3_FLG2_SUN_570X) {
-		/* All SUN chips are built-in LOMs. */
-		tp->tg3_flags |= TG3_FLAG_EEPROM_WRITE_PROT;
-		return;
-	}
+	/* Assume an onboard device by default.  */
+	tp->tg3_flags |= TG3_FLAG_EEPROM_WRITE_PROT;
 
 	tg3_read_mem(tp, NIC_SRAM_DATA_SIG, &val);
 	if (val == NIC_SRAM_DATA_SIG_MAGIC) {
@@ -9686,6 +9680,8 @@ static void __devinit tg3_get_eeprom_hw_cfg(struct tg3 *tp)
 
 		if (nic_cfg & NIC_SRAM_DATA_CFG_EEPROM_WP)
 			tp->tg3_flags |= TG3_FLAG_EEPROM_WRITE_PROT;
+		else
+			tp->tg3_flags &= ~TG3_FLAG_EEPROM_WRITE_PROT;
 
 		if (nic_cfg & NIC_SRAM_DATA_CFG_ASF_ENABLE) {
 			tp->tg3_flags |= TG3_FLAG_ENABLE_ASF;
@@ -9834,16 +9830,8 @@ static void __devinit tg3_read_partno(struct tg3 *tp)
 	int i;
 	u32 magic;
 
-	if (tp->tg3_flags2 & TG3_FLG2_SUN_570X) {
-		/* Sun decided not to put the necessary bits in the
-		 * NVRAM of their onboard tg3 parts :(
-		 */
-		strcpy(tp->board_part_number, "Sun 570X");
-		return;
-	}
-
 	if (tg3_nvram_read_swab(tp, 0x0, &magic))
-		return;
+		goto out_not_found;
 
 	if (magic == TG3_EEPROM_MAGIC) {
 		for (i = 0; i < 256; i += 4) {
@@ -9874,6 +9862,9 @@ static void __devinit tg3_read_partno(struct tg3 *tp)
 					break;
 				msleep(1);
 			}
+			if (!(tmp16 & 0x8000))
+				goto out_not_found;
+
 			pci_read_config_dword(tp->pdev, vpd_cap + PCI_VPD_DATA,
 					      &tmp);
 			tmp = cpu_to_le32(tmp);
@@ -9965,37 +9956,6 @@ static void __devinit tg3_read_fw_ver(struct tg3 *tp)
 	}
 }
 
-#ifdef CONFIG_SPARC64
-static int __devinit tg3_is_sun_570X(struct tg3 *tp)
-{
-	struct pci_dev *pdev = tp->pdev;
-	struct pcidev_cookie *pcp = pdev->sysdata;
-
-	if (pcp != NULL) {
-		int node = pcp->prom_node;
-		u32 venid;
-		int err;
-
-		err = prom_getproperty(node, "subsystem-vendor-id",
-				       (char *) &venid, sizeof(venid));
-		if (err == 0 || err == -1)
-			return 0;
-		if (venid == PCI_VENDOR_ID_SUN)
-			return 1;
-
-		/* TG3 chips onboard the SunBlade-2500 don't have the
-		 * subsystem-vendor-id set to PCI_VENDOR_ID_SUN but they
-		 * are distinguishable from non-Sun variants by being
-		 * named "network" by the firmware.  Non-Sun cards will
-		 * show up as being named "ethernet".
-		 */
-		if (!strcmp(pcp->prom_name, "network"))
-			return 1;
-	}
-	return 0;
-}
-#endif
-
 static int __devinit tg3_get_invariants(struct tg3 *tp)
 {
 	static struct pci_device_id write_reorder_chipsets[] = {
@@ -10012,11 +9972,6 @@ static int __devinit tg3_get_invariants(struct tg3 *tp)
 	u16 pci_cmd;
 	int err;
 
-#ifdef CONFIG_SPARC64
-	if (tg3_is_sun_570X(tp))
-		tp->tg3_flags2 |= TG3_FLG2_SUN_570X;
-#endif
-
 	/* Force memory write invalidate off.  If we leave it on,
 	 * then on 5700_BX chips we have to enable a workaround.
 	 * The workaround is to set the TG3PCI_DMA_RW_CTRL boundary
@@ -10312,8 +10267,7 @@ static int __devinit tg3_get_invariants(struct tg3 *tp)
 	if (tp->write32 == tg3_write_indirect_reg32 ||
 	    ((tp->tg3_flags & TG3_FLAG_PCIX_MODE) &&
 	     (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5700 ||
-	      GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5701)) ||
-	    (tp->tg3_flags2 & TG3_FLG2_SUN_570X))
+	      GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5701)))
 		tp->tg3_flags |= TG3_FLAG_SRAM_USE_CONFIG;
 
 	/* Get eeprom hw config before calling tg3_set_power_state().
@@ -10594,8 +10548,7 @@ static int __devinit tg3_get_device_address(struct tg3 *tp)
 #endif
 
 	mac_offset = 0x7c;
-	if ((GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5704 &&
-	     !(tp->tg3_flags & TG3_FLG2_SUN_570X)) ||
+	if ((GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5704) ||
 	    (tp->tg3_flags2 & TG3_FLG2_5780_CLASS)) {
 		if (tr32(TG3PCI_DUAL_MAC_CTRL) & DUAL_MAC_CTRL_ID)
 			mac_offset = 0xcc;
@@ -10622,8 +10575,7 @@ static int __devinit tg3_get_device_address(struct tg3 *tp)
 	}
 	if (!addr_ok) {
 		/* Next, try NVRAM. */
-		if (!(tp->tg3_flags & TG3_FLG2_SUN_570X) &&
-		    !tg3_nvram_read(tp, mac_offset + 0, &hi) &&
+		if (!tg3_nvram_read(tp, mac_offset + 0, &hi) &&
 		    !tg3_nvram_read(tp, mac_offset + 4, &lo)) {
 			dev->dev_addr[0] = ((hi >> 16) & 0xff);
 			dev->dev_addr[1] = ((hi >> 24) & 0xff);

drivers/net/tg3.h

@@ -2184,7 +2184,7 @@ struct tg3 {
 #define TG3_FLAG_INIT_COMPLETE		0x80000000
 	u32				tg3_flags2;
 #define TG3_FLG2_RESTART_TIMER		0x00000001
-#define TG3_FLG2_SUN_570X		0x00000002
+/*					0x00000002 available */
 #define TG3_FLG2_NO_ETH_WIRE_SPEED	0x00000004
 #define TG3_FLG2_IS_5788		0x00000008
 #define TG3_FLG2_MAX_RXPEND_64		0x00000010
@@ -2216,6 +2216,7 @@ struct tg3 {
 #define TG3_FLG2_HW_TSO			(TG3_FLG2_HW_TSO_1 | TG3_FLG2_HW_TSO_2)
 #define TG3_FLG2_1SHOT_MSI		0x10000000
 #define TG3_FLG2_PHY_JITTER_BUG		0x20000000
+#define TG3_FLG2_NO_FWARE_REPORTED	0x40000000
 
 	u32				split_mode_max_reqs;
 #define SPLIT_MODE_5704_MAX_REQ		3

drivers/pci/pci-driver.c

@@ -285,9 +285,9 @@ static int pci_device_suspend(struct device * dev, pm_message_t state)
  * Default resume method for devices that have no driver provided resume,
  * or not even a driver at all.
  */
-static void pci_default_resume(struct pci_dev *pci_dev)
+static int pci_default_resume(struct pci_dev *pci_dev)
 {
-	int retval;
+	int retval = 0;
 
 	/* restore the PCI config space */
 	pci_restore_state(pci_dev);
@@ -297,18 +297,21 @@ static void pci_default_resume(struct pci_dev *pci_dev)
 	/* if the device was busmaster before the suspend, make it busmaster again */
 	if (pci_dev->is_busmaster)
 		pci_set_master(pci_dev);
+
+	return retval;
 }
 
 static int pci_device_resume(struct device * dev)
 {
+	int error;
 	struct pci_dev * pci_dev = to_pci_dev(dev);
 	struct pci_driver * drv = pci_dev->driver;
 
 	if (drv && drv->resume)
-		drv->resume(pci_dev);
+		error = drv->resume(pci_dev);
 	else
-		pci_default_resume(pci_dev);
-	return 0;
+		error = pci_default_resume(pci_dev);
+	return error;
 }
 
 static void pci_device_shutdown(struct device *dev)

drivers/pci/pci.c

@@ -460,9 +460,23 @@ int
 pci_restore_state(struct pci_dev *dev)
 {
 	int i;
+	int val;
 
-	for (i = 0; i < 16; i++)
-		pci_write_config_dword(dev,i * 4, dev->saved_config_space[i]);
+	/*
+	 * The Base Address register should be programmed before the command
+	 * register(s)
+	 */
+	for (i = 15; i >= 0; i--) {
+		pci_read_config_dword(dev, i * 4, &val);
+		if (val != dev->saved_config_space[i]) {
+			printk(KERN_DEBUG "PM: Writing back config space on "
+				"device %s at offset %x (was %x, writing %x)\n",
+				pci_name(dev), i,
+				val, (int)dev->saved_config_space[i]);
+			pci_write_config_dword(dev,i * 4,
+				dev->saved_config_space[i]);
+		}
+	}
 	pci_restore_msi_state(dev);
 	pci_restore_msix_state(dev);
 	return 0;

drivers/scsi/sata_mv.c

@@ -2035,6 +2035,7 @@ static void mv_phy_reset(struct ata_port *ap)
 static void mv_eng_timeout(struct ata_port *ap)
 {
 	struct ata_queued_cmd *qc;
+	unsigned long flags;
 
 	printk(KERN_ERR "ata%u: Entering mv_eng_timeout\n",ap->id);
 	DPRINTK("All regs @ start of eng_timeout\n");
@@ -2046,8 +2047,10 @@ static void mv_eng_timeout(struct ata_port *ap)
 	       ap->host_set->mmio_base, ap, qc, qc->scsicmd,
 	       &qc->scsicmd->cmnd);
 
+	spin_lock_irqsave(&ap->host_set->lock, flags);
 	mv_err_intr(ap, 0);
 	mv_stop_and_reset(ap);
+	spin_unlock_irqrestore(&ap->host_set->lock, flags);
 
 	WARN_ON(!(qc->flags & ATA_QCFLAG_ACTIVE));
 	if (qc->flags & ATA_QCFLAG_ACTIVE) {

drivers/usb/host/ohci-pxa27x.c

@@ -185,6 +185,9 @@ int usb_hcd_pxa27x_probe (const struct hc_driver *driver, struct platform_device
 	/* Select Power Management Mode */
 	pxa27x_ohci_select_pmm(inf->port_mode);
 
+	if (inf->power_budget)
+		hcd->power_budget = inf->power_budget;
+
 	ohci_hcd_init(hcd_to_ohci(hcd));
 
 	retval = usb_add_hcd(hcd, pdev->resource[1].start, SA_INTERRUPT);

drivers/video/console/fbcon.c

@@ -1745,7 +1745,7 @@ static int fbcon_scroll(struct vc_data *vc, int t, int b, int dir,
 					fbcon_redraw_move(vc, p, 0, t, count);
 				ypan_up_redraw(vc, t, count);
 				if (vc->vc_rows - b > 0)
-					fbcon_redraw_move(vc, p, b - count,
+					fbcon_redraw_move(vc, p, b,
 							  vc->vc_rows - b, b);
 			} else
 				fbcon_redraw_move(vc, p, t + count, b - t - count, t);

fs/debugfs/inode.c

@@ -67,12 +67,13 @@ static struct inode *debugfs_get_inode(struct super_block *sb, int mode, dev_t d
 static int debugfs_mknod(struct inode *dir, struct dentry *dentry,
 			 int mode, dev_t dev)
 {
-	struct inode *inode = debugfs_get_inode(dir->i_sb, mode, dev);
+	struct inode *inode;
 	int error = -EPERM;
 
 	if (dentry->d_inode)
 		return -EEXIST;
 
+	inode = debugfs_get_inode(dir->i_sb, mode, dev);
 	if (inode) {
 		d_instantiate(dentry, inode);
 		dget(dentry);

include/asm-arm/arch-pxa/ohci.h

@@ -11,6 +11,8 @@ struct pxaohci_platform_data {
 #define PMM_NPS_MODE           1
 #define PMM_GLOBAL_MODE        2
 #define PMM_PERPORT_MODE       3
+
+	int power_budget;
 };
 
 extern void pxa_set_ohci_info(struct pxaohci_platform_data *info);

include/asm-s390/futex.h

@@ -11,23 +11,24 @@
 #define __futex_atomic_fixup \
 		     ".section __ex_table,\"a\"\n"			\
 		     "   .align 4\n"					\
-		     "   .long  0b,2b,1b,2b\n"				\
+		     "   .long  0b,4b,2b,4b,3b,4b\n"			\
 		     ".previous"
 #else /* __s390x__ */
 #define __futex_atomic_fixup \
 		     ".section __ex_table,\"a\"\n"			\
 		     "   .align 8\n"					\
-		     "   .quad  0b,2b,1b,2b\n"				\
+		     "   .quad  0b,4b,2b,4b,3b,4b\n"			\
 		     ".previous"
 #endif /* __s390x__ */
 
 #define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg)	\
-	asm volatile("   l   %1,0(%6)\n"				\
-		     "0: " insn						\
-		     "   cs  %1,%2,0(%6)\n"				\
-		     "1: jl  0b\n"					\
+	asm volatile("   sacf 256\n"					\
+		     "0: l   %1,0(%6)\n"				\
+		     "1: " insn						\
+		     "2: cs  %1,%2,0(%6)\n"				\
+		     "3: jl  1b\n"					\
 		     "   lhi %0,0\n"					\
-		     "2:\n"						\
+		     "4: sacf 0\n"					\
 		     __futex_atomic_fixup				\
 		     : "=d" (ret), "=&d" (oldval), "=&d" (newval),	\
 		       "=m" (*uaddr)					\

include/linux/elevator.h

@@ -21,7 +21,7 @@ typedef void (elevator_put_req_fn) (request_queue_t *, struct request *);
 typedef void (elevator_activate_req_fn) (request_queue_t *, struct request *);
 typedef void (elevator_deactivate_req_fn) (request_queue_t *, struct request *);
 
-typedef int (elevator_init_fn) (request_queue_t *, elevator_t *);
+typedef void *(elevator_init_fn) (request_queue_t *, elevator_t *);
 typedef void (elevator_exit_fn) (elevator_t *);
 
 struct elevator_ops

include/linux/i2o.h

@@ -1114,8 +1114,11 @@ static inline struct i2o_message *i2o_msg_get(struct i2o_controller *c)
 
 	mmsg->mfa = readl(c->in_port);
 	if (unlikely(mmsg->mfa >= c->in_queue.len)) {
+		u32 mfa = mmsg->mfa;
+
 		mempool_free(mmsg, c->in_msg.mempool);
-		if(mmsg->mfa == I2O_QUEUE_EMPTY)
+
+		if (mfa == I2O_QUEUE_EMPTY)
 			return ERR_PTR(-EBUSY);
 		return ERR_PTR(-EFAULT);
 	}

include/linux/mempolicy.h

@@ -36,6 +36,7 @@
 #include <linux/nodemask.h>
 
 struct vm_area_struct;
+struct mm_struct;
 
 #ifdef CONFIG_NUMA
 

include/linux/pci-acpi.h

@@ -50,7 +50,7 @@
 extern acpi_status pci_osc_control_set(acpi_handle handle, u32 flags);
 extern acpi_status pci_osc_support_set(u32 flags);
 #else
-#if !defined(acpi_status)
+#if !defined(AE_ERROR)
 typedef u32 		acpi_status;
 #define AE_ERROR      	(acpi_status) (0x0001)
 #endif    

mm/shmem.c

@@ -1780,6 +1780,7 @@ static int shmem_rmdir(struct inode *dir, struct dentry *dentry)
 	if (!simple_empty(dentry))
 		return -ENOTEMPTY;
 
+	dentry->d_inode->i_nlink--;
 	dir->i_nlink--;
 	return shmem_unlink(dir, dentry);
 }
@@ -2102,6 +2103,7 @@ static int shmem_fill_super(struct super_block *sb,
 	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
 	sb->s_magic = TMPFS_MAGIC;
 	sb->s_op = &shmem_ops;
+	sb->s_time_gran = 1;
 
 	inode = shmem_get_inode(sb, S_IFDIR | mode, 0);
 	if (!inode)

mm/vmscan.c

@@ -1061,7 +1061,7 @@ static unsigned long balance_pgdat(pg_data_t *pgdat, unsigned long nr_pages,
 loop_again:
 	total_scanned = 0;
 	nr_reclaimed = 0;
-	sc.may_writepage = !laptop_mode,
+	sc.may_writepage = !laptop_mode;
 	sc.nr_mapped = read_page_state(nr_mapped);
 
 	inc_page_state(pageoutrun);

net/dccp/ackvec.c

@@ -452,6 +452,7 @@ static void dccp_ackvec_check_rcv_ackvector(struct dccp_ackvec *av,
 					      (unsigned long long)
 					      avr->dccpavr_ack_ackno);
 				dccp_ackvec_throw_record(av, avr);
+				break;
 			}
 			/*
 			 * If it wasn't received, continue scanning... we might

net/ipv4/ip_forward.c

@@ -116,6 +116,7 @@ int ip_forward(struct sk_buff *skb)
 
 too_many_hops:
         /* Tell the sender its packet died... */
+        IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS);
         icmp_send(skb, ICMP_TIME_EXCEEDED, ICMP_EXC_TTL, 0);
 drop:
 	kfree_skb(skb);

net/ipv4/tcp_input.c

@@ -1649,7 +1649,7 @@ static void tcp_update_scoreboard(struct sock *sk, struct tcp_sock *tp)
 	 * Hence, we can detect timed out packets during fast
 	 * retransmit without falling to slow start.
 	 */
-	if (tcp_head_timedout(sk, tp)) {
+	if (!IsReno(tp) && tcp_head_timedout(sk, tp)) {
 		struct sk_buff *skb;
 
 		skb = tp->scoreboard_skb_hint ? tp->scoreboard_skb_hint
@@ -1662,8 +1662,6 @@ static void tcp_update_scoreboard(struct sock *sk, struct tcp_sock *tp)
 			if (!(TCP_SKB_CB(skb)->sacked&TCPCB_TAGBITS)) {
 				TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
 				tp->lost_out += tcp_skb_pcount(skb);
-				if (IsReno(tp))
-					tcp_remove_reno_sacks(sk, tp, tcp_skb_pcount(skb) + 1);
 
 				/* clear xmit_retrans hint */
 				if (tp->retransmit_skb_hint &&