Merge branch 's3c-fixes-rc4' of git://aeryn.fluff.org.uk/bjdooks/linux

Documentation/filesystems/sysfs.txt

@@ -23,7 +23,8 @@ interface.
 Using sysfs
 ~~~~~~~~~~~
 
-sysfs is always compiled in. You can access it by doing:
+sysfs is always compiled in if CONFIG_SYSFS is defined. You can access
+it by doing:
 
     mount -t sysfs sysfs /sys 
 

Documentation/sound/alsa/Procfile.txt

@@ -101,6 +101,8 @@ card*/pcm*/xrun_debug
 	  bit 0 = Enable XRUN/jiffies debug messages
 	  bit 1 = Show stack trace at XRUN / jiffies check
 	  bit 2 = Enable additional jiffies check
+	  bit 3 = Log hwptr update at each period interrupt
+	  bit 4 = Log hwptr update at each snd_pcm_update_hw_ptr()
 
 	When the bit 0 is set, the driver will show the messages to
 	kernel log when an xrun is detected.  The debug message is
@@ -117,6 +119,9 @@ card*/pcm*/xrun_debug
 	buggy) hardware that doesn't give smooth pointer updates.
 	This feature is enabled via the bit 2.
 
+	Bits 3 and 4 are for logging the hwptr records.  Note that
+	these will give flood of kernel messages.
+
 card*/pcm*/sub*/info
 	The general information of this PCM sub-stream.
 

Documentation/video4linux/CARDLIST.em28xx

@@ -20,7 +20,7 @@
  19 -> EM2860/SAA711X Reference Design          (em2860)
  20 -> AMD ATI TV Wonder HD 600                 (em2880)        [0438:b002]
  21 -> eMPIA Technology, Inc. GrabBeeX+ Video Encoder (em2800)        [eb1a:2801]
- 22 -> Unknown EM2750/EM2751 webcam grabber     (em2750)        [eb1a:2750,eb1a:2751]
+ 22 -> EM2710/EM2750/EM2751 webcam grabber      (em2750)        [eb1a:2750,eb1a:2751]
  23 -> Huaqi DLCW-130                           (em2750)
  24 -> D-Link DUB-T210 TV Tuner                 (em2820/em2840) [2001:f112]
  25 -> Gadmei UTV310                            (em2820/em2840)

Documentation/video4linux/gspca.txt

@@ -44,7 +44,9 @@ zc3xx		0458:7007	Genius VideoCam V2
 zc3xx		0458:700c	Genius VideoCam V3
 zc3xx		0458:700f	Genius VideoCam Web V2
 sonixj		0458:7025	Genius Eye 311Q
+sn9c20x		0458:7029	Genius Look 320s
 sonixj		0458:702e	Genius Slim 310 NB
+sn9c20x		045e:00f4	LifeCam VX-6000 (SN9C20x + OV9650)
 sonixj		045e:00f5	MicroSoft VX3000
 sonixj		045e:00f7	MicroSoft VX1000
 ov519		045e:028c	Micro$oft xbox cam
@@ -282,6 +284,28 @@ sonixj		0c45:613a	Microdia Sonix PC Camera
 sonixj		0c45:613b	Surfer SN-206
 sonixj		0c45:613c	Sonix Pccam168
 sonixj		0c45:6143	Sonix Pccam168
+sn9c20x		0c45:6240	PC Camera (SN9C201 + MT9M001)
+sn9c20x		0c45:6242	PC Camera (SN9C201 + MT9M111)
+sn9c20x		0c45:6248	PC Camera (SN9C201 + OV9655)
+sn9c20x		0c45:624e	PC Camera (SN9C201 + SOI968)
+sn9c20x		0c45:624f	PC Camera (SN9C201 + OV9650)
+sn9c20x		0c45:6251	PC Camera (SN9C201 + OV9650)
+sn9c20x		0c45:6253	PC Camera (SN9C201 + OV9650)
+sn9c20x		0c45:6260	PC Camera (SN9C201 + OV7670)
+sn9c20x		0c45:6270	PC Camera (SN9C201 + MT9V011/MT9V111/MT9V112)
+sn9c20x		0c45:627b	PC Camera (SN9C201 + OV7660)
+sn9c20x		0c45:627c	PC Camera (SN9C201 + HV7131R)
+sn9c20x		0c45:627f	PC Camera (SN9C201 + OV9650)
+sn9c20x		0c45:6280	PC Camera (SN9C202 + MT9M001)
+sn9c20x		0c45:6282	PC Camera (SN9C202 + MT9M111)
+sn9c20x		0c45:6288	PC Camera (SN9C202 + OV9655)
+sn9c20x		0c45:628e	PC Camera (SN9C202 + SOI968)
+sn9c20x		0c45:628f	PC Camera (SN9C202 + OV9650)
+sn9c20x		0c45:62a0	PC Camera (SN9C202 + OV7670)
+sn9c20x		0c45:62b0	PC Camera (SN9C202 + MT9V011/MT9V111/MT9V112)
+sn9c20x		0c45:62b3	PC Camera (SN9C202 + OV9655)
+sn9c20x		0c45:62bb	PC Camera (SN9C202 + OV7660)
+sn9c20x		0c45:62bc	PC Camera (SN9C202 + HV7131R)
 sunplus		0d64:0303	Sunplus FashionCam DXG
 etoms		102c:6151	Qcam Sangha CIF
 etoms		102c:6251	Qcam xxxxxx VGA
@@ -290,6 +314,7 @@ spca561		10fd:7e50	FlyCam Usb 100
 zc3xx		10fd:8050	Typhoon Webshot II USB 300k
 ov534		1415:2000	Sony HD Eye for PS3 (SLEH 00201)
 pac207		145f:013a	Trust WB-1300N
+sn9c20x		145f:013d	Trust WB-3600R
 vc032x		15b8:6001	HP 2.0 Megapixel
 vc032x		15b8:6002	HP 2.0 Megapixel rz406aa
 spca501		1776:501c	Arowana 300K CMOS Camera
@@ -300,4 +325,11 @@ spca500		2899:012c	Toptro Industrial
 spca508		8086:0110	Intel Easy PC Camera
 spca500		8086:0630	Intel Pocket PC Camera
 spca506		99fa:8988	Grandtec V.cap
+sn9c20x		a168:0610	Dino-Lite Digital Microscope (SN9C201 + HV7131R)
+sn9c20x		a168:0611	Dino-Lite Digital Microscope (SN9C201 + HV7131R)
+sn9c20x		a168:0613	Dino-Lite Digital Microscope (SN9C201 + HV7131R)
+sn9c20x		a168:0618	Dino-Lite Digital Microscope (SN9C201 + HV7131R)
+sn9c20x		a168:0614	Dino-Lite Digital Microscope (SN9C201 + MT9M111)
+sn9c20x		a168:0615	Dino-Lite Digital Microscope (SN9C201 + MT9M111)
+sn9c20x		a168:0617	Dino-Lite Digital Microscope (SN9C201 + MT9M111)
 spca561		abcd:cdee	Petcam

arch/alpha/include/asm/tlb.h

@@ -9,7 +9,7 @@
 
 #include <asm-generic/tlb.h>
 
-#define __pte_free_tlb(tlb, pte)			pte_free((tlb)->mm, pte)
+#define __pte_free_tlb(tlb, pte, address)		pte_free((tlb)->mm, pte)
-#define __pmd_free_tlb(tlb, pmd)			pmd_free((tlb)->mm, pmd)
+#define __pmd_free_tlb(tlb, pmd, address)		pmd_free((tlb)->mm, pmd)
  
 #endif

arch/arm/include/asm/tlb.h

@@ -102,8 +102,8 @@ tlb_end_vma(struct mmu_gather *tlb, struct vm_area_struct *vma)
 }
 
 #define tlb_remove_page(tlb,page)	free_page_and_swap_cache(page)
-#define pte_free_tlb(tlb, ptep)		pte_free((tlb)->mm, ptep)
+#define pte_free_tlb(tlb, ptep, addr)	pte_free((tlb)->mm, ptep)
-#define pmd_free_tlb(tlb, pmdp)		pmd_free((tlb)->mm, pmdp)
+#define pmd_free_tlb(tlb, pmdp, addr)	pmd_free((tlb)->mm, pmdp)
 
 #define tlb_migrate_finish(mm)		do { } while (0)
 

arch/arm/mach-s3c2410/include/mach/gpio-core.h

@@ -28,7 +28,7 @@ static inline struct s3c_gpio_chip *s3c_gpiolib_getchip(unsigned int pin)
 		return NULL;
 
 	chip = &s3c24xx_gpios[pin/32];
-	return (S3C2410_GPIO_OFFSET(pin) > chip->chip.ngpio) ? chip : NULL;
+	return (S3C2410_GPIO_OFFSET(pin) < chip->chip.ngpio) ? chip : NULL;
 }
 
 #endif /* __ASM_ARCH_GPIO_CORE_H */

arch/arm/plat-s3c64xx/s3c6400-clock.c

@@ -153,7 +153,7 @@ static unsigned long s3c64xx_clk_arm_round_rate(struct clk *clk,
 	u32 div;
 
 	if (parent < rate)
-		return rate;
+		return parent;
 
 	div = (parent / rate) - 1;
 	if (div > armclk_mask)
@@ -175,7 +175,7 @@ static int s3c64xx_clk_arm_set_rate(struct clk *clk, unsigned long rate)
 	div = clk_get_rate(clk->parent) / rate;
 
 	val = __raw_readl(S3C_CLK_DIV0);
-	val &= armclk_mask;
+	val &= ~armclk_mask;
 	val |= (div - 1);
 	__raw_writel(val, S3C_CLK_DIV0);
 

arch/avr32/include/asm/pgalloc.h

@@ -83,7 +83,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	quicklist_free_page(QUICK_PT, NULL, pte);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,addr)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb), pte);			\

arch/cris/include/asm/pgalloc.h

@@ -47,7 +47,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	__free_page(pte);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,address)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb), pte);			\

arch/frv/include/asm/pgalloc.h

@@ -49,7 +49,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	__free_page(pte);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,address)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb),(pte));			\
@@ -62,7 +62,7 @@ do {							\
  */
 #define pmd_alloc_one(mm, addr)		({ BUG(); ((pmd_t *) 2); })
 #define pmd_free(mm, x)			do { } while (0)
-#define __pmd_free_tlb(tlb,x)		do { } while (0)
+#define __pmd_free_tlb(tlb,x,a)		do { } while (0)
 
 #endif /* CONFIG_MMU */
 

arch/frv/include/asm/pgtable.h

@@ -225,7 +225,7 @@ static inline pud_t *pud_offset(pgd_t *pgd, unsigned long address)
  */
 #define pud_alloc_one(mm, address)		NULL
 #define pud_free(mm, x)				do { } while (0)
-#define __pud_free_tlb(tlb, x)			do { } while (0)
+#define __pud_free_tlb(tlb, x, address)		do { } while (0)
 
 /*
  * The "pud_xxx()" functions here are trivial for a folded two-level

arch/ia64/include/asm/pgalloc.h

@@ -48,7 +48,7 @@ static inline void pud_free(struct mm_struct *mm, pud_t *pud)
 {
 	quicklist_free(0, NULL, pud);
 }
-#define __pud_free_tlb(tlb, pud)	pud_free((tlb)->mm, pud)
+#define __pud_free_tlb(tlb, pud, address)	pud_free((tlb)->mm, pud)
 #endif /* CONFIG_PGTABLE_4 */
 
 static inline void
@@ -67,7 +67,7 @@ static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
 	quicklist_free(0, NULL, pmd);
 }
 
-#define __pmd_free_tlb(tlb, pmd)	pmd_free((tlb)->mm, pmd)
+#define __pmd_free_tlb(tlb, pmd, address)	pmd_free((tlb)->mm, pmd)
 
 static inline void
 pmd_populate(struct mm_struct *mm, pmd_t * pmd_entry, pgtable_t pte)
@@ -117,6 +117,6 @@ static inline void check_pgt_cache(void)
 	quicklist_trim(0, NULL, 25, 16);
 }
 
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, pte)
+#define __pte_free_tlb(tlb, pte, address)	pte_free((tlb)->mm, pte)
 
 #endif				/* _ASM_IA64_PGALLOC_H */

arch/ia64/include/asm/tlb.h

@@ -236,22 +236,22 @@ do {							\
 	__tlb_remove_tlb_entry(tlb, ptep, addr);	\
 } while (0)
 
-#define pte_free_tlb(tlb, ptep)				\
+#define pte_free_tlb(tlb, ptep, address)		\
 do {							\
 	tlb->need_flush = 1;				\
-	__pte_free_tlb(tlb, ptep);			\
+	__pte_free_tlb(tlb, ptep, address);		\
 } while (0)
 
-#define pmd_free_tlb(tlb, ptep)				\
+#define pmd_free_tlb(tlb, ptep, address)		\
 do {							\
 	tlb->need_flush = 1;				\
-	__pmd_free_tlb(tlb, ptep);			\
+	__pmd_free_tlb(tlb, ptep, address);		\
 } while (0)
 
-#define pud_free_tlb(tlb, pudp)				\
+#define pud_free_tlb(tlb, pudp, address)		\
 do {							\
 	tlb->need_flush = 1;				\
-	__pud_free_tlb(tlb, pudp);			\
+	__pud_free_tlb(tlb, pudp, address);		\
 } while (0)
 
 #endif /* _ASM_IA64_TLB_H */

arch/m32r/include/asm/pgalloc.h

@@ -58,7 +58,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	__free_page(pte);
 }
 
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, (pte))
+#define __pte_free_tlb(tlb, pte, addr)	pte_free((tlb)->mm, (pte))
 
 /*
  * allocating and freeing a pmd is trivial: the 1-entry pmd is
@@ -68,7 +68,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 
 #define pmd_alloc_one(mm, addr)		({ BUG(); ((pmd_t *)2); })
 #define pmd_free(mm, x)			do { } while (0)
-#define __pmd_free_tlb(tlb, x)		do { } while (0)
+#define __pmd_free_tlb(tlb, x, addr)	do { } while (0)
 #define pgd_populate(mm, pmd, pte)	BUG()
 
 #define check_pgt_cache()	do { } while (0)

arch/m68k/include/asm/motorola_pgalloc.h

@@ -54,7 +54,8 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t page)
 	__free_page(page);
 }
 
-static inline void __pte_free_tlb(struct mmu_gather *tlb, pgtable_t page)
+static inline void __pte_free_tlb(struct mmu_gather *tlb, pgtable_t page,
+				  unsigned long address)
 {
 	pgtable_page_dtor(page);
 	cache_page(kmap(page));
@@ -73,7 +74,8 @@ static inline int pmd_free(struct mm_struct *mm, pmd_t *pmd)
 	return free_pointer_table(pmd);
 }
 
-static inline int __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
+static inline int __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd,
+				 unsigned long address)
 {
 	return free_pointer_table(pmd);
 }

arch/m68k/include/asm/sun3_pgalloc.h

@@ -32,7 +32,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t page)
         __free_page(page);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,addr)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb), pte);			\
@@ -80,7 +80,7 @@ static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd, pgtable_t page
  * inside the pgd, so has no extra memory associated with it.
  */
 #define pmd_free(mm, x)			do { } while (0)
-#define __pmd_free_tlb(tlb, x)		do { } while (0)
+#define __pmd_free_tlb(tlb, x, addr)	do { } while (0)
 
 static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
 {

arch/microblaze/Makefile

@@ -6,14 +6,16 @@ endif
 
 # What CPU vesion are we building for, and crack it open
 # as major.minor.rev
-CPU_VER=$(subst ",,$(CONFIG_XILINX_MICROBLAZE0_HW_VER) )
+CPU_VER   := $(shell echo $(CONFIG_XILINX_MICROBLAZE0_HW_VER))
-CPU_MAJOR=$(shell echo $(CPU_VER) | cut -d '.' -f 1)
+CPU_MAJOR := $(shell echo $(CPU_VER) | cut -d '.' -f 1)
-CPU_MINOR=$(shell echo $(CPU_VER) | cut -d '.' -f 2)
+CPU_MINOR := $(shell echo $(CPU_VER) | cut -d '.' -f 2)
-CPU_REV=$(shell echo $(CPU_VER) | cut -d '.' -f 3)
+CPU_REV   := $(shell echo $(CPU_VER) | cut -d '.' -f 3)
 
 export CPU_VER CPU_MAJOR CPU_MINOR CPU_REV
 
 # Use cpu-related CONFIG_ vars to set compile options.
+# The various CONFIG_XILINX cpu features options are integers 0/1/2...
+# rather than bools y/n
 
 # Work out HW multipler support.  This is icky.
 # 1. Spartan2 has no HW multiplers.
@@ -34,30 +36,29 @@ CPUFLAGS-$(CONFIG_XILINX_MICROBLAZE0_USE_PCMP_INSTR) += -mxl-pattern-compare
 
 CPUFLAGS-1 += $(call cc-option,-mcpu=v$(CPU_VER))
 
-# The various CONFIG_XILINX cpu features options are integers 0/1/2...
-# rather than bools y/n
-
 # r31 holds current when in kernel mode
-CFLAGS_KERNEL += -ffixed-r31 $(CPUFLAGS-1) $(CPUFLAGS-2)
+KBUILD_KERNEL += -ffixed-r31 $(CPUFLAGS-1) $(CPUFLAGS-2)
 
 LDFLAGS		:=
 LDFLAGS_vmlinux	:=
-LDFLAGS_BLOB := --format binary --oformat elf32-microblaze
 
-LIBGCC := $(shell $(CC) $(CFLAGS_KERNEL) -print-libgcc-file-name)
+LIBGCC := $(shell $(CC) $(KBUILD_KERNEL) -print-libgcc-file-name)
 
 head-y := arch/microblaze/kernel/head.o
-libs-y		+= arch/microblaze/lib/ $(LIBGCC)
+libs-y += arch/microblaze/lib/
-core-y		+= arch/microblaze/kernel/ arch/microblaze/mm/ \
+libs-y += $(LIBGCC)
-		   arch/microblaze/platform/
+core-y += arch/microblaze/kernel/
+core-y += arch/microblaze/mm/
+core-y += arch/microblaze/platform/
 
-boot := arch/$(ARCH)/boot
+boot := arch/microblaze/boot
 
 # defines filename extension depending memory management type
 ifeq ($(CONFIG_MMU),)
-MMUEXT		:= -nommu
+MMU := -nommu
 endif
-export	MMUEXT
+
+export MMU
 
 all: linux.bin
 

arch/microblaze/include/asm/io.h

@@ -14,7 +14,6 @@
 #include <asm/byteorder.h>
 #include <asm/page.h>
 #include <linux/types.h>
-#include <asm/byteorder.h>
 #include <linux/mm.h>          /* Get struct page {...} */
 
 

arch/microblaze/include/asm/pgalloc.h

@@ -180,7 +180,7 @@ extern inline void pte_free(struct mm_struct *mm, struct page *ptepage)
 	__free_page(ptepage);
 }
 
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, (pte))
+#define __pte_free_tlb(tlb, pte, addr)	pte_free((tlb)->mm, (pte))
 
 #define pmd_populate(mm, pmd, pte)	(pmd_val(*(pmd)) = page_address(pte))
 
@@ -193,7 +193,7 @@ extern inline void pte_free(struct mm_struct *mm, struct page *ptepage)
  */
 #define pmd_alloc_one(mm, address)	({ BUG(); ((pmd_t *)2); })
 /*#define pmd_free(mm, x)			do { } while (0)*/
-#define __pmd_free_tlb(tlb, x)		do { } while (0)
+#define __pmd_free_tlb(tlb, x, addr)	do { } while (0)
 #define pgd_populate(mm, pmd, pte)	BUG()
 
 extern int do_check_pgt_cache(int, int);

arch/microblaze/include/asm/pgtable.h

@@ -185,6 +185,7 @@ static inline pte_t pte_mkspecial(pte_t pte)	{ return pte; }
 
 /* Definitions for MicroBlaze. */
 #define	_PAGE_GUARDED	0x001	/* G: page is guarded from prefetch */
+#define _PAGE_FILE	0x001	/* when !present: nonlinear file mapping */
 #define _PAGE_PRESENT	0x002	/* software: PTE contains a translation */
 #define	_PAGE_NO_CACHE	0x004	/* I: caching is inhibited */
 #define	_PAGE_WRITETHRU	0x008	/* W: caching is write-through */
@@ -320,8 +321,7 @@ static inline int pte_write(pte_t pte) { return pte_val(pte) & _PAGE_RW; }
 static inline int pte_exec(pte_t pte)  { return pte_val(pte) & _PAGE_EXEC; }
 static inline int pte_dirty(pte_t pte) { return pte_val(pte) & _PAGE_DIRTY; }
 static inline int pte_young(pte_t pte) { return pte_val(pte) & _PAGE_ACCESSED; }
-/* FIXME */
+static inline int pte_file(pte_t pte)  { return pte_val(pte) & _PAGE_FILE; }
-static inline int pte_file(pte_t pte)		{ return 0; }
 
 static inline void pte_uncache(pte_t pte) { pte_val(pte) |= _PAGE_NO_CACHE; }
 static inline void pte_cache(pte_t pte)   { pte_val(pte) &= ~_PAGE_NO_CACHE; }
@@ -488,7 +488,7 @@ static inline pmd_t *pmd_offset(pgd_t *dir, unsigned long address)
 /* Encode and decode a nonlinear file mapping entry */
 #define PTE_FILE_MAX_BITS	29
 #define pte_to_pgoff(pte)	(pte_val(pte) >> 3)
-#define pgoff_to_pte(off)	((pte_t) { ((off) << 3) })
+#define pgoff_to_pte(off)	((pte_t) { ((off) << 3) | _PAGE_FILE })
 
 extern pgd_t swapper_pg_dir[PTRS_PER_PGD];
 

arch/microblaze/include/asm/prom.h

@@ -16,6 +16,18 @@
 #define _ASM_MICROBLAZE_PROM_H
 #ifdef __KERNEL__
 
+/* Definitions used by the flattened device tree */
+#define OF_DT_HEADER		0xd00dfeed /* marker */
+#define OF_DT_BEGIN_NODE	0x1 /* Start of node, full name */
+#define OF_DT_END_NODE		0x2 /* End node */
+#define OF_DT_PROP		0x3 /* Property: name off, size, content */
+#define OF_DT_NOP		0x4 /* nop */
+#define OF_DT_END		0x9
+
+#define OF_DT_VERSION		0x10
+
+#ifndef __ASSEMBLY__
+
 #include <linux/types.h>
 #include <linux/proc_fs.h>
 #include <linux/platform_device.h>
@@ -29,16 +41,6 @@
 #define of_prop_cmp(s1, s2)		strcmp((s1), (s2))
 #define of_node_cmp(s1, s2)		strcasecmp((s1), (s2))
 
-/* Definitions used by the flattened device tree */
-#define OF_DT_HEADER		0xd00dfeed /* marker */
-#define OF_DT_BEGIN_NODE	0x1 /* Start of node, full name */
-#define OF_DT_END_NODE		0x2 /* End node */
-#define OF_DT_PROP		0x3 /* Property: name off, size, content */
-#define OF_DT_NOP		0x4 /* nop */
-#define OF_DT_END		0x9
-
-#define OF_DT_VERSION		0x10
-
 /*
  * This is what gets passed to the kernel by prom_init or kexec
  *
@@ -309,5 +311,6 @@ extern void __iomem *of_iomap(struct device_node *device, int index);
  */
 #include <linux/of.h>
 
+#endif /* __ASSEMBLY__ */
 #endif /* __KERNEL__ */
 #endif /* _ASM_MICROBLAZE_PROM_H */

arch/microblaze/include/asm/tlb.h

@@ -11,7 +11,7 @@
 #ifndef _ASM_MICROBLAZE_TLB_H
 #define _ASM_MICROBLAZE_TLB_H
 
-#define tlb_flush(tlb)	do {} while (0)
+#define tlb_flush(tlb)	flush_tlb_mm((tlb)->mm)
 
 #include <asm-generic/tlb.h>
 

arch/microblaze/include/asm/uaccess.h

@@ -189,7 +189,7 @@ extern long strnlen_user(const char *src, long count);
 
 #define __put_user(x, ptr)						\
 ({									\
-	__typeof__(*(ptr)) __gu_val = x;				\
+	__typeof__(*(ptr)) volatile __gu_val = (x);			\
 	long __gu_err = 0;						\
 	switch (sizeof(__gu_val)) {					\
 	case 1:								\

arch/microblaze/kernel/Makefile

@@ -17,4 +17,4 @@ obj-$(CONFIG_HEART_BEAT)	+= heartbeat.o
 obj-$(CONFIG_MODULES)		+= microblaze_ksyms.o module.o
 obj-$(CONFIG_MMU)		+= misc.o
 
-obj-y	+= entry$(MMUEXT).o
+obj-y	+= entry$(MMU).o

arch/microblaze/kernel/cpu/cpuinfo-pvr-full.c

@@ -22,7 +22,7 @@
 
 #define CI(c, p) { ci->c = PVR_##p(pvr); }
 #define err_printk(x) \
-	early_printk("ERROR: Microblaze " x " - different for PVR and DTS\n");
+	early_printk("ERROR: Microblaze " x "-different for PVR and DTS\n");
 
 void set_cpuinfo_pvr_full(struct cpuinfo *ci, struct device_node *cpu)
 {

arch/microblaze/kernel/cpu/cpuinfo-static.c

@@ -18,7 +18,7 @@ static const char family_string[] = CONFIG_XILINX_MICROBLAZE0_FAMILY;
 static const char cpu_ver_string[] = CONFIG_XILINX_MICROBLAZE0_HW_VER;
 
 #define err_printk(x) \
-	early_printk("ERROR: Microblaze " x "- different for kernel and DTS\n");
+	early_printk("ERROR: Microblaze " x "-different for kernel and DTS\n");
 
 void __init set_cpuinfo_static(struct cpuinfo *ci, struct device_node *cpu)
 {

arch/microblaze/kernel/cpu/cpuinfo.c

@@ -26,6 +26,8 @@ const struct cpu_ver_key cpu_ver_lookup[] = {
 	{"7.10.b", 0x09},
 	{"7.10.c", 0x0a},
 	{"7.10.d", 0x0b},
+	{"7.20.a", 0x0c},
+	{"7.20.b", 0x0d},
 	/* FIXME There is no keycode defined in MBV for these versions */
 	{"2.10.a", 0x10},
 	{"3.00.a", 0x20},

arch/microblaze/kernel/head.S

@@ -31,6 +31,7 @@
 #include <linux/linkage.h>
 #include <asm/thread_info.h>
 #include <asm/page.h>
+#include <asm/prom.h>		/* for OF_DT_HEADER */
 
 #ifdef CONFIG_MMU
 #include <asm/setup.h> /* COMMAND_LINE_SIZE */
@@ -54,11 +55,19 @@ ENTRY(_start)
 	andi	r1, r1, ~2
 	mts	rmsr, r1
 
-/* save fdt to kernel location */
+/* r7 may point to an FDT, or there may be one linked in.
-/* r7 stores pointer to fdt blob */
+   if it's in r7, we've got to save it away ASAP.
-	beqi	r7, no_fdt_arg
+   We ensure r7 points to a valid FDT, just in case the bootloader
+   is broken or non-existent */
+	beqi	r7, no_fdt_arg			/* NULL pointer?  don't copy */
+	lw	r11, r0, r7			/* Does r7 point to a */
+	rsubi	r11, r11, OF_DT_HEADER		/* valid FDT? */
+	beqi	r11, _prepare_copy_fdt
+	or	r7, r0, r0		/* clear R7 when not valid DTB */
+	bnei	r11, no_fdt_arg			/* No - get out of here */
+_prepare_copy_fdt:
 	or	r11, r0, r0 /* incremment */
-	ori	r4, r0, TOPHYS(_fdt_start) /* save bram context */
+	ori	r4, r0, TOPHYS(_fdt_start)
 	ori	r3, r0, (0x4000 - 4)
 _copy_fdt:
 	lw	r12, r7, r11 /* r12 = r7 + r11 */

arch/microblaze/kernel/hw_exception_handler.S

@@ -74,6 +74,7 @@
 
 #include <asm/mmu.h>
 #include <asm/pgtable.h>
+#include <asm/signal.h>
 #include <asm/asm-offsets.h>
 
 /* Helpful Macros */
@@ -428,19 +429,9 @@ handle_unaligned_ex:
 	mfs	r17, rbtr;	/* ESR[DS] set - return address in BTR */
 	nop
 _no_delayslot:
-#endif
+	/* jump to high level unaligned handler */
-
-#ifdef CONFIG_MMU
-	/* Check if unaligned address is last on a 4k page */
-		andi	r5, r4, 0xffc
-		xori	r5, r5, 0xffc
-		bnei	r5, _unaligned_ex2
-	_unaligned_ex1:
 	RESTORE_STATE;
-/* Another page must be accessed or physical address not in page table */
 	bri	unaligned_data_trap
-
-	_unaligned_ex2:
 #endif
 	andi	r6, r3, 0x3E0; /* Mask and extract the register operand */
 	srl	r6, r6; /* r6 >> 5 */
@@ -450,45 +441,6 @@ _no_delayslot:
 	srl	r6, r6;
 	/* Store the register operand in a temporary location */
 	sbi	r6, r0, TOPHYS(ex_reg_op);
-#ifdef CONFIG_MMU
-	/* Get physical address */
-	/* If we are faulting a kernel address, we have to use the
-	 * kernel page tables.
-	 */
-	ori	r5, r0, CONFIG_KERNEL_START
-	cmpu	r5, r4, r5
-	bgti	r5, _unaligned_ex3
-	ori	r5, r0, swapper_pg_dir
-	bri	_unaligned_ex4
-
-	/* Get the PGD for the current thread. */
-_unaligned_ex3: /* user thread */
-	addi	r5 ,CURRENT_TASK, TOPHYS(0); /* get current task address */
-	lwi	r5, r5, TASK_THREAD + PGDIR
-_unaligned_ex4:
-	tophys(r5,r5)
-	BSRLI(r6,r4,20)			/* Create L1 (pgdir/pmd) address */
-	andi	r6, r6, 0xffc
-/* Assume pgdir aligned on 4K boundary, no need for "andi r5,r5,0xfffff003" */
-	or	r5, r5, r6
-	lwi	r6, r5, 0		/* Get L1 entry */
-	andi	r5, r6, 0xfffff000	/* Extract L2 (pte) base address. */
-	beqi	r5, _unaligned_ex1	/* Bail if no table */
-
-	tophys(r5,r5)
-	BSRLI(r6,r4,10)			/* Compute PTE address */
-	andi	r6, r6, 0xffc
-	andi	r5, r5, 0xfffff003
-	or	r5, r5, r6
-	lwi	r5, r5, 0		/* Get Linux PTE */
-
-	andi	r6, r5, _PAGE_PRESENT
-	beqi	r6, _unaligned_ex1	/* Bail if no page */
-
-	andi	r5, r5, 0xfffff000	/* Extract RPN */
-	andi	r4, r4, 0x00000fff	/* Extract offset */
-	or	r4, r4, r5		/* Create physical address */
-#endif /* CONFIG_MMU */
 
 	andi	r6, r3, 0x400; /* Extract ESR[S] */
 	bnei	r6, ex_sw;
@@ -959,15 +911,15 @@ _unaligned_data_exception:
 	andi	r6, r3, 0x800;	/* Extract ESR[W] - delay slot */
 ex_lw_vm:
 	beqid	r6, ex_lhw_vm;
-	lbui	r5, r4, 0;	/* Exception address in r4 - delay slot */
+load1:	lbui	r5, r4, 0;	/* Exception address in r4 - delay slot */
 /* Load a word, byte-by-byte from destination address and save it in tmp space*/
 	la	r6, r0, ex_tmp_data_loc_0;
 	sbi	r5, r6, 0;
-	lbui	r5, r4, 1;
+load2:	lbui	r5, r4, 1;
 	sbi	r5, r6, 1;
-	lbui	r5, r4, 2;
+load3:	lbui	r5, r4, 2;
 	sbi	r5, r6, 2;
-	lbui	r5, r4, 3;
+load4:	lbui	r5, r4, 3;
 	sbi	r5, r6, 3;
 	brid	ex_lw_tail_vm;
 /* Get the destination register value into r3 - delay slot */
@@ -977,7 +929,7 @@ ex_lhw_vm:
 	 * save it in tmp space */
 	la	r6, r0, ex_tmp_data_loc_0;
 	sbi	r5, r6, 0;
-	lbui	r5, r4, 1;
+load5:	lbui	r5, r4, 1;
 	sbi	r5, r6, 1;
 	lhui	r3, r6, 0;	/* Get the destination register value into r3 */
 ex_lw_tail_vm:
@@ -996,22 +948,53 @@ ex_sw_tail_vm:
 	swi	r3, r5, 0;	/* Get the word - delay slot */
 	/* Store the word, byte-by-byte into destination address */
 	lbui	r3, r5, 0;
-	sbi	r3, r4, 0;
+store1:	sbi	r3, r4, 0;
 	lbui	r3, r5, 1;
-	sbi	r3, r4, 1;
+store2:	sbi	r3, r4, 1;
 	lbui	r3, r5, 2;
-	sbi	r3, r4, 2;
+store3:	sbi	r3, r4, 2;
 	lbui	r3, r5, 3;
 	brid	ret_from_exc;
-	sbi	r3, r4, 3;	/* Delay slot */
+store4:	sbi	r3, r4, 3;	/* Delay slot */
 ex_shw_vm:
 	/* Store the lower half-word, byte-by-byte into destination address */
 	lbui	r3, r5, 2;
-	sbi	r3, r4, 0;
+store5:	sbi	r3, r4, 0;
 	lbui	r3, r5, 3;
 	brid	ret_from_exc;
-	sbi	r3, r4, 1;	/* Delay slot */
+store6:	sbi	r3, r4, 1;	/* Delay slot */
 ex_sw_end_vm:			/* Exception handling of store word, ends. */
+
+/* We have to prevent cases that get/put_user macros get unaligned pointer
+ * to bad page area. We have to find out which origin instruction caused it
+ * and called fixup for that origin instruction not instruction in unaligned
+ * handler */
+ex_unaligned_fixup:
+	ori	r5, r7, 0 /* setup pointer to pt_regs */
+	lwi	r6, r7, PT_PC; /* faulting address is one instruction above */
+	addik	r6, r6, -4 /* for finding proper fixup */
+	swi	r6, r7, PT_PC; /* a save back it to PT_PC */
+	addik	r7, r0, SIGSEGV
+	/* call bad_page_fault for finding aligned fixup, fixup address is saved
+	 * in PT_PC which is used as return address from exception */
+	la	r15, r0, ret_from_exc-8 /* setup return address */
+	brid	bad_page_fault
+	nop
+
+/* We prevent all load/store because it could failed any attempt to access */
+.section __ex_table,"a";
+	.word	load1,ex_unaligned_fixup;
+	.word	load2,ex_unaligned_fixup;
+	.word	load3,ex_unaligned_fixup;
+	.word	load4,ex_unaligned_fixup;
+	.word	load5,ex_unaligned_fixup;
+	.word	store1,ex_unaligned_fixup;
+	.word	store2,ex_unaligned_fixup;
+	.word	store3,ex_unaligned_fixup;
+	.word	store4,ex_unaligned_fixup;
+	.word	store5,ex_unaligned_fixup;
+	.word	store6,ex_unaligned_fixup;
+.previous;
 .end _unaligned_data_exception
 #endif /* CONFIG_MMU */
 

arch/microblaze/kernel/module.c

@@ -57,7 +57,6 @@ int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
 	Elf32_Rela *rela = (void *)sechdrs[relsec].sh_addr;
 	Elf32_Sym *sym;
 	unsigned long int *location;
-	unsigned long int locoffs;
 	unsigned long int value;
 #if __GNUC__ < 4
 	unsigned long int old_value;
@@ -113,10 +112,12 @@ int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
 			break;
 
 		case R_MICROBLAZE_64_PCREL:
-			locoffs = (location[0] & 0xFFFF) << 16 |
+#if __GNUC__ < 4
+			old_value = (location[0] & 0xFFFF) << 16 |
 				(location[1] & 0xFFFF);
-			value -= (unsigned long int)(location) + 4 +
+			value -= old_value;
-				locoffs;
+#endif
+			value -= (unsigned long int)(location) + 4;
 			location[0] = (location[0] & 0xFFFF0000) |
 					(value >> 16);
 			location[1] = (location[1] & 0xFFFF0000) |
@@ -125,6 +126,14 @@ int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
 				value);
 			break;
 
+		case R_MICROBLAZE_32_PCREL_LO:
+			pr_debug("R_MICROBLAZE_32_PCREL_LO\n");
+			break;
+
+		case R_MICROBLAZE_64_NONE:
+			pr_debug("R_MICROBLAZE_NONE\n");
+			break;
+
 		case R_MICROBLAZE_NONE:
 			pr_debug("R_MICROBLAZE_NONE\n");
 			break;
@@ -133,7 +142,7 @@ int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
 			printk(KERN_ERR "module %s: "
 				"Unknown relocation: %u\n",
 				module->name,
-				ELF32_R_TYPE(rela->r_info));
+				ELF32_R_TYPE(rela[i].r_info));
 			return -ENOEXEC;
 		}
 	}

arch/microblaze/kernel/setup.c

@@ -138,8 +138,12 @@ void __init machine_early_init(const char *cmdline, unsigned int ram,
 	setup_early_printk(NULL);
 #endif
 
-	early_printk("Ramdisk addr 0x%08x, FDT 0x%08x\n", ram, fdt);
+	early_printk("Ramdisk addr 0x%08x, ", ram);
-	printk(KERN_NOTICE "Found FDT at 0x%08x\n", fdt);
+	if (fdt)
+		early_printk("FDT at 0x%08x\n", fdt);
+	else
+		early_printk("Compiled-in FDT at 0x%08x\n",
+					(unsigned int)_fdt_start);
 
 #ifdef CONFIG_MTD_UCLINUX
 	early_printk("Found romfs @ 0x%08x (0x%08x)\n",

arch/microblaze/kernel/sys_microblaze.c

@@ -33,105 +33,6 @@
 #include <linux/unistd.h>
 
 #include <asm/syscalls.h>
-/*
- * sys_ipc() is the de-multiplexer for the SysV IPC calls..
- *
- * This is really horribly ugly. This will be remove with new toolchain.
- */
-asmlinkage long
-sys_ipc(uint call, int first, int second, int third, void *ptr, long fifth)
-{
-	int version, ret;
-
-	version = call >> 16; /* hack for backward compatibility */
-	call &= 0xffff;
-
-	ret = -EINVAL;
-	switch (call) {
-	case SEMOP:
-		ret = sys_semop(first, (struct sembuf *)ptr, second);
-		break;
-	case SEMGET:
-		ret = sys_semget(first, second, third);
-		break;
-	case SEMCTL:
-	{
-		union semun fourth;
-
-		if (!ptr)
-			break;
-		ret = (access_ok(VERIFY_READ, ptr, sizeof(long)) ? 0 : -EFAULT)
-				|| (get_user(fourth.__pad, (void **)ptr)) ;
-		if (ret)
-			break;
-		ret = sys_semctl(first, second, third, fourth);
-		break;
-	}
-	case MSGSND:
-		ret = sys_msgsnd(first, (struct msgbuf *) ptr, second, third);
-		break;
-	case MSGRCV:
-		switch (version) {
-		case 0: {
-			struct ipc_kludge tmp;
-
-			if (!ptr)
-				break;
-			ret = (access_ok(VERIFY_READ, ptr, sizeof(tmp))
-				? 0 : -EFAULT) || copy_from_user(&tmp,
-				(struct ipc_kludge *) ptr, sizeof(tmp));
-			if (ret)
-				break;
-			ret = sys_msgrcv(first, tmp.msgp, second, tmp.msgtyp,
-					third);
-			break;
-			}
-		default:
-			ret = sys_msgrcv(first, (struct msgbuf *) ptr,
-					second, fifth, third);
-			break;
-		}
-		break;
-	case MSGGET:
-		ret = sys_msgget((key_t) first, second);
-		break;
-	case MSGCTL:
-		ret = sys_msgctl(first, second, (struct msqid_ds *) ptr);
-		break;
-	case SHMAT:
-		switch (version) {
-		default: {
-			ulong raddr;
-			ret = access_ok(VERIFY_WRITE, (ulong *) third,
-					sizeof(ulong)) ? 0 : -EFAULT;
-			if (ret)
-				break;
-			ret = do_shmat(first, (char *) ptr, second, &raddr);
-			if (ret)
-				break;
-			ret = put_user(raddr, (ulong *) third);
-			break;
-			}
-		case 1:	/* iBCS2 emulator entry point */
-			if (!segment_eq(get_fs(), get_ds()))
-				break;
-			ret = do_shmat(first, (char *) ptr, second,
-					(ulong *) third);
-			break;
-		}
-		break;
-	case SHMDT:
-		ret = sys_shmdt((char *)ptr);
-		break;
-	case SHMGET:
-		ret = sys_shmget(first, second, third);
-		break;
-	case SHMCTL:
-		ret = sys_shmctl(first, second, (struct shmid_ds *) ptr);
-		break;
-	}
-	return ret;
-}
 
 asmlinkage long microblaze_vfork(struct pt_regs *regs)
 {

arch/microblaze/kernel/syscall_table.S

@@ -121,7 +121,7 @@ ENTRY(sys_call_table)
 	.long sys_wait4
 	.long sys_swapoff		/* 115 */
 	.long sys_sysinfo
-	.long sys_ipc
+	.long sys_ni_syscall		/* old sys_ipc */
 	.long sys_fsync
 	.long sys_ni_syscall		/* sys_sigreturn_wrapper */
 	.long sys_clone		/* 120 */

arch/microblaze/mm/fault.c

@@ -69,7 +69,7 @@ static int store_updates_sp(struct pt_regs *regs)
  * It is called from do_page_fault above and from some of the procedures
  * in traps.c.
  */
-static void bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
+void bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
 {
 	const struct exception_table_entry *fixup;
 /* MS: no context */
@@ -122,15 +122,10 @@ void do_page_fault(struct pt_regs *regs, unsigned long address,
 	}
 #endif /* CONFIG_KGDB */
 
-	if (in_atomic() || mm == NULL) {
+	if (in_atomic() || !mm) {
-		/* FIXME */
+		if (kernel_mode(regs))
-		if (kernel_mode(regs)) {
+			goto bad_area_nosemaphore;
-			printk(KERN_EMERG
+
-				"Page fault in kernel mode - Oooou!!! pid %d\n",
-				current->pid);
-			_exception(SIGSEGV, regs, code, address);
-			return;
-		}
 		/* in_atomic() in user mode is really bad,
 		   as is current->mm == NULL. */
 		printk(KERN_EMERG "Page fault in user mode with "

arch/mips/include/asm/pgalloc.h

@@ -98,23 +98,12 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	__free_pages(pte, PTE_ORDER);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,address)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb), pte);			\
 } while (0)
 
-#ifdef CONFIG_32BIT
-
-/*
- * allocating and freeing a pmd is trivial: the 1-entry pmd is
- * inside the pgd, so has no extra memory associated with it.
- */
-#define pmd_free(mm, x)			do { } while (0)
-#define __pmd_free_tlb(tlb, x)		do { } while (0)
-
-#endif
-
 #ifdef CONFIG_64BIT
 
 static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address)
@@ -132,7 +121,7 @@ static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
 	free_pages((unsigned long)pmd, PMD_ORDER);
 }
 
-#define __pmd_free_tlb(tlb, x)	pmd_free((tlb)->mm, x)
+#define __pmd_free_tlb(tlb, x, addr)	pmd_free((tlb)->mm, x)
 
 #endif
 

arch/mn10300/include/asm/pgalloc.h

@@ -51,6 +51,6 @@ static inline void pte_free(struct mm_struct *mm, struct page *pte)
 }
 
 
-#define __pte_free_tlb(tlb, pte) tlb_remove_page((tlb), (pte))
+#define __pte_free_tlb(tlb, pte, addr) tlb_remove_page((tlb), (pte))
 
 #endif /* _ASM_PGALLOC_H */

arch/parisc/include/asm/tlb.h

@@ -21,7 +21,7 @@ do {	if (!(tlb)->fullmm)	\
 
 #include <asm-generic/tlb.h>
 
-#define __pmd_free_tlb(tlb, pmd)	pmd_free((tlb)->mm, pmd)
+#define __pmd_free_tlb(tlb, pmd, addr)	pmd_free((tlb)->mm, pmd)
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, pte)
+#define __pte_free_tlb(tlb, pte, addr)	pte_free((tlb)->mm, pte)
 
 #endif

arch/powerpc/include/asm/pgalloc-32.h

@@ -16,7 +16,7 @@ extern void pgd_free(struct mm_struct *mm, pgd_t *pgd);
  */
 /* #define pmd_alloc_one(mm,address)       ({ BUG(); ((pmd_t *)2); }) */
 #define pmd_free(mm, x) 		do { } while (0)
-#define __pmd_free_tlb(tlb,x)		do { } while (0)
+#define __pmd_free_tlb(tlb,x,a)		do { } while (0)
 /* #define pgd_populate(mm, pmd, pte)      BUG() */
 
 #ifndef CONFIG_BOOKE

arch/powerpc/include/asm/pgalloc-64.h

@@ -118,11 +118,11 @@ static inline void pgtable_free(pgtable_free_t pgf)
 		kmem_cache_free(pgtable_cache[cachenum], p);
 }
 
-#define __pmd_free_tlb(tlb, pmd) 	\
+#define __pmd_free_tlb(tlb, pmd,addr)		      \
 	pgtable_free_tlb(tlb, pgtable_free_cache(pmd, \
 		PMD_CACHE_NUM, PMD_TABLE_SIZE-1))
 #ifndef CONFIG_PPC_64K_PAGES
-#define __pud_free_tlb(tlb, pud)	\
+#define __pud_free_tlb(tlb, pud, addr)		      \
 	pgtable_free_tlb(tlb, pgtable_free_cache(pud, \
 		PUD_CACHE_NUM, PUD_TABLE_SIZE-1))
 #endif /* CONFIG_PPC_64K_PAGES */

arch/powerpc/include/asm/pgalloc.h

@@ -38,14 +38,14 @@ static inline pgtable_free_t pgtable_free_cache(void *p, int cachenum,
 extern void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf);
 
 #ifdef CONFIG_SMP
-#define __pte_free_tlb(tlb,ptepage)	\
+#define __pte_free_tlb(tlb,ptepage,address)		\
 do { \
 	pgtable_page_dtor(ptepage); \
 	pgtable_free_tlb(tlb, pgtable_free_cache(page_address(ptepage), \
 					PTE_NONCACHE_NUM, PTE_TABLE_SIZE-1)); \
 } while (0)
 #else
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, (pte))
+#define __pte_free_tlb(tlb, pte, address)	pte_free((tlb)->mm, (pte))
 #endif
 
 

arch/powerpc/mm/hugetlbpage.c

@@ -305,7 +305,7 @@ static void hugetlb_free_pmd_range(struct mmu_gather *tlb, pud_t *pud,
 
 	pmd = pmd_offset(pud, start);
 	pud_clear(pud);
-	pmd_free_tlb(tlb, pmd);
+	pmd_free_tlb(tlb, pmd, start);
 }
 
 static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd,
@@ -348,7 +348,7 @@ static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd,
 
 	pud = pud_offset(pgd, start);
 	pgd_clear(pgd);
-	pud_free_tlb(tlb, pud);
+	pud_free_tlb(tlb, pud, start);
 }
 
 /*

arch/s390/include/asm/tlb.h

@@ -96,7 +96,8 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
  * pte_free_tlb frees a pte table and clears the CRSTE for the
  * page table from the tlb.
  */
-static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte)
+static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte,
+				unsigned long address)
 {
 	if (!tlb->fullmm) {
 		tlb->array[tlb->nr_ptes++] = pte;
@@ -113,7 +114,8 @@ static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte)
  * as the pgd. pmd_free_tlb checks the asce_limit against 2GB
  * to avoid the double free of the pmd in this case.
  */
-static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
+static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd,
+				unsigned long address)
 {
 #ifdef __s390x__
 	if (tlb->mm->context.asce_limit <= (1UL << 31))
@@ -134,7 +136,8 @@ static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
  * as the pgd. pud_free_tlb checks the asce_limit against 4TB
  * to avoid the double free of the pud in this case.
  */
-static inline void pud_free_tlb(struct mmu_gather *tlb, pud_t *pud)
+static inline void pud_free_tlb(struct mmu_gather *tlb, pud_t *pud,
+				unsigned long address)
 {
 #ifdef __s390x__
 	if (tlb->mm->context.asce_limit <= (1UL << 42))

arch/s390/kernel/early.c

@@ -210,7 +210,7 @@ static noinline __init void detect_machine_type(void)
 		machine_flags |= MACHINE_FLAG_VM;
 }
 
-static void early_pgm_check_handler(void)
+static __init void early_pgm_check_handler(void)
 {
 	unsigned long addr;
 	const struct exception_table_entry *fixup;
@@ -222,7 +222,7 @@ static void early_pgm_check_handler(void)
 	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
 }
 
-void setup_lowcore_early(void)
+static noinline __init void setup_lowcore_early(void)
 {
 	psw_t psw;
 

arch/s390/kernel/smp.c

@@ -687,13 +687,14 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 #ifndef CONFIG_64BIT
 	if (MACHINE_HAS_IEEE)
 		lowcore->extended_save_area_addr = (u32) save_area;
-#else
-	if (vdso_alloc_per_cpu(smp_processor_id(), lowcore))
-		BUG();
 #endif
 	set_prefix((u32)(unsigned long) lowcore);
 	local_mcck_enable();
 	local_irq_enable();
+#ifdef CONFIG_64BIT
+	if (vdso_alloc_per_cpu(smp_processor_id(), &S390_lowcore))
+		BUG();
+#endif
 	for_each_possible_cpu(cpu)
 		if (cpu != smp_processor_id())
 			smp_create_idle(cpu);

arch/s390/kernel/vdso64/clock_gettime.S

@@ -88,10 +88,17 @@ __kernel_clock_gettime:
 	llilh	%r4,0x0100
 	sar	%a4,%r4
 	lghi	%r4,0
+	epsw	%r5,0
 	sacf	512				/* Magic ectg instruction */
 	.insn	ssf,0xc80100000000,__VDSO_ECTG_BASE(4),__VDSO_ECTG_USER(4),4
-	sacf	0
+	tml	%r5,0x4000
-	sar	%a4,%r2
+	jo	11f
+	tml	%r5,0x8000
+	jno	10f
+	sacf	256
+	j	11f
+10:	sacf	0
+11:	sar	%a4,%r2
 	algr	%r1,%r0				/* r1 = cputime as TOD value */
 	mghi	%r1,1000			/* convert to nanoseconds */
 	srlg	%r1,%r1,12			/* r1 = cputime in nanosec */

arch/s390/power/swsusp.c

@@ -7,24 +7,36 @@
  *
  */
 
+#include <asm/system.h>
 
-/*
- * save CPU registers before creating a hibernation image and before
- * restoring the memory state from it
- */
 void save_processor_state(void)
 {
-	/* implentation contained in the
+	/* swsusp_arch_suspend() actually saves all cpu register contents.
-	 * swsusp_arch_suspend function
+	 * Machine checks must be disabled since swsusp_arch_suspend() stores
+	 * register contents to their lowcore save areas. That's the same
+	 * place where register contents on machine checks would be saved.
+	 * To avoid register corruption disable machine checks.
+	 * We must also disable machine checks in the new psw mask for
+	 * program checks, since swsusp_arch_suspend() may generate program
+	 * checks. Disabling machine checks for all other new psw masks is
+	 * just paranoia.
 	 */
+	local_mcck_disable();
+	/* Disable lowcore protection */
+	__ctl_clear_bit(0,28);
+	S390_lowcore.external_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.svc_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.io_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask &= ~PSW_MASK_MCHECK;
 }
 
-/*
- * restore the contents of CPU registers
- */
 void restore_processor_state(void)
 {
-	/* implentation contained in the
+	S390_lowcore.external_new_psw.mask |= PSW_MASK_MCHECK;
-	 * swsusp_arch_resume function
+	S390_lowcore.svc_new_psw.mask |= PSW_MASK_MCHECK;
-	 */
+	S390_lowcore.io_new_psw.mask |= PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask |= PSW_MASK_MCHECK;
+	/* Enable lowcore protection */
+	__ctl_set_bit(0,28);
+	local_mcck_enable();
 }

arch/s390/power/swsusp_asm64.S

@@ -32,19 +32,14 @@ swsusp_arch_suspend:
 	/* Deactivate DAT */
 	stnsm	__SF_EMPTY(%r15),0xfb
 
-	/* Switch off lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	ni	__SF_EMPTY+4(%r15),0xef
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Store prefix register on stack */
 	stpx	__SF_EMPTY(%r15)
 
-	/* Setup base register for lowcore (absolute 0) */
+	/* Save prefix register contents for lowcore */
-	llgf	%r1,__SF_EMPTY(%r15)
+	llgf	%r4,__SF_EMPTY(%r15)
 
 	/* Get pointer to save area */
-	aghi	%r1,0x1000
+	lghi	%r1,0x1000
 
 	/* Store registers */
 	mvc	0x318(4,%r1),__SF_EMPTY(%r15)	/* move prefix to lowcore */
@@ -79,17 +74,15 @@ swsusp_arch_suspend:
 	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
 	spx	__SF_EMPTY(%r15)
 
-	/* Setup lowcore */
+	lghi	%r2,0
-	brasl	%r14,setup_lowcore_early
+	lghi	%r3,2*PAGE_SIZE
+	lghi	%r5,2*PAGE_SIZE
+1:	mvcle	%r2,%r4,0
+	jo	1b
 
 	/* Save image */
 	brasl	%r14,swsusp_save
 
-	/* Switch on lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	oi	__SF_EMPTY+4(%r15),0x10
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Restore prefix register and return */
 	lghi	%r1,0x1000
 	spx	0x318(%r1)
@@ -117,11 +110,6 @@ swsusp_arch_resume:
 	/* Deactivate DAT */
 	stnsm	__SF_EMPTY(%r15),0xfb
 
-	/* Switch off lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	ni	__SF_EMPTY+4(%r15),0xef
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Set prefix page to zero */
 	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
 	spx	__SF_EMPTY(%r15)
@@ -175,7 +163,7 @@ swsusp_arch_resume:
 	/* Load old stack */
 	lg	%r15,0x2f8(%r13)
 
-	/* Pointer to save arae */
+	/* Pointer to save area */
 	lghi	%r13,0x1000
 
 #ifdef CONFIG_SMP
@@ -187,11 +175,6 @@ swsusp_arch_resume:
 	/* Restore prefix register */
 	spx	0x318(%r13)
 
-	/* Switch on lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	oi	__SF_EMPTY+4(%r15),0x10
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Activate DAT */
 	stosm	__SF_EMPTY(%r15),0x04
 

arch/sh/include/asm/pgalloc.h

@@ -73,20 +73,12 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	quicklist_free_page(QUICK_PT, NULL, pte);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte,addr)			\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb), (pte));			\
 } while (0)
 
-/*
- * allocating and freeing a pmd is trivial: the 1-entry pmd is
- * inside the pgd, so has no extra memory associated with it.
- */
-
-#define pmd_free(mm, x)			do { } while (0)
-#define __pmd_free_tlb(tlb,x)		do { } while (0)
-
 static inline void check_pgt_cache(void)
 {
 	quicklist_trim(QUICK_PGD, NULL, 25, 16);

arch/sh/include/asm/tlb.h

@@ -91,9 +91,9 @@ tlb_end_vma(struct mmu_gather *tlb, struct vm_area_struct *vma)
 }
 
 #define tlb_remove_page(tlb,page)	free_page_and_swap_cache(page)
-#define pte_free_tlb(tlb, ptep)		pte_free((tlb)->mm, ptep)
+#define pte_free_tlb(tlb, ptep, addr)	pte_free((tlb)->mm, ptep)
-#define pmd_free_tlb(tlb, pmdp)		pmd_free((tlb)->mm, pmdp)
+#define pmd_free_tlb(tlb, pmdp, addr)	pmd_free((tlb)->mm, pmdp)
-#define pud_free_tlb(tlb, pudp)		pud_free((tlb)->mm, pudp)
+#define pud_free_tlb(tlb, pudp, addr)	pud_free((tlb)->mm, pudp)
 
 #define tlb_migrate_finish(mm)		do { } while (0)
 

arch/sparc/include/asm/pgalloc_32.h

@@ -45,7 +45,7 @@ BTFIXUPDEF_CALL(void, free_pmd_fast, pmd_t *)
 #define free_pmd_fast(pmd)	BTFIXUP_CALL(free_pmd_fast)(pmd)
 
 #define pmd_free(mm, pmd)		free_pmd_fast(pmd)
-#define __pmd_free_tlb(tlb, pmd) pmd_free((tlb)->mm, pmd)
+#define __pmd_free_tlb(tlb, pmd, addr)	pmd_free((tlb)->mm, pmd)
 
 BTFIXUPDEF_CALL(void, pmd_populate, pmd_t *, struct page *)
 #define pmd_populate(MM, PMD, PTE)        BTFIXUP_CALL(pmd_populate)(PMD, PTE)
@@ -63,6 +63,6 @@ BTFIXUPDEF_CALL(void, free_pte_fast, pte_t *)
 
 BTFIXUPDEF_CALL(void, pte_free, pgtable_t )
 #define pte_free(mm, pte)		BTFIXUP_CALL(pte_free)(pte)
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, pte)
+#define __pte_free_tlb(tlb, pte, addr)	pte_free((tlb)->mm, pte)
 
 #endif /* _SPARC_PGALLOC_H */

arch/sparc/include/asm/tlb_64.h

@@ -100,9 +100,9 @@ static inline void tlb_remove_page(struct mmu_gather *mp, struct page *page)
 }
 
 #define tlb_remove_tlb_entry(mp,ptep,addr) do { } while (0)
-#define pte_free_tlb(mp, ptepage) pte_free((mp)->mm, ptepage)
+#define pte_free_tlb(mp, ptepage, addr) pte_free((mp)->mm, ptepage)
-#define pmd_free_tlb(mp, pmdp) pmd_free((mp)->mm, pmdp)
+#define pmd_free_tlb(mp, pmdp, addr) pmd_free((mp)->mm, pmdp)
-#define pud_free_tlb(tlb,pudp) __pud_free_tlb(tlb,pudp)
+#define pud_free_tlb(tlb,pudp, addr) __pud_free_tlb(tlb,pudp,addr)
 
 #define tlb_migrate_finish(mm)	do { } while (0)
 #define tlb_start_vma(tlb, vma) do { } while (0)

arch/um/include/asm/pgalloc.h

@@ -40,7 +40,7 @@ static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
 	__free_page(pte);
 }
 
-#define __pte_free_tlb(tlb,pte)				\
+#define __pte_free_tlb(tlb,pte, address)		\
 do {							\
 	pgtable_page_dtor(pte);				\
 	tlb_remove_page((tlb),(pte));			\
@@ -53,7 +53,7 @@ static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
 	free_page((unsigned long)pmd);
 }
 
-#define __pmd_free_tlb(tlb,x)   tlb_remove_page((tlb),virt_to_page(x))
+#define __pmd_free_tlb(tlb,x, address)   tlb_remove_page((tlb),virt_to_page(x))
 #endif
 
 #define check_pgt_cache()	do { } while (0)

arch/um/include/asm/tlb.h

@@ -116,11 +116,11 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
 		__tlb_remove_tlb_entry(tlb, ptep, address);	\
 	} while (0)
 
-#define pte_free_tlb(tlb, ptep) __pte_free_tlb(tlb, ptep)
+#define pte_free_tlb(tlb, ptep, addr) __pte_free_tlb(tlb, ptep, addr)
 
-#define pud_free_tlb(tlb, pudp) __pud_free_tlb(tlb, pudp)
+#define pud_free_tlb(tlb, pudp, addr) __pud_free_tlb(tlb, pudp, addr)
 
-#define pmd_free_tlb(tlb, pmdp) __pmd_free_tlb(tlb, pmdp)
+#define pmd_free_tlb(tlb, pmdp, addr) __pmd_free_tlb(tlb, pmdp, addr)
 
 #define tlb_migrate_finish(mm) do {} while (0)
 

arch/x86/include/asm/pgalloc.h

@@ -46,7 +46,13 @@ static inline void pte_free(struct mm_struct *mm, struct page *pte)
 	__free_page(pte);
 }
 
-extern void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte);
+extern void ___pte_free_tlb(struct mmu_gather *tlb, struct page *pte);
+
+static inline void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte,
+				  unsigned long address)
+{
+	___pte_free_tlb(tlb, pte);
+}
 
 static inline void pmd_populate_kernel(struct mm_struct *mm,
 				       pmd_t *pmd, pte_t *pte)
@@ -78,7 +84,13 @@ static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
 	free_page((unsigned long)pmd);
 }
 
-extern void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd);
+extern void ___pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd);
+
+static inline void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd,
+				  unsigned long adddress)
+{
+	___pmd_free_tlb(tlb, pmd);
+}
 
 #ifdef CONFIG_X86_PAE
 extern void pud_populate(struct mm_struct *mm, pud_t *pudp, pmd_t *pmd);
@@ -108,7 +120,14 @@ static inline void pud_free(struct mm_struct *mm, pud_t *pud)
 	free_page((unsigned long)pud);
 }
 
-extern void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud);
+extern void ___pud_free_tlb(struct mmu_gather *tlb, pud_t *pud);
+
+static inline void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud,
+				  unsigned long address)
+{
+	___pud_free_tlb(tlb, pud);
+}
+
 #endif	/* PAGETABLE_LEVELS > 3 */
 #endif	/* PAGETABLE_LEVELS > 2 */
 

arch/x86/include/asm/uaccess.h

@@ -212,9 +212,9 @@ extern int __get_user_bad(void);
 		     : "A" ((typeof(*(ptr)))(x)), "c" (ptr) : "ebx")
 #else
 #define __put_user_asm_u64(x, ptr, retval, errret) \
-	__put_user_asm(x, ptr, retval, "q", "", "Zr", errret)
+	__put_user_asm(x, ptr, retval, "q", "", "er", errret)
 #define __put_user_asm_ex_u64(x, addr)	\
-	__put_user_asm_ex(x, addr, "q", "", "Zr")
+	__put_user_asm_ex(x, addr, "q", "", "er")
 #define __put_user_x8(x, ptr, __ret_pu) __put_user_x(8, x, ptr, __ret_pu)
 #endif
 

arch/x86/include/asm/uaccess_64.h

@@ -88,11 +88,11 @@ int __copy_to_user(void __user *dst, const void *src, unsigned size)
 			      ret, "l", "k", "ir", 4);
 		return ret;
 	case 8:__put_user_asm(*(u64 *)src, (u64 __user *)dst,
-			      ret, "q", "", "ir", 8);
+			      ret, "q", "", "er", 8);
 		return ret;
 	case 10:
 		__put_user_asm(*(u64 *)src, (u64 __user *)dst,
-			       ret, "q", "", "ir", 10);
+			       ret, "q", "", "er", 10);
 		if (unlikely(ret))
 			return ret;
 		asm("":::"memory");
@@ -101,12 +101,12 @@ int __copy_to_user(void __user *dst, const void *src, unsigned size)
 		return ret;
 	case 16:
 		__put_user_asm(*(u64 *)src, (u64 __user *)dst,
-			       ret, "q", "", "ir", 16);
+			       ret, "q", "", "er", 16);
 		if (unlikely(ret))
 			return ret;
 		asm("":::"memory");
 		__put_user_asm(1[(u64 *)src], 1 + (u64 __user *)dst,
-			       ret, "q", "", "ir", 8);
+			       ret, "q", "", "er", 8);
 		return ret;
 	default:
 		return copy_user_generic((__force void *)dst, src, size);
@@ -157,7 +157,7 @@ int __copy_in_user(void __user *dst, const void __user *src, unsigned size)
 			       ret, "q", "", "=r", 8);
 		if (likely(!ret))
 			__put_user_asm(tmp, (u64 __user *)dst,
-				       ret, "q", "", "ir", 8);
+				       ret, "q", "", "er", 8);
 		return ret;
 	}
 	default:

arch/x86/kernel/cpu/amd.c

@@ -356,7 +356,7 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c)
 #endif
 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_PCI)
 	/* check CPU config space for extended APIC ID */
-	if (c->x86 >= 0xf) {
+	if (cpu_has_apic && c->x86 >= 0xf) {
 		unsigned int val;
 		val = read_pci_config(0, 24, 0, 0x68);
 		if ((val & ((1 << 17) | (1 << 18))) == ((1 << 17) | (1 << 18)))

arch/x86/kernel/cpu/mcheck/mce.c

@@ -1692,17 +1692,15 @@ static ssize_t set_trigger(struct sys_device *s, struct sysdev_attribute *attr,
 				const char *buf, size_t siz)
 {
 	char *p;
-	int len;
 
 	strncpy(mce_helper, buf, sizeof(mce_helper));
 	mce_helper[sizeof(mce_helper)-1] = 0;
-	len = strlen(mce_helper);
 	p = strchr(mce_helper, '\n');
 
-	if (*p)
+	if (p)
 		*p = 0;
 
-	return len;
+	return strlen(mce_helper) + !!p;
 }
 
 static ssize_t set_ignore_ce(struct sys_device *s,

arch/x86/kernel/irqinit.c

@@ -187,7 +187,7 @@ static void __init apic_intr_init(void)
 #ifdef CONFIG_X86_THERMAL_VECTOR
 	alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt);
 #endif
-#ifdef CONFIG_X86_THRESHOLD
+#ifdef CONFIG_X86_MCE_THRESHOLD
 	alloc_intr_gate(THRESHOLD_APIC_VECTOR, threshold_interrupt);
 #endif
 #if defined(CONFIG_X86_NEW_MCE) && defined(CONFIG_X86_LOCAL_APIC)

arch/x86/kernel/mfgpt_32.c

@@ -347,7 +347,7 @@ static irqreturn_t mfgpt_tick(int irq, void *dev_id)
 
 static struct irqaction mfgptirq  = {
 	.handler = mfgpt_tick,
-	.flags = IRQF_DISABLED | IRQF_NOBALANCING,
+	.flags = IRQF_DISABLED | IRQF_NOBALANCING | IRQF_TIMER,
 	.name = "mfgpt-timer"
 };
 

arch/x86/kernel/reboot.c

@@ -249,6 +249,14 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = {
 			DMI_MATCH(DMI_PRODUCT_NAME, "VGN-Z540N"),
 		},
 	},
+	{	/* Handle problems with rebooting on CompuLab SBC-FITPC2 */
+		.callback = set_bios_reboot,
+		.ident = "CompuLab SBC-FITPC2",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "CompuLab"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "SBC-FITPC2"),
+		},
+	},
 	{ }
 };
 

arch/x86/kernel/setup.c

@@ -672,6 +672,19 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = {
 			DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"),
 		},
 	},
+	{
+	/*
+	 * AMI BIOS with low memory corruption was found on Intel DG45ID board.
+	 * It hase different DMI_BIOS_VENDOR = "Intel Corp.", for now we will
+	 * match only DMI_BOARD_NAME and see if there is more bad products
+	 * with this vendor.
+	 */
+		.callback = dmi_low_memory_corruption,
+		.ident = "AMI BIOS",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_NAME, "DG45ID"),
+		},
+	},
 #endif
 	{}
 };

arch/x86/kernel/vmlinux.lds.S

@@ -112,11 +112,6 @@ SECTIONS
 		_sdata = .;
 		DATA_DATA
 		CONSTRUCTORS
-
-#ifdef CONFIG_X86_64
-		/* End of data section */
-		_edata = .;
-#endif
 	} :data
 
 #ifdef CONFIG_X86_32
@@ -156,10 +151,8 @@ SECTIONS
 	.data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
 		*(.data.read_mostly)
 
-#ifdef CONFIG_X86_32
 		/* End of data section */
 		_edata = .;
-#endif
 	}
 
 #ifdef CONFIG_X86_64

arch/x86/mm/highmem_32.c

@@ -103,6 +103,7 @@ EXPORT_SYMBOL(kmap);
 EXPORT_SYMBOL(kunmap);
 EXPORT_SYMBOL(kmap_atomic);
 EXPORT_SYMBOL(kunmap_atomic);
+EXPORT_SYMBOL(kmap_atomic_prot);
 
 void __init set_highmem_pages_init(void)
 {

arch/x86/mm/pgtable.c

@@ -25,7 +25,7 @@ pgtable_t pte_alloc_one(struct mm_struct *mm, unsigned long address)
 	return pte;
 }
 
-void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte)
+void ___pte_free_tlb(struct mmu_gather *tlb, struct page *pte)
 {
 	pgtable_page_dtor(pte);
 	paravirt_release_pte(page_to_pfn(pte));
@@ -33,14 +33,14 @@ void __pte_free_tlb(struct mmu_gather *tlb, struct page *pte)
 }
 
 #if PAGETABLE_LEVELS > 2
-void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
+void ___pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
 {
 	paravirt_release_pmd(__pa(pmd) >> PAGE_SHIFT);
 	tlb_remove_page(tlb, virt_to_page(pmd));
 }
 
 #if PAGETABLE_LEVELS > 3
-void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud)
+void ___pud_free_tlb(struct mmu_gather *tlb, pud_t *pud)
 {
 	paravirt_release_pud(__pa(pud) >> PAGE_SHIFT);
 	tlb_remove_page(tlb, virt_to_page(pud));

arch/x86/mm/srat_64.c

@@ -79,8 +79,10 @@ static __init void bad_srat(void)
 	acpi_numa = -1;
 	for (i = 0; i < MAX_LOCAL_APIC; i++)
 		apicid_to_node[i] = NUMA_NO_NODE;
-	for (i = 0; i < MAX_NUMNODES; i++)
+	for (i = 0; i < MAX_NUMNODES; i++) {
-		nodes_add[i].start = nodes[i].end = 0;
+		nodes[i].start = nodes[i].end = 0;
+		nodes_add[i].start = nodes_add[i].end = 0;
+	}
 	remove_all_active_ranges();
 }
 

arch/xtensa/include/asm/tlb.h

@@ -42,6 +42,6 @@
 
 #include <asm-generic/tlb.h>
 
-#define __pte_free_tlb(tlb, pte)		pte_free((tlb)->mm, pte)
+#define __pte_free_tlb(tlb, pte, address)	pte_free((tlb)->mm, pte)
 
 #endif	/* _XTENSA_TLB_H */

drivers/ata/ahci.c

@@ -515,10 +515,14 @@ static const struct pci_device_id ahci_pci_tbl[] = {
 	{ PCI_VDEVICE(INTEL, 0x3a05), board_ahci }, /* ICH10 */
 	{ PCI_VDEVICE(INTEL, 0x3a22), board_ahci }, /* ICH10 */
 	{ PCI_VDEVICE(INTEL, 0x3a25), board_ahci }, /* ICH10 */
+	{ PCI_VDEVICE(INTEL, 0x3b22), board_ahci }, /* PCH AHCI */
+	{ PCI_VDEVICE(INTEL, 0x3b23), board_ahci }, /* PCH AHCI */
 	{ PCI_VDEVICE(INTEL, 0x3b24), board_ahci }, /* PCH RAID */
 	{ PCI_VDEVICE(INTEL, 0x3b25), board_ahci }, /* PCH RAID */
+	{ PCI_VDEVICE(INTEL, 0x3b29), board_ahci }, /* PCH AHCI */
 	{ PCI_VDEVICE(INTEL, 0x3b2b), board_ahci }, /* PCH RAID */
 	{ PCI_VDEVICE(INTEL, 0x3b2c), board_ahci }, /* PCH RAID */
+	{ PCI_VDEVICE(INTEL, 0x3b2f), board_ahci }, /* PCH AHCI */
 
 	/* JMicron 360/1/3/5/6, match class to avoid IDE function */
 	{ PCI_VENDOR_ID_JMICRON, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID,

drivers/ata/ata_piix.c

@@ -596,9 +596,12 @@ static const struct ich_laptop ich_laptop[] = {
 	{ 0x27DF, 0x0005, 0x0280 },	/* ICH7 on Acer 5602WLMi */
 	{ 0x27DF, 0x1025, 0x0102 },	/* ICH7 on Acer 5602aWLMi */
 	{ 0x27DF, 0x1025, 0x0110 },	/* ICH7 on Acer 3682WLMi */
+	{ 0x27DF, 0x1028, 0x02b0 },	/* ICH7 on unknown Dell */
 	{ 0x27DF, 0x1043, 0x1267 },	/* ICH7 on Asus W5F */
 	{ 0x27DF, 0x103C, 0x30A1 },	/* ICH7 on HP Compaq nc2400 */
+	{ 0x27DF, 0x103C, 0x361a },	/* ICH7 on unkown HP  */
 	{ 0x27DF, 0x1071, 0xD221 },	/* ICH7 on Hercules EC-900 */
+	{ 0x27DF, 0x152D, 0x0778 },	/* ICH7 on unknown Intel */
 	{ 0x24CA, 0x1025, 0x0061 },	/* ICH4 on ACER Aspire 2023WLMi */
 	{ 0x24CA, 0x1025, 0x003d },	/* ICH4 on ACER TM290 */
 	{ 0x266F, 0x1025, 0x0066 },	/* ICH6 on ACER Aspire 1694WLMi */

drivers/ata/libata-core.c

@@ -1515,6 +1515,7 @@ static int ata_hpa_resize(struct ata_device *dev)
 
 		return rc;
 	}
+	dev->n_native_sectors = native_sectors;
 
 	/* nothing to do? */
 	if (native_sectors <= sectors || !ata_ignore_hpa) {
@@ -4099,6 +4100,7 @@ int ata_dev_revalidate(struct ata_device *dev, unsigned int new_class,
 		       unsigned int readid_flags)
 {
 	u64 n_sectors = dev->n_sectors;
+	u64 n_native_sectors = dev->n_native_sectors;
 	int rc;
 
 	if (!ata_dev_enabled(dev))
@@ -4128,17 +4130,31 @@ int ata_dev_revalidate(struct ata_device *dev, unsigned int new_class,
 	/* verify n_sectors hasn't changed */
 	if (dev->class == ATA_DEV_ATA && n_sectors &&
 	    dev->n_sectors != n_sectors) {
-		ata_dev_printk(dev, KERN_INFO, "n_sectors mismatch "
+		ata_dev_printk(dev, KERN_WARNING, "n_sectors mismatch "
 			       "%llu != %llu\n",
 			       (unsigned long long)n_sectors,
 			       (unsigned long long)dev->n_sectors);
-
+		/*
-		/* restore original n_sectors */
+		 * Something could have caused HPA to be unlocked
+		 * involuntarily.  If n_native_sectors hasn't changed
+		 * and the new size matches it, keep the device.
+		 */
+		if (dev->n_native_sectors == n_native_sectors &&
+		    dev->n_sectors > n_sectors &&
+		    dev->n_sectors == n_native_sectors) {
+			ata_dev_printk(dev, KERN_WARNING,
+				       "new n_sectors matches native, probably "
+				       "late HPA unlock, continuing\n");
+			/* keep using the old n_sectors */
+			dev->n_sectors = n_sectors;
+		} else {
+			/* restore original n_[native]_sectors and fail */
+			dev->n_native_sectors = n_native_sectors;
 			dev->n_sectors = n_sectors;
-
 			rc = -ENODEV;
 			goto fail;
 		}
+	}
 
 	return 0;
 

drivers/ata/libata-eh.c

@@ -2327,7 +2327,7 @@ int ata_eh_reset(struct ata_link *link, int classify,
 	struct ata_port *ap = link->ap;
 	struct ata_link *slave = ap->slave_link;
 	struct ata_eh_context *ehc = &link->eh_context;
-	struct ata_eh_context *sehc = &slave->eh_context;
+	struct ata_eh_context *sehc = slave ? &slave->eh_context : NULL;
 	unsigned int *classes = ehc->classes;
 	unsigned int lflags = link->flags;
 	int verbose = !(ehc->i.flags & ATA_EHI_QUIET);

drivers/ata/pata_at91.c

@@ -26,9 +26,7 @@
 #include <linux/platform_device.h>
 #include <linux/ata_platform.h>
 
-#include <mach/at91sam9260_matrix.h>
 #include <mach/at91sam9_smc.h>
-#include <mach/at91sam9260.h>
 #include <mach/board.h>
 #include <mach/gpio.h>
 
@@ -44,14 +42,16 @@ struct at91_ide_info {
 	unsigned long mode;
 	unsigned int cs;
 
+	struct clk *mck;
+
 	void __iomem *ide_addr;
 	void __iomem *alt_addr;
 };
 
-const struct ata_timing initial_timing =
+static const struct ata_timing initial_timing =
 	{XFER_PIO_0, 70, 290, 240, 600, 165, 150, 600, 0};
 
-static unsigned int calc_mck_cycles(unsigned int ns, unsigned int mck_hz)
+static unsigned long calc_mck_cycles(unsigned long ns, unsigned long mck_hz)
 {
 	unsigned long mul;
 
@@ -77,32 +77,27 @@ static void set_smc_mode(struct at91_ide_info *info)
 static void set_smc_timing(struct device *dev,
 		struct at91_ide_info *info, const struct ata_timing *ata)
 {
-	int read_cycle, write_cycle, active, recover;
+	unsigned long read_cycle, write_cycle, active, recover;
-	int nrd_setup, nrd_pulse, nrd_recover;
+	unsigned long nrd_setup, nrd_pulse, nrd_recover;
-	int nwe_setup, nwe_pulse;
+	unsigned long nwe_setup, nwe_pulse;
 
-	int ncs_write_setup, ncs_write_pulse;
+	unsigned long ncs_write_setup, ncs_write_pulse;
-	int ncs_read_setup, ncs_read_pulse;
+	unsigned long ncs_read_setup, ncs_read_pulse;
 
-	unsigned int mck_hz;
+	unsigned long mck_hz;
-	struct clk *mck;
 
 	read_cycle  = ata->cyc8b;
 	nrd_setup   = ata->setup;
 	nrd_pulse   = ata->act8b;
 	nrd_recover = ata->rec8b;
 
-	mck = clk_get(NULL, "mck");
+	mck_hz = clk_get_rate(info->mck);
-	BUG_ON(IS_ERR(mck));
-	mck_hz = clk_get_rate(mck);
 
 	read_cycle  = calc_mck_cycles(read_cycle, mck_hz);
 	nrd_setup   = calc_mck_cycles(nrd_setup, mck_hz);
 	nrd_pulse   = calc_mck_cycles(nrd_pulse, mck_hz);
 	nrd_recover = calc_mck_cycles(nrd_recover, mck_hz);
 
-	clk_put(mck);
-
 	active  = nrd_setup + nrd_pulse;
 	recover = read_cycle - active;
 
@@ -121,13 +116,13 @@ static void set_smc_timing(struct device *dev,
 	ncs_write_setup = ncs_read_setup;
 	ncs_write_pulse = ncs_read_pulse;
 
-	dev_dbg(dev, "ATA timings: nrd_setup = %d nrd_pulse = %d nrd_cycle = %d\n",
+	dev_dbg(dev, "ATA timings: nrd_setup = %lu nrd_pulse = %lu nrd_cycle = %lu\n",
 			nrd_setup, nrd_pulse, read_cycle);
-	dev_dbg(dev, "ATA timings: nwe_setup = %d nwe_pulse = %d nwe_cycle = %d\n",
+	dev_dbg(dev, "ATA timings: nwe_setup = %lu nwe_pulse = %lu nwe_cycle = %lu\n",
 			nwe_setup, nwe_pulse, write_cycle);
-	dev_dbg(dev, "ATA timings: ncs_read_setup = %d ncs_read_pulse = %d\n",
+	dev_dbg(dev, "ATA timings: ncs_read_setup = %lu ncs_read_pulse = %lu\n",
 			ncs_read_setup, ncs_read_pulse);
-	dev_dbg(dev, "ATA timings: ncs_write_setup = %d ncs_write_pulse = %d\n",
+	dev_dbg(dev, "ATA timings: ncs_write_setup = %lu ncs_write_pulse = %lu\n",
 			ncs_write_setup, ncs_write_pulse);
 
 	at91_sys_write(AT91_SMC_SETUP(info->cs),
@@ -217,6 +212,7 @@ static int __devinit pata_at91_probe(struct platform_device *pdev)
 	struct resource *mem_res;
 	struct ata_host *host;
 	struct ata_port *ap;
+
 	int irq_flags = 0;
 	int irq = 0;
 	int ret;
@@ -261,6 +257,13 @@ static int __devinit pata_at91_probe(struct platform_device *pdev)
 		return -ENOMEM;
 	}
 
+	info->mck = clk_get(NULL, "mck");
+
+	if (IS_ERR(info->mck)) {
+		dev_err(dev, "failed to get access to mck clock\n");
+		return -ENODEV;
+	}
+
 	info->cs    = board->chipselect;
 	info->mode  = AT91_SMC_READMODE | AT91_SMC_WRITEMODE |
 		AT91_SMC_EXNWMODE_READY | AT91_SMC_BAT_SELECT |
@@ -304,6 +307,7 @@ err_alt_ioremap:
 	devm_iounmap(dev, info->ide_addr);
 
 err_ide_ioremap:
+	clk_put(info->mck);
 	kfree(info);
 
 	return ret;
@@ -326,6 +330,7 @@ static int __devexit pata_at91_remove(struct platform_device *pdev)
 
 	devm_iounmap(dev, info->ide_addr);
 	devm_iounmap(dev, info->alt_addr);
+	clk_put(info->mck);
 
 	kfree(info);
 	return 0;

drivers/ata/pata_octeon_cf.c

@@ -653,7 +653,8 @@ static irqreturn_t octeon_cf_interrupt(int irq, void *dev_instance)
 
 		ap = host->ports[i];
 		ocd = ap->dev->platform_data;
-		if (!ap || (ap->flags & ATA_FLAG_DISABLED))
+
+		if (ap->flags & ATA_FLAG_DISABLED)
 			continue;
 
 		ocd = ap->dev->platform_data;

drivers/ata/pata_pcmcia.c

@@ -411,6 +411,7 @@ static struct pcmcia_device_id pcmcia_devices[] = {
 	PCMCIA_DEVICE_PROD_ID123("PCMCIA", "IDE CARD", "F1", 0x281f1c5d, 0x1907960c, 0xf7fde8b9),
 	PCMCIA_DEVICE_PROD_ID12("ARGOSY", "CD-ROM", 0x78f308dc, 0x66536591),
 	PCMCIA_DEVICE_PROD_ID12("ARGOSY", "PnPIDE", 0x78f308dc, 0x0c694728),
+	PCMCIA_DEVICE_PROD_ID12("CNF   ", "CD-ROM", 0x46d7db81, 0x66536591),
 	PCMCIA_DEVICE_PROD_ID12("CNF CD-M", "CD-ROM", 0x7d93b852, 0x66536591),
 	PCMCIA_DEVICE_PROD_ID12("Creative Technology Ltd.", "PCMCIA CD-ROM Interface Card", 0xff8c8a45, 0xfe8020c4),
 	PCMCIA_DEVICE_PROD_ID12("Digital Equipment Corporation.", "Digital Mobile Media CD-ROM", 0x17692a66, 0xef1dcbde),

drivers/ata/sata_mv.c

@@ -2514,7 +2514,7 @@ static void mv_unexpected_intr(struct ata_port *ap, int edma_was_enabled)
 	char *when = "idle";
 
 	ata_ehi_clear_desc(ehi);
-	if (!ap || (ap->flags & ATA_FLAG_DISABLED)) {
+	if (ap->flags & ATA_FLAG_DISABLED) {
 		when = "disabled";
 	} else if (edma_was_enabled) {
 		when = "EDMA enabled";

drivers/ata/sata_sil.c

@@ -532,7 +532,7 @@ static irqreturn_t sil_interrupt(int irq, void *dev_instance)
 		struct ata_port *ap = host->ports[i];
 		u32 bmdma2 = readl(mmio_base + sil_port[ap->port_no].bmdma2);
 
-		if (unlikely(!ap || ap->flags & ATA_FLAG_DISABLED))
+		if (unlikely(ap->flags & ATA_FLAG_DISABLED))
 			continue;
 
 		/* turn off SATA_IRQ if not supported */

drivers/base/firmware_class.c

@@ -180,7 +180,6 @@ static ssize_t firmware_loading_store(struct device *dev,
 				goto err;
 			}
 			/* Pages will be freed by vfree() */
-			fw_priv->pages = NULL;
 			fw_priv->page_array_size = 0;
 			fw_priv->nr_pages = 0;
 			complete(&fw_priv->completion);

drivers/base/sys.c

@@ -275,9 +275,9 @@ int sysdev_register(struct sys_device *sysdev)
 				drv->add(sysdev);
 		}
 		mutex_unlock(&sysdev_drivers_lock);
+		kobject_uevent(&sysdev->kobj, KOBJ_ADD);
 	}
 
-	kobject_uevent(&sysdev->kobj, KOBJ_ADD);
 	return error;
 }
 

drivers/char/n_tty.c

@@ -1583,6 +1583,7 @@ static int n_tty_open(struct tty_struct *tty)
 
 static inline int input_available_p(struct tty_struct *tty, int amt)
 {
+	tty_flush_to_ldisc(tty);
 	if (tty->icanon) {
 		if (tty->canon_data)
 			return 1;

drivers/char/tty_buffer.c

@@ -461,6 +461,19 @@ static void flush_to_ldisc(struct work_struct *work)
 	tty_ldisc_deref(disc);
 }
 
+/**
+ *	tty_flush_to_ldisc
+ *	@tty: tty to push
+ *
+ *	Push the terminal flip buffers to the line discipline.
+ *
+ *	Must not be called from IRQ context.
+ */
+void tty_flush_to_ldisc(struct tty_struct *tty)
+{
+	flush_to_ldisc(&tty->buf.work.work);
+}
+
 /**
  *	tty_flip_buffer_push	-	terminal
  *	@tty: tty to push

drivers/edac/amd64_edac.c

@@ -970,7 +970,7 @@ static void amd64_read_dct_base_mask(struct amd64_pvt *pvt)
 	}
 
 	for (cs = 0; cs < pvt->num_dcsm; cs++) {
-		reg = K8_DCSB0 + (cs * 4);
+		reg = K8_DCSM0 + (cs * 4);
 		err = pci_read_config_dword(pvt->dram_f2_ctl, reg,
 					&pvt->dcsm0[cs]);
 		if (unlikely(err))

drivers/gpu/drm/radeon/Makefile

@@ -13,7 +13,8 @@ radeon-$(CONFIG_DRM_RADEON_KMS) += radeon_device.o radeon_kms.o \
 	radeon_encoders.o radeon_display.o radeon_cursor.o radeon_i2c.o \
 	radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \
 	radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
-	rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o
+	rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o \
+	radeon_test.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 

drivers/gpu/drm/radeon/atombios_crtc.c

@@ -31,6 +31,132 @@
 #include "atom.h"
 #include "atom-bits.h"
 
+static void atombios_overscan_setup(struct drm_crtc *crtc,
+				    struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	SET_CRTC_OVERSCAN_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, SetCRTC_OverScan);
+	int a1, a2;
+
+	memset(&args, 0, sizeof(args));
+
+	args.usOverscanRight = 0;
+	args.usOverscanLeft = 0;
+	args.usOverscanBottom = 0;
+	args.usOverscanTop = 0;
+	args.ucCRTC = radeon_crtc->crtc_id;
+
+	switch (radeon_crtc->rmx_type) {
+	case RMX_CENTER:
+		args.usOverscanTop = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2;
+		args.usOverscanBottom = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2;
+		args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2;
+		args.usOverscanRight = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2;
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+		break;
+	case RMX_ASPECT:
+		a1 = mode->crtc_vdisplay * adjusted_mode->crtc_hdisplay;
+		a2 = adjusted_mode->crtc_vdisplay * mode->crtc_hdisplay;
+
+		if (a1 > a2) {
+			args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2;
+			args.usOverscanRight = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2;
+		} else if (a2 > a1) {
+			args.usOverscanLeft = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2;
+			args.usOverscanRight = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2;
+		}
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+		break;
+	case RMX_FULL:
+	default:
+		args.usOverscanRight = 0;
+		args.usOverscanLeft = 0;
+		args.usOverscanBottom = 0;
+		args.usOverscanTop = 0;
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+		break;
+	}
+}
+
+static void atombios_scaler_setup(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	ENABLE_SCALER_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, EnableScaler);
+	/* fixme - fill in enc_priv for atom dac */
+	enum radeon_tv_std tv_std = TV_STD_NTSC;
+
+	if (!ASIC_IS_AVIVO(rdev) && radeon_crtc->crtc_id)
+		return;
+
+	memset(&args, 0, sizeof(args));
+
+	args.ucScaler = radeon_crtc->crtc_id;
+
+	if (radeon_crtc->devices & (ATOM_DEVICE_TV_SUPPORT)) {
+		switch (tv_std) {
+		case TV_STD_NTSC:
+		default:
+			args.ucTVStandard = ATOM_TV_NTSC;
+			break;
+		case TV_STD_PAL:
+			args.ucTVStandard = ATOM_TV_PAL;
+			break;
+		case TV_STD_PAL_M:
+			args.ucTVStandard = ATOM_TV_PALM;
+			break;
+		case TV_STD_PAL_60:
+			args.ucTVStandard = ATOM_TV_PAL60;
+			break;
+		case TV_STD_NTSC_J:
+			args.ucTVStandard = ATOM_TV_NTSCJ;
+			break;
+		case TV_STD_SCART_PAL:
+			args.ucTVStandard = ATOM_TV_PAL; /* ??? */
+			break;
+		case TV_STD_SECAM:
+			args.ucTVStandard = ATOM_TV_SECAM;
+			break;
+		case TV_STD_PAL_CN:
+			args.ucTVStandard = ATOM_TV_PALCN;
+			break;
+		}
+		args.ucEnable = SCALER_ENABLE_MULTITAP_MODE;
+	} else if (radeon_crtc->devices & (ATOM_DEVICE_CV_SUPPORT)) {
+		args.ucTVStandard = ATOM_TV_CV;
+		args.ucEnable = SCALER_ENABLE_MULTITAP_MODE;
+	} else {
+		switch (radeon_crtc->rmx_type) {
+		case RMX_FULL:
+			args.ucEnable = ATOM_SCALER_EXPANSION;
+			break;
+		case RMX_CENTER:
+			args.ucEnable = ATOM_SCALER_CENTER;
+			break;
+		case RMX_ASPECT:
+			args.ucEnable = ATOM_SCALER_EXPANSION;
+			break;
+		default:
+			if (ASIC_IS_AVIVO(rdev))
+				args.ucEnable = ATOM_SCALER_DISABLE;
+			else
+				args.ucEnable = ATOM_SCALER_CENTER;
+			break;
+		}
+	}
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+	if (radeon_crtc->devices & (ATOM_DEVICE_CV_SUPPORT | ATOM_DEVICE_TV_SUPPORT)
+	    && rdev->family >= CHIP_RV515 && rdev->family <= CHIP_RV570) {
+		atom_rv515_force_tv_scaler(rdev);
+	}
+}
+
 static void atombios_lock_crtc(struct drm_crtc *crtc, int lock)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
@@ -203,6 +329,12 @@ void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode)
 	if (ASIC_IS_AVIVO(rdev)) {
 		uint32_t ss_cntl;
 
+		if ((rdev->family == CHIP_RS600) ||
+		    (rdev->family == CHIP_RS690) ||
+		    (rdev->family == CHIP_RS740))
+			pll_flags |= (RADEON_PLL_USE_FRAC_FB_DIV |
+				      RADEON_PLL_PREFER_CLOSEST_LOWER);
+
 		if (ASIC_IS_DCE32(rdev) && mode->clock > 200000)	/* range limits??? */
 			pll_flags |= RADEON_PLL_PREFER_HIGH_FB_DIV;
 		else
@@ -321,7 +453,7 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y,
 	struct drm_gem_object *obj;
 	struct drm_radeon_gem_object *obj_priv;
 	uint64_t fb_location;
-	uint32_t fb_format, fb_pitch_pixels;
+	uint32_t fb_format, fb_pitch_pixels, tiling_flags;
 
 	if (!crtc->fb)
 		return -EINVAL;
@@ -358,7 +490,14 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y,
 		return -EINVAL;
 	}
 
-	/* TODO tiling */
+	radeon_object_get_tiling_flags(obj->driver_private,
+				       &tiling_flags, NULL);
+	if (tiling_flags & RADEON_TILING_MACRO)
+		fb_format |= AVIVO_D1GRPH_MACRO_ADDRESS_MODE;
+
+	if (tiling_flags & RADEON_TILING_MICRO)
+		fb_format |= AVIVO_D1GRPH_TILED;
+
 	if (radeon_crtc->crtc_id == 0)
 		WREG32(AVIVO_D1VGA_CONTROL, 0);
 	else
@@ -509,6 +648,9 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 		radeon_crtc_set_base(crtc, x, y, old_fb);
 		radeon_legacy_atom_set_surface(crtc);
 	}
+	atombios_overscan_setup(crtc, mode, adjusted_mode);
+	atombios_scaler_setup(crtc);
+	radeon_bandwidth_update(rdev);
 	return 0;
 }
 
@@ -516,6 +658,8 @@ static bool atombios_crtc_mode_fixup(struct drm_crtc *crtc,
 				     struct drm_display_mode *mode,
 				     struct drm_display_mode *adjusted_mode)
 {
+	if (!radeon_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode))
+		return false;
 	return true;
 }
 
@@ -548,148 +692,3 @@ void radeon_atombios_init_crtc(struct drm_device *dev,
 		    AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL;
 	drm_crtc_helper_add(&radeon_crtc->base, &atombios_helper_funcs);
 }
-
-void radeon_init_disp_bw_avivo(struct drm_device *dev,
-			       struct drm_display_mode *mode1,
-			       uint32_t pixel_bytes1,
-			       struct drm_display_mode *mode2,
-			       uint32_t pixel_bytes2)
-{
-	struct radeon_device *rdev = dev->dev_private;
-	fixed20_12 min_mem_eff;
-	fixed20_12 peak_disp_bw, mem_bw, pix_clk, pix_clk2, temp_ff;
-	fixed20_12 sclk_ff, mclk_ff;
-	uint32_t dc_lb_memory_split, temp;
-
-	min_mem_eff.full = rfixed_const_8(0);
-	if (rdev->disp_priority == 2) {
-		uint32_t mc_init_misc_lat_timer = 0;
-		if (rdev->family == CHIP_RV515)
-			mc_init_misc_lat_timer =
-			    RREG32_MC(RV515_MC_INIT_MISC_LAT_TIMER);
-		else if (rdev->family == CHIP_RS690)
-			mc_init_misc_lat_timer =
-			    RREG32_MC(RS690_MC_INIT_MISC_LAT_TIMER);
-
-		mc_init_misc_lat_timer &=
-		    ~(R300_MC_DISP1R_INIT_LAT_MASK <<
-		      R300_MC_DISP1R_INIT_LAT_SHIFT);
-		mc_init_misc_lat_timer &=
-		    ~(R300_MC_DISP0R_INIT_LAT_MASK <<
-		      R300_MC_DISP0R_INIT_LAT_SHIFT);
-
-		if (mode2)
-			mc_init_misc_lat_timer |=
-			    (1 << R300_MC_DISP1R_INIT_LAT_SHIFT);
-		if (mode1)
-			mc_init_misc_lat_timer |=
-			    (1 << R300_MC_DISP0R_INIT_LAT_SHIFT);
-
-		if (rdev->family == CHIP_RV515)
-			WREG32_MC(RV515_MC_INIT_MISC_LAT_TIMER,
-				  mc_init_misc_lat_timer);
-		else if (rdev->family == CHIP_RS690)
-			WREG32_MC(RS690_MC_INIT_MISC_LAT_TIMER,
-				  mc_init_misc_lat_timer);
-	}
-
-	/*
-	 * determine is there is enough bw for current mode
-	 */
-	temp_ff.full = rfixed_const(100);
-	mclk_ff.full = rfixed_const(rdev->clock.default_mclk);
-	mclk_ff.full = rfixed_div(mclk_ff, temp_ff);
-	sclk_ff.full = rfixed_const(rdev->clock.default_sclk);
-	sclk_ff.full = rfixed_div(sclk_ff, temp_ff);
-
-	temp = (rdev->mc.vram_width / 8) * (rdev->mc.vram_is_ddr ? 2 : 1);
-	temp_ff.full = rfixed_const(temp);
-	mem_bw.full = rfixed_mul(mclk_ff, temp_ff);
-	mem_bw.full = rfixed_mul(mem_bw, min_mem_eff);
-
-	pix_clk.full = 0;
-	pix_clk2.full = 0;
-	peak_disp_bw.full = 0;
-	if (mode1) {
-		temp_ff.full = rfixed_const(1000);
-		pix_clk.full = rfixed_const(mode1->clock);	/* convert to fixed point */
-		pix_clk.full = rfixed_div(pix_clk, temp_ff);
-		temp_ff.full = rfixed_const(pixel_bytes1);
-		peak_disp_bw.full += rfixed_mul(pix_clk, temp_ff);
-	}
-	if (mode2) {
-		temp_ff.full = rfixed_const(1000);
-		pix_clk2.full = rfixed_const(mode2->clock);	/* convert to fixed point */
-		pix_clk2.full = rfixed_div(pix_clk2, temp_ff);
-		temp_ff.full = rfixed_const(pixel_bytes2);
-		peak_disp_bw.full += rfixed_mul(pix_clk2, temp_ff);
-	}
-
-	if (peak_disp_bw.full >= mem_bw.full) {
-		DRM_ERROR
-		    ("You may not have enough display bandwidth for current mode\n"
-		     "If you have flickering problem, try to lower resolution, refresh rate, or color depth\n");
-		printk("peak disp bw %d, mem_bw %d\n",
-		       rfixed_trunc(peak_disp_bw), rfixed_trunc(mem_bw));
-	}
-
-	/*
-	 * Line Buffer Setup
-	 * There is a single line buffer shared by both display controllers.
-	 * DC_LB_MEMORY_SPLIT controls how that line buffer is shared between the display
-	 * controllers.  The paritioning can either be done manually or via one of four
-	 * preset allocations specified in bits 1:0:
-	 * 0 - line buffer is divided in half and shared between each display controller
-	 * 1 - D1 gets 3/4 of the line buffer, D2 gets 1/4
-	 * 2 - D1 gets the whole buffer
-	 * 3 - D1 gets 1/4 of the line buffer, D2 gets 3/4
-	 * Setting bit 2 of DC_LB_MEMORY_SPLIT controls switches to manual allocation mode.
-	 * In manual allocation mode, D1 always starts at 0, D1 end/2 is specified in bits
-	 * 14:4; D2 allocation follows D1.
-	 */
-
-	/* is auto or manual better ? */
-	dc_lb_memory_split =
-	    RREG32(AVIVO_DC_LB_MEMORY_SPLIT) & ~AVIVO_DC_LB_MEMORY_SPLIT_MASK;
-	dc_lb_memory_split &= ~AVIVO_DC_LB_MEMORY_SPLIT_SHIFT_MODE;
-#if 1
-	/* auto */
-	if (mode1 && mode2) {
-		if (mode1->hdisplay > mode2->hdisplay) {
-			if (mode1->hdisplay > 2560)
-				dc_lb_memory_split |=
-				    AVIVO_DC_LB_MEMORY_SPLIT_D1_3Q_D2_1Q;
-			else
-				dc_lb_memory_split |=
-				    AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF;
-		} else if (mode2->hdisplay > mode1->hdisplay) {
-			if (mode2->hdisplay > 2560)
-				dc_lb_memory_split |=
-				    AVIVO_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q;
-			else
-				dc_lb_memory_split |=
-				    AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF;
-		} else
-			dc_lb_memory_split |=
-			    AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF;
-	} else if (mode1) {
-		dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_D1_ONLY;
-	} else if (mode2) {
-		dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q;
-	}
-#else
-	/* manual */
-	dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_SHIFT_MODE;
-	dc_lb_memory_split &=
-	    ~(AVIVO_DC_LB_DISP1_END_ADR_MASK <<
-	      AVIVO_DC_LB_DISP1_END_ADR_SHIFT);
-	if (mode1) {
-		dc_lb_memory_split |=
-		    ((((mode1->hdisplay / 2) + 64) & AVIVO_DC_LB_DISP1_END_ADR_MASK)
-		     << AVIVO_DC_LB_DISP1_END_ADR_SHIFT);
-	} else if (mode2) {
-		dc_lb_memory_split |= (0 << AVIVO_DC_LB_DISP1_END_ADR_SHIFT);
-	}
-#endif
-	WREG32(AVIVO_DC_LB_MEMORY_SPLIT, dc_lb_memory_split);
-}

drivers/gpu/drm/radeon/r100.c

@@ -110,7 +110,7 @@ int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
 	}
-	rdev->gart.table.ram.ptr[i] = cpu_to_le32((uint32_t)addr);
+	rdev->gart.table.ram.ptr[i] = cpu_to_le32(lower_32_bits(addr));
 	return 0;
 }
 
@@ -173,8 +173,12 @@ void r100_mc_setup(struct radeon_device *rdev)
 		DRM_ERROR("Failed to register debugfs file for R100 MC !\n");
 	}
 	/* Write VRAM size in case we are limiting it */
-	WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size);
+	WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size);
-	tmp = rdev->mc.vram_location + rdev->mc.vram_size - 1;
+	/* Novell bug 204882 for RN50/M6/M7 with 8/16/32MB VRAM,
+	 * if the aperture is 64MB but we have 32MB VRAM
+	 * we report only 32MB VRAM but we have to set MC_FB_LOCATION
+	 * to 64MB, otherwise the gpu accidentially dies */
+	tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1;
 	tmp = REG_SET(RADEON_MC_FB_TOP, tmp >> 16);
 	tmp |= REG_SET(RADEON_MC_FB_START, rdev->mc.vram_location >> 16);
 	WREG32(RADEON_MC_FB_LOCATION, tmp);
@@ -215,7 +219,6 @@ int r100_mc_init(struct radeon_device *rdev)
 	r100_pci_gart_disable(rdev);
 
 	/* Setup GPU memory space */
-	rdev->mc.vram_location = 0xFFFFFFFFUL;
 	rdev->mc.gtt_location = 0xFFFFFFFFUL;
 	if (rdev->flags & RADEON_IS_AGP) {
 		r = radeon_agp_init(rdev);
@@ -752,6 +755,102 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p,
 	return 0;
 }
 
+/**
+ * r100_cs_packet_next_vline() - parse userspace VLINE packet
+ * @parser:		parser structure holding parsing context.
+ *
+ * Userspace sends a special sequence for VLINE waits.
+ * PACKET0 - VLINE_START_END + value
+ * PACKET0 - WAIT_UNTIL +_value
+ * RELOC (P3) - crtc_id in reloc.
+ *
+ * This function parses this and relocates the VLINE START END
+ * and WAIT UNTIL packets to the correct crtc.
+ * It also detects a switched off crtc and nulls out the
+ * wait in that case.
+ */
+int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
+{
+	struct radeon_cs_chunk *ib_chunk;
+	struct drm_mode_object *obj;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+	struct radeon_cs_packet p3reloc, waitreloc;
+	int crtc_id;
+	int r;
+	uint32_t header, h_idx, reg;
+
+	ib_chunk = &p->chunks[p->chunk_ib_idx];
+
+	/* parse the wait until */
+	r = r100_cs_packet_parse(p, &waitreloc, p->idx);
+	if (r)
+		return r;
+
+	/* check its a wait until and only 1 count */
+	if (waitreloc.reg != RADEON_WAIT_UNTIL ||
+	    waitreloc.count != 0) {
+		DRM_ERROR("vline wait had illegal wait until segment\n");
+		r = -EINVAL;
+		return r;
+	}
+
+	if (ib_chunk->kdata[waitreloc.idx + 1] != RADEON_WAIT_CRTC_VLINE) {
+		DRM_ERROR("vline wait had illegal wait until\n");
+		r = -EINVAL;
+		return r;
+	}
+
+	/* jump over the NOP */
+	r = r100_cs_packet_parse(p, &p3reloc, p->idx);
+	if (r)
+		return r;
+
+	h_idx = p->idx - 2;
+	p->idx += waitreloc.count;
+	p->idx += p3reloc.count;
+
+	header = ib_chunk->kdata[h_idx];
+	crtc_id = ib_chunk->kdata[h_idx + 5];
+	reg = ib_chunk->kdata[h_idx] >> 2;
+	mutex_lock(&p->rdev->ddev->mode_config.mutex);
+	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
+	if (!obj) {
+		DRM_ERROR("cannot find crtc %d\n", crtc_id);
+		r = -EINVAL;
+		goto out;
+	}
+	crtc = obj_to_crtc(obj);
+	radeon_crtc = to_radeon_crtc(crtc);
+	crtc_id = radeon_crtc->crtc_id;
+
+	if (!crtc->enabled) {
+		/* if the CRTC isn't enabled - we need to nop out the wait until */
+		ib_chunk->kdata[h_idx + 2] = PACKET2(0);
+		ib_chunk->kdata[h_idx + 3] = PACKET2(0);
+	} else if (crtc_id == 1) {
+		switch (reg) {
+		case AVIVO_D1MODE_VLINE_START_END:
+			header &= R300_CP_PACKET0_REG_MASK;
+			header |= AVIVO_D2MODE_VLINE_START_END >> 2;
+			break;
+		case RADEON_CRTC_GUI_TRIG_VLINE:
+			header &= R300_CP_PACKET0_REG_MASK;
+			header |= RADEON_CRTC2_GUI_TRIG_VLINE >> 2;
+			break;
+		default:
+			DRM_ERROR("unknown crtc reloc\n");
+			r = -EINVAL;
+			goto out;
+		}
+		ib_chunk->kdata[h_idx] = header;
+		ib_chunk->kdata[h_idx + 3] |= RADEON_ENG_DISPLAY_SELECT_CRTC1;
+	}
+out:
+	mutex_unlock(&p->rdev->ddev->mode_config.mutex);
+	return r;
+}
+
 /**
  * r100_cs_packet_next_reloc() - parse next packet which should be reloc packet3
  * @parser:		parser structure holding parsing context.
@@ -814,6 +913,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	unsigned idx;
 	bool onereg;
 	int r;
+	u32 tile_flags = 0;
 
 	ib = p->ib->ptr;
 	ib_chunk = &p->chunks[p->chunk_ib_idx];
@@ -825,6 +925,15 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	}
 	for (i = 0; i <= pkt->count; i++, idx++, reg += 4) {
 		switch (reg) {
+		case RADEON_CRTC_GUI_TRIG_VLINE:
+			r = r100_cs_packet_parse_vline(p);
+			if (r) {
+				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+						idx, reg);
+				r100_cs_dump_packet(p, pkt);
+				return r;
+			}
+			break;
 		/* FIXME: only allow PACKET3 blit? easier to check for out of
 		 * range access */
 		case RADEON_DST_PITCH_OFFSET:
@@ -838,7 +947,20 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			}
 			tmp = ib_chunk->kdata[idx] & 0x003fffff;
 			tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
-			ib[idx] = (ib_chunk->kdata[idx] & 0xffc00000) | tmp;
+
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= RADEON_DST_TILE_MACRO;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+				if (reg == RADEON_SRC_PITCH_OFFSET) {
+					DRM_ERROR("Cannot src blit from microtiled surface\n");
+					r100_cs_dump_packet(p, pkt);
+					return -EINVAL;
+				}
+				tile_flags |= RADEON_DST_TILE_MICRO;
+			}
+
+			tmp |= tile_flags;
+			ib[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp;
 			break;
 		case RADEON_RB3D_DEPTHOFFSET:
 		case RADEON_RB3D_COLOROFFSET:
@@ -869,6 +991,11 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		case R300_TX_OFFSET_0+52:
 		case R300_TX_OFFSET_0+56:
 		case R300_TX_OFFSET_0+60:
+			/* rn50 has no 3D engine so fail on any 3d setup */
+			if (ASIC_IS_RN50(p->rdev)) {
+				DRM_ERROR("attempt to use RN50 3D engine failed\n");
+				return -EINVAL;
+			}
 			r = r100_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
@@ -878,6 +1005,25 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			}
 			ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
 			break;
+		case R300_RB3D_COLORPITCH0:
+		case RADEON_RB3D_COLORPITCH:
+			r = r100_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+					  idx, reg);
+				r100_cs_dump_packet(p, pkt);
+				return r;
+			}
+
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= RADEON_COLOR_TILE_ENABLE;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+				tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
+
+			tmp = ib_chunk->kdata[idx] & ~(0x7 << 16);
+			tmp |= tile_flags;
+			ib[idx] = tmp;
+			break;
 		default:
 			/* FIXME: we don't want to allow anyothers packet */
 			break;
@@ -1256,29 +1402,100 @@ static void r100_vram_get_type(struct radeon_device *rdev)
 	}
 }
 
-void r100_vram_info(struct radeon_device *rdev)
+static u32 r100_get_accessible_vram(struct radeon_device *rdev)
 {
-	r100_vram_get_type(rdev);
+	u32 aper_size;
+	u8 byte;
+
+	aper_size = RREG32(RADEON_CONFIG_APER_SIZE);
+
+	/* Set HDP_APER_CNTL only on cards that are known not to be broken,
+	 * that is has the 2nd generation multifunction PCI interface
+	 */
+	if (rdev->family == CHIP_RV280 ||
+	    rdev->family >= CHIP_RV350) {
+		WREG32_P(RADEON_HOST_PATH_CNTL, RADEON_HDP_APER_CNTL,
+		       ~RADEON_HDP_APER_CNTL);
+		DRM_INFO("Generation 2 PCI interface, using max accessible memory\n");
+		return aper_size * 2;
+	}
+
+	/* Older cards have all sorts of funny issues to deal with. First
+	 * check if it's a multifunction card by reading the PCI config
+	 * header type... Limit those to one aperture size
+	 */
+	pci_read_config_byte(rdev->pdev, 0xe, &byte);
+	if (byte & 0x80) {
+		DRM_INFO("Generation 1 PCI interface in multifunction mode\n");
+		DRM_INFO("Limiting VRAM to one aperture\n");
+		return aper_size;
+	}
+
+	/* Single function older card. We read HDP_APER_CNTL to see how the BIOS
+	 * have set it up. We don't write this as it's broken on some ASICs but
+	 * we expect the BIOS to have done the right thing (might be too optimistic...)
+	 */
+	if (RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL)
+		return aper_size * 2;
+	return aper_size;
+}
+
+void r100_vram_init_sizes(struct radeon_device *rdev)
+{
+	u64 config_aper_size;
+	u32 accessible;
+
+	config_aper_size = RREG32(RADEON_CONFIG_APER_SIZE);
 
 	if (rdev->flags & RADEON_IS_IGP) {
 		uint32_t tom;
 		/* read NB_TOM to get the amount of ram stolen for the GPU */
 		tom = RREG32(RADEON_NB_TOM);
-		rdev->mc.vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16);
+		rdev->mc.real_vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16);
-		WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size);
+		/* for IGPs we need to keep VRAM where it was put by the BIOS */
+		rdev->mc.vram_location = (tom & 0xffff) << 16;
+		WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size);
+		rdev->mc.mc_vram_size = rdev->mc.real_vram_size;
 	} else {
-		rdev->mc.vram_size = RREG32(RADEON_CONFIG_MEMSIZE);
+		rdev->mc.real_vram_size = RREG32(RADEON_CONFIG_MEMSIZE);
 		/* Some production boards of m6 will report 0
 		 * if it's 8 MB
 		 */
-		if (rdev->mc.vram_size == 0) {
+		if (rdev->mc.real_vram_size == 0) {
-			rdev->mc.vram_size = 8192 * 1024;
+			rdev->mc.real_vram_size = 8192 * 1024;
-			WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size);
+			WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size);
 		}
+		/* let driver place VRAM */
+		rdev->mc.vram_location = 0xFFFFFFFFUL;
+		 /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - 
+		  * Novell bug 204882 + along with lots of ubuntu ones */
+		if (config_aper_size > rdev->mc.real_vram_size)
+			rdev->mc.mc_vram_size = config_aper_size;
+		else
+			rdev->mc.mc_vram_size = rdev->mc.real_vram_size;
 	}
 
+	/* work out accessible VRAM */
+	accessible = r100_get_accessible_vram(rdev);
+
 	rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0);
 	rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0);
+
+	if (accessible > rdev->mc.aper_size)
+		accessible = rdev->mc.aper_size;
+
+	if (rdev->mc.mc_vram_size > rdev->mc.aper_size)
+		rdev->mc.mc_vram_size = rdev->mc.aper_size;
+
+	if (rdev->mc.real_vram_size > rdev->mc.aper_size)
+		rdev->mc.real_vram_size = rdev->mc.aper_size;
+}
+
+void r100_vram_info(struct radeon_device *rdev)
+{
+	r100_vram_get_type(rdev);
+
+	r100_vram_init_sizes(rdev);
 }
 
 
@@ -1533,3 +1750,530 @@ int r100_debugfs_mc_info_init(struct radeon_device *rdev)
 	return 0;
 #endif
 }
+
+int r100_set_surface_reg(struct radeon_device *rdev, int reg,
+			 uint32_t tiling_flags, uint32_t pitch,
+			 uint32_t offset, uint32_t obj_size)
+{
+	int surf_index = reg * 16;
+	int flags = 0;
+
+	/* r100/r200 divide by 16 */
+	if (rdev->family < CHIP_R300)
+		flags = pitch / 16;
+	else
+		flags = pitch / 8;
+
+	if (rdev->family <= CHIP_RS200) {
+		if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
+				 == (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
+			flags |= RADEON_SURF_TILE_COLOR_BOTH;
+		if (tiling_flags & RADEON_TILING_MACRO)
+			flags |= RADEON_SURF_TILE_COLOR_MACRO;
+	} else if (rdev->family <= CHIP_RV280) {
+		if (tiling_flags & (RADEON_TILING_MACRO))
+			flags |= R200_SURF_TILE_COLOR_MACRO;
+		if (tiling_flags & RADEON_TILING_MICRO)
+			flags |= R200_SURF_TILE_COLOR_MICRO;
+	} else {
+		if (tiling_flags & RADEON_TILING_MACRO)
+			flags |= R300_SURF_TILE_MACRO;
+		if (tiling_flags & RADEON_TILING_MICRO)
+			flags |= R300_SURF_TILE_MICRO;
+	}
+
+	DRM_DEBUG("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1);
+	WREG32(RADEON_SURFACE0_INFO + surf_index, flags);
+	WREG32(RADEON_SURFACE0_LOWER_BOUND + surf_index, offset);
+	WREG32(RADEON_SURFACE0_UPPER_BOUND + surf_index, offset + obj_size - 1);
+	return 0;
+}
+
+void r100_clear_surface_reg(struct radeon_device *rdev, int reg)
+{
+	int surf_index = reg * 16;
+	WREG32(RADEON_SURFACE0_INFO + surf_index, 0);
+}
+
+void r100_bandwidth_update(struct radeon_device *rdev)
+{
+	fixed20_12 trcd_ff, trp_ff, tras_ff, trbs_ff, tcas_ff;
+	fixed20_12 sclk_ff, mclk_ff, sclk_eff_ff, sclk_delay_ff;
+	fixed20_12 peak_disp_bw, mem_bw, pix_clk, pix_clk2, temp_ff, crit_point_ff;
+	uint32_t temp, data, mem_trcd, mem_trp, mem_tras;
+	fixed20_12 memtcas_ff[8] = {
+		fixed_init(1),
+		fixed_init(2),
+		fixed_init(3),
+		fixed_init(0),
+		fixed_init_half(1),
+		fixed_init_half(2),
+		fixed_init(0),
+	};
+	fixed20_12 memtcas_rs480_ff[8] = {
+		fixed_init(0),
+		fixed_init(1),
+		fixed_init(2),
+		fixed_init(3),
+		fixed_init(0),
+		fixed_init_half(1),
+		fixed_init_half(2),
+		fixed_init_half(3),
+	};
+	fixed20_12 memtcas2_ff[8] = {
+		fixed_init(0),
+		fixed_init(1),
+		fixed_init(2),
+		fixed_init(3),
+		fixed_init(4),
+		fixed_init(5),
+		fixed_init(6),
+		fixed_init(7),
+	};
+	fixed20_12 memtrbs[8] = {
+		fixed_init(1),
+		fixed_init_half(1),
+		fixed_init(2),
+		fixed_init_half(2),
+		fixed_init(3),
+		fixed_init_half(3),
+		fixed_init(4),
+		fixed_init_half(4)
+	};
+	fixed20_12 memtrbs_r4xx[8] = {
+		fixed_init(4),
+		fixed_init(5),
+		fixed_init(6),
+		fixed_init(7),
+		fixed_init(8),
+		fixed_init(9),
+		fixed_init(10),
+		fixed_init(11)
+	};
+	fixed20_12 min_mem_eff;
+	fixed20_12 mc_latency_sclk, mc_latency_mclk, k1;
+	fixed20_12 cur_latency_mclk, cur_latency_sclk;
+	fixed20_12 disp_latency, disp_latency_overhead, disp_drain_rate,
+		disp_drain_rate2, read_return_rate;
+	fixed20_12 time_disp1_drop_priority;
+	int c;
+	int cur_size = 16;       /* in octawords */
+	int critical_point = 0, critical_point2;
+/* 	uint32_t read_return_rate, time_disp1_drop_priority; */
+	int stop_req, max_stop_req;
+	struct drm_display_mode *mode1 = NULL;
+	struct drm_display_mode *mode2 = NULL;
+	uint32_t pixel_bytes1 = 0;
+	uint32_t pixel_bytes2 = 0;
+
+	if (rdev->mode_info.crtcs[0]->base.enabled) {
+		mode1 = &rdev->mode_info.crtcs[0]->base.mode;
+		pixel_bytes1 = rdev->mode_info.crtcs[0]->base.fb->bits_per_pixel / 8;
+	}
+	if (rdev->mode_info.crtcs[1]->base.enabled) {
+		mode2 = &rdev->mode_info.crtcs[1]->base.mode;
+		pixel_bytes2 = rdev->mode_info.crtcs[1]->base.fb->bits_per_pixel / 8;
+	}
+
+	min_mem_eff.full = rfixed_const_8(0);
+	/* get modes */
+	if ((rdev->disp_priority == 2) && ASIC_IS_R300(rdev)) {
+		uint32_t mc_init_misc_lat_timer = RREG32(R300_MC_INIT_MISC_LAT_TIMER);
+		mc_init_misc_lat_timer &= ~(R300_MC_DISP1R_INIT_LAT_MASK << R300_MC_DISP1R_INIT_LAT_SHIFT);
+		mc_init_misc_lat_timer &= ~(R300_MC_DISP0R_INIT_LAT_MASK << R300_MC_DISP0R_INIT_LAT_SHIFT);
+		/* check crtc enables */
+		if (mode2)
+			mc_init_misc_lat_timer |= (1 << R300_MC_DISP1R_INIT_LAT_SHIFT);
+		if (mode1)
+			mc_init_misc_lat_timer |= (1 << R300_MC_DISP0R_INIT_LAT_SHIFT);
+		WREG32(R300_MC_INIT_MISC_LAT_TIMER, mc_init_misc_lat_timer);
+	}
+
+	/*
+	 * determine is there is enough bw for current mode
+	 */
+	mclk_ff.full = rfixed_const(rdev->clock.default_mclk);
+	temp_ff.full = rfixed_const(100);
+	mclk_ff.full = rfixed_div(mclk_ff, temp_ff);
+	sclk_ff.full = rfixed_const(rdev->clock.default_sclk);
+	sclk_ff.full = rfixed_div(sclk_ff, temp_ff);
+
+	temp = (rdev->mc.vram_width / 8) * (rdev->mc.vram_is_ddr ? 2 : 1);
+	temp_ff.full = rfixed_const(temp);
+	mem_bw.full = rfixed_mul(mclk_ff, temp_ff);
+
+	pix_clk.full = 0;
+	pix_clk2.full = 0;
+	peak_disp_bw.full = 0;
+	if (mode1) {
+		temp_ff.full = rfixed_const(1000);
+		pix_clk.full = rfixed_const(mode1->clock); /* convert to fixed point */
+		pix_clk.full = rfixed_div(pix_clk, temp_ff);
+		temp_ff.full = rfixed_const(pixel_bytes1);
+		peak_disp_bw.full += rfixed_mul(pix_clk, temp_ff);
+	}
+	if (mode2) {
+		temp_ff.full = rfixed_const(1000);
+		pix_clk2.full = rfixed_const(mode2->clock); /* convert to fixed point */
+		pix_clk2.full = rfixed_div(pix_clk2, temp_ff);
+		temp_ff.full = rfixed_const(pixel_bytes2);
+		peak_disp_bw.full += rfixed_mul(pix_clk2, temp_ff);
+	}
+
+	mem_bw.full = rfixed_mul(mem_bw, min_mem_eff);
+	if (peak_disp_bw.full >= mem_bw.full) {
+		DRM_ERROR("You may not have enough display bandwidth for current mode\n"
+			  "If you have flickering problem, try to lower resolution, refresh rate, or color depth\n");
+	}
+
+	/*  Get values from the EXT_MEM_CNTL register...converting its contents. */
+	temp = RREG32(RADEON_MEM_TIMING_CNTL);
+	if ((rdev->family == CHIP_RV100) || (rdev->flags & RADEON_IS_IGP)) { /* RV100, M6, IGPs */
+		mem_trcd = ((temp >> 2) & 0x3) + 1;
+		mem_trp  = ((temp & 0x3)) + 1;
+		mem_tras = ((temp & 0x70) >> 4) + 1;
+	} else if (rdev->family == CHIP_R300 ||
+		   rdev->family == CHIP_R350) { /* r300, r350 */
+		mem_trcd = (temp & 0x7) + 1;
+		mem_trp = ((temp >> 8) & 0x7) + 1;
+		mem_tras = ((temp >> 11) & 0xf) + 4;
+	} else if (rdev->family == CHIP_RV350 ||
+		   rdev->family <= CHIP_RV380) {
+		/* rv3x0 */
+		mem_trcd = (temp & 0x7) + 3;
+		mem_trp = ((temp >> 8) & 0x7) + 3;
+		mem_tras = ((temp >> 11) & 0xf) + 6;
+	} else if (rdev->family == CHIP_R420 ||
+		   rdev->family == CHIP_R423 ||
+		   rdev->family == CHIP_RV410) {
+		/* r4xx */
+		mem_trcd = (temp & 0xf) + 3;
+		if (mem_trcd > 15)
+			mem_trcd = 15;
+		mem_trp = ((temp >> 8) & 0xf) + 3;
+		if (mem_trp > 15)
+			mem_trp = 15;
+		mem_tras = ((temp >> 12) & 0x1f) + 6;
+		if (mem_tras > 31)
+			mem_tras = 31;
+	} else { /* RV200, R200 */
+		mem_trcd = (temp & 0x7) + 1;
+		mem_trp = ((temp >> 8) & 0x7) + 1;
+		mem_tras = ((temp >> 12) & 0xf) + 4;
+	}
+	/* convert to FF */
+	trcd_ff.full = rfixed_const(mem_trcd);
+	trp_ff.full = rfixed_const(mem_trp);
+	tras_ff.full = rfixed_const(mem_tras);
+
+	/* Get values from the MEM_SDRAM_MODE_REG register...converting its */
+	temp = RREG32(RADEON_MEM_SDRAM_MODE_REG);
+	data = (temp & (7 << 20)) >> 20;
+	if ((rdev->family == CHIP_RV100) || rdev->flags & RADEON_IS_IGP) {
+		if (rdev->family == CHIP_RS480) /* don't think rs400 */
+			tcas_ff = memtcas_rs480_ff[data];
+		else
+			tcas_ff = memtcas_ff[data];
+	} else
+		tcas_ff = memtcas2_ff[data];
+
+	if (rdev->family == CHIP_RS400 ||
+	    rdev->family == CHIP_RS480) {
+		/* extra cas latency stored in bits 23-25 0-4 clocks */
+		data = (temp >> 23) & 0x7;
+		if (data < 5)
+			tcas_ff.full += rfixed_const(data);
+	}
+
+	if (ASIC_IS_R300(rdev) && !(rdev->flags & RADEON_IS_IGP)) {
+		/* on the R300, Tcas is included in Trbs.
+		 */
+		temp = RREG32(RADEON_MEM_CNTL);
+		data = (R300_MEM_NUM_CHANNELS_MASK & temp);
+		if (data == 1) {
+			if (R300_MEM_USE_CD_CH_ONLY & temp) {
+				temp = RREG32(R300_MC_IND_INDEX);
+				temp &= ~R300_MC_IND_ADDR_MASK;
+				temp |= R300_MC_READ_CNTL_CD_mcind;
+				WREG32(R300_MC_IND_INDEX, temp);
+				temp = RREG32(R300_MC_IND_DATA);
+				data = (R300_MEM_RBS_POSITION_C_MASK & temp);
+			} else {
+				temp = RREG32(R300_MC_READ_CNTL_AB);
+				data = (R300_MEM_RBS_POSITION_A_MASK & temp);
+			}
+		} else {
+			temp = RREG32(R300_MC_READ_CNTL_AB);
+			data = (R300_MEM_RBS_POSITION_A_MASK & temp);
+		}
+		if (rdev->family == CHIP_RV410 ||
+		    rdev->family == CHIP_R420 ||
+		    rdev->family == CHIP_R423)
+			trbs_ff = memtrbs_r4xx[data];
+		else
+			trbs_ff = memtrbs[data];
+		tcas_ff.full += trbs_ff.full;
+	}
+
+	sclk_eff_ff.full = sclk_ff.full;
+
+	if (rdev->flags & RADEON_IS_AGP) {
+		fixed20_12 agpmode_ff;
+		agpmode_ff.full = rfixed_const(radeon_agpmode);
+		temp_ff.full = rfixed_const_666(16);
+		sclk_eff_ff.full -= rfixed_mul(agpmode_ff, temp_ff);
+	}
+	/* TODO PCIE lanes may affect this - agpmode == 16?? */
+
+	if (ASIC_IS_R300(rdev)) {
+		sclk_delay_ff.full = rfixed_const(250);
+	} else {
+		if ((rdev->family == CHIP_RV100) ||
+		    rdev->flags & RADEON_IS_IGP) {
+			if (rdev->mc.vram_is_ddr)
+				sclk_delay_ff.full = rfixed_const(41);
+			else
+				sclk_delay_ff.full = rfixed_const(33);
+		} else {
+			if (rdev->mc.vram_width == 128)
+				sclk_delay_ff.full = rfixed_const(57);
+			else
+				sclk_delay_ff.full = rfixed_const(41);
+		}
+	}
+
+	mc_latency_sclk.full = rfixed_div(sclk_delay_ff, sclk_eff_ff);
+
+	if (rdev->mc.vram_is_ddr) {
+		if (rdev->mc.vram_width == 32) {
+			k1.full = rfixed_const(40);
+			c  = 3;
+		} else {
+			k1.full = rfixed_const(20);
+			c  = 1;
+		}
+	} else {
+		k1.full = rfixed_const(40);
+		c  = 3;
+	}
+
+	temp_ff.full = rfixed_const(2);
+	mc_latency_mclk.full = rfixed_mul(trcd_ff, temp_ff);
+	temp_ff.full = rfixed_const(c);
+	mc_latency_mclk.full += rfixed_mul(tcas_ff, temp_ff);
+	temp_ff.full = rfixed_const(4);
+	mc_latency_mclk.full += rfixed_mul(tras_ff, temp_ff);
+	mc_latency_mclk.full += rfixed_mul(trp_ff, temp_ff);
+	mc_latency_mclk.full += k1.full;
+
+	mc_latency_mclk.full = rfixed_div(mc_latency_mclk, mclk_ff);
+	mc_latency_mclk.full += rfixed_div(temp_ff, sclk_eff_ff);
+
+	/*
+	  HW cursor time assuming worst case of full size colour cursor.
+	*/
+	temp_ff.full = rfixed_const((2 * (cur_size - (rdev->mc.vram_is_ddr + 1))));
+	temp_ff.full += trcd_ff.full;
+	if (temp_ff.full < tras_ff.full)
+		temp_ff.full = tras_ff.full;
+	cur_latency_mclk.full = rfixed_div(temp_ff, mclk_ff);
+
+	temp_ff.full = rfixed_const(cur_size);
+	cur_latency_sclk.full = rfixed_div(temp_ff, sclk_eff_ff);
+	/*
+	  Find the total latency for the display data.
+	*/
+	disp_latency_overhead.full = rfixed_const(80);
+	disp_latency_overhead.full = rfixed_div(disp_latency_overhead, sclk_ff);
+	mc_latency_mclk.full += disp_latency_overhead.full + cur_latency_mclk.full;
+	mc_latency_sclk.full += disp_latency_overhead.full + cur_latency_sclk.full;
+
+	if (mc_latency_mclk.full > mc_latency_sclk.full)
+		disp_latency.full = mc_latency_mclk.full;
+	else
+		disp_latency.full = mc_latency_sclk.full;
+
+	/* setup Max GRPH_STOP_REQ default value */
+	if (ASIC_IS_RV100(rdev))
+		max_stop_req = 0x5c;
+	else
+		max_stop_req = 0x7c;
+
+	if (mode1) {
+		/*  CRTC1
+		    Set GRPH_BUFFER_CNTL register using h/w defined optimal values.
+		    GRPH_STOP_REQ <= MIN[ 0x7C, (CRTC_H_DISP + 1) * (bit depth) / 0x10 ]
+		*/
+		stop_req = mode1->hdisplay * pixel_bytes1 / 16;
+
+		if (stop_req > max_stop_req)
+			stop_req = max_stop_req;
+
+		/*
+		  Find the drain rate of the display buffer.
+		*/
+		temp_ff.full = rfixed_const((16/pixel_bytes1));
+		disp_drain_rate.full = rfixed_div(pix_clk, temp_ff);
+
+		/*
+		  Find the critical point of the display buffer.
+		*/
+		crit_point_ff.full = rfixed_mul(disp_drain_rate, disp_latency);
+		crit_point_ff.full += rfixed_const_half(0);
+
+		critical_point = rfixed_trunc(crit_point_ff);
+
+		if (rdev->disp_priority == 2) {
+			critical_point = 0;
+		}
+
+		/*
+		  The critical point should never be above max_stop_req-4.  Setting
+		  GRPH_CRITICAL_CNTL = 0 will thus force high priority all the time.
+		*/
+		if (max_stop_req - critical_point < 4)
+			critical_point = 0;
+
+		if (critical_point == 0 && mode2 && rdev->family == CHIP_R300) {
+			/* some R300 cards have problem with this set to 0, when CRTC2 is enabled.*/
+			critical_point = 0x10;
+		}
+
+		temp = RREG32(RADEON_GRPH_BUFFER_CNTL);
+		temp &= ~(RADEON_GRPH_STOP_REQ_MASK);
+		temp |= (stop_req << RADEON_GRPH_STOP_REQ_SHIFT);
+		temp &= ~(RADEON_GRPH_START_REQ_MASK);
+		if ((rdev->family == CHIP_R350) &&
+		    (stop_req > 0x15)) {
+			stop_req -= 0x10;
+		}
+		temp |= (stop_req << RADEON_GRPH_START_REQ_SHIFT);
+		temp |= RADEON_GRPH_BUFFER_SIZE;
+		temp &= ~(RADEON_GRPH_CRITICAL_CNTL   |
+			  RADEON_GRPH_CRITICAL_AT_SOF |
+			  RADEON_GRPH_STOP_CNTL);
+		/*
+		  Write the result into the register.
+		*/
+		WREG32(RADEON_GRPH_BUFFER_CNTL, ((temp & ~RADEON_GRPH_CRITICAL_POINT_MASK) |
+						       (critical_point << RADEON_GRPH_CRITICAL_POINT_SHIFT)));
+
+#if 0
+		if ((rdev->family == CHIP_RS400) ||
+		    (rdev->family == CHIP_RS480)) {
+			/* attempt to program RS400 disp regs correctly ??? */
+			temp = RREG32(RS400_DISP1_REG_CNTL);
+			temp &= ~(RS400_DISP1_START_REQ_LEVEL_MASK |
+				  RS400_DISP1_STOP_REQ_LEVEL_MASK);
+			WREG32(RS400_DISP1_REQ_CNTL1, (temp |
+						       (critical_point << RS400_DISP1_START_REQ_LEVEL_SHIFT) |
+						       (critical_point << RS400_DISP1_STOP_REQ_LEVEL_SHIFT)));
+			temp = RREG32(RS400_DMIF_MEM_CNTL1);
+			temp &= ~(RS400_DISP1_CRITICAL_POINT_START_MASK |
+				  RS400_DISP1_CRITICAL_POINT_STOP_MASK);
+			WREG32(RS400_DMIF_MEM_CNTL1, (temp |
+						      (critical_point << RS400_DISP1_CRITICAL_POINT_START_SHIFT) |
+						      (critical_point << RS400_DISP1_CRITICAL_POINT_STOP_SHIFT)));
+		}
+#endif
+
+		DRM_DEBUG("GRPH_BUFFER_CNTL from to %x\n",
+			  /* 	  (unsigned int)info->SavedReg->grph_buffer_cntl, */
+			  (unsigned int)RREG32(RADEON_GRPH_BUFFER_CNTL));
+	}
+
+	if (mode2) {
+		u32 grph2_cntl;
+		stop_req = mode2->hdisplay * pixel_bytes2 / 16;
+
+		if (stop_req > max_stop_req)
+			stop_req = max_stop_req;
+
+		/*
+		  Find the drain rate of the display buffer.
+		*/
+		temp_ff.full = rfixed_const((16/pixel_bytes2));
+		disp_drain_rate2.full = rfixed_div(pix_clk2, temp_ff);
+
+		grph2_cntl = RREG32(RADEON_GRPH2_BUFFER_CNTL);
+		grph2_cntl &= ~(RADEON_GRPH_STOP_REQ_MASK);
+		grph2_cntl |= (stop_req << RADEON_GRPH_STOP_REQ_SHIFT);
+		grph2_cntl &= ~(RADEON_GRPH_START_REQ_MASK);
+		if ((rdev->family == CHIP_R350) &&
+		    (stop_req > 0x15)) {
+			stop_req -= 0x10;
+		}
+		grph2_cntl |= (stop_req << RADEON_GRPH_START_REQ_SHIFT);
+		grph2_cntl |= RADEON_GRPH_BUFFER_SIZE;
+		grph2_cntl &= ~(RADEON_GRPH_CRITICAL_CNTL   |
+			  RADEON_GRPH_CRITICAL_AT_SOF |
+			  RADEON_GRPH_STOP_CNTL);
+
+		if ((rdev->family == CHIP_RS100) ||
+		    (rdev->family == CHIP_RS200))
+			critical_point2 = 0;
+		else {
+			temp = (rdev->mc.vram_width * rdev->mc.vram_is_ddr + 1)/128;
+			temp_ff.full = rfixed_const(temp);
+			temp_ff.full = rfixed_mul(mclk_ff, temp_ff);
+			if (sclk_ff.full < temp_ff.full)
+				temp_ff.full = sclk_ff.full;
+
+			read_return_rate.full = temp_ff.full;
+
+			if (mode1) {
+				temp_ff.full = read_return_rate.full - disp_drain_rate.full;
+				time_disp1_drop_priority.full = rfixed_div(crit_point_ff, temp_ff);
+			} else {
+				time_disp1_drop_priority.full = 0;
+			}
+			crit_point_ff.full = disp_latency.full + time_disp1_drop_priority.full + disp_latency.full;
+			crit_point_ff.full = rfixed_mul(crit_point_ff, disp_drain_rate2);
+			crit_point_ff.full += rfixed_const_half(0);
+
+			critical_point2 = rfixed_trunc(crit_point_ff);
+
+			if (rdev->disp_priority == 2) {
+				critical_point2 = 0;
+			}
+
+			if (max_stop_req - critical_point2 < 4)
+				critical_point2 = 0;
+
+		}
+
+		if (critical_point2 == 0 && rdev->family == CHIP_R300) {
+			/* some R300 cards have problem with this set to 0 */
+			critical_point2 = 0x10;
+		}
+
+		WREG32(RADEON_GRPH2_BUFFER_CNTL, ((grph2_cntl & ~RADEON_GRPH_CRITICAL_POINT_MASK) |
+						  (critical_point2 << RADEON_GRPH_CRITICAL_POINT_SHIFT)));
+
+		if ((rdev->family == CHIP_RS400) ||
+		    (rdev->family == CHIP_RS480)) {
+#if 0
+			/* attempt to program RS400 disp2 regs correctly ??? */
+			temp = RREG32(RS400_DISP2_REQ_CNTL1);
+			temp &= ~(RS400_DISP2_START_REQ_LEVEL_MASK |
+				  RS400_DISP2_STOP_REQ_LEVEL_MASK);
+			WREG32(RS400_DISP2_REQ_CNTL1, (temp |
+						       (critical_point2 << RS400_DISP1_START_REQ_LEVEL_SHIFT) |
+						       (critical_point2 << RS400_DISP1_STOP_REQ_LEVEL_SHIFT)));
+			temp = RREG32(RS400_DISP2_REQ_CNTL2);
+			temp &= ~(RS400_DISP2_CRITICAL_POINT_START_MASK |
+				  RS400_DISP2_CRITICAL_POINT_STOP_MASK);
+			WREG32(RS400_DISP2_REQ_CNTL2, (temp |
+						       (critical_point2 << RS400_DISP2_CRITICAL_POINT_START_SHIFT) |
+						       (critical_point2 << RS400_DISP2_CRITICAL_POINT_STOP_SHIFT)));
+#endif
+			WREG32(RS400_DISP2_REQ_CNTL1, 0x105DC1CC);
+			WREG32(RS400_DISP2_REQ_CNTL2, 0x2749D000);
+			WREG32(RS400_DMIF_MEM_CNTL1,  0x29CA71DC);
+			WREG32(RS400_DISP1_REQ_CNTL1, 0x28FBC3AC);
+		}
+
+		DRM_DEBUG("GRPH2_BUFFER_CNTL from to %x\n",
+			  (unsigned int)RREG32(RADEON_GRPH2_BUFFER_CNTL));
+	}
+}

drivers/gpu/drm/radeon/r300.c

@@ -30,6 +30,8 @@
 #include "drm.h"
 #include "radeon_reg.h"
 #include "radeon.h"
+#include "radeon_drm.h"
+#include "radeon_share.h"
 
 /* r300,r350,rv350,rv370,rv380 depends on : */
 void r100_hdp_reset(struct radeon_device *rdev);
@@ -44,6 +46,7 @@ int r100_gui_wait_for_idle(struct radeon_device *rdev);
 int r100_cs_packet_parse(struct radeon_cs_parser *p,
 			 struct radeon_cs_packet *pkt,
 			 unsigned idx);
+int r100_cs_packet_parse_vline(struct radeon_cs_parser *p);
 int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
 			      struct radeon_cs_reloc **cs_reloc);
 int r100_cs_parse_packet0(struct radeon_cs_parser *p,
@@ -150,8 +153,13 @@ int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
 	}
-	addr = (((u32)addr) >> 8) | ((upper_32_bits(addr) & 0xff) << 4) | 0xC;
+	addr = (lower_32_bits(addr) >> 8) |
-	writel(cpu_to_le32(addr), ((void __iomem *)ptr) + (i * 4));
+	       ((upper_32_bits(addr) & 0xff) << 24) |
+	       0xc;
+	/* on x86 we want this to be CPU endian, on powerpc
+	 * on powerpc without HW swappers, it'll get swapped on way
+	 * into VRAM - so no need for cpu_to_le32 on VRAM tables */
+	writel(addr, ((void __iomem *)ptr) + (i * 4));
 	return 0;
 }
 
@@ -579,10 +587,8 @@ void r300_vram_info(struct radeon_device *rdev)
 	} else {
 		rdev->mc.vram_width = 64;
 	}
-	rdev->mc.vram_size = RREG32(RADEON_CONFIG_MEMSIZE);
 
-	rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0);
+	r100_vram_init_sizes(rdev);
-	rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0);
 }
 
 
@@ -970,7 +976,7 @@ static inline void r300_cs_track_clear(struct r300_cs_track *track)
 
 static const unsigned r300_reg_safe_bm[159] = {
 	0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
-	0xFFFFFFBF, 0xFFFFFFFF, 0xFFFFFFBF, 0xFFFFFFFF,
+	0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
 	0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
 	0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
 	0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
@@ -1019,7 +1025,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 	struct radeon_cs_reloc *reloc;
 	struct r300_cs_track *track;
 	volatile uint32_t *ib;
-	uint32_t tmp;
+	uint32_t tmp, tile_flags = 0;
 	unsigned i;
 	int r;
 
@@ -1027,6 +1033,16 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 	ib_chunk = &p->chunks[p->chunk_ib_idx];
 	track = (struct r300_cs_track*)p->track;
 	switch(reg) {
+	case AVIVO_D1MODE_VLINE_START_END:
+	case RADEON_CRTC_GUI_TRIG_VLINE:
+		r = r100_cs_packet_parse_vline(p);
+		if (r) {
+			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+					idx, reg);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+		break;
 	case RADEON_DST_PITCH_OFFSET:
 	case RADEON_SRC_PITCH_OFFSET:
 		r = r100_cs_packet_next_reloc(p, &reloc);
@@ -1038,7 +1054,19 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		}
 		tmp = ib_chunk->kdata[idx] & 0x003fffff;
 		tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
-		ib[idx] = (ib_chunk->kdata[idx] & 0xffc00000) | tmp;
+
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= RADEON_DST_TILE_MACRO;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+			if (reg == RADEON_SRC_PITCH_OFFSET) {
+				DRM_ERROR("Cannot src blit from microtiled surface\n");
+				r100_cs_dump_packet(p, pkt);
+				return -EINVAL;
+			}
+			tile_flags |= RADEON_DST_TILE_MICRO;
+		}
+		tmp |= tile_flags;
+		ib[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp;
 		break;
 	case R300_RB3D_COLOROFFSET0:
 	case R300_RB3D_COLOROFFSET1:
@@ -1127,6 +1155,23 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		/* RB3D_COLORPITCH1 */
 		/* RB3D_COLORPITCH2 */
 		/* RB3D_COLORPITCH3 */
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+				  idx, reg);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= R300_COLOR_TILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= R300_COLOR_MICROTILE_ENABLE;
+
+		tmp = ib_chunk->kdata[idx] & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
+
 		i = (reg - 0x4E38) >> 2;
 		track->cb[i].pitch = ib_chunk->kdata[idx] & 0x3FFE;
 		switch (((ib_chunk->kdata[idx] >> 21) & 0xF)) {
@@ -1182,6 +1227,23 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		break;
 	case 0x4F24:
 		/* ZB_DEPTHPITCH */
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+				  idx, reg);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= R300_DEPTHMACROTILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= R300_DEPTHMICROTILE_TILED;;
+
+		tmp = ib_chunk->kdata[idx] & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
+
 		track->zb.pitch = ib_chunk->kdata[idx] & 0x3FFC;
 		break;
 	case 0x4104:

drivers/gpu/drm/radeon/r300_reg.h

@@ -27,7 +27,9 @@
 #ifndef _R300_REG_H_
 #define _R300_REG_H_
 
-
+#define R300_SURF_TILE_MACRO (1<<16)
+#define R300_SURF_TILE_MICRO (2<<16)
+#define R300_SURF_TILE_BOTH (3<<16)
 
 
 #define R300_MC_INIT_MISC_LAT_TIMER	0x180

drivers/gpu/drm/radeon/r500_reg.h

@@ -445,6 +445,7 @@
 #define AVIVO_D1MODE_DATA_FORMAT                0x6528
 #       define AVIVO_D1MODE_INTERLEAVE_EN       (1 << 0)
 #define AVIVO_D1MODE_DESKTOP_HEIGHT             0x652C
+#define AVIVO_D1MODE_VLINE_START_END            0x6538
 #define AVIVO_D1MODE_VIEWPORT_START             0x6580
 #define AVIVO_D1MODE_VIEWPORT_SIZE              0x6584
 #define AVIVO_D1MODE_EXT_OVERSCAN_LEFT_RIGHT    0x6588
@@ -496,6 +497,7 @@
 #define AVIVO_D2CUR_SIZE                        0x6c10
 #define AVIVO_D2CUR_POSITION                    0x6c14
 
+#define AVIVO_D2MODE_VLINE_START_END            0x6d38
 #define AVIVO_D2MODE_VIEWPORT_START             0x6d80
 #define AVIVO_D2MODE_VIEWPORT_SIZE              0x6d84
 #define AVIVO_D2MODE_EXT_OVERSCAN_LEFT_RIGHT    0x6d88

drivers/gpu/drm/radeon/r520.c

@@ -28,6 +28,7 @@
 #include "drmP.h"
 #include "radeon_reg.h"
 #include "radeon.h"
+#include "radeon_share.h"
 
 /* r520,rv530,rv560,rv570,r580 depends on : */
 void r100_hdp_reset(struct radeon_device *rdev);
@@ -94,8 +95,8 @@ int r520_mc_init(struct radeon_device *rdev)
 		       "programming pipes. Bad things might happen.\n");
 	}
 	/* Write VRAM size in case we are limiting it */
-	WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size);
+	WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size);
-	tmp = rdev->mc.vram_location + rdev->mc.vram_size - 1;
+	tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1;
 	tmp = REG_SET(R520_MC_FB_TOP, tmp >> 16);
 	tmp |= REG_SET(R520_MC_FB_START, rdev->mc.vram_location >> 16);
 	WREG32_MC(R520_MC_FB_LOCATION, tmp);
@@ -226,9 +227,20 @@ static void r520_vram_get_type(struct radeon_device *rdev)
 
 void r520_vram_info(struct radeon_device *rdev)
 {
-	r520_vram_get_type(rdev);
+	fixed20_12 a;
-	rdev->mc.vram_size = RREG32(RADEON_CONFIG_MEMSIZE);
 
-	rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0);
+	r520_vram_get_type(rdev);
-	rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0);
+
+	r100_vram_init_sizes(rdev);
+	/* FIXME: we should enforce default clock in case GPU is not in
+	 * default setup
+	 */
+	a.full = rfixed_const(100);
+	rdev->pm.sclk.full = rfixed_const(rdev->clock.default_sclk);
+	rdev->pm.sclk.full = rfixed_div(rdev->pm.sclk, a);
+}
+
+void r520_bandwidth_update(struct radeon_device *rdev)
+{
+	rv515_bandwidth_avivo_update(rdev);
 }

drivers/gpu/drm/radeon/r600.c

@@ -67,7 +67,7 @@ int r600_mc_init(struct radeon_device *rdev)
 		       "programming pipes. Bad things might happen.\n");
 	}
 
-	tmp = rdev->mc.vram_location + rdev->mc.vram_size - 1;
+	tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1;
 	tmp = REG_SET(R600_MC_FB_TOP, tmp >> 24);
 	tmp |= REG_SET(R600_MC_FB_BASE, rdev->mc.vram_location >> 24);
 	WREG32(R600_MC_VM_FB_LOCATION, tmp);
@@ -140,7 +140,8 @@ void r600_vram_get_type(struct radeon_device *rdev)
 void r600_vram_info(struct radeon_device *rdev)
 {
 	r600_vram_get_type(rdev);
-	rdev->mc.vram_size = RREG32(R600_CONFIG_MEMSIZE);
+	rdev->mc.real_vram_size = RREG32(R600_CONFIG_MEMSIZE);
+	rdev->mc.mc_vram_size = rdev->mc.real_vram_size;
 
 	/* Could aper size report 0 ? */
 	rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0);

drivers/gpu/drm/radeon/radeon.h

@@ -64,6 +64,7 @@ extern int radeon_agpmode;
 extern int radeon_vram_limit;
 extern int radeon_gart_size;
 extern int radeon_benchmarking;
+extern int radeon_testing;
 extern int radeon_connector_table;
 
 /*
@@ -113,6 +114,7 @@ enum radeon_family {
 	CHIP_RV770,
 	CHIP_RV730,
 	CHIP_RV710,
+	CHIP_RS880,
 	CHIP_LAST,
 };
 
@@ -201,6 +203,14 @@ int radeon_fence_wait_last(struct radeon_device *rdev);
 struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence);
 void radeon_fence_unref(struct radeon_fence **fence);
 
+/*
+ * Tiling registers
+ */
+struct radeon_surface_reg {
+	struct radeon_object *robj;
+};
+
+#define RADEON_GEM_MAX_SURFACES 8
 
 /*
  * Radeon buffer.
@@ -213,6 +223,7 @@ struct radeon_object_list {
 	uint64_t		gpu_offset;
 	unsigned		rdomain;
 	unsigned		wdomain;
+	uint32_t                tiling_flags;
 };
 
 int radeon_object_init(struct radeon_device *rdev);
@@ -242,8 +253,15 @@ void radeon_object_list_clean(struct list_head *head);
 int radeon_object_fbdev_mmap(struct radeon_object *robj,
 			     struct vm_area_struct *vma);
 unsigned long radeon_object_size(struct radeon_object *robj);
-
+void radeon_object_clear_surface_reg(struct radeon_object *robj);
-
+int radeon_object_check_tiling(struct radeon_object *robj, bool has_moved,
+			       bool force_drop);
+void radeon_object_set_tiling_flags(struct radeon_object *robj,
+				    uint32_t tiling_flags, uint32_t pitch);
+void radeon_object_get_tiling_flags(struct radeon_object *robj, uint32_t *tiling_flags, uint32_t *pitch);
+void radeon_bo_move_notify(struct ttm_buffer_object *bo,
+			   struct ttm_mem_reg *mem);
+void radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 /*
  * GEM objects.
  */
@@ -315,8 +333,11 @@ struct radeon_mc {
 	unsigned		gtt_location;
 	unsigned		gtt_size;
 	unsigned		vram_location;
-	unsigned		vram_size;
+	/* for some chips with <= 32MB we need to lie
+	 * about vram size near mc fb location */
+	unsigned		mc_vram_size;
 	unsigned		vram_width;
+	unsigned		real_vram_size;
 	int			vram_mtrr;
 	bool			vram_is_ddr;
 };
@@ -474,6 +495,39 @@ struct radeon_wb {
 	uint64_t		gpu_addr;
 };
 
+/**
+ * struct radeon_pm - power management datas
+ * @max_bandwidth:      maximum bandwidth the gpu has (MByte/s)
+ * @igp_sideport_mclk:  sideport memory clock Mhz (rs690,rs740,rs780,rs880)
+ * @igp_system_mclk:    system clock Mhz (rs690,rs740,rs780,rs880)
+ * @igp_ht_link_clk:    ht link clock Mhz (rs690,rs740,rs780,rs880)
+ * @igp_ht_link_width:  ht link width in bits (rs690,rs740,rs780,rs880)
+ * @k8_bandwidth:       k8 bandwidth the gpu has (MByte/s) (IGP)
+ * @sideport_bandwidth: sideport bandwidth the gpu has (MByte/s) (IGP)
+ * @ht_bandwidth:       ht bandwidth the gpu has (MByte/s) (IGP)
+ * @core_bandwidth:     core GPU bandwidth the gpu has (MByte/s) (IGP)
+ * @sclk:          	GPU clock Mhz (core bandwith depends of this clock)
+ * @needed_bandwidth:   current bandwidth needs
+ *
+ * It keeps track of various data needed to take powermanagement decision.
+ * Bandwith need is used to determine minimun clock of the GPU and memory.
+ * Equation between gpu/memory clock and available bandwidth is hw dependent
+ * (type of memory, bus size, efficiency, ...)
+ */
+struct radeon_pm {
+	fixed20_12		max_bandwidth;
+	fixed20_12		igp_sideport_mclk;
+	fixed20_12		igp_system_mclk;
+	fixed20_12		igp_ht_link_clk;
+	fixed20_12		igp_ht_link_width;
+	fixed20_12		k8_bandwidth;
+	fixed20_12		sideport_bandwidth;
+	fixed20_12		ht_bandwidth;
+	fixed20_12		core_bandwidth;
+	fixed20_12		sclk;
+	fixed20_12		needed_bandwidth;
+};
+
 
 /*
  * Benchmarking
@@ -481,6 +535,12 @@ struct radeon_wb {
 void radeon_benchmark(struct radeon_device *rdev);
 
 
+/*
+ * Testing
+ */
+void radeon_test_moves(struct radeon_device *rdev);
+
+
 /*
  * Debugfs
  */
@@ -535,6 +595,11 @@ struct radeon_asic {
 	void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock);
 	void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
 	void (*set_clock_gating)(struct radeon_device *rdev, int enable);
+	int (*set_surface_reg)(struct radeon_device *rdev, int reg,
+			       uint32_t tiling_flags, uint32_t pitch,
+			       uint32_t offset, uint32_t obj_size);
+	int (*clear_surface_reg)(struct radeon_device *rdev, int reg);
+	void (*bandwidth_update)(struct radeon_device *rdev);
 };
 
 union radeon_asic_config {
@@ -566,6 +631,10 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data,
 int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *filp);
 int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp);
+int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *filp);
+int radeon_gem_get_tiling_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *filp);
 
 
 /*
@@ -594,8 +663,8 @@ struct radeon_device {
 	struct radeon_object		*fbdev_robj;
 	struct radeon_framebuffer	*fbdev_rfb;
 	/* Register mmio */
-	unsigned long			rmmio_base;
+	resource_size_t			rmmio_base;
-	unsigned long			rmmio_size;
+	resource_size_t			rmmio_size;
 	void				*rmmio;
 	radeon_rreg_t			mm_rreg;
 	radeon_wreg_t			mm_wreg;
@@ -619,11 +688,14 @@ struct radeon_device {
 	struct radeon_irq		irq;
 	struct radeon_asic		*asic;
 	struct radeon_gem		gem;
+	struct radeon_pm		pm;
 	struct mutex			cs_mutex;
 	struct radeon_wb		wb;
 	bool				gpu_lockup;
 	bool				shutdown;
 	bool				suspend;
+	bool				need_dma32;
+	struct radeon_surface_reg surface_regs[RADEON_GEM_MAX_SURFACES];
 };
 
 int radeon_device_init(struct radeon_device *rdev,
@@ -670,6 +742,8 @@ void r100_pll_errata_after_index(struct radeon_device *rdev);
 /*
  * ASICs helpers.
  */
+#define ASIC_IS_RN50(rdev) ((rdev->pdev->device == 0x515e) || \
+			    (rdev->pdev->device == 0x5969))
 #define ASIC_IS_RV100(rdev) ((rdev->family == CHIP_RV100) || \
 		(rdev->family == CHIP_RV200) || \
 		(rdev->family == CHIP_RS100) || \
@@ -796,5 +870,8 @@ static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v)
 #define radeon_set_memory_clock(rdev, e) (rdev)->asic->set_engine_clock((rdev), (e))
 #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->set_pcie_lanes((rdev), (l))
 #define radeon_set_clock_gating(rdev, e) (rdev)->asic->set_clock_gating((rdev), (e))
+#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->set_surface_reg((rdev), (r), (f), (p), (o), (s)))
+#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->clear_surface_reg((rdev), (r)))
+#define radeon_bandwidth_update(rdev) (rdev)->asic->bandwidth_update((rdev))
 
 #endif

drivers/gpu/drm/radeon/radeon_asic.h

@@ -71,6 +71,11 @@ int r100_copy_blit(struct radeon_device *rdev,
 		   uint64_t dst_offset,
 		   unsigned num_pages,
 		   struct radeon_fence *fence);
+int r100_set_surface_reg(struct radeon_device *rdev, int reg,
+			 uint32_t tiling_flags, uint32_t pitch,
+			 uint32_t offset, uint32_t obj_size);
+int r100_clear_surface_reg(struct radeon_device *rdev, int reg);
+void r100_bandwidth_update(struct radeon_device *rdev);
 
 static struct radeon_asic r100_asic = {
 	.init = &r100_init,
@@ -100,6 +105,9 @@ static struct radeon_asic r100_asic = {
 	.set_memory_clock = NULL,
 	.set_pcie_lanes = NULL,
 	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
 };
 
 
@@ -128,6 +136,7 @@ int r300_copy_dma(struct radeon_device *rdev,
 		  uint64_t dst_offset,
 		  unsigned num_pages,
 		  struct radeon_fence *fence);
+
 static struct radeon_asic r300_asic = {
 	.init = &r300_init,
 	.errata = &r300_errata,
@@ -156,6 +165,9 @@ static struct radeon_asic r300_asic = {
 	.set_memory_clock = NULL,
 	.set_pcie_lanes = &rv370_set_pcie_lanes,
 	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
 };
 
 /*
@@ -193,6 +205,9 @@ static struct radeon_asic r420_asic = {
 	.set_memory_clock = &radeon_atom_set_memory_clock,
 	.set_pcie_lanes = &rv370_set_pcie_lanes,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
 };
 
 
@@ -237,6 +252,9 @@ static struct radeon_asic rs400_asic = {
 	.set_memory_clock = NULL,
 	.set_pcie_lanes = NULL,
 	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
 };
 
 
@@ -254,6 +272,7 @@ void rs600_gart_tlb_flush(struct radeon_device *rdev);
 int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr);
 uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+void rs600_bandwidth_update(struct radeon_device *rdev);
 static struct radeon_asic rs600_asic = {
 	.init = &r300_init,
 	.errata = &rs600_errata,
@@ -282,6 +301,7 @@ static struct radeon_asic rs600_asic = {
 	.set_memory_clock = &radeon_atom_set_memory_clock,
 	.set_pcie_lanes = NULL,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.bandwidth_update = &rs600_bandwidth_update,
 };
 
 
@@ -294,6 +314,7 @@ int rs690_mc_init(struct radeon_device *rdev);
 void rs690_mc_fini(struct radeon_device *rdev);
 uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs690_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+void rs690_bandwidth_update(struct radeon_device *rdev);
 static struct radeon_asic rs690_asic = {
 	.init = &r300_init,
 	.errata = &rs690_errata,
@@ -322,6 +343,9 @@ static struct radeon_asic rs690_asic = {
 	.set_memory_clock = &radeon_atom_set_memory_clock,
 	.set_pcie_lanes = NULL,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rs690_bandwidth_update,
 };
 
 
@@ -339,6 +363,7 @@ void rv515_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 void rv515_ring_start(struct radeon_device *rdev);
 uint32_t rv515_pcie_rreg(struct radeon_device *rdev, uint32_t reg);
 void rv515_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+void rv515_bandwidth_update(struct radeon_device *rdev);
 static struct radeon_asic rv515_asic = {
 	.init = &rv515_init,
 	.errata = &rv515_errata,
@@ -367,6 +392,9 @@ static struct radeon_asic rv515_asic = {
 	.set_memory_clock = &radeon_atom_set_memory_clock,
 	.set_pcie_lanes = &rv370_set_pcie_lanes,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rv515_bandwidth_update,
 };
 
 
@@ -377,6 +405,7 @@ void r520_errata(struct radeon_device *rdev);
 void r520_vram_info(struct radeon_device *rdev);
 int r520_mc_init(struct radeon_device *rdev);
 void r520_mc_fini(struct radeon_device *rdev);
+void r520_bandwidth_update(struct radeon_device *rdev);
 static struct radeon_asic r520_asic = {
 	.init = &rv515_init,
 	.errata = &r520_errata,
@@ -405,6 +434,9 @@ static struct radeon_asic r520_asic = {
 	.set_memory_clock = &radeon_atom_set_memory_clock,
 	.set_pcie_lanes = &rv370_set_pcie_lanes,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r520_bandwidth_update,
 };
 
 /*

drivers/gpu/drm/radeon/radeon_atombios.c

@@ -103,7 +103,8 @@ static inline struct radeon_i2c_bus_rec radeon_lookup_gpio(struct drm_device
 static bool radeon_atom_apply_quirks(struct drm_device *dev,
 				     uint32_t supported_device,
 				     int *connector_type,
-				     struct radeon_i2c_bus_rec *i2c_bus)
+				     struct radeon_i2c_bus_rec *i2c_bus,
+				     uint8_t *line_mux)
 {
 
 	/* Asus M2A-VM HDMI board lists the DVI port as HDMI */
@@ -127,8 +128,10 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 	if ((dev->pdev->device == 0x5653) &&
 	    (dev->pdev->subsystem_vendor == 0x1462) &&
 	    (dev->pdev->subsystem_device == 0x0291)) {
-		if (*connector_type == DRM_MODE_CONNECTOR_LVDS)
+		if (*connector_type == DRM_MODE_CONNECTOR_LVDS) {
 			i2c_bus->valid = false;
+			*line_mux = 53;
+		}
 	}
 
 	/* Funky macbooks */
@@ -526,7 +529,7 @@ bool radeon_get_atom_connector_info_from_supported_devices_table(struct
 
 		if (!radeon_atom_apply_quirks
 		    (dev, (1 << i), &bios_connectors[i].connector_type,
-		     &bios_connectors[i].ddc_bus))
+		     &bios_connectors[i].ddc_bus, &bios_connectors[i].line_mux))
 			continue;
 
 		bios_connectors[i].valid = true;

drivers/gpu/drm/radeon/radeon_benchmark.c

@@ -63,7 +63,7 @@ void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize,
 		if (r) {
 			goto out_cleanup;
 		}
-		r = radeon_copy_dma(rdev, saddr, daddr, size >> 14, fence);
+		r = radeon_copy_dma(rdev, saddr, daddr, size / 4096, fence);
 		if (r) {
 			goto out_cleanup;
 		}
@@ -88,7 +88,7 @@ void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize,
 		if (r) {
 			goto out_cleanup;
 		}
-		r = radeon_copy_blit(rdev, saddr, daddr, size >> 14, fence);
+		r = radeon_copy_blit(rdev, saddr, daddr, size / 4096, fence);
 		if (r) {
 			goto out_cleanup;
 		}

drivers/gpu/drm/radeon/radeon_cs.c

@@ -127,17 +127,23 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 				       sizeof(struct drm_radeon_cs_chunk))) {
 			return -EFAULT;
 		}
+		p->chunks[i].length_dw = user_chunk.length_dw;
+		p->chunks[i].kdata = NULL;
 		p->chunks[i].chunk_id = user_chunk.chunk_id;
+
 		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) {
 			p->chunk_relocs_idx = i;
 		}
 		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
 			p->chunk_ib_idx = i;
+			/* zero length IB isn't useful */
+			if (p->chunks[i].length_dw == 0)
+				return -EINVAL;
 		}
+
 		p->chunks[i].length_dw = user_chunk.length_dw;
 		cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data;
 
-		p->chunks[i].kdata = NULL;
 		size = p->chunks[i].length_dw * sizeof(uint32_t);
 		p->chunks[i].kdata = kzalloc(size, GFP_KERNEL);
 		if (p->chunks[i].kdata == NULL) {

drivers/gpu/drm/radeon/radeon_cursor.c

@@ -111,9 +111,11 @@ static void radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj,
 
 	if (ASIC_IS_AVIVO(rdev))
 		WREG32(AVIVO_D1CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset, gpu_addr);
-	else
+	else {
+		radeon_crtc->legacy_cursor_offset = gpu_addr - radeon_crtc->legacy_display_base_addr;
 		/* offset is from DISP(2)_BASE_ADDRESS */
-		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, gpu_addr);
+		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, radeon_crtc->legacy_cursor_offset);
+	}
 }
 
 int radeon_crtc_cursor_set(struct drm_crtc *crtc,
@@ -245,6 +247,9 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 		       (RADEON_CUR_LOCK
 			| ((xorigin ? 0 : x) << 16)
 			| (yorigin ? 0 : y)));
+		/* offset is from DISP(2)_BASE_ADDRESS */
+		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, (radeon_crtc->legacy_cursor_offset +
+								      (yorigin * 256)));
 	}
 	radeon_lock_cursor(crtc, false);
 

drivers/gpu/drm/radeon/radeon_device.c

@@ -48,6 +48,8 @@ static void radeon_surface_init(struct radeon_device *rdev)
 			       i * (RADEON_SURFACE1_INFO - RADEON_SURFACE0_INFO),
 			       0);
 		}
+		/* enable surfaces */
+		WREG32(RADEON_SURFACE_CNTL, 0);
 	}
 }
 
@@ -119,7 +121,7 @@ int radeon_mc_setup(struct radeon_device *rdev)
 	if (rdev->mc.vram_location != 0xFFFFFFFFUL) {
 		/* vram location was already setup try to put gtt after
 		 * if it fits */
-		tmp = rdev->mc.vram_location + rdev->mc.vram_size;
+		tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size;
 		tmp = (tmp + rdev->mc.gtt_size - 1) & ~(rdev->mc.gtt_size - 1);
 		if ((0xFFFFFFFFUL - tmp) >= rdev->mc.gtt_size) {
 			rdev->mc.gtt_location = tmp;
@@ -134,13 +136,13 @@ int radeon_mc_setup(struct radeon_device *rdev)
 	} else if (rdev->mc.gtt_location != 0xFFFFFFFFUL) {
 		/* gtt location was already setup try to put vram before
 		 * if it fits */
-		if (rdev->mc.vram_size < rdev->mc.gtt_location) {
+		if (rdev->mc.mc_vram_size < rdev->mc.gtt_location) {
 			rdev->mc.vram_location = 0;
 		} else {
 			tmp = rdev->mc.gtt_location + rdev->mc.gtt_size;
-			tmp += (rdev->mc.vram_size - 1);
+			tmp += (rdev->mc.mc_vram_size - 1);
-			tmp &= ~(rdev->mc.vram_size - 1);
+			tmp &= ~(rdev->mc.mc_vram_size - 1);
-			if ((0xFFFFFFFFUL - tmp) >= rdev->mc.vram_size) {
+			if ((0xFFFFFFFFUL - tmp) >= rdev->mc.mc_vram_size) {
 				rdev->mc.vram_location = tmp;
 			} else {
 				printk(KERN_ERR "[drm] vram too big to fit "
@@ -150,12 +152,14 @@ int radeon_mc_setup(struct radeon_device *rdev)
 		}
 	} else {
 		rdev->mc.vram_location = 0;
-		rdev->mc.gtt_location = rdev->mc.vram_size;
+		rdev->mc.gtt_location = rdev->mc.mc_vram_size;
 	}
-	DRM_INFO("radeon: VRAM %uM\n", rdev->mc.vram_size >> 20);
+	DRM_INFO("radeon: VRAM %uM\n", rdev->mc.real_vram_size >> 20);
 	DRM_INFO("radeon: VRAM from 0x%08X to 0x%08X\n",
 		 rdev->mc.vram_location,
-		 rdev->mc.vram_location + rdev->mc.vram_size - 1);
+		 rdev->mc.vram_location + rdev->mc.mc_vram_size - 1);
+	if (rdev->mc.real_vram_size != rdev->mc.mc_vram_size)
+		DRM_INFO("radeon: VRAM less than aperture workaround enabled\n");
 	DRM_INFO("radeon: GTT %uM\n", rdev->mc.gtt_size >> 20);
 	DRM_INFO("radeon: GTT from 0x%08X to 0x%08X\n",
 		 rdev->mc.gtt_location,
@@ -450,6 +454,7 @@ int radeon_device_init(struct radeon_device *rdev,
 		       uint32_t flags)
 {
 	int r, ret;
+	int dma_bits;
 
 	DRM_INFO("radeon: Initializing kernel modesetting.\n");
 	rdev->shutdown = false;
@@ -492,8 +497,20 @@ int radeon_device_init(struct radeon_device *rdev,
 		return r;
 	}
 
-	/* Report DMA addressing limitation */
+	/* set DMA mask + need_dma32 flags.
-	r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(32));
+	 * PCIE - can handle 40-bits.
+	 * IGP - can handle 40-bits (in theory)
+	 * AGP - generally dma32 is safest
+	 * PCI - only dma32
+	 */
+	rdev->need_dma32 = false;
+	if (rdev->flags & RADEON_IS_AGP)
+		rdev->need_dma32 = true;
+	if (rdev->flags & RADEON_IS_PCI)
+		rdev->need_dma32 = true;
+
+	dma_bits = rdev->need_dma32 ? 32 : 40;
+	r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits));
 	if (r) {
 		printk(KERN_WARNING "radeon: No suitable DMA available.\n");
 	}
@@ -546,27 +563,22 @@ int radeon_device_init(struct radeon_device *rdev,
 			radeon_combios_asic_init(rdev->ddev);
 		}
 	}
-	/* Get vram informations */
-	radeon_vram_info(rdev);
-	/* Device is severly broken if aper size > vram size.
-	 * for RN50/M6/M7 - Novell bug 204882 ?
-	 */
-	if (rdev->mc.vram_size < rdev->mc.aper_size) {
-		rdev->mc.aper_size = rdev->mc.vram_size;
-	}
-	/* Add an MTRR for the VRAM */
-	rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
-				      MTRR_TYPE_WRCOMB, 1);
-	DRM_INFO("Detected VRAM RAM=%uM, BAR=%uM\n",
-		 rdev->mc.vram_size >> 20,
-		 (unsigned)rdev->mc.aper_size >> 20);
-	DRM_INFO("RAM width %dbits %cDR\n",
-		 rdev->mc.vram_width, rdev->mc.vram_is_ddr ? 'D' : 'S');
 	/* Initialize clocks */
 	r = radeon_clocks_init(rdev);
 	if (r) {
 		return r;
 	}
+	/* Get vram informations */
+	radeon_vram_info(rdev);
+
+	/* Add an MTRR for the VRAM */
+	rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
+				      MTRR_TYPE_WRCOMB, 1);
+	DRM_INFO("Detected VRAM RAM=%uM, BAR=%uM\n",
+		 rdev->mc.real_vram_size >> 20,
+		 (unsigned)rdev->mc.aper_size >> 20);
+	DRM_INFO("RAM width %dbits %cDR\n",
+		 rdev->mc.vram_width, rdev->mc.vram_is_ddr ? 'D' : 'S');
 	/* Initialize memory controller (also test AGP) */
 	r = radeon_mc_init(rdev);
 	if (r) {
@@ -626,6 +638,9 @@ int radeon_device_init(struct radeon_device *rdev,
 	if (!ret) {
 		DRM_INFO("radeon: kernel modesetting successfully initialized.\n");
 	}
+	if (radeon_testing) {
+		radeon_test_moves(rdev);
+	}
 	if (radeon_benchmarking) {
 		radeon_benchmark(rdev);
 	}

drivers/gpu/drm/radeon/radeon_display.c

@@ -187,6 +187,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 
 	drm_mode_crtc_set_gamma_size(&radeon_crtc->base, 256);
 	radeon_crtc->crtc_id = index;
+	rdev->mode_info.crtcs[index] = radeon_crtc;
 
 	radeon_crtc->mode_set.crtc = &radeon_crtc->base;
 	radeon_crtc->mode_set.connectors = (struct drm_connector **)(radeon_crtc + 1);
@@ -491,6 +492,10 @@ void radeon_compute_pll(struct radeon_pll *pll,
 					tmp += (uint64_t)pll->reference_freq * 1000 * frac_feedback_div;
 					current_freq = radeon_div(tmp, ref_div * post_div);
 
+					if (flags & RADEON_PLL_PREFER_CLOSEST_LOWER) {
+						error = freq - current_freq;
+						error = error < 0 ? 0xffffffff : error;
+					} else
 						error = abs(current_freq - freq);
 					vco_diff = abs(vco - best_vco);
 
@@ -657,36 +662,51 @@ void radeon_modeset_fini(struct radeon_device *rdev)
 	}
 }
 
-void radeon_init_disp_bandwidth(struct drm_device *dev)
+bool radeon_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
 {
-	struct radeon_device *rdev = dev->dev_private;
+	struct drm_device *dev = crtc->dev;
-	struct drm_display_mode *modes[2];
+	struct drm_encoder *encoder;
-	int pixel_bytes[2];
-	struct drm_crtc *crtc;
-
-	pixel_bytes[0] = pixel_bytes[1] = 0;
-	modes[0] = modes[1] = NULL;
-
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct radeon_encoder *radeon_encoder;
+	bool first = true;
 
-		if (crtc->enabled && crtc->fb) {
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-			modes[radeon_crtc->crtc_id] = &crtc->mode;
+		radeon_encoder = to_radeon_encoder(encoder);
-			pixel_bytes[radeon_crtc->crtc_id] = crtc->fb->bits_per_pixel / 8;
+		if (encoder->crtc != crtc)
-		}
+			continue;
-	}
+		if (first) {
-
+			radeon_crtc->rmx_type = radeon_encoder->rmx_type;
-	if (ASIC_IS_AVIVO(rdev)) {
+			radeon_crtc->devices = radeon_encoder->devices;
-		radeon_init_disp_bw_avivo(dev,
+			memcpy(&radeon_crtc->native_mode,
-					  modes[0],
+				&radeon_encoder->native_mode,
-					  pixel_bytes[0],
+				sizeof(struct radeon_native_mode));
-					  modes[1],
+			first = false;
-					  pixel_bytes[1]);
 		} else {
-		radeon_init_disp_bw_legacy(dev,
+			if (radeon_crtc->rmx_type != radeon_encoder->rmx_type) {
-					   modes[0],
+				/* WARNING: Right now this can't happen but
-					   pixel_bytes[0],
+				 * in the future we need to check that scaling
-					   modes[1],
+				 * are consistent accross different encoder
-					   pixel_bytes[1]);
+				 * (ie all encoder can work with the same
+				 *  scaling).
+				 */
+				DRM_ERROR("Scaling not consistent accross encoder.\n");
+				return false;
 			}
+		}
+	}
+	if (radeon_crtc->rmx_type != RMX_OFF) {
+		fixed20_12 a, b;
+		a.full = rfixed_const(crtc->mode.vdisplay);
+		b.full = rfixed_const(radeon_crtc->native_mode.panel_xres);
+		radeon_crtc->vsc.full = rfixed_div(a, b);
+		a.full = rfixed_const(crtc->mode.hdisplay);
+		b.full = rfixed_const(radeon_crtc->native_mode.panel_yres);
+		radeon_crtc->hsc.full = rfixed_div(a, b);
+	} else {
+		radeon_crtc->vsc.full = rfixed_const(1);
+		radeon_crtc->hsc.full = rfixed_const(1);
+	}
+	return true;
 }

drivers/gpu/drm/radeon/radeon_drv.c

@@ -89,6 +89,7 @@ int radeon_agpmode = 0;
 int radeon_vram_limit = 0;
 int radeon_gart_size = 512; /* default gart size */
 int radeon_benchmarking = 0;
+int radeon_testing = 0;
 int radeon_connector_table = 0;
 #endif
 
@@ -117,6 +118,9 @@ module_param_named(gartsize, radeon_gart_size, int, 0600);
 MODULE_PARM_DESC(benchmark, "Run benchmark");
 module_param_named(benchmark, radeon_benchmarking, int, 0444);
 
+MODULE_PARM_DESC(test, "Run tests");
+module_param_named(test, radeon_testing, int, 0444);
+
 MODULE_PARM_DESC(connector_table, "Force connector table");
 module_param_named(connector_table, radeon_connector_table, int, 0444);
 #endif