diff options
Diffstat (limited to 'arch/x86/include/asm')
24 files changed, 313 insertions, 316 deletions
diff --git a/arch/x86/include/asm/asm.h b/arch/x86/include/asm/asm.h index 9412d6558c88..1c2d247f65ce 100644 --- a/arch/x86/include/asm/asm.h +++ b/arch/x86/include/asm/asm.h @@ -4,11 +4,9 @@ #ifdef __ASSEMBLY__ # define __ASM_FORM(x) x # define __ASM_FORM_COMMA(x) x, -# define __ASM_EX_SEC .section __ex_table, "a" #else # define __ASM_FORM(x) " " #x " " # define __ASM_FORM_COMMA(x) " " #x "," -# define __ASM_EX_SEC " .section __ex_table,\"a\"\n" #endif #ifdef CONFIG_X86_32 @@ -42,17 +40,33 @@ /* Exception table entry */ #ifdef __ASSEMBLY__ -# define _ASM_EXTABLE(from,to) \ - __ASM_EX_SEC ; \ - _ASM_ALIGN ; \ - _ASM_PTR from , to ; \ - .previous +# define _ASM_EXTABLE(from,to) \ + .pushsection "__ex_table","a" ; \ + .balign 8 ; \ + .long (from) - . ; \ + .long (to) - . ; \ + .popsection + +# define _ASM_EXTABLE_EX(from,to) \ + .pushsection "__ex_table","a" ; \ + .balign 8 ; \ + .long (from) - . ; \ + .long (to) - . + 0x7ffffff0 ; \ + .popsection #else -# define _ASM_EXTABLE(from,to) \ - __ASM_EX_SEC \ - _ASM_ALIGN "\n" \ - _ASM_PTR #from "," #to "\n" \ - " .previous\n" +# define _ASM_EXTABLE(from,to) \ + " .pushsection \"__ex_table\",\"a\"\n" \ + " .balign 8\n" \ + " .long (" #from ") - .\n" \ + " .long (" #to ") - .\n" \ + " .popsection\n" + +# define _ASM_EXTABLE_EX(from,to) \ + " .pushsection \"__ex_table\",\"a\"\n" \ + " .balign 8\n" \ + " .long (" #from ") - .\n" \ + " .long (" #to ") - . + 0x7ffffff0\n" \ + " .popsection\n" #endif #endif /* _ASM_X86_ASM_H */ diff --git a/arch/x86/include/asm/bootparam.h b/arch/x86/include/asm/bootparam.h index 2f90c51cc49d..eb45aa6b1f27 100644 --- a/arch/x86/include/asm/bootparam.h +++ b/arch/x86/include/asm/bootparam.h @@ -112,7 +112,8 @@ struct boot_params { __u8 e820_entries; /* 0x1e8 */ __u8 eddbuf_entries; /* 0x1e9 */ __u8 edd_mbr_sig_buf_entries; /* 0x1ea */ - __u8 _pad6[6]; /* 0x1eb */ + __u8 kbd_status; /* 0x1eb */ + __u8 _pad6[5]; /* 0x1ec */ struct setup_header hdr; /* setup header */ /* 0x1f1 */ __u8 _pad7[0x290-0x1f1-sizeof(struct setup_header)]; __u32 edd_mbr_sig_buffer[EDD_MBR_SIG_MAX]; /* 0x290 */ diff --git a/arch/x86/include/asm/device.h b/arch/x86/include/asm/device.h index 63a2a03d7d51..93e1c55f14ab 100644 --- a/arch/x86/include/asm/device.h +++ b/arch/x86/include/asm/device.h @@ -5,8 +5,8 @@ struct dev_archdata { #ifdef CONFIG_ACPI void *acpi_handle; #endif -#ifdef CONFIG_X86_64 -struct dma_map_ops *dma_ops; +#ifdef CONFIG_X86_DEV_DMA_OPS + struct dma_map_ops *dma_ops; #endif #if defined(CONFIG_INTEL_IOMMU) || defined(CONFIG_AMD_IOMMU) void *iommu; /* hook for IOMMU specific extension */ diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h index 4b4331d71935..61c0bd25845a 100644 --- a/arch/x86/include/asm/dma-mapping.h +++ b/arch/x86/include/asm/dma-mapping.h @@ -30,7 +30,7 @@ extern struct dma_map_ops *dma_ops; static inline struct dma_map_ops *get_dma_ops(struct device *dev) { -#ifdef CONFIG_X86_32 +#ifndef CONFIG_X86_DEV_DMA_OPS return dma_ops; #else if (unlikely(!dev) || !dev->archdata.dma_ops) @@ -62,6 +62,12 @@ extern void *dma_generic_alloc_coherent(struct device *dev, size_t size, dma_addr_t *dma_addr, gfp_t flag, struct dma_attrs *attrs); +#ifdef CONFIG_X86_DMA_REMAP /* Platform code defines bridge-specific code */ +extern bool dma_capable(struct device *dev, dma_addr_t addr, size_t size); +extern dma_addr_t phys_to_dma(struct device *dev, phys_addr_t paddr); +extern phys_addr_t dma_to_phys(struct device *dev, dma_addr_t daddr); +#else + static inline bool dma_capable(struct device *dev, dma_addr_t addr, size_t size) { if (!dev->dma_mask) @@ -79,6 +85,7 @@ static inline phys_addr_t dma_to_phys(struct device *dev, dma_addr_t daddr) { return daddr; } +#endif /* CONFIG_X86_DMA_REMAP */ static inline void dma_cache_sync(struct device *dev, void *vaddr, size_t size, diff --git a/arch/x86/include/asm/kbdleds.h b/arch/x86/include/asm/kbdleds.h new file mode 100644 index 000000000000..f27ac5ff597d --- /dev/null +++ b/arch/x86/include/asm/kbdleds.h @@ -0,0 +1,17 @@ +#ifndef _ASM_X86_KBDLEDS_H +#define _ASM_X86_KBDLEDS_H + +/* + * Some laptops take the 789uiojklm,. keys as number pad when NumLock is on. + * This seems a good reason to start with NumLock off. That's why on X86 we + * ask the bios for the correct state. + */ + +#include <asm/setup.h> + +static inline int kbd_defleds(void) +{ + return boot_params.kbd_status & 0x20 ? (1 << VC_NUMLOCK) : 0; +} + +#endif /* _ASM_X86_KBDLEDS_H */ diff --git a/arch/x86/include/asm/kdebug.h b/arch/x86/include/asm/kdebug.h index d73f1571bde7..2c37aadcbc35 100644 --- a/arch/x86/include/asm/kdebug.h +++ b/arch/x86/include/asm/kdebug.h @@ -24,7 +24,6 @@ enum die_val { extern void printk_address(unsigned long address, int reliable); extern void die(const char *, struct pt_regs *,long); extern int __must_check __die(const char *, struct pt_regs *, long); -extern void show_registers(struct pt_regs *regs); extern void show_trace(struct task_struct *t, struct pt_regs *regs, unsigned long *sp, unsigned long bp); extern void __show_regs(struct pt_regs *regs, int all); diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index e216ba066e79..e5b97be12d2a 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -27,6 +27,7 @@ #include <asm/desc.h> #include <asm/mtrr.h> #include <asm/msr-index.h> +#include <asm/asm.h> #define KVM_MAX_VCPUS 254 #define KVM_SOFT_MAX_VCPUS 160 @@ -921,9 +922,7 @@ extern bool kvm_rebooting; __ASM_SIZE(push) " $666b \n\t" \ "call kvm_spurious_fault \n\t" \ ".popsection \n\t" \ - ".pushsection __ex_table, \"a\" \n\t" \ - _ASM_PTR " 666b, 667b \n\t" \ - ".popsection" + _ASM_EXTABLE(666b, 667b) #define __kvm_handle_fault_on_reboot(insn) \ ____kvm_handle_fault_on_reboot(insn, "") diff --git a/arch/x86/include/asm/mca.h b/arch/x86/include/asm/mca.h deleted file mode 100644 index eedbb6cc1efb..000000000000 --- a/arch/x86/include/asm/mca.h +++ /dev/null @@ -1,43 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8 -*- */ - -/* Platform specific MCA defines */ -#ifndef _ASM_X86_MCA_H -#define _ASM_X86_MCA_H - -/* Maximal number of MCA slots - actually, some machines have less, but - * they all have sufficient number of POS registers to cover 8. - */ -#define MCA_MAX_SLOT_NR 8 - -/* Most machines have only one MCA bus. The only multiple bus machines - * I know have at most two */ -#define MAX_MCA_BUSSES 2 - -#define MCA_PRIMARY_BUS 0 -#define MCA_SECONDARY_BUS 1 - -/* Dummy slot numbers on primary MCA for integrated functions */ -#define MCA_INTEGSCSI (MCA_MAX_SLOT_NR) -#define MCA_INTEGVIDEO (MCA_MAX_SLOT_NR+1) -#define MCA_MOTHERBOARD (MCA_MAX_SLOT_NR+2) - -/* Dummy POS values for integrated functions */ -#define MCA_DUMMY_POS_START 0x10000 -#define MCA_INTEGSCSI_POS (MCA_DUMMY_POS_START+1) -#define MCA_INTEGVIDEO_POS (MCA_DUMMY_POS_START+2) -#define MCA_MOTHERBOARD_POS (MCA_DUMMY_POS_START+3) - -/* MCA registers */ - -#define MCA_MOTHERBOARD_SETUP_REG 0x94 -#define MCA_ADAPTER_SETUP_REG 0x96 -#define MCA_POS_REG(n) (0x100+(n)) - -#define MCA_ENABLED 0x01 /* POS 2, set if adapter enabled */ - -/* Max number of adapters, including both slots and various integrated - * things. - */ -#define MCA_NUMADAPTERS (MCA_MAX_SLOT_NR+3) - -#endif /* _ASM_X86_MCA_H */ diff --git a/arch/x86/include/asm/mca_dma.h b/arch/x86/include/asm/mca_dma.h deleted file mode 100644 index 45271aef82dd..000000000000 --- a/arch/x86/include/asm/mca_dma.h +++ /dev/null @@ -1,201 +0,0 @@ -#ifndef _ASM_X86_MCA_DMA_H -#define _ASM_X86_MCA_DMA_H - -#include <asm/io.h> -#include <linux/ioport.h> - -/* - * Microchannel specific DMA stuff. DMA on an MCA machine is fairly similar to - * standard PC dma, but it certainly has its quirks. DMA register addresses - * are in a different place and there are some added functions. Most of this - * should be pretty obvious on inspection. Note that the user must divide - * count by 2 when using 16-bit dma; that is not handled by these functions. - * - * Ramen Noodles are yummy. - * - * 1998 Tymm Twillman <[email protected]> - */ - -/* - * Registers that are used by the DMA controller; FN is the function register - * (tell the controller what to do) and EXE is the execution register (how - * to do it) - */ - -#define MCA_DMA_REG_FN 0x18 -#define MCA_DMA_REG_EXE 0x1A - -/* - * Functions that the DMA controller can do - */ - -#define MCA_DMA_FN_SET_IO 0x00 -#define MCA_DMA_FN_SET_ADDR 0x20 -#define MCA_DMA_FN_GET_ADDR 0x30 -#define MCA_DMA_FN_SET_COUNT 0x40 -#define MCA_DMA_FN_GET_COUNT 0x50 -#define MCA_DMA_FN_GET_STATUS 0x60 -#define MCA_DMA_FN_SET_MODE 0x70 -#define MCA_DMA_FN_SET_ARBUS 0x80 -#define MCA_DMA_FN_MASK 0x90 -#define MCA_DMA_FN_RESET_MASK 0xA0 -#define MCA_DMA_FN_MASTER_CLEAR 0xD0 - -/* - * Modes (used by setting MCA_DMA_FN_MODE in the function register) - * - * Note that the MODE_READ is read from memory (write to device), and - * MODE_WRITE is vice-versa. - */ - -#define MCA_DMA_MODE_XFER 0x04 /* read by default */ -#define MCA_DMA_MODE_READ 0x04 /* same as XFER */ -#define MCA_DMA_MODE_WRITE 0x08 /* OR with MODE_XFER to use */ -#define MCA_DMA_MODE_IO 0x01 /* DMA from IO register */ -#define MCA_DMA_MODE_16 0x40 /* 16 bit xfers */ - - -/** - * mca_enable_dma - channel to enable DMA on - * @dmanr: DMA channel - * - * Enable the MCA bus DMA on a channel. This can be called from - * IRQ context. - */ - -static inline void mca_enable_dma(unsigned int dmanr) -{ - outb(MCA_DMA_FN_RESET_MASK | dmanr, MCA_DMA_REG_FN); -} - -/** - * mca_disble_dma - channel to disable DMA on - * @dmanr: DMA channel - * - * Enable the MCA bus DMA on a channel. This can be called from - * IRQ context. - */ - -static inline void mca_disable_dma(unsigned int dmanr) -{ - outb(MCA_DMA_FN_MASK | dmanr, MCA_DMA_REG_FN); -} - -/** - * mca_set_dma_addr - load a 24bit DMA address - * @dmanr: DMA channel - * @a: 24bit bus address - * - * Load the address register in the DMA controller. This has a 24bit - * limitation (16Mb). - */ - -static inline void mca_set_dma_addr(unsigned int dmanr, unsigned int a) -{ - outb(MCA_DMA_FN_SET_ADDR | dmanr, MCA_DMA_REG_FN); - outb(a & 0xff, MCA_DMA_REG_EXE); - outb((a >> 8) & 0xff, MCA_DMA_REG_EXE); - outb((a >> 16) & 0xff, MCA_DMA_REG_EXE); -} - -/** - * mca_get_dma_addr - load a 24bit DMA address - * @dmanr: DMA channel - * - * Read the address register in the DMA controller. This has a 24bit - * limitation (16Mb). The return is a bus address. - */ - -static inline unsigned int mca_get_dma_addr(unsigned int dmanr) -{ - unsigned int addr; - - outb(MCA_DMA_FN_GET_ADDR | dmanr, MCA_DMA_REG_FN); - addr = inb(MCA_DMA_REG_EXE); - addr |= inb(MCA_DMA_REG_EXE) << 8; - addr |= inb(MCA_DMA_REG_EXE) << 16; - - return addr; -} - -/** - * mca_set_dma_count - load a 16bit transfer count - * @dmanr: DMA channel - * @count: count - * - * Set the DMA count for this channel. This can be up to 64Kbytes. - * Setting a count of zero will not do what you expect. - */ - -static inline void mca_set_dma_count(unsigned int dmanr, unsigned int count) -{ - count--; /* transfers one more than count -- correct for this */ - - outb(MCA_DMA_FN_SET_COUNT | dmanr, MCA_DMA_REG_FN); - outb(count & 0xff, MCA_DMA_REG_EXE); - outb((count >> 8) & 0xff, MCA_DMA_REG_EXE); -} - -/** - * mca_get_dma_residue - get the remaining bytes to transfer - * @dmanr: DMA channel - * - * This function returns the number of bytes left to transfer - * on this DMA channel. - */ - -static inline unsigned int mca_get_dma_residue(unsigned int dmanr) -{ - unsigned short count; - - outb(MCA_DMA_FN_GET_COUNT | dmanr, MCA_DMA_REG_FN); - count = 1 + inb(MCA_DMA_REG_EXE); - count += inb(MCA_DMA_REG_EXE) << 8; - - return count; -} - -/** - * mca_set_dma_io - set the port for an I/O transfer - * @dmanr: DMA channel - * @io_addr: an I/O port number - * - * Unlike the ISA bus DMA controllers the DMA on MCA bus can transfer - * with an I/O port target. - */ - -static inline void mca_set_dma_io(unsigned int dmanr, unsigned int io_addr) -{ - /* - * DMA from a port address -- set the io address - */ - - outb(MCA_DMA_FN_SET_IO | dmanr, MCA_DMA_REG_FN); - outb(io_addr & 0xff, MCA_DMA_REG_EXE); - outb((io_addr >> 8) & 0xff, MCA_DMA_REG_EXE); -} - -/** - * mca_set_dma_mode - set the DMA mode - * @dmanr: DMA channel - * @mode: mode to set - * - * The DMA controller supports several modes. The mode values you can - * set are- - * - * %MCA_DMA_MODE_READ when reading from the DMA device. - * - * %MCA_DMA_MODE_WRITE to writing to the DMA device. - * - * %MCA_DMA_MODE_IO to do DMA to or from an I/O port. - * - * %MCA_DMA_MODE_16 to do 16bit transfers. - */ - -static inline void mca_set_dma_mode(unsigned int dmanr, unsigned int mode) -{ - outb(MCA_DMA_FN_SET_MODE | dmanr, MCA_DMA_REG_FN); - outb(mode, MCA_DMA_REG_EXE); -} - -#endif /* _ASM_X86_MCA_DMA_H */ diff --git a/arch/x86/include/asm/mmzone_32.h b/arch/x86/include/asm/mmzone_32.h index 55728e121473..eb05fb3b02fb 100644 --- a/arch/x86/include/asm/mmzone_32.h +++ b/arch/x86/include/asm/mmzone_32.h @@ -61,10 +61,4 @@ static inline int pfn_valid(int pfn) #endif /* CONFIG_DISCONTIGMEM */ -#ifdef CONFIG_NEED_MULTIPLE_NODES -/* always use node 0 for bootmem on this numa platform */ -#define bootmem_arch_preferred_node(__bdata, size, align, goal, limit) \ - (NODE_DATA(0)->bdata) -#endif /* CONFIG_NEED_MULTIPLE_NODES */ - #endif /* _ASM_X86_MMZONE_32_H */ diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index 9c7d95f6174b..3e2f42a4b872 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -40,7 +40,7 @@ extern int quad_local_to_mp_bus_id [NR_CPUS/4][4]; #endif /* CONFIG_X86_64 */ -#if defined(CONFIG_MCA) || defined(CONFIG_EISA) +#ifdef CONFIG_EISA extern int mp_bus_id_to_type[MAX_MP_BUSSES]; #endif diff --git a/arch/x86/include/asm/mpspec_def.h b/arch/x86/include/asm/mpspec_def.h index c0a955a9a087..b31f8c098271 100644 --- a/arch/x86/include/asm/mpspec_def.h +++ b/arch/x86/include/asm/mpspec_def.h @@ -84,7 +84,7 @@ struct mpc_bus { #define BUSTYPE_EISA "EISA" #define BUSTYPE_ISA "ISA" #define BUSTYPE_INTERN "INTERN" /* Internal BUS */ -#define BUSTYPE_MCA "MCA" +#define BUSTYPE_MCA "MCA" /* Obsolete */ #define BUSTYPE_VL "VL" /* Local bus */ #define BUSTYPE_PCI "PCI" #define BUSTYPE_PCMCIA "PCMCIA" @@ -169,6 +169,5 @@ enum mp_bustype { MP_BUS_ISA = 1, MP_BUS_EISA, MP_BUS_PCI, - MP_BUS_MCA, }; #endif /* _ASM_X86_MPSPEC_DEF_H */ diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h index 95203d40ffdd..084ef95274cd 100644 --- a/arch/x86/include/asm/msr.h +++ b/arch/x86/include/asm/msr.h @@ -169,14 +169,7 @@ static inline int wrmsr_safe(unsigned msr, unsigned low, unsigned high) return native_write_msr_safe(msr, low, high); } -/* - * rdmsr with exception handling. - * - * Please note that the exception handling works only after we've - * switched to the "smart" #GP handler in trap_init() which knows about - * exception tables - using this macro earlier than that causes machine - * hangs on boxes which do not implement the @msr in the first argument. - */ +/* rdmsr with exception handling */ #define rdmsr_safe(msr, p1, p2) \ ({ \ int __err; \ diff --git a/arch/x86/include/asm/nops.h b/arch/x86/include/asm/nops.h index 405b4032a60b..aff2b3356101 100644 --- a/arch/x86/include/asm/nops.h +++ b/arch/x86/include/asm/nops.h @@ -87,7 +87,11 @@ #define P6_NOP8 0x0f,0x1f,0x84,0x00,0,0,0,0 #define P6_NOP5_ATOMIC P6_NOP5 +#ifdef __ASSEMBLY__ +#define _ASM_MK_NOP(x) .byte x +#else #define _ASM_MK_NOP(x) ".byte " __stringify(x) "\n" +#endif #if defined(CONFIG_MK7) #define ASM_NOP1 _ASM_MK_NOP(K7_NOP1) diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h index aa0f91308367..6cbbabf52707 100644 --- a/arch/x86/include/asm/paravirt.h +++ b/arch/x86/include/asm/paravirt.h @@ -1023,10 +1023,8 @@ extern void default_banner(void); call PARA_INDIRECT(pv_cpu_ops+PV_CPU_swapgs) \ ) -#define GET_CR2_INTO_RCX \ - call PARA_INDIRECT(pv_mmu_ops+PV_MMU_read_cr2); \ - movq %rax, %rcx; \ - xorq %rax, %rax; +#define GET_CR2_INTO_RAX \ + call PARA_INDIRECT(pv_mmu_ops+PV_MMU_read_cr2) #define PARAVIRT_ADJUST_EXCEPTION_FRAME \ PARA_SITE(PARA_PATCH(pv_irq_ops, PV_IRQ_adjust_exception_frame), \ diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index ccbb1ea99ccb..7745b257f035 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -579,9 +579,6 @@ extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags); /* Free all resources held by a thread. */ extern void release_thread(struct task_struct *); -/* Prepare to copy thread state - unlazy all lazy state */ -extern void prepare_to_copy(struct task_struct *tsk); - unsigned long get_wchan(struct task_struct *p); /* diff --git a/arch/x86/include/asm/segment.h b/arch/x86/include/asm/segment.h index 165466233ab0..c48a95035a77 100644 --- a/arch/x86/include/asm/segment.h +++ b/arch/x86/include/asm/segment.h @@ -205,13 +205,15 @@ #define IDT_ENTRIES 256 #define NUM_EXCEPTION_VECTORS 32 +/* Bitmask of exception vectors which push an error code on the stack */ +#define EXCEPTION_ERRCODE_MASK 0x00027d00 #define GDT_SIZE (GDT_ENTRIES * 8) #define GDT_ENTRY_TLS_ENTRIES 3 #define TLS_SIZE (GDT_ENTRY_TLS_ENTRIES * 8) #ifdef __KERNEL__ #ifndef __ASSEMBLY__ -extern const char early_idt_handlers[NUM_EXCEPTION_VECTORS][10]; +extern const char early_idt_handlers[NUM_EXCEPTION_VECTORS][2+2+5]; /* * Load a segment. Fall back on loading the zero diff --git a/arch/x86/include/asm/spinlock.h b/arch/x86/include/asm/spinlock.h index 76bfa2cf301d..b315a33867f2 100644 --- a/arch/x86/include/asm/spinlock.h +++ b/arch/x86/include/asm/spinlock.h @@ -20,10 +20,8 @@ #ifdef CONFIG_X86_32 # define LOCK_PTR_REG "a" -# define REG_PTR_MODE "k" #else # define LOCK_PTR_REG "D" -# define REG_PTR_MODE "q" #endif #if defined(CONFIG_X86_32) && \ diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h index 1620d23f14d7..36a1a2ab87d2 100644 --- a/arch/x86/include/asm/tlbflush.h +++ b/arch/x86/include/asm/tlbflush.h @@ -62,11 +62,7 @@ static inline void __flush_tlb_one(unsigned long addr) __flush_tlb(); } -#ifdef CONFIG_X86_32 -# define TLB_FLUSH_ALL 0xffffffff -#else -# define TLB_FLUSH_ALL -1ULL -#endif +#define TLB_FLUSH_ALL -1UL /* * TLB flushing: diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h index e0544597cfe7..851fe0dc13bc 100644 --- a/arch/x86/include/asm/uaccess.h +++ b/arch/x86/include/asm/uaccess.h @@ -79,11 +79,12 @@ #define access_ok(type, addr, size) (likely(__range_not_ok(addr, size) == 0)) /* - * The exception table consists of pairs of addresses: the first is the - * address of an instruction that is allowed to fault, and the second is - * the address at which the program should continue. No registers are - * modified, so it is entirely up to the continuation code to figure out - * what to do. + * The exception table consists of pairs of addresses relative to the + * exception table enty itself: the first is the address of an + * instruction that is allowed to fault, and the second is the address + * at which the program should continue. No registers are modified, + * so it is entirely up to the continuation code to figure out what to + * do. * * All the routines below use bits of fixup code that are out of line * with the main instruction path. This means when everything is well, @@ -92,10 +93,14 @@ */ struct exception_table_entry { - unsigned long insn, fixup; + int insn, fixup; }; +/* This is not the generic standard exception_table_entry format */ +#define ARCH_HAS_SORT_EXTABLE +#define ARCH_HAS_SEARCH_EXTABLE extern int fixup_exception(struct pt_regs *regs); +extern int early_fixup_exception(unsigned long *ip); /* * These are the main single-value transfer routines. They automatically @@ -202,8 +207,8 @@ extern int __get_user_bad(void); asm volatile("1: movl %%eax,0(%1)\n" \ "2: movl %%edx,4(%1)\n" \ "3:\n" \ - _ASM_EXTABLE(1b, 2b - 1b) \ - _ASM_EXTABLE(2b, 3b - 2b) \ + _ASM_EXTABLE_EX(1b, 2b) \ + _ASM_EXTABLE_EX(2b, 3b) \ : : "A" (x), "r" (addr)) #define __put_user_x8(x, ptr, __ret_pu) \ @@ -408,7 +413,7 @@ do { \ #define __get_user_asm_ex(x, addr, itype, rtype, ltype) \ asm volatile("1: mov"itype" %1,%"rtype"0\n" \ "2:\n" \ - _ASM_EXTABLE(1b, 2b - 1b) \ + _ASM_EXTABLE_EX(1b, 2b) \ : ltype(x) : "m" (__m(addr))) #define __put_user_nocheck(x, ptr, size) \ @@ -450,7 +455,7 @@ struct __large_struct { unsigned long buf[100]; }; #define __put_user_asm_ex(x, addr, itype, rtype, ltype) \ asm volatile("1: mov"itype" %"rtype"0,%1\n" \ "2:\n" \ - _ASM_EXTABLE(1b, 2b - 1b) \ + _ASM_EXTABLE_EX(1b, 2b) \ : : ltype(x), "m" (__m(addr))) /* diff --git a/arch/x86/include/asm/xor_32.h b/arch/x86/include/asm/xor_32.h index 133b40a0f495..454570891bdc 100644 --- a/arch/x86/include/asm/xor_32.h +++ b/arch/x86/include/asm/xor_32.h @@ -861,6 +861,9 @@ static struct xor_block_template xor_block_pIII_sse = { .do_5 = xor_sse_5, }; +/* Also try the AVX routines */ +#include "xor_avx.h" + /* Also try the generic routines. */ #include <asm-generic/xor.h> @@ -871,6 +874,7 @@ do { \ xor_speed(&xor_block_8regs_p); \ xor_speed(&xor_block_32regs); \ xor_speed(&xor_block_32regs_p); \ + AVX_XOR_SPEED; \ if (cpu_has_xmm) \ xor_speed(&xor_block_pIII_sse); \ if (cpu_has_mmx) { \ @@ -883,6 +887,6 @@ do { \ We may also be able to load into the L1 only depending on how the cpu deals with a load to a line that is being prefetched. */ #define XOR_SELECT_TEMPLATE(FASTEST) \ - (cpu_has_xmm ? &xor_block_pIII_sse : FASTEST) + AVX_SELECT(cpu_has_xmm ? &xor_block_pIII_sse : FASTEST) #endif /* _ASM_X86_XOR_32_H */ diff --git a/arch/x86/include/asm/xor_64.h b/arch/x86/include/asm/xor_64.h index 1549b5e261f6..b9b2323e90fe 100644 --- a/arch/x86/include/asm/xor_64.h +++ b/arch/x86/include/asm/xor_64.h @@ -347,15 +347,21 @@ static struct xor_block_template xor_block_sse = { .do_5 = xor_sse_5, }; + +/* Also try the AVX routines */ +#include "xor_avx.h" + #undef XOR_TRY_TEMPLATES #define XOR_TRY_TEMPLATES \ do { \ + AVX_XOR_SPEED; \ xor_speed(&xor_block_sse); \ } while (0) /* We force the use of the SSE xor block because it can write around L2. We may also be able to load into the L1 only depending on how the cpu deals with a load to a line that is being prefetched. */ -#define XOR_SELECT_TEMPLATE(FASTEST) (&xor_block_sse) +#define XOR_SELECT_TEMPLATE(FASTEST) \ + AVX_SELECT(&xor_block_sse) #endif /* _ASM_X86_XOR_64_H */ diff --git a/arch/x86/include/asm/xor_avx.h b/arch/x86/include/asm/xor_avx.h new file mode 100644 index 000000000000..2510d35f480e --- /dev/null +++ b/arch/x86/include/asm/xor_avx.h @@ -0,0 +1,214 @@ +#ifndef _ASM_X86_XOR_AVX_H +#define _ASM_X86_XOR_AVX_H + +/* + * Optimized RAID-5 checksumming functions for AVX + * + * Copyright (C) 2012 Intel Corporation + * Author: Jim Kukunas <[email protected]> + * + * Based on Ingo Molnar and Zach Brown's respective MMX and SSE routines + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; version 2 + * of the License. + */ + +#ifdef CONFIG_AS_AVX + +#include <linux/compiler.h> +#include <asm/i387.h> + +#define ALIGN32 __aligned(32) + +#define YMM_SAVED_REGS 4 + +#define YMMS_SAVE \ +do { \ + preempt_disable(); \ + cr0 = read_cr0(); \ + clts(); \ + asm volatile("vmovaps %%ymm0, %0" : "=m" (ymm_save[0]) : : "memory"); \ + asm volatile("vmovaps %%ymm1, %0" : "=m" (ymm_save[32]) : : "memory"); \ + asm volatile("vmovaps %%ymm2, %0" : "=m" (ymm_save[64]) : : "memory"); \ + asm volatile("vmovaps %%ymm3, %0" : "=m" (ymm_save[96]) : : "memory"); \ +} while (0); + +#define YMMS_RESTORE \ +do { \ + asm volatile("sfence" : : : "memory"); \ + asm volatile("vmovaps %0, %%ymm3" : : "m" (ymm_save[96])); \ + asm volatile("vmovaps %0, %%ymm2" : : "m" (ymm_save[64])); \ + asm volatile("vmovaps %0, %%ymm1" : : "m" (ymm_save[32])); \ + asm volatile("vmovaps %0, %%ymm0" : : "m" (ymm_save[0])); \ + write_cr0(cr0); \ + preempt_enable(); \ +} while (0); + +#define BLOCK4(i) \ + BLOCK(32 * i, 0) \ + BLOCK(32 * (i + 1), 1) \ + BLOCK(32 * (i + 2), 2) \ + BLOCK(32 * (i + 3), 3) + +#define BLOCK16() \ + BLOCK4(0) \ + BLOCK4(4) \ + BLOCK4(8) \ + BLOCK4(12) + +static void xor_avx_2(unsigned long bytes, unsigned long *p0, unsigned long *p1) +{ + unsigned long cr0, lines = bytes >> 9; + char ymm_save[32 * YMM_SAVED_REGS] ALIGN32; + + YMMS_SAVE + + while (lines--) { +#undef BLOCK +#define BLOCK(i, reg) \ +do { \ + asm volatile("vmovdqa %0, %%ymm" #reg : : "m" (p1[i / sizeof(*p1)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p0[i / sizeof(*p0)])); \ + asm volatile("vmovdqa %%ymm" #reg ", %0" : \ + "=m" (p0[i / sizeof(*p0)])); \ +} while (0); + + BLOCK16() + + p0 = (unsigned long *)((uintptr_t)p0 + 512); + p1 = (unsigned long *)((uintptr_t)p1 + 512); + } + + YMMS_RESTORE +} + +static void xor_avx_3(unsigned long bytes, unsigned long *p0, unsigned long *p1, + unsigned long *p2) +{ + unsigned long cr0, lines = bytes >> 9; + char ymm_save[32 * YMM_SAVED_REGS] ALIGN32; + + YMMS_SAVE + + while (lines--) { +#undef BLOCK +#define BLOCK(i, reg) \ +do { \ + asm volatile("vmovdqa %0, %%ymm" #reg : : "m" (p2[i / sizeof(*p2)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p1[i / sizeof(*p1)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p0[i / sizeof(*p0)])); \ + asm volatile("vmovdqa %%ymm" #reg ", %0" : \ + "=m" (p0[i / sizeof(*p0)])); \ +} while (0); + + BLOCK16() + + p0 = (unsigned long *)((uintptr_t)p0 + 512); + p1 = (unsigned long *)((uintptr_t)p1 + 512); + p2 = (unsigned long *)((uintptr_t)p2 + 512); + } + + YMMS_RESTORE +} + +static void xor_avx_4(unsigned long bytes, unsigned long *p0, unsigned long *p1, + unsigned long *p2, unsigned long *p3) +{ + unsigned long cr0, lines = bytes >> 9; + char ymm_save[32 * YMM_SAVED_REGS] ALIGN32; + + YMMS_SAVE + + while (lines--) { +#undef BLOCK +#define BLOCK(i, reg) \ +do { \ + asm volatile("vmovdqa %0, %%ymm" #reg : : "m" (p3[i / sizeof(*p3)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p2[i / sizeof(*p2)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p1[i / sizeof(*p1)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p0[i / sizeof(*p0)])); \ + asm volatile("vmovdqa %%ymm" #reg ", %0" : \ + "=m" (p0[i / sizeof(*p0)])); \ +} while (0); + + BLOCK16(); + + p0 = (unsigned long *)((uintptr_t)p0 + 512); + p1 = (unsigned long *)((uintptr_t)p1 + 512); + p2 = (unsigned long *)((uintptr_t)p2 + 512); + p3 = (unsigned long *)((uintptr_t)p3 + 512); + } + + YMMS_RESTORE +} + +static void xor_avx_5(unsigned long bytes, unsigned long *p0, unsigned long *p1, + unsigned long *p2, unsigned long *p3, unsigned long *p4) +{ + unsigned long cr0, lines = bytes >> 9; + char ymm_save[32 * YMM_SAVED_REGS] ALIGN32; + + YMMS_SAVE + + while (lines--) { +#undef BLOCK +#define BLOCK(i, reg) \ +do { \ + asm volatile("vmovdqa %0, %%ymm" #reg : : "m" (p4[i / sizeof(*p4)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p3[i / sizeof(*p3)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p2[i / sizeof(*p2)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p1[i / sizeof(*p1)])); \ + asm volatile("vxorps %0, %%ymm" #reg ", %%ymm" #reg : : \ + "m" (p0[i / sizeof(*p0)])); \ + asm volatile("vmovdqa %%ymm" #reg ", %0" : \ + "=m" (p0[i / sizeof(*p0)])); \ +} while (0); + + BLOCK16() + + p0 = (unsigned long *)((uintptr_t)p0 + 512); + p1 = (unsigned long *)((uintptr_t)p1 + 512); + p2 = (unsigned long *)((uintptr_t)p2 + 512); + p3 = (unsigned long *)((uintptr_t)p3 + 512); + p4 = (unsigned long *)((uintptr_t)p4 + 512); + } + + YMMS_RESTORE +} + +static struct xor_block_template xor_block_avx = { + .name = "avx", + .do_2 = xor_avx_2, + .do_3 = xor_avx_3, + .do_4 = xor_avx_4, + .do_5 = xor_avx_5, +}; + +#define AVX_XOR_SPEED \ +do { \ + if (cpu_has_avx) \ + xor_speed(&xor_block_avx); \ +} while (0) + +#define AVX_SELECT(FASTEST) \ + (cpu_has_avx ? &xor_block_avx : FASTEST) + +#else + +#define AVX_XOR_SPEED {} + +#define AVX_SELECT(FASTEST) (FASTEST) + +#endif +#endif diff --git a/arch/x86/include/asm/xsave.h b/arch/x86/include/asm/xsave.h index c6ce2452f10c..8a1b6f9b594a 100644 --- a/arch/x86/include/asm/xsave.h +++ b/arch/x86/include/asm/xsave.h @@ -80,10 +80,7 @@ static inline int xsave_user(struct xsave_struct __user *buf) "3: movl $-1,%[err]\n" " jmp 2b\n" ".previous\n" - ".section __ex_table,\"a\"\n" - _ASM_ALIGN "\n" - _ASM_PTR "1b,3b\n" - ".previous" + _ASM_EXTABLE(1b,3b) : [err] "=r" (err) : "D" (buf), "a" (-1), "d" (-1), "0" (0) : "memory"); @@ -106,10 +103,7 @@ static inline int xrestore_user(struct xsave_struct __user *buf, u64 mask) "3: movl $-1,%[err]\n" " jmp 2b\n" ".previous\n" - ".section __ex_table,\"a\"\n" - _ASM_ALIGN "\n" - _ASM_PTR "1b,3b\n" - ".previous" + _ASM_EXTABLE(1b,3b) : [err] "=r" (err) : "D" (xstate), "a" (lmask), "d" (hmask), "0" (0) : "memory"); /* memory required? */ |