mips: sync asm/mipsregs.h with Linux 5.7

Sync asm/mipsregs.h with Linux 5.7. Also replace the custom
symbols EBASE_CPUNUM and EBASE_WG with the according symbols
from Linux.

Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>

3 files changed, 1101 insertions, 194 deletions

diff --git a/arch/mips/include/asm/compiler.h b/arch/mips/include/asm/compiler.h
new file mode 100644
index 0000000..c498b42
--- /dev/null
+++ b/arch/mips/include/asm/compiler.h
@@ -0,0 +1,69 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2004, 2007  Maciej W. Rozycki
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+#ifndef _ASM_COMPILER_H
+#define _ASM_COMPILER_H
+
+/*
+ * With GCC 4.5 onwards we can use __builtin_unreachable to indicate to the
+ * compiler that a particular code path will never be hit. This allows it to be
+ * optimised out of the generated binary.
+ *
+ * Unfortunately at least GCC 4.6.3 through 7.3.0 inclusive suffer from a bug
+ * that can lead to instructions from beyond an unreachable statement being
+ * incorrectly reordered into earlier delay slots if the unreachable statement
+ * is the only content of a case in a switch statement. This can lead to
+ * seemingly random behaviour, such as invalid memory accesses from incorrectly
+ * reordered loads or stores. See this potential GCC fix for details:
+ *
+ *   https://gcc.gnu.org/ml/gcc-patches/2015-09/msg00360.html
+ *
+ * It is unclear whether GCC 8 onwards suffer from the same issue - nothing
+ * relevant is mentioned in GCC 8 release notes and nothing obviously relevant
+ * stands out in GCC commit logs, but these newer GCC versions generate very
+ * different code for the testcase which doesn't exhibit the bug.
+ *
+ * GCC also handles stack allocation suboptimally when calling noreturn
+ * functions or calling __builtin_unreachable():
+ *
+ *   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82365
+ *
+ * We work around both of these issues by placing a volatile asm statement,
+ * which GCC is prevented from reordering past, prior to __builtin_unreachable
+ * calls.
+ *
+ * The .insn statement is required to ensure that any branches to the
+ * statement, which sadly must be kept due to the asm statement, are known to
+ * be branches to code and satisfy linker requirements for microMIPS kernels.
+ */
+#undef barrier_before_unreachable
+#define barrier_before_unreachable() asm volatile(".insn")
+
+#if !defined(CONFIG_CC_IS_GCC) || \
+    (__GNUC__ > 4) || (__GNUC__ == 4 && __GNUC_MINOR__ >= 9)
+# define GCC_OFF_SMALL_ASM() "ZC"
+#elif defined(CONFIG_CPU_MICROMIPS)
+# error "microMIPS compilation unsupported with GCC older than 4.9"
+#else
+# define GCC_OFF_SMALL_ASM() "R"
+#endif
+
+#ifdef CONFIG_CPU_MIPSR6
+#define MIPS_ISA_LEVEL "mips64r6"
+#define MIPS_ISA_ARCH_LEVEL MIPS_ISA_LEVEL
+#define MIPS_ISA_LEVEL_RAW mips64r6
+#define MIPS_ISA_ARCH_LEVEL_RAW MIPS_ISA_LEVEL_RAW
+#else
+/* MIPS64 is a superset of MIPS32 */
+#define MIPS_ISA_LEVEL "mips64r2"
+#define MIPS_ISA_ARCH_LEVEL "arch=r4000"
+#define MIPS_ISA_LEVEL_RAW mips64r2
+#define MIPS_ISA_ARCH_LEVEL_RAW MIPS_ISA_LEVEL_RAW
+#endif /* CONFIG_CPU_MIPSR6 */
+
+#endif /* _ASM_COMPILER_H */
diff --git a/arch/mips/include/asm/isa-rev.h b/arch/mips/include/asm/isa-rev.h
new file mode 100644
index 0000000..683ea34
--- /dev/null
+++ b/arch/mips/include/asm/isa-rev.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2018 MIPS Tech, LLC
+ * Author: Matt Redfearn <matt.redfearn@mips.com>
+ */
+
+#ifndef __MIPS_ASM_ISA_REV_H__
+#define __MIPS_ASM_ISA_REV_H__
+
+/*
+ * The ISA revision level. This is 0 for MIPS I to V and N for
+ * MIPS{32,64}rN.
+ */
+
+/* If the compiler has defined __mips_isa_rev, believe it. */
+#ifdef __mips_isa_rev
+#define MIPS_ISA_REV __mips_isa_rev
+#else
+/* The compiler hasn't defined the isa rev so assume it's MIPS I - V (0) */
+#define MIPS_ISA_REV 0
+#endif
+
+
+#endif /* __MIPS_ASM_ISA_REV_H__ */
diff --git a/arch/mips/include/asm/mipsregs.h b/arch/mips/include/asm/mipsregs.h
index 1738130..e65485b 100644
--- a/arch/mips/include/asm/mipsregs.h
+++ b/arch/mips/include/asm/mipsregs.h
@@ -10,6 +10,8 @@
 #ifndef _ASM_MIPSREGS_H
 #define _ASM_MIPSREGS_H
 
+#include <asm/compiler.h>
+#include <asm/isa-rev.h>
 /*
  * The following macros are especially useful for __asm__
  * inline assembler.
@@ -26,9 +28,10 @@
  */
 #ifdef __ASSEMBLY__
 #define _ULCAST_
+#define _U64CAST_
 #else
-#include <linux/bitops.h>
 #define _ULCAST_ (unsigned long)
+#define _U64CAST_ (u64)
 #endif
 
 /*
@@ -42,15 +45,25 @@
 #define CP0_GLOBALNUMBER $3, 1
 #define CP0_CONTEXT $4
 #define CP0_PAGEMASK $5
+#define CP0_PAGEGRAIN $5, 1
+#define CP0_SEGCTL0 $5, 2
+#define CP0_SEGCTL1 $5, 3
+#define CP0_SEGCTL2 $5, 4
 #define CP0_WIRED $6
 #define CP0_INFO $7
-#define CP0_HWRENA $7, 0
+#define CP0_HWRENA $7
 #define CP0_BADVADDR $8
 #define CP0_BADINSTR $8, 1
 #define CP0_COUNT $9
 #define CP0_ENTRYHI $10
+#define CP0_GUESTCTL1 $10, 4
+#define CP0_GUESTCTL2 $10, 5
+#define CP0_GUESTCTL3 $10, 6
 #define CP0_COMPARE $11
+#define CP0_GUESTCTL0EXT $11, 4
 #define CP0_STATUS $12
+#define CP0_GUESTCTL0 $12, 6
+#define CP0_GTOFFSET $12, 7
 #define CP0_CAUSE $13
 #define CP0_EPC $14
 #define CP0_PRID $15
@@ -59,6 +72,7 @@
 #define CP0_CONFIG $16
 #define CP0_CONFIG3 $16, 3
 #define CP0_CONFIG5 $16, 5
+#define CP0_CONFIG6 $16, 6
 #define CP0_LLADDR $17
 #define CP0_WATCHLO $18
 #define CP0_WATCHHI $19
@@ -131,6 +145,16 @@
 #define MIPS_ENTRYLO_RI		(_ULCAST_(1) << (BITS_PER_LONG - 1))
 
 /*
+ * MIPSr6+ GlobalNumber register definitions
+ */
+#define MIPS_GLOBALNUMBER_VP_SHF	0
+#define MIPS_GLOBALNUMBER_VP		(_ULCAST_(0xff) << MIPS_GLOBALNUMBER_VP_SHF)
+#define MIPS_GLOBALNUMBER_CORE_SHF	8
+#define MIPS_GLOBALNUMBER_CORE		(_ULCAST_(0xff) << MIPS_GLOBALNUMBER_CORE_SHF)
+#define MIPS_GLOBALNUMBER_CLUSTER_SHF	16
+#define MIPS_GLOBALNUMBER_CLUSTER	(_ULCAST_(0xf) << MIPS_GLOBALNUMBER_CLUSTER_SHF)
+
+/*
  * Values for PageMask register
  */
 #ifdef CONFIG_CPU_VR41XX
@@ -189,6 +213,8 @@
 
 /* MIPS32/64 EntryHI bit definitions */
 #define MIPS_ENTRYHI_EHINV	(_ULCAST_(1) << 10)
+#define MIPS_ENTRYHI_ASIDX	(_ULCAST_(0x3) << 8)
+#define MIPS_ENTRYHI_ASID	(_ULCAST_(0xff) << 0)
 
 /*
  * R4x00 interrupt enable / cause bits
@@ -351,10 +377,14 @@
 #define	 CAUSEF_IP7		(_ULCAST_(1)   << 15)
 #define CAUSEB_FDCI		21
 #define CAUSEF_FDCI		(_ULCAST_(1)   << 21)
+#define CAUSEB_WP		22
+#define CAUSEF_WP		(_ULCAST_(1)   << 22)
 #define CAUSEB_IV		23
 #define CAUSEF_IV		(_ULCAST_(1)   << 23)
 #define CAUSEB_PCI		26
 #define CAUSEF_PCI		(_ULCAST_(1)   << 26)
+#define CAUSEB_DC		27
+#define CAUSEF_DC		(_ULCAST_(1)   << 27)
 #define CAUSEB_CE		28
 #define CAUSEF_CE		(_ULCAST_(3)   << 28)
 #define CAUSEB_TI		30
@@ -363,10 +393,36 @@
 #define CAUSEF_BD		(_ULCAST_(1)   << 31)
 
 /*
- * Bits in the coprocessor 0 EBase register.
+ * Cause.ExcCode trap codes.
  */
-#define EBASE_CPUNUM		0x3ff
-#define EBASE_WG		(_ULCAST_(1) << 11)
+#define EXCCODE_INT		0	/* Interrupt pending */
+#define EXCCODE_MOD		1	/* TLB modified fault */
+#define EXCCODE_TLBL		2	/* TLB miss on load or ifetch */
+#define EXCCODE_TLBS		3	/* TLB miss on a store */
+#define EXCCODE_ADEL		4	/* Address error on a load or ifetch */
+#define EXCCODE_ADES		5	/* Address error on a store */
+#define EXCCODE_IBE		6	/* Bus error on an ifetch */
+#define EXCCODE_DBE		7	/* Bus error on a load or store */
+#define EXCCODE_SYS		8	/* System call */
+#define EXCCODE_BP		9	/* Breakpoint */
+#define EXCCODE_RI		10	/* Reserved instruction exception */
+#define EXCCODE_CPU		11	/* Coprocessor unusable */
+#define EXCCODE_OV		12	/* Arithmetic overflow */
+#define EXCCODE_TR		13	/* Trap instruction */
+#define EXCCODE_MSAFPE		14	/* MSA floating point exception */
+#define EXCCODE_FPE		15	/* Floating point exception */
+#define EXCCODE_TLBRI		19	/* TLB Read-Inhibit exception */
+#define EXCCODE_TLBXI		20	/* TLB Execution-Inhibit exception */
+#define EXCCODE_MSADIS		21	/* MSA disabled exception */
+#define EXCCODE_MDMX		22	/* MDMX unusable exception */
+#define EXCCODE_WATCH		23	/* Watch address reference */
+#define EXCCODE_MCHECK		24	/* Machine check */
+#define EXCCODE_THREAD		25	/* Thread exceptions (MT) */
+#define EXCCODE_DSPDIS		26	/* DSP disabled exception */
+#define EXCCODE_GE		27	/* Virtualized guest exception (VZ) */
+
+/* Implementation specific trap codes used by MIPS cores */
+#define MIPS_EXCCODE_TLBPAR	16	/* TLB parity error exception */
 
 /*
  * Bits in the coprocessor 0 config register.
@@ -394,7 +450,7 @@
 #define CONF_SM			(_ULCAST_(1) << 16)
 #define CONF_SC			(_ULCAST_(1) << 17)
 #define CONF_EW			(_ULCAST_(3) << 18)
-#define CONF_EP			(_ULCAST_(15) << 24)
+#define CONF_EP			(_ULCAST_(15)<< 24)
 #define CONF_EC			(_ULCAST_(7) << 28)
 #define CONF_CM			(_ULCAST_(1) << 31)
 
@@ -420,7 +476,7 @@
 #define R10K_CONF_CT		(_ULCAST_(1) <<	 5)
 #define R10K_CONF_PE		(_ULCAST_(1) <<	 6)
 #define R10K_CONF_PM		(_ULCAST_(3) <<	 7)
-#define R10K_CONF_EC		(_ULCAST_(15) << 9)
+#define R10K_CONF_EC		(_ULCAST_(15)<<	 9)
 #define R10K_CONF_SB		(_ULCAST_(1) << 13)
 #define R10K_CONF_SK		(_ULCAST_(1) << 14)
 #define R10K_CONF_SS		(_ULCAST_(7) << 16)
@@ -453,6 +509,7 @@
 #define TX49_CONF_CWFON		(_ULCAST_(1) << 27)
 
 /* Bits specific to the MIPS32/64 PRA.	*/
+#define MIPS_CONF_VI		(_ULCAST_(1) <<  3)
 #define MIPS_CONF_MT		(_ULCAST_(7) <<	 7)
 #define MIPS_CONF_MT_TLB	(_ULCAST_(1) <<  7)
 #define MIPS_CONF_MT_FTLB	(_ULCAST_(4) <<  7)
@@ -549,7 +606,8 @@
 #define MIPS_CONF4_MMUEXTDEF_MMUSIZEEXT (_ULCAST_(1) << 14)
 #define MIPS_CONF4_MMUEXTDEF_FTLBSIZEEXT	(_ULCAST_(2) << 14)
 #define MIPS_CONF4_MMUEXTDEF_VTLBSIZEEXT	(_ULCAST_(3) << 14)
-#define MIPS_CONF4_KSCREXIST	(_ULCAST_(255) << 16)
+#define MIPS_CONF4_KSCREXIST_SHIFT	(16)
+#define MIPS_CONF4_KSCREXIST	(_ULCAST_(255) << MIPS_CONF4_KSCREXIST_SHIFT)
 #define MIPS_CONF4_VTLBSIZEEXT_SHIFT	(24)
 #define MIPS_CONF4_VTLBSIZEEXT	(_ULCAST_(15) << MIPS_CONF4_VTLBSIZEEXT_SHIFT)
 #define MIPS_CONF4_AE		(_ULCAST_(1) << 28)
@@ -562,9 +620,13 @@
 #define MIPS_CONF5_LLB		(_ULCAST_(1) << 4)
 #define MIPS_CONF5_MVH		(_ULCAST_(1) << 5)
 #define MIPS_CONF5_VP		(_ULCAST_(1) << 7)
+#define MIPS_CONF5_SBRI		(_ULCAST_(1) << 6)
 #define MIPS_CONF5_FRE		(_ULCAST_(1) << 8)
 #define MIPS_CONF5_UFE		(_ULCAST_(1) << 9)
 #define MIPS_CONF5_L2C		(_ULCAST_(1) << 10)
+#define MIPS_CONF5_CA2		(_ULCAST_(1) << 14)
+#define MIPS_CONF5_MI		(_ULCAST_(1) << 17)
+#define MIPS_CONF5_CRCP		(_ULCAST_(1) << 18)
 #define MIPS_CONF5_MSAEN	(_ULCAST_(1) << 27)
 #define MIPS_CONF5_EVA		(_ULCAST_(1) << 28)
 #define MIPS_CONF5_CV		(_ULCAST_(1) << 29)
@@ -573,6 +635,8 @@
 #define MIPS_CONF6_SYND		(_ULCAST_(1) << 13)
 /* proAptiv FTLB on/off bit */
 #define MIPS_CONF6_FTLBEN	(_ULCAST_(1) << 15)
+/* Loongson-3 FTLB on/off bit */
+#define MIPS_CONF6_FTLBDIS	(_ULCAST_(1) << 22)
 /* FTLB probability bits */
 #define MIPS_CONF6_FTLBP_SHIFT	(16)
 
@@ -582,19 +646,95 @@
 
 #define MIPS_CONF7_IAR		(_ULCAST_(1) << 10)
 #define MIPS_CONF7_AR		(_ULCAST_(1) << 16)
-/* FTLB probability bits for R6 */
-#define MIPS_CONF7_FTLBP_SHIFT	(18)
+
+/* Ingenic HPTLB off bits */
+#define XBURST_PAGECTRL_HPTLB_DIS 0xa9000000
+
+/* Ingenic Config7 bits */
+#define MIPS_CONF7_BTB_LOOP_EN	(_ULCAST_(1) << 4)
+
+/* Config7 Bits specific to MIPS Technologies. */
+
+/* Performance counters implemented Per TC */
+#define MTI_CONF7_PTC		(_ULCAST_(1) << 19)
+
+/* WatchLo* register definitions */
+#define MIPS_WATCHLO_IRW	(_ULCAST_(0x7) << 0)
+
+/* WatchHi* register definitions */
+#define MIPS_WATCHHI_M		(_ULCAST_(1) << 31)
+#define MIPS_WATCHHI_G		(_ULCAST_(1) << 30)
+#define MIPS_WATCHHI_WM		(_ULCAST_(0x3) << 28)
+#define MIPS_WATCHHI_WM_R_RVA	(_ULCAST_(0) << 28)
+#define MIPS_WATCHHI_WM_R_GPA	(_ULCAST_(1) << 28)
+#define MIPS_WATCHHI_WM_G_GVA	(_ULCAST_(2) << 28)
+#define MIPS_WATCHHI_EAS	(_ULCAST_(0x3) << 24)
+#define MIPS_WATCHHI_ASID	(_ULCAST_(0xff) << 16)
+#define MIPS_WATCHHI_MASK	(_ULCAST_(0x1ff) << 3)
+#define MIPS_WATCHHI_I		(_ULCAST_(1) << 2)
+#define MIPS_WATCHHI_R		(_ULCAST_(1) << 1)
+#define MIPS_WATCHHI_W		(_ULCAST_(1) << 0)
+#define MIPS_WATCHHI_IRW	(_ULCAST_(0x7) << 0)
+
+/* PerfCnt control register definitions */
+#define MIPS_PERFCTRL_EXL	(_ULCAST_(1) << 0)
+#define MIPS_PERFCTRL_K		(_ULCAST_(1) << 1)
+#define MIPS_PERFCTRL_S		(_ULCAST_(1) << 2)
+#define MIPS_PERFCTRL_U		(_ULCAST_(1) << 3)
+#define MIPS_PERFCTRL_IE	(_ULCAST_(1) << 4)
+#define MIPS_PERFCTRL_EVENT_S	5
+#define MIPS_PERFCTRL_EVENT	(_ULCAST_(0x3ff) << MIPS_PERFCTRL_EVENT_S)
+#define MIPS_PERFCTRL_PCTD	(_ULCAST_(1) << 15)
+#define MIPS_PERFCTRL_EC	(_ULCAST_(0x3) << 23)
+#define MIPS_PERFCTRL_EC_R	(_ULCAST_(0) << 23)
+#define MIPS_PERFCTRL_EC_RI	(_ULCAST_(1) << 23)
+#define MIPS_PERFCTRL_EC_G	(_ULCAST_(2) << 23)
+#define MIPS_PERFCTRL_EC_GRI	(_ULCAST_(3) << 23)
+#define MIPS_PERFCTRL_W		(_ULCAST_(1) << 30)
+#define MIPS_PERFCTRL_M		(_ULCAST_(1) << 31)
+
+/* PerfCnt control register MT extensions used by MIPS cores */
+#define MIPS_PERFCTRL_VPEID_S	16
+#define MIPS_PERFCTRL_VPEID	(_ULCAST_(0xf) << MIPS_PERFCTRL_VPEID_S)
+#define MIPS_PERFCTRL_TCID_S	22
+#define MIPS_PERFCTRL_TCID	(_ULCAST_(0xff) << MIPS_PERFCTRL_TCID_S)
+#define MIPS_PERFCTRL_MT_EN	(_ULCAST_(0x3) << 20)
+#define MIPS_PERFCTRL_MT_EN_ALL	(_ULCAST_(0) << 20)
+#define MIPS_PERFCTRL_MT_EN_VPE	(_ULCAST_(1) << 20)
+#define MIPS_PERFCTRL_MT_EN_TC	(_ULCAST_(2) << 20)
+
+/* PerfCnt control register MT extensions used by BMIPS5000 */
+#define BRCM_PERFCTRL_TC	(_ULCAST_(1) << 30)
+
+/* PerfCnt control register MT extensions used by Netlogic XLR */
+#define XLR_PERFCTRL_ALLTHREADS	(_ULCAST_(1) << 13)
 
 /* MAAR bit definitions */
+#define MIPS_MAAR_VH		(_U64CAST_(1) << 63)
 #define MIPS_MAAR_ADDR		((BIT_ULL(BITS_PER_LONG - 12) - 1) << 12)
 #define MIPS_MAAR_ADDR_SHIFT	12
 #define MIPS_MAAR_S		(_ULCAST_(1) << 1)
-#define MIPS_MAAR_V		(_ULCAST_(1) << 0)
+#define MIPS_MAAR_VL		(_ULCAST_(1) << 0)
+
+/* MAARI bit definitions */
+#define MIPS_MAARI_INDEX	(_ULCAST_(0x3f) << 0)
+
+/* EBase bit definitions */
+#define MIPS_EBASE_CPUNUM_SHIFT	0
+#define MIPS_EBASE_CPUNUM	(_ULCAST_(0x3ff) << 0)
+#define MIPS_EBASE_WG_SHIFT	11
+#define MIPS_EBASE_WG		(_ULCAST_(1) << 11)
+#define MIPS_EBASE_BASE_SHIFT	12
+#define MIPS_EBASE_BASE		(~_ULCAST_((1 << MIPS_EBASE_BASE_SHIFT) - 1))
 
 /* CMGCRBase bit definitions */
 #define MIPS_CMGCRB_BASE	11
 #define MIPS_CMGCRF_BASE	(~_ULCAST_((1 << MIPS_CMGCRB_BASE) - 1))
 
+/* LLAddr bit definitions */
+#define MIPS_LLADDR_LLB_SHIFT	0
+#define MIPS_LLADDR_LLB		(_ULCAST_(1) << MIPS_LLADDR_LLB_SHIFT)
+
 /*
  * Bits in the MIPS32 Memory Segmentation registers.
  */
@@ -626,6 +766,8 @@
 #define MIPS_PWFIELD_PTEI_SHIFT	0
 #define MIPS_PWFIELD_PTEI_MASK	0x0000003f
 
+#define MIPS_PWSIZE_PS_SHIFT	30
+#define MIPS_PWSIZE_PS_MASK	0x40000000
 #define MIPS_PWSIZE_GDW_SHIFT	24
 #define MIPS_PWSIZE_GDW_MASK	0x3f000000
 #define MIPS_PWSIZE_UDW_SHIFT	18
@@ -639,6 +781,12 @@
 
 #define MIPS_PWCTL_PWEN_SHIFT	31
 #define MIPS_PWCTL_PWEN_MASK	0x80000000
+#define MIPS_PWCTL_XK_SHIFT	28
+#define MIPS_PWCTL_XK_MASK	0x10000000
+#define MIPS_PWCTL_XS_SHIFT	27
+#define MIPS_PWCTL_XS_MASK	0x08000000
+#define MIPS_PWCTL_XU_SHIFT	26
+#define MIPS_PWCTL_XU_MASK	0x04000000
 #define MIPS_PWCTL_DPH_SHIFT	7
 #define MIPS_PWCTL_DPH_MASK	0x00000080
 #define MIPS_PWCTL_HUGEPG_SHIFT	6
@@ -646,6 +794,94 @@
 #define MIPS_PWCTL_PSN_SHIFT	0
 #define MIPS_PWCTL_PSN_MASK	0x0000003f
 
+/* GuestCtl0 fields */
+#define MIPS_GCTL0_GM_SHIFT	31
+#define MIPS_GCTL0_GM		(_ULCAST_(1) << MIPS_GCTL0_GM_SHIFT)
+#define MIPS_GCTL0_RI_SHIFT	30
+#define MIPS_GCTL0_RI		(_ULCAST_(1) << MIPS_GCTL0_RI_SHIFT)
+#define MIPS_GCTL0_MC_SHIFT	29
+#define MIPS_GCTL0_MC		(_ULCAST_(1) << MIPS_GCTL0_MC_SHIFT)
+#define MIPS_GCTL0_CP0_SHIFT	28
+#define MIPS_GCTL0_CP0		(_ULCAST_(1) << MIPS_GCTL0_CP0_SHIFT)
+#define MIPS_GCTL0_AT_SHIFT	26
+#define MIPS_GCTL0_AT		(_ULCAST_(0x3) << MIPS_GCTL0_AT_SHIFT)
+#define MIPS_GCTL0_GT_SHIFT	25
+#define MIPS_GCTL0_GT		(_ULCAST_(1) << MIPS_GCTL0_GT_SHIFT)
+#define MIPS_GCTL0_CG_SHIFT	24
+#define MIPS_GCTL0_CG		(_ULCAST_(1) << MIPS_GCTL0_CG_SHIFT)
+#define MIPS_GCTL0_CF_SHIFT	23
+#define MIPS_GCTL0_CF		(_ULCAST_(1) << MIPS_GCTL0_CF_SHIFT)
+#define MIPS_GCTL0_G1_SHIFT	22
+#define MIPS_GCTL0_G1		(_ULCAST_(1) << MIPS_GCTL0_G1_SHIFT)
+#define MIPS_GCTL0_G0E_SHIFT	19
+#define MIPS_GCTL0_G0E		(_ULCAST_(1) << MIPS_GCTL0_G0E_SHIFT)
+#define MIPS_GCTL0_PT_SHIFT	18
+#define MIPS_GCTL0_PT		(_ULCAST_(1) << MIPS_GCTL0_PT_SHIFT)
+#define MIPS_GCTL0_RAD_SHIFT	9
+#define MIPS_GCTL0_RAD		(_ULCAST_(1) << MIPS_GCTL0_RAD_SHIFT)
+#define MIPS_GCTL0_DRG_SHIFT	8
+#define MIPS_GCTL0_DRG		(_ULCAST_(1) << MIPS_GCTL0_DRG_SHIFT)
+#define MIPS_GCTL0_G2_SHIFT	7
+#define MIPS_GCTL0_G2		(_ULCAST_(1) << MIPS_GCTL0_G2_SHIFT)
+#define MIPS_GCTL0_GEXC_SHIFT	2
+#define MIPS_GCTL0_GEXC		(_ULCAST_(0x1f) << MIPS_GCTL0_GEXC_SHIFT)
+#define MIPS_GCTL0_SFC2_SHIFT	1
+#define MIPS_GCTL0_SFC2		(_ULCAST_(1) << MIPS_GCTL0_SFC2_SHIFT)
+#define MIPS_GCTL0_SFC1_SHIFT	0
+#define MIPS_GCTL0_SFC1		(_ULCAST_(1) << MIPS_GCTL0_SFC1_SHIFT)
+
+/* GuestCtl0.AT Guest address translation control */
+#define MIPS_GCTL0_AT_ROOT	1  /* Guest MMU under Root control */
+#define MIPS_GCTL0_AT_GUEST	3  /* Guest MMU under Guest control */
+
+/* GuestCtl0.GExcCode Hypervisor exception cause codes */
+#define MIPS_GCTL0_GEXC_GPSI	0  /* Guest Privileged Sensitive Instruction */
+#define MIPS_GCTL0_GEXC_GSFC	1  /* Guest Software Field Change */
+#define MIPS_GCTL0_GEXC_HC	2  /* Hypercall */
+#define MIPS_GCTL0_GEXC_GRR	3  /* Guest Reserved Instruction Redirect */
+#define MIPS_GCTL0_GEXC_GVA	8  /* Guest Virtual Address available */
+#define MIPS_GCTL0_GEXC_GHFC	9  /* Guest Hardware Field Change */
+#define MIPS_GCTL0_GEXC_GPA	10 /* Guest Physical Address available */
+
+/* GuestCtl0Ext fields */
+#define MIPS_GCTL0EXT_RPW_SHIFT	8
+#define MIPS_GCTL0EXT_RPW	(_ULCAST_(0x3) << MIPS_GCTL0EXT_RPW_SHIFT)
+#define MIPS_GCTL0EXT_NCC_SHIFT	6
+#define MIPS_GCTL0EXT_NCC	(_ULCAST_(0x3) << MIPS_GCTL0EXT_NCC_SHIFT)
+#define MIPS_GCTL0EXT_CGI_SHIFT	4
+#define MIPS_GCTL0EXT_CGI	(_ULCAST_(1) << MIPS_GCTL0EXT_CGI_SHIFT)
+#define MIPS_GCTL0EXT_FCD_SHIFT	3
+#define MIPS_GCTL0EXT_FCD	(_ULCAST_(1) << MIPS_GCTL0EXT_FCD_SHIFT)
+#define MIPS_GCTL0EXT_OG_SHIFT	2
+#define MIPS_GCTL0EXT_OG	(_ULCAST_(1) << MIPS_GCTL0EXT_OG_SHIFT)
+#define MIPS_GCTL0EXT_BG_SHIFT	1
+#define MIPS_GCTL0EXT_BG	(_ULCAST_(1) << MIPS_GCTL0EXT_BG_SHIFT)
+#define MIPS_GCTL0EXT_MG_SHIFT	0
+#define MIPS_GCTL0EXT_MG	(_ULCAST_(1) << MIPS_GCTL0EXT_MG_SHIFT)
+
+/* GuestCtl0Ext.RPW Root page walk configuration */
+#define MIPS_GCTL0EXT_RPW_BOTH	0  /* Root PW for GPA->RPA and RVA->RPA */
+#define MIPS_GCTL0EXT_RPW_GPA	2  /* Root PW for GPA->RPA */
+#define MIPS_GCTL0EXT_RPW_RVA	3  /* Root PW for RVA->RPA */
+
+/* GuestCtl0Ext.NCC Nested cache coherency attributes */
+#define MIPS_GCTL0EXT_NCC_IND	0  /* Guest CCA independent of Root CCA */
+#define MIPS_GCTL0EXT_NCC_MOD	1  /* Guest CCA modified by Root CCA */
+
+/* GuestCtl1 fields */
+#define MIPS_GCTL1_ID_SHIFT	0
+#define MIPS_GCTL1_ID_WIDTH	8
+#define MIPS_GCTL1_ID		(_ULCAST_(0xff) << MIPS_GCTL1_ID_SHIFT)
+#define MIPS_GCTL1_RID_SHIFT	16
+#define MIPS_GCTL1_RID_WIDTH	8
+#define MIPS_GCTL1_RID		(_ULCAST_(0xff) << MIPS_GCTL1_RID_SHIFT)
+#define MIPS_GCTL1_EID_SHIFT	24
+#define MIPS_GCTL1_EID_WIDTH	8
+#define MIPS_GCTL1_EID		(_ULCAST_(0xff) << MIPS_GCTL1_EID_SHIFT)
+
+/* GuestID reserved for root context */
+#define MIPS_GCTL1_ROOT_GUESTID	0
+
 /* CDMMBase register bit definitions */
 #define MIPS_CDMMBASE_SIZE_SHIFT 0
 #define MIPS_CDMMBASE_SIZE	(_ULCAST_(511) << MIPS_CDMMBASE_SIZE_SHIFT)
@@ -654,6 +890,24 @@
 #define MIPS_CDMMBASE_ADDR_SHIFT 11
 #define MIPS_CDMMBASE_ADDR_START 15
 
+/* RDHWR register numbers */
+#define MIPS_HWR_CPUNUM		0	/* CPU number */
+#define MIPS_HWR_SYNCISTEP	1	/* SYNCI step size */
+#define MIPS_HWR_CC		2	/* Cycle counter */
+#define MIPS_HWR_CCRES		3	/* Cycle counter resolution */
+#define MIPS_HWR_ULR		29	/* UserLocal */
+#define MIPS_HWR_IMPL1		30	/* Implementation dependent */
+#define MIPS_HWR_IMPL2		31	/* Implementation dependent */
+
+/* Bits in HWREna register */
+#define MIPS_HWRENA_CPUNUM	(_ULCAST_(1) << MIPS_HWR_CPUNUM)
+#define MIPS_HWRENA_SYNCISTEP	(_ULCAST_(1) << MIPS_HWR_SYNCISTEP)
+#define MIPS_HWRENA_CC		(_ULCAST_(1) << MIPS_HWR_CC)
+#define MIPS_HWRENA_CCRES	(_ULCAST_(1) << MIPS_HWR_CCRES)
+#define MIPS_HWRENA_ULR		(_ULCAST_(1) << MIPS_HWR_ULR)
+#define MIPS_HWRENA_IMPL1	(_ULCAST_(1) << MIPS_HWR_IMPL1)
+#define MIPS_HWRENA_IMPL2	(_ULCAST_(1) << MIPS_HWR_IMPL2)
+
 /*
  * Bitfields in the TX39 family CP0 Configuration Register 3
  */
@@ -697,6 +951,31 @@
 /* Disable Branch Return Cache */
 #define R10K_DIAG_D_BRC		(_ULCAST_(1) << 22)
 
+/* Flush ITLB */
+#define LOONGSON_DIAG_ITLB	(_ULCAST_(1) << 2)
+/* Flush DTLB */
+#define LOONGSON_DIAG_DTLB	(_ULCAST_(1) << 3)
+/* Flush VTLB */
+#define LOONGSON_DIAG_VTLB	(_ULCAST_(1) << 12)
+/* Flush FTLB */
+#define LOONGSON_DIAG_FTLB	(_ULCAST_(1) << 13)
+
+/* CvmCtl register field definitions */
+#define CVMCTL_IPPCI_SHIFT	7
+#define CVMCTL_IPPCI		(_U64CAST_(0x7) << CVMCTL_IPPCI_SHIFT)
+#define CVMCTL_IPTI_SHIFT	4
+#define CVMCTL_IPTI		(_U64CAST_(0x7) << CVMCTL_IPTI_SHIFT)
+
+/* CvmMemCtl2 register field definitions */
+#define CVMMEMCTL2_INHIBITTS	(_U64CAST_(1) << 17)
+
+/* CvmVMConfig register field definitions */
+#define CVMVMCONF_DGHT		(_U64CAST_(1) << 60)
+#define CVMVMCONF_MMUSIZEM1_S	12
+#define CVMVMCONF_MMUSIZEM1	(_U64CAST_(0xff) << CVMVMCONF_MMUSIZEM1_S)
+#define CVMVMCONF_RMMUSIZEM1_S	0
+#define CVMVMCONF_RMMUSIZEM1	(_U64CAST_(0xff) << CVMVMCONF_RMMUSIZEM1_S)
+
 /*
  * Coprocessor 1 (FPU) register names
  */
@@ -780,9 +1059,12 @@
 /*
  * Bits 22:20 of the FPU Status Register will be read as 0,
  * and should be written as zero.
+ * MAC2008 was removed in Release 5 so we still treat it as
+ * reserved.
  */
 #define FPU_CSR_RSVD	(_ULCAST_(7) << 20)
 
+#define FPU_CSR_MAC2008	(_ULCAST_(1) << 20)
 #define FPU_CSR_ABS2008	(_ULCAST_(1) << 19)
 #define FPU_CSR_NAN2008	(_ULCAST_(1) << 18)
 
@@ -827,14 +1109,14 @@
  * Macros for handling the ISA mode bit for MIPS16 and microMIPS.
  */
 #if defined(CONFIG_SYS_SUPPORTS_MIPS16) || \
-	defined(CONFIG_SYS_SUPPORTS_MICROMIPS)
+    defined(CONFIG_SYS_SUPPORTS_MICROMIPS)
 #define get_isa16_mode(x)		((x) & 0x1)
 #define msk_isa16_mode(x)		((x) & ~0x1)
-#define set_isa16_mode(x)		do { (x) |= 0x1; } while (0)
+#define set_isa16_mode(x)		do { (x) |= 0x1; } while(0)
 #else
 #define get_isa16_mode(x)		0
 #define msk_isa16_mode(x)		(x)
-#define set_isa16_mode(x)		do { } while (0)
+#define set_isa16_mode(x)		do { } while(0)
 #endif
 
 /*
@@ -849,6 +1131,123 @@ static inline int mm_insn_16bit(u16 insn)
 }
 
 /*
+ * Helper macros for generating raw instruction encodings in inline asm.
+ */
+#ifdef CONFIG_CPU_MICROMIPS
+#define _ASM_INSN16_IF_MM(_enc)			\
+	".insn\n\t"				\
+	".hword (" #_enc ")\n\t"
+#define _ASM_INSN32_IF_MM(_enc)			\
+	".insn\n\t"				\
+	".hword ((" #_enc ") >> 16)\n\t"	\
+	".hword ((" #_enc ") & 0xffff)\n\t"
+#else
+#define _ASM_INSN_IF_MIPS(_enc)			\
+	".insn\n\t"				\
+	".word (" #_enc ")\n\t"
+#endif
+
+#ifndef _ASM_INSN16_IF_MM
+#define _ASM_INSN16_IF_MM(_enc)
+#endif
+#ifndef _ASM_INSN32_IF_MM
+#define _ASM_INSN32_IF_MM(_enc)
+#endif
+#ifndef _ASM_INSN_IF_MIPS
+#define _ASM_INSN_IF_MIPS(_enc)
+#endif
+
+/*
+ * parse_r var, r - Helper assembler macro for parsing register names.
+ *
+ * This converts the register name in $n form provided in \r to the
+ * corresponding register number, which is assigned to the variable \var. It is
+ * needed to allow explicit encoding of instructions in inline assembly where
+ * registers are chosen by the compiler in $n form, allowing us to avoid using
+ * fixed register numbers.
+ *
+ * It also allows newer instructions (not implemented by the assembler) to be
+ * transparently implemented using assembler macros, instead of needing separate
+ * cases depending on toolchain support.
+ *
+ * Simple usage example:
+ * __asm__ __volatile__("parse_r __rt, %0\n\t"
+ *			".insn\n\t"
+ *			"# di    %0\n\t"
+ *			".word   (0x41606000 | (__rt << 16))"
+ *			: "=r" (status);
+ */
+
+/* Match an individual register number and assign to \var */
+#define _IFC_REG(n)				\
+	".ifc	\\r, $" #n "\n\t"		\
+	"\\var	= " #n "\n\t"			\
+	".endif\n\t"
+
+__asm__(".macro	parse_r var r\n\t"
+	"\\var	= -1\n\t"
+	_IFC_REG(0)  _IFC_REG(1)  _IFC_REG(2)  _IFC_REG(3)
+	_IFC_REG(4)  _IFC_REG(5)  _IFC_REG(6)  _IFC_REG(7)
+	_IFC_REG(8)  _IFC_REG(9)  _IFC_REG(10) _IFC_REG(11)
+	_IFC_REG(12) _IFC_REG(13) _IFC_REG(14) _IFC_REG(15)
+	_IFC_REG(16) _IFC_REG(17) _IFC_REG(18) _IFC_REG(19)
+	_IFC_REG(20) _IFC_REG(21) _IFC_REG(22) _IFC_REG(23)
+	_IFC_REG(24) _IFC_REG(25) _IFC_REG(26) _IFC_REG(27)
+	_IFC_REG(28) _IFC_REG(29) _IFC_REG(30) _IFC_REG(31)
+	".iflt	\\var\n\t"
+	".error	\"Unable to parse register name \\r\"\n\t"
+	".endif\n\t"
+	".endm");
+
+#undef _IFC_REG
+
+/*
+ * C macros for generating assembler macros for common instruction formats.
+ *
+ * The names of the operands can be chosen by the caller, and the encoding of
+ * register operand \<Rn> is assigned to __<Rn> where it can be accessed from
+ * the ENC encodings.
+ */
+
+/* Instructions with no operands */
+#define _ASM_MACRO_0(OP, ENC)						\
+	__asm__(".macro	" #OP "\n\t"					\
+		ENC							\
+		".endm")
+
+/* Instructions with 1 register operand & 1 immediate operand */
+#define _ASM_MACRO_1R1I(OP, R1, I2, ENC)				\
+	__asm__(".macro	" #OP " " #R1 ", " #I2 "\n\t"			\
+		"parse_r __" #R1 ", \\" #R1 "\n\t"			\
+		ENC							\
+		".endm")
+
+/* Instructions with 2 register operands */
+#define _ASM_MACRO_2R(OP, R1, R2, ENC)					\
+	__asm__(".macro	" #OP " " #R1 ", " #R2 "\n\t"			\
+		"parse_r __" #R1 ", \\" #R1 "\n\t"			\
+		"parse_r __" #R2 ", \\" #R2 "\n\t"			\
+		ENC							\
+		".endm")
+
+/* Instructions with 3 register operands */
+#define _ASM_MACRO_3R(OP, R1, R2, R3, ENC)				\
+	__asm__(".macro	" #OP " " #R1 ", " #R2 ", " #R3 "\n\t"		\
+		"parse_r __" #R1 ", \\" #R1 "\n\t"			\
+		"parse_r __" #R2 ", \\" #R2 "\n\t"			\
+		"parse_r __" #R3 ", \\" #R3 "\n\t"			\
+		ENC							\
+		".endm")
+
+/* Instructions with 2 register operands and 1 optional select operand */
+#define _ASM_MACRO_2R_1S(OP, R1, R2, SEL3, ENC)				\
+	__asm__(".macro	" #OP " " #R1 ", " #R2 ", " #SEL3 " = 0\n\t"	\
+		"parse_r __" #R1 ", \\" #R1 "\n\t"			\
+		"parse_r __" #R2 ", \\" #R2 "\n\t"			\
+		ENC							\
+		".endm")
+
+/*
  * TLB Invalidate Flush
  */
 static inline void tlbinvf(void)
@@ -856,7 +1255,9 @@ static inline void tlbinvf(void)
 	__asm__ __volatile__(
 		".set push\n\t"
 		".set noreorder\n\t"
-		".word 0x42000004\n\t" /* tlbinvf */
+		"# tlbinvf\n\t"
+		_ASM_INSN_IF_MIPS(0x42000004)
+		_ASM_INSN32_IF_MM(0x0000537c)
 		".set pop");
 }
 
@@ -911,40 +1312,55 @@ do {								\
  * Macros to access the system control coprocessor
  */
 
-#define __read_32bit_c0_register(source, sel)				\
+#define ___read_32bit_c0_register(source, sel, vol)			\
 ({ unsigned int __res;							\
 	if (sel == 0)							\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
 			"mfc0\t%0, " #source "\n\t"			\
 			: "=r" (__res));				\
 	else								\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
+			".set\tpush\n\t"				\
 			".set\tmips32\n\t"				\
 			"mfc0\t%0, " #source ", " #sel "\n\t"		\
-			".set\tmips0\n\t"				\
+			".set\tpop\n\t"					\
 			: "=r" (__res));				\
 	__res;								\
 })
 
-#define __read_64bit_c0_register(source, sel)				\
+#define ___read_64bit_c0_register(source, sel, vol)			\
 ({ unsigned long long __res;						\
 	if (sizeof(unsigned long) == 4)					\
-		__res = __read_64bit_c0_split(source, sel);		\
+		__res = __read_64bit_c0_split(source, sel, vol);	\
 	else if (sel == 0)						\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
+			".set\tpush\n\t"				\
 			".set\tmips3\n\t"				\
 			"dmfc0\t%0, " #source "\n\t"			\
-			".set\tmips0"					\
+			".set\tpop"					\
 			: "=r" (__res));				\
 	else								\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
 			"dmfc0\t%0, " #source ", " #sel "\n\t"		\
-			".set\tmips0"					\
+			".set\tpop"					\
 			: "=r" (__res));				\
 	__res;								\
 })
 
+#define __read_32bit_c0_register(source, sel)				\
+	___read_32bit_c0_register(source, sel, __volatile__)
+
+#define __read_const_32bit_c0_register(source, sel)			\
+	___read_32bit_c0_register(source, sel,)
+
+#define __read_64bit_c0_register(source, sel)				\
+	___read_64bit_c0_register(source, sel, __volatile__)
+
+#define __read_const_64bit_c0_register(source, sel)			\
+	___read_64bit_c0_register(source, sel,)
+
 #define __write_32bit_c0_register(register, sel, value)			\
 do {									\
 	if (sel == 0)							\
@@ -953,9 +1369,10 @@ do {									\
 			: : "Jr" ((unsigned int)(value)));		\
 	else								\
 		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
 			".set\tmips32\n\t"				\
 			"mtc0\t%z0, " #register ", " #sel "\n\t"	\
-			".set\tmips0"					\
+			".set\tpop"					\
 			: : "Jr" ((unsigned int)(value)));		\
 } while (0)
 
@@ -965,15 +1382,17 @@ do {									\
 		__write_64bit_c0_split(register, sel, value);		\
 	else if (sel == 0)						\
 		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
 			".set\tmips3\n\t"				\
 			"dmtc0\t%z0, " #register "\n\t"			\
-			".set\tmips0"					\
+			".set\tpop"					\
 			: : "Jr" (value));				\
 	else								\
 		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
 			"dmtc0\t%z0, " #register ", " #sel "\n\t"	\
-			".set\tmips0"					\
+			".set\tpop"					\
 			: : "Jr" (value));				\
 } while (0)
 
@@ -982,6 +1401,11 @@ do {									\
 	(unsigned long) __read_32bit_c0_register(reg, sel) :		\
 	(unsigned long) __read_64bit_c0_register(reg, sel))
 
+#define __read_const_ulong_c0_register(reg, sel)			\
+	((sizeof(unsigned long) == 4) ?					\
+	(unsigned long) __read_const_32bit_c0_register(reg, sel) :	\
+	(unsigned long) __read_const_64bit_c0_register(reg, sel))
+
 #define __write_ulong_c0_register(reg, sel, val)			\
 do {									\
 	if (sizeof(unsigned long) == 4)					\
@@ -1012,27 +1436,27 @@ do {									\
  * These versions are only needed for systems with more than 38 bits of
  * physical address space running the 32-bit kernel.  That's none atm :-)
  */
-#define __read_64bit_c0_split(source, sel)				\
+#define __read_64bit_c0_split(source, sel, vol)				\
 ({									\
 	unsigned long long __val;					\
 									\
 	if (sel == 0)							\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
-			"dmfc0\t%M0, " #source "\n\t"			\
-			"dsll\t%L0, %M0, 32\n\t"			\
-			"dsra\t%M0, %M0, 32\n\t"			\
-			"dsra\t%L0, %L0, 32\n\t"			\
-			".set\tmips0"					\
+			"dmfc0\t%L0, " #source "\n\t"			\
+			"dsra\t%M0, %L0, 32\n\t"			\
+			"sll\t%L0, %L0, 0\n\t"				\
+			".set\tpop"					\
 			: "=r" (__val));				\
 	else								\
-		__asm__ __volatile__(					\
+		__asm__ vol(						\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
-			"dmfc0\t%M0, " #source ", " #sel "\n\t"		\
-			"dsll\t%L0, %M0, 32\n\t"			\
-			"dsra\t%M0, %M0, 32\n\t"			\
-			"dsra\t%L0, %L0, 32\n\t"			\
-			".set\tmips0"					\
+			"dmfc0\t%L0, " #source ", " #sel "\n\t"		\
+			"dsra\t%M0, %L0, 32\n\t"			\
+			"sll\t%L0, %L0, 0\n\t"				\
+			".set\tpop"					\
 			: "=r" (__val));				\
 									\
 	__val;								\
@@ -1040,60 +1464,77 @@ do {									\
 
 #define __write_64bit_c0_split(source, sel, val)			\
 do {									\
-	if (sel == 0)							\
+	unsigned long long __tmp = (val);				\
+	if (MIPS_ISA_REV >= 2)						\
 		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
+			".set\t" MIPS_ISA_LEVEL "\n\t"			\
+			"dins\t%L0, %M0, 32, 32\n\t"			\
+			"dmtc0\t%L0, " #source ", " #sel "\n\t"		\
+			".set\tpop"					\
+			: "+r" (__tmp));				\
+	else if (sel == 0)						\
+		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
 			"dsll\t%L0, %L0, 32\n\t"			\
 			"dsrl\t%L0, %L0, 32\n\t"			\
 			"dsll\t%M0, %M0, 32\n\t"			\
 			"or\t%L0, %L0, %M0\n\t"				\
 			"dmtc0\t%L0, " #source "\n\t"			\
-			".set\tmips0"					\
-			: : "r" (val));					\
+			".set\tpop"					\
+			: "+r" (__tmp));				\
 	else								\
 		__asm__ __volatile__(					\
+			".set\tpush\n\t"				\
 			".set\tmips64\n\t"				\
 			"dsll\t%L0, %L0, 32\n\t"			\
 			"dsrl\t%L0, %L0, 32\n\t"			\
 			"dsll\t%M0, %M0, 32\n\t"			\
 			"or\t%L0, %L0, %M0\n\t"				\
 			"dmtc0\t%L0, " #source ", " #sel "\n\t"		\
-			".set\tmips0"					\
-			: : "r" (val));					\
+			".set\tpop"					\
+			: "+r" (__tmp));				\
 } while (0)
 
-#define __readx_32bit_c0_register(source)				\
+#ifndef TOOLCHAIN_SUPPORTS_XPA
+_ASM_MACRO_2R_1S(mfhc0, rt, rs, sel,
+	_ASM_INSN_IF_MIPS(0x40400000 | __rt << 16 | __rs << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x000000f4 | __rt << 21 | __rs << 16 | \\sel << 11));
+_ASM_MACRO_2R_1S(mthc0, rt, rd, sel,
+	_ASM_INSN_IF_MIPS(0x40c00000 | __rt << 16 | __rd << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x000002f4 | __rt << 21 | __rd << 16 | \\sel << 11));
+#define _ASM_SET_XPA ""
+#else	/* !TOOLCHAIN_SUPPORTS_XPA */
+#define _ASM_SET_XPA ".set\txpa\n\t"
+#endif
+
+#define __readx_32bit_c0_register(source, sel)				\
 ({									\
 	unsigned int __res;						\
 									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
-	"	.set	noat					\n"	\
 	"	.set	mips32r2				\n"	\
-	"	.insn						\n"	\
-	"	# mfhc0 $1, %1					\n"	\
-	"	.word	(0x40410000 | ((%1 & 0x1f) << 11))	\n"	\
-	"	move	%0, $1					\n"	\
+	_ASM_SET_XPA							\
+	"	mfhc0	%0, " #source ", %1			\n"	\
 	"	.set	pop					\n"	\
 	: "=r" (__res)							\
-	: "i" (source));						\
+	: "i" (sel));							\
 	__res;								\
 })
 
-#define __writex_32bit_c0_register(register, value)			\
-({									\
+#define __writex_32bit_c0_register(register, sel, value)		\
+do {									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
-	"	.set	noat					\n"	\
 	"	.set	mips32r2				\n"	\
-	"	move	$1, %0					\n"	\
-	"	# mthc0 $1, %1					\n"	\
-	"	.insn						\n"	\
-	"	.word	(0x40c10000 | ((%1 & 0x1f) << 11))	\n"	\
+	_ASM_SET_XPA							\
+	"	mthc0	%z0, " #register ", %1			\n"	\
 	"	.set	pop					\n"	\
 	:								\
-	: "r" (value), "i" (register));					\
-})
+	: "Jr" (value), "i" (sel));					\
+} while (0)
 
 #define read_c0_index()		__read_32bit_c0_register($0, 0)
 #define write_c0_index(val)	__write_32bit_c0_register($0, 0, val)
@@ -1104,24 +1545,35 @@ do {									\
 #define read_c0_entrylo0()	__read_ulong_c0_register($2, 0)
 #define write_c0_entrylo0(val)	__write_ulong_c0_register($2, 0, val)
 
-#define readx_c0_entrylo0()	__readx_32bit_c0_register(2)
-#define writex_c0_entrylo0(val)	__writex_32bit_c0_register(2, val)
+#define readx_c0_entrylo0()	__readx_32bit_c0_register($2, 0)
+#define writex_c0_entrylo0(val)	__writex_32bit_c0_register($2, 0, val)
 
 #define read_c0_entrylo1()	__read_ulong_c0_register($3, 0)
 #define write_c0_entrylo1(val)	__write_ulong_c0_register($3, 0, val)
 
-#define readx_c0_entrylo1()	__readx_32bit_c0_register(3)
-#define writex_c0_entrylo1(val)	__writex_32bit_c0_register(3, val)
+#define readx_c0_entrylo1()	__readx_32bit_c0_register($3, 0)
+#define writex_c0_entrylo1(val)	__writex_32bit_c0_register($3, 0, val)
 
 #define read_c0_conf()		__read_32bit_c0_register($3, 0)
 #define write_c0_conf(val)	__write_32bit_c0_register($3, 0, val)
 
+#define read_c0_globalnumber()	__read_32bit_c0_register($3, 1)
+
 #define read_c0_context()	__read_ulong_c0_register($4, 0)
 #define write_c0_context(val)	__write_ulong_c0_register($4, 0, val)
 
+#define read_c0_contextconfig()		__read_32bit_c0_register($4, 1)
+#define write_c0_contextconfig(val)	__write_32bit_c0_register($4, 1, val)
+
 #define read_c0_userlocal()	__read_ulong_c0_register($4, 2)
 #define write_c0_userlocal(val) __write_ulong_c0_register($4, 2, val)
 
+#define read_c0_xcontextconfig()	__read_ulong_c0_register($4, 3)
+#define write_c0_xcontextconfig(val)	__write_ulong_c0_register($4, 3, val)
+
+#define read_c0_memorymapid()		__read_32bit_c0_register($4, 5)
+#define write_c0_memorymapid(val)	__write_32bit_c0_register($4, 5, val)
+
 #define read_c0_pagemask()	__read_32bit_c0_register($5, 0)
 #define write_c0_pagemask(val)	__write_32bit_c0_register($5, 0, val)
 
@@ -1139,6 +1591,9 @@ do {									\
 #define read_c0_badvaddr()	__read_ulong_c0_register($8, 0)
 #define write_c0_badvaddr(val)	__write_ulong_c0_register($8, 0, val)
 
+#define read_c0_badinstr()	__read_32bit_c0_register($8, 1)
+#define read_c0_badinstrp()	__read_32bit_c0_register($8, 2)
+
 #define read_c0_count()		__read_32bit_c0_register($9, 0)
 #define write_c0_count(val)	__write_32bit_c0_register($9, 0, val)
 
@@ -1151,9 +1606,21 @@ do {									\
 #define read_c0_entryhi()	__read_ulong_c0_register($10, 0)
 #define write_c0_entryhi(val)	__write_ulong_c0_register($10, 0, val)
 
+#define read_c0_guestctl1()	__read_32bit_c0_register($10, 4)
+#define write_c0_guestctl1(val)	__write_32bit_c0_register($10, 4, val)
+
+#define read_c0_guestctl2()	__read_32bit_c0_register($10, 5)
+#define write_c0_guestctl2(val)	__write_32bit_c0_register($10, 5, val)
+
+#define read_c0_guestctl3()	__read_32bit_c0_register($10, 6)
+#define write_c0_guestctl3(val)	__write_32bit_c0_register($10, 6, val)
+
 #define read_c0_compare()	__read_32bit_c0_register($11, 0)
 #define write_c0_compare(val)	__write_32bit_c0_register($11, 0, val)
 
+#define read_c0_guestctl0ext()	__read_32bit_c0_register($11, 4)
+#define write_c0_guestctl0ext(val) __write_32bit_c0_register($11, 4, val)
+
 #define read_c0_compare2()	__read_32bit_c0_register($11, 6) /* pnx8550 */
 #define write_c0_compare2(val)	__write_32bit_c0_register($11, 6, val)
 
@@ -1164,13 +1631,19 @@ do {									\
 
 #define write_c0_status(val)	__write_32bit_c0_register($12, 0, val)
 
+#define read_c0_guestctl0()	__read_32bit_c0_register($12, 6)
+#define write_c0_guestctl0(val)	__write_32bit_c0_register($12, 6, val)
+
+#define read_c0_gtoffset()	__read_32bit_c0_register($12, 7)
+#define write_c0_gtoffset(val)	__write_32bit_c0_register($12, 7, val)
+
 #define read_c0_cause()		__read_32bit_c0_register($13, 0)
 #define write_c0_cause(val)	__write_32bit_c0_register($13, 0, val)
 
 #define read_c0_epc()		__read_ulong_c0_register($14, 0)
 #define write_c0_epc(val)	__write_ulong_c0_register($14, 0, val)
 
-#define read_c0_prid()		__read_32bit_c0_register($15, 0)
+#define read_c0_prid()		__read_const_32bit_c0_register($15, 0)
 
 #define read_c0_cmgcrbase()	__read_ulong_c0_register($15, 3)
 
@@ -1349,6 +1822,9 @@ do {									\
 #define read_c0_ebase()		__read_32bit_c0_register($15, 1)
 #define write_c0_ebase(val)	__write_32bit_c0_register($15, 1, val)
 
+#define read_c0_ebase_64()	__read_64bit_c0_register($15, 1)
+#define write_c0_ebase_64(val)	__write_64bit_c0_register($15, 1, val)
+
 #define read_c0_cdmmbase()	__read_ulong_c0_register($15, 2)
 #define write_c0_cdmmbase(val)	__write_ulong_c0_register($15, 2, val)
 
@@ -1375,6 +1851,12 @@ do {									\
 #define read_c0_pwctl()		__read_32bit_c0_register($6, 6)
 #define write_c0_pwctl(val)	__write_32bit_c0_register($6, 6, val)
 
+#define read_c0_pgd()		__read_64bit_c0_register($9, 7)
+#define write_c0_pgd(val)	__write_64bit_c0_register($9, 7, val)
+
+#define read_c0_kpgd()		__read_64bit_c0_register($31, 7)
+#define write_c0_kpgd(val)	__write_64bit_c0_register($31, 7, val)
+
 /* Cavium OCTEON (cnMIPS) */
 #define read_c0_cvmcount()	__read_ulong_c0_register($9, 6)
 #define write_c0_cvmcount(val)	__write_ulong_c0_register($9, 6, val)
@@ -1384,6 +1866,13 @@ do {									\
 
 #define read_c0_cvmmemctl()	__read_64bit_c0_register($11, 7)
 #define write_c0_cvmmemctl(val) __write_64bit_c0_register($11, 7, val)
+
+#define read_c0_cvmmemctl2()	__read_64bit_c0_register($16, 6)
+#define write_c0_cvmmemctl2(val) __write_64bit_c0_register($16, 6, val)
+
+#define read_c0_cvmvmconfig()	__read_64bit_c0_register($16, 7)
+#define write_c0_cvmvmconfig(val) __write_64bit_c0_register($16, 7, val)
+
 /*
  * The cacheerr registers are not standardized.	 On OCTEON, they are
  * 64 bits wide.
@@ -1439,6 +1928,306 @@ do {									\
 #define read_c0_brcm_sleepcount()	__read_32bit_c0_register($22, 7)
 #define write_c0_brcm_sleepcount(val)	__write_32bit_c0_register($22, 7, val)
 
+/* Ingenic page ctrl register */
+#define write_c0_page_ctrl(val)	__write_32bit_c0_register($5, 4, val)
+
+/*
+ * Macros to access the guest system control coprocessor
+ */
+
+#ifndef TOOLCHAIN_SUPPORTS_VIRT
+_ASM_MACRO_2R_1S(mfgc0, rt, rs, sel,
+	_ASM_INSN_IF_MIPS(0x40600000 | __rt << 16 | __rs << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x000004fc | __rt << 21 | __rs << 16 | \\sel << 11));
+_ASM_MACRO_2R_1S(dmfgc0, rt, rs, sel,
+	_ASM_INSN_IF_MIPS(0x40600100 | __rt << 16 | __rs << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x580004fc | __rt << 21 | __rs << 16 | \\sel << 11));
+_ASM_MACRO_2R_1S(mtgc0, rt, rd, sel,
+	_ASM_INSN_IF_MIPS(0x40600200 | __rt << 16 | __rd << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x000006fc | __rt << 21 | __rd << 16 | \\sel << 11));
+_ASM_MACRO_2R_1S(dmtgc0, rt, rd, sel,
+	_ASM_INSN_IF_MIPS(0x40600300 | __rt << 16 | __rd << 11 | \\sel)
+	_ASM_INSN32_IF_MM(0x580006fc | __rt << 21 | __rd << 16 | \\sel << 11));
+_ASM_MACRO_0(tlbgp,    _ASM_INSN_IF_MIPS(0x42000010)
+		       _ASM_INSN32_IF_MM(0x0000017c));
+_ASM_MACRO_0(tlbgr,    _ASM_INSN_IF_MIPS(0x42000009)
+		       _ASM_INSN32_IF_MM(0x0000117c));
+_ASM_MACRO_0(tlbgwi,   _ASM_INSN_IF_MIPS(0x4200000a)
+		       _ASM_INSN32_IF_MM(0x0000217c));
+_ASM_MACRO_0(tlbgwr,   _ASM_INSN_IF_MIPS(0x4200000e)
+		       _ASM_INSN32_IF_MM(0x0000317c));
+_ASM_MACRO_0(tlbginvf, _ASM_INSN_IF_MIPS(0x4200000c)
+		       _ASM_INSN32_IF_MM(0x0000517c));
+#define _ASM_SET_VIRT ""
+#else	/* !TOOLCHAIN_SUPPORTS_VIRT */
+#define _ASM_SET_VIRT ".set\tvirt\n\t"
+#endif
+
+#define __read_32bit_gc0_register(source, sel)				\
+({ int __res;								\
+	__asm__ __volatile__(						\
+		".set\tpush\n\t"					\
+		".set\tmips32r2\n\t"					\
+		_ASM_SET_VIRT						\
+		"mfgc0\t%0, " #source ", %1\n\t"			\
+		".set\tpop"						\
+		: "=r" (__res)						\
+		: "i" (sel));						\
+	__res;								\
+})
+
+#define __read_64bit_gc0_register(source, sel)				\
+({ unsigned long long __res;						\
+	__asm__ __volatile__(						\
+		".set\tpush\n\t"					\
+		".set\tmips64r2\n\t"					\
+		_ASM_SET_VIRT						\
+		"dmfgc0\t%0, " #source ", %1\n\t"			\
+		".set\tpop"						\
+		: "=r" (__res)						\
+		: "i" (sel));						\
+	__res;								\
+})
+
+#define __write_32bit_gc0_register(register, sel, value)		\
+do {									\
+	__asm__ __volatile__(						\
+		".set\tpush\n\t"					\
+		".set\tmips32r2\n\t"					\
+		_ASM_SET_VIRT						\
+		"mtgc0\t%z0, " #register ", %1\n\t"			\
+		".set\tpop"						\
+		: : "Jr" ((unsigned int)(value)),			\
+		    "i" (sel));						\
+} while (0)
+
+#define __write_64bit_gc0_register(register, sel, value)		\
+do {									\
+	__asm__ __volatile__(						\
+		".set\tpush\n\t"					\
+		".set\tmips64r2\n\t"					\
+		_ASM_SET_VIRT						\
+		"dmtgc0\t%z0, " #register ", %1\n\t"			\
+		".set\tpop"						\
+		: : "Jr" (value),					\
+		    "i" (sel));						\
+} while (0)
+
+#define __read_ulong_gc0_register(reg, sel)				\
+	((sizeof(unsigned long) == 4) ?					\
+	(unsigned long) __read_32bit_gc0_register(reg, sel) :		\
+	(unsigned long) __read_64bit_gc0_register(reg, sel))
+
+#define __write_ulong_gc0_register(reg, sel, val)			\
+do {									\
+	if (sizeof(unsigned long) == 4)					\
+		__write_32bit_gc0_register(reg, sel, val);		\
+	else								\
+		__write_64bit_gc0_register(reg, sel, val);		\
+} while (0)
+
+#define read_gc0_index()		__read_32bit_gc0_register($0, 0)
+#define write_gc0_index(val)		__write_32bit_gc0_register($0, 0, val)
+
+#define read_gc0_entrylo0()		__read_ulong_gc0_register($2, 0)
+#define write_gc0_entrylo0(val)		__write_ulong_gc0_register($2, 0, val)
+
+#define read_gc0_entrylo1()		__read_ulong_gc0_register($3, 0)
+#define write_gc0_entrylo1(val)		__write_ulong_gc0_register($3, 0, val)
+
+#define read_gc0_context()		__read_ulong_gc0_register($4, 0)
+#define write_gc0_context(val)		__write_ulong_gc0_register($4, 0, val)
+
+#define read_gc0_contextconfig()	__read_32bit_gc0_register($4, 1)
+#define write_gc0_contextconfig(val)	__write_32bit_gc0_register($4, 1, val)
+
+#define read_gc0_userlocal()		__read_ulong_gc0_register($4, 2)
+#define write_gc0_userlocal(val)	__write_ulong_gc0_register($4, 2, val)
+
+#define read_gc0_xcontextconfig()	__read_ulong_gc0_register($4, 3)
+#define write_gc0_xcontextconfig(val)	__write_ulong_gc0_register($4, 3, val)
+
+#define read_gc0_pagemask()		__read_32bit_gc0_register($5, 0)
+#define write_gc0_pagemask(val)		__write_32bit_gc0_register($5, 0, val)
+
+#define read_gc0_pagegrain()		__read_32bit_gc0_register($5, 1)
+#define write_gc0_pagegrain(val)	__write_32bit_gc0_register($5, 1, val)
+
+#define read_gc0_segctl0()		__read_ulong_gc0_register($5, 2)
+#define write_gc0_segctl0(val)		__write_ulong_gc0_register($5, 2, val)
+
+#define read_gc0_segctl1()		__read_ulong_gc0_register($5, 3)
+#define write_gc0_segctl1(val)		__write_ulong_gc0_register($5, 3, val)
+
+#define read_gc0_segctl2()		__read_ulong_gc0_register($5, 4)
+#define write_gc0_segctl2(val)		__write_ulong_gc0_register($5, 4, val)
+
+#define read_gc0_pwbase()		__read_ulong_gc0_register($5, 5)
+#define write_gc0_pwbase(val)		__write_ulong_gc0_register($5, 5, val)
+
+#define read_gc0_pwfield()		__read_ulong_gc0_register($5, 6)
+#define write_gc0_pwfield(val)		__write_ulong_gc0_register($5, 6, val)
+
+#define read_gc0_pwsize()		__read_ulong_gc0_register($5, 7)
+#define write_gc0_pwsize(val)		__write_ulong_gc0_register($5, 7, val)
+
+#define read_gc0_wired()		__read_32bit_gc0_register($6, 0)
+#define write_gc0_wired(val)		__write_32bit_gc0_register($6, 0, val)
+
+#define read_gc0_pwctl()		__read_32bit_gc0_register($6, 6)
+#define write_gc0_pwctl(val)		__write_32bit_gc0_register($6, 6, val)
+
+#define read_gc0_hwrena()		__read_32bit_gc0_register($7, 0)
+#define write_gc0_hwrena(val)		__write_32bit_gc0_register($7, 0, val)
+
+#define read_gc0_badvaddr()		__read_ulong_gc0_register($8, 0)
+#define write_gc0_badvaddr(val)		__write_ulong_gc0_register($8, 0, val)
+
+#define read_gc0_badinstr()		__read_32bit_gc0_register($8, 1)
+#define write_gc0_badinstr(val)		__write_32bit_gc0_register($8, 1, val)
+
+#define read_gc0_badinstrp()		__read_32bit_gc0_register($8, 2)
+#define write_gc0_badinstrp(val)	__write_32bit_gc0_register($8, 2, val)
+
+#define read_gc0_count()		__read_32bit_gc0_register($9, 0)
+
+#define read_gc0_entryhi()		__read_ulong_gc0_register($10, 0)
+#define write_gc0_entryhi(val)		__write_ulong_gc0_register($10, 0, val)
+
+#define read_gc0_compare()		__read_32bit_gc0_register($11, 0)
+#define write_gc0_compare(val)		__write_32bit_gc0_register($11, 0, val)
+
+#define read_gc0_status()		__read_32bit_gc0_register($12, 0)
+#define write_gc0_status(val)		__write_32bit_gc0_register($12, 0, val)
+
+#define read_gc0_intctl()		__read_32bit_gc0_register($12, 1)
+#define write_gc0_intctl(val)		__write_32bit_gc0_register($12, 1, val)
+
+#define read_gc0_cause()		__read_32bit_gc0_register($13, 0)
+#define write_gc0_cause(val)		__write_32bit_gc0_register($13, 0, val)
+
+#define read_gc0_epc()			__read_ulong_gc0_register($14, 0)
+#define write_gc0_epc(val)		__write_ulong_gc0_register($14, 0, val)
+
+#define read_gc0_prid()			__read_32bit_gc0_register($15, 0)
+
+#define read_gc0_ebase()		__read_32bit_gc0_register($15, 1)
+#define write_gc0_ebase(val)		__write_32bit_gc0_register($15, 1, val)
+
+#define read_gc0_ebase_64()		__read_64bit_gc0_register($15, 1)
+#define write_gc0_ebase_64(val)		__write_64bit_gc0_register($15, 1, val)
+
+#define read_gc0_config()		__read_32bit_gc0_register($16, 0)
+#define read_gc0_config1()		__read_32bit_gc0_register($16, 1)
+#define read_gc0_config2()		__read_32bit_gc0_register($16, 2)
+#define read_gc0_config3()		__read_32bit_gc0_register($16, 3)
+#define read_gc0_config4()		__read_32bit_gc0_register($16, 4)
+#define read_gc0_config5()		__read_32bit_gc0_register($16, 5)
+#define read_gc0_config6()		__read_32bit_gc0_register($16, 6)
+#define read_gc0_config7()		__read_32bit_gc0_register($16, 7)
+#define write_gc0_config(val)		__write_32bit_gc0_register($16, 0, val)
+#define write_gc0_config1(val)		__write_32bit_gc0_register($16, 1, val)
+#define write_gc0_config2(val)		__write_32bit_gc0_register($16, 2, val)
+#define write_gc0_config3(val)		__write_32bit_gc0_register($16, 3, val)
+#define write_gc0_config4(val)		__write_32bit_gc0_register($16, 4, val)
+#define write_gc0_config5(val)		__write_32bit_gc0_register($16, 5, val)
+#define write_gc0_config6(val)		__write_32bit_gc0_register($16, 6, val)
+#define write_gc0_config7(val)		__write_32bit_gc0_register($16, 7, val)
+
+#define read_gc0_lladdr()		__read_ulong_gc0_register($17, 0)
+#define write_gc0_lladdr(val)		__write_ulong_gc0_register($17, 0, val)
+
+#define read_gc0_watchlo0()		__read_ulong_gc0_register($18, 0)
+#define read_gc0_watchlo1()		__read_ulong_gc0_register($18, 1)
+#define read_gc0_watchlo2()		__read_ulong_gc0_register($18, 2)
+#define read_gc0_watchlo3()		__read_ulong_gc0_register($18, 3)
+#define read_gc0_watchlo4()		__read_ulong_gc0_register($18, 4)
+#define read_gc0_watchlo5()		__read_ulong_gc0_register($18, 5)
+#define read_gc0_watchlo6()		__read_ulong_gc0_register($18, 6)
+#define read_gc0_watchlo7()		__read_ulong_gc0_register($18, 7)
+#define write_gc0_watchlo0(val)		__write_ulong_gc0_register($18, 0, val)
+#define write_gc0_watchlo1(val)		__write_ulong_gc0_register($18, 1, val)
+#define write_gc0_watchlo2(val)		__write_ulong_gc0_register($18, 2, val)
+#define write_gc0_watchlo3(val)		__write_ulong_gc0_register($18, 3, val)
+#define write_gc0_watchlo4(val)		__write_ulong_gc0_register($18, 4, val)
+#define write_gc0_watchlo5(val)		__write_ulong_gc0_register($18, 5, val)
+#define write_gc0_watchlo6(val)		__write_ulong_gc0_register($18, 6, val)
+#define write_gc0_watchlo7(val)		__write_ulong_gc0_register($18, 7, val)
+
+#define read_gc0_watchhi0()		__read_32bit_gc0_register($19, 0)
+#define read_gc0_watchhi1()		__read_32bit_gc0_register($19, 1)
+#define read_gc0_watchhi2()		__read_32bit_gc0_register($19, 2)
+#define read_gc0_watchhi3()		__read_32bit_gc0_register($19, 3)
+#define read_gc0_watchhi4()		__read_32bit_gc0_register($19, 4)
+#define read_gc0_watchhi5()		__read_32bit_gc0_register($19, 5)
+#define read_gc0_watchhi6()		__read_32bit_gc0_register($19, 6)
+#define read_gc0_watchhi7()		__read_32bit_gc0_register($19, 7)
+#define write_gc0_watchhi0(val)		__write_32bit_gc0_register($19, 0, val)
+#define write_gc0_watchhi1(val)		__write_32bit_gc0_register($19, 1, val)
+#define write_gc0_watchhi2(val)		__write_32bit_gc0_register($19, 2, val)
+#define write_gc0_watchhi3(val)		__write_32bit_gc0_register($19, 3, val)
+#define write_gc0_watchhi4(val)		__write_32bit_gc0_register($19, 4, val)
+#define write_gc0_watchhi5(val)		__write_32bit_gc0_register($19, 5, val)
+#define write_gc0_watchhi6(val)		__write_32bit_gc0_register($19, 6, val)
+#define write_gc0_watchhi7(val)		__write_32bit_gc0_register($19, 7, val)
+
+#define read_gc0_xcontext()		__read_ulong_gc0_register($20, 0)
+#define write_gc0_xcontext(val)		__write_ulong_gc0_register($20, 0, val)
+
+#define read_gc0_perfctrl0()		__read_32bit_gc0_register($25, 0)
+#define write_gc0_perfctrl0(val)	__write_32bit_gc0_register($25, 0, val)
+#define read_gc0_perfcntr0()		__read_32bit_gc0_register($25, 1)
+#define write_gc0_perfcntr0(val)	__write_32bit_gc0_register($25, 1, val)
+#define read_gc0_perfcntr0_64()		__read_64bit_gc0_register($25, 1)
+#define write_gc0_perfcntr0_64(val)	__write_64bit_gc0_register($25, 1, val)
+#define read_gc0_perfctrl1()		__read_32bit_gc0_register($25, 2)
+#define write_gc0_perfctrl1(val)	__write_32bit_gc0_register($25, 2, val)
+#define read_gc0_perfcntr1()		__read_32bit_gc0_register($25, 3)
+#define write_gc0_perfcntr1(val)	__write_32bit_gc0_register($25, 3, val)
+#define read_gc0_perfcntr1_64()		__read_64bit_gc0_register($25, 3)
+#define write_gc0_perfcntr1_64(val)	__write_64bit_gc0_register($25, 3, val)
+#define read_gc0_perfctrl2()		__read_32bit_gc0_register($25, 4)
+#define write_gc0_perfctrl2(val)	__write_32bit_gc0_register($25, 4, val)
+#define read_gc0_perfcntr2()		__read_32bit_gc0_register($25, 5)
+#define write_gc0_perfcntr2(val)	__write_32bit_gc0_register($25, 5, val)
+#define read_gc0_perfcntr2_64()		__read_64bit_gc0_register($25, 5)
+#define write_gc0_perfcntr2_64(val)	__write_64bit_gc0_register($25, 5, val)
+#define read_gc0_perfctrl3()		__read_32bit_gc0_register($25, 6)
+#define write_gc0_perfctrl3(val)	__write_32bit_gc0_register($25, 6, val)
+#define read_gc0_perfcntr3()		__read_32bit_gc0_register($25, 7)
+#define write_gc0_perfcntr3(val)	__write_32bit_gc0_register($25, 7, val)
+#define read_gc0_perfcntr3_64()		__read_64bit_gc0_register($25, 7)
+#define write_gc0_perfcntr3_64(val)	__write_64bit_gc0_register($25, 7, val)
+
+#define read_gc0_errorepc()		__read_ulong_gc0_register($30, 0)
+#define write_gc0_errorepc(val)		__write_ulong_gc0_register($30, 0, val)
+
+#define read_gc0_kscratch1()		__read_ulong_gc0_register($31, 2)
+#define read_gc0_kscratch2()		__read_ulong_gc0_register($31, 3)
+#define read_gc0_kscratch3()		__read_ulong_gc0_register($31, 4)
+#define read_gc0_kscratch4()		__read_ulong_gc0_register($31, 5)
+#define read_gc0_kscratch5()		__read_ulong_gc0_register($31, 6)
+#define read_gc0_kscratch6()		__read_ulong_gc0_register($31, 7)
+#define write_gc0_kscratch1(val)	__write_ulong_gc0_register($31, 2, val)
+#define write_gc0_kscratch2(val)	__write_ulong_gc0_register($31, 3, val)
+#define write_gc0_kscratch3(val)	__write_ulong_gc0_register($31, 4, val)
+#define write_gc0_kscratch4(val)	__write_ulong_gc0_register($31, 5, val)
+#define write_gc0_kscratch5(val)	__write_ulong_gc0_register($31, 6, val)
+#define write_gc0_kscratch6(val)	__write_ulong_gc0_register($31, 7, val)
+
+/* Cavium OCTEON (cnMIPS) */
+#define read_gc0_cvmcount()		__read_ulong_gc0_register($9, 6)
+#define write_gc0_cvmcount(val)		__write_ulong_gc0_register($9, 6, val)
+
+#define read_gc0_cvmctl()		__read_64bit_gc0_register($9, 7)
+#define write_gc0_cvmctl(val)		__write_64bit_gc0_register($9, 7, val)
+
+#define read_gc0_cvmmemctl()		__read_64bit_gc0_register($11, 7)
+#define write_gc0_cvmmemctl(val)	__write_64bit_gc0_register($11, 7, val)
+
+#define read_gc0_cvmmemctl2()		__read_64bit_gc0_register($16, 6)
+#define write_gc0_cvmmemctl2(val)	__write_64bit_gc0_register($16, 6, val)
+
 /*
  * Macros to access the floating point coprocessor control registers
  */
@@ -1460,7 +2249,7 @@ do {									\
 })
 
 #define _write_32bit_cp1_register(dest, val, gas_hardfloat)		\
-({									\
+do {									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
 	"	.set	reorder					\n"	\
@@ -1468,7 +2257,7 @@ do {									\
 	"	ctc1	%0,"STR(dest)"				\n"	\
 	"	.set	pop					\n"	\
 	: : "r" (val));							\
-})
+} while (0)
 
 #ifdef GAS_HAS_SET_HARDFLOAT
 #define read_32bit_cp1_register(source)					\
@@ -1482,13 +2271,14 @@ do {									\
 	_write_32bit_cp1_register(dest, val, )
 #endif
 
-#ifdef HAVE_AS_DSP
+#ifdef TOOLCHAIN_SUPPORTS_DSP
 #define rddsp(mask)							\
 ({									\
 	unsigned int __dspctl;						\
 									\
 	__asm__ __volatile__(						\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	rddsp	%0, %x1					\n"	\
 	"	.set pop					\n"	\
@@ -1498,25 +2288,27 @@ do {									\
 })
 
 #define wrdsp(val, mask)						\
-({									\
+do {									\
 	__asm__ __volatile__(						\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	wrdsp	%0, %x1					\n"	\
 	"	.set pop					\n"	\
 	:								\
 	: "r" (val), "i" (mask));					\
-})
+} while (0)
 
 #define mflo0()								\
 ({									\
 	long mflo0;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac0					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mflo0));						\
+	"	.set pop					\n" 	\
+	: "=r" (mflo0)); 						\
 	mflo0;								\
 })
 
@@ -1525,10 +2317,11 @@ do {									\
 	long mflo1;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac1					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mflo1));						\
+	"	.set pop					\n" 	\
+	: "=r" (mflo1)); 						\
 	mflo1;								\
 })
 
@@ -1537,10 +2330,11 @@ do {									\
 	long mflo2;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac2					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mflo2));						\
+	"	.set pop					\n" 	\
+	: "=r" (mflo2)); 						\
 	mflo2;								\
 })
 
@@ -1549,10 +2343,11 @@ do {									\
 	long mflo3;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac3					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mflo3));						\
+	"	.set pop					\n" 	\
+	: "=r" (mflo3)); 						\
 	mflo3;								\
 })
 
@@ -1561,10 +2356,11 @@ do {									\
 	long mfhi0;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac0					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mfhi0));						\
+	"	.set pop					\n" 	\
+	: "=r" (mfhi0)); 						\
 	mfhi0;								\
 })
 
@@ -1573,10 +2369,11 @@ do {									\
 	long mfhi1;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac1					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mfhi1));						\
+	"	.set pop					\n" 	\
+	: "=r" (mfhi1)); 						\
 	mfhi1;								\
 })
 
@@ -1585,10 +2382,11 @@ do {									\
 	long mfhi2;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac2					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mfhi2));						\
+	"	.set pop					\n" 	\
+	: "=r" (mfhi2)); 						\
 	mfhi2;								\
 })
 
@@ -1597,10 +2395,11 @@ do {									\
 	long mfhi3;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac3					\n"	\
-	"	.set pop					\n"	\
-	: "=r" (mfhi3));						\
+	"	.set pop					\n" 	\
+	: "=r" (mfhi3)); 						\
 	mfhi3;								\
 })
 
@@ -1609,6 +2408,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac0					\n"	\
 	"	.set pop					\n"	\
@@ -1620,6 +2420,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac1					\n"	\
 	"	.set pop					\n"	\
@@ -1631,6 +2432,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac2					\n"	\
 	"	.set pop					\n"	\
@@ -1642,6 +2444,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac3					\n"	\
 	"	.set pop					\n"	\
@@ -1653,6 +2456,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac0					\n"	\
 	"	.set pop					\n"	\
@@ -1664,6 +2468,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac1					\n"	\
 	"	.set pop					\n"	\
@@ -1675,6 +2480,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac2					\n"	\
 	"	.set pop					\n"	\
@@ -1686,6 +2492,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac3					\n"	\
 	"	.set pop					\n"	\
@@ -1695,7 +2502,6 @@ do {									\
 
 #else
 
-#ifdef CONFIG_CPU_MICROMIPS
 #define rddsp(mask)							\
 ({									\
 	unsigned int __res;						\
@@ -1704,8 +2510,8 @@ do {									\
 	"	.set	push					\n"	\
 	"	.set	noat					\n"	\
 	"	# rddsp $1, %x1					\n"	\
-	"	.hword	((0x0020067c | (%x1 << 14)) >> 16)	\n"	\
-	"	.hword	((0x0020067c | (%x1 << 14)) & 0xffff)	\n"	\
+	_ASM_INSN_IF_MIPS(0x7c000cb8 | (%x1 << 16))			\
+	_ASM_INSN32_IF_MM(0x0020067c | (%x1 << 14))			\
 	"	move	%0, $1					\n"	\
 	"	.set	pop					\n"	\
 	: "=r" (__res)							\
@@ -1714,28 +2520,28 @@ do {									\
 })
 
 #define wrdsp(val, mask)						\
-({									\
+do {									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
 	"	.set	noat					\n"	\
 	"	move	$1, %0					\n"	\
 	"	# wrdsp $1, %x1					\n"	\
-	"	.hword	((0x0020167c | (%x1 << 14)) >> 16)	\n"	\
-	"	.hword	((0x0020167c | (%x1 << 14)) & 0xffff)	\n"	\
+	_ASM_INSN_IF_MIPS(0x7c2004f8 | (%x1 << 11))			\
+	_ASM_INSN32_IF_MM(0x0020167c | (%x1 << 14))			\
 	"	.set	pop					\n"	\
 	:								\
 	: "r" (val), "i" (mask));					\
-})
+} while (0)
 
-#define _umips_dsp_mfxxx(ins)						\
+#define _dsp_mfxxx(ins)							\
 ({									\
 	unsigned long __treg;						\
 									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
 	"	.set	noat					\n"	\
-	"	.hword	0x0001					\n"	\
-	"	.hword	%x1					\n"	\
+	_ASM_INSN_IF_MIPS(0x00000810 | %X1)				\
+	_ASM_INSN32_IF_MM(0x0001007c | %x1)				\
 	"	move	%0, $1					\n"	\
 	"	.set	pop					\n"	\
 	: "=r" (__treg)							\
@@ -1743,101 +2549,28 @@ do {									\
 	__treg;								\
 })
 
-#define _umips_dsp_mtxxx(val, ins)					\
-({									\
+#define _dsp_mtxxx(val, ins)						\
+do {									\
 	__asm__ __volatile__(						\
 	"	.set	push					\n"	\
 	"	.set	noat					\n"	\
 	"	move	$1, %0					\n"	\
-	"	.hword	0x0001					\n"	\
-	"	.hword	%x1					\n"	\
+	_ASM_INSN_IF_MIPS(0x00200011 | %X1)				\
+	_ASM_INSN32_IF_MM(0x0001207c | %x1)				\
 	"	.set	pop					\n"	\
 	:								\
 	: "r" (val), "i" (ins));					\
-})
-
-#define _umips_dsp_mflo(reg) _umips_dsp_mfxxx((reg << 14) | 0x107c)
-#define _umips_dsp_mfhi(reg) _umips_dsp_mfxxx((reg << 14) | 0x007c)
-
-#define _umips_dsp_mtlo(val, reg) _umips_dsp_mtxxx(val, ((reg << 14) | 0x307c))
-#define _umips_dsp_mthi(val, reg) _umips_dsp_mtxxx(val, ((reg << 14) | 0x207c))
-
-#define mflo0() _umips_dsp_mflo(0)
-#define mflo1() _umips_dsp_mflo(1)
-#define mflo2() _umips_dsp_mflo(2)
-#define mflo3() _umips_dsp_mflo(3)
+} while (0)
 
-#define mfhi0() _umips_dsp_mfhi(0)
-#define mfhi1() _umips_dsp_mfhi(1)
-#define mfhi2() _umips_dsp_mfhi(2)
-#define mfhi3() _umips_dsp_mfhi(3)
+#ifdef CONFIG_CPU_MICROMIPS
 
-#define mtlo0(x) _umips_dsp_mtlo(x, 0)
-#define mtlo1(x) _umips_dsp_mtlo(x, 1)
-#define mtlo2(x) _umips_dsp_mtlo(x, 2)
-#define mtlo3(x) _umips_dsp_mtlo(x, 3)
+#define _dsp_mflo(reg) _dsp_mfxxx((reg << 14) | 0x1000)
+#define _dsp_mfhi(reg) _dsp_mfxxx((reg << 14) | 0x0000)
 
-#define mthi0(x) _umips_dsp_mthi(x, 0)
-#define mthi1(x) _umips_dsp_mthi(x, 1)
-#define mthi2(x) _umips_dsp_mthi(x, 2)
-#define mthi3(x) _umips_dsp_mthi(x, 3)
+#define _dsp_mtlo(val, reg) _dsp_mtxxx(val, ((reg << 14) | 0x1000))
+#define _dsp_mthi(val, reg) _dsp_mtxxx(val, ((reg << 14) | 0x0000))
 
 #else  /* !CONFIG_CPU_MICROMIPS */
-#define rddsp(mask)							\
-({									\
-	unsigned int __res;						\
-									\
-	__asm__ __volatile__(						\
-	"	.set	push				\n"		\
-	"	.set	noat				\n"		\
-	"	# rddsp $1, %x1				\n"		\
-	"	.word	0x7c000cb8 | (%x1 << 16)	\n"		\
-	"	move	%0, $1				\n"		\
-	"	.set	pop				\n"		\
-	: "=r" (__res)							\
-	: "i" (mask));							\
-	__res;								\
-})
-
-#define wrdsp(val, mask)						\
-({									\
-	__asm__ __volatile__(						\
-	"	.set	push					\n"	\
-	"	.set	noat					\n"	\
-	"	move	$1, %0					\n"	\
-	"	# wrdsp $1, %x1					\n"	\
-	"	.word	0x7c2004f8 | (%x1 << 11)		\n"	\
-	"	.set	pop					\n"	\
-	:								\
-	: "r" (val), "i" (mask));					\
-})
-
-#define _dsp_mfxxx(ins)							\
-({									\
-	unsigned long __treg;						\
-									\
-	__asm__ __volatile__(						\
-	"	.set	push					\n"	\
-	"	.set	noat					\n"	\
-	"	.word	(0x00000810 | %1)			\n"	\
-	"	move	%0, $1					\n"	\
-	"	.set	pop					\n"	\
-	: "=r" (__treg)							\
-	: "i" (ins));							\
-	__treg;								\
-})
-
-#define _dsp_mtxxx(val, ins)						\
-({									\
-	__asm__ __volatile__(						\
-	"	.set	push					\n"	\
-	"	.set	noat					\n"	\
-	"	move	$1, %0					\n"	\
-	"	.word	(0x00200011 | %1)			\n"	\
-	"	.set	pop					\n"	\
-	:								\
-	: "r" (val), "i" (ins));					\
-})
 
 #define _dsp_mflo(reg) _dsp_mfxxx((reg << 21) | 0x0002)
 #define _dsp_mfhi(reg) _dsp_mfxxx((reg << 21) | 0x0000)
@@ -1845,6 +2578,8 @@ do {									\
 #define _dsp_mtlo(val, reg) _dsp_mtxxx(val, ((reg << 11) | 0x0002))
 #define _dsp_mthi(val, reg) _dsp_mtxxx(val, ((reg << 11) | 0x0000))
 
+#endif /* CONFIG_CPU_MICROMIPS */
+
 #define mflo0() _dsp_mflo(0)
 #define mflo1() _dsp_mflo(1)
 #define mflo2() _dsp_mflo(2)
@@ -1865,7 +2600,6 @@ do {									\
 #define mthi2(x) _dsp_mthi(x, 2)
 #define mthi3(x) _dsp_mthi(x, 3)
 
-#endif /* CONFIG_CPU_MICROMIPS */
 #endif
 
 /*
@@ -1935,54 +2669,123 @@ static inline void tlb_write_random(void)
 }
 
 /*
- * Manipulate bits in a c0 register.
+ * Guest TLB operations.
+ *
+ * It is responsibility of the caller to take care of any TLB hazards.
  */
-#define __BUILD_SET_C0(name)					\
+static inline void guest_tlb_probe(void)
+{
+	__asm__ __volatile__(
+		".set push\n\t"
+		".set noreorder\n\t"
+		_ASM_SET_VIRT
+		"tlbgp\n\t"
+		".set pop");
+}
+
+static inline void guest_tlb_read(void)
+{
+	__asm__ __volatile__(
+		".set push\n\t"
+		".set noreorder\n\t"
+		_ASM_SET_VIRT
+		"tlbgr\n\t"
+		".set pop");
+}
+
+static inline void guest_tlb_write_indexed(void)
+{
+	__asm__ __volatile__(
+		".set push\n\t"
+		".set noreorder\n\t"
+		_ASM_SET_VIRT
+		"tlbgwi\n\t"
+		".set pop");
+}
+
+static inline void guest_tlb_write_random(void)
+{
+	__asm__ __volatile__(
+		".set push\n\t"
+		".set noreorder\n\t"
+		_ASM_SET_VIRT
+		"tlbgwr\n\t"
+		".set pop");
+}
+
+/*
+ * Guest TLB Invalidate Flush
+ */
+static inline void guest_tlbinvf(void)
+{
+	__asm__ __volatile__(
+		".set push\n\t"
+		".set noreorder\n\t"
+		_ASM_SET_VIRT
+		"tlbginvf\n\t"
+		".set pop");
+}
+
+/*
+ * Manipulate bits in a register.
+ */
+#define __BUILD_SET_COMMON(name)				\
 static inline unsigned int					\
-set_c0_##name(unsigned int set)					\
+set_##name(unsigned int set)					\
 {								\
 	unsigned int res, new;					\
 								\
-	res = read_c0_##name();					\
+	res = read_##name();					\
 	new = res | set;					\
-	write_c0_##name(new);					\
+	write_##name(new);					\
 								\
 	return res;						\
 }								\
 								\
 static inline unsigned int					\
-clear_c0_##name(unsigned int clear)				\
+clear_##name(unsigned int clear)				\
 {								\
 	unsigned int res, new;					\
 								\
-	res = read_c0_##name();					\
+	res = read_##name();					\
 	new = res & ~clear;					\
-	write_c0_##name(new);					\
+	write_##name(new);					\
 								\
 	return res;						\
 }								\
 								\
 static inline unsigned int					\
-change_c0_##name(unsigned int change, unsigned int val)		\
+change_##name(unsigned int change, unsigned int val)		\
 {								\
 	unsigned int res, new;					\
 								\
-	res = read_c0_##name();					\
+	res = read_##name();					\
 	new = res & ~change;					\
 	new |= (val & change);					\
-	write_c0_##name(new);					\
+	write_##name(new);					\
 								\
 	return res;						\
 }
 
+/*
+ * Manipulate bits in a c0 register.
+ */
+#define __BUILD_SET_C0(name)	__BUILD_SET_COMMON(c0_##name)
+
 __BUILD_SET_C0(status)
 __BUILD_SET_C0(cause)
 __BUILD_SET_C0(config)
 __BUILD_SET_C0(config5)
+__BUILD_SET_C0(config7)
 __BUILD_SET_C0(intcontrol)
 __BUILD_SET_C0(intctl)
 __BUILD_SET_C0(srsmap)
 __BUILD_SET_C0(pagegrain)
+__BUILD_SET_C0(guestctl0)
+__BUILD_SET_C0(guestctl0ext)
+__BUILD_SET_C0(guestctl1)
+__BUILD_SET_C0(guestctl2)
+__BUILD_SET_C0(guestctl3)
 __BUILD_SET_C0(brcm_config_0)
 __BUILD_SET_C0(brcm_bus_pll)
 __BUILD_SET_C0(brcm_reset)
@@ -1992,12 +2795,23 @@ __BUILD_SET_C0(brcm_config)
 __BUILD_SET_C0(brcm_mode)
 
 /*
+ * Manipulate bits in a guest c0 register.
+ */
+#define __BUILD_SET_GC0(name)	__BUILD_SET_COMMON(gc0_##name)
+
+__BUILD_SET_GC0(wired)
+__BUILD_SET_GC0(status)
+__BUILD_SET_GC0(cause)
+__BUILD_SET_GC0(ebase)
+__BUILD_SET_GC0(config1)
+
+/*
  * Return low 10 bits of ebase.
  * Note that under KVM (MIPSVZ) this returns vcpu id.
  */
 static inline unsigned int get_ebase_cpunum(void)
 {
-	return read_c0_ebase() & 0x3ff;
+	return read_c0_ebase() & MIPS_EBASE_CPUNUM;
 }
 
 static inline void write_one_tlb(int index, u32 pagemask, u32 hi, u32 low0,