Merge tag 'arc-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc

Pull ARC updates from Vineet Gupta:

 - long due rewrite of do_page_fault

 - refactoring of entry/exit code to utilize the double load/store
   instructions

 - hsdk platform updates

* tag 'arc-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  ARC: [plat-hsdk]: Enable AXI DW DMAC in defconfig
  ARC: [plat-hsdk]: enable DW SPI controller
  ARC: hide unused function unw_hdr_alloc
  ARC: [haps] Add Virtio support
  ARCv2: entry: simplify return to Delay Slot via interrupt
  ARC: entry: EV_Trap expects r10 (vs. r9) to have exception cause
  ARCv2: entry: rewrite to enable use of double load/stores LDD/STD
  ARCv2: entry: avoid a branch
  ARCv2: entry: push out the Z flag unclobber from common EXCEPTION_PROLOGUE
  ARCv2: entry: comments about hardware auto-save on taken interrupts
  ARC: mm: do_page_fault refactor #8: release mmap_sem sooner
  ARC: mm: do_page_fault refactor #7: fold the various error handling
  ARC: mm: do_page_fault refactor #6: error handlers to use same pattern
  ARC: mm: do_page_fault refactor #5: scoot no_context to end
  ARC: mm: do_page_fault refactor #4: consolidate retry related logic
  ARC: mm: do_page_fault refactor #3: tidyup vma access permission code
  ARC: mm: do_page_fault refactor #2: remove short lived variable
  ARC: mm: do_page_fault refactor #1: remove label @good_area

14 files changed, 377 insertions, 340 deletions

diff --git a/arch/arc/boot/dts/haps_hs.dts b/arch/arc/boot/dts/haps_hs.dts
index 1ebfa046492b..44bc522fdec8 100644
--- a/arch/arc/boot/dts/haps_hs.dts
+++ b/arch/arc/boot/dts/haps_hs.dts
@@ -62,5 +62,35 @@
                         #interrupt-cells = <1>;
                         interrupts = <20>;
                 };
+
+                virtio0: virtio@f0100000 {
+                        compatible = "virtio,mmio";
+                        reg = <0xf0100000 0x2000>;
+                        interrupts = <31>;
+                };
+
+                virtio1: virtio@f0102000 {
+                        compatible = "virtio,mmio";
+                        reg = <0xf0102000 0x2000>;
+                        interrupts = <32>;
+                };
+
+                virtio2: virtio@f0104000 {
+                        compatible = "virtio,mmio";
+                        reg = <0xf0104000 0x2000>;
+                        interrupts = <33>;
+                };
+
+                virtio3: virtio@f0106000 {
+                        compatible = "virtio,mmio";
+                        reg = <0xf0106000 0x2000>;
+                        interrupts = <34>;
+                };
+
+                virtio4: virtio@f0108000 {
+                        compatible = "virtio,mmio";
+                        reg = <0xf0108000 0x2000>;
+                        interrupts = <35>;
+                };
         };
 };
diff --git a/arch/arc/boot/dts/hsdk.dts b/arch/arc/boot/dts/hsdk.dts
index 9a45cb093096..bfc7f5f5d6f2 100644
--- a/arch/arc/boot/dts/hsdk.dts
+++ b/arch/arc/boot/dts/hsdk.dts
@@ -8,6 +8,7 @@
  */
 /dts-v1/;
 
+#include <dt-bindings/gpio/gpio.h>
 #include <dt-bindings/reset/snps,hsdk-reset.h>
 
 / {
@@ -252,6 +253,19 @@
                         dma-coherent;
                 };
 
+                spi0: spi@20000 {
+                        compatible = "snps,dw-apb-ssi";
+                        reg = <0x20000 0x100>;
+                        #address-cells = <1>;
+                        #size-cells = <0>;
+                        interrupts = <16>;
+                        num-cs = <2>;
+                        reg-io-width = <4>;
+                        clocks = <&input_clk>;
+                        cs-gpios = <&creg_gpio 0 GPIO_ACTIVE_LOW>,
+                                   <&creg_gpio 1 GPIO_ACTIVE_LOW>;
+                };
+
                 creg_gpio: gpio@14b0 {
                         compatible = "snps,creg-gpio-hsdk";
                         reg = <0x14b0 0x4>;
diff --git a/arch/arc/configs/haps_hs_defconfig b/arch/arc/configs/haps_hs_defconfig
index b117e6c16d41..436f2135bdc1 100644
--- a/arch/arc/configs/haps_hs_defconfig
+++ b/arch/arc/configs/haps_hs_defconfig
@@ -35,10 +35,12 @@ CONFIG_INET=y
 # CONFIG_IPV6 is not set
 # CONFIG_WIRELESS is not set
 CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y
 # CONFIG_STANDALONE is not set
 # CONFIG_PREVENT_FIRMWARE_BUILD is not set
-# CONFIG_BLK_DEV is not set
+CONFIG_VIRTIO_BLK=y
 CONFIG_NETDEVICES=y
+CONFIG_VIRTIO_NET=y
 # CONFIG_NET_VENDOR_ARC is not set
 # CONFIG_NET_VENDOR_BROADCOM is not set
 # CONFIG_NET_VENDOR_INTEL is not set
@@ -68,6 +70,7 @@ CONFIG_FRAMEBUFFER_CONSOLE=y
 CONFIG_LOGO=y
 # CONFIG_HID is not set
 # CONFIG_USB_SUPPORT is not set
+CONFIG_VIRTIO_MMIO=y
 # CONFIG_IOMMU_SUPPORT is not set
 CONFIG_EXT2_FS=y
 CONFIG_EXT2_FS_XATTR=y
diff --git a/arch/arc/configs/hsdk_defconfig b/arch/arc/configs/hsdk_defconfig
index c8fb5d60c53f..403125d9c9a3 100644
--- a/arch/arc/configs/hsdk_defconfig
+++ b/arch/arc/configs/hsdk_defconfig
@@ -46,6 +46,9 @@ CONFIG_SERIAL_8250_CONSOLE=y
 CONFIG_SERIAL_8250_DW=y
 CONFIG_SERIAL_OF_PLATFORM=y
 # CONFIG_HW_RANDOM is not set
+CONFIG_SPI=y
+CONFIG_SPI_DESIGNWARE=y
+CONFIG_SPI_DW_MMIO=y
 CONFIG_GPIOLIB=y
 CONFIG_GPIO_SYSFS=y
 CONFIG_GPIO_DWAPB=y
@@ -66,6 +69,8 @@ CONFIG_MMC=y
 CONFIG_MMC_SDHCI=y
 CONFIG_MMC_SDHCI_PLTFM=y
 CONFIG_MMC_DW=y
+CONFIG_DMADEVICES=y
+CONFIG_DW_AXI_DMAC=y
 CONFIG_EXT3_FS=y
 CONFIG_VFAT_FS=y
 CONFIG_TMPFS=y
diff --git a/arch/arc/include/asm/entry-arcv2.h b/arch/arc/include/asm/entry-arcv2.h
index 225e7df2d8ed..f5ae394ebe06 100644
--- a/arch/arc/include/asm/entry-arcv2.h
+++ b/arch/arc/include/asm/entry-arcv2.h
@@ -7,232 +7,251 @@
 #include <asm/irqflags-arcv2.h>
 #include <asm/thread_info.h>    /* For THREAD_SIZE */
 
+/*
+ * Interrupt/Exception stack layout (pt_regs) for ARCv2
+ *   (End of struct aligned to end of page [unless nested])
+ *
+ *  INTERRUPT                          EXCEPTION
+ *
+ *    manual    ---------------------  manual
+ *              |      orig_r0      |
+ *              |      event/ECR    |
+ *              |      bta          |
+ *              |      user_r25     |
+ *              |      gp           |
+ *              |      fp           |
+ *              |      sp           |
+ *              |      r12          |
+ *              |      r30          |
+ *              |      r58          |
+ *              |      r59          |
+ *  hw autosave ---------------------
+ *    optional  |      r0           |
+ *              |      r1           |
+ *              ~                   ~
+ *              |      r9           |
+ *              |      r10          |
+ *              |      r11          |
+ *              |      blink        |
+ *              |      lpe          |
+ *              |      lps          |
+ *              |      lpc          |
+ *              |      ei base      |
+ *              |      ldi base     |
+ *              |      jli base     |
+ *              ---------------------
+ *  hw autosave |       pc / eret   |
+ *   mandatory  | stat32 / erstatus |
+ *              ---------------------
+ */
+
 /*------------------------------------------------------------------------*/
-.macro INTERRUPT_PROLOGUE       called_from
+.macro INTERRUPT_PROLOGUE
 
-        ; Before jumping to Interrupt Vector, hardware micro-ops did following:
+        ; (A) Before jumping to Interrupt Vector, hardware micro-ops did following:
         ;   1. SP auto-switched to kernel mode stack
-        ;   2. STATUS32.Z flag set to U mode at time of interrupt (U:1, K:0)
-        ;   3. Auto saved: r0-r11, blink, LPE,LPS,LPC, JLI,LDI,EI, PC, STAT32
+        ;   2. STATUS32.Z flag set if in U mode at time of interrupt (U:1,K:0)
+        ;   3. Auto save: (mandatory) Push PC and STAT32 on stack
+        ;                 hardware does even if CONFIG_ARC_IRQ_NO_AUTOSAVE
+        ;   4. Auto save: (optional) r0-r11, blink, LPE,LPS,LPC, JLI,LDI,EI
         ;
-        ; Now manually save: r12, sp, fp, gp, r25
+        ; (B) Manually saved some regs: r12,r25,r30, sp,fp,gp, ACCL pair
 
 #ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
-.ifnc \called_from, exception
-        st.as   r9, [sp, -10]   ; save r9 in it's final stack slot
-        sub     sp, sp, 12      ; skip JLI, LDI, EI
-
-        PUSH    lp_count
-        PUSHAX  lp_start
-        PUSHAX  lp_end
-        PUSH    blink
-
-        PUSH    r11
-        PUSH    r10
-
-        sub     sp, sp, 4       ; skip r9
-
-        PUSH    r8
-        PUSH    r7
-        PUSH    r6
-        PUSH    r5
-        PUSH    r4
-        PUSH    r3
-        PUSH    r2
-        PUSH    r1
-        PUSH    r0
-.endif
-#endif
+        ; carve pt_regs on stack (case #3), PC/STAT32 already on stack
+        sub     sp, sp, SZ_PT_REGS - 8
 
-#ifdef CONFIG_ARC_HAS_ACCL_REGS
-        PUSH    r59
-        PUSH    r58
+        __SAVE_REGFILE_HARD
+#else
+        ; carve pt_regs on stack (case #4), which grew partially already
+        sub     sp, sp, PT_r0
 #endif
 
-        PUSH    r30
-        PUSH    r12
+        __SAVE_REGFILE_SOFT
+.endm
+
+/*------------------------------------------------------------------------*/
+.macro EXCEPTION_PROLOGUE
+
+        ; (A) Before jumping to Exception Vector, hardware micro-ops did following:
+        ;   1. SP auto-switched to kernel mode stack
+        ;   2. STATUS32.Z flag set if in U mode at time of exception (U:1,K:0)
+        ;
+        ; (B) Manually save the complete reg file below
+
+        sub     sp, sp, SZ_PT_REGS      ; carve pt_regs
+
+        ; _HARD saves r10 clobbered by _SOFT as scratch hence comes first
+
+        __SAVE_REGFILE_HARD
+        __SAVE_REGFILE_SOFT
+
+        st      r0, [sp]        ; orig_r0
+
+        lr      r10, [eret]
+        lr      r11, [erstatus]
+        ST2     r10, r11, PT_ret
+
+        lr      r10, [ecr]
+        lr      r11, [erbta]
+        ST2     r10, r11, PT_event
+
+        ; OUTPUT: r10 has ECR expected by EV_Trap
+.endm
+
+/*------------------------------------------------------------------------
+ * This macro saves the registers manually which would normally be autosaved
+ * by hardware on taken interrupts. It is used by
+ *   - exception handlers (which don't have autosave)
+ *   - interrupt autosave disabled due to CONFIG_ARC_IRQ_NO_AUTOSAVE
+ */
+.macro __SAVE_REGFILE_HARD
+
+        ST2     r0,  r1,  PT_r0
+        ST2     r2,  r3,  PT_r2
+        ST2     r4,  r5,  PT_r4
+        ST2     r6,  r7,  PT_r6
+        ST2     r8,  r9,  PT_r8
+        ST2     r10, r11, PT_r10
+
+        st      blink, [sp, PT_blink]
+
+        lr      r10, [lp_end]
+        lr      r11, [lp_start]
+        ST2     r10, r11, PT_lpe
+
+        st      lp_count, [sp, PT_lpc]
+
+        ; skip JLI, LDI, EI for now
+.endm
+
+/*------------------------------------------------------------------------
+ * This macros saves a bunch of other registers which can't be autosaved for
+ * various reasons:
+ *   - r12: the last caller saved scratch reg since hardware saves in pairs so r0-r11
+ *   - r30: free reg, used by gcc as scratch
+ *   - ACCL/ACCH pair when they exist
+ */
+.macro __SAVE_REGFILE_SOFT
+
+        ST2     gp, fp, PT_r26          ; gp (r26), fp (r27)
+
+        st      r12, [sp, PT_sp + 4]
+        st      r30, [sp, PT_sp + 8]
 
         ; Saving pt_regs->sp correctly requires some extra work due to the way
         ; Auto stack switch works
         ;  - U mode: retrieve it from AUX_USER_SP
         ;  - K mode: add the offset from current SP where H/w starts auto push
         ;
-        ; Utilize the fact that Z bit is set if Intr taken in U mode
-        mov.nz  r9, sp
-        add.nz  r9, r9, SZ_PT_REGS - PT_sp - 4
-        bnz     1f
+        ; 1. Utilize the fact that Z bit is set if Intr taken in U mode
+        ; 2. Upon entry SP is always saved (for any inspection, unwinding etc),
+        ;    but on return, restored only if U mode
 
-        lr      r9, [AUX_USER_SP]
-1:
-        PUSH    r9      ; SP
+        lr      r10, [AUX_USER_SP]      ; U mode SP
+
+        ; ISA requires ADD.nz to have same dest and src reg operands
+        mov.nz  r10, sp
+        add.nz  r10, r10, SZ_PT_REGS    ; K mode SP
 
-        PUSH    fp
-        PUSH    gp
+        st      r10, [sp, PT_sp]        ; SP (pt_regs->sp)
 
 #ifdef CONFIG_ARC_CURR_IN_REG
-        PUSH    r25                     ; user_r25
+        st      r25, [sp, PT_user_r25]
         GET_CURR_TASK_ON_CPU    r25
-#else
-        sub     sp, sp, 4
 #endif
 
-.ifnc \called_from, exception
-        sub     sp, sp, 12      ; BTA/ECR/orig_r0 placeholder per pt_regs
-.endif
+#ifdef CONFIG_ARC_HAS_ACCL_REGS
+        ST2     r58, r59, PT_sp + 12
+#endif
 
 .endm
 
 /*------------------------------------------------------------------------*/
-.macro INTERRUPT_EPILOGUE       called_from
+.macro __RESTORE_REGFILE_SOFT
 
-.ifnc \called_from, exception
-        add     sp, sp, 12      ; skip BTA/ECR/orig_r0 placeholderss
-.endif
+        LD2     gp, fp, PT_r26          ; gp (r26), fp (r27)
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-        POP     r25
-#else
-        add     sp, sp, 4
-#endif
+        ld      r12, [sp, PT_sp + 4]
+        ld      r30, [sp, PT_sp + 8]
 
-        POP     gp
-        POP     fp
-
-        ; Don't touch AUX_USER_SP if returning to K mode (Z bit set)
-        ; (Z bit set on K mode is inverse of INTERRUPT_PROLOGUE)
-        add.z   sp, sp, 4
+        ; Restore SP (into AUX_USER_SP) only if returning to U mode
+        ;  - for K mode, it will be implicitly restored as stack is unwound
+        ;  - Z flag set on K is inverse of what hardware does on interrupt entry
+        ;    but that doesn't really matter
         bz      1f
 
-        POPAX   AUX_USER_SP
+        ld      r10, [sp, PT_sp]        ; SP (pt_regs->sp)
+        sr      r10, [AUX_USER_SP]
 1:
-        POP     r12
-        POP     r30
 
-#ifdef CONFIG_ARC_HAS_ACCL_REGS
-        POP     r58
-        POP     r59
+#ifdef CONFIG_ARC_CURR_IN_REG
+        ld      r25, [sp, PT_user_r25]
 #endif
 
-#ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
-.ifnc \called_from, exception
-        POP     r0
-        POP     r1
-        POP     r2
-        POP     r3
-        POP     r4
-        POP     r5
-        POP     r6
-        POP     r7
-        POP     r8
-        POP     r9
-        POP     r10
-        POP     r11
-
-        POP     blink
-        POPAX   lp_end
-        POPAX   lp_start
-
-        POP     r9
-        mov     lp_count, r9
-
-        add     sp, sp, 12      ; skip JLI, LDI, EI
-        ld.as   r9, [sp, -10]   ; reload r9 which got clobbered
-.endif
+#ifdef CONFIG_ARC_HAS_ACCL_REGS
+        LD2     r58, r59, PT_sp + 12
 #endif
+.endm
 
+/*------------------------------------------------------------------------*/
+.macro __RESTORE_REGFILE_HARD
+
+        ld      blink, [sp, PT_blink]
+
+        LD2     r10, r11, PT_lpe
+        sr      r10, [lp_end]
+        sr      r11, [lp_start]
+
+        ld      r10, [sp, PT_lpc]       ; lp_count can't be target of LD
+        mov     lp_count, r10
+
+        LD2     r0,  r1,  PT_r0
+        LD2     r2,  r3,  PT_r2
+        LD2     r4,  r5,  PT_r4
+        LD2     r6,  r7,  PT_r6
+        LD2     r8,  r9,  PT_r8
+        LD2     r10, r11, PT_r10
 .endm
 
+
 /*------------------------------------------------------------------------*/
-.macro EXCEPTION_PROLOGUE
+.macro INTERRUPT_EPILOGUE
 
-        ; Before jumping to Exception Vector, hardware micro-ops did following:
-        ;   1. SP auto-switched to kernel mode stack
-        ;   2. STATUS32.Z flag set to U mode at time of interrupt (U:1,K:0)
-        ;
-        ; Now manually save the complete reg file
-
-        PUSH    r9              ; freeup a register: slot of erstatus
-
-        PUSHAX  eret
-        sub     sp, sp, 12      ; skip JLI, LDI, EI
-        PUSH    lp_count
-        PUSHAX  lp_start
-        PUSHAX  lp_end
-        PUSH    blink
-
-        PUSH    r11
-        PUSH    r10
-
-        ld.as   r9,  [sp, 10]   ; load stashed r9 (status32 stack slot)
-        lr      r10, [erstatus]
-        st.as   r10, [sp, 10]   ; save status32 at it's right stack slot
-
-        PUSH    r9
-        PUSH    r8
-        PUSH    r7
-        PUSH    r6
-        PUSH    r5
-        PUSH    r4
-        PUSH    r3
-        PUSH    r2
-        PUSH    r1
-        PUSH    r0
-
-        ; -- for interrupts, regs above are auto-saved by h/w in that order --
-        ; Now do what ISR prologue does (manually save r12, sp, fp, gp, r25)
-        ;
-        ; Set Z flag if this was from U mode (expected by INTERRUPT_PROLOGUE)
-        ; Although H/w exception micro-ops do set Z flag for U mode (just like
-        ; for interrupts), it could get clobbered in case we soft land here from
-        ; a TLB Miss exception handler (tlbex.S)
+        ; INPUT: r0 has STAT32 of calling context
+        ; INPUT: Z flag set if returning to K mode
 
-        and     r10, r10, STATUS_U_MASK
-        xor.f   0, r10, STATUS_U_MASK
+        ; _SOFT clobbers r10 restored by _HARD hence the order
 
-        INTERRUPT_PROLOGUE  exception
+        __RESTORE_REGFILE_SOFT
 
-        PUSHAX  erbta
-        PUSHAX  ecr             ; r9 contains ECR, expected by EV_Trap
+#ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
+        __RESTORE_REGFILE_HARD
+        add     sp, sp, SZ_PT_REGS - 8
+#else
+        add     sp, sp, PT_r0
+#endif
 
-        PUSH    r0              ; orig_r0
 .endm
 
 /*------------------------------------------------------------------------*/
 .macro EXCEPTION_EPILOGUE
 
-        ; Assumes r0 has PT_status32
-        btst   r0, STATUS_U_BIT ; Z flag set if K, used in INTERRUPT_EPILOGUE
-
-        add     sp, sp, 8       ; orig_r0/ECR don't need restoring
-        POPAX   erbta
-
-        INTERRUPT_EPILOGUE  exception
+        ; INPUT: r0 has STAT32 of calling context
 
-        POP     r0
-        POP     r1
-        POP     r2
-        POP     r3
-        POP     r4
-        POP     r5
-        POP     r6
-        POP     r7
-        POP     r8
-        POP     r9
-        POP     r10
-        POP     r11
+        btst    r0, STATUS_U_BIT        ; Z flag set if K, used in restoring SP
 
-        POP     blink
-        POPAX   lp_end
-        POPAX   lp_start
+        ld      r10, [sp, PT_event + 4]
+        sr      r10, [erbta]
 
-        POP     r9
-        mov     lp_count, r9
+        LD2     r10, r11, PT_ret
+        sr      r10, [eret]
+        sr      r11, [erstatus]
 
-        add     sp, sp, 12      ; skip JLI, LDI, EI
-        POPAX   eret
-        POPAX   erstatus
+        __RESTORE_REGFILE_SOFT
+        __RESTORE_REGFILE_HARD
 
-        ld.as   r9, [sp, -12]   ; reload r9 which got clobbered
+        add     sp, sp, SZ_PT_REGS
 .endm
 
 .macro FAKE_RET_FROM_EXCPN
diff --git a/arch/arc/include/asm/entry-compact.h b/arch/arc/include/asm/entry-compact.h
index 66ba1bf21d28..66a292335ee6 100644
--- a/arch/arc/include/asm/entry-compact.h
+++ b/arch/arc/include/asm/entry-compact.h
@@ -195,8 +195,8 @@
         PUSHAX  CTOP_AUX_EFLAGS
 #endif
 
-        lr      r9, [ecr]
-        st      r9, [sp, PT_event]    /* EV_Trap expects r9 to have ECR */
+        lr      r10, [ecr]
+        st      r10, [sp, PT_event]    /* EV_Trap expects r10 to have ECR */
 .endm
 
 /*--------------------------------------------------------------
diff --git a/arch/arc/include/asm/linkage.h b/arch/arc/include/asm/linkage.h
index 54f5ec5c1759..a0eeb9f8f0a9 100644
--- a/arch/arc/include/asm/linkage.h
+++ b/arch/arc/include/asm/linkage.h
@@ -10,6 +10,24 @@
 
 #ifdef __ASSEMBLY__
 
+.macro ST2 e, o, off
+#ifdef CONFIG_ARC_HAS_LL64
+        std     \e, [sp, \off]
+#else
+        st      \e, [sp, \off]
+        st      \o, [sp, \off+4]
+#endif
+.endm
+
+.macro LD2 e, o, off
+#ifdef CONFIG_ARC_HAS_LL64
+        ldd     \e, [sp, \off]
+#else
+        ld      \e, [sp, \off]
+        ld      \o, [sp, \off+4]
+#endif
+.endm
+
 #define ASM_NL           `      /* use '`' to mark new line in macro */
 
 /* annotation for data we want in DCCM - if enabled in .config */
diff --git a/arch/arc/kernel/asm-offsets.c b/arch/arc/kernel/asm-offsets.c
index dba116535005..1f621e416521 100644
--- a/arch/arc/kernel/asm-offsets.c
+++ b/arch/arc/kernel/asm-offsets.c
@@ -55,7 +55,14 @@ int main(void)
         DEFINE(PT_r5, offsetof(struct pt_regs, r5));
         DEFINE(PT_r6, offsetof(struct pt_regs, r6));
         DEFINE(PT_r7, offsetof(struct pt_regs, r7));
+        DEFINE(PT_r8, offsetof(struct pt_regs, r8));
+        DEFINE(PT_r10, offsetof(struct pt_regs, r10));
+        DEFINE(PT_r26, offsetof(struct pt_regs, r26));
         DEFINE(PT_ret, offsetof(struct pt_regs, ret));
+        DEFINE(PT_blink, offsetof(struct pt_regs, blink));
+        DEFINE(PT_lpe, offsetof(struct pt_regs, lp_end));
+        DEFINE(PT_lpc, offsetof(struct pt_regs, lp_count));
+        DEFINE(PT_user_r25, offsetof(struct pt_regs, user_r25));
 
         DEFINE(SZ_CALLEE_REGS, sizeof(struct callee_regs));
         DEFINE(SZ_PT_REGS, sizeof(struct pt_regs));
diff --git a/arch/arc/kernel/entry-arcv2.S b/arch/arc/kernel/entry-arcv2.S
index 14254b866fdc..12d5f12d10d2 100644
--- a/arch/arc/kernel/entry-arcv2.S
+++ b/arch/arc/kernel/entry-arcv2.S
@@ -67,7 +67,7 @@ reserved:
 
 ENTRY(handle_interrupt)
 
-        INTERRUPT_PROLOGUE  irq
+        INTERRUPT_PROLOGUE
 
         # irq control APIs local_irq_save/restore/disable/enable fiddle with
         # global interrupt enable bits in STATUS32 (.IE for 1 prio, .E[] for 2 prio)
@@ -79,7 +79,7 @@ ENTRY(handle_interrupt)
         #
         # Note this disable is only for consistent book-keeping as further interrupts
         # will be disabled anyways even w/o this. Hardware tracks active interrupts
-        # seperately in AUX_IRQ_ACTIVE.active and will not take new interrupts
+        # seperately in AUX_IRQ_ACT.active and will not take new interrupts
         # unless this one returns (or higher prio becomes pending in 2-prio scheme)
 
         IRQ_DISABLE
@@ -200,17 +200,18 @@ restore_regs:
         ld      r0, [sp, PT_status32]   ; U/K mode at time of entry
         lr      r10, [AUX_IRQ_ACT]
 
-        bmsk    r11, r10, 15    ; AUX_IRQ_ACT.ACTIVE
+        bmsk    r11, r10, 15            ; extract AUX_IRQ_ACT.active
         breq    r11, 0, .Lexcept_ret    ; No intr active, ret from Exception
 
 ;####### Return from Intr #######
 
+.Lisr_ret:
+
 debug_marker_l1:
         ; bbit1.nt r0, STATUS_DE_BIT, .Lintr_ret_to_delay_slot
         btst    r0, STATUS_DE_BIT               ; Z flag set if bit clear
         bnz     .Lintr_ret_to_delay_slot        ; branch if STATUS_DE_BIT set
 
-.Lisr_ret_fast_path:
         ; Handle special case #1: (Entry via Exception, Return via IRQ)
         ;
         ; Exception in U mode, preempted in kernel, Intr taken (K mode), orig
@@ -223,7 +224,7 @@ debug_marker_l1:
         bset.nz r11, r11, AUX_IRQ_ACT_BIT_U     ; NZ means U
         sr      r11, [AUX_IRQ_ACT]
 
-        INTERRUPT_EPILOGUE  irq
+        INTERRUPT_EPILOGUE
         rtie
 
 ;####### Return from Exception / pure kernel mode #######
@@ -244,8 +245,8 @@ debug_marker_syscall:
 ;
 ; IRQ RTIE won't reliably restore DE bit and/or BTA, needs workaround
 ;
-; Solution is return from Intr w/o any delay slot quirks into a kernel trampoline
-; and from pure kernel mode return to delay slot which handles DS bit/BTA correctly
+; Solution is to drop out of interrupt context into pure kernel mode
+; and return from pure kernel mode which does right things for delay slot
 
 .Lintr_ret_to_delay_slot:
 debug_marker_ds:
@@ -254,48 +255,9 @@ debug_marker_ds:
         add     r2, r2, 1
         st      r2, [@intr_to_DE_cnt]
 
-        ld      r2, [sp, PT_ret]
-        ld      r3, [sp, PT_status32]
-
-        ; STAT32 for Int return created from scratch
-        ; (No delay dlot, disable Further intr in trampoline)
-
-        bic     r0, r3, STATUS_U_MASK|STATUS_DE_MASK|STATUS_IE_MASK|STATUS_L_MASK
-        st      r0, [sp, PT_status32]
-
-        mov     r1, .Lintr_ret_to_delay_slot_2
-        st      r1, [sp, PT_ret]
-
-        ; Orig exception PC/STAT32 safekept @orig_r0 and @event stack slots
-        st      r2, [sp, 0]
-        st      r3, [sp, 4]
-
-        b       .Lisr_ret_fast_path
-
-.Lintr_ret_to_delay_slot_2:
-        ; Trampoline to restore orig exception PC/STAT32/BTA/AUX_USER_SP
-        sub     sp, sp, SZ_PT_REGS
-        st      r9, [sp, -4]
-
-        ld      r9, [sp, 0]
-        sr      r9, [eret]
-
-        ld      r9, [sp, 4]
-        sr      r9, [erstatus]
-
-        ; restore AUX_USER_SP if returning to U mode
-        bbit0   r9, STATUS_U_BIT, 1f
-        ld      r9, [sp, PT_sp]
-        sr      r9, [AUX_USER_SP]
-
-1:
-        ld      r9, [sp, 8]
-        sr      r9, [erbta]
-
-        ld      r9, [sp, -4]
-        add     sp, sp, SZ_PT_REGS
-
-        ; return from pure kernel mode to delay slot
-        rtie
+        ; drop out of interrupt context (clear AUX_IRQ_ACT.active)
+        bmskn   r11, r10, 15
+        sr      r11, [AUX_IRQ_ACT]
+        b       .Lexcept_ret
 
 END(ret_from_exception)
diff --git a/arch/arc/kernel/entry-compact.S b/arch/arc/kernel/entry-compact.S
index 7fe59880c16b..5cb0cd7e4eab 100644
--- a/arch/arc/kernel/entry-compact.S
+++ b/arch/arc/kernel/entry-compact.S
@@ -256,7 +256,7 @@ ENTRY(EV_TLBProtV)
 
         EXCEPTION_PROLOGUE
 
-        mov r2, r9      ; ECR set into r9 already
+        mov r2, r10     ; ECR set into r10 already
         lr  r0, [efa]   ; Faulting Data address (not part of pt_regs saved above)
 
         ; Exception auto-disables further Intr/exceptions.
diff --git a/arch/arc/kernel/entry.S b/arch/arc/kernel/entry.S
index a2bfacbcfce1..72be01270e24 100644
--- a/arch/arc/kernel/entry.S
+++ b/arch/arc/kernel/entry.S
@@ -232,8 +232,8 @@ ENTRY(EV_Trap)
         EXCEPTION_PROLOGUE
 
         ;============ TRAP 1   :breakpoints
-        ; Check ECR for trap with arg (PROLOGUE ensures r9 has ECR)
-        bmsk.f 0, r9, 7
+        ; Check ECR for trap with arg (PROLOGUE ensures r10 has ECR)
+        bmsk.f 0, r10, 7
         bnz    trap_with_param
 
         ;============ TRAP  (no param): syscall top level
diff --git a/arch/arc/kernel/unwind.c b/arch/arc/kernel/unwind.c
index 182ce67dfe10..c2663fce7f6c 100644
--- a/arch/arc/kernel/unwind.c
+++ b/arch/arc/kernel/unwind.c
@@ -181,11 +181,6 @@ static void *__init unw_hdr_alloc_early(unsigned long sz)
         return memblock_alloc_from(sz, sizeof(unsigned int), MAX_DMA_ADDRESS);
 }
 
-static void *unw_hdr_alloc(unsigned long sz)
-{
-        return kmalloc(sz, GFP_KERNEL);
-}
-
 static void init_unwind_table(struct unwind_table *table, const char *name,
                               const void *core_start, unsigned long core_size,
                               const void *init_start, unsigned long init_size,
@@ -366,6 +361,10 @@ ret_err:
 }
 
 #ifdef CONFIG_MODULES
+static void *unw_hdr_alloc(unsigned long sz)
+{
+        return kmalloc(sz, GFP_KERNEL);
+}
 
 static struct unwind_table *last_table;
 
diff --git a/arch/arc/mm/fault.c b/arch/arc/mm/fault.c
index 81e84426fe21..3861543b66a0 100644
--- a/arch/arc/mm/fault.c
+++ b/arch/arc/mm/fault.c
@@ -63,24 +63,19 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
         struct vm_area_struct *vma = NULL;
         struct task_struct *tsk = current;
         struct mm_struct *mm = tsk->mm;
-        int si_code = SEGV_MAPERR;
-        int ret;
-        vm_fault_t fault;
-        int write = regs->ecr_cause & ECR_C_PROTV_STORE;  /* ST/EX */
-        unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
+        int sig, si_code = SEGV_MAPERR;
+        unsigned int write = 0, exec = 0, mask;
+        vm_fault_t fault = VM_FAULT_SIGSEGV;    /* handle_mm_fault() output */
+        unsigned int flags;                     /* handle_mm_fault() input */
 
         /*
-         * We fault-in kernel-space virtual memory on-demand. The
-         * 'reference' page table is init_mm.pgd.
-         *
          * NOTE! We MUST NOT take any locks for this case. We may
          * be in an interrupt or a critical region, and should
          * only copy the information from the master page table,
          * nothing more.
          */
         if (address >= VMALLOC_START && !user_mode(regs)) {
-                ret = handle_kernel_vaddr_fault(address);
-                if (unlikely(ret))
+                if (unlikely(handle_kernel_vaddr_fault(address)))
                         goto no_context;
                 else
                         return;
@@ -93,143 +88,117 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
         if (faulthandler_disabled() || !mm)
                 goto no_context;
 
+        if (regs->ecr_cause & ECR_C_PROTV_STORE)        /* ST/EX */
+                write = 1;
+        else if ((regs->ecr_vec == ECR_V_PROTV) &&
+                 (regs->ecr_cause == ECR_C_PROTV_INST_FETCH))
+                exec = 1;
+
+        flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
         if (user_mode(regs))
                 flags |= FAULT_FLAG_USER;
+        if (write)
+                flags |= FAULT_FLAG_WRITE;
+
 retry:
         down_read(&mm->mmap_sem);
+
         vma = find_vma(mm, address);
         if (!vma)
                 goto bad_area;
-        if (vma->vm_start <= address)
-                goto good_area;
-        if (!(vma->vm_flags & VM_GROWSDOWN))
-                goto bad_area;
-        if (expand_stack(vma, address))
-                goto bad_area;
+        if (unlikely(address < vma->vm_start)) {
+                if (!(vma->vm_flags & VM_GROWSDOWN) || expand_stack(vma, address))
+                        goto bad_area;
+        }
 
         /*
-         * Ok, we have a good vm_area for this memory access, so
-         * we can handle it..
+         * vm_area is good, now check permissions for this memory access
          */
-good_area:
-        si_code = SEGV_ACCERR;
-
-        /* Handle protection violation, execute on heap or stack */
-
-        if ((regs->ecr_vec == ECR_V_PROTV) &&
-            (regs->ecr_cause == ECR_C_PROTV_INST_FETCH))
+        mask = VM_READ;
+        if (write)
+                mask = VM_WRITE;
+        if (exec)
+                mask = VM_EXEC;
+
+        if (!(vma->vm_flags & mask)) {
+                si_code = SEGV_ACCERR;
                 goto bad_area;
-
-        if (write) {
-                if (!(vma->vm_flags & VM_WRITE))
-                        goto bad_area;
-                flags |= FAULT_FLAG_WRITE;
-        } else {
-                if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
-                        goto bad_area;
         }
 
-        /*
-         * If for any reason at all we couldn't handle the fault,
-         * make sure we exit gracefully rather than endlessly redo
-         * the fault.
-         */
         fault = handle_mm_fault(vma, address, flags);
 
-        if (fatal_signal_pending(current)) {
+        /*
+         * Fault retry nuances
+         */
+        if (unlikely(fault & VM_FAULT_RETRY)) {
 
                 /*
-                 * if fault retry, mmap_sem already relinquished by core mm
-                 * so OK to return to user mode (with signal handled first)
+                 * If fault needs to be retried, handle any pending signals
+                 * first (by returning to user mode).
+                 * mmap_sem already relinquished by core mm for RETRY case
                  */
-                if (fault & VM_FAULT_RETRY) {
+                if (fatal_signal_pending(current)) {
                         if (!user_mode(regs))
                                 goto no_context;
                         return;
                 }
-        }
-
-        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
-
-        if (likely(!(fault & VM_FAULT_ERROR))) {
+                /*
+                 * retry state machine
+                 */
                 if (flags & FAULT_FLAG_ALLOW_RETRY) {
-                        /* To avoid updating stats twice for retry case */
-                        if (fault & VM_FAULT_MAJOR) {
-                                tsk->maj_flt++;
-                                perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
-                                              regs, address);
-                        } else {
-                                tsk->min_flt++;
-                                perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
-                                              regs, address);
-                        }
-
-                        if (fault & VM_FAULT_RETRY) {
-                                flags &= ~FAULT_FLAG_ALLOW_RETRY;
-                                flags |= FAULT_FLAG_TRIED;
-                                goto retry;
-                        }
+                        flags &= ~FAULT_FLAG_ALLOW_RETRY;
+                        flags |= FAULT_FLAG_TRIED;
+                        goto retry;
                 }
-
-                /* Fault Handled Gracefully */
-                up_read(&mm->mmap_sem);
-                return;
         }
 
-        if (fault & VM_FAULT_OOM)
-                goto out_of_memory;
-        else if (fault & VM_FAULT_SIGSEGV)
-                goto bad_area;
-        else if (fault & VM_FAULT_SIGBUS)
-                goto do_sigbus;
-
-        /* no man's land */
-        BUG();
+bad_area:
+        up_read(&mm->mmap_sem);
 
         /*
-         * Something tried to access memory that isn't in our memory map..
-         * Fix it, but check if it's kernel or user first..
+         * Major/minor page fault accounting
+         * (in case of retry we only land here once)
          */
-bad_area:
-        up_read(&mm->mmap_sem);
+        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
-        /* User mode accesses just cause a SIGSEGV */
-        if (user_mode(regs)) {
-                tsk->thread.fault_address = address;
-                force_sig_fault(SIGSEGV, si_code, (void __user *)address);
-                return;
-        }
+        if (likely(!(fault & VM_FAULT_ERROR))) {
+                if (fault & VM_FAULT_MAJOR) {
+                        tsk->maj_flt++;
+                        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
+                                      regs, address);
+                } else {
+                        tsk->min_flt++;
+                        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
+                                      regs, address);
+                }
 
-no_context:
-        /* Are we prepared to handle this kernel fault?
-         *
-         * (The kernel has valid exception-points in the source
-         *  when it accesses user-memory. When it fails in one
-         *  of those points, we find it in a table and do a jump
-         *  to some fixup code that loads an appropriate error
-         *  code)
-         */
-        if (fixup_exception(regs))
+                /* Normal return path: fault Handled Gracefully */
                 return;
+        }
 
-        die("Oops", regs, address);
-
-out_of_memory:
-        up_read(&mm->mmap_sem);
+        if (!user_mode(regs))
+                goto no_context;
 
-        if (user_mode(regs)) {
+        if (fault & VM_FAULT_OOM) {
                 pagefault_out_of_memory();
                 return;
         }
 
-        goto no_context;
+        if (fault & VM_FAULT_SIGBUS) {
+                sig = SIGBUS;
+                si_code = BUS_ADRERR;
+        }
+        else {
+                sig = SIGSEGV;
+        }
 
-do_sigbus:
-        up_read(&mm->mmap_sem);
+        tsk->thread.fault_address = address;
+        force_sig_fault(sig, si_code, (void __user *)address);
+        return;
 
-        if (!user_mode(regs))
-                goto no_context;
+no_context:
+        if (fixup_exception(regs))
+                return;
 
-        tsk->thread.fault_address = address;
-        force_sig_fault(SIGBUS, BUS_ADRERR, (void __user *)address);
+        die("Oops", regs, address);
 }
diff --git a/arch/arc/mm/tlbex.S b/arch/arc/mm/tlbex.S
index 471a97bf492d..c55d95dd2f39 100644
--- a/arch/arc/mm/tlbex.S
+++ b/arch/arc/mm/tlbex.S
@@ -393,6 +393,17 @@ EV_TLBMissD_fast_ret:	; additional label for VDK OS-kit instrumentation
 ;-------- Common routine to call Linux Page Fault Handler -----------
 do_slow_path_pf:
 
+#ifdef CONFIG_ISA_ARCV2
+        ; Set Z flag if exception in U mode. Hardware micro-ops do this on any
+        ; taken interrupt/exception, and thus is already the case at the entry
+        ; above, but ensuing code would have already clobbered.
+        ; EXCEPTION_PROLOGUE called in slow path, relies on correct Z flag set
+
+        lr      r2, [erstatus]
+        and     r2, r2, STATUS_U_MASK
+        bxor.f  0, r2, STATUS_U_BIT
+#endif
+
         ; Restore the 4-scratch regs saved by fast path miss handler
         TLBMISS_RESTORE_REGS