Merge branch 'linus' into perf/core, to pick up perf/urgent fixes

Pick up perf/urgent fixes that are upstream already, but not
yet in the perf/core development branch.

Signed-off-by: Ingo Molnar <mingo@kernel.org>

15 files changed, 97 insertions, 87 deletions

diff --git a/arch/arm64/boot/dts/freescale/imx8-ss-conn.dtsi b/arch/arm64/boot/dts/freescale/imx8-ss-conn.dtsi
index 3c42240e78e2..4aaf5a0c1ed8 100644
--- a/arch/arm64/boot/dts/freescale/imx8-ss-conn.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8-ss-conn.dtsi
@@ -41,7 +41,7 @@ conn_subsys: bus@5b000000 {
 		interrupts = <GIC_SPI 267 IRQ_TYPE_LEVEL_HIGH>;
 		fsl,usbphy = <&usbphy1>;
 		fsl,usbmisc = <&usbmisc1 0>;
-		clocks = <&usb2_lpcg 0>;
+		clocks = <&usb2_lpcg IMX_LPCG_CLK_6>;
 		ahb-burst-config = <0x0>;
 		tx-burst-size-dword = <0x10>;
 		rx-burst-size-dword = <0x10>;
@@ -58,7 +58,7 @@ conn_subsys: bus@5b000000 {
 	usbphy1: usbphy@5b100000 {
 		compatible = "fsl,imx7ulp-usbphy";
 		reg = <0x5b100000 0x1000>;
-		clocks = <&usb2_lpcg 1>;
+		clocks = <&usb2_lpcg IMX_LPCG_CLK_7>;
 		power-domains = <&pd IMX_SC_R_USB_0_PHY>;
 		status = "disabled";
 	};
@@ -67,8 +67,8 @@ conn_subsys: bus@5b000000 {
 		interrupts = <GIC_SPI 232 IRQ_TYPE_LEVEL_HIGH>;
 		reg = <0x5b010000 0x10000>;
 		clocks = <&sdhc0_lpcg IMX_LPCG_CLK_4>,
-			 <&sdhc0_lpcg IMX_LPCG_CLK_0>,
-			 <&sdhc0_lpcg IMX_LPCG_CLK_5>;
+			 <&sdhc0_lpcg IMX_LPCG_CLK_5>,
+			 <&sdhc0_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "ahb", "per";
 		power-domains = <&pd IMX_SC_R_SDHC_0>;
 		status = "disabled";
@@ -78,8 +78,8 @@ conn_subsys: bus@5b000000 {
 		interrupts = <GIC_SPI 233 IRQ_TYPE_LEVEL_HIGH>;
 		reg = <0x5b020000 0x10000>;
 		clocks = <&sdhc1_lpcg IMX_LPCG_CLK_4>,
-			 <&sdhc1_lpcg IMX_LPCG_CLK_0>,
-			 <&sdhc1_lpcg IMX_LPCG_CLK_5>;
+			 <&sdhc1_lpcg IMX_LPCG_CLK_5>,
+			 <&sdhc1_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "ahb", "per";
 		power-domains = <&pd IMX_SC_R_SDHC_1>;
 		fsl,tuning-start-tap = <20>;
@@ -91,8 +91,8 @@ conn_subsys: bus@5b000000 {
 		interrupts = <GIC_SPI 234 IRQ_TYPE_LEVEL_HIGH>;
 		reg = <0x5b030000 0x10000>;
 		clocks = <&sdhc2_lpcg IMX_LPCG_CLK_4>,
-			 <&sdhc2_lpcg IMX_LPCG_CLK_0>,
-			 <&sdhc2_lpcg IMX_LPCG_CLK_5>;
+			 <&sdhc2_lpcg IMX_LPCG_CLK_5>,
+			 <&sdhc2_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "ahb", "per";
 		power-domains = <&pd IMX_SC_R_SDHC_2>;
 		status = "disabled";
diff --git a/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi b/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
index cab3468b1875..f7a91d43a0ff 100644
--- a/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
@@ -28,8 +28,8 @@ dma_subsys: bus@5a000000 {
 		#size-cells = <0>;
 		interrupts = <GIC_SPI 336 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&spi0_lpcg 0>,
-			 <&spi0_lpcg 1>;
+		clocks = <&spi0_lpcg IMX_LPCG_CLK_0>,
+			 <&spi0_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_SPI_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <60000000>;
@@ -44,8 +44,8 @@ dma_subsys: bus@5a000000 {
 		#size-cells = <0>;
 		interrupts = <GIC_SPI 337 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&spi1_lpcg 0>,
-			 <&spi1_lpcg 1>;
+		clocks = <&spi1_lpcg IMX_LPCG_CLK_0>,
+			 <&spi1_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_SPI_1 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <60000000>;
@@ -60,8 +60,8 @@ dma_subsys: bus@5a000000 {
 		#size-cells = <0>;
 		interrupts = <GIC_SPI 338 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&spi2_lpcg 0>,
-			 <&spi2_lpcg 1>;
+		clocks = <&spi2_lpcg IMX_LPCG_CLK_0>,
+			 <&spi2_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_SPI_2 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <60000000>;
@@ -76,8 +76,8 @@ dma_subsys: bus@5a000000 {
 		#size-cells = <0>;
 		interrupts = <GIC_SPI 339 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&spi3_lpcg 0>,
-			 <&spi3_lpcg 1>;
+		clocks = <&spi3_lpcg IMX_LPCG_CLK_0>,
+			 <&spi3_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_SPI_3 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <60000000>;
@@ -145,8 +145,8 @@ dma_subsys: bus@5a000000 {
 		compatible = "fsl,imx8qxp-pwm", "fsl,imx27-pwm";
 		reg = <0x5a190000 0x1000>;
 		interrupts = <GIC_SPI 127 IRQ_TYPE_LEVEL_HIGH>;
-		clocks = <&adma_pwm_lpcg 1>,
-			 <&adma_pwm_lpcg 0>;
+		clocks = <&adma_pwm_lpcg IMX_LPCG_CLK_4>,
+			 <&adma_pwm_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "per";
 		assigned-clocks = <&clk IMX_SC_R_LCD_0_PWM_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
@@ -355,8 +355,8 @@ dma_subsys: bus@5a000000 {
 		reg = <0x5a880000 0x10000>;
 		interrupts = <GIC_SPI 240 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&adc0_lpcg 0>,
-			 <&adc0_lpcg 1>;
+		clocks = <&adc0_lpcg IMX_LPCG_CLK_0>,
+			 <&adc0_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_ADC_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
@@ -370,8 +370,8 @@ dma_subsys: bus@5a000000 {
 		reg = <0x5a890000 0x10000>;
 		interrupts = <GIC_SPI 241 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&adc1_lpcg 0>,
-			 <&adc1_lpcg 1>;
+		clocks = <&adc1_lpcg IMX_LPCG_CLK_0>,
+			 <&adc1_lpcg IMX_LPCG_CLK_4>;
 		clock-names = "per", "ipg";
 		assigned-clocks = <&clk IMX_SC_R_ADC_1 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
@@ -384,8 +384,8 @@ dma_subsys: bus@5a000000 {
 		reg = <0x5a8d0000 0x10000>;
 		interrupts = <GIC_SPI 235 IRQ_TYPE_LEVEL_HIGH>;
 		interrupt-parent = <&gic>;
-		clocks = <&can0_lpcg 1>,
-			 <&can0_lpcg 0>;
+		clocks = <&can0_lpcg IMX_LPCG_CLK_4>,
+			 <&can0_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "per";
 		assigned-clocks = <&clk IMX_SC_R_CAN_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <40000000>;
@@ -405,8 +405,8 @@ dma_subsys: bus@5a000000 {
 		 * CAN1 shares CAN0's clock and to enable CAN0's clock it
 		 * has to be powered on.
 		 */
-		clocks = <&can0_lpcg 1>,
-			 <&can0_lpcg 0>;
+		clocks = <&can0_lpcg IMX_LPCG_CLK_4>,
+			 <&can0_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "per";
 		assigned-clocks = <&clk IMX_SC_R_CAN_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <40000000>;
@@ -426,8 +426,8 @@ dma_subsys: bus@5a000000 {
 		 * CAN2 shares CAN0's clock and to enable CAN0's clock it
 		 * has to be powered on.
 		 */
-		clocks = <&can0_lpcg 1>,
-			 <&can0_lpcg 0>;
+		clocks = <&can0_lpcg IMX_LPCG_CLK_4>,
+			 <&can0_lpcg IMX_LPCG_CLK_0>;
 		clock-names = "ipg", "per";
 		assigned-clocks = <&clk IMX_SC_R_CAN_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <40000000>;
diff --git a/arch/arm64/boot/dts/freescale/imx8-ss-lsio.dtsi b/arch/arm64/boot/dts/freescale/imx8-ss-lsio.dtsi
index 7e510b21bbac..764c1a08e3b1 100644
--- a/arch/arm64/boot/dts/freescale/imx8-ss-lsio.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8-ss-lsio.dtsi
@@ -25,8 +25,8 @@ lsio_subsys: bus@5d000000 {
 		compatible = "fsl,imx27-pwm";
 		reg = <0x5d000000 0x10000>;
 		clock-names = "ipg", "per";
-		clocks = <&pwm0_lpcg 4>,
-			 <&pwm0_lpcg 1>;
+		clocks = <&pwm0_lpcg IMX_LPCG_CLK_6>,
+			 <&pwm0_lpcg IMX_LPCG_CLK_1>;
 		assigned-clocks = <&clk IMX_SC_R_PWM_0 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
 		#pwm-cells = <3>;
@@ -38,8 +38,8 @@ lsio_subsys: bus@5d000000 {
 		compatible = "fsl,imx27-pwm";
 		reg = <0x5d010000 0x10000>;
 		clock-names = "ipg", "per";
-		clocks = <&pwm1_lpcg 4>,
-			 <&pwm1_lpcg 1>;
+		clocks = <&pwm1_lpcg IMX_LPCG_CLK_6>,
+			 <&pwm1_lpcg IMX_LPCG_CLK_1>;
 		assigned-clocks = <&clk IMX_SC_R_PWM_1 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
 		#pwm-cells = <3>;
@@ -51,8 +51,8 @@ lsio_subsys: bus@5d000000 {
 		compatible = "fsl,imx27-pwm";
 		reg = <0x5d020000 0x10000>;
 		clock-names = "ipg", "per";
-		clocks = <&pwm2_lpcg 4>,
-			 <&pwm2_lpcg 1>;
+		clocks = <&pwm2_lpcg IMX_LPCG_CLK_6>,
+			 <&pwm2_lpcg IMX_LPCG_CLK_1>;
 		assigned-clocks = <&clk IMX_SC_R_PWM_2 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
 		#pwm-cells = <3>;
@@ -64,8 +64,8 @@ lsio_subsys: bus@5d000000 {
 		compatible = "fsl,imx27-pwm";
 		reg = <0x5d030000 0x10000>;
 		clock-names = "ipg", "per";
-		clocks = <&pwm3_lpcg 4>,
-			 <&pwm3_lpcg 1>;
+		clocks = <&pwm3_lpcg IMX_LPCG_CLK_6>,
+			 <&pwm3_lpcg IMX_LPCG_CLK_1>;
 		assigned-clocks = <&clk IMX_SC_R_PWM_3 IMX_SC_PM_CLK_PER>;
 		assigned-clock-rates = <24000000>;
 		#pwm-cells = <3>;
diff --git a/arch/arm64/boot/dts/freescale/imx8mp-venice-gw72xx.dtsi b/arch/arm64/boot/dts/freescale/imx8mp-venice-gw72xx.dtsi
index 41c79d2ebdd6..f24b14744799 100644
--- a/arch/arm64/boot/dts/freescale/imx8mp-venice-gw72xx.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8mp-venice-gw72xx.dtsi
@@ -14,6 +14,7 @@
 		pinctrl-0 = <&pinctrl_usbcon1>;
 		type = "micro";
 		label = "otg";
+		vbus-supply = <&reg_usb1_vbus>;
 		id-gpios = <&gpio3 21 GPIO_ACTIVE_HIGH>;
 
 		port {
@@ -183,7 +184,6 @@
 };
 
 &usb3_phy0 {
-	vbus-supply = <&reg_usb1_vbus>;
 	status = "okay";
 };
 
diff --git a/arch/arm64/boot/dts/freescale/imx8mp-venice-gw73xx.dtsi b/arch/arm64/boot/dts/freescale/imx8mp-venice-gw73xx.dtsi
index d5c400b355af..f5491a608b2f 100644
--- a/arch/arm64/boot/dts/freescale/imx8mp-venice-gw73xx.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8mp-venice-gw73xx.dtsi
@@ -14,6 +14,7 @@
 		pinctrl-0 = <&pinctrl_usbcon1>;
 		type = "micro";
 		label = "otg";
+		vbus-supply = <&reg_usb1_vbus>;
 		id-gpios = <&gpio3 21 GPIO_ACTIVE_HIGH>;
 
 		port {
@@ -202,7 +203,6 @@
 };
 
 &usb3_phy0 {
-	vbus-supply = <&reg_usb1_vbus>;
 	status = "okay";
 };
 
diff --git a/arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi b/arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi
index 11626fae5f97..aa9f28c4431d 100644
--- a/arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi
@@ -153,15 +153,15 @@
 };
 
 &flexcan2 {
-	clocks = <&can1_lpcg 1>,
-		 <&can1_lpcg 0>;
+	clocks = <&can1_lpcg IMX_LPCG_CLK_4>,
+		 <&can1_lpcg IMX_LPCG_CLK_0>;
 	assigned-clocks = <&clk IMX_SC_R_CAN_1 IMX_SC_PM_CLK_PER>;
 	fsl,clk-source = /bits/ 8 <1>;
 };
 
 &flexcan3 {
-	clocks = <&can2_lpcg 1>,
-		 <&can2_lpcg 0>;
+	clocks = <&can2_lpcg IMX_LPCG_CLK_4>,
+		 <&can2_lpcg IMX_LPCG_CLK_0>;
 	assigned-clocks = <&clk IMX_SC_R_CAN_2 IMX_SC_PM_CLK_PER>;
 	fsl,clk-source = /bits/ 8 <1>;
 };
diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi
index f3a6da8b2890..5260c63db007 100644
--- a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi
+++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi
@@ -944,6 +944,8 @@ ap_spi_fp: &spi10 {
 		vddrf-supply = <&pp1300_l2c>;
 		vddch0-supply = <&pp3300_l10c>;
 		max-speed = <3200000>;
+
+		qcom,local-bd-address-broken;
 	};
 };
 
diff --git a/arch/arm64/include/asm/tlbflush.h b/arch/arm64/include/asm/tlbflush.h
index 3b0e8248e1a4..a75de2665d84 100644
--- a/arch/arm64/include/asm/tlbflush.h
+++ b/arch/arm64/include/asm/tlbflush.h
@@ -161,12 +161,18 @@ static inline unsigned long get_trans_granule(void)
 #define MAX_TLBI_RANGE_PAGES		__TLBI_RANGE_PAGES(31, 3)
 
 /*
- * Generate 'num' values from -1 to 30 with -1 rejected by the
- * __flush_tlb_range() loop below.
+ * Generate 'num' values from -1 to 31 with -1 rejected by the
+ * __flush_tlb_range() loop below. Its return value is only
+ * significant for a maximum of MAX_TLBI_RANGE_PAGES pages. If
+ * 'pages' is more than that, you must iterate over the overall
+ * range.
  */
-#define TLBI_RANGE_MASK			GENMASK_ULL(4, 0)
-#define __TLBI_RANGE_NUM(pages, scale)	\
-	((((pages) >> (5 * (scale) + 1)) & TLBI_RANGE_MASK) - 1)
+#define __TLBI_RANGE_NUM(pages, scale)					\
+	({								\
+		int __pages = min((pages),				\
+				  __TLBI_RANGE_PAGES(31, (scale)));	\
+		(__pages >> (5 * (scale) + 1)) - 1;			\
+	})
 
 /*
  *	TLB Invalidation
@@ -379,10 +385,6 @@ static inline void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch)
  * 3. If there is 1 page remaining, flush it through non-range operations. Range
  *    operations can only span an even number of pages. We save this for last to
  *    ensure 64KB start alignment is maintained for the LPA2 case.
- *
- * Note that certain ranges can be represented by either num = 31 and
- * scale or num = 0 and scale + 1. The loop below favours the latter
- * since num is limited to 30 by the __TLBI_RANGE_NUM() macro.
  */
 #define __flush_tlb_range_op(op, start, pages, stride,			\
 				asid, tlb_level, tlbi_user, lpa2)	\
diff --git a/arch/arm64/kernel/head.S b/arch/arm64/kernel/head.S
index ce08b744aaab..06234c3a15f3 100644
--- a/arch/arm64/kernel/head.S
+++ b/arch/arm64/kernel/head.S
@@ -291,6 +291,21 @@ SYM_INNER_LABEL(init_el2, SYM_L_LOCAL)
 	blr	x2
 0:
 	mov_q	x0, HCR_HOST_NVHE_FLAGS
+
+	/*
+	 * Compliant CPUs advertise their VHE-onlyness with
+	 * ID_AA64MMFR4_EL1.E2H0 < 0. HCR_EL2.E2H can be
+	 * RES1 in that case. Publish the E2H bit early so that
+	 * it can be picked up by the init_el2_state macro.
+	 *
+	 * Fruity CPUs seem to have HCR_EL2.E2H set to RAO/WI, but
+	 * don't advertise it (they predate this relaxation).
+	 */
+	mrs_s	x1, SYS_ID_AA64MMFR4_EL1
+	tbz	x1, #(ID_AA64MMFR4_EL1_E2H0_SHIFT + ID_AA64MMFR4_EL1_E2H0_WIDTH - 1), 1f
+
+	orr	x0, x0, #HCR_E2H
+1:
 	msr	hcr_el2, x0
 	isb
 
@@ -303,22 +318,10 @@ SYM_INNER_LABEL(init_el2, SYM_L_LOCAL)
 
 	mov_q	x1, INIT_SCTLR_EL1_MMU_OFF
 
-	/*
-	 * Compliant CPUs advertise their VHE-onlyness with
-	 * ID_AA64MMFR4_EL1.E2H0 < 0. HCR_EL2.E2H can be
-	 * RES1 in that case.
-	 *
-	 * Fruity CPUs seem to have HCR_EL2.E2H set to RES1, but
-	 * don't advertise it (they predate this relaxation).
-	 */
-	mrs_s	x0, SYS_ID_AA64MMFR4_EL1
-	ubfx	x0, x0, #ID_AA64MMFR4_EL1_E2H0_SHIFT, #ID_AA64MMFR4_EL1_E2H0_WIDTH
-	tbnz	x0, #(ID_AA64MMFR4_EL1_E2H0_SHIFT + ID_AA64MMFR4_EL1_E2H0_WIDTH - 1), 1f
-
 	mrs	x0, hcr_el2
 	and	x0, x0, #HCR_E2H
 	cbz	x0, 2f
-1:
+
 	/* Set a sane SCTLR_EL1, the VHE way */
 	pre_disable_mmu_workaround
 	msr_s	SYS_SCTLR_EL12, x1
diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 162b030ab9da..0d022599eb61 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -761,7 +761,6 @@ static void sve_init_header_from_task(struct user_sve_header *header,
 {
 	unsigned int vq;
 	bool active;
-	bool fpsimd_only;
 	enum vec_type task_type;
 
 	memset(header, 0, sizeof(*header));
@@ -777,12 +776,10 @@ static void sve_init_header_from_task(struct user_sve_header *header,
 	case ARM64_VEC_SVE:
 		if (test_tsk_thread_flag(target, TIF_SVE_VL_INHERIT))
 			header->flags |= SVE_PT_VL_INHERIT;
-		fpsimd_only = !test_tsk_thread_flag(target, TIF_SVE);
 		break;
 	case ARM64_VEC_SME:
 		if (test_tsk_thread_flag(target, TIF_SME_VL_INHERIT))
 			header->flags |= SVE_PT_VL_INHERIT;
-		fpsimd_only = false;
 		break;
 	default:
 		WARN_ON_ONCE(1);
@@ -790,7 +787,7 @@ static void sve_init_header_from_task(struct user_sve_header *header,
 	}
 
 	if (active) {
-		if (fpsimd_only) {
+		if (target->thread.fp_type == FP_STATE_FPSIMD) {
 			header->flags |= SVE_PT_REGS_FPSIMD;
 		} else {
 			header->flags |= SVE_PT_REGS_SVE;
diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index 3dee5490eea9..c4a0a35e02c7 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
@@ -2597,14 +2597,11 @@ static __init int kvm_arm_init(void)
 	if (err)
 		goto out_hyp;
 
-	if (is_protected_kvm_enabled()) {
-		kvm_info("Protected nVHE mode initialized successfully\n");
-	} else if (in_hyp_mode) {
-		kvm_info("VHE mode initialized successfully\n");
-	} else {
-		char mode = cpus_have_final_cap(ARM64_KVM_HVHE) ? 'h' : 'n';
-		kvm_info("Hyp mode (%cVHE) initialized successfully\n", mode);
-	}
+	kvm_info("%s%sVHE mode initialized successfully\n",
+		 in_hyp_mode ? "" : (is_protected_kvm_enabled() ?
+				     "Protected " : "Hyp "),
+		 in_hyp_mode ? "" : (cpus_have_final_cap(ARM64_KVM_HVHE) ?
+				     "h" : "n"));
 
 	/*
 	 * FIXME: Do something reasonable if kvm_init() fails after pKVM
diff --git a/arch/arm64/kvm/hyp/nvhe/tlb.c b/arch/arm64/kvm/hyp/nvhe/tlb.c
index a60fb13e2192..2fc68da4036d 100644
--- a/arch/arm64/kvm/hyp/nvhe/tlb.c
+++ b/arch/arm64/kvm/hyp/nvhe/tlb.c
@@ -154,7 +154,8 @@ void __kvm_tlb_flush_vmid_range(struct kvm_s2_mmu *mmu,
 	/* Switch to requested VMID */
 	__tlb_switch_to_guest(mmu, &cxt, false);
 
-	__flush_s2_tlb_range_op(ipas2e1is, start, pages, stride, 0);
+	__flush_s2_tlb_range_op(ipas2e1is, start, pages, stride,
+				TLBI_TTL_UNKNOWN);
 
 	dsb(ish);
 	__tlbi(vmalle1is);
diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
index 3fae5830f8d2..5a59ef88b646 100644
--- a/arch/arm64/kvm/hyp/pgtable.c
+++ b/arch/arm64/kvm/hyp/pgtable.c
@@ -528,7 +528,7 @@ static int hyp_unmap_walker(const struct kvm_pgtable_visit_ctx *ctx,
 
 		kvm_clear_pte(ctx->ptep);
 		dsb(ishst);
-		__tlbi_level(vae2is, __TLBI_VADDR(ctx->addr, 0), ctx->level);
+		__tlbi_level(vae2is, __TLBI_VADDR(ctx->addr, 0), TLBI_TTL_UNKNOWN);
 	} else {
 		if (ctx->end - ctx->addr < granule)
 			return -EINVAL;
@@ -843,12 +843,15 @@ static bool stage2_try_break_pte(const struct kvm_pgtable_visit_ctx *ctx,
 		 * Perform the appropriate TLB invalidation based on the
 		 * evicted pte value (if any).
 		 */
-		if (kvm_pte_table(ctx->old, ctx->level))
-			kvm_tlb_flush_vmid_range(mmu, ctx->addr,
-						kvm_granule_size(ctx->level));
-		else if (kvm_pte_valid(ctx->old))
+		if (kvm_pte_table(ctx->old, ctx->level)) {
+			u64 size = kvm_granule_size(ctx->level);
+			u64 addr = ALIGN_DOWN(ctx->addr, size);
+
+			kvm_tlb_flush_vmid_range(mmu, addr, size);
+		} else if (kvm_pte_valid(ctx->old)) {
 			kvm_call_hyp(__kvm_tlb_flush_vmid_ipa, mmu,
 				     ctx->addr, ctx->level);
+		}
 	}
 
 	if (stage2_pte_is_counted(ctx->old))
@@ -896,9 +899,13 @@ static void stage2_unmap_put_pte(const struct kvm_pgtable_visit_ctx *ctx,
 	if (kvm_pte_valid(ctx->old)) {
 		kvm_clear_pte(ctx->ptep);
 
-		if (!stage2_unmap_defer_tlb_flush(pgt))
-			kvm_call_hyp(__kvm_tlb_flush_vmid_ipa, mmu,
-					ctx->addr, ctx->level);
+		if (kvm_pte_table(ctx->old, ctx->level)) {
+			kvm_call_hyp(__kvm_tlb_flush_vmid_ipa, mmu, ctx->addr,
+				     TLBI_TTL_UNKNOWN);
+		} else if (!stage2_unmap_defer_tlb_flush(pgt)) {
+			kvm_call_hyp(__kvm_tlb_flush_vmid_ipa, mmu, ctx->addr,
+				     ctx->level);
+		}
 	}
 
 	mm_ops->put_page(ctx->ptep);
diff --git a/arch/arm64/kvm/hyp/vhe/tlb.c b/arch/arm64/kvm/hyp/vhe/tlb.c
index b32e2940df7d..1a60b95381e8 100644
--- a/arch/arm64/kvm/hyp/vhe/tlb.c
+++ b/arch/arm64/kvm/hyp/vhe/tlb.c
@@ -171,7 +171,8 @@ void __kvm_tlb_flush_vmid_range(struct kvm_s2_mmu *mmu,
 	/* Switch to requested VMID */
 	__tlb_switch_to_guest(mmu, &cxt);
 
-	__flush_s2_tlb_range_op(ipas2e1is, start, pages, stride, 0);
+	__flush_s2_tlb_range_op(ipas2e1is, start, pages, stride,
+				TLBI_TTL_UNKNOWN);
 
 	dsb(ish);
 	__tlbi(vmalle1is);
diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c
index 18680771cdb0..dc04bc767865 100644
--- a/arch/arm64/kvm/mmu.c
+++ b/arch/arm64/kvm/mmu.c
@@ -1637,7 +1637,7 @@ int kvm_handle_guest_abort(struct kvm_vcpu *vcpu)
 	fault_ipa = kvm_vcpu_get_fault_ipa(vcpu);
 	is_iabt = kvm_vcpu_trap_is_iabt(vcpu);
 
-	if (esr_fsc_is_permission_fault(esr)) {
+	if (esr_fsc_is_translation_fault(esr)) {
 		/* Beyond sanitised PARange (which is the IPA limit) */
 		if (fault_ipa >= BIT_ULL(get_kvm_ipa_limit())) {
 			kvm_inject_size_fault(vcpu);