arm64: entry: Add fake CPU feature for unmapping the kernel at EL0

Allow explicit disabling of the entry trampoline on the kernel command
line (kpti=off) by adding a fake CPU feature (ARM64_UNMAP_KERNEL_AT_EL0)
that can be used to toggle the alternative sequences in our entry code and
avoid use of the trampoline altogether if desired. This also allows us to
make use of a static key in arm64_kernel_unmapped_at_el0().
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>

arch/arm64/include/asm/cpucaps.h

@@ -41,7 +41,8 @@
 #define ARM64_WORKAROUND_CAVIUM_30115		20
 #define ARM64_HAS_DCPOP				21
 #define ARM64_SVE				22
+#define ARM64_UNMAP_KERNEL_AT_EL0		23
 
-#define ARM64_NCAPS				23
+#define ARM64_NCAPS				24
 
 #endif /* __ASM_CPUCAPS_H */

arch/arm64/include/asm/mmu.h

@@ -36,7 +36,8 @@ typedef struct {
 
 static inline bool arm64_kernel_unmapped_at_el0(void)
 {
-	return IS_ENABLED(CONFIG_UNMAP_KERNEL_AT_EL0);
+	return IS_ENABLED(CONFIG_UNMAP_KERNEL_AT_EL0) &&
+	       cpus_have_const_cap(ARM64_UNMAP_KERNEL_AT_EL0);
 }
 
 extern void paging_init(void);

arch/arm64/kernel/cpufeature.c

@@ -845,6 +845,40 @@ static bool has_no_fpsimd(const struct arm64_cpu_capabilities *entry, int __unus
 					ID_AA64PFR0_FP_SHIFT) < 0;
 }
 
+#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
+static int __kpti_forced; /* 0: not forced, >0: forced on, <0: forced off */
+
+static bool unmap_kernel_at_el0(const struct arm64_cpu_capabilities *entry,
+				int __unused)
+{
+	/* Forced on command line? */
+	if (__kpti_forced) {
+		pr_info_once("kernel page table isolation forced %s by command line option\n",
+			     __kpti_forced > 0 ? "ON" : "OFF");
+		return __kpti_forced > 0;
+	}
+
+	/* Useful for KASLR robustness */
+	if (IS_ENABLED(CONFIG_RANDOMIZE_BASE))
+		return true;
+
+	return false;
+}
+
+static int __init parse_kpti(char *str)
+{
+	bool enabled;
+	int ret = strtobool(str, &enabled);
+
+	if (ret)
+		return ret;
+
+	__kpti_forced = enabled ? 1 : -1;
+	return 0;
+}
+__setup("kpti=", parse_kpti);
+#endif	/* CONFIG_UNMAP_KERNEL_AT_EL0 */
+
 static const struct arm64_cpu_capabilities arm64_features[] = {
 	{
 		.desc = "GIC system register CPU interface",
@@ -931,6 +965,13 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
 		.def_scope = SCOPE_SYSTEM,
 		.matches = hyp_offset_low,
 	},
+#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
+	{
+		.capability = ARM64_UNMAP_KERNEL_AT_EL0,
+		.def_scope = SCOPE_SYSTEM,
+		.matches = unmap_kernel_at_el0,
+	},
+#endif
 	{
 		/* FP/SIMD is not implemented */
 		.capability = ARM64_HAS_NO_FPSIMD,

arch/arm64/kernel/entry.S

@@ -74,6 +74,7 @@
 	.macro kernel_ventry, el, label, regsize = 64
 	.align 7
 #ifdef CONFIG_UNMAP_KERNEL_AT_EL0
+alternative_if ARM64_UNMAP_KERNEL_AT_EL0
 	.if	\el == 0
 	.if	\regsize == 64
 	mrs	x30, tpidrro_el0
@@ -82,6 +83,7 @@
 	mov	x30, xzr
 	.endif
 	.endif
+alternative_else_nop_endif
 #endif
 
 	sub	sp, sp, #S_FRAME_SIZE
@@ -323,10 +325,9 @@ alternative_else_nop_endif
 	ldr	lr, [sp, #S_LR]
 	add	sp, sp, #S_FRAME_SIZE		// restore sp
 
-#ifndef CONFIG_UNMAP_KERNEL_AT_EL0
-	eret
-#else
 	.if	\el == 0
+alternative_insn eret, nop, ARM64_UNMAP_KERNEL_AT_EL0
+#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
 	bne	4f
 	msr	far_el1, x30
 	tramp_alias	x30, tramp_exit_native
@@ -334,10 +335,10 @@ alternative_else_nop_endif
 4:
 	tramp_alias	x30, tramp_exit_compat
 	br	x30
+#endif
 	.else
 	eret
 	.endif
-#endif
 	.endm
 
 	.macro	irq_stack_entry