arm64: alternatives: have callbacks take a cap

Today, callback alternatives are special-cased within
__apply_alternatives(), and are applied alongside patching for system
capabilities as ARM64_NCAPS is not part of the boot_capabilities feature
mask.

This special-casing is less than ideal. Giving special meaning to
ARM64_NCAPS for this requires some structures and loops to use
ARM64_NCAPS + 1 (AKA ARM64_NPATCHABLE), while others use ARM64_NCAPS.
It's also not immediately clear callback alternatives are only applied
when applying alternatives for system-wide features.

To make this a bit clearer, changes the way that callback alternatives
are identified to remove the special-casing of ARM64_NCAPS, and to allow
callback alternatives to be associated with a cpucap as with all other
alternatives.

New cpucaps, ARM64_ALWAYS_BOOT and ARM64_ALWAYS_SYSTEM are added which
are always detected alongside boot cpu capabilities and system
capabilities respectively. All existing callback alternatives are made
to use ARM64_ALWAYS_SYSTEM, and so will be patched at the same point
during the boot flow as before.

Subsequent patches will make more use of these new cpucaps.

There should be no functional change as a result of this patch.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Joey Gouly <joey.gouly@arm.com>
Cc: Marc Zyngier <maz@kernel.org>
Cc: Will Deacon <will@kernel.org>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220912162210.3626215-7-mark.rutland@arm.comSigned-off-by: Catalin Marinas <catalin.marinas@arm.com>

arch/arm64/include/asm/alternative-macros.h

@@ -2,10 +2,16 @@
 #ifndef __ASM_ALTERNATIVE_MACROS_H
 #define __ASM_ALTERNATIVE_MACROS_H
 
+#include <linux/const.h>
+
 #include <asm/cpucaps.h>
 #include <asm/insn-def.h>
 
-#define ARM64_CB_PATCH ARM64_NCAPS
+#define ARM64_CB_BIT	(UL(1) << 15)
+
+#if ARM64_NCAPS >= ARM64_CB_BIT
+#error "cpucaps have overflown ARM64_CB_BIT"
+#endif
 
 #ifndef __ASSEMBLY__
 
@@ -73,8 +79,8 @@
 #define _ALTERNATIVE_CFG(oldinstr, newinstr, feature, cfg, ...)	\
 	__ALTERNATIVE_CFG(oldinstr, newinstr, feature, IS_ENABLED(cfg))
 
-#define ALTERNATIVE_CB(oldinstr, cb) \
-	__ALTERNATIVE_CFG_CB(oldinstr, ARM64_CB_PATCH, 1, cb)
+#define ALTERNATIVE_CB(oldinstr, feature, cb) \
+	__ALTERNATIVE_CFG_CB(oldinstr, ARM64_CB_BIT | (feature), 1, cb)
 #else
 
 #include <asm/assembler.h>
@@ -82,7 +88,7 @@
 .macro altinstruction_entry orig_offset alt_offset feature orig_len alt_len
 	.word \orig_offset - .
 	.word \alt_offset - .
-	.hword \feature
+	.hword (\feature)
 	.byte \orig_len
 	.byte \alt_len
 .endm
@@ -141,10 +147,10 @@
 661:
 .endm
 
-.macro alternative_cb cb
+.macro alternative_cb cap, cb
 	.set .Lasm_alt_mode, 0
 	.pushsection .altinstructions, "a"
-	altinstruction_entry 661f, \cb, ARM64_CB_PATCH, 662f-661f, 0
+	altinstruction_entry 661f, \cb, ARM64_CB_BIT | \cap, 662f-661f, 0
 	.popsection
 661:
 .endm

arch/arm64/include/asm/assembler.h

@@ -293,7 +293,7 @@ alternative_endif
 alternative_if_not ARM64_KVM_PROTECTED_MODE
 	ASM_BUG()
 alternative_else_nop_endif
-alternative_cb kvm_compute_final_ctr_el0
+alternative_cb ARM64_ALWAYS_SYSTEM, kvm_compute_final_ctr_el0
 	movz	\reg, #0
 	movk	\reg, #0, lsl #16
 	movk	\reg, #0, lsl #32
@@ -877,7 +877,7 @@ alternative_endif
 
 	.macro __mitigate_spectre_bhb_loop      tmp
 #ifdef CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY
-alternative_cb  spectre_bhb_patch_loop_iter
+alternative_cb ARM64_ALWAYS_SYSTEM, spectre_bhb_patch_loop_iter
 	mov	\tmp, #32		// Patched to correct the immediate
 alternative_cb_end
 .Lspectre_bhb_loop\@:
@@ -890,7 +890,7 @@ alternative_cb_end
 
 	.macro mitigate_spectre_bhb_loop	tmp
 #ifdef CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY
-alternative_cb	spectre_bhb_patch_loop_mitigation_enable
+alternative_cb ARM64_ALWAYS_SYSTEM, spectre_bhb_patch_loop_mitigation_enable
 	b	.L_spectre_bhb_loop_done\@	// Patched to NOP
 alternative_cb_end
 	__mitigate_spectre_bhb_loop	\tmp
@@ -904,7 +904,7 @@ alternative_cb_end
 	stp	x0, x1, [sp, #-16]!
 	stp	x2, x3, [sp, #-16]!
 	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_3
-alternative_cb	smccc_patch_fw_mitigation_conduit
+alternative_cb ARM64_ALWAYS_SYSTEM, smccc_patch_fw_mitigation_conduit
 	nop					// Patched to SMC/HVC #0
 alternative_cb_end
 	ldp	x2, x3, [sp], #16
@@ -914,7 +914,7 @@ alternative_cb_end
 
 	.macro mitigate_spectre_bhb_clear_insn
 #ifdef CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY
-alternative_cb	spectre_bhb_patch_clearbhb
+alternative_cb ARM64_ALWAYS_SYSTEM, spectre_bhb_patch_clearbhb
 	/* Patched to NOP when not supported */
 	clearbhb
 	isb

arch/arm64/include/asm/cpufeature.h

@@ -422,9 +422,7 @@ extern DECLARE_BITMAP(cpu_hwcaps, ARM64_NCAPS);
 extern struct static_key_false cpu_hwcap_keys[ARM64_NCAPS];
 extern struct static_key_false arm64_const_caps_ready;
 
-/* ARM64 CAPS + alternative_cb */
-#define ARM64_NPATCHABLE (ARM64_NCAPS + 1)
-extern DECLARE_BITMAP(boot_capabilities, ARM64_NPATCHABLE);
+extern DECLARE_BITMAP(boot_capabilities, ARM64_NCAPS);
 
 #define for_each_available_cap(cap)		\
 	for_each_set_bit(cap, cpu_hwcaps, ARM64_NCAPS)

arch/arm64/include/asm/kvm_mmu.h

@@ -63,7 +63,7 @@
  * specific registers encoded in the instructions).
  */
 .macro kern_hyp_va	reg
-alternative_cb kvm_update_va_mask
+alternative_cb ARM64_ALWAYS_SYSTEM, kvm_update_va_mask
 	and     \reg, \reg, #1		/* mask with va_mask */
 	ror	\reg, \reg, #1		/* rotate to the first tag bit */
 	add	\reg, \reg, #0		/* insert the low 12 bits of the tag */
@@ -97,7 +97,7 @@ alternative_cb_end
 	hyp_pa	\reg, \tmp
 
 	/* Load kimage_voffset. */
-alternative_cb kvm_get_kimage_voffset
+alternative_cb ARM64_ALWAYS_SYSTEM, kvm_get_kimage_voffset
 	movz	\tmp, #0
 	movk	\tmp, #0, lsl #16
 	movk	\tmp, #0, lsl #32
@@ -131,6 +131,7 @@ static __always_inline unsigned long __kern_hyp_va(unsigned long v)
 				    "add %0, %0, #0\n"
 				    "add %0, %0, #0, lsl 12\n"
 				    "ror %0, %0, #63\n",
+				    ARM64_ALWAYS_SYSTEM,
 				    kvm_update_va_mask)
 		     : "+r" (v));
 	return v;

arch/arm64/kernel/alternative.c

@@ -21,6 +21,9 @@
 #define ALT_ORIG_PTR(a)		__ALT_PTR(a, orig_offset)
 #define ALT_REPL_PTR(a)		__ALT_PTR(a, alt_offset)
 
+#define ALT_CAP(a)		((a)->cpufeature & ~ARM64_CB_BIT)
+#define ALT_HAS_CB(a)		((a)->cpufeature & ARM64_CB_BIT)
+
 /* Volatile, as we may be patching the guts of READ_ONCE() */
 static volatile int all_alternatives_applied;
 
@@ -143,16 +146,15 @@ static void __nocfi __apply_alternatives(const struct alt_region *region,
 
 	for (alt = region->begin; alt < region->end; alt++) {
 		int nr_inst;
+		int cap = ALT_CAP(alt);
 
-		if (!test_bit(alt->cpufeature, feature_mask))
+		if (!test_bit(cap, feature_mask))
 			continue;
 
-		/* Use ARM64_CB_PATCH as an unconditional patch */
-		if (alt->cpufeature < ARM64_CB_PATCH &&
-		    !cpus_have_cap(alt->cpufeature))
+		if (!cpus_have_cap(cap))
 			continue;
 
-		if (alt->cpufeature == ARM64_CB_PATCH)
+		if (ALT_HAS_CB(alt))
 			BUG_ON(alt->alt_len != 0);
 		else
 			BUG_ON(alt->alt_len != alt->orig_len);
@@ -161,10 +163,10 @@ static void __nocfi __apply_alternatives(const struct alt_region *region,
 		updptr = is_module ? origptr : lm_alias(origptr);
 		nr_inst = alt->orig_len / AARCH64_INSN_SIZE;
 
-		if (alt->cpufeature < ARM64_CB_PATCH)
-			alt_cb = patch_alternative;
-		else
+		if (ALT_HAS_CB(alt))
 			alt_cb  = ALT_REPL_PTR(alt);
+		else
+			alt_cb = patch_alternative;
 
 		alt_cb(alt, origptr, updptr, nr_inst);
 
@@ -208,10 +210,10 @@ static int __apply_alternatives_multi_stop(void *unused)
 			cpu_relax();
 		isb();
 	} else {
-		DECLARE_BITMAP(remaining_capabilities, ARM64_NPATCHABLE);
+		DECLARE_BITMAP(remaining_capabilities, ARM64_NCAPS);
 
 		bitmap_complement(remaining_capabilities, boot_capabilities,
-				  ARM64_NPATCHABLE);
+				  ARM64_NCAPS);
 
 		BUG_ON(all_alternatives_applied);
 		__apply_alternatives(&kernel_alternatives, false,
@@ -254,9 +256,9 @@ void apply_alternatives_module(void *start, size_t length)
 		.begin	= start,
 		.end	= start + length,
 	};
-	DECLARE_BITMAP(all_capabilities, ARM64_NPATCHABLE);
+	DECLARE_BITMAP(all_capabilities, ARM64_NCAPS);
 
-	bitmap_fill(all_capabilities, ARM64_NPATCHABLE);
+	bitmap_fill(all_capabilities, ARM64_NCAPS);
 
 	__apply_alternatives(&region, true, &all_capabilities[0]);
 }

arch/arm64/kernel/cpufeature.c

@@ -108,8 +108,7 @@ DECLARE_BITMAP(cpu_hwcaps, ARM64_NCAPS);
 EXPORT_SYMBOL(cpu_hwcaps);
 static struct arm64_cpu_capabilities const __ro_after_init *cpu_hwcaps_ptrs[ARM64_NCAPS];
 
-/* Need also bit for ARM64_CB_PATCH */
-DECLARE_BITMAP(boot_capabilities, ARM64_NPATCHABLE);
+DECLARE_BITMAP(boot_capabilities, ARM64_NCAPS);
 
 bool arm64_use_ng_mappings = false;
 EXPORT_SYMBOL(arm64_use_ng_mappings);
@@ -1391,6 +1390,12 @@ u64 __read_sysreg_by_encoding(u32 sys_id)
 
 #include <linux/irqchip/arm-gic-v3.h>
 
+static bool
+has_always(const struct arm64_cpu_capabilities *entry, int scope)
+{
+	return true;
+}
+
 static bool
 feature_matches(u64 reg, const struct arm64_cpu_capabilities *entry)
 {
@@ -2087,6 +2092,16 @@ cpucap_panic_on_conflict(const struct arm64_cpu_capabilities *cap)
 }
 
 static const struct arm64_cpu_capabilities arm64_features[] = {
+	{
+		.capability = ARM64_ALWAYS_BOOT,
+		.type = ARM64_CPUCAP_BOOT_CPU_FEATURE,
+		.matches = has_always,
+	},
+	{
+		.capability = ARM64_ALWAYS_SYSTEM,
+		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
+		.matches = has_always,
+	},
 	{
 		.desc = "GIC system register CPU interface",
 		.capability = ARM64_HAS_SYSREG_GIC_CPUIF,

arch/arm64/kernel/entry.S

@@ -114,7 +114,7 @@
 	 * them if required.
 	 */
 	.macro	apply_ssbd, state, tmp1, tmp2
-alternative_cb	spectre_v4_patch_fw_mitigation_enable
+alternative_cb	ARM64_ALWAYS_SYSTEM, spectre_v4_patch_fw_mitigation_enable
 	b	.L__asm_ssbd_skip\@		// Patched to NOP
 alternative_cb_end
 	ldr_this_cpu	\tmp2, arm64_ssbd_callback_required, \tmp1
@@ -123,7 +123,7 @@ alternative_cb_end
 	tbnz	\tmp2, #TIF_SSBD, .L__asm_ssbd_skip\@
 	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_2
 	mov	w1, #\state
-alternative_cb	smccc_patch_fw_mitigation_conduit
+alternative_cb	ARM64_ALWAYS_SYSTEM, smccc_patch_fw_mitigation_conduit
 	nop					// Patched to SMC/HVC #0
 alternative_cb_end
 .L__asm_ssbd_skip\@:
@@ -175,7 +175,7 @@ alternative_else_nop_endif
 
 	.macro mte_set_kernel_gcr, tmp, tmp2
 #ifdef CONFIG_KASAN_HW_TAGS
-alternative_cb	kasan_hw_tags_enable
+alternative_cb	ARM64_ALWAYS_SYSTEM, kasan_hw_tags_enable
 	b	1f
 alternative_cb_end
 	mov	\tmp, KERNEL_GCR_EL1
@@ -186,7 +186,7 @@ alternative_cb_end
 
 	.macro mte_set_user_gcr, tsk, tmp, tmp2
 #ifdef CONFIG_KASAN_HW_TAGS
-alternative_cb	kasan_hw_tags_enable
+alternative_cb	ARM64_ALWAYS_SYSTEM, kasan_hw_tags_enable
 	b	1f
 alternative_cb_end
 	ldr	\tmp, [\tsk, #THREAD_MTE_CTRL]

arch/arm64/kvm/hyp/hyp-entry.S

@@ -196,7 +196,7 @@ SYM_CODE_END(__kvm_hyp_vector)
 	sub	sp, sp, #(8 * 4)
 	stp	x2, x3, [sp, #(8 * 0)]
 	stp	x0, x1, [sp, #(8 * 2)]
-	alternative_cb spectre_bhb_patch_wa3
+	alternative_cb ARM64_ALWAYS_SYSTEM, spectre_bhb_patch_wa3
 	/* Patched to mov WA3 when supported */
 	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_1
 	alternative_cb_end
@@ -216,7 +216,7 @@ SYM_CODE_END(__kvm_hyp_vector)
 	mitigate_spectre_bhb_clear_insn
 	.endif
 	.if \indirect != 0
-	alternative_cb  kvm_patch_vector_branch
+	alternative_cb ARM64_ALWAYS_SYSTEM, kvm_patch_vector_branch
 	/*
 	 * For ARM64_SPECTRE_V3A configurations, these NOPs get replaced with:
 	 *

arch/arm64/tools/cpucaps

@@ -2,6 +2,8 @@
 #
 # Internal CPU capabilities constants, keep this list sorted
 
+ALWAYS_BOOT
+ALWAYS_SYSTEM
 BTI
 # Unreliable: use system_supports_32bit_el0() instead.
 HAS_32BIT_EL0_DO_NOT_USE