powerpc/book3s64/pkeys: Optimize KUAP and KUEP feature disabled case

If FTR_BOOK3S_KUAP is disabled, kernel will continue to run with the same AMR
value with which it was entered. Hence there is a high chance that
we can return without restoring the AMR value. This also helps the case
when applications are not using the pkey feature. In this case, different
applications will have the same AMR values and hence we can avoid restoring
AMR in this case too.

Also avoid isync() if not really needed.

Do the same for IAMR.

null-syscall benchmark results:

With smap/smep disabled:
Without patch:
	957.95 ns    2778.17 cycles
With patch:
	858.38 ns    2489.30 cycles

With smap/smep enabled:
Without patch:
	1017.26 ns    2950.36 cycles
With patch:
	1021.51 ns    2962.44 cycles

Signed-off-by: Aneesh Kumar K.V <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]

arch/powerpc/include/asm/book3s/64/kup.h

@@ -12,28 +12,54 @@
 
 #ifdef __ASSEMBLY__
 
-.macro kuap_user_restore gpr1
+.macro kuap_user_restore gpr1, gpr2
 #if defined(CONFIG_PPC_PKEY)
 	BEGIN_MMU_FTR_SECTION_NESTED(67)
+	b	100f  // skip_restore_amr
+	END_MMU_FTR_SECTION_NESTED_IFCLR(MMU_FTR_PKEY, 67)
 	/*
 	 * AMR and IAMR are going to be different when
 	 * returning to userspace.
 	 */
 	ld	\gpr1, STACK_REGS_AMR(r1)
+
+	/*
+	 * If kuap feature is not enabled, do the mtspr
+	 * only if AMR value is different.
+	 */
+	BEGIN_MMU_FTR_SECTION_NESTED(68)
+	mfspr	\gpr2, SPRN_AMR
+	cmpd	\gpr1, \gpr2
+	beq	99f
+	END_MMU_FTR_SECTION_NESTED_IFCLR(MMU_FTR_BOOK3S_KUAP, 68)
+
 	isync
 	mtspr	SPRN_AMR, \gpr1
+99:
 	/*
 	 * Restore IAMR only when returning to userspace
 	 */
 	ld	\gpr1, STACK_REGS_IAMR(r1)
+
+	/*
+	 * If kuep feature is not enabled, do the mtspr
+	 * only if IAMR value is different.
+	 */
+	BEGIN_MMU_FTR_SECTION_NESTED(69)
+	mfspr	\gpr2, SPRN_IAMR
+	cmpd	\gpr1, \gpr2
+	beq	100f
+	END_MMU_FTR_SECTION_NESTED_IFCLR(MMU_FTR_BOOK3S_KUEP, 69)
+
+	isync
 	mtspr	SPRN_IAMR, \gpr1
 
+100: //skip_restore_amr
 	/* No isync required, see kuap_user_restore() */
-	END_MMU_FTR_SECTION_NESTED_IFSET(MMU_FTR_PKEY, 67)
 #endif
 .endm
 
-.macro kuap_kernel_restore	gpr1, gpr2
+.macro kuap_kernel_restore gpr1, gpr2
 #if defined(CONFIG_PPC_PKEY)
 
 	BEGIN_MMU_FTR_SECTION_NESTED(67)
@@ -199,18 +225,43 @@ static inline u64 current_thread_iamr(void)
 
 static inline void kuap_user_restore(struct pt_regs *regs)
 {
+	bool restore_amr = false, restore_iamr = false;
+	unsigned long amr, iamr;
+
 	if (!mmu_has_feature(MMU_FTR_PKEY))
 		return;
 
-	isync();
-	mtspr(SPRN_AMR, regs->amr);
-	mtspr(SPRN_IAMR, regs->iamr);
+	if (!mmu_has_feature(MMU_FTR_BOOK3S_KUAP)) {
+		amr = mfspr(SPRN_AMR);
+		if (amr != regs->amr)
+			restore_amr = true;
+	} else {
+		restore_amr = true;
+	}
+
+	if (!mmu_has_feature(MMU_FTR_BOOK3S_KUEP)) {
+		iamr = mfspr(SPRN_IAMR);
+		if (iamr != regs->iamr)
+			restore_iamr = true;
+	} else {
+		restore_iamr = true;
+	}
+
+
+	if (restore_amr || restore_iamr) {
+		isync();
+		if (restore_amr)
+			mtspr(SPRN_AMR, regs->amr);
+		if (restore_iamr)
+			mtspr(SPRN_IAMR, regs->iamr);
+	}
 	/*
 	 * No isync required here because we are about to rfi
 	 * back to previous context before any user accesses
 	 * would be made, which is a CSI.
 	 */
 }
+
 static inline void kuap_kernel_restore(struct pt_regs *regs,
 					   unsigned long amr)
 {

arch/powerpc/kernel/entry_64.S

@@ -675,7 +675,7 @@ _ASM_NOKPROBE_SYMBOL(interrupt_return)
 	bne-	.Lrestore_nvgprs
 
 .Lfast_user_interrupt_return_amr:
-	kuap_user_restore r3
+	kuap_user_restore r3, r4
 .Lfast_user_interrupt_return:
 	ld	r11,_NIP(r1)
 	ld	r12,_MSR(r1)

arch/powerpc/kernel/syscall_64.c

@@ -38,6 +38,7 @@ notrace long system_call_exception(long r3, long r4, long r5,
 #ifdef CONFIG_PPC_PKEY
 	if (mmu_has_feature(MMU_FTR_PKEY)) {
 		unsigned long amr, iamr;
+		bool flush_needed = false;
 		/*
 		 * When entering from userspace we mostly have the AMR/IAMR
 		 * different from kernel default values. Hence don't compare.
@@ -46,11 +47,16 @@ notrace long system_call_exception(long r3, long r4, long r5,
 		iamr = mfspr(SPRN_IAMR);
 		regs->amr  = amr;
 		regs->iamr = iamr;
-		if (mmu_has_feature(MMU_FTR_BOOK3S_KUAP))
+		if (mmu_has_feature(MMU_FTR_BOOK3S_KUAP)) {
 			mtspr(SPRN_AMR, AMR_KUAP_BLOCKED);
-		if (mmu_has_feature(MMU_FTR_BOOK3S_KUEP))
+			flush_needed = true;
+		}
+		if (mmu_has_feature(MMU_FTR_BOOK3S_KUEP)) {
 			mtspr(SPRN_IAMR, AMR_KUEP_BLOCKED);
-		isync();
+			flush_needed = true;
+		}
+		if (flush_needed)
+			isync();
 	} else
 #endif
 		kuap_check_amr();