BACKPORT: ARM: 8991/1: use VFP assembler mnemonics if available

commit 2cbd1cc3dcd3e84be1fc1987da24b190ddf24a70 upstream.

The integrated assembler of Clang 10 and earlier do not allow to access
the VFP registers through the coprocessor load/store instructions:
arch/arm/vfp/vfpmodule.c:342:2: error: invalid operand for instruction
        fmxr(FPEXC, fpexc & ~(FPEXC_EX|FPEXC_DEX|FPEXC_FP2V|FPEXC_VV|FPEXC_TRAP_MASK));
        ^
arch/arm/vfp/vfpinstr.h:79:6: note: expanded from macro 'fmxr'
        asm("mcr p10, 7, %0, " vfpreg(_vfp_) ", cr0, 0 @ fmxr   " #_vfp_ ", %0"
            ^
<inline asm>:1:6: note: instantiated into assembly here
        mcr p10, 7, r0, cr8, cr0, 0 @ fmxr      FPEXC, r0
            ^

This has been addressed with Clang 11 [0]. However, to support earlier
versions of Clang and for better readability use of VFP assembler
mnemonics still is preferred.

Ideally we would replace this code with the unified assembler language
mnemonics vmrs/vmsr on call sites along with .fpu assembler directives.
The GNU assembler supports the .fpu directive at least since 2.17 (when
documentation has been added). Since Linux requires binutils 2.21 it is
safe to use .fpu directive. However, binutils does not allow to use
FPINST or FPINST2 as an argument to vmrs/vmsr instructions up to
binutils 2.24 (see binutils commit 16d02dc907c5):
arch/arm/vfp/vfphw.S: Assembler messages:
arch/arm/vfp/vfphw.S:162: Error: operand 0 must be FPSID or FPSCR pr FPEXC -- `vmsr FPINST,r6'
arch/arm/vfp/vfphw.S:165: Error: operand 0 must be FPSID or FPSCR pr FPEXC -- `vmsr FPINST2,r8'
arch/arm/vfp/vfphw.S:235: Error: operand 1 must be a VFP extension System Register -- `vmrs r3,FPINST'
arch/arm/vfp/vfphw.S:238: Error: operand 1 must be a VFP extension System Register -- `vmrs r12,FPINST2'

Use as-instr in Kconfig to check if FPINST/FPINST2 can be used. If they
can be used make use of .fpu directives and UAL VFP mnemonics for
register access.

This allows to build vfpmodule.c with Clang and its integrated assembler.

[0] https://reviews.llvm.org/D59733

Link: ClangBuiltLinux/linux#905

Signed-off-by: Stefan Agner <[email protected]>
Signed-off-by: Russell King <[email protected]>
[nd: Adjusted hunk from arch/arm/Kconfig due to missing
     commit 8a90a3228b6a ("arm: Unplug KVM from the build system").
     Replace __ACCESS_VFP with fmrx/fmxr for arch/arm/kvm/switch.c.
     CONFIG_KVM was removed in
     commit 541ad0150ca4 ("arm: Remove 32bit KVM host support") in
     v5.7.1. CONFIG_KVM is not enabled as part of allmodconfig or
     alldefconfig, but axm55xx_defconfig explicitly does enable it.
     Adjusted arch/arm/vfp/vfphw.S due to conflict with
     commit a47b395 ("ARM: 9030/1: entry: omit FP emulation for UND
     exceptions taken in kernel mode").]
Bug: 141693040
Signed-off-by: Nick Desaulniers <[email protected]>
Change-Id: I84733bd8649784e1e636e107c1dac03ef7d48136

Author: agners

arch/arm/Kconfig

@@ -2167,3 +2167,4 @@ source "arch/arm/crypto/Kconfig"
 endif
 
 source "arch/arm/kvm/Kconfig"
+source "arch/arm/Kconfig.assembler"

arch/arm/Kconfig.assembler

@@ -0,0 +1,6 @@
+# SPDX-License-Identifier: GPL-2.0
+
+config AS_VFP_VMRS_FPINST
+	def_bool $(as-instr,.fpu vfpv2\nvmrs r0$(comma)FPINST)
+	help
+	  Supported by binutils >= 2.24 and LLVM integrated assembler.

arch/arm/include/asm/kvm_hyp.h

@@ -14,9 +14,6 @@
 
 #define __hyp_text __section(.hyp.text) notrace
 
-#define __ACCESS_VFP(CRn)			\
-	"mrc", "mcr", __stringify(p10, 7, %0, CRn, cr0, 0), u32
-
 #define write_special(v, r)					\
 	asm volatile("msr " __stringify(r) ", %0" : : "r" (v))
 #define read_special(r) ({					\
@@ -79,8 +76,6 @@
 #define CNTV_CTL	__ACCESS_CP15(c14, 0, c3, 1)
 #define CNTHCTL		__ACCESS_CP15(c14, 4, c1, 0)
 
-#define VFP_FPEXC	__ACCESS_VFP(FPEXC)
-
 /* AArch64 compatibility macros, only for the timer so far */
 #define read_sysreg_el0(r)		read_sysreg(r##_EL0)
 #define write_sysreg_el0(v, r)		write_sysreg(v, r##_EL0)

arch/arm/include/asm/vfp.h

@@ -9,13 +9,15 @@
 #ifndef __ASM_VFP_H
 #define __ASM_VFP_H
 
+#ifndef CONFIG_AS_VFP_VMRS_FPINST
 #define FPSID			cr0
 #define FPSCR			cr1
 #define MVFR1			cr6
 #define MVFR0			cr7
 #define FPEXC			cr8
 #define FPINST			cr9
 #define FPINST2			cr10
+#endif
 
 /* FPSID bits */
 #define FPSID_IMPLEMENTER_BIT	(24)

arch/arm/include/asm/vfpmacros.h

@@ -8,14 +8,24 @@
 
 #include <asm/vfp.h>
 
-@ Macros to allow building with old toolkits (with no VFP support)
+#ifdef CONFIG_AS_VFP_VMRS_FPINST
+	.macro	VFPFMRX, rd, sysreg, cond
+	vmrs\cond	\rd, \sysreg
+	.endm
+
+	.macro	VFPFMXR, sysreg, rd, cond
+	vmsr\cond	\sysreg, \rd
+	.endm
+#else
+	@ Macros to allow building with old toolkits (with no VFP support)
 	.macro	VFPFMRX, rd, sysreg, cond
 	MRC\cond	p10, 7, \rd, \sysreg, cr0, 0	@ FMRX	\rd, \sysreg
 	.endm
 
 	.macro	VFPFMXR, sysreg, rd, cond
 	MCR\cond	p10, 7, \rd, \sysreg, cr0, 0	@ FMXR	\sysreg, \rd
 	.endm
+#endif
 
 	@ read all the working registers back into the VFP
 	.macro	VFPFLDMIA, base, tmp

arch/arm/kvm/Makefile

@@ -3,7 +3,7 @@
 # Makefile for Kernel-based Virtual Machine module
 #
 
-plus_virt := $(call as-instr,.arch_extension virt,+virt)
+plus_virt := $(call as-instr,".cpu cortex-a7\n\n.arch_extension virt",+virt)
 ifeq ($(plus_virt),+virt)
 	plus_virt_def := -DREQUIRES_VIRT=1
 endif

arch/arm/kvm/hyp/switch.c

@@ -8,6 +8,7 @@
 #include <asm/kvm_asm.h>
 #include <asm/kvm_hyp.h>
 #include <asm/kvm_mmu.h>
+#include "../../vfp/vfpinstr.h"
 
 __asm__(".arch_extension     virt");
 
@@ -26,10 +27,10 @@ static void __hyp_text __activate_traps(struct kvm_vcpu *vcpu, u32 *fpexc_host)
 	 * trap to SVC.  Therefore, always make sure that for 32-bit guests,
 	 * we set FPEXC.EN to prevent traps to SVC, when setting the TCP bits.
 	 */
-	val = read_sysreg(VFP_FPEXC);
+	val = fmrx(FPEXC);
 	*fpexc_host = val;
 	if (!(val & FPEXC_EN)) {
-		write_sysreg(val | FPEXC_EN, VFP_FPEXC);
+		fmxr(FPEXC, val | FPEXC_EN);
 		isb();
 	}
 
@@ -196,7 +197,7 @@ int __hyp_text __kvm_vcpu_run_nvhe(struct kvm_vcpu *vcpu)
 		__vfp_restore_state(&host_ctxt->vfp);
 	}
 
-	write_sysreg(fpexc, VFP_FPEXC);
+	fmxr(FPEXC, fpexc);
 
 	return exit_code;
 }

arch/arm/kvm/hyp/vfp.S

@@ -8,6 +8,7 @@
 #include <asm/vfpmacros.h>
 
 	.text
+	.fpu vfpv2
 	.pushsection	.hyp.text, "ax"
 
 /* void __vfp_save_state(struct vfp_hard_struct *vfp); */

arch/arm/kvm/init.S

@@ -35,6 +35,7 @@
  */
 
 	.text
+	.arch_extension virt
 	.pushsection    .hyp.idmap.text,"ax"
 	.align 5
 __kvm_hyp_init:

arch/arm/kvm/interrupts.S

@@ -7,6 +7,7 @@
 #include <linux/linkage.h>
 
 	.text
+	.arch_extension virt
 
 /********************************************************************
  *  Call function in Hyp mode