s390: add support for vector extension

The vector extension introduces 32 128-bit vector registers and a set of
instruction to operate on the vector registers.

The kernel can control the use of vector registers for the problem state
program with a bit in control register 0. Once enabled for a process the
kernel needs to retain the content of the vector registers on context
switch. The signal frame is extended to include the vector registers.
Two new register sets NT_S390_VXRS_LOW and NT_S390_VXRS_HIGH are added
to the regset interface for the debugger and core dumps.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

6 files changed, 58 insertions, 7 deletions

diff --git a/arch/s390/include/asm/elf.h b/arch/s390/include/asm/elf.h
index 78f4f8711d58..27735ae06a78 100644
--- a/arch/s390/include/asm/elf.h
+++ b/arch/s390/include/asm/elf.h
@@ -102,6 +102,7 @@
 #define HWCAP_S390_ETF3EH	256
 #define HWCAP_S390_HIGH_GPRS	512
 #define HWCAP_S390_TE		1024
+#define HWCAP_S390_VXRS		2048
 
 /*
  * These are used to set parameters in the core dumps.
diff --git a/arch/s390/include/asm/lowcore.h b/arch/s390/include/asm/lowcore.h
index 4349197ab9df..d812cf1a8177 100644
--- a/arch/s390/include/asm/lowcore.h
+++ b/arch/s390/include/asm/lowcore.h
@@ -310,7 +310,10 @@ struct _lowcore {
 
 	/* Extended facility list */
 	__u64	stfle_fac_list[32];		/* 0x0f00 */
-	__u8	pad_0x1000[0x11b8-0x1000];	/* 0x1000 */
+	__u8	pad_0x1000[0x11b0-0x1000];	/* 0x1000 */
+
+	/* Pointer to vector register save area */
+	__u64	vector_save_area_addr;		/* 0x11b0 */
 
 	/* 64 bit extparam used for pfault/diag 250: defined by architecture */
 	__u64	ext_params2;			/* 0x11B8 */
@@ -334,9 +337,10 @@ struct _lowcore {
 
 	/* Transaction abort diagnostic block */
 	__u8	pgm_tdb[256];			/* 0x1800 */
+	__u8	pad_0x1900[0x1c00-0x1900];	/* 0x1900 */
 
-	/* align to the top of the prefix area */
-	__u8	pad_0x1900[0x2000-0x1900];	/* 0x1900 */
+	/* Software defined save area for vector registers */
+	__u8	vector_save_area[1024];		/* 0x1c00 */
 } __packed;
 
 #endif /* CONFIG_32BIT */
diff --git a/arch/s390/include/asm/nmi.h b/arch/s390/include/asm/nmi.h
index 35f8ec185616..3027a5a72b74 100644
--- a/arch/s390/include/asm/nmi.h
+++ b/arch/s390/include/asm/nmi.h
@@ -38,7 +38,7 @@ struct mci {
 	__u32 pm :  1; /* 22 psw program mask and cc validity */
 	__u32 ia :  1; /* 23 psw instruction address validity */
 	__u32 fa :  1; /* 24 failing storage address validity */
-	__u32	 :  1; /* 25 */
+	__u32 vr :  1; /* 25 vector register validity */
 	__u32 ec :  1; /* 26 external damage code validity */
 	__u32 fp :  1; /* 27 floating point register validity */
 	__u32 gr :  1; /* 28 general register validity */
diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h
index 3d0871058306..d559bdb03d18 100644
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -117,6 +117,7 @@ struct thread_struct {
 	int ri_signum;
 #ifdef CONFIG_64BIT
 	unsigned char trap_tdb[256];	/* Transaction abort diagnose block */
+	__vector128 *vxrs;		/* Vector register save area */
 #endif
 };
 
diff --git a/arch/s390/include/asm/setup.h b/arch/s390/include/asm/setup.h
index dbde7c236577..7736fdd72595 100644
--- a/arch/s390/include/asm/setup.h
+++ b/arch/s390/include/asm/setup.h
@@ -56,6 +56,7 @@ extern void detect_memory_memblock(void);
 #define MACHINE_FLAG_TOPOLOGY	(1UL << 14)
 #define MACHINE_FLAG_TE		(1UL << 15)
 #define MACHINE_FLAG_TLB_LC	(1UL << 17)
+#define MACHINE_FLAG_VX		(1UL << 18)
 
 #define MACHINE_IS_VM		(S390_lowcore.machine_flags & MACHINE_FLAG_VM)
 #define MACHINE_IS_KVM		(S390_lowcore.machine_flags & MACHINE_FLAG_KVM)
@@ -78,6 +79,7 @@ extern void detect_memory_memblock(void);
 #define MACHINE_HAS_TOPOLOGY	(0)
 #define MACHINE_HAS_TE		(0)
 #define MACHINE_HAS_TLB_LC	(0)
+#define MACHINE_HAS_VX		(0)
 #else /* CONFIG_64BIT */
 #define MACHINE_HAS_IEEE	(1)
 #define MACHINE_HAS_CSP		(1)
@@ -90,6 +92,7 @@ extern void detect_memory_memblock(void);
 #define MACHINE_HAS_TOPOLOGY	(S390_lowcore.machine_flags & MACHINE_FLAG_TOPOLOGY)
 #define MACHINE_HAS_TE		(S390_lowcore.machine_flags & MACHINE_FLAG_TE)
 #define MACHINE_HAS_TLB_LC	(S390_lowcore.machine_flags & MACHINE_FLAG_TLB_LC)
+#define MACHINE_HAS_VX		(S390_lowcore.machine_flags & MACHINE_FLAG_VX)
 #endif /* CONFIG_64BIT */
 
 /*
diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h
index 18ea9e3f8142..0e0109578021 100644
--- a/arch/s390/include/asm/switch_to.h
+++ b/arch/s390/include/asm/switch_to.h
@@ -103,6 +103,48 @@ static inline void restore_fp_regs(freg_t *fprs)
 	asm volatile("ld 15,%0" : : "Q" (fprs[15]));
 }
 
+static inline void save_vx_regs(__vector128 *vxrs)
+{
+	typedef struct { __vector128 _[__NUM_VXRS]; } addrtype;
+
+	asm volatile(
+		"	la	1,%0\n"
+		"	.word	0xe70f,0x1000,0x003e\n"	/* vstm 0,15,0(1) */
+		"	.word	0xe70f,0x1100,0x0c3e\n"	/* vstm 16,31,256(1) */
+		: "=Q" (*(addrtype *) vxrs) : : "1");
+}
+
+static inline void restore_vx_regs(__vector128 *vxrs)
+{
+	typedef struct { __vector128 _[__NUM_VXRS]; } addrtype;
+
+	asm volatile(
+		"	la	1,%0\n"
+		"	.word	0xe70f,0x1000,0x0036\n"	/* vlm 0,15,0(1) */
+		"	.word	0xe70f,0x1100,0x0c36\n"	/* vlm 16,31,256(1) */
+		: : "Q" (*(addrtype *) vxrs) : "1");
+}
+
+static inline void save_fp_vx_regs(struct task_struct *task)
+{
+#ifdef CONFIG_64BIT
+	if (task->thread.vxrs)
+		save_vx_regs(task->thread.vxrs);
+	else
+#endif
+	save_fp_regs(task->thread.fp_regs.fprs);
+}
+
+static inline void restore_fp_vx_regs(struct task_struct *task)
+{
+#ifdef CONFIG_64BIT
+	if (task->thread.vxrs)
+		restore_vx_regs(task->thread.vxrs);
+	else
+#endif
+	restore_fp_regs(task->thread.fp_regs.fprs);
+}
+
 static inline void save_access_regs(unsigned int *acrs)
 {
 	typedef struct { int _[NUM_ACRS]; } acrstype;
@@ -120,16 +162,16 @@ static inline void restore_access_regs(unsigned int *acrs)
 #define switch_to(prev,next,last) do {					\
 	if (prev->mm) {							\
 		save_fp_ctl(&prev->thread.fp_regs.fpc);			\
-		save_fp_regs(prev->thread.fp_regs.fprs);		\
+		save_fp_vx_regs(prev);					\
 		save_access_regs(&prev->thread.acrs[0]);		\
 		save_ri_cb(prev->thread.ri_cb);				\
 	}								\
 	if (next->mm) {							\
+		update_cr_regs(next);					\
 		restore_fp_ctl(&next->thread.fp_regs.fpc);		\
-		restore_fp_regs(next->thread.fp_regs.fprs);		\
+		restore_fp_vx_regs(next);				\
 		restore_access_regs(&next->thread.acrs[0]);		\
 		restore_ri_cb(next->thread.ri_cb, prev->thread.ri_cb);	\
-		update_cr_regs(next);					\
 	}								\
 	prev = __switch_to(prev,next);					\
 } while (0)