s390/fpu: improve kernel_fpu_[begin|end]

In case of nested user of the FPU or vector registers in the kernel
the current code uses the mask of the FPU/vector registers of the
previous contexts to decide which registers to save and restore.
E.g. if the previous context used KERNEL_VXR_V0V7 and the next
context wants to use KERNEL_VXR_V24V31 the first 8 vector registers
are stored to the FPU state structure. But this is not necessary
as the next context does not use these registers.

Rework the FPU/vector register save and restore code. The new code
does a few things differently:
1) A lowcore field is used instead of a per-cpu variable.
2) The kernel_fpu_end function now has two parameters just like
   kernel_fpu_begin. The register flags are required by both
   functions to save / restore the minimal register set.
3) The inline functions kernel_fpu_begin/kernel_fpu_end now do the
   update of the register masks. If the user space FPU registers
   have already been stored neither save_fpu_regs nor the
   __kernel_fpu_begin/__kernel_fpu_end functions have to be called
   for the first context. In this case kernel_fpu_begin adds 7
   instructions and kernel_fpu_end adds 4 instructions.
3) The inline assemblies in __kernel_fpu_begin / __kernel_fpu_end
   to save / restore the vector registers are simplified a bit.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

arch/s390/crypto/crc32-vx.c

@@ -67,7 +67,7 @@ u32 crc32c_le_vgfm_16(u32 crc, unsigned char const *buf, size_t size);
 									    \
 		kernel_fpu_begin(&vxstate, KERNEL_VXR_LOW);		    \
 		crc = ___crc32_vx(crc, data, aligned);			    \
-		kernel_fpu_end(&vxstate);				    \
+		kernel_fpu_end(&vxstate, KERNEL_VXR_LOW);		    \
 									    \
 		if (remaining)						    \
 			crc = ___crc32_sw(crc, data + aligned, remaining);  \

arch/s390/include/asm/fpu/api.h

@@ -64,18 +64,18 @@ static inline int test_fp_ctl(u32 fpc)
 	return rc;
 }
 
-#define KERNEL_VXR_V0V7		1
-#define KERNEL_VXR_V8V15	2
-#define KERNEL_VXR_V16V23	4
-#define KERNEL_VXR_V24V31	8
-#define KERNEL_FPR		16
-#define KERNEL_FPC		256
+#define KERNEL_FPC		1
+#define KERNEL_VXR_V0V7		2
+#define KERNEL_VXR_V8V15	4
+#define KERNEL_VXR_V16V23	8
+#define KERNEL_VXR_V24V31	16
 
 #define KERNEL_VXR_LOW		(KERNEL_VXR_V0V7|KERNEL_VXR_V8V15)
 #define KERNEL_VXR_MID		(KERNEL_VXR_V8V15|KERNEL_VXR_V16V23)
 #define KERNEL_VXR_HIGH		(KERNEL_VXR_V16V23|KERNEL_VXR_V24V31)
 
-#define KERNEL_FPU_MASK		(KERNEL_VXR_LOW|KERNEL_VXR_HIGH|KERNEL_FPR)
+#define KERNEL_VXR		(KERNEL_VXR_LOW|KERNEL_VXR_HIGH)
+#define KERNEL_FPR		(KERNEL_FPC|KERNEL_VXR_V0V7)
 
 struct kernel_fpu;
 
@@ -87,18 +87,28 @@ struct kernel_fpu;
  * Prefer using the kernel_fpu_begin()/kernel_fpu_end() pair of functions.
  */
 void __kernel_fpu_begin(struct kernel_fpu *state, u32 flags);
-void __kernel_fpu_end(struct kernel_fpu *state);
+void __kernel_fpu_end(struct kernel_fpu *state, u32 flags);
 
 
 static inline void kernel_fpu_begin(struct kernel_fpu *state, u32 flags)
 {
 	preempt_disable();
+	state->mask = S390_lowcore.fpu_flags;
+	if (!test_cpu_flag(CIF_FPU))
+		/* Save user space FPU state and register contents */
+		save_fpu_regs();
+	else if (state->mask & flags)
+		/* Save FPU/vector register in-use by the kernel */
 		__kernel_fpu_begin(state, flags);
+	S390_lowcore.fpu_flags |= flags;
 }
 
-static inline void kernel_fpu_end(struct kernel_fpu *state)
+static inline void kernel_fpu_end(struct kernel_fpu *state, u32 flags)
 {
-	__kernel_fpu_end(state);
+	S390_lowcore.fpu_flags = state->mask;
+	if (state->mask & flags)
+		/* Restore FPU/vector register in-use by the kernel */
+		__kernel_fpu_end(state, flags);
 	preempt_enable();
 }
 

arch/s390/include/asm/lowcore.h

@@ -129,7 +129,8 @@ struct lowcore {
 	__u8	pad_0x0390[0x0398-0x0390];	/* 0x0390 */
 	__u64	gmap;				/* 0x0398 */
 	__u32	spinlock_lockval;		/* 0x03a0 */
-	__u8	pad_0x03a0[0x0400-0x03a4];	/* 0x03a4 */
+	__u32	fpu_flags;			/* 0x03a4 */
+	__u8	pad_0x03a8[0x0400-0x03a8];	/* 0x03a8 */
 
 	/* Per cpu primary space access list */
 	__u32	paste[16];			/* 0x0400 */

arch/s390/kernel/sysinfo.c

@@ -454,7 +454,7 @@ void s390_adjust_jiffies(void)
 			: "Q" (info->capability), "d" (10000000), "d" (0)
 			: "cc"
 			);
-		kernel_fpu_end(&fpu);
+		kernel_fpu_end(&fpu, KERNEL_FPR);
 	} else
 		/*
 		 * Really old machine without stsi block for basic