[ARM] 3105/4: ARM EABI: new syscall entry convention

Patch from Nicolas Pitre

For a while we wanted to change the way syscalls were called on ARM.
Instead of encoding the syscall number in the swi instruction which
requires reading back the instruction from memory to extract that number
and polluting the data cache, it was decided that simply storing the
syscall number into r7 would be more efficient. Since this represents
an ABI change then making that change at the same time as EABI support
is the right thing to do.

It is now expected that EABI user space binaries put the syscall number
into r7 and use "swi 0" to call the kernel. Syscall register argument
are also expected to have "EABI arrangement" i.e. 64-bit arguments
should be put in a pair of registers from an even register number.

Example with long ftruncate64(unsigned int fd, loff_t length):

legacy ABI:
- put fd into r0
- put length into r1-r2
- use "swi #(0x900000 + 194)" to call the kernel

new ARM EABI:
- put fd into r0
- put length into r2-r3 (skipping over r1)
- put 194 into r7
- use "swi 0" to call the kernel

Note that it is important to use 0 for the swi argument as backward
compatibility with legacy ABI user space relies on this.
The syscall macros in asm-arm/unistd.h were also updated to support
both ABIs and implement the right call method automatically.

Signed-off-by: Nicolas Pitre <[e-mail]>
Signed-off-by: Russell King <[e-mail]>

arch/arm/kernel/entry-common.S

···
 	   run on an ARM7 and we can save a couple of instructions.  
 								--pb */
 #ifdef CONFIG_CPU_ARM710
-	.macro	arm710_bug_check, instr, temp
-	and	\temp, \instr, #0x0f000000	@ check for SWI
-	teq	\temp, #0x0f000000
-	bne	.Larm700bug
-	.endm
-
-.Larm700bug:
+#define A710(code...) code
+.Larm710bug:
 	ldmia	sp, {r0 - lr}^			@ Get calling r0 - lr
 	mov	r0, r0
 	add	sp, sp, #S_FRAME_SIZE
 	subs	pc, lr, #4
 #else
-	.macro	arm710_bug_check, instr, temp
-	.endm
+#define A710(code...)
 #endif
 
 	.align	5
···
 	/*
 	 * Get the system call number.
 	 */
-#ifdef CONFIG_ARM_THUMB
+#if defined(CONFIG_AEABI)
+
+	@ syscall number is in scno (r7) already.
+
+  A710(	ldr	ip, [lr, #-4]			@ get SWI instruction	)
+  A710(	and	ip, ip, #0x0f000000		@ check for SWI		)
+  A710(	teq	ip, #0x0f000000						)
+  A710(	bne	.Larm710bug						)
+#elif defined(CONFIG_ARM_THUMB)
 	tst	r8, #PSR_T_BIT			@ this is SPSR from save_user_regs
 	addne	scno, r7, #__NR_SYSCALL_BASE	@ put OS number in
 	ldreq	scno, [lr, #-4]
 #else
 	ldr	scno, [lr, #-4]			@ get SWI instruction
+  A710(	and	ip, scno, #0x0f000000		@ check for SWI		)
+  A710(	teq	ip, #0x0f000000						)
+  A710(	bne	.Larm710bug						)
 #endif
-	arm710_bug_check scno, ip
 
 #ifdef CONFIG_ALIGNMENT_TRAP
 	ldr	ip, __cr_alignment
···
 #endif
 	enable_irq
 
-	stmdb	sp!, {r4, r5}			@ push fifth and sixth args
-
 	get_thread_info tsk
 	ldr	ip, [tsk, #TI_FLAGS]		@ check for syscall tracing
+#ifndef CONFIG_AEABI
 	bic	scno, scno, #0xff000000		@ mask off SWI op-code
 	eor	scno, scno, #__NR_SYSCALL_BASE	@ check OS number
+#endif
 	adr	tbl, sys_call_table		@ load syscall table pointer
+	stmdb	sp!, {r4, r5}			@ push fifth and sixth args
 	tst	ip, #_TIF_SYSCALL_TRACE		@ are we tracing syscalls?
 	bne	__sys_trace
 
-	adr	lr, ret_fast_syscall		@ return address
 	cmp	scno, #NR_syscalls		@ check upper syscall limit
+	adr	lr, ret_fast_syscall		@ return address
 	ldrcc	pc, [tbl, scno, lsl #2]		@ call sys_* routine
 
 	add	r1, sp, #S_OFF
···
 @ r8 = syscall table
 		.type	sys_syscall, #function
 sys_syscall:
+#ifndef CONFIG_AEABI
 		eor	scno, r0, #__NR_SYSCALL_BASE
 		cmp	scno, #__NR_syscall - __NR_SYSCALL_BASE
 		cmpne	scno, #NR_syscalls	@ check range
···
 		movlo	r2, r3
 		movlo	r3, r4
 		ldrlo	pc, [tbl, scno, lsl #2]
+#endif
 		b	sys_ni_syscall
 
 sys_fork_wrapper:

arch/arm/kernel/traps.c

···
 	struct thread_info *thread = current_thread_info();
 	siginfo_t info;
 
-	if ((no >> 16) != 0x9f)
+	if ((no >> 16) != (__ARM_NR_BASE>> 16))
 		return bad_syscall(no, regs);
 
 	switch (no & 0xffff) {

include/asm-arm/unistd.h

···
 
 #include <linux/linkage.h>
 
-#if defined(__thumb__)
+#define __NR_OABI_SYSCALL_BASE	0x900000
+
+#if defined(__thumb__) || defined(__ARM_EABI__)
 #define __NR_SYSCALL_BASE	0
 #else
-#define __NR_SYSCALL_BASE	0x900000
+#define __NR_SYSCALL_BASE	__NR_OABI_SYSCALL_BASE
 #endif
 
 /*
···
 #define __sys1(x) __sys2(x)
 
 #ifndef __syscall
-#if defined(__thumb__)
-#define __syscall(name)					\
-	"push	{r7}\n\t"				\
-	"mov	r7, #" __sys1(__NR_##name) "\n\t"	\
-	"swi	0\n\t"					\
-	"pop	{r7}"
+#if defined(__thumb__) || defined(__ARM_EABI__)
+#define __SYS_REG(name) register long __sysreg __asm__("r7") = __NR_##name;
+#define __SYS_REG_LIST(regs...) "r" (__sysreg) , ##regs
+#define __syscall(name) "swi\t0"
 #else
+#define __SYS_REG(name)
+#define __SYS_REG_LIST(regs...) regs
 #define __syscall(name) "swi\t" __sys1(__NR_##name) ""
 #endif
 #endif
···
 
 #define _syscall0(type,name)						\
 type name(void) {							\
+  __SYS_REG(name)							\
   register long __res_r0 __asm__("r0");					\
   long __res;								\
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	:								\
-	: "lr");							\
+	: __SYS_REG_LIST() );						\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
 
 #define _syscall1(type,name,type1,arg1) 				\
 type name(type1 arg1) { 						\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __res_r0 __asm__("r0");					\
   long __res;								\
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0)							\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0) ) );				\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
 
 #define _syscall2(type,name,type1,arg1,type2,arg2)			\
 type name(type1 arg1,type2 arg2) {					\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __r1 __asm__("r1") = (long)arg2;			\
   register long __res_r0 __asm__("r0");					\
···
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0),"r" (__r1) 					\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0), "r" (__r1) ) );			\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
···
 
 #define _syscall3(type,name,type1,arg1,type2,arg2,type3,arg3)		\
 type name(type1 arg1,type2 arg2,type3 arg3) {				\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __r1 __asm__("r1") = (long)arg2;			\
   register long __r2 __asm__("r2") = (long)arg3;			\
···
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0),"r" (__r1),"r" (__r2)				\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0), "r" (__r1), "r" (__r2) ) );	\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
···
 
 #define _syscall4(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4)\
 type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4) {		\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __r1 __asm__("r1") = (long)arg2;			\
   register long __r2 __asm__("r2") = (long)arg3;			\
···
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0),"r" (__r1),"r" (__r2),"r" (__r3)			\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0), "r" (__r1), "r" (__r2), "r" (__r3) ) ); \
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
···
 
 #define _syscall5(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4,type5,arg5)	\
 type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4, type5 arg5) {	\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __r1 __asm__("r1") = (long)arg2;			\
   register long __r2 __asm__("r2") = (long)arg3;			\
···
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0),"r" (__r1),"r" (__r2),"r" (__r3),"r" (__r4)	\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0), "r" (__r1), "r" (__r2),		\
+			  "r" (__r3), "r" (__r4) ) );			\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }
 
 #define _syscall6(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4,type5,arg5,type6,arg6)	\
 type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4, type5 arg5, type6 arg6) {	\
+  __SYS_REG(name)							\
   register long __r0 __asm__("r0") = (long)arg1;			\
   register long __r1 __asm__("r1") = (long)arg2;			\
   register long __r2 __asm__("r2") = (long)arg3;			\
···
   __asm__ __volatile__ (						\
   __syscall(name)							\
 	: "=r" (__res_r0)						\
-	: "r" (__r0),"r" (__r1),"r" (__r2),"r" (__r3), "r" (__r4),"r" (__r5)		\
-	: "lr");							\
+	: __SYS_REG_LIST( "0" (__r0), "r" (__r1), "r" (__r2),		\
+			  "r" (__r3), "r" (__r4), "r" (__r5) ) );	\
   __res = __res_r0;							\
   __syscall_return(type,__res);						\
 }