[PATCH -mm] x86_64 EFI runtime service support : Calling convention fix (resend, cc LKML)

[PATCH -mm] x86_64 EFI runtime service support : Calling convention fix (resend, cc LKML)

[Huang, Ying]

In EFI calling convention, %xmm0 - %xmm5 are specified as the scratch
registers (UEFI Specification 2.1, 2.3.4.2). To conforms to EFI
specification, this patch save/restore %xmm0 - %xmm5 registers
before/after invoking EFI runtime service. At the same time, the stack
is aligned in 16 bytes, and TS in CR0 in clear/restore to make it
possible to use SSE2 in EFI runtime service.

This patch is based on 2.6.24-rc4-mm1. And it has been tested on Intel
platforms with 64-bit UEFI 2.0 firmware.

Signed-off-by: Huang Ying <ying.huang@intel.com>

---
 arch/x86/kernel/efi_stub_64.S |   71 +++++++++++++++++++++++++++++++++---------
 1 file changed, 56 insertions(+), 15 deletions(-)

--- a/arch/x86/kernel/efi_stub_64.S
+++ b/arch/x86/kernel/efi_stub_64.S
@@ -8,61 +8,102 @@
 
 #include <linux/linkage.h>
 
+#define SAVE_XMM			\
+	mov %rsp, %rax;			\
+	subq $0x70, %rsp;		\
+	and $~0xf, %rsp;		\
+	mov %rax, (%rsp);		\
+	mov %cr0, %rax;			\
+	clts;				\
+	mov %rax, 0x8(%rsp);		\
+	movaps %xmm0, 0x60(%rsp);	\
+	movaps %xmm1, 0x50(%rsp);	\
+	movaps %xmm2, 0x40(%rsp);	\
+	movaps %xmm3, 0x30(%rsp);	\
+	movaps %xmm4, 0x20(%rsp);	\
+	movaps %xmm5, 0x10(%rsp)
+
+#define RESTORE_XMM			\
+	movaps 0x60(%rsp), %xmm0;	\
+	movaps 0x50(%rsp), %xmm1;	\
+	movaps 0x40(%rsp), %xmm2;	\
+	movaps 0x30(%rsp), %xmm3;	\
+	movaps 0x20(%rsp), %xmm4;	\
+	movaps 0x10(%rsp), %xmm5;	\
+	mov 0x8(%rsp), %rsi;		\
+	mov %rsi, %cr0;			\
+	mov (%rsp), %rsp
+
 ENTRY(efi_call0)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $32, %rsp
 	call *%rdi
-	addq $40, %rsp
+	addq $32, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call1)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $32, %rsp
 	mov  %rsi, %rcx
 	call *%rdi
-	addq $40, %rsp
+	addq $32, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call2)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $32, %rsp
 	mov  %rsi, %rcx
 	call *%rdi
-	addq $40, %rsp
+	addq $32, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call3)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $32, %rsp
 	mov  %rcx, %r8
 	mov  %rsi, %rcx
 	call *%rdi
-	addq $40, %rsp
+	addq $32, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call4)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $32, %rsp
 	mov %r8, %r9
 	mov %rcx, %r8
 	mov %rsi, %rcx
 	call *%rdi
-	addq $40, %rsp
+	addq $32, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call5)
-	subq $40, %rsp
+	SAVE_XMM
+	subq $48, %rsp
 	mov %r9, 32(%rsp)
 	mov %r8, %r9
 	mov %rcx, %r8
 	mov %rsi, %rcx
 	call *%rdi
-	addq $40, %rsp
+	addq $48, %rsp
+	RESTORE_XMM
 	ret
 
 ENTRY(efi_call6)
-	subq $56, %rsp
-	mov 56+8(%rsp), %rax
+	SAVE_XMM
+	mov (%rsp), %rax
+	mov 8(%rax), %rax
+	subq $48, %rsp
 	mov %r9, 32(%rsp)
 	mov %rax, 40(%rsp)
 	mov %r8, %r9
 	mov %rcx, %r8
 	mov %rsi, %rcx
 	call *%rdi
-	addq $56, %rsp
+	addq $48, %rsp
+	RESTORE_XMM
 	ret