Preserve xmm registers on x86_64

Summary: public

`%xmm` registers weren't being preserved what would cause eventual weird issues
for methods using floats / doubles / etc.

Reviewed By: jspahrsummers

Differential Revision: D2581358

fb-gh-sync-id: 701498def0f05716c665f4749e5154b828bf41ec

React/Profiler/RCTProfileTrampoline-x86_64.S

@@ -22,6 +22,15 @@ SYMBOL_NAME(RCTProfileTrampoline):
    * We have to save %r12 since its value should be preserved across function
    * calls and we'll use it to keep the stack pointer
    */
+  subq $0x80+8, %rsp // 8 x 16-bytes xmm registers + 8-bytes alignment
+  movdqa  %xmm0, 0x70(%rsp)
+  movdqa  %xmm1, 0x60(%rsp)
+  movdqa  %xmm2, 0x50(%rsp)
+  movdqa  %xmm3, 0x40(%rsp)
+  movdqa  %xmm4, 0x30(%rsp)
+  movdqa  %xmm5, 0x20(%rsp)
+  movdqa  %xmm6, 0x10(%rsp)
+  movdqa  %xmm7, 0x00(%rsp)
   pushq %rdi
   pushq %rsi
   pushq %rdx
@@ -82,7 +91,7 @@ SYMBOL_NAME(RCTProfileTrampoline):
   movq %r14, 0x8(%rax)
 
   // mov the pointers we need to the callee saved registers
-  movq 0x48(%rsp), %r13 // caller of RCTProfileTrampoline
+  movq 0xd8(%rsp), %r13 // caller of RCTProfileTrampoline (0xd8 is stack top)
   movq %rax, %r14 // allocated memory's address
 
   /**
@@ -108,6 +117,15 @@ SYMBOL_NAME(RCTProfileTrampoline):
   popq %rdx
   popq %rsi
   popq %rdi
+  movdqa 0x00(%rsp), %xmm7
+  movdqa 0x10(%rsp), %xmm6
+  movdqa 0x20(%rsp), %xmm5
+  movdqa 0x30(%rsp), %xmm4
+  movdqa 0x40(%rsp), %xmm3
+  movdqa 0x50(%rsp), %xmm2
+  movdqa 0x60(%rsp), %xmm1
+  movdqa 0x70(%rsp), %xmm0
+  addq $0x80+8, %rsp
 
   /**
    * delete parent caller (saved in %r13) `call` will add the new address so
@@ -118,6 +136,8 @@ SYMBOL_NAME(RCTProfileTrampoline):
   // call the actual function and save the return value
   callq *%r11
   pushq %rax
+  subq $0x10+8, %rsp //16-bytes xmm register + 8-bytes for alignment
+  movdqa %xmm0, (%rsp)
 
   // align stack
   pushq %r12
@@ -131,15 +151,12 @@ SYMBOL_NAME(RCTProfileTrampoline):
   movq %r12, %rsp
   popq %r12
 
-  // save the return of the actual function call
-  popq %rax
-
   /**
    * Restore the initial value of the callee saved registers, saved in the
    * memory allocated.
    */
   movq %r13, %rcx
-  movq %r14, %rdx
+  movq %r14, %rdi
   movq 0x0(%r14), %r13
   movq 0x8(%r14), %r14
 
@@ -148,13 +165,11 @@ SYMBOL_NAME(RCTProfileTrampoline):
    * memory) and align the stack
    */
   pushq %rcx
-  pushq %rax
   pushq %r12
   movq %rsp, %r12
   andq $-0x10, %rsp
 
   // Free the memory allocated to stash callee saved registers
-  movq %rdx, %rdi
   callq SYMBOL_NAME(free)
 
   // unalign  stack and restore %r12
@@ -165,8 +180,10 @@ SYMBOL_NAME(RCTProfileTrampoline):
    * pop the caller address to %rcx and the actual function return value to
    * %rax, so it's the return value of RCTProfileTrampoline
    */
-  popq %rax
   popq %rcx
+  movdqa (%rsp), %xmm0
+  addq $0x10+8, %rsp
+  popq %rax
 
   // jump to caller
   jmpq *%rcx