mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-22 12:33:33 +02:00
e2b0519ed8
This patch adds support for AArch64 (ARM's 64-bit architecture) to LLVM in the "experimental" category. Currently, it won't be built unless requested explicitly. This initial commit should have support for: + Assembly of all scalar (i.e. non-NEON, non-Crypto) instructions (except the late addition CRC instructions). + CodeGen features required for C++03 and C99. + Compilation for the "small" memory model: code+static data < 4GB. + Absolute and position-independent code. + GNU-style (i.e. "__thread") TLS. + Debugging information. The principal omission, currently, is performance tuning. This patch excludes the NEON support also reviewed due to an outbreak of batshit insanity in our legal department. That will be committed soon bringing the changes to precisely what has been approved. Further reviews would be gratefully received. llvm-svn: 174054
24 lines
800 B
LLVM
24 lines
800 B
LLVM
; RUN: llc -march=aarch64 -relocation-model=pic < %s | FileCheck %s
|
|
; RUN: llc -march=aarch64 -relocation-model=pic -filetype=obj < %s
|
|
|
|
; LLVM gives well-defined semantics to this horrible construct (though C says
|
|
; it's undefined). Regardless, we shouldn't crash. The important feature here is
|
|
; that in general the only way to access a GOT symbol is via a 64-bit
|
|
; load. Neither of these alternatives has the ELF relocations required to
|
|
; support it:
|
|
; + ldr wD, [xN, #:got_lo12:func]
|
|
; + add xD, xN, #:got_lo12:func
|
|
|
|
declare void @consume(i32)
|
|
declare void @func()
|
|
|
|
define void @foo() nounwind {
|
|
; CHECK: foo:
|
|
entry:
|
|
call void @consume(i32 ptrtoint (void ()* @func to i32))
|
|
; CHECK: adrp x[[ADDRHI:[0-9]+]], :got:func
|
|
; CHECK: ldr {{x[0-9]+}}, [x[[ADDRHI]], #:got_lo12:func]
|
|
ret void
|
|
}
|
|
|