This is an archive of the discontinued LLVM Phabricator instance.

Differential D150196

[LoongArch] Remove AssemblerPredicate for features: f/d/lsx/lasx/lvz/lbt
ClosedPublic

Authored by SixWeining on May 9 2023, 6:08 AM.

Details

Reviewers
xen0n
xry111
hev
wangleiat
tangyouling
Commits
rG9a8727df35da: [LoongArch] Remove AssemblerPredicate for features: f/d/lsx/lasx/lvz/lbt
Summary

Linux LoongArch port [1] uses -msoft-float (implies no FPU) in its
cflags while it also uses floating-point insns in asm sources [2].
GAS allows this usage while IAS currently does not.

This patch removes AssemblerPredicates for floating-point insns so
that to make IAS compitable with GAS. Similarly, also remove
AssemblerPredicates for other ISA extensions, i.e. lsx/las/lvz/lbt.

[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/loongarch/Makefile?h=v6.4-rc1#n49
[2]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/loongarch/kernel/fpu.S?h=v6.4-rc1#n29

Diff Detail

Event Timeline

SixWeining created this revision.May 9 2023, 6:08 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 9 2023, 6:08 AM
Herald added a subscriber: hiraditya. · View Herald Transcript
SixWeining requested review of this revision.May 9 2023, 6:08 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 9 2023, 6:08 AM
Herald added a subscriber: llvm-commits. · View Herald Transcript
xry111 added a comment.May 9 2023, 6:38 AM

Hmm, I think it's a kernel issue and should be fixed by using -mabi=lp64s instead of -msoft-float for this specific file...

Harbormaster completed remote builds in B230852: Diff 520671.May 9 2023, 6:41 AM
SixWeining added a comment.May 9 2023, 8:08 AM
In D150196#4329518, @xry111 wrote:

Hmm, I think it's a kernel issue and should be fixed by using -mabi=lp64s instead of -msoft-float for this specific file...

Makes sense. I'm not sure whether kernel‘s Makefile can distinguish .c and .S (or .c with floating-point insn inlineasm). I'll report this issue to Huacai tomorrow.

Seems there is another approach: add an assembler option to IAS called -mfpu=64/32/0/none.
Then kernel Makefile can be:

if LLVM == 1 && LLVM_IAS ==1
  cflags-y = -msolft-float -Wa,-mfpu=64
endif
xry111 added a comment.May 9 2023, 8:17 AM
In D150196#4329730, @SixWeining wrote:
In D150196#4329518, @xry111 wrote:

Hmm, I think it's a kernel issue and should be fixed by using -mabi=lp64s instead of -msoft-float for this specific file...

Makes sense. I'm not sure whether kernel‘s Makefile can distinguish .c and .S (or .c with floating-point insn inlineasm). I'll report this issue to Huacai tomorrow.

AFAIK we just need

CFLAGS_fpu.o  = -mabi=lp64s
CFLAGS_REMOVE_fpu.o = -msoft-float

KBuild has a fine-grained control on CFLAGS for TUs.

Seems there is another approach: add an assembler option to IAS called -mfpu=64/32/0/none.
Then kernel Makefile can be:

if LLVM == 1 && LLVM_IAS ==1
  cflags-y = -msolft-float -Wa,-mfpu=64
endif
SixWeining added a comment.May 9 2023, 8:20 AM

AFAIK we just need

CFLAGS_fpu.o  = -mabi=lp64s
CFLAGS_REMOVE_fpu.o = -msoft-float

KBuild has a fine-grained control on CFLAGS for TUs.

Looks great!

hev added a comment.May 9 2023, 8:31 AM

AFAIK we just need

CFLAGS_fpu.o  = -mabi=lp64s
CFLAGS_REMOVE_fpu.o = -msoft-float

Does this work for performing floating-point operations using inline assembly in C source files?

hev added a comment.Jun 7 2023, 8:34 AM

Just like in the Linux kernel, In bare-metal development with Rust, there is also need for compiling targets with software floating-point support to access the floating-point context through hardware floating-point instructions, potentially even within Rust's inline assembly. This should be one of the ways to address the problem.

Thanks.

hev accepted this revision.Jun 14 2023, 8:29 AM
This revision is now accepted and ready to land.Jun 14 2023, 8:29 AM
xen0n accepted this revision.Jun 21 2023, 8:30 PM
This revision was landed with ongoing or failed builds.Jun 25 2023, 7:38 PM
Closed by commit rG9a8727df35da: [LoongArch] Remove AssemblerPredicate for features: f/d/lsx/lasx/lvz/lbt (authored by SixWeining). · Explain Why
This revision was automatically updated to reflect the committed changes.
SixWeining added a commit: rG9a8727df35da: [LoongArch] Remove AssemblerPredicate for features: f/d/lsx/lasx/lvz/lbt.

Revision Contents

PathSize
llvm/
lib/
Target/
LoongArch/
30 lines
test/
MC/
LoongArch/
Basic/
Float/
8 lines
8 lines
8 lines
8 lines
8 lines
2 lines
8 lines
8 lines
8 lines
8 lines
8 lines
8 lines
8 lines
8 lines
8 lines
Integer/
12 lines

Diff 534409

llvm/lib/Target/LoongArch/LoongArch.td

Show All 30 Lines
defvar LA32 = DefaultMode; defvar LA32 = DefaultMode;
def LA64 : HwMode<"+64bit", [IsLA64]>; def LA64 : HwMode<"+64bit", [IsLA64]>;
// Single Precision floating point // Single Precision floating point
def FeatureBasicF def FeatureBasicF
: SubtargetFeature<"f", "HasBasicF", "true", : SubtargetFeature<"f", "HasBasicF", "true",
"'F' (Single-Precision Floating-Point)">; "'F' (Single-Precision Floating-Point)">;
def HasBasicF def HasBasicF : Predicate<"Subtarget->hasBasicF()">;
: Predicate<"Subtarget->hasBasicF()">,
AssemblerPredicate<(all_of FeatureBasicF),
"'F' (Single-Precision Floating-Point)">;
// Double Precision floating point // Double Precision floating point
def FeatureBasicD def FeatureBasicD
: SubtargetFeature<"d", "HasBasicD", "true", : SubtargetFeature<"d", "HasBasicD", "true",
"'D' (Double-Precision Floating-Point)", "'D' (Double-Precision Floating-Point)",
[FeatureBasicF]>; [FeatureBasicF]>;
def HasBasicD def HasBasicD : Predicate<"Subtarget->hasBasicD()">;
: Predicate<"Subtarget->hasBasicD()">,
AssemblerPredicate<(all_of FeatureBasicD),
"'D' (Double-Precision Floating-Point)">;
// Loongson SIMD eXtension (LSX) // Loongson SIMD eXtension (LSX)
def FeatureExtLSX def FeatureExtLSX
: SubtargetFeature<"lsx", "HasExtLSX", "true", : SubtargetFeature<"lsx", "HasExtLSX", "true",
"'LSX' (Loongson SIMD Extension)", [FeatureBasicD]>; "'LSX' (Loongson SIMD Extension)", [FeatureBasicD]>;
def HasExtLSX def HasExtLSX : Predicate<"Subtarget->hasExtLSX()">;
: Predicate<"Subtarget->hasExtLSX()">,
AssemblerPredicate<(all_of FeatureExtLSX),
"'LSX' (Loongson SIMD Extension)">;
// Loongson Advanced SIMD eXtension (LASX) // Loongson Advanced SIMD eXtension (LASX)
def FeatureExtLASX def FeatureExtLASX
: SubtargetFeature<"lasx", "HasExtLASX", "true", : SubtargetFeature<"lasx", "HasExtLASX", "true",
"'LASX' (Loongson Advanced SIMD Extension)", "'LASX' (Loongson Advanced SIMD Extension)",
[FeatureExtLSX]>; [FeatureExtLSX]>;
def HasExtLASX def HasExtLASX : Predicate<"Subtarget->hasExtLASX()">;
: Predicate<"Subtarget->hasExtLASX()">,
AssemblerPredicate<(all_of FeatureExtLASX),
"'LASX' (Loongson Advanced SIMD Extension)">;
// Loongson VirtualiZation (LVZ) // Loongson VirtualiZation (LVZ)
def FeatureExtLVZ def FeatureExtLVZ
: SubtargetFeature<"lvz", "HasExtLVZ", "true", : SubtargetFeature<"lvz", "HasExtLVZ", "true",
"'LVZ' (Loongson Virtualization Extension)">; "'LVZ' (Loongson Virtualization Extension)">;
def HasExtLVZ def HasExtLVZ : Predicate<"Subtarget->hasExtLVZ()">;
: Predicate<"Subtarget->hasExtLVZ()">,
AssemblerPredicate<(all_of FeatureExtLVZ),
"'LVZ' (Loongson Virtualization Extension)">;
// Loongson Binary Translation (LBT) // Loongson Binary Translation (LBT)
def FeatureExtLBT def FeatureExtLBT
: SubtargetFeature<"lbt", "HasExtLBT", "true", : SubtargetFeature<"lbt", "HasExtLBT", "true",
"'LBT' (Loongson Binary Translation Extension)">; "'LBT' (Loongson Binary Translation Extension)">;
def HasExtLBT def HasExtLBT : Predicate<"Subtarget->hasExtLBT()">;
: Predicate<"Subtarget->hasExtLBT()">,
AssemblerPredicate<(all_of FeatureExtLBT),
"'LBT' (Loongson Binary Translation Extension)">;
// Expand la.global as la.pcrel // Expand la.global as la.pcrel
def LaGlobalWithPcrel def LaGlobalWithPcrel
: SubtargetFeature<"la-global-with-pcrel", "HasLaGlobalWithPcrel", "true", : SubtargetFeature<"la-global-with-pcrel", "HasLaGlobalWithPcrel", "true",
"Expand la.global as la.pcrel">; "Expand la.global as la.pcrel">;
def HasLaGlobalWithPcrel def HasLaGlobalWithPcrel
: Predicate<"Subtarget->hasLaGlobalWithPcrel()">, : Predicate<"Subtarget->hasLaGlobalWithPcrel()">,
AssemblerPredicate<(all_of LaGlobalWithPcrel), AssemblerPredicate<(all_of LaGlobalWithPcrel),
▲ Show 20 Lines • Show All 82 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/d-arith.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: fadd.s $fs5, $ft7, $fs1 # ASM-AND-OBJ: fadd.s $fs5, $ft7, $fs1
# ASM: encoding: [0xfd,0xe5,0x00,0x01] # ASM: encoding: [0xfd,0xe5,0x00,0x01]
fadd.s $fs5, $ft7, $fs1 fadd.s $fs5, $ft7, $fs1
▲ Show 20 Lines • Show All 83 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/d-bound-check.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: fldgt.s $fa3, $s4, $t1 # ASM-AND-OBJ: fldgt.s $fa3, $s4, $t1
# ASM: encoding: [0x63,0x37,0x74,0x38] # ASM: encoding: [0x63,0x37,0x74,0x38]
fldgt.s $fa3, $s4, $t1 fldgt.s $fa3, $s4, $t1
Show All 15 Lines

llvm/test/MC/LoongArch/Basic/Float/d-branch.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: bceqz $fcc6, 12 # ASM-AND-OBJ: bceqz $fcc6, 12
# ASM: encoding: [0xc0,0x0c,0x00,0x48] # ASM: encoding: [0xc0,0x0c,0x00,0x48]
bceqz $fcc6, 12 bceqz $fcc6, 12

llvm/test/MC/LoongArch/Basic/Float/d-comp.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: fcmp.caf.s $fcc0, $fa0, $fa1 # ASM-AND-OBJ: fcmp.caf.s $fcc0, $fa0, $fa1
# ASM: encoding: [0x00,0x04,0x10,0x0c] # ASM: encoding: [0x00,0x04,0x10,0x0c]
fcmp.caf.s $fcc0, $fa0, $fa1 fcmp.caf.s $fcc0, $fa0, $fa1
▲ Show 20 Lines • Show All 87 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/d-conv.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: frint.s $fa5, $ft9 # ASM-AND-OBJ: frint.s $fa5, $ft9
# ASM: encoding: [0x25,0x46,0x1e,0x01] # ASM: encoding: [0x25,0x46,0x1e,0x01]
frint.s $fa5, $ft9 frint.s $fa5, $ft9
▲ Show 20 Lines • Show All 83 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/d-invalid.s

# RUN: not llvm-mc --triple=loongarch32 -mattr=+d %s 2>&1 | FileCheck %s # RUN: not llvm-mc --triple=loongarch32 %s 2>&1 | FileCheck %s
# CHECK: :[[#@LINE+1]]:1: error: instruction requires the following: LA64 Basic Integer and Privilege Instruction Set # CHECK: :[[#@LINE+1]]:1: error: instruction requires the following: LA64 Basic Integer and Privilege Instruction Set
movgr2fr.d $fa0, $a0 movgr2fr.d $fa0, $a0
# CHECK: :[[#@LINE+1]]:1: error: instruction requires the following: LA64 Basic Integer and Privilege Instruction Set # CHECK: :[[#@LINE+1]]:1: error: instruction requires the following: LA64 Basic Integer and Privilege Instruction Set
movfr2gr.d $a0, $fa0 movfr2gr.d $a0, $fa0

llvm/test/MC/LoongArch/Basic/Float/d-memory.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: fld.s $ft15, $t3, 250 # ASM-AND-OBJ: fld.s $ft15, $t3, 250
# ASM: encoding: [0xf7,0xe9,0x03,0x2b] # ASM: encoding: [0xf7,0xe9,0x03,0x2b]
fld.s $ft15, $t3, 250 fld.s $ft15, $t3, 250
Show All 15 Lines

llvm/test/MC/LoongArch/Basic/Float/d-move.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --show-encoding --defsym=LA64=1 \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding --defsym=LA64=1 \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM,ASM-AND-OBJ64,ASM64 %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM,ASM-AND-OBJ64,ASM64 %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+d --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+d --filetype=obj --defsym=LA64=1 \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj --defsym=LA64=1 \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM-AND-OBJ64 %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM-AND-OBJ64 %s
## Support for the 'D' extension implies support for 'F' ## Support for the 'D' extension implies support for 'F'
# ASM-AND-OBJ: fmov.s $ft5, $ft15 # ASM-AND-OBJ: fmov.s $ft5, $ft15
# ASM: encoding: [0xed,0x96,0x14,0x01] # ASM: encoding: [0xed,0x96,0x14,0x01]
fmov.s $ft5, $ft15 fmov.s $ft5, $ft15
Show All 23 Lines

llvm/test/MC/LoongArch/Basic/Float/f-arith.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: fadd.s $fs5, $ft7, $fs1 # ASM-AND-OBJ: fadd.s $fs5, $ft7, $fs1
# ASM: encoding: [0xfd,0xe5,0x00,0x01] # ASM: encoding: [0xfd,0xe5,0x00,0x01]
fadd.s $fs5, $ft7, $fs1 fadd.s $fs5, $ft7, $fs1
# ASM-AND-OBJ: fsub.s $ft6, $fa6, $fs7 # ASM-AND-OBJ: fsub.s $ft6, $fa6, $fs7
▲ Show 20 Lines • Show All 78 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/f-bound-check.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: fldgt.s $fa3, $s4, $t1 # ASM-AND-OBJ: fldgt.s $fa3, $s4, $t1
# ASM: encoding: [0x63,0x37,0x74,0x38] # ASM: encoding: [0x63,0x37,0x74,0x38]
fldgt.s $fa3, $s4, $t1 fldgt.s $fa3, $s4, $t1
# ASM-AND-OBJ: fldle.s $fs0, $s6, $t5 # ASM-AND-OBJ: fldle.s $fs0, $s6, $t5
Show All 10 Lines

llvm/test/MC/LoongArch/Basic/Float/f-branch.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: bceqz $fcc6, 12 # ASM-AND-OBJ: bceqz $fcc6, 12
# ASM: encoding: [0xc0,0x0c,0x00,0x48] # ASM: encoding: [0xc0,0x0c,0x00,0x48]
bceqz $fcc6, 12 bceqz $fcc6, 12
# ASM-AND-OBJ: bcnez $fcc6, 72 # ASM-AND-OBJ: bcnez $fcc6, 72
# ASM: encoding: [0xc0,0x49,0x00,0x48] # ASM: encoding: [0xc0,0x49,0x00,0x48]
bcnez $fcc6, 72 bcnez $fcc6, 72

llvm/test/MC/LoongArch/Basic/Float/f-comp.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: fcmp.caf.s $fcc0, $fa0, $fa1 # ASM-AND-OBJ: fcmp.caf.s $fcc0, $fa0, $fa1
# ASM: encoding: [0x00,0x04,0x10,0x0c] # ASM: encoding: [0x00,0x04,0x10,0x0c]
fcmp.caf.s $fcc0, $fa0, $fa1 fcmp.caf.s $fcc0, $fa0, $fa1
# ASM-AND-OBJ: fcmp.cun.s $fcc0, $fa0, $fa1 # ASM-AND-OBJ: fcmp.cun.s $fcc0, $fa0, $fa1
▲ Show 20 Lines • Show All 82 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Float/f-conv.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: ffint.s.w $fs6, $fa5 # ASM-AND-OBJ: ffint.s.w $fs6, $fa5
# ASM: encoding: [0xbe,0x10,0x1d,0x01] # ASM: encoding: [0xbe,0x10,0x1d,0x01]
ffint.s.w $fs6, $fa5 ffint.s.w $fs6, $fa5
# ASM-AND-OBJ: ftint.w.s $ft13, $ft5 # ASM-AND-OBJ: ftint.w.s $ft13, $ft5
Show All 22 Lines

llvm/test/MC/LoongArch/Basic/Float/f-invalid.s

  • This file was deleted.
# RUN: not llvm-mc --triple=loongarch32 -mattr=+f %s 2>&1 | FileCheck %s
# CHECK: :[[#@LINE+1]]:1: error: instruction requires the following: 'D' (Double-Precision Floating-Point)
fadd.d $fa0, $fa0, $fa0

llvm/test/MC/LoongArch/Basic/Float/f-memory.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: fld.s $ft15, $t3, 250 # ASM-AND-OBJ: fld.s $ft15, $t3, 250
# ASM: encoding: [0xf7,0xe9,0x03,0x2b] # ASM: encoding: [0xf7,0xe9,0x03,0x2b]
fld.s $ft15, $t3, 250 fld.s $ft15, $t3, 250
# ASM-AND-OBJ: fst.s $fs6, $t7, 230 # ASM-AND-OBJ: fst.s $fs6, $t7, 230
Show All 10 Lines

llvm/test/MC/LoongArch/Basic/Float/f-move.s

# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch32 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --show-encoding \ # RUN: llvm-mc %s --triple=loongarch64 --show-encoding \
# RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s # RUN: | FileCheck --check-prefixes=ASM-AND-OBJ,ASM %s
# RUN: llvm-mc %s --triple=loongarch32 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch32 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# RUN: llvm-mc %s --triple=loongarch64 --mattr=+f --filetype=obj \ # RUN: llvm-mc %s --triple=loongarch64 --filetype=obj \
# RUN: | llvm-objdump -d - \ # RUN: | llvm-objdump -d - \
# RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s # RUN: | FileCheck --check-prefix=ASM-AND-OBJ %s
# ASM-AND-OBJ: fmov.s $ft5, $ft15 # ASM-AND-OBJ: fmov.s $ft5, $ft15
# ASM: encoding: [0xed,0x96,0x14,0x01] # ASM: encoding: [0xed,0x96,0x14,0x01]
fmov.s $ft5, $ft15 fmov.s $ft5, $ft15
# ASM-AND-OBJ: fsel $ft10, $ft12, $ft13, $fcc4 # ASM-AND-OBJ: fsel $ft10, $ft12, $ft13, $fcc4
▲ Show 20 Lines • Show All 58 Lines • Show Last 20 Lines

llvm/test/MC/LoongArch/Basic/Integer/invalid.s

## Test invalid instructions on both loongarch32 and loongarch64 target. ## Test invalid instructions on both loongarch32 and loongarch64 target.
# RUN: not llvm-mc --triple=loongarch32 --mattr=-f %s 2>&1 \ # RUN: not llvm-mc --triple=loongarch32 %s 2>&1 | FileCheck %s --check-prefixes=CHECK,CHECK64
# RUN: | FileCheck %s --check-prefixes=CHECK,CHECK64 # RUN: not llvm-mc --triple=loongarch64 %s 2>&1 --defsym=LA64=1 | FileCheck %s
# RUN: not llvm-mc --triple=loongarch64 --mattr=-f %s 2>&1 --defsym=LA64=1 \
# RUN: | FileCheck %s
## Out of range immediates ## Out of range immediates
## uimm2 ## uimm2
bytepick.w $a0, $a0, $a0, -1 bytepick.w $a0, $a0, $a0, -1
# CHECK: :[[#@LINE-1]]:27: error: immediate must be an integer in the range [0, 3] # CHECK: :[[#@LINE-1]]:27: error: immediate must be an integer in the range [0, 3]
bytepick.w $a0, $a0, $a0, 4 bytepick.w $a0, $a0, $a0, 4
# CHECK: :[[#@LINE-1]]:27: error: immediate must be an integer in the range [0, 3] # CHECK: :[[#@LINE-1]]:27: error: immediate must be an integer in the range [0, 3]
▲ Show 20 Lines • Show All 161 Lines • ▼ Show 20 Lines
and $a0, $a0 and $a0, $a0
# CHECK: :[[#@LINE-1]]:1: error: too few operands for instruction # CHECK: :[[#@LINE-1]]:1: error: too few operands for instruction
andi $a0, $a0 andi $a0, $a0
# CHECK: :[[#@LINE-1]]:1: error: too few operands for instruction # CHECK: :[[#@LINE-1]]:1: error: too few operands for instruction
## Instructions outside the base integer ISA ## Instructions outside the base integer ISA
## TODO: Test instructions in LSX/LASX/LBT/LVZ after their introduction. ## TODO: Test instructions in LSX/LASX/LBT/LVZ after their introduction.
## Floating-Point mnemonics
fadd.s $fa0, $fa0, $fa0
# CHECK: :[[#@LINE-1]]:1: error: instruction requires the following: 'F' (Single-Precision Floating-Point)
fadd.d $fa0, $fa0, $fa0
# CHECK: :[[#@LINE-1]]:1: error: instruction requires the following: 'D' (Double-Precision Floating-Point)
## Using floating point registers when integer registers are expected ## Using floating point registers when integer registers are expected
sll.w $a0, $a0, $fa0 sll.w $a0, $a0, $fa0
# CHECK: :[[#@LINE-1]]:18: error: invalid operand for instruction # CHECK: :[[#@LINE-1]]:18: error: invalid operand for instruction
## msbw < lsbw ## msbw < lsbw
# CHECK: :[[#@LINE+1]]:21: error: msb is less than lsb # CHECK: :[[#@LINE+1]]:21: error: msb is less than lsb
bstrins.w $a0, $a0, 1, 2 bstrins.w $a0, $a0, 1, 2
# CHECK: ^~~~ # CHECK: ^~~~
# CHECK: :[[#@LINE+1]]:22: error: msb is less than lsb # CHECK: :[[#@LINE+1]]:22: error: msb is less than lsb
bstrpick.w $a0, $a0, 30, 31 bstrpick.w $a0, $a0, 30, 31
# CHECK: ^~~~~~ # CHECK: ^~~~~~
llvm/test/MC/LoongArch/Basic/Float/f-comp.s