@@ -16,13 +16,13 @@ you can use this crate to get those intrinsics and solve the linker errors. To
1616do that, add this crate somewhere in the dependency graph of the crate you are
1717building:
1818
19- ``` toml
19+ ``` toml
2020# Cargo.toml
2121[dependencies ]
2222compiler_builtins = { git = " https://github.com/rust-lang/compiler-builtins" }
2323```
2424
25- ``` rust
25+ ``` rust
2626extern crate compiler_builtins;
2727
2828// ...
@@ -91,7 +91,7 @@ obtained with the following:
9191``` sh
9292curl -L -o rustc-llvm-18.0.tar.gz https://github.com/rust-lang/llvm-project/archive/rustc/18.0-2024-02-13.tar.gz
9393tar xzf rustc-llvm-18.0.tar.gz --strip-components 1 llvm-project-rustc-18.0-2024-02-13/compiler-rt
94- ````
94+ ```
9595
9696Local targets may also be tested with ` ./ci/run.sh [target] ` .
9797
@@ -100,9 +100,11 @@ rely on CI.
100100
101101## Progress
102102
103+ - [x] aarch64/chkstk.S
103104- [x] adddf3.c
104105- [x] addsf3.c
105106- [x] arm/adddf3vfp.S
107+ - [x] arm/addsf3.S
106108- [x] arm/addsf3vfp.S
107109- [x] arm/aeabi_dcmp.S
108110- [x] arm/aeabi_fcmp.S
@@ -113,6 +115,7 @@ rely on CI.
113115- [x] arm/aeabi_memset.S
114116- [x] arm/aeabi_uidivmod.S
115117- [x] arm/aeabi_uldivmod.S
118+ - [ ] arm/chkstk.S
116119- [x] arm/divdf3vfp.S
117120- [ ] arm/divmodsi4.S (generic version is done)
118121- [x] arm/divsf3vfp.S
@@ -154,6 +157,12 @@ rely on CI.
154157- [ ] arm/unordsf2vfp.S
155158- [x] ashldi3.c
156159- [x] ashrdi3.c
160+ - [ ] avr/divmodhi4.S
161+ - [ ] avr/divmodqi4.S
162+ - [ ] avr/mulhi3.S
163+ - [ ] avr/mulqi3.S
164+ - [ ] avr/udivmodhi4.S
165+ - [ ] avr/udivmodqi4.S
157166- [x] bswapdi2.c
158167- [x] bswapsi2.c
159168- [x] bswapti2.c
@@ -166,6 +175,7 @@ rely on CI.
166175- [x] divdi3.c
167176- [x] divmoddi4.c
168177- [x] divmodsi4.c
178+ - [x] divmodti4.c
169179- [x] divsf3.c
170180- [x] divsi3.c
171181- [x] extendsfdf2.c
@@ -204,6 +214,8 @@ rely on CI.
204214- [x] mulsf3.c
205215- [x] powidf2.c
206216- [x] powisf2.c
217+ - [ ] riscv/muldi3.S
218+ - [ ] riscv/mulsi3.S
207219- [x] subdf3.c
208220- [x] subsf3.c
209221- [x] truncdfsf2.c
@@ -236,7 +248,8 @@ These builtins are needed to support 128-bit integers.
236248- [x] udivti3.c
237249- [x] umodti3.c
238250
239- These builtins are needed to support ` f16` and ` f128` , which are in the process of being added to Rust.
251+ These builtins are needed to support ` f16 ` and ` f128 ` , which are in the process
252+ of being added to Rust.
240253
241254- [x] addtf3.c
242255- [x] comparetf2.c
@@ -253,8 +266,10 @@ These builtins are needed to support `f16` and `f128`, which are in the process
253266- [x] fixunstfti.c
254267- [ ] floatditf.c
255268- [ ] floatsitf.c
269+ - [ ] floattitf.c
256270- [ ] floatunditf.c
257271- [ ] floatunsitf.c
272+ - [ ] floatuntitf.c
258273- [x] multf3.c
259274- [ ] powitf2.c
260275- [x] subtf3.c
@@ -264,11 +279,42 @@ These builtins are needed to support `f16` and `f128`, which are in the process
264279- [x] trunctfhf2.c
265280- [x] trunctfsf2.c
266281
282+
283+ These builtins are used by the Hexagon DSP
284+
285+ - [ ] hexagon/common_entry_exit_abi1.S
286+ - [ ] hexagon/common_entry_exit_abi2.S
287+ - [ ] hexagon/common_entry_exit_legacy.S
288+ - [x] hexagon/dfaddsub.S~~
289+ - [x] hexagon/dfdiv.S~~
290+ - [x] hexagon/dffma.S~~
291+ - [x] hexagon/dfminmax.S~~
292+ - [x] hexagon/dfmul.S~~
293+ - [x] hexagon/dfsqrt.S~~
294+ - [x] hexagon/divdi3.S~~
295+ - [x] hexagon/divsi3.S~~
296+ - [x] hexagon/fastmath2_dlib_asm.S~~
297+ - [x] hexagon/fastmath2_ldlib_asm.S~~
298+ - [x] hexagon/fastmath_dlib_asm.S~~
299+ - [x] hexagon/memcpy_forward_vp4cp4n2.S~~
300+ - [x] hexagon/memcpy_likely_aligned.S~~
301+ - [x] hexagon/moddi3.S~~
302+ - [x] hexagon/modsi3.S~~
303+ - [x] hexagon/sfdiv_opt.S~~
304+ - [x] hexagon/sfsqrt_opt.S~~
305+ - [x] hexagon/udivdi3.S~~
306+ - [x] hexagon/udivmoddi4.S~~
307+ - [x] hexagon/udivmodsi4.S~~
308+ - [x] hexagon/udivsi3.S~~
309+ - [x] hexagon/umoddi3.S~~
310+ - [x] hexagon/umodsi3.S~~
311+
267312## Unimplemented functions
268313
269314These builtins are for x87 ` f80 ` floating-point numbers that are not supported
270315by Rust.
271316
317+ - ~~ extendxftf2.c~~
272318- ~~ fixunsxfdi.c~~
273319- ~~ fixunsxfsi.c~~
274320- ~~ fixunsxfti.c~~
@@ -288,15 +334,25 @@ numbers.
288334
289335- ~~ ppc/divtc3.c~~
290336- ~~ ppc/fixtfdi.c~~
337+ - ~~ ppc/fixtfti.c~~
291338- ~~ ppc/fixunstfdi.c~~
339+ - ~~ ppc/fixunstfti.c~~
292340- ~~ ppc/floatditf.c~~
341+ - ~~ ppc/floattitf.c~~
293342- ~~ ppc/floatunditf.c~~
294343- ~~ ppc/gcc_qadd.c~~
295344- ~~ ppc/gcc_qdiv.c~~
296345- ~~ ppc/gcc_qmul.c~~
297346- ~~ ppc/gcc_qsub.c~~
298347- ~~ ppc/multc3.c~~
299348
349+ These builtins are for 16-bit brain floating-point numbers that are not
350+ supported by Rust.
351+
352+ - ~~ truncdfbf2.c~~
353+ - ~~ truncsfbf2.c~~
354+ - ~~ trunctfxf2.c~~
355+
300356These builtins involve complex floating-point types that are not supported by
301357Rust.
302358
@@ -342,6 +398,7 @@ These builtins are never called by LLVM.
342398- ~~ ctzdi2.c~~
343399- ~~ ctzsi2.c~~
344400- ~~ ctzti2.c~~
401+ - ~~ ffssi2.c~~
345402- ~~ ffsdi2.c~~ - this is [ called by gcc] [ jemalloc-fail ] though!
346403- ~~ ffsti2.c~~
347404- ~~ mulvdi3.c~~
@@ -404,13 +461,34 @@ Rust only exposes atomic types on platforms that support them, and therefore doe
404461
405462Miscellaneous functionality that is not used by Rust.
406463
464+ - ~~ aarch64/fp_mode.c~~
465+ - ~~ aarch64/lse.S~~ (LSE atomics)
466+ - ~~ aarch64/sme-abi-init.c~~ (matrix extension)
467+ - ~~ aarch64/sme-abi.S~~ (matrix extension)
468+ - ~~ aarch64/sme-libc-routines.c~~ (matrix extension)
407469- ~~ apple_versioning.c~~
470+ - ~~ arm/fp_mode.c~~
471+ - ~~ avr/exit.S~~
408472- ~~ clear_cache.c~~
473+ - ~~ cpu_model/aarch64.c~~
474+ - ~~ cpu_model/x86.c~~
475+ - ~~ crtbegin.c~~
476+ - ~~ crtend.c~~
409477- ~~ emutls.c~~
410478- ~~ enable_execute_stack.c~~
411479- ~~ eprintf.c~~
480+ - ~~ fp_mode.c~~ (float exception handling)
412481- ~~ gcc_personality_v0.c~~
482+ - ~~ i386/fp_mode.c~~
483+ - ~~ int_util.c~~
484+ - ~~ loongarch/fp_mode.c~~
485+ - ~~ os_version_check.c~~
486+ - ~~ riscv/fp_mode.c~~
487+ - ~~ riscv/restore.S~~ (callee-saved registers)
488+ - ~~ riscv/save.S~~ (callee-saved registers)
413489- ~~ trampoline_setup.c~~
490+ - ~~ ve/grow_stack.S~~
491+ - ~~ ve/grow_stack_align.S~~
414492
415493Floating-point implementations of builtins that are only called from soft-float code. It would be better to simply use the generic soft-float versions in this case.
416494
0 commit comments