Commit Graph

826 Commits

Author SHA1 Message Date
Paolo Teti 28c57e7212 Add __ledf2vfp and __lesf2vfp
Now that `73884ae` is in some nightly release We can add ledf2vfp/leds2vfp
and so these two functions be aliased to aeabi_fcmple/aeabi_dcmple on soft-float targets.
2018-02-25 17:50:56 +01:00
bors cfecfa2918 Auto merge of #232 - paoloteti:fix-thumb-ci, r=alexcrichton
Fix CI for thumb* targets

Main fix is inside `utest` (implement Termination to avoid ICE) all the rest are minor fixes due to the new `testcrate`.

Thumb6m needs some extra fixes: some undefined reference and a panic raised removing assembly implementation not in the list.

Still allow failures on TravisCI and keep my `utest` fork for a while. Let's see if is stable.
2018-02-25 12:07:46 +00:00
Paolo Teti 9e99aa2bbe Remove no_std features
Everything default unconditionally to #![no_std].
2018-02-25 09:19:34 +01:00
Paolo Teti 7b81af044e Keep mangled-names as a default feature 2018-02-25 01:11:49 +01:00
Paolo Teti 86612733be testcrate/lib.rs: use #![no_std] unconditionally 2018-02-25 00:22:39 +01:00
Paolo Teti 8289e626fe Use upstream for utest
Fix has been merged into upstream.
2018-02-24 22:48:27 +01:00
Paolo Teti 7f92a8ee7b Fix CI for thumb* targets
Main fix is inside 'utest' all the rest are minor fixes due to the
new 'testcrate'.

Still allow failures on TravisCI for a while.
2018-02-18 18:15:57 +01:00
Paolo Teti 280012b1df Fix thumbv6m build (feature=c)
1. Avoid undefined references as:
   undefined reference to `__modsi3'
   undefined reference to `__umodsi3'

2. We can't remove assembly implementations that are not in the list
2018-02-16 17:34:01 +01:00
bors 266ea0740a Auto merge of #231 - paoloteti:vfp, r=alexcrichton
Collection of VFP intrinsics

Nothing really exciting here, just a list of trivial VFP intrinsics.

First of all set  `mfloat-abi=hard` not only for thumb targets, then add support for the following intrinsics:

```
 __gesf2vfp
 __gedf2vfp
 __gtsf2vfp
 __gtdf2vfp
 __ltsf2vfp
 __ltdf2vfp
 __nesf2vfp
 __nedf2vfp
 __eqsf2vfp
 __eqdf2vfp
 __extendsfdf2vfp
```
Resulting implementation is really trivial thanks to native code generated by LLVM on hard-float targets
2018-02-12 18:57:43 +00:00
Alex Crichton 7c11f5c018 Merge branch 'llvm-60' 2018-02-12 10:57:11 -08:00
Paolo Teti 2467bd2cbf Update Progress list
Already implemented, but not marked as done.
2018-02-12 18:23:55 +01:00
Paolo Teti e7bf03019c Collection of VFP intrinsics
Nothing really exciting here. LLVM on hard-float target use native instructions
for all listed VFP intrinsics and so resulting implementation is really trivial.

Implemented intrinsics:

 __gesf2vfp
 __gedf2vfp
 __gtsf2vfp
 __gtdf2vfp
 __ltsf2vfp
 __ltdf2vfp
 __nesf2vfp
 __nedf2vfp
 __eqsf2vfp
 __eqdf2vfp
 __extendsfdf2vfp
2018-02-11 23:46:56 +01:00
Paolo Teti e15ab79173 Set -mfloat-abi=hard not only for thumb targets 2018-02-11 21:23:59 +01:00
bors 63de7c62c1 Auto merge of #230 - paoloteti:fix-extend, r=alexcrichton
Fix issue extending f32::MIN/MAX to f64 and improve testcrate.

I was able to trigger an issue extending f32::MAX or f32::MIN to a f64.
Issue was **not** triggered by `testcrate` mainly because f32::MAX/MIN are
not in the list of special values to generate.

This PR fix the issue and improve `testcrate` adding MAX/MIN/MIN_POSITIVE
in the list of special values.
2018-02-09 20:50:27 +00:00
Paolo Teti 2cb290afa3 Fix __subsf3 and __subdf3 on x86
Be sure to do not mix hard-float and soft-float calls.
Disassembled code show exactly this.
So replace add with an explicit call to __addsf3/__adddf3

This seems the root cause of some sporadic failures.
2018-02-09 20:35:55 +01:00
Paolo Teti 306764b091 Fix issue extending f32::MIN/MAX to f64 and improve testcrate.
I was able to trigger an issue extending f32::MAX or f32::MIN to a f64.
Issue was not triggered by `testcrate` mainly because f32::MAX/MIN are
not in the list of special values to generate.

This PR fix the issue and improve `testcrate` adding MAX/MIN/MIN_POSITIVE
in the list of special values.
2018-02-09 17:23:16 +01:00
bors 2801c99793 Auto merge of #229 - paoloteti:extend, r=alexcrichton
Add generic conversion from a narrower to a wider FP type

Add `extend` module to implement conversion from a narrower to a wider floating-point type.

This implementation is only intended to support *widening* operations.
Module to convert a *narrower* floating-point will be added in the future.
2018-02-08 21:55:46 +00:00
Paolo Teti 45f924ece5 Add generic conversion from a narrower to a wider FP type
Add `extend` module to implement conversion from a narrower to a wider
floating-point type.

This implementation is only intended to support *widening* operations.
Module to convert a *narrower* floating-point will be added in the future.
2018-02-08 18:20:45 +01:00
bors 3e7fba7669 Auto merge of #228 - paoloteti:eabihf, r=alexcrichton
Add support for sub*f3vfp and add*f3vfp

As done before for `mul` and `div` let's use extern "C" to generate `"aapcs"` or `"aapcs-vfp"` depending on target configuration.
2018-02-05 05:18:36 +00:00
Paolo Teti 13db8bf436 Add support for sub*f3vfp and add*f3vfp
As done before for mul and div let's use extern "C" to generate `"aapcs"`
or `"aapcs-vfp"` depending on target configuration.
2018-02-04 18:34:31 +01:00
Alex Crichton 2d8f137801
Merge pull request #227 from paoloteti/softfloat
Fix __aeabi_fcmple and add test cases
2018-02-04 08:25:47 -06:00
Paolo Teti 73884ae671 Fix __aeabi_fcmple and add test cases
`le` in __aeabi_fcmple means `less or equal`
2018-02-04 10:56:02 +01:00
Alex Crichton 1129b62ca2
Merge pull request #226 from alexcrichton/refactor-tests
Simplify how testing is done
2018-01-31 14:22:51 -06:00
Alex Crichton 2a13475197 Simplify how testing is done
All tests are moved to a separate crate in this repository to enable features by
default. Additionally the test generation is moved to a seprate build script and
simplified to reduce the amount of boilerplate needed per test.

Overall this should still be testing everything, just in a different location!
2018-01-31 11:32:20 -08:00
Alex Crichton 032424971d Allow thumb failures
They've been broken for some time anyway
2018-01-31 07:26:44 -08:00
Alex Crichton 93977e65a0
Merge pull request #225 from paoloteti/armhf
Add support for mul[s/d]f3vfp and div[s/d]f3vfp
2018-01-31 09:25:51 -06:00
Paolo Teti 1269ad1438 Fix build errors on arm*-linux-gnueabi
I removed std::mem, but We need it.
2018-01-30 23:10:56 +01:00
Paolo Teti 9e04723812 Add test cases for Mul[s/d]f3vfp and Div[s/d]f3vfp
for ARM targets only
2018-01-30 20:15:37 +01:00
Paolo Teti f42b9e0bad __[mul/div]sf3vfp and __[mul/div]df3vfp only on ARM 2018-01-30 18:48:20 +01:00
Paolo Teti 8f018562ca Add support for mul[s/d]f3vfp and div[s/d]f3vfp
Here using `"C"` the compiler will use `"aapcs"` or `"aapcs-vfp"`
depending on target configuration.

Of course this translates in a call to `__aeabi_fdiv` / `__aeabi_fmul`
on non-HF targets.

On `eabi` targets with +vfpv2/vfpv3 LLVM generate:

   vmov	s0, r1
   vmov	s2, r0
   vdiv.f32	s0, s2, s0
   vmov	r0, s0
   bx	lr

On `eabihf` targets with +vfpv3-d16/d32/f32 +fp-only-sp LLVM generate:

  vdiv.f32	s0, s0, s1
  bx	lr

That's exactly what We need for [div/mul][s/d]f3vfp.S
2018-01-29 20:49:55 +01:00
Alex Crichton 345447948f Update to release_60 of LLVM's compiler-rt 2018-01-24 21:49:11 -08:00
Alex Crichton 0a95675bab
Merge pull request #224 from alexcrichton/fix-float-cmp
Fix some typos/bugs with float comparison intrinsics
2018-01-24 20:26:10 -06:00
Alex Crichton 7d5fb94560 Fix some typos/bugs with float comparison intrinsics
* I believe `__gtdf2` erroneously used `f32` instead of `f64`
* Most of these needed `#[arm_aeabi_alias]` to ensure they're correctly called
  through the alias
* Some existing aliases were corrected with the right names
2018-01-24 17:01:04 -08:00
Alex Crichton 22f76e35fa
Merge pull request #223 from paoloteti/be
Skip aeabi_cdcmp and aeabi_cfcmp on big-endian targets
2018-01-19 13:58:10 -06:00
Paolo Teti 083f107c5e Fix build issue and comments
Fix an unclosed if block
2018-01-19 19:27:25 +01:00
Paolo Teti a75a49990c Skip aeabi_cdcmp and aeabi_cfcmp on big-endian targets
First of all aeabi_cdcmp and aeabi_cfcmp are never called LLVM.
Second are little-endian only, so build fail on big-endian targets.
Temporally workaround: exclude these files for big-endian targets.
2018-01-19 09:19:05 +01:00
Alex Crichton bf912e607e
Merge pull request #222 from whitequark/master
Implement comparesf2/comparedf2 intrinsics
2018-01-13 17:27:55 -06:00
whitequark 205322baf0 comparesf2: do not build the assembly versions of intrinsics on ARM. 2018-01-12 21:33:24 +00:00
whitequark 07b446a9ab comparesf2/comparedf2: fix a signedness bug and add tests. 2017-12-29 07:58:34 +00:00
whitequark 4cf6571362 README: update to explain how to run quickcheck-style tests. 2017-12-29 06:36:09 +00:00
whitequark 3d1600b654 comparesf2/comparedf2: do not build the C versions of intrinsics. 2017-12-29 06:16:35 +00:00
whitequark 6adcc9e9cb comparesf2/comparedf2: use i32 instead of bool for return type.
Note that this changes semantics:
    pub extern "C" fn __eqsf2(a: f32, b: f32) -> bool {
        cmp(a, b).to_le_abi() != 0
    }

is not the same as

    pub extern "C" fn __eqsf2(a: f32, b: f32) -> i32 {
        cmp(a, b).to_le_abi()
    }

However, compiler-rt does the latter, so this is actually
an improvement.
2017-12-29 06:14:51 +00:00
whitequark f5d86476a8 Implement comparesf2/comparedf2 intrinsics. 2017-12-28 03:45:32 +00:00
whitequark 82ee1dd760 comparesf2/comparedf2 are called on soft-float targets.
E.g. take a look at the assembly output for:

  pub fn comparesf2(a: f32, b: f32) -> bool { a > b }
  pub fn comparedf2(a: f64, b: f64) -> bool { a > b }

which will include calls to __gtsf2 and __gtdf2.
2017-12-28 02:57:42 +00:00
bors 0ba07e4926 Auto merge of #220 - malbarbo:armv4t, r=alexcrichton
Use kernel user helpers for armv4t

Rename armv5te cfg to kernel_user_helpers and enable it for armv4t.
2017-12-26 20:21:11 +00:00
Alex Crichton cc9a70f61c Work around japaric/xargo#186 temporarily 2017-12-26 12:20:17 -08:00
Alex Crichton a076c9aaa7 Disable incremental + LTO
This causes errors in more recent versions of rustc
2017-12-26 10:14:11 -08:00
Marco A L Barbosa f4edcf7b24 Use kernel user helpers for armv4t
Rename armv5te cfg to kernel_user_helpers and enable it for armv4t
2017-12-26 13:02:05 -02:00
bors 3a9ba44406 Auto merge of #219 - Amanieu:sync_val_compare_and_swap, r=alexcrichton
Fix argument order for __sync_val_compare_and_swap

Fixes https://github.com/rust-lang/rust/issues/46822#issuecomment-353698138
2017-12-23 15:36:04 +00:00
Amanieu d'Antras e8ea848785 Fix argument order for __sync_val_compare_and_swap 2017-12-23 08:45:25 +01:00