Commit Graph

702 Commits

Author SHA1 Message Date
Ralf Jung
b0cf4c28ea turns out Layout has some more things to worry about -- move ABI comparison into helper function
like is_bool, and some special magic extra fields
2023-09-08 09:14:07 +02:00
Ralf Jung
c3e14edd8b accept some differences for rustc_abi(assert_eq), so that we can test more things to be compatible 2023-09-08 08:59:55 +02:00
Ralf Jung
fa5f13775a interpret: make MemPlace, Place, Operand types private to the interpreter 2023-09-04 17:53:38 +02:00
Ralf Jung
efc759238d miri ABI check: fix handling of 1-ZST; don't accept sign differences 2023-08-31 17:27:23 +02:00
Ralf Jung
c37bd09d88 miri function ABI check: specifically look for repr(transparent) 2023-08-30 17:07:25 +02:00
Ralf Jung
1e95aa0c49 interpret: make sure we accept transparent newtypes as ABI-compatible
also we were missing the case for Vector arguments, so handle those as well
2023-08-30 17:04:54 +02:00
Ralf Jung
6d1ce9bd13 storage_live: avoid computing the layout unless necessary 2023-08-30 13:46:55 +02:00
Ralf Jung
f87e91de7d unify passing of sized and unsized function arguments :-) 2023-08-30 13:46:55 +02:00
Ralf Jung
7cdeff266c a bit of meta-related cleanup on Projectable 2023-08-30 13:46:55 +02:00
Ralf Jung
a09df43d9f move marking-locals-live out of push_stack_frame, so it happens with argument passing
this entirely avoids even creating unsized locals in Immediate::Uninitialized state
2023-08-30 13:46:54 +02:00
Ralf Jung
bdd5855b8e interpret: fix projecting into an unsized field of a local
new invariant: Place::Local never refers to something unsized
2023-08-30 13:43:34 +02:00
Matthias Krüger
23f86255ef
Rollup merge of #115272 - RalfJung:miri-error-print, r=saethlin
miri/diagnostics: don't forget to print_backtrace when ICEing on unexpected errors

This should fix the missing output encountered [here](https://github.com/rust-lang/rust/issues/115145#issuecomment-1694334410).

r? `@saethlin`
2023-08-30 07:18:11 +02:00
Ralf Jung
b2ebf1c23f const_eval and codegen: audit uses of is_zst 2023-08-29 09:03:46 +02:00
bors
191dc54dbf Auto merge of #115182 - RalfJung:abi-compat-sign, r=b-naber
miri ABI compatibility check: accept u32 and i32

If only the sign differs, then surely these types are compatible. (We do still check that `arg_ext` is the same, just in case.)

Also I made it so that the ABI check must *imply* that size and alignment are the same, but it doesn't actively check that itself. With how crazy ABI constraints get, having equal size and align really shouldn't be used as a signal for anything I think...
2023-08-28 22:56:10 +00:00
Ralf Jung
beeb2b13cc miri/diagnostics: don't forget to print_backtrace when ICEing on unexpected errors
then also use the new helper in a few other places
2023-08-27 15:42:25 +02:00
bors
296c7a683c Auto merge of #115184 - saethlin:local-allocated-spans, r=RalfJung
Record allocation spans inside force_allocation

This expands https://github.com/rust-lang/miri/pull/2940 to cover locals

r? `@RalfJung`
2023-08-25 17:03:33 +00:00
Ben Kimock
8ecdefb3db
Add a doc comment for the new hook
Co-authored-by: Ralf Jung <post@ralfj.de>
2023-08-25 11:58:31 -04:00
Ben Kimock
ec21d584ee Record allocation spans inside force_allocation 2023-08-25 11:16:52 -04:00
Ralf Jung
5194060ded miri ABI compatibility check: accept u32 and i32 2023-08-24 21:02:21 +02:00
Ralf Jung
4c53783f3c when terminating during unwinding, show the reason why 2023-08-24 13:28:26 +02:00
Ralf Jung
ac3bca24b7 interpret: have assert_* intrinsics call the panic machinery instead of a direct abort 2023-08-20 15:52:40 +02:00
Ralf Jung
788fd44a3b interpret/miri: call panic_cannot_unwind lang item instead of hard-coding the same message 2023-08-20 15:52:40 +02:00
Ralf Jung
818ec8e23a give some unwind-related terminators a more clear name 2023-08-20 15:52:38 +02:00
Matthias Krüger
5c5ae6c5f3
Rollup merge of #114615 - RalfJung:interpret-invalid-where, r=lcnr
interpret: remove incomplete protection against invalid where clauses

Cc https://github.com/rust-lang/rust/issues/97477, https://github.com/rust-lang/project-const-generics/issues/37

r? ``@lcnr``
2023-08-08 21:44:45 +02:00
Ralf Jung
a7132bf387 interpret: remove incomplete protection against invalid where clauses 2023-08-08 10:35:22 +02:00
bors
6d55184d05 Auto merge of #114520 - RalfJung:unsized-valtrees, r=oli-obk
simplify handling of valtrees for unsized types
2023-08-08 07:48:01 +00:00
Matthias Krüger
cbe2522652
Rollup merge of #114382 - scottmcm:compare-bytes-intrinsic, r=cjgillot
Add a new `compare_bytes` intrinsic instead of calling `memcmp` directly

As discussed in #113435, this lets the backends be the place that can have the "don't call the function if n == 0" logic, if it's needed for the target.  (I didn't actually *add* those checks, though, since as I understood it we didn't actually need them on known targets?)

Doing this also let me make it `const` (unstable), which I don't think `extern "C" fn memcmp` can be.

cc `@RalfJung` `@Amanieu`
2023-08-07 05:29:12 +02:00
Scott McMurray
502af03445 Add a new compare_bytes intrinsic instead of calling memcmp directly 2023-08-06 15:47:40 -07:00
Matthias Krüger
74dce18639
Rollup merge of #114542 - RalfJung:const-prop-nonsense, r=compiler-errors
interpret: use ConstPropNonsense for more const-prop induced issues
2023-08-07 00:06:06 +02:00
Ralf Jung
997ec63fb1 simplify handling of valtrees for unsized types 2023-08-06 21:25:49 +02:00
Matthias Krüger
13de583583
Rollup merge of #114505 - ouz-a:cleanup_mir, r=RalfJung
Add documentation to has_deref

Documentation of `has_deref` needed some polish to be more clear about where it should be used and what's it's purpose.

cc https://github.com/rust-lang/rust/issues/114401

r? `@RalfJung`
2023-08-06 17:26:29 +02:00
ouz-a
6df546281b cleanup misinformation regarding has_deref 2023-08-06 17:29:09 +03:00
Ralf Jung
efd54ccf5a interpret: use ConstPropNonsense for more const-prop induced issues 2023-08-06 15:20:03 +02:00
Ralf Jung
3345077b42 interpret: add mplace_to_ref helper method 2023-08-04 15:00:57 +02:00
Matthias Krüger
00dcc7b97c
Rollup merge of #114372 - RalfJung:const-pointer-as-int, r=oli-obk
const validation: point at where we found a pointer but expected an integer

Instead of validation just printing "unable to turn pointer into bytes", make this a regular validation error that says where in the value the bad pointer was found. Also distinguish "expected integer, got pointer" from "expected pointer, got partial pointer or mix of pointers".

To avoid duplicating things too much I refactored the diagnostics for validity a bit, so that "got uninit, expected X" and "got pointer, expected X" can share the "X" part. Also all the errors emitted for validation are now grouped under `const_eval_validation` so that they are in a single group in the ftl file.

r? `@oli-obk`
2023-08-03 17:29:08 +02:00
Matthias Krüger
01fdb9d148
Rollup merge of #114363 - RalfJung:interpret-not-miri, r=jackh726
avoid 'miri' when refering to the shared interpreter

This is basically the rustc source code version of https://github.com/rust-lang/rustc-dev-guide/pull/1471.
2023-08-03 08:12:39 +02:00
Ralf Jung
7767cbb3b0 const validation: point at where we found a pointer but expected an integer 2023-08-02 18:51:50 +02:00
Ralf Jung
2984670cea avoid 'miri' when refering to the shared interpreter 2023-08-02 16:52:03 +02:00
bors
64ad036307 Auto merge of #114333 - RalfJung:dangling-ptr-offset, r=oli-obk
Miri: fix error on dangling pointer inbounds offset

We used to claim that the pointer was "dereferenced", but that is just not true.

Can be reviewed commit-by-commit. The first commit is an unrelated rename that didn't seem worth splitting into its own PR.

r? `@oli-obk`
2023-08-02 09:12:32 +00:00
Ralf Jung
8496292dda properly track why we checked whether a pointer is in-bounds
also simplify the in-bounds checking in Miri's borrow trackers
2023-08-01 17:57:13 +02:00
Ralf Jung
7d5886504c rename deref_operand → deref_pointer and some Miri helper functions 2023-08-01 13:40:29 +02:00
Ralf Jung
b169ee7c1a fix alignment handling for Repeat expressions 2023-07-31 19:22:14 +02:00
Deadbeef
e6b423aebb Remove constness from ParamEnv 2023-07-27 15:50:42 +00:00
Matthias Krüger
fa21a8c6f8
Rollup merge of #114075 - matthiaskrgr:fmt_args_rustc_3, r=wesleywiser
inline format!() args from rustc_codegen_llvm to the end (4)

r? `@WaffleLapkin`
2023-07-27 06:04:13 +02:00
Matthias Krüger
c64ef5e070 inline format!() args from rustc_codegen_llvm to the end (4)
r? @WaffleLapkin
2023-07-25 23:20:28 +02:00
Ralf Jung
da3f0d0eb7 make MPlaceTy non-Copy 2023-07-25 22:35:07 +02:00
Ralf Jung
77ff1b83cd interpret: make read functions generic over operand type 2023-07-25 22:33:59 +02:00
Ralf Jung
00fb45dccd interpret: make write functions generic over the place type 2023-07-25 22:33:58 +02:00
Ralf Jung
d127600511 add some sanity checks in write_immediate_no_validate 2023-07-25 14:30:58 +02:00
Ralf Jung
14a674c88c interpret: read_discriminant: only return VariantIdx 2023-07-25 14:30:58 +02:00