87 Commits

Author SHA1 Message Date
Timm Baeder
adba24aa3c
[clang][bytecode] Add missing __builtin_memcpy checks (#135975)
Add a test for type punning and tests and the necessary checks for
non-trivially-copyable types and incomplete types.
2025-04-17 06:19:37 +02:00
Timm Baeder
ab7e0c0fc0
[clang][bytecode] Implement __builtin_wmem{cpy,move} (#135969) 2025-04-16 17:08:10 +02:00
Timm Baeder
2d63faead4
[clang][bytecode][NFC] Remove PT_FnPtr (#135947)
We don't need this anymore since we don't return it from classify()
anymore.
2025-04-16 13:21:25 +02:00
Timm Baeder
98ea512f72
[clang][bytecode] Clear inactive union fields when copying (#134982)
When copying unions, we need to only copy the active field of the source
union, which we were already doing. However, we also need to zero out
the (now) inactive fields, so we don't end up with dangling pointers in
those inactive fields.
2025-04-10 06:12:00 +02:00
Timm Baeder
a0e1e680d2
[clang][bytecode] Return Invalid() on non-constexpr builtins (#133700)
So the diagnostic output matches with the current interpreter
2025-03-31 18:53:12 +02:00
Timm Baeder
0bc2c5b2a4
Reapply "[clang][bytecode] Implement __builtin_{wcscmp,wcsncmp} (#132… (#132963)
…723)"

This reverts commit 1e2ad6793ac205607e7c809283cf69e1cc36a69a.


Fix the previous commit on big-endian hosts by _not_ falling through to
the `uint8_t` code path.
2025-03-26 08:19:31 +01:00
Timm Baeder
bcedb368e3
[clang][bytecode] Support composite arrays in memcpy op (#132775)
See the attached test case.
2025-03-25 07:17:10 +01:00
Timm Bäder
1e2ad6793a Revert "[clang][bytecode] Implement __builtin_{wcscmp,wcsncmp} (#132723)"
This reverts commit f7aea4d081f77dba48b0fc019f59b678fb679aa8.

This broke the clang-solaris11-sparcv9 builder:
https://lab.llvm.org/buildbot/#/builders/13/builds/6151
2025-03-25 07:15:30 +01:00
Timm Baeder
f7aea4d081
[clang][bytecode] Implement __builtin_{wcscmp,wcsncmp} (#132723) 2025-03-24 15:03:49 +01:00
Timm Baeder
9ab3b6a006
[clang][bytecode] Diagnose integral source/dest in memcpy (#132715)
Like the current interpreter does.
2025-03-24 12:44:35 +01:00
Timm Baeder
c7f14f601f
[clang][bytecode] Implement __builtin_wcschr (#132708)
This is already almost implemented, just need to enable support for it.
2025-03-24 12:28:23 +01:00
Timm Baeder
849e5ea94f
[clang][bytecode] Add Descriptor::getDataType() (#132681)
This returns the type of data in the Block, which might be different
than the type of the expression or declaration we created the block for.
This lets us remove some special cases from CheckNewDeleteForms() and
CheckNewTypeMismatch().
2025-03-24 09:47:52 +01:00
Timm Baeder
5fc891b965
[clang][bytecode][NFC] Use getElemType() in __builtin_memchr as well (#132550)
For consistency.
2025-03-22 16:34:22 +01:00
Timm Baeder
c51d396f4d
[clang][bytecode] Fix __builtin_memmove type diagnostics (#132544)
Set the source type when allocating primitives so we can later retrieve
it.
2025-03-22 14:58:32 +01:00
Timm Baeder
6c8e9a6192
[clang][bytecode][NFC] Add assert to ptrauth_string_discriminator (#132527)
As pointed out by @shafik, this confuses static analysis and most
probably humans as well. Add an assertion to ensure the given array has
at least one element.
2025-03-22 08:05:28 +01:00
Timm Baeder
7492666482
[clang][bytecode] Implement __builtin_wmemchr (#132254) 2025-03-20 21:01:14 +01:00
Timm Baeder
cfa07ccdfc
[clang][bytecode] Fix builtin_memchr with non-0 start index (#131633) 2025-03-17 19:02:55 +01:00
Timm Baeder
cf6a520a7a
[clang][bytecode] Fix builtin_memcmp buffer sizes for pointers (#130570)
Don't use the pointer size, but the number of elements multiplied by the
element size.
2025-03-10 15:51:31 +01:00
Timm Baeder
0f732481ac
[clang][bytecode] Fix getting pointer element type in __builtin_memcmp (#130485)
When such a pointer is heap allocated, the type we get is a pointer
type. Take the pointee type in that case.
2025-03-09 12:57:42 +01:00
Timm Baeder
227f9544a4
[clang][bytecode][NFC] Bail out on non constant evaluated builtins (#130431)
If the ASTContext says so, don't bother trying to constant evaluate the
given builtin.
2025-03-09 06:35:19 +01:00
Timm Baeder
46d218d1af
[clang][bytecode] Implement __builtin_{memchr,strchr,char_memchr} (#130420)
llvm has recently started to use `__builitn_memchr` at compile time, so
implement this. Still needs some work but the basics are done.
2025-03-08 16:52:06 +01:00
Timm Baeder
d08cf7900d
[clang][bytecode] Implement __builtin_constant_p (#130143)
Use the regular code paths for interpreting.

Add new instructions: `StartSpeculation` will reset the diagnostics
pointers to `nullptr`, which will keep us from reporting any diagnostics
during speculation. `EndSpeculation` will undo this.

The rest depends on what `Emitter` we use.

For `EvalEmitter`, we have no bytecode, so we implement `speculate()` by
simply visiting the first argument of `__builtin_constant_p`. If the
evaluation fails, we push a `0` on the stack, otherwise a `1`.

For `ByteCodeEmitter`, add another instrucion called `BCP`, that
interprets all the instructions following it until the next
`EndSpeculation` instruction. If any of those instructions fails, we
jump to the `EndLabel`, which brings us right before the
`EndSpeculation`. We then push the result on the stack.
2025-03-08 06:06:14 +01:00
Timm Baeder
e9fe95acf3
[clang][bytecode][NFC] Use maxnum/minnum for fmax/fmin (#129643)
Equivalent of https://github.com/llvm/llvm-project/pull/129630 for the
bytecode interpreter.
2025-03-04 07:22:54 +01:00
Timm Baeder
e8d848824b
[clang][bytecode] Fix dynamic array allocation return values (#127387)
We need to return a pointer to the first element, not the array itself.
2025-02-16 12:24:35 +01:00
Timm Baeder
20506a0a15
[clang][bytecode] Fix operator new source expression (#126870)
... for composite element types. Looks like I forgot this in
e6030d389571b3f1b0f0c5a35b7fa45937ed0f6c
2025-02-12 10:05:49 +01:00
Timm Baeder
51c7338cc6
[clang][bytecode] Fix dummy handling for p2280r4 (#124396)
This makes some other problems show up like the fact that we didn't
suppress diagnostics during __builtin_constant_p evaluation.
2025-01-29 09:32:35 +01:00
Timm Baeder
e6030d3895
[clang][bytecode] Use std::allocator calls for Descriptor source (#123900)
... for the dynamic blocks created for operator new calls. This way we
get the type of memory allocated right. As a side-effect, the
diagnostics now point to the std::allocator calls, which is an
improvement.
2025-01-24 11:50:56 +01:00
Chandler Carruth
dc3cd2e95e
Factor common code for quoting a builtin name (#120835)
This shows up in several places in order to match the quoting of other
uses of the same diagnostic. Handling it centrally simplifies the code
and reduces changes if the storage for builtin names changes.

This refactoring is extracted out of #120534 as requested in code
review.
2025-01-03 19:23:42 -08:00
Timm Baeder
02b30128e8
[clang][bytecode] Always return false for invalid bcp results (#121467) 2025-01-02 13:44:45 +01:00
Timm Baeder
34097c07e1
[clang][bytecode] Consider unknown-size arrays in memcpy/memcmp (#121462)
When emitting diagnostics for the number of elements.
2025-01-02 10:59:08 +01:00
Timm Baeder
45e874e390
[clang][bytecode] Check for memcpy/memmove dummy pointers earlier (#121453) 2025-01-02 09:15:14 +01:00
Timm Baeder
f0d60170cc
[clang][bytecode] Check memove/memcpy for available elements (#121383)
Both destination and source pointer need to have at least as many
elements as requested.
2024-12-31 12:14:27 +01:00
Timm Baeder
1f2d934525
[clang][bytecode] Support pointers in __builtin_mem{move,cpy} (#120560)
Unfortunately, that means we can't use the __builtin_bit_cast
implementation for this.
2024-12-19 16:38:58 +01:00
Timm Baeder
dd6f6a096a
[clang][bytecode] Handle builtin_wmemcmp (#120070) 2024-12-16 13:41:14 +01:00
Timm Baeder
30cbd09f4b
[clang][bytecode] Fix memcmp/bcmp failures on big-endian hosts (#119851)
See the discussion in

https://github.com/llvm/llvm-project/pull/119678#issuecomment-2539680746
and
https://github.com/llvm/llvm-project/pull/119544#issuecomment-2539678561
2024-12-13 12:19:27 +01:00
Timm Baeder
98470c0b2e
[clang][bytecode] Handle __builtin_bcmp (#119678)
... the same as `__builtin_memcmp`. Also fix a bug we still had when we
couldn't find a difference in the two inputs after `Size` bytes.
2024-12-12 10:57:39 +01:00
Timm Bäder
cfad8f14f8 [clang][bytecode] Fix a build failure on aarch64
This broke e.g.
https://lab.llvm.org/buildbot/#/builders/190/builds/11216
2024-12-12 09:08:30 +01:00
Timm Baeder
8713914d76
[clang][bytecode] Handle __builtin_memcmp (#119544) 2024-12-12 08:59:35 +01:00
Timm Baeder
d416cae180
[clang][bytecode][NFC] Use Pointer::pointToSameBlock (#119552)
block() requires the pointer to be a block pointer.
2024-12-11 16:14:19 +01:00
Timm Baeder
4dde52d76b
[clang][bytecode] Check for overlapping memcpy regions (#119535) 2024-12-11 12:06:56 +01:00
Timm Baeder
6168739f00
[clang][bytecode] Reject memcpy sizes with element size remainder (#119209) 2024-12-09 16:57:49 +01:00
Timm Baeder
d17e51f306
[clang][bytecode] Handle __builtin_strncmp (#119208) 2024-12-09 16:17:39 +01:00
Jie Fu
4f08fa1a77 [clang] Fix -Wunused-variable in InterpBuiltin.cpp (NFC)
/llvm-project/clang/lib/AST/ByteCode/InterpBuiltin.cpp:262:23:
error: unused variable 'AC' [-Werror,-Wunused-variable]
    const ASTContext &AC = S.getASTContext();
                      ^
1 error generated.
2024-12-09 21:23:08 +08:00
Timm Baeder
00b50c917b
[clang][bytecode] Handle __builtin_wcslen (#119187)
Handle different char widths in builtin_strlen.
2024-12-09 12:34:25 +01:00
Timm Baeder
abc27039be
[clang][bytecode] Pass __builtin_memcpy size along (#118649)
To DoBitCastPtr, so we know how many bytes we want to read.
2024-12-05 06:55:18 +01:00
Timm Bäder
61c2ac03d8 Revert "[clang][bytecode] Handle __builtin_wcslen (#118446)"
This reverts commit 89a0ee89973c3d213c4bc11c26b41eab67e06da0.

This breaks builders:
https://lab.llvm.org/buildbot/#/builders/13/builds/3885
2024-12-03 13:35:57 +01:00
Timm Baeder
3f39c5df08
[clang][bytecode] Reject memcpy dummy pointers after null check (#118460)
To match the diagnostic output of the current interpreter.
2024-12-03 11:51:49 +01:00
Timm Baeder
0f4dc4276f
[clang][bytecode] Initialize elements in __builtin_elementwise_popcount (#118457) 2024-12-03 11:14:06 +01:00
Timm Baeder
89a0ee8997
[clang][bytecode] Handle __builtin_wcslen (#118446)
... just like strlen.
2024-12-03 10:20:30 +01:00
Timm Baeder
46446bb2d3
[clang][bytecode][NFC] Diagnose non-constexpr builtin strcmp calls (#118442) 2024-12-03 09:49:26 +01:00