gingerBill
2992ca5df1
Basic support for new procedure code (non-polymorphic, non-proc-group)
2023-06-15 01:37:55 +01:00
gingerBill
151396507e
Merge pull request #1309 from Yawning/fix/freebsd-syscall
...
src: "Fix" the system call intrinsic for FreeBSD
2023-06-07 12:04:57 +01:00
gingerBill
ca6cef9a7d
Fix wasm intrinsics; fix len for strings and slices
2023-06-06 23:11:42 +01:00
gingerBill
4a75a1e839
Merge branch 'master' into separate-int-word-sizes
2023-06-06 22:42:04 +01:00
gingerBill
8bf32ac697
Minor change to handling of propagation of errors with --- as a value
2023-05-22 12:53:29 +01:00
gingerBill
8693a045bb
Revert "Change intrinsics.read_cycle_counter on ARM64"
...
This reverts commit b567679eb6
2023-05-12 11:50:49 +01:00
gingerBill
b567679eb6
Change intrinsics.read_cycle_counter on ARM64
2023-05-12 11:34:01 +01:00
gingerBill
685f7d0fea
Rename word_size to ptr_size internally to make it clearer
2023-04-20 12:18:13 +01:00
gingerBill
02a58c1247
Merge pull request #2464 from ap29600/simd_unaligned_load
...
Improve code generation for `intrinsics.unaligned_load/store` on `#simd` types
2023-04-20 10:33:32 +01:00
gingerBill
72a7b35513
Merge pull request #2450 from destroycomputers/master
...
Fix intrinsics.alloca code generation
2023-04-20 10:25:43 +01:00
Andrea Piseri
af63eff8d7
improve code generation for intrinsics.unaligned_load/store on #simd types
...
the default implementation calls memcpy on an `alloca` constant, which
seems to heavily confuse the optimizer and produces overall suboptimal
code.
Introducing this specialization simplifies the intermediate
representation produced, resulting in more efficient code.
2023-04-16 15:01:30 +02:00
gingerBill
dc55e88588
Add @(deferred_*_by_ptr=<proc>)
2023-04-15 16:04:04 +01:00
destroycomputers
b6f356c211
Fix intrinsics.alloca code generation
...
There was a disconnect between the declared return type for alloca
intrinsic in check_builtin.cpp (multi_pointer(t_u8)) and the generated
result type in llvm_backend_proc.cpp (t_u8_ptr).
This allowed slicing the return type, but in the code generation process
the type of the expression wasn't sliceable, which triggered the assert.
Fixes #2139
2023-04-11 01:11:01 +02:00
gingerBill
8dc70f797c
Increase use of temporary_allocator() where possible
2023-03-16 15:16:17 +00:00
gingerBill
c1c7128634
Minimize severe memory usage by enforcing the heap_allocator() in places
2023-03-16 15:04:57 +00:00
gingerBill
233f47cc99
Fix #2329
2023-02-19 12:47:14 +00:00
gingerBill
c08809e29d
Improve handling of passing constants to implicit immutable const ref parameters
2023-02-17 14:49:37 +00:00
gingerBill
8a16fd7699
Rename built-in procedure to expand_values
2023-02-07 15:39:39 +00:00
gingerBill
0324281634
Enforce dynamic map calls for the time being
2023-02-03 15:17:30 +00:00
gingerBill
de0a3e0ab9
Minor change to byval for readonly parameters
2023-02-03 15:07:44 +00:00
gingerBill
881ef69063
Use ~{} syntax
2023-01-23 10:14:21 +00:00
gingerBill
761a19689d
Escape $ in asm
2023-01-23 10:09:22 +00:00
gingerBill
5d397804f7
Fix #2286 by printing an error correctly
2023-01-16 13:22:37 +00:00
gingerBill
68b2d4b9e2
Fix #2305
2023-01-16 11:41:58 +00:00
gingerBill
3b22c6620c
Begin to generalize modules away from AstPackage * in -use-separate-modules
2023-01-12 17:13:25 +00:00
gingerBill
402a165b60
Correct missing procedures in other build modules which cause a linkage problem
2023-01-12 16:59:16 +00:00
gingerBill
252be0fb41
Make all maps use heap allocator implicitly
2023-01-03 11:59:52 +00:00
gingerBill
fa562ec5d6
Remove unneeded local_entity_map
2023-01-02 15:40:25 +00:00
gingerBill
529383f5b1
Correct a race condition when checking the procedure body
2023-01-02 15:30:04 +00:00
gingerBill
9b278db993
Revert "Change tav to be a pointer internally"
...
This reverts commit e98f1a28e6
2022-12-22 12:01:41 +00:00
gingerBill
e98f1a28e6
Change tav to be a pointer internally
2022-12-22 11:53:13 +00:00
gingerBill
c1f5be24e2
Remove dead code in the compiler
2022-12-18 22:49:10 +00:00
gingerBill
6cdec65ca1
gb_internal LLVM backend
2022-12-18 22:32:05 +00:00
gingerBill
1f8f94276e
Initialize the multiple return value map in lb_create_dummy_procedure
2022-12-07 16:44:26 +00:00
gingerBill
d88b052d2d
Naïve optimization of named _split_ multiple return valued when defer is never used
...
This is a naïve optimization but it helps a lot in the general case where callee temporary stack variables
are not allocated to represent the named return values by using that specific memory.
In the future, try to check if a specific named return value is ever used a `defer` within a procedure or not,
or is ever passed to a nested procedure call (e.g. possibly escapes).
2022-11-25 23:57:55 +00:00
gingerBill
615eccb6d1
Correct return ptr semantics for split returns
2022-11-24 14:26:45 +00:00
gingerBill
0befadde1d
Basic copy elision support for multiple return values
2022-11-24 01:27:39 +00:00
gingerBill
aef8b25a8e
Listen to past Bill's wisdom
2022-11-23 23:54:12 +00:00
gingerBill
7ab591667a
Basic support for new ABI experiment on Win64
2022-11-23 16:25:09 +00:00
gingerBill
0a0db23b17
Remove copy elision code
2022-11-22 15:49:27 +00:00
gingerBill
1baa47c78e
Fix #2179
2022-11-21 11:31:35 +00:00
gingerBill
5bbab05161
Fix #2199
2022-11-21 11:02:52 +00:00
gingerBill
25bec19b1f
Revert "Minor improvement to multi return value reducing stack usage"
2022-11-13 23:56:05 +00:00
gingerBill
d2019e3e4d
Enforce pointer cast
2022-11-13 23:50:45 +00:00
gingerBill
a705a2e38b
Minor improvement to multi return value reducing stack usage
2022-11-13 22:55:32 +00:00
gingerBill
2b83f27f06
Merge branch 'master' into map-dev
2022-11-11 15:54:33 +00:00
gingerBill
fcd8860990
Make intrinsics.ptr_sub use explicit integer arithmetic internally
2022-11-11 15:52:49 +00:00
gingerBill
8852d090b6
Correct static map get; make get take a pointer to simplify compiler internals
2022-11-10 12:46:53 +00:00
gingerBill
a74093784c
Add intrinsics.map_cell_info and intrinsics.map_info
2022-11-08 12:24:00 +00:00
gingerBill
ed58374964
Make Map_Info store pointers to cell info rather than inline
2022-11-08 12:18:36 +00:00
