Introduction/Motivation:
LLVM-IR supports integers of non-power-of-2 bitwidth, in the iN syntax.
Integers of non-power-of-two aren't particularly interesting or useful
on most hardware, so much so that no language in Clang has been
motivated to expose it before.
However, in the case of FPGA hardware normal integer types where the
full bitwidth isn't used, is extremely wasteful and has severe
performance/space concerns. Because of this, Intel has introduced this
functionality in the High Level Synthesis compiler[0]
under the name "Arbitrary Precision Integer" (ap_int for short). This
has been extremely useful and effective for our users, permitting them
to optimize their storage and operation space on an architecture where
both can be extremely expensive.
We are proposing upstreaming a more palatable version of this to the
community, in the form of this proposal and accompanying patch. We are
proposing the syntax _ExtInt(N). We intend to propose this to the WG14
committee[1], and the underscore-capital seems like the active direction
for a WG14 paper's acceptance. An alternative that Richard Smith
suggested on the initial review was __int(N), however we believe that
is much less acceptable by WG14. We considered _Int, however _Int is
used as an identifier in libstdc++ and there is no good way to fall
back to an identifier (since _Int(5) is indistinguishable from an
unnamed initializer of a template type named _Int).
[0]https://www.intel.com/content/www/us/en/software/programmable/quartus-prime/hls-compiler.html)
[1]http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2472.pdf
Differential Revision: https://reviews.llvm.org/D73967
Summary:
This patch adds a mechanism to easily add range checks for a builtin's
immediate operands. This patch is tested with the qdech intrinsic, which takes
both an enum for the predicate pattern, as well as an immediate for the
multiplier.
Reviewers: efriedma, SjoerdMeijer, rovka
Reviewed By: efriedma, SjoerdMeijer
Subscribers: mgorny, tschuett, mgrang, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76678
WG14 has adopted N2480 (http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2480.pdf)
into C2x at the meetings last week, allowing parameter names of a function
definition to be elided. This patch relaxes the error so that C++ and C2x do not
diagnose this situation, and modes before C2x will allow it as an extension.
This also adds the same feature to ObjC blocks under the assumption that ObjC
wishes to follow the C standard in this regard.
instead of recursing on the stack.
This doesn't actually resolve PR45333, because we now hit stack overflow
somewhere else, but it does get us further. I've not found any way of
testing this that doesn't still crash elsewhere.
Summary: If the size parameter of `__builtin_memcpy_inline` comes from an un-instantiated template parameter current code would crash.
Reviewers: efriedma, courbet
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76504
Summary:
This patch generalizes the existing code to support CDE intrinsics
which will share some properties with existing MVE intrinsics
(some of the intrinsics will be polymorphic and accept/return values
of MVE vector types).
Specifically the patch:
* Adds new tablegen backends -gen-arm-cde-builtin-def,
-gen-arm-cde-builtin-codegen, -gen-arm-cde-builtin-sema,
-gen-arm-cde-builtin-aliases, -gen-arm-cde-builtin-header based on
existing MVE backends.
* Renames the '__clang_arm_mve_alias' attribute into
'__clang_arm_builtin_alias' (it will be used with CDE intrinsics as
well as MVE intrinsics)
* Implements semantic checks for the coprocessor argument of the CDE
intrinsics as well as the existing coprocessor intrinsics.
* Adds one CDE intrinsic __arm_cx1 to test the above changes
Reviewers: simon_tatham, MarkMurrayARM, ostannard, dmgreen
Reviewed By: simon_tatham
Subscribers: sdesmalen, mgorny, kristof.beyls, danielkiss, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75850
Clang is missing a warning for
builtin_return_address/builtin_frame_address called with > 0 argument.
Gcc provides a warning for this via -Wframe-address:
https://gcc.gnu.org/onlinedocs/gcc/Return-Address.html
As calling these functions with argument > 0 has caused several crashes
for us, we would like to have the same warning as gcc here. This diff
adds the warning and makes it part of -Wmost.
Differential Revision: https://reviews.llvm.org/D75768
__builtin_os_log_format
This is needed to keep all the objects, including temporaries returned
by function calls, written to the buffer alive until os_log_pack_send is
called.
rdar://problem/60105410
Summary:
There was even a TODO for this.
The main motivation is to make use of call-site based
`__attribute__((alloc_align(param_idx)))` validation (D72996).
Reviewers: rsmith, erichkeane, aaron.ballman, jdoerfert
Reviewed By: rsmith
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73020
Verifies that an argument passed to __builtin_frame_address or __builtin_return_address is within the range [0, 0xFFFF]
Differential revision: https://reviews.llvm.org/D66839
Re-committed after fixed: c93112dc4f
Summary:
As @rsmith notes in https://reviews.llvm.org/D73020#inline-672219
while that is certainly UB land, it may not be actually reachable at runtime, e.g.:
```
template<int N> void *make() {
if ((N & (N-1)) == 0)
return operator new(N, std::align_val_t(N));
else
return operator new(N);
}
void *p = make<7>();
```
and we shouldn't really error-out there.
That being said, i'm not really following the logic here.
Which ones of these cases should remain being an error?
Reviewers: rsmith, erichkeane
Reviewed By: erichkeane
Subscribers: cfe-commits, rsmith
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73996
New intrinisics are implemented for when we need to port SIMD code from other
arhitectures and only load or store portions of MSA registers.
Following intriniscs are added which only load/store element 0 of a vector:
v4i32 __builtin_msa_ldrq_w (const void *, imm_n2048_2044);
v2i64 __builtin_msa_ldr_d (const void *, imm_n4096_4088);
void __builtin_msa_strq_w (v4i32, void *, imm_n2048_2044);
void __builtin_msa_str_d (v2i64, void *, imm_n4096_4088);
Differential Revision: https://reviews.llvm.org/D73644
Implement a pessimistic evaluator of the minimal required size for a buffer
based on the format string, and couple that with the fortified version to emit a
warning when the buffer size is lower than the lower bound computed from the
format string.
Differential Revision: https://reviews.llvm.org/D71566
There is llvm::Value::MaximumAlignment, which is numerically
equivalent to these constants, but we can't use it directly
because we can't include llvm IR headers in clang Sema.
So instead, copy-paste the constant, and fixup the places to use it.
This was initially reviewed in https://reviews.llvm.org/D72998
Summary:
`alloc_align` attribute takes parameter number, not the alignment itself,
so given **just** the attribute/function declaration we can't do any
sanity checking for said alignment.
However, at call site, given the actual `Expr` that is passed
into that parameter, we //might// be able to evaluate said `Expr`
as Integer Constant Expression, and perform the sanity checks.
But since there is no requirement for that argument to be an immediate,
we may fail, and that's okay.
However if we did evaluate, we should enforce the same constraints
as with `__builtin_assume_aligned()`/`__attribute__((assume_aligned(imm)))`:
said alignment is a power of two, and is not greater than our magic threshold
This was initially committed in c2a9061ac5
but reverted in 00756b1823 because of
suspicious bot failures.
Reviewers: erichkeane, aaron.ballman, hfinkel, rsmith, jdoerfert
Reviewed By: erichkeane
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72996
Summary:
I initially encountered those assertions when trying to create
this IR `alignment` attribute from clang's `__attribute__((assume_aligned(imm)))`,
because until D72994 there is no sanity checking for the value of `imm`.
But even then, we have `llvm::Value::MaximumAlignment` constant (which is `536870912`),
which is enforced for clang attributes, and then there are some other magical constant
(`0x40000000` i.e. `1073741824` i.e. `2 * 536870912`) in
`Attribute::getWithAlignment()`/`AttrBuilder::addAlignmentAttr()`.
I strongly suspect that `0x40000000` is incorrect,
and that also should be `llvm::Value::MaximumAlignment`.
Reviewers: erichkeane, hfinkel, jdoerfert, gchatelet, courbet
Reviewed By: erichkeane
Subscribers: hiraditya, cfe-commits, llvm-commits
Tags: #llvm, #clang
Differential Revision: https://reviews.llvm.org/D72998
Summary:
`alloc_align` attribute takes parameter number, not the alignment itself,
so given **just** the attribute/function declaration we can't do any
sanity checking for said alignment.
However, at call site, given the actual `Expr` that is passed
into that parameter, we //might// be able to evaluate said `Expr`
as Integer Constant Expression, and perform the sanity checks.
But since there is no requirement for that argument to be an immediate,
we may fail, and that's okay.
However if we did evaluate, we should enforce the same constraints
as with `__builtin_assume_aligned()`/`__attribute__((assume_aligned(imm)))`:
said alignment is a power of two, and is not greater than our magic threshold
Reviewers: erichkeane, aaron.ballman, hfinkel, rsmith, jdoerfert
Reviewed By: erichkeane
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72996
Summary:
Immediate vmvnq is code-generated as a simple vector constant in IR,
and left to the backend to recognize that it can be created with an
MVE VMVN instruction. The predicated version is represented as a
select between the input and the same constant, and I've added a
Tablegen isel rule to turn that into a predicated VMVN. (That should
be better than the previous VMVN + VPSEL: it's the same number of
instructions but now it can fold into an adjacent VPT block.)
The unpredicated forms of VBIC and VORR are done by enabling the same
isel lowering as for NEON, recognizing appropriate immediates and
rewriting them as ARMISD::VBICIMM / ARMISD::VORRIMM SDNodes, which I
then instruction-select into the right MVE instructions (now that I've
also reworked those instructions to use the same MC operand encoding).
In order to do that, I had to promote the Tablegen SDNode instance
`NEONvorrImm` to a general `ARMvorrImm` available in MVE as well, and
similarly for `NEONvbicImm`.
The predicated forms of VBIC and VORR are represented as a vector
select between the original input vector and the output of the
unpredicated operation. The main convenience of this is that it still
lets me use the existing isel lowering for VBICIMM/VORRIMM, and not
have to write another copy of the operand encoding translation code.
This intrinsic family is the first to use the `imm_simd` system I put
into the MveEmitter tablegen backend. So, naturally, it showed up a
bug or two (emitting bogus range checks and the like). Fixed those,
and added a full set of tests for the permissible immediates in the
existing Sema test.
Also adjusted the isel pattern for `vmovlb.u8`, which stopped matching
because lowering started turning its input into a VBICIMM. Now it
recognizes the VBICIMM instead.
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72934
This change introduces three new builtins (which work on both pointers
and integers) that can be used instead of common bitwise arithmetic:
__builtin_align_up(x, alignment), __builtin_align_down(x, alignment) and
__builtin_is_aligned(x, alignment).
I originally added these builtins to the CHERI fork of LLVM a few years ago
to handle the slightly different C semantics that we use for CHERI [1].
Until recently these builtins (or sequences of other builtins) were
required to generate correct code. I have since made changes to the default
C semantics so that they are no longer strictly necessary (but using them
does generate slightly more efficient code). However, based on our experience
using them in various projects over the past few years, I believe that adding
these builtins to clang would be useful.
These builtins have the following benefit over bit-manipulation and casts
via uintptr_t:
- The named builtins clearly convey the semantics of the operation. While
checking alignment using __builtin_is_aligned(x, 16) versus
((x & 15) == 0) is probably not a huge win in readably, I personally find
__builtin_align_up(x, N) a lot easier to read than (x+(N-1))&~(N-1).
- They preserve the type of the argument (including const qualifiers). When
using casts via uintptr_t, it is easy to cast to the wrong type or strip
qualifiers such as const.
- If the alignment argument is a constant value, clang can check that it is
a power-of-two and within the range of the type. Since the semantics of
these builtins is well defined compared to arbitrary bit-manipulation,
it is possible to add a UBSAN checker that the run-time value is a valid
power-of-two. I intend to add this as a follow-up to this change.
- The builtins avoids int-to-pointer casts both in C and LLVM IR.
In the future (i.e. once most optimizations handle it), we could use the new
llvm.ptrmask intrinsic to avoid the ptrtoint instruction that would normally
be generated.
- They can be used to round up/down to the next aligned value for both
integers and pointers without requiring two separate macros.
- In many projects the alignment operations are already wrapped in macros (e.g.
roundup2 and rounddown2 in FreeBSD), so by replacing the macro implementation
with a builtin call, we get improved diagnostics for many call-sites while
only having to change a few lines.
- Finally, the builtins also emit assume_aligned metadata when used on pointers.
This can improve code generation compared to the uintptr_t casts.
[1] In our CHERI compiler we have compilation mode where all pointers are
implemented as capabilities (essentially unforgeable 128-bit fat pointers).
In our original model, casts from uintptr_t (which is a 128-bit capability)
to an integer value returned the "offset" of the capability (i.e. the
difference between the virtual address and the base of the allocation).
This causes problems for cases such as checking the alignment: for example, the
expression `if ((uintptr_t)ptr & 63) == 0` is generally used to check if the
pointer is aligned to a multiple of 64 bytes. The problem with offsets is that
any pointer to the beginning of an allocation will have an offset of zero, so
this check always succeeds in that case (even if the address is not correctly
aligned). The same issues also exist when aligning up or down. Using the
alignment builtins ensures that the address is used instead of the offset. While
I have since changed the default C semantics to return the address instead of
the offset when casting, this offset compilation mode can still be used by
passing a command-line flag.
Reviewers: rsmith, aaron.ballman, theraven, fhahn, lebedev.ri, nlopes, aqjune
Reviewed By: aaron.ballman, lebedev.ri
Differential Revision: https://reviews.llvm.org/D71499
The current handling of the operators ||, && and ?: has a number of false
positive and false negative. The issues for operator || and && are:
1. We need to add sequencing regions for the LHS and RHS as is done for the
comma operator. Not doing so causes false positives in expressions like
`((a++, false) || (a++, false))` (from PR39779, see PR22197 for another
example).
2. In the current implementation when the evaluation of the LHS fails, the RHS
is added to a worklist to be processed later. This results in false negatives
in expressions like `(a && a++) + a`.
Fix these issues by introducing sequencing regions for the LHS and RHS, and by
not deferring the visitation of the RHS.
The issues with the ternary operator ?: are similar, with the added twist that
we should not warn on expressions like `(x ? y += 1 : y += 2)` since exactly
one of the 2nd and 3rd expression is going to be evaluated, but we should still
warn on expressions like `(x ? y += 1 : y += 2) = y`.
Differential Revision: https://reviews.llvm.org/D57747
Reviewed By: rsmith
NFCs factored out of the following patches:
- Change all of the `Expr *` to `const Expr *` in SequenceChecker for
const-correctness. SequenceChecker should not modify AST nodes.
- Add some comments.
- clang-format
Differential Revision: https://reviews.llvm.org/D57659
Reviewed By: xbolva00
The FP-classification builtins (__builtin_isfinite, etc) use variadic
packs in the definition file to mean an overload set. Because of that,
floats were converted to doubles, which is incorrect. There WAS a patch
to remove the cast after the fact.
THis patch switches these builtins to just be custom type checking,
calls the implicit conversions for the integer members, and makes sure
the correct L->R casts are put into place, then does type checking like
normal.
A future direction (that wouldn't be NFC) would consider making
conversions for the floating point parameter legal.
Note: The initial patch for this missed that certain systems need to
still convert half to float, since they dont' support that type.
This covers:
* usual arithmetic conversions (comparisons, arithmetic, conditionals)
between different enumeration types
* usual arithmetic conversions between enums and floating-point types
* comparisons between two operands of array type
The deprecation warnings are on-by-default (in C++20 compilations); it
seems likely that these forms will become ill-formed in C++23, so
warning on them now by default seems wise.
For the first two bullets, off-by-default warnings were also added for
all the cases where we didn't already have warnings (covering language
modes prior to C++20). These warnings are in subgroups of the existing
-Wenum-conversion (except that the first case is not warned on if either
enumeration type is anonymous, consistent with our existing
-Wenum-conversion warnings).
This reverts commit b1e542f302.
The original 'hack' didn't chop out fp-16 to double conversions, so
systems that use FP16ConversionIntrinsics end up in IR-CodeGen with an
i16 type isntead of a float type (like PPC64-BE). The bots noticed
this.
Reverting until I figure out how to fix this
The FP-classification builtins (__builtin_isfinite, etc) use variadic
packs in the definition file to mean an overload set. Because of that,
floats were converted to doubles, which is incorrect. There WAS a patch
to remove the cast after the fact.
THis patch switches these builtins to just be custom type checking,
calls the implicit conversions for the integer members, and makes sure
the correct L->R casts are put into place, then does type checking like
normal.
A future direction (that wouldn't be NFC) would consider making
conversions for the floating point parameter legal.
Remove implicit conversion that promotes half to double
for the target that support fp16. If the target doesn't
support fp16, fp16 will be converted to fp16 intrinsic.
Summary:
It shouldn't promote half to double or any larger precision types for fp classification builtins.
Because fp classification builtins would get incorrect result with promoted argument.
For example, __builtin_isnormal with a subnormal half value should return false, but it is not.
That the subnormal half value is promoted to a normal double value.
Reviewers: aaron.ballman
Reviewed By: aaron.ballman
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71049
This patch reapplies commit 759948467e. Patch was reverted due to a
clang-tidy test fail on Windows. The test has been modified. There
are no additional code changes.
Patch was tested with ninja check-all on Windows and Linux.
Summary of code changes:
Clang currently crashes for switch statements inside a template when the
condition is a non-integer field member because contextual implicit
conversion is skipped when parsing the condition. This conversion is
however later checked in an assert when the case statement is handled.
The conversion is skipped when parsing the condition because
the field member is set as type-dependent based on its containing class.
This patch sets the type dependency based on the field's type instead.
This patch fixes Bug 40982.
Now Clang does not check that features required by built-in functions
are enabled. That causes errors in the backend reported in PR44018.
This patch fixes this bug by checking that required features
are enabled.
This should fix PR44018.
Differential Revision: https://reviews.llvm.org/D70808
We seem to have been gradually growing support for atomic min/max operations
(exposing longstanding IR atomicrmw instructions). But until now there have
been gaps in the expected intrinsics. This adds support for the C11-style
intrinsics (i.e. taking _Atomic, rather than individually blessed by C11
standard), and the variants that return the new value instead of the original
one.
That way, people won't be misled by trying one form and it not working, and the
front-end is more friendly to people using _Atomic types, as we recommend.
Some clients of this function want to know about any expression that is known
to produce a 0/1 value, and others care about expressions that are semantically
boolean.
This fixes a -Wswitch-bool regression I introduced in 8bfb353bb3, pointed out
by Chris Hamilton!
This patch reapplies commit 76945821b9. The first version broke
buildbots due to clang-tidy test fails. The fails are because some
errors in templates are now diagnosed earlier (does not wait till
instantiation). I have modified the tests to add checks for these
diagnostics/prevent these diagnostics. There are no additional code
changes.
Summary of code changes:
Clang currently crashes for switch statements inside a template when the
condition is a non-integer field member because contextual implicit
conversion is skipped when parsing the condition. This conversion is
however later checked in an assert when the case statement is handled.
The conversion is skipped when parsing the condition because
the field member is set as type-dependent based on its containing class.
This patch sets the type dependency based on the field's type instead.
This patch fixes Bug 40982.
Reviewers: rnk, gribozavr2
Patch by: Elizabeth Andrews (eandrews)
Differential revision: https://reviews.llvm.org/D69950
Erich Keane