- This is an assumption that is currently hardwired into the backend, we need
to do this in order for the frontend and backend to agree.
llvm-svn: 166428
are no known current users of column info. Robustify and fix up
a few tests in the process. Reduces the size of debug information
by a small amount.
Part of PR14106
llvm-svn: 166236
crtfastmath.o contains routines to set the floating point flags to a faster,
unsafe mode. Linking it in speeds up code dealing with denormals significantly
(PR14024).
For now this is only enabled on linux where I can test it and crtfastmath.o is
widely available. We may want to provide a similar file with compiler-rt
eventually and/or enable it on other platforms too.
llvm-svn: 165240
clang specifying a temporary file that it later cleans up so that it can survive
the linking stage. However, when we compile object files during LTO we don't
call 'dsymutil'. That's done at a different stage (if at all). We rely upon the
linker to specify a unique name for the temporary file it generates.
<rdar://problem/12401423>
llvm-svn: 165028
passing -fretain-comments-from-system-headers. By default, the
compiler no longer parses such documentation comments, as they
can result in a noticeable compile time/PCH slowdown.
Fixes <rdar://problem/11860820>.
llvm-svn: 163778
Android uses the same flavour of crt*.o for PIE and non-PIE executables, and a
different one for DSOs. GNU/Linux, on the other hand, uses one set of crt*.o
for non-PIE executables, and another for both PIE executables and DSOs.
llvm-svn: 163500
Most of the code guarded with ANDROIDEABI are not
ARM-specific, and having no relation with arm-eabi.
Thus, it will be more natural to call this
environment "Android" instead of "ANDROIDEABI".
Note: We are not using ANDROID because several projects
are using "-DANDROID" as the conditional compilation
flag.
llvm-svn: 163088
This improves compatibility with gcc in this regard, and this file generation
can be ameliorated with GCOV_PREFIX and GCOV_PREFIX_STRIP. It's also useful if
your build directory doesn't specify -o <abspath> and it uses a recursive make
structure, so it's not relative to the toplevel.
Patch by Joshua Cranmer!
<rdar://problem/12179524>
llvm-svn: 162884
diagnostics for bad deployment targets and adding a few
more predicates. Includes a patch by Jonathan Schleifer
to enable ARC for ObjFW.
llvm-svn: 162252
If you build with -fobjc-arc, then -fobjc-link-runtime is implied but we
don't need to warn about it being unused in that case. rdar://12039965
llvm-svn: 161444
assembly.
By default, we don't emit IR for MS-style inline assembly (see r158833 as to
why). This is strictly for testing purposes and should not be enabled with the
expectation that things will work. This is a temporary flag and will be removed
once MS-style inline assembly is fully supported.
llvm-svn: 160573
This macro was being unconditionally set to zero, preceded by a FIXME comment.
This fixes <rdar://problem/11845441>. Patch by Michael Gottesman!
llvm-svn: 160491
- Split pedantic driver flag test into separate test file, and XFAIL on cygwin,mingw32
- Fix bug in tablegen logic where a missing '{' caused errors to be included in -Wpedantic.
llvm-svn: 159892
I suspect FileCheck might match assertion failure, even if clang/test/Misc/warning-flags.c passed the test.
> 0. Program arguments: bin/./clang -### -pedantic -Wpedantic clang/test/Driver/warning-options.cpp
llvm-svn: 159886
This patch introduces some magic in tablegen to create a "Pedantic" diagnostic
group which automagically includes all warnings that are extensions. This
allows a user to suppress specific warnings traditionally under -pedantic used
an ordinary warning flag. This also allows users to use #pragma to silence
specific -pedantic warnings, or promote them to errors, within blocks of text
(just like any other warning).
-Wpedantic is NOT an alias for -pedantic. Instead, it provides another way
to (a) activate -pedantic warnings and (b) disable them. Where they differ
is that -pedantic changes the behavior of the preprocessor slightly, whereas
-Wpedantic does not (it just turns on the warnings).
The magic in the tablegen diagnostic emitter has to do with computing the minimal
set of diagnostic groups and diagnostics that should go into -Wpedantic, as those
diagnostics that already members of groups that themselves are (transitively) members
of -Wpedantic do not need to be included in the Pedantic group directly. I went
back and forth on whether or not to magically generate this group, and the invariant
was that we always wanted extension warnings to be included in -Wpedantic "some how",
but the bookkeeping would be very onerous to manage by hand.
-no-pedantic (and --no-pedantic) is included for completeness, and matches many of the
same kind of flags the compiler already supports. It does what it says: cancels out
-pedantic. One discrepancy is that if one specifies --no-pedantic and -Weverything or
-Wpedantic the pedantic warnings are still enabled (essentially the -W flags win). We
can debate the correct behavior here.
Along the way, this patch nukes some code in TextDiagnosticPrinter.cpp and CXStoredDiagnostic.cpp
that determine whether to include the "-pedantic" flag in the warning output. This is
no longer needed, as all extensions now have a -W flag.
This patch also significantly reduces the number of warnings not under flags from 229
to 158 (all extension warnings). That's a 31% reduction.
llvm-svn: 159875
This flag sets the 'fp-contract' mode, which controls the formation of fused
floating point operations. Available modes are:
- Fast: Form fused operations anywhere.
- On: Form fused operations where allowed by FP_CONTRACT. This is the default
mode.
- Off: Don't form fused operations (in future this may be relaxed to forming
fused operations where it can be proved that the result won't be
affected).
Currently clang doesn't support the FP_CONTRACT pragma, so the 'On' and 'Off'
modes are equivalent.
llvm-svn: 159794
By default on OS X 10.8, we don't link with a crt1.o file and the linker
knows to use _main as the entry point. But, when compiling with -pg, we
need to link with the gcrt1.o file, and the linker needs to be told to use
the "start" symbol as the entry point. The -no_new_main linker option does
that last part. <rdar://problem/11491405>
llvm-svn: 159683
comparison between two templated types when they both appear in a diagnostic.
Type elision will remove indentical template arguments, which can be disabled
with -fno-elide-type. Cyan highlighting is applied to the differing types.
For more formatting, -fdiagnostic-show-template-tree will output the template
type as an indented text tree, with differences appearing inline. Template
tree works with or without type elision.
llvm-svn: 159216
target Objective-C runtime down to the frontend: break this
down into a single target runtime kind and version, and compute
all the relevant information from that. This makes it
relatively painless to add support for new runtimes to the
compiler. Make the new -cc1 flag, -fobjc-runtime=blah-x.y.z,
available at the driver level as a better and more general
alternative to -fgnu-runtime and -fnext-runtime. This new
concept of an Objective-C runtime also encompasses what we
were previously separating out as the "Objective-C ABI", so
fragile vs. non-fragile runtimes are now really modelled as
different kinds of runtime, paving the way for better overall
differentiation.
As a sort of special case, continue to accept the -cc1 flag
-fobjc-runtime-has-weak, as a sop to PLCompatibilityWeak.
I won't go so far as to say "no functionality change", even
ignoring the new driver flag, but subtle changes in driver
semantics are almost certainly not intended.
llvm-svn: 158793
option. On the driver, check if we are using libraries from gcc 4.7 or newer
and if so pass -fuse-init-array to the frontend.
The crtbegin*.o files in gcc 4.7 no longer call the constructors listed in
.ctors, so we have to use .init_array.
llvm-svn: 158694
This functionality is based on what is done on ARM, and enables selecting PPC CPUs
in a way compatible with gcc's driver. Also, mirroring gcc (and what is done on x86),
-mcpu=native support was added. This uses the host cpu detection from LLVM
(which will also soon be updated by refactoring code currently in backend).
In order for this to work, the target needs a list of valid CPUs -- we now accept all CPUs accepted by LLVM.
A few preprocessor defines for common CPU types have been added.
llvm-svn: 158334
used by the preprocessor. Apple's GCC also supported a -A option for linking.
The ld man page has the following:
-A basefile - Obsolete incremental load format. This option is obsolete.
Nick Kledzik confirms this option is no longer needed/supported.
rdar://11455614
llvm-svn: 156965
When enabled, clang generates bounds checks for array and pointers dereferences. Work to follow in LLVM's backend.
OK'ed by Chad; thanks for the review.
llvm-svn: 156431
It reduces the amount of emitted debug information:
1) DIEs in .debug_info have types DW_TAG_compile_unit, DW_TAG_subprogram,
DW_TAG_inlined_subroutine (for opt builds) and DW_TAG_lexical_block only.
2) .debug_str contains only function names.
3) No debug data for types/namespaces/variables is emitted.
4) The data in .debug_line is enough to produce valid stack traces with
function names and line numbers.
Reviewed by Eric Christopher.
llvm-svn: 156160
For now -fno-math-errno is the default on BSD-derived platforms (Darwin,
DragonFlyBSD, FreeBSD, NetBSD, OpenBSD). If the default is not right for
your platform, please yell. I only verified the result with the default
compilers on Darwin and FreeBSD.
llvm-svn: 155990
-fpack-struct's handling has changed in CC1 (one of only two flags that needed changing) because the driver treats "-fpack-struct" as a boolean flag, and CC1 (did) treat it as an option with a separated value.
This change causes -fpack-struct=X to be forwarded correctly to -fpack-struct=X instead of erroneously to "-fpack-struct X"
llvm-svn: 155981
Linux and other (non-Darwin) platforms and have it use -fmath-errno by
default (for better or worse).
Darwin has seen the light here and uses -fno-math-errno by default, this
patch preserves that.
If any maintainers for a non-Linux platform would also like to opt-in to
-fno-math-errno by default, I'm happy to add folks, but we're currently
getting buts and misleading comparisons with GCC due to this difference
in behavior on Linux at least.
llvm-svn: 155607
r155047. See the LLVM log for the primary motivation:
http://llvm.org/viewvc/llvm-project?rev=155047&view=rev
Primary commit r154828:
- Several issues were raised in review, and fixed in subsequent
commits.
- Follow-up commits also reverted, and which should be folded into the
original before reposting:
- r154837: Re-add the 'undef BUILTIN' thing to fix the build.
- r154928: Fix build warnings, re-add (and correct) header and
license
- r154937: Typo fix.
Please resubmit this patch with the relevant LLVM resubmission.
llvm-svn: 155048
requires the -plugin to come before any -plugin-opt options, we were passing
them the other way around. With this one can run (for example):
clang -o foo foo.c -O4 -Wl,-plugin-opt=generate-api-file
llvm-svn: 154357
First, this patch cleans up the parsing of the PIC and PIE family of
options in the driver. The existing logic failed to claim arguments all
over the place resulting in kludges that marked the options as unused.
Instead actually walk all of the arguments and claim them properly.
We now treat -f{,no-}{pic,PIC,pie,PIE} as a single set, accepting the
last one on the commandline. Previously there were lots of ordering bugs
that could creep in due to the nature of the parsing. Let me know if
folks would like weird things such as "-fPIE -fno-pic" to turn on PIE,
but disable full PIC. This doesn't make any sense to me, but we could in
theory support it.
Options that seem to have intentional "trump" status (-static, -mkernel,
etc) continue to do so and are commented as such.
Next, a -pie-level flag is threaded into the frontend, rigged to
a language option, and handled preprocessor, setting up the appropriate
defines. We'll now have the correct defines when compiling with -fpie.
The one place outside of the preprocessor that was inspecting the PIC
level (as opposed to the relocation model, which is set and handled
separately, yay!) is in the GNU ObjC runtime. I changed it to exactly
preserve existing behavior. If folks want to change its behavior in the
face of PIE, they can do that in a separate patch.
Essentially the only functionality changed here is the preprocessor
defines and bug-fixes to the argument management.
Tests have been updated and extended to test all of this a bit more
thoroughly.
llvm-svn: 154291
uses Neon instructions for single-precision FP.
-mfpmath=neon is analogous to passing llc -mattr=+neonfp.
-mfpmath=[vfp|vfp2|vfp3|vfp4] is analogous to passing llc -mattr=-neonfp.
rdar://11108618
llvm-svn: 154046
llvm-gcc doesn't handle --serialize-diagnostics so when compiling i386
kernel/kext code with -Werror, you get an error about that option being
unused. Claim the argument to prevent this from breaking builds.
<rdar://problem/11161933>
llvm-svn: 153854
flag as GCC uses: -fstrict-enums). There is a *lot* of code making
unwarranted assumptions about the underlying type of enums, and it
doesn't seem entirely reasonable to eagerly break all of it.
Much more importantly, the current state of affairs is *very* good at
optimizing based upon this information, which causes failures that are
very distant from the actual enum. Before we push for enabling this by
default, I think we need to implement -fcatch-undefined-behavior support
for instrumenting and trapping whenever we store or load a value outside
of the range. That way we can track down the misbehaving code very
quickly.
I discussed this with Rafael, and currently the only important cases he
is aware of are the bool range-based optimizations which are staying
hard enabled. We've not seen any issue with those either, and they are
much more important for performance.
llvm-svn: 153550
1. Don't short-circuit conditional statements that are checking flags.
Otherwise, the driver emits warnings about unused arguments.
2. -mkernel and -fapple-kext imply no exceptions, so claim exception related
arguments now to avoid warnings about unused arguments.
rdar://11120518
llvm-svn: 153478
The getARMTargetCPU and getLLVMArchSuffixForARM functions exist in both
Toolchain.cpp and Tools.cpp. This stuff needs a thorough overhaul. In the
meantime, this patch at least makes them consistent. One version had been
converted to use StringSwitch, and the other version had new Cortex M-series
processors added.
llvm-svn: 153202
Original commit message:
Provide -Wnull-conversion separately from -Wconversion.
Like GCC, provide a NULL conversion to non-pointer conversion as a separate
flag, on by default. GCC's flag is "conversion-null" which we provide for
cross compatibility, but in the interests of consistency (with
-Wint-conversion, -Wbool-conversion, etc) the canonical Clang flag is called
-Wnull-conversion.
Patch by Lubos Lunak.
Review feedback by myself, Chandler Carruth, and Chad Rosier.
llvm-svn: 152774
Like GCC, provide a NULL conversion to non-pointer conversion as a separate
flag, on by default. GCC's flag is "conversion-null" which we provide for
cross compatibility, but in the interests of consistency (with
-Wint-conversion, -Wbool-conversion, etc) the canonical Clang flag is called
-Wnull-conversion.
Patch by Lubos Lunak.
Review feedback by myself, Chandler Carruth, and Chad Rosier.
llvm-svn: 152745
The LIBRARY_PATH environment variable should be honored by clang. Have the
driver pass the directories to the linker.
<rdar://problem/9743567> and PR10296.
llvm-svn: 152578
the new Objective-C NSArray/NSDictionary/NSNumber literal syntax.
This introduces a new library, libEdit, which provides a new way to support
migration of code that improves on the original ARC migrator. We now believe
that most of its functionality can be refactored into the existing libraries,
and thus this new library may shortly disappear.
llvm-svn: 152141
NSNumber, and boolean literals. This includes both Sema and Codegen support.
Included is also support for new Objective-C container subscripting.
My apologies for the large patch. It was very difficult to break apart.
The patch introduces changes to the driver as well to cause clang to link
in additional runtime support when needed to support the new language features.
Docs are forthcoming to document the implementation and behavior of these features.
llvm-svn: 152137
This flag enables ThreadSanitizer instrumentation committed to llvm as r150423.
The patch includes one test for -fthread-sanitizer and one similar test for -faddress-sanitizer.
This patch does not modify the linker flags (as we do it for -faddress-sanitizer) because the run-time library is not yet
committed and it's structure in compiler-rt is not 100% clear.
The users manual wil be changed in a separate commit.
llvm-svn: 151846
This option was added in r129614 and doesn't have any use case that I'm aware
of. It's possible that external tools are using these names - and if that's
the case we can certainly reassess the functionality, but for now it lets us
shave out a few unneeded bits from clang.
Move the "StaticDiagNameIndex" table into the only remaining consumer, diagtool.
This removes the actual diagnostic name strings from clang entirely.
Reviewed by Chris Lattner & Ted Kremenek.
llvm-svn: 150612
world on Solaris 11 for both x86 and x86-64 using the built-in assembler and
Solaris (not GNU) ld, however it currently relies on a hard-coded GCC location
to find crtbegin.o and crtend.o, as well as libgcc and libgcc_eh.
llvm-svn: 150580
When creating the MCSubtargetInfo, the assembler driver uses the CPU and
feature string to construct a more accurate model of what instructions
are and are not legal.
rdar://10840476
llvm-svn: 150273
That llvm change removed the -trap-func backend option, so that using
-ftrap-function with clang would cause the backend to complain. Fix it
by adding the trap function name to the CodeGenOptions and passing it through
to the TargetOptions.
llvm-svn: 149679
the recent refactoring. All interesting NetBSD release have a GNU as
version on i386 that supports --32, so don't bother with the conditional
setting of it.
llvm-svn: 149087
did anything. The two big pieces of functionality it tried to provide
was to cache the ToolChain objects for each target, and to figure out
the exact target based on the flag set coming in to an invocation.
However, it had a lot of flaws even with those goals:
- Neither of these have anything to do with the host, or its info.
- The HostInfo class was setup as a full blown class *hierarchy* with
a separate implementation for each "host" OS. This required
dispatching just to create the objects in the first place.
- The hierarchy claimed to represent the host, when in fact it was
based on the target OS.
- Each leaf in the hierarchy was responsible for implementing the flag
processing and caching, resulting in a *lot* of copy-paste code and
quite a few bugs.
- The caching was consistently done based on architecture alone, even
though *any* aspect of the targeted triple might change the behavior
of the configured toolchain.
- Flag processing was already being done in the Driver proper,
separating the flag handling even more than it already is.
Instead of this, we can simply have the dispatch logic in the Driver
which previously created a HostInfo object create the ToolChain objects.
Adding caching in the Driver layer is a tiny amount of code. Finally,
pulling the flag processing into the Driver puts it where it belongs and
consolidates it in one location.
The result is that two functions, and maybe 100 lines of new code
replace over 10 classes and 800 lines of code. Woot.
This also paves the way to introduce more detailed ToolChain objects for
various OSes without threading through a new HostInfo type as well, and
the accompanying boiler plate. That, of course, was the yak I started to
shave that began this entire refactoring escapade. Wheee!
llvm-svn: 148950
Patch from Jyotsna Verma:
I have made the changes to remove assertions in the Hexagon backend
specific clang driver. Instead of asserting on invalid arch name, it has
been modified to use the default value.
I have changed the implementation of the CPU flag validation for the
Hexagon backend. Earlier, the clang driver performed the check and
asserted on invalid inputs. In the new implementation, the driver passes
the last CPU flag (or sets to "v4" if not specified) to the compiler (and
also to the assembler and linker which perform their own check) instead of
asserting on incorrect values. This patch changes the setCPU function for
the Hexagon backend in clang/lib/Basic/Targets.cpp which causes the
compiler to error out on incorrect CPU flag values.
llvm-svn: 148139
- Support gcc-compatible vfpv3 name in addition to vfp3.
- Support vfpv3-d16.
- Disable neon feature for -mfpu=vfp* (yes, we were emitting Neon instructions
for those!).
llvm-svn: 147943
for the arm-linux-androideabi triple in particular.
Also use this to do a better job of selecting soft FP settings.
Patch by Evgeniy Stepanov.
llvm-svn: 147872
source file. Otherwise -g -save-temps will error out on the compile
of any .c file.
Fixes about 4000 of the errors in the clang-tests gdb test suite.
llvm-svn: 147819
module imports from -fauto-module-import to -fmodules. The new name
will eventually be used to enable modules, and the #include/#import
mapping is a crucial part of the feature.
llvm-svn: 147447
Clang driver. This involves a bunch of silly option parsing code to try
to carefully emulate GCC's options. Currently, this takes a conservative
approach, and unless all of the unsafe optimizations are enabled, none
of them are. The fine grained control doesn't seem particularly useful.
If it ever becomes useful, we can add that to LLVM first, and then
expose it here.
This also fixes a few tiny bugs in the flag management around
-fhonor-infinities and -fhonor-nans; the flags now form proper sets both
for enabling and disabling, with the last flag winning.
I've also implemented a moderately terrifying GCC feature where
a language change is also provided by the '-ffast-math' flag by defining
the __FAST_MATH__ preprocessor macro. This feature is tracked and
serialized in the frontend but it isn't used yet. A subsequent patch
will add the preprocessor macro and tests for it.
I've manually tested that codegen appears to respect this, but I've not
dug in enough to see if there is an easy way to test codegen options w/o
relying on the particulars of LLVM's optimizations.
llvm-svn: 147434
fails within a call to a constexpr function. Add -fconstexpr-backtrace-limit
argument to driver and frontend, to control the maximum number of notes so
produced (default 10). Fix APValue printing to be able to pretty-print all
APValue types, and move the testing for this functionality from a unittest to
a -verify test now that it's visible in clang's output.
llvm-svn: 146749
. move compiler-rt to a separate directory so the -L argument only includes compiler-rt (thanks joerg)
. build all clang subdirs
. switches the Minix platform to ELF
. normalizes toolchain invocation
Patch by Ben Gras.
llvm-svn: 146206
it to GNU assembler. In addition, change function getMipsArchFromCPU() so that
it can be reused in ConstructJob().
Patch by Simon Atanasyan.
llvm-svn: 145509
semantics and defaults as the corresponding g++ arguments. The historical g++
argument -ftemplate-depth-N is kept for compatibility, but modern g++ versions
no longer document that option.
Add -cc1 argument -fconstexpr-depth N to implement the corresponding
functionality.
The -ftemplate-depth=N part of this fixes PR9890.
llvm-svn: 145045
output files that are valid regardless of whether the compilation
succeeded or failed (but not if we crash). Add depfiles to the
failure result file list.
llvm-svn: 145018
This is a partial revert of r143846. While cleaning up after a crash is
probably a good idea, we were also deleting .d files if the compilation failed
due to invalid input, which is not the desired behavior. The test is XFAIL'd
until the cleanup code can be reworked to do the right thing.
llvm-svn: 144590
the first (and diff-noisiest) step to making Linux header searching
tremendously more principled and less brittle. Note that this step
should have essentially no functional impact. We still search the exact
same set of paths in the exact same order. The only change here is where
the code implementing such a search lives.
This has one obvious negative impact -- we now pass a ludicrous number
of flags to the CC1 layer. That should go away as I re-base this logic
on the logic to detect a GCC installation. I want to do this in two
phases so the bots can tell me if this step alone breaks something, and
so that the diffs of the refactoring make more sense.
llvm-svn: 143822
handling logic of the generic ToolChain. This flag, despite its name,
has *nothing* to do with the GCC flag '-nostdlib' that relates
(exclusively) to the linking behavior. It is a most unfortunate name in
that regard...
It is used to tell InitHeaderSearch.cpp *which* set of C++ standard
library header search paths to use -- those for libstdc++ from GCC's
installation, or those from a libc++ installation. As this logic is
hoisted out of the Frontend, and into the Driver as part of this
ToolChain, the generic method will be overridden for the platform, where
it can implement this logic directly. As such, hiding the CC1 option
passing in the generic space is a natural fit despite the odd naming.
Also, expand on the comments to clarify whats going on, and tidy up the
Tools.cpp code now that its simpler.
llvm-svn: 143687
implementation in the driver. This cleans up the signature and semantics
of the include flag adding component of the toolchain. Another step to
ready it for holding all the InitHeaderSearch logic.
llvm-svn: 143686
the rest of the mess in InitHeaderSearch.cpp. We could hoist it into the
driver profitably, removing more noise from the driver -> frontend
communication.
llvm-svn: 143685
and the C++ include management routine from the proper place when
forming preprocessor options in the driver. This is the first step to
teaching the driver to manage all of the header search paths. Currently,
these methods remain just stubs in the abstract toolchain. Subsequent
patches will flesh them out with implementations for various toolchains
based on the current code in InitHeaderSearch.cpp.
llvm-svn: 143684
The -g and --gdwarf2 options are currently synonyms to the Darwin assembler.
But clang itself does not recognize --gdwarf2, so if we want to experiment
with using clang, with its integrated assembler, to replace the default
assembler, it is necessary to use -g. <rdar://problem/10349486>
llvm-svn: 143533
Daniel Dunbar