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
Now that we're only using -frewrite-includes rather than full preprocessing
when producing repro source files, we should also include command line macro
definitions in the repro script.
I don't have a test case for this because I'm not sure if/how I can open the
crash report file when the name is only known by scraping the crash report
output. Suggestions welcome if anyone thinks it'd be helpful.
llvm-svn: 159592
In future changes we should:
* use __builtin_trap rather than derefing 'random' volatile pointers.
* avoid dumping temporary files into /tmp when running tests, instead
preferring a location that is properly cleaned up by lit.
Review by Chandler Carruth.
llvm-svn: 159469
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
express library-level dependencies within Clang.
This is no more verbose really, and plays nicer with the rest of the
CMake facilities. It should also have no change in functionality.
llvm-svn: 158888
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
architecture; this was happening for tools such as lipo and dsymutil.
Also, if no -arch option has been specified, set the architecture based
on the TC default.
rdar://11329656
llvm-svn: 155730
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
overwriting the input file. For example,
clang -c foo.s -o foo.o -save-temps
Unfortunately, the original patch didn't compare the paths of the input and
output files. Thus, something like the following would fail to create foo.s.
cd /tmp/obj
clang -c ../src/foo.s -o foo.o -save-temps
rdar://11252615
llvm-svn: 155224
flags. We have preprocessed source, so we don't need these.
No test case as it's fairly difficult to make the compiler crash on demand. I'll
patiently wait for Ben to tell me how to do this in 2 lines of code. :)
rdar://11283560
llvm-svn: 155180
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
However, the '-x' option has special handling and wasn't following this
paradigm. Fix it to do so by claiming the arg as we parse the '-x' option.
rdar://11203340
llvm-svn: 154231
inside of a sysroot targeting a system+sysroot which is "similar" or
"compatible" with the host system. This shows up when trying to build
system images on largely compatible hardware as-if fully cross compiled.
The problem is that previously we *perfectly* mimiced GCC here, and it
turns out GCC has a bug that no one has really stumbled across. GCC will
try to look in thy system prefix ('/usr/local' f.ex.) into which it is
instaled to find libraries installed along side GCC that should be
preferred to the base system libraries ('/usr' f.ex.). This seems not
unreasonable, but it has a very unfortunate consequence when combined
with a '--sysroot' which does *not* contain the GCC installation we're
using to complete the toolchain. That results in some of the host
system's library directories being searched during the link.
Now, it so happens that most folks doing stuff like this use
'--with-sysroot' and '--disable-multilib' when configuring GCC. Even
better, they're usually not cross-compiling to a target that is similar
to the host. As a result, searching the host for libraries doesn't
really matter -- most of the time weird directories get appended that
don't exist (no arm triple lib directory, etc). Even if you're
cross-compiling from 32-bit to 64-bit x86 or vice-versa, disabling
multilib makes it less likely that you'll actually find viable libraries
on the host. But that's just luck. We shouldn't rely on this, and this
patch disables looking in the system prefix containing the GCC
installation if that system prefix is *outside* of the sysroot. For
empty sysroots, this has no effect. Similarly, when using the GCC
*inside* of the sysroot, we still track wherever it is installed within
the sysroot and look there for libraries. But now we can use a cross
compiler GCC installation outside the system root, and only look for the
crtbegin.o in the GCC installation, and look for all the other libraries
inside the system root.
This should fix PR12478, allowing Clang to be used when building
a ChromiumOS image without polluting the image with libraries from the
host system.
llvm-svn: 154176
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
On Darwin the architecture and the corresponding Mach-O slice is typically
specified with -arch. If not, it defaults to the current host architecture.
Do not use -mcpu to override the -arch value. This is only an issue when
people need to use specialized code for a non-default CPU (hopefully guarded
by run-time checks to detect the current processor). The -mcpu option is
still used for the -target-cpu option to clang, but this patch causes it to
not be used to set the architecture in the target triple.
llvm-svn: 153197
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
Previously, only diagnostics thrown by the cc1 process were
actually honoring the diagnostic options given on the command line,
like -Werror.
Reuse the existing code in Frontend currently used for cc1,
adjusting it to not interpret -Wl, linker flags as warnings.
Also fix a faulty test exposed by this change.
It wasn't actually testing anything, and was giving this warning:
clang-3: warning: argument unused during compilation: '-verify'
Which -Werror didn't turn into an error because it was output
by the driver, not the cc1 process, and diagnostic options
weren't parsed by the driver. And you couldn't see the warning
when running the test suite.
Fixes PR12181.
Patch by Dylan Noblesmith <nobled@dreamwidth.org>.
llvm-svn: 152660
FYI,
On VS10, %INCLUDE% contains;
(VS10)\VC\INCLUDE
(VS10)\VC\ATLMFC\INCLUDE
(SDK70A)\include
On VS11,
(VS11)\VC\INCLUDE
(VS11)\VC\ATLMFC\INCLUDE
(SDK80)\include\shared
(SDK80)\include\um
(SDK80)\include\winrt
FIXME: It may be enabled also on mingw.
llvm-svn: 152589
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
Debian multiarch libraries, this should in theory add support for those
platform's header search rules. I don't have a system to check this
with, so review appreciated. I've added the corresponding tests
referring to the debian multiarch tree.
We are starting to have a relatively completely tested Linux platform
for header search and library search, with several interesting
peculiarities. We should point people at the debian_multiarch_tree when
suggesting new tests. Folks with Debian systems that can check this for
correctness, it would be much appreciated. The missing chunks I know of
are testing bi-arch peudo-cross-compiling toolchains betwen 32-bit and
64-bit variants of platforms, and the MIPS and ARM Debian toolchains.
llvm-svn: 151484
Patch from Michel Dänzer, sent our way via Jeremy Huddleston who added
64-bit support. I just added one other place where powerpc64-linux-gnu
was missing (we only had powerpc64-unknown-linux-gnu).
I've also added a tree to test out the debian multiarch stuff. I don't
use debian regularly, so I'm not certain this is entirely accurate. If
anyone wants to check it against a debian system and fix any
inaccuracies, fire away. This way at least folks can see how this is
*supposed* to be tested.
It'd be particularly good to get the Debian MIPS toolchains tested in
this way.
llvm-svn: 151482
by -target and similar options. As discussed in PR 12026, the change
broke support for target-prefixed tools, i.e. calling x86_64--linux-ld
when compiling for x86_64--linux. Improve the test cases added
originally in r149083 to not require execution, just executable files.
Document the hack with appropiate FIXME comments.
llvm-svn: 151185
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
And remove HAVE_CLANG_CONFIG_H, now that the header is generated
in the autoconf build, too.
Reverts r149571/restores r149504, now that config.h is generated
correctly by LLVM's configure in all build configurations.
llvm-svn: 150487
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
This was from way-back-when (r82583) when Clang's C++ support wasn't prime-time
yet. Production quality C++ was tested experimentally from r100119 and turned
on by default in r141063.
Patch by Justin Bogner.
llvm-svn: 150148
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
that just uses the new toolchain probing logic. This fixes linking with -m32 on
64 bit systems (the /32 dir was not being added to the search).
llvm-svn: 149652
And remove HAVE_CLANG_CONFIG_H, now that the header is generated
in the autoconf build, too. (clang r149497 / llvm r149498)
Also include the config.h header after all other headers, per
the LLVM coding standards.
It also turns out WindowsToolChain.cpp wasn't using the config
header at all, so that include's just deleted now.
llvm-svn: 149504
The Darwin toolchain constructor was assuming that all Darwin triples would
have an OS string starting with "darwin". Triples starting with "macosx"
would misinterpret the version number, and "ios" triples would completely
miss the version number (or worse) because the OS name is not 6 characters
long. We lose some sanity checking of triple strings here, since the
Triple.getOSVersion function doesn't do all the checking that the previous
code did, but this still seems like a step in the right direction.
llvm-svn: 149422
driver based on discussions with Doug Gregor. There are several issues:
1) The patch was not reviewed prior to commit and there were review comments.
2) The design of the functionality (triple-prefixed tool invocation)
isn't the design we want for Clang going forward: it focuses on the
"user triple" rather than on the "toolchain triple", and forces that
bit of state into the API of every single toolchain instead of
handling it automatically in the common base classes.
3) The tests provided are not stable. They fail on a few Linux variants
(Gentoo among them) and on mingw32 and some other environments.
I *am* interested in the Clang driver being able to invoke
triple-prefixed tools, but we need to design that feature the right way.
This patch just extends the previous hack without fixing the underlying
problems with it. I'm working on a new design for this that I will mail
for review by tomorrow.
I am aware that this removes functionality that NetBSD relies on, but
this is ToT, not a release. This functionality hasn't been properly
designed, implemented, and tested yet. We can't "regress" until we get
something that really works, both with the immediate use cases and with
long term maintenance of the Clang driver.
For reference, the original commit log:
Keep track of the original target the user specified before
normalization. This used to be captured in DefaultTargetTriple and is
used for the (optional) $triple-$tool lookup for cross-compilation.
Do this properly by making it an attribute of the toolchain and use it
in combination with the computed triple as index for the toolchain
lookup.
llvm-svn: 149337
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
normalization. This used to be captured in DefaultTargetTriple and is
used for the (optional) $triple-$tool lookup for cross-compilation.
Do this properly by making it an attribute of the toolchain and use it
in combination with the computed triple as index for the toolchain
lookup.
llvm-svn: 149083
both actually tests what it wants to, doesn't have bogus and broken
assertions in it, and is also formatted much more cleanly and
consistently. Probably still some more that can be improved here, but
its much better.
Original commit message:
----
Try to unbreak the FreeBSD toolchain's detection of 32-bit targets
inside a 64-bit freebsd machine with the 32-bit compatibility layer
installed. The FreeBSD image always has the /usr/lib32 directory, so
test for the more concrete existence of crt1.o. Also enhance the tests
for freebsd to clarify what these trees look like and exercise the new
code.
Thanks to all the FreeBSD folks for helping me understand what caused
the failure and how we might fix it. =] That helps a lot. Also, yay
build bots.
llvm-svn: 149011
Original log:
Author: chandlerc <chandlerc@91177308-0d34-0410-b5e6-96231b3b80d8>
Date: Wed Jan 25 21:32:31 2012 +0000
Try to unbreak the FreeBSD toolchain's detection of 32-bit targets
inside a 64-bit freebsd machine with the 32-bit compatibility layer
installed. The FreeBSD image always has the /usr/lib32 directory, so
test for the more concrete existence of crt1.o. Also enhance the tests
for freebsd to clarify what these trees look like and exercise the new
code.
Thanks to all the FreeBSD folks for helping me understand what caused
the failure and how we might fix it. =] That helps a lot. Also, yay
build bots.
llvm-svn: 148993
inside a 64-bit freebsd machine with the 32-bit compatibility layer
installed. The FreeBSD image always has the /usr/lib32 directory, so
test for the more concrete existence of crt1.o. Also enhance the tests
for freebsd to clarify what these trees look like and exercise the new
code.
Thanks to all the FreeBSD folks for helping me understand what caused
the failure and how we might fix it. =] That helps a lot. Also, yay
build bots.
llvm-svn: 148981
Linux toolchain selection -- sorry folks. =] This should fix the Hexagon
toolchain.
However, I would point out that I see why my testing didn't catch this
-- we have no tests for Hexagon. ;]
llvm-svn: 148977
to suit the FreeBSD folks. Take them back to something closer to the old
behavior. We test whether the /usr/lib32 directory exists (within the
SysRoot), and use it if so, otherwise use /usr/lib.
FreeBSD folks, let me know if this causes any problems, or if you have
further tweaks.
llvm-svn: 148953
gross hack to provide it from my previous patch removing HostInfo. This
was enshrining (and hiding from my searches) the concept of storing and
diff-ing the host and target triples. We don't have the host triple
reliably available, so we need to merely inspect the target system. I've
changed the logic in selecting library search paths for NetBSD to match
what I provided for FreeBSD -- we include both search paths, but put the
32-bit-on-64-bit-host path first so it trumps.
NetBSD maintainers, you may want to tweak this, or feel free to ask me
to tweak it. I've left a FIXME here about the challeng I see in fixing
this properly.
llvm-svn: 148952
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
a HostInfo reference. Nothing about the HostInfo was used by any
toolchain except digging out the driver from it. This just makes that
a lot more direct. The change was accomplished entirely mechanically.
It's one step closer to removing the shim full of buggy copy/paste code
that is HostInfo.
llvm-svn: 148945
helped stage the refactoring of things a bit, but really isn't the right
place for it. The driver may be responsible for compilations with many
different targets. In those cases, having a target triple in the driver
is actively misleading because for many of those compilations that is
not actually the triple being targeted.
This moves the last remaining users of the Driver's target triple to
instead use the ToolChain's target triple. The toolchain has a single,
concrete target it operates over, making this a more stable and natural
home for it.
llvm-svn: 148942
adding search paths. Add them only when they exist, and prefix the paths
with the sysroot. This will allow targeting a FreeBSD sysroot on
a non-FreeBSD host machine, and perhaps more importantly should allow
testing the FreeBSD driver's behavior similarly to the Linux tests with
a fake tree of files in the regression test suite.
I don't have FreeBSD systems handy to build up the list of files that
should be used here, but this is the basic functionality and I'm hoping
Roman or someone from the community can contribute the actual test
cases.
llvm-svn: 148940
search paths for 32-bit targets. This avoids having to detect which is
expected for the target system, and the linker should DTRT, and take the
32-bit libraries from the first one when applicable. Thanks to Roman
Divacky for sanity checking this.
llvm-svn: 148939
The fundamental shift here is to stop making *any* assumptions about the
*host* triple. Where these assumptions you ask? Why, they were in one of
the two target triples referenced of course. This was the single biggest
place where the previously named "host triple" was actually used as
such. ;] The reason we were reasoning about the host is in order to
detect the use of '-m32' or '-m64' flags to change the target. These
flags shift the default target only slightly, which typically means
a slight deviation from the host. When using these flags, the GCC
installation is under a different triple from the one actually targeted
in the compilation, and we used the host triple to find it.
Too bad that wasn't even correct. Consider an x86 Linux host which has
a PPC64 cross-compiling GCC toolchain installed. This toolchain is also
configured for multiarch compiling and can target PPC32 with eth '-m32'
flag. When targeting 'powerpc-linux-gnu' or some other PPC32 triple, we
have to look for the PPC64 variant of the triple to find the GCC
install, and that triple is neither the host nor target.
The new logic computes the multiarch's alternate triple from the target
triple, and looks under both sides. It also looks more aggressively for
the correct subdirectory of the GCC installation, and exposes the
subdirectory in a nice programmatic way. This '/32' or '/64' suffix is
something we can reuse in many other parts of the toolchain.
An important note -- while this likely fixes a large category of
cross-compile use cases, that's not my primary goal, and I've not done
testing (or added test cases) for scenarios that may now work. If
someone else wants to try more interesting PPC cross compiles, I'd love
to have reports. But my focus is on factoring away the references to the
"host" triple. The refactoring is my goal, and so I'm mostly relying on
the existing (pretty good) test coverage we have here.
Future patches will leverage this new functionality to factor out more
and more of the toolchain's triple manipulation.
llvm-svn: 148935
of the target triple to stand in for the "host" triple.
Thanks to a great conversation with Richard Smith, I'm now much more
confident in how this is proceeding. In all of the places where we
currently reason about the "host" architecture or triple, what we really
want to reason about in the detected GCC installation architecture or
triple, and the ways in which that differs from the target. When we find
a GCC installation with a different triple from our target *but capable
of targeting our target* through an option such as '-m64', we want to
detect *that* case and change the paths within the GCC installation (and
libstdc++ installation) to reflect this difference.
This patch makes one function do this correctly. Subsequent commits will
hoist the logic used here into the GCCInstallation utility, and then
reuse it through the rest of the toolchains to fix the remaining places
where this is currently happening.
llvm-svn: 148852
inside of GCCInstallation to be a proper llvm::Triple. This is still
a touch ugly because we have to use it as a string in so many places,
but I think on the whole the more structured representation is better.
Comments of course welcome if this tradeoff isn't working for folks.
llvm-svn: 148843
function. The logic for this, and I want to emphasize that this is the
logic for computing the *target* triple, is currently scattered
throughout various different HostInfo classes ToolChain factoring
functions. Best part, it is largely *duplicated* there. The goal is to
hoist all of that up to here where we can deal with it once, and in
a consistent manner.
Unfortunately, this uncovers more fun problems: the ToolChains assume
that the *actual* target triple is the one passed into them by these
factory functions, while the *host* triple is the one in the driver.
This already was a lie, and a damn lie, when the '-target' flag was
specified. It only really worked when the difference stemmed from '-m32'
and '-m64' flags. I'll have to fix that (and remove all the FIXMEs I've
introduced here to document the problem) before I can finish hoisting
the target-calculation logic.
It's bugs all the way down today it seems...
llvm-svn: 148839
inside the innards of the Driver implementation, and only ever
implemented to return 'true' for the Darwin OSes. Instead use a more
direct query on the target triple and a comment to document why the
target matters here.
If anyone is worried about this predicate getting wider use or improper
use, I can make it a local or private predicate in the driver.
llvm-svn: 148797
The Driver has a fixed target, whether we like it or not, the
DefaultTargetTriple is not a default. This at least makes things more
honest. I'll eventually get rid of most (if not all) of
DefaultTargetTriple with this proper triple object. Bit of a WIP.
llvm-svn: 148796
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
especially nice as the Windows toolchain needs the windows header files,
and has lots of platform specific hooks in it.
To facilitate the split, hoist a bunch of file-level static helpers into
class-level static helpers. Spiff up their doxygen comments while there
as they're now more likely to be looked up via docs.
Hopefully, this will be followed by further breaking apart of the
toolchain definitions. Most of the large and complex ones should likely
live on their own. I'm looking at you Darwin. ;]
llvm-svn: 146840
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
clang/lib/Driver/Driver.cpp: Don't pass through negative exit status, or parent would be confused.
llvm::sys::Program::Wait(): Suppose 0x8000XXXX and 0xC000XXXX as abnormal exit code and pass it as negative value.
Win32 Exception Handler: Exit with ExceptionCode on an unhandle exception.
llvm-svn: 145389
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
- With the current implementation of sys::Program this always printed "2".
- The command execution code will output the right number anyway (including the signal name).
llvm-svn: 144993
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
version of Ubuntu. It has a very broken multiarch configuration, and so
we need special logic to handle it correctly. Fixing and testing this
uncovered a few other trivial issues with the logic that are fixed as
well.
I added tests to cover this as it is hard to notice if you install
recent versions of the OS.
llvm-svn: 144165
useful when using Clang as a system-compiler, but its harmless. When
using Clang as a cross-compiler, this can be very handy as quite a few
toolchains ship their libc headers here rather than under
'/usr/include'.
For reference, this is the beginning of my work to also make the Clang
driver more suitable as a cross-compiler.
llvm-svn: 144089
We don't actually need a separate flag for non-sysrooted paths as the
driver has to manage the sysroot anyways. The driver is not infrequently
adding paths to the header search based on their existence on the
filesystem. For that, it has to add the sysroot anyways, we should pass
it on down to CC1 already joined. More importantly, the driver cannot in
all cases distinguish between sysrooted paths and paths that are
relative to the Clang binary's installation directory. Essentially, we
always need to ignore the system root for these internal header search
options. It turns out in most of the places we were already providing
the system root in the driver, and then another one in CC1 so this fixes
several bugs.
llvm-svn: 143917
toolchain. The logic is mostly generic already, and where possible
should be made more generic. Also, it has no impact other than to expose
a set of methods which each toolchain can then query to setup their
desired configuration. These should be available to toolchains beyond
just Linux.
llvm-svn: 143899
detection system that is providing the library paths and crt object
files.
This, modulo any bugs that need to be shaken out, resolves numerous bugs
with how we handle header paths. Here are a few that I know of:
- We no longer need to enumerate all GCC versions searched.
- OpenSUSE searched GCC versions in the wrong order.
- There were typos when selecting various patterns, etc.
- We aren't stating quite some many directories now.
- SysRoot didn't always work in a reasonable way.
I'm working on tests for this, but the tests are making me and Lit sad.
The real testing for this type of driver change is to try it out on
various distributions. I'll hit the common ones right away, and start
more thorough testing tomorrow after some sleep.
llvm-svn: 143874
the detected GCC installation. This allows us to expose another aspect
of what we detected: the GCC version. This will be used shortly.
llvm-svn: 143871
David Chisnall