Summary:
Use TSC emulation in cases where RDTSCP isn't available on the host
running an XRay instrumented binary. We can then fall back into
emulation instead of not even installing XRay's runtime functionality.
We only do this for now in the naive/basic logging implementation, but
should be useful in even FDR mode.
Should fix http://llvm.org/PR32148.
Reviewers: pelikan, rnk, sdardis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30677
llvm-svn: 297800
The tests for libc++ specify -target on the command-line to the
compiler, but this is problematic for a few reasons.
Firstly, the -target option isn't supported on Apple platforms. Parts
of the triple get dropped and ignored. Instead, software should be
compiled with a combination of the -arch and -m<name>-version-min
options.
Secondly, the generic "darwin" target references a kernel version
instead of a platform version. Each platform has its own independent
versions (with different versions of libc++.1.dylib), independent of the
version of the Darwin kernel.
This commit adds support to the LIT infrastructure for testing against
Apple platforms using -arch and -platform options.
If the host is not on OS X, or the compiler type is not clang or apple-clang, then this commit has NFC.
If the host is on OS X and --param=target_triple=... is specified, then a warning is emitted to use arch and platform instead. Besides the warning, there's NFC.
If the host is on OS X and *no* target-triple is specified, then use the new deployment target logic. This uses two new lit parameters, --param=arch=<arch> and --param=platform=<platform>. <platform> has the form <name>[<version>].
By default, arch is auto-detected from clang -dumpmachine, and platform is "macosx".
If the platform doesn't have a version:
For "macosx", the version is auto-detected from the host system using sw_vers. This may give a different version than the SDK, since new SDKs can be installed on older hosts.
Otherwise, the version is auto-detected from the SDK version using xcrun --show-sdk-path.
-arch <arch> -m<name>-version-min=<version> is added to the compiler flags.
The target triple is computed as <arch>-apple-<platform>. It is *not* passed to clang, but it is available for XFAIL and UNSUPPORTED (as is with_system_cxx_lib=<target>).
For convenience, apple-darwin and <arch>-apple-darwin are added to the set of available features.
There were a number of tests marked to XFAIL on x86_64-apple-darwin11
and x86_64-apple-darwin12. I updated these to
x86_64-apple-macosx10.7 and x86_64-apple-macosx10.8.
llvm-svn: 297798
Summary:
<1 x Ty> is not a legal vector type in LLT, we shouldn’t build G_MERGE_VALUES
instruction for them.
Reviewers: qcolombet, aditya_nandakumar, dsanders, t.p.northover, ab, javed.absar
Reviewed By: qcolombet
Subscribers: dberris, rovka, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D30948
llvm-svn: 297792
This fixes lookup mismatches that could happen when the module cache
path contained a '/./' component.
<rdar://problem/30413458>
Differential Revision: https://reviews.llvm.org/D30915
llvm-svn: 297790
Being passed -z notext is a pretty strong indication that the user is
OK with text relocations. This is not the same behavior as bfd, but
bfd defaults to -z notext, so it has to try to avoid text relocations
and use them as a last resort.
llvm-svn: 297789
This patch modifies the Get/Set Register Value/Context functions for Intel to not duplicate code for reading non-AVX registers. This is similar to other transformations I've been making to the AVX register handling code.
llvm-svn: 297787
Summary:
Adds a new kind of MachineOperand: MO_Placeholder.
This operand must not appear in the MIR and only exists as a way of
creating an 'uninitialized' operand until a matcher function overwrites it.
Depends on D30046, D29712
Reviewers: t.p.northover, ab, rovka, aditya_nandakumar, javed.absar, qcolombet
Reviewed By: qcolombet
Subscribers: dberris, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D30089
llvm-svn: 297782
Reduced version of D26357 - based on the discussion on llvm-dev about canonicalization of UMIN/UMAX/SMIN/SMAX as well as ABS I've reduced that patch to just the ABS ISD node (with x86/sse support) to improve basic combines and lowering.
ARM/AArch64, Hexagon, PowerPC and NVPTX all have similar instructions allowing us to make this a generic opcode and move away from the hard coded tablegen patterns which makes it tricky to match more complex patterns.
At the moment this patch doesn't attempt legalization as we only create an ABS node if its legal/custom.
Differential Revision: https://reviews.llvm.org/D29639
llvm-svn: 297780
Clangd is often waiting for input on getline as it awaits requests. If the
getline is interrupted, it causes the system call (read) to fail and the EINTR
error to be set. This can be seen when attaching a debugger such as LLDB on
macOS. On macOS (and possibly other operating systems), this system call is not
restarted after interruption but on Linux it is restarted which is why
attaching a debugger does work correctly there.
The solution is to work around the non-restarting system call by checking the
errno for EINTR when the stream fails and try again. This should be safe on all
Unixish platforms.
See also http://bugs.llvm.org/show_bug.cgi?id=32149 for some background
discussion.
Patch by Marc-Andre Laperle!
Differential Revision: https://reviews.llvm.org/D30675
llvm-svn: 297779
Previously we were using the encoded LEB hex values
for the value types. This change uses the decoded
negative value and the LEB encoder to write them out.
Differential Revision: https://reviews.llvm.org/D30847
Patch by Sam Clegg
llvm-svn: 297777
This patch adds support to the test suite for overriding the path to debugserver, and uses the override to point to the build tree's debugserver on Darwin.
llvm-svn: 297776
On Solaris ld (and some other tools that use the underlying utility
libraries, such as elfdump) chokes on an archive library that has no
symbol table. The Solaris tools always create one, even if it's empty.
That bug has been fixed in the latest development line, and can
probably be backported to a supported release, but it would be nice if
LLVM's archiver could emit the empty symbol table, too.
Patch by Danek Duvall!
llvm-svn: 297773
Since bitcode uses VBR encoding, large numbers are more expensive than
small ones. Instead of emitting a UINT_MAX sentinel after each sequence
of state-change pairs, emit the size of the sequence as a prefix.
This should have no functionality change besides saving bits from the
encoding.
llvm-svn: 297770
Make MCSectionELF::AssociatedSection be a link to a symbol, because
that's how it works in the assembly, and use it in the asm printer.
llvm-svn: 297769
check for the existence of RTLD_DEEPBIND, since this constant is only
supported for glibc >= 2.3.4. This fixes builds for FreeBSD and other
platforms that do not have RTLD_DEEPBIND.
llvm-svn: 297763
This adds -Wbitfield-enum-conversion, which warns on implicit
conversions that happen on bitfield assignment that change the value of
some enumerators.
Values of enum type typically take on a very small range of values, so
they are frequently stored in bitfields. Unfortunately, there is no
convenient way to calculate the minimum number of bits necessary to
store all possible values at compile time, so users usually hard code a
bitwidth that works today and widen it as necessary to pass basic
testing and validation. This is very error-prone, and leads to stale
widths as enums grow. This warning aims to catch such bugs.
This would have found two real bugs in clang and two instances of
questionable code. See r297680 and r297654 for the full description of
the issues.
This warning is currently disabled by default while we investigate its
usefulness outside of LLVM.
The major cause of false positives with this warning is this kind of
enum:
enum E { W, X, Y, Z, SENTINEL_LAST };
The last enumerator is an invalid value used to validate inputs or size
an array. Depending on the prevalance of this style of enum across a
codebase, this warning may be more or less feasible to deploy. It also
has trouble on sentinel values such as ~0U.
Reviewers: rsmith, rtrieu, thakis
Reviewed By: thakis
Subscribers: hfinkel, voskresensky.vladimir, sashab, cfe-commits
Differential Revision: https://reviews.llvm.org/D30923
llvm-svn: 297761
symbols dynamically
The CoreFoundation symbols uses by __isOSVersionAtLeast should be loaded at
runtime to ensure that the programs that don't use @available won't have to be
linked to CoreFoundation.
The Clang frontend IRGen library will need to emit a CoreFoundation symbol
when @available is used to ensure that programs that actually use @available
are linked to CoreFoundation.
rdar://31039554
llvm-svn: 297760
Summary:
In SamplePGO, if the profile is collected from non-LTO binary, and used to drive ThinLTO, the indirect call promotion may fail because ThinLTO adjusts local function names to avoid conflicts. There are two places of where the mismatch can happen:
1. thin-link prepends SourceFileName to front of FuncName to build the GUID (GlobalValue::getGlobalIdentifier). Unlike instrumentation FDO, SamplePGO does not use the PGOFuncName scheme and therefore the indirect call target profile data contains a hash of the OriginalName.
2. backend compiler promotes some local functions to global and appends .llvm.{$ModuleHash} to the end of the FuncName to derive PromotedFunctionName
This patch tries at the best effort to find the GUID from the original local function name (in profile), and use that in ICP promotion, and in SamplePGO matching that happens in the backend after importing/inlining:
1. in thin-link, it builds the map from OriginalName to GUID so that when thin-link reads in indirect call target profile (represented by OriginalName), it knows which GUID to import.
2. in backend compiler, if sample profile reader cannot find a profile match for PromotedFunctionName, it will try to find if there is a match for OriginalFunctionName.
3. in backend compiler, we build symbol table entry for OriginalFunctionName and pointer to the same symbol of PromotedFunctionName, so that ICP can find the correct target to promote.
Reviewers: mehdi_amini, tejohnson
Reviewed By: tejohnson
Subscribers: llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D30754
llvm-svn: 297757
Summary: Fix the test by adding missing -target flags with a 'linux' triple.
Reviewers: rnk, srhines
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D30947
llvm-svn: 297754
Summary:
This patch adds -f[no-]rtlib-add-rpath, which if enabled, embeds the
arch-specific subdirectory in resource directory using -rpath (instead
of doing so only during native compilation).
This patch also re-enables test arch-specific-libdir.c which was
silently unsupported because of the REQUIRES tag 'linux'.
Reviewers: bkramer, rnk, mgorny
Subscribers: srhines, cfe-commits
Differential Revision: https://reviews.llvm.org/D30700
llvm-svn: 297751
This is a follow-up to r297700 (Add a nullability sanitizer).
It addresses some FIXME's re: using nullability-specific diagnostic
handlers from compiler-rt, now that the necessary handlers exist.
check-ubsan test updates to follow.
llvm-svn: 297750
Dean Michael Berris