Summary:
FreeBSD N64 MIPS systems can include 32-bit libraries for O32 in
/usr/lib32 similar to the 32-bit compatibility libraries provided
for FreeBSD/amd64 and FreeBSD/powerpc64.
Reviewers: dim
Reviewed By: dim
Differential Revision: https://reviews.llvm.org/D42972
llvm-svn: 324948
As a first step, pass '-c/--compile-only' to ptxas so that it
doesn't complain about references to external function. This
will successfully generate object files, but they won't work
at runtime because the registration routines need to adapted.
Differential Revision: https://reviews.llvm.org/D42921
llvm-svn: 324878
The difference from the previous try is that we no longer directly
access function declarations from position independent executables. It
should work, but currently doesn't with some linkers.
It now includes a fix to not mark available_externally definitions as
dso_local.
Original message:
Start setting dso_local in clang.
This starts adding dso_local to clang.
The hope is to eventually have TargetMachine::shouldAssumeDsoLocal go
away. My objective for now is to move enough of it to clang to remove
the need for the TargetMachine one to handle PIE copy relocations and
-fno-plt. With that it should then be easy to implement a
-fno-copy-reloc in clang.
This patch just adds the cases where we assume a symbol to be local
based on the file being compiled for an executable or a shared
library.
llvm-svn: 324535
This reverts commit r324500.
The bots found two failures:
ThreadSanitizer-x86_64 :: Linux/pie_no_aslr.cc
ThreadSanitizer-x86_64 :: pie_test.cc
when using gold. The issue is a limitation in gold when building pie
binaries. I will investigate how to work around it.
llvm-svn: 324505
It now includes a fix to not mark available_externally definitions as
dso_local.
Original message:
Start setting dso_local in clang.
This starts adding dso_local to clang.
The hope is to eventually have TargetMachine::shouldAssumeDsoLocal go
away. My objective for now is to move enough of it to clang to remove
the need for the TargetMachine one to handle PIE copy relocations and
-fno-plt. With that it should then be easy to implement a
-fno-copy-reloc in clang.
This patch just adds the cases where we assume a symbol to be local
based on the file being compiled for an executable or a shared
library.
llvm-svn: 324500
Summary:
Currently, assertion-disabled Clang builds emit value names when generating LLVM IR. This is controlled by the `NDEBUG` macro, and is not easily overridable. In order to get IR output containing names from a release build of Clang, the user must manually construct the CC1 invocation w/o the `-discard-value-names` option. This is less than ideal.
For example, Godbolt uses a release build of Clang, and so when asked to emit LLVM IR the result lacks names, making it harder to read. Manually invoking CC1 on Compiler Explorer is not feasible.
This patch adds the driver options `-fdiscard-value-names` and `-fno-discard-value-names` which allow the user to override the default behavior. If neither is specified, the old behavior remains.
Reviewers: erichkeane, aaron.ballman, lebedev.ri
Reviewed By: aaron.ballman
Subscribers: bogner, cfe-commits
Differential Revision: https://reviews.llvm.org/D42887
llvm-svn: 324498
This starts adding dso_local to clang.
The hope is to eventually have TargetMachine::shouldAssumeDsoLocal go
away. My objective for now is to move enough of it to clang to remove
the need for the TargetMachine one to handle PIE copy relocations and
-fno-plt. With that it should then be easy to implement a
-fno-copy-reloc in clang.
This patch just adds the cases where we assume a symbol to be local
based on the file being compiled for an executable or a shared
library.
llvm-svn: 324107
If the CUDA toolkit is not installed to its default locations
in /usr/local/cuda, the user is forced to specify --cuda-path.
This is tedious and the driver can be smarter if well-known tools
(like ptxas) can already be found in the PATH environment variable.
Add option --cuda-path-ignore-env if the user wants to ignore
set environment variables. Also use it in the tests to make sure
the driver always finds the same CUDA installation, regardless
of the user's environment.
Differential Revision: https://reviews.llvm.org/D42642
llvm-svn: 323848
For /arch:AVX512F:
clang-cl and cl.exe both defines __AVX512F__ __AVX512CD__.
clang-cl also defines __AVX512ER__ __AVX512PF__.
64-bit cl.exe also defines (according to /Bz) _NO_PREFETCHW.
For /arch:AVX512:
clang-cl and cl.exe both define
__AVX512F__ __AVX512CD__ __AVX512BW__ __AVX512DQ__ __AVX512VL__.
64-bit cl.exe also defines _NO_PREFETCHW.
So not 100% identical, but pretty close.
Also refactor the existing AVX / AVX2 code to not repeat itself in both the
32-bit and 64-bit cases.
https://reviews.llvm.org/D42538
llvm-svn: 323433
r213083 initially implemented /arch: support by mapping it to CPU features.
Then r241077 additionally mapped it to CPU, which made the feature flags
redundant (if harmless). This change here removes the redundant mapping to
feature flags, and rewrites test/Driver/cl-x86-flags.c to be a bit more of an
integration test that checks for preprocessor defines like AVX (like documented
on MSDN) instead of for driver flags.
To keep emitting warn_drv_unused_argument, use getLastArgNoClaim() followed by an explicit claim() if needed.
This is in preparation for adding support for /arch:AVX512(F).
No intended behavior change.
https://reviews.llvm.org/D42497
llvm-svn: 323426
r317337 missed that scudo is supported on MIPS32, so permit that option for
MIPS32.
Reviewers: cryptoad, atanasyan
Differential Revision: https://reviews.llvm.org/D42416
llvm-svn: 323412
NVPTX does not have runtime support necessary for profiling to work
and even call arc collection is prohibitively expensive. Furthermore,
there's no easy way to collect the samples. NVPTX also does not
support global constructors that clang generates if sample/arc collection
is enabled.
Differential Revision: https://reviews.llvm.org/D42452
llvm-svn: 323345
- Test needs to be able to handle "clang.exe" on Windows
- Test needs to be able to handle either '/' or '\\' used as the path separator
Reviewed by Paul Robinson
llvm-svn: 323285
The tests are targeting Windows but do not specify an environment. When
executed on Linux, they would use an ELF output rather than the COFF
output. Explicitly provide an environment.
llvm-svn: 323225
Summary:
General idea is to utilize generic (mostly Generic_GCC) code
and get rid of Solaris-specific handling as much as possible.
In particular:
- scanLibDirForGCCTripleSolaris was removed, relying on generic
CollectLibDirsAndTriples
- findBiarchMultilibs is now properly utilized to switch between
m32 and m64 include & lib paths on Solaris
- C system include handling copied from Linux (bar multilib hacks)
Fixes PR24606.
Reviewers: dlj, rafael, jyknight, theraven, tstellar
Reviewed By: jyknight
Subscribers: aaron.ballman, mgorny, krytarowski, ro, joerg, cfe-commits
Differential Revision: https://reviews.llvm.org/D35755
llvm-svn: 323193
Summary:
First, we need to explain the core of the vulnerability. Note that this
is a very incomplete description, please see the Project Zero blog post
for details:
https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html
The basis for branch target injection is to direct speculative execution
of the processor to some "gadget" of executable code by poisoning the
prediction of indirect branches with the address of that gadget. The
gadget in turn contains an operation that provides a side channel for
reading data. Most commonly, this will look like a load of secret data
followed by a branch on the loaded value and then a load of some
predictable cache line. The attacker then uses timing of the processors
cache to determine which direction the branch took *in the speculative
execution*, and in turn what one bit of the loaded value was. Due to the
nature of these timing side channels and the branch predictor on Intel
processors, this allows an attacker to leak data only accessible to
a privileged domain (like the kernel) back into an unprivileged domain.
The goal is simple: avoid generating code which contains an indirect
branch that could have its prediction poisoned by an attacker. In many
cases, the compiler can simply use directed conditional branches and
a small search tree. LLVM already has support for lowering switches in
this way and the first step of this patch is to disable jump-table
lowering of switches and introduce a pass to rewrite explicit indirectbr
sequences into a switch over integers.
However, there is no fully general alternative to indirect calls. We
introduce a new construct we call a "retpoline" to implement indirect
calls in a non-speculatable way. It can be thought of loosely as
a trampoline for indirect calls which uses the RET instruction on x86.
Further, we arrange for a specific call->ret sequence which ensures the
processor predicts the return to go to a controlled, known location. The
retpoline then "smashes" the return address pushed onto the stack by the
call with the desired target of the original indirect call. The result
is a predicted return to the next instruction after a call (which can be
used to trap speculative execution within an infinite loop) and an
actual indirect branch to an arbitrary address.
On 64-bit x86 ABIs, this is especially easily done in the compiler by
using a guaranteed scratch register to pass the target into this device.
For 32-bit ABIs there isn't a guaranteed scratch register and so several
different retpoline variants are introduced to use a scratch register if
one is available in the calling convention and to otherwise use direct
stack push/pop sequences to pass the target address.
This "retpoline" mitigation is fully described in the following blog
post: https://support.google.com/faqs/answer/7625886
We also support a target feature that disables emission of the retpoline
thunk by the compiler to allow for custom thunks if users want them.
These are particularly useful in environments like kernels that
routinely do hot-patching on boot and want to hot-patch their thunk to
different code sequences. They can write this custom thunk and use
`-mretpoline-external-thunk` *in addition* to `-mretpoline`. In this
case, on x86-64 thu thunk names must be:
```
__llvm_external_retpoline_r11
```
or on 32-bit:
```
__llvm_external_retpoline_eax
__llvm_external_retpoline_ecx
__llvm_external_retpoline_edx
__llvm_external_retpoline_push
```
And the target of the retpoline is passed in the named register, or in
the case of the `push` suffix on the top of the stack via a `pushl`
instruction.
There is one other important source of indirect branches in x86 ELF
binaries: the PLT. These patches also include support for LLD to
generate PLT entries that perform a retpoline-style indirection.
The only other indirect branches remaining that we are aware of are from
precompiled runtimes (such as crt0.o and similar). The ones we have
found are not really attackable, and so we have not focused on them
here, but eventually these runtimes should also be replicated for
retpoline-ed configurations for completeness.
For kernels or other freestanding or fully static executables, the
compiler switch `-mretpoline` is sufficient to fully mitigate this
particular attack. For dynamic executables, you must compile *all*
libraries with `-mretpoline` and additionally link the dynamic
executable and all shared libraries with LLD and pass `-z retpolineplt`
(or use similar functionality from some other linker). We strongly
recommend also using `-z now` as non-lazy binding allows the
retpoline-mitigated PLT to be substantially smaller.
When manually apply similar transformations to `-mretpoline` to the
Linux kernel we observed very small performance hits to applications
running typical workloads, and relatively minor hits (approximately 2%)
even for extremely syscall-heavy applications. This is largely due to
the small number of indirect branches that occur in performance
sensitive paths of the kernel.
When using these patches on statically linked applications, especially
C++ applications, you should expect to see a much more dramatic
performance hit. For microbenchmarks that are switch, indirect-, or
virtual-call heavy we have seen overheads ranging from 10% to 50%.
However, real-world workloads exhibit substantially lower performance
impact. Notably, techniques such as PGO and ThinLTO dramatically reduce
the impact of hot indirect calls (by speculatively promoting them to
direct calls) and allow optimized search trees to be used to lower
switches. If you need to deploy these techniques in C++ applications, we
*strongly* recommend that you ensure all hot call targets are statically
linked (avoiding PLT indirection) and use both PGO and ThinLTO. Well
tuned servers using all of these techniques saw 5% - 10% overhead from
the use of retpoline.
We will add detailed documentation covering these components in
subsequent patches, but wanted to make the core functionality available
as soon as possible. Happy for more code review, but we'd really like to
get these patches landed and backported ASAP for obvious reasons. We're
planning to backport this to both 6.0 and 5.0 release streams and get
a 5.0 release with just this cherry picked ASAP for distros and vendors.
This patch is the work of a number of people over the past month: Eric, Reid,
Rui, and myself. I'm mailing it out as a single commit due to the time
sensitive nature of landing this and the need to backport it. Huge thanks to
everyone who helped out here, and everyone at Intel who helped out in
discussions about how to craft this. Also, credit goes to Paul Turner (at
Google, but not an LLVM contributor) for much of the underlying retpoline
design.
Reviewers: echristo, rnk, ruiu, craig.topper, DavidKreitzer
Subscribers: sanjoy, emaste, mcrosier, mgorny, mehdi_amini, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D41723
llvm-svn: 323155
When using -fno-integrated-as flag, the gnu assembler produces code
with some default march/mabi which later causes linker failure due
to incompatible mabi/march.
In this patch we explicitly propagate -mabi and -march flags to the
GNU assembler.
In this patch we explicitly propagate -mabi and -march flags to the GNU assembler.
Differential Revision: https://reviews.llvm.org/D41271
llvm-svn: 322769
Summary:
There are only two valid integrated Clang driver tools: `-cc1` and
`-cc1as`. If a user asks for an unknown tool, such as `-cc1asphalt`,
an error message is displayed to indicate that there is no such tool,
but the message doesn't indicate what the valid options are.
Include the valid options in the error message.
Test Plan: `check-clang`
Reviewers: sepavloff, bkramer, phosek
Reviewed By: bkramer
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D42004
llvm-svn: 322517
RISCVABIInfo is implemented in terms of XLen, supporting both RV32 and RV64.
Unfortunately we need to count argument registers in the frontend in order to
determine when to emit signext and zeroext attributes. Integer scalars are
extended according to their type up to 32-bits and then sign-extended to XLen
when passed in registers, but are anyext when passed on the stack. This patch
only implements the base integer (soft float) ABIs.
For more information on the RISC-V ABI, see [the ABI
doc](https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md),
my [golden model](https://github.com/lowRISC/riscv-calling-conv-model), and
the [LLVM RISC-V calling convention
patch](https://reviews.llvm.org/D39898#2d1595b4) (specifically the comment
documenting frontend expectations).
Differential Revision: https://reviews.llvm.org/D40023
llvm-svn: 322494
Summary:
In https://reviews.llvm.org/D41733, the driver was modified such that,
when a user provided a mispelled option such as `-hel`, it would
suggest a valid option with a nearby edit distance: "did you mean
'-help'?".
Add these suggestions to invocations of `clang -cc1as` as well.
Test Plan: `check-clang`
Reviewers: v.g.vassilev, bruno
Reviewed By: v.g.vassilev
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D42001
llvm-svn: 322445
Petr Hosek reported an external buildbot was failing on riscv32-toolchain.c,
seemingly as it set CLANG_DEFAULT_LINKER to lld. Address this by explicitly
setting -fuse-ld=ld in the tests.
llvm-svn: 322435
Referenced implementation from Fuchsia and Darwin Toolchain.
Still only support CST_Libcxx. Now checks that the argument
is really '-stdlib=libc++', and display error.
Also, now will pass -lc++ and -lc++abi to the linker.
Patch by Patrick Cheng!
Differential Revision: https://reviews.llvm.org/D41937
llvm-svn: 322382
We were seeing test failures of riscv32-toolchain.c on windows due to the \
path separator being used for the linker. Add {{/|\\\\}} pattern (made
horrible due to escaping), just like introduced in r214931.
llvm-svn: 322286
The dummy crtbegin.o files were left out in r322276 (as they were ignored by
svn add of test/Driver/Inputs/multilib_riscv_linux_sdk) and are necessary for
the driver test to work.
llvm-svn: 322277
As RV64 codegen has not yet been upstreamed into LLVM, we focus on RV32 driver
support (RV64 to follow).
Differential Revision: https://reviews.llvm.org/D39963
llvm-svn: 322276
John Baldwin