[runtimes] remove workaround for old CMake when setting `--unwindlib=none` #93429

h-vetinari · 2024-05-26T22:43:19Z

Noticed a left-over for old CMake, irrelevant now that the baseline version moved to 3.20.

h-vetinari · 2024-05-26T23:11:21Z

Hm, this seems to affect some feature detection (and constraints) later on, though I don't see why/how removing the specification of --unwindlib=none to the driver (leaving it only to the linker) would cause that:

llvm-project/libunwind/src/CMakeLists.txt

Lines 99 to 106 in 1c046ca

    
           if (LIBUNWIND_ENABLE_SHARED AND 
        
               NOT (CXX_SUPPORTS_FNO_EXCEPTIONS_FLAG AND 
        
                    CXX_SUPPORTS_FUNWIND_TABLES_FLAG)) 
        
             message(FATAL_ERROR 
        
                     "Compiler doesn't support generation of unwind tables if exception " 
        
                     "support is disabled.  Building libunwind DSO with runtime dependency " 
        
                     "on C++ ABI library is not supported.") 
        
           endif()

ldionne · 2024-06-11T21:46:03Z

I think something else is going on. The CMake configuration as a whole is failing, the logs show:

-- Performing Test CXX_SUPPORTS_NOSTDLIBXX_FLAG - Failed
-- Performing Test CXX_SUPPORTS_NOSTDINCXX_FLAG
-- Performing Test CXX_SUPPORTS_NOSTDINCXX_FLAG - Failed

This shouldn't be affected by this patch, but somehow it is.

ldionne

BTW this LGTM if we can make it work -- this is a great cleanup, thanks for noticing.

h-vetinari · 2024-06-20T10:24:54Z

The CMake configuration as a whole is failing, the logs show:

The interesting thing is that everything past that check is failing, which somehow looks like it's corrupting some (CMake's?) state somehow:

-- Performing Test CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG
-- Performing Test CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG - Success
-- Performing Test CXX_SUPPORTS_NOSTDLIBXX_FLAG
-- Performing Test CXX_SUPPORTS_NOSTDLIBXX_FLAG - Failed
-- Performing Test CXX_SUPPORTS_NOSTDINCXX_FLAG
-- Performing Test CXX_SUPPORTS_NOSTDINCXX_FLAG - Failed
-- [everything after that fails]

llvmbot · 2024-06-20T10:47:27Z

@llvm/pr-subscribers-libunwind
@llvm/pr-subscribers-offload
@llvm/pr-subscribers-libcxx

@llvm/pr-subscribers-clang

Author: None (h-vetinari)

Changes

Noticed a left-over for old CMake, irrelevant now that the baseline version moved to 3.20.

Full diff: https://github.com/llvm/llvm-project/pull/93429.diff

6 Files Affected:

(modified) clang/tools/driver/CMakeLists.txt (+2-2)
(modified) llvm/cmake/modules/AddLLVM.cmake (+2-2)
(modified) llvm/cmake/modules/HandleLLVMOptions.cmake (+3-3)
(modified) llvm/cmake/modules/HandleLLVMStdlib.cmake (+3-3)
(removed) llvm/cmake/modules/LLVMCheckLinkerFlag.cmake (-28)
(modified) runtimes/CMakeLists.txt (+1-21)

diff --git a/clang/tools/driver/CMakeLists.txt b/clang/tools/driver/CMakeLists.txt
index 290bf2a42536d..018605c2fd4f2 100644
--- a/clang/tools/driver/CMakeLists.txt
+++ b/clang/tools/driver/CMakeLists.txt
@@ -107,7 +107,7 @@ endif()
 
 if(CLANG_ORDER_FILE AND
     (LLVM_LINKER_IS_APPLE OR LLVM_LINKER_IS_GOLD OR LLVM_LINKER_IS_LLD))
-  include(LLVMCheckLinkerFlag)
+  include(CheckLinkerFlag)
 
   if (LLVM_LINKER_IS_APPLE OR (LLVM_LINKER_IS_LLD AND APPLE))
     set(LINKER_ORDER_FILE_OPTION "-Wl,-order_file,${CLANG_ORDER_FILE}")
@@ -118,7 +118,7 @@ if(CLANG_ORDER_FILE AND
   endif()
 
   # This is a test to ensure the actual order file works with the linker.
-  llvm_check_linker_flag(CXX ${LINKER_ORDER_FILE_OPTION} LINKER_ORDER_FILE_WORKS)
+  check_linker_flag(CXX ${LINKER_ORDER_FILE_OPTION} LINKER_ORDER_FILE_WORKS)
 
   # Passing an empty order file disables some linker layout optimizations.
   # To work around this and enable workflows for re-linking when the order file
diff --git a/llvm/cmake/modules/AddLLVM.cmake b/llvm/cmake/modules/AddLLVM.cmake
index 03f4e1f190fd9..cac5470435d91 100644
--- a/llvm/cmake/modules/AddLLVM.cmake
+++ b/llvm/cmake/modules/AddLLVM.cmake
@@ -327,8 +327,8 @@ function(add_link_opts target_name)
       elseif(${CMAKE_SYSTEM_NAME} MATCHES "SunOS" AND LLVM_LINKER_IS_SOLARISLD)
         # Support for ld -z discard-unused=sections was only added in
         # Solaris 11.4.  GNU ld ignores it, but warns every time.
-        include(LLVMCheckLinkerFlag)
-        llvm_check_linker_flag(CXX "-Wl,-z,discard-unused=sections" LINKER_SUPPORTS_Z_DISCARD_UNUSED)
+        include(CheckLinkerFlag)
+        check_linker_flag(CXX "-Wl,-z,discard-unused=sections" LINKER_SUPPORTS_Z_DISCARD_UNUSED)
         if (LINKER_SUPPORTS_Z_DISCARD_UNUSED)
           set_property(TARGET ${target_name} APPEND_STRING PROPERTY
                        LINK_FLAGS " -Wl,-z,discard-unused=sections")
diff --git a/llvm/cmake/modules/HandleLLVMOptions.cmake b/llvm/cmake/modules/HandleLLVMOptions.cmake
index 99d848ba3d853..4ec9dc3a3b571 100644
--- a/llvm/cmake/modules/HandleLLVMOptions.cmake
+++ b/llvm/cmake/modules/HandleLLVMOptions.cmake
@@ -1061,8 +1061,8 @@ if (LLVM_USE_SPLIT_DWARF AND
   if (CMAKE_CXX_COMPILER_ID MATCHES "Clang" OR
       CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
     add_compile_options($<$<COMPILE_LANGUAGE:C,CXX>:-gsplit-dwarf>)
-    include(LLVMCheckLinkerFlag)
-    llvm_check_linker_flag(CXX "-Wl,--gdb-index" LINKER_SUPPORTS_GDB_INDEX)
+    include(CheckLinkerFlag)
+    check_linker_flag(CXX "-Wl,--gdb-index" LINKER_SUPPORTS_GDB_INDEX)
     append_if(LINKER_SUPPORTS_GDB_INDEX "-Wl,--gdb-index"
       CMAKE_EXE_LINKER_FLAGS CMAKE_MODULE_LINKER_FLAGS CMAKE_SHARED_LINKER_FLAGS)
   endif()
@@ -1084,7 +1084,7 @@ endif()
 # lld doesn't print colored diagnostics when invoked from Ninja
 if (UNIX AND CMAKE_GENERATOR MATCHES "Ninja")
   include(LLVMCheckLinkerFlag)
-  llvm_check_linker_flag(CXX "-Wl,--color-diagnostics" LINKER_SUPPORTS_COLOR_DIAGNOSTICS)
+  check_linker_flag(CXX "-Wl,--color-diagnostics" LINKER_SUPPORTS_COLOR_DIAGNOSTICS)
   append_if(LINKER_SUPPORTS_COLOR_DIAGNOSTICS "-Wl,--color-diagnostics"
     CMAKE_EXE_LINKER_FLAGS CMAKE_MODULE_LINKER_FLAGS CMAKE_SHARED_LINKER_FLAGS)
 endif()
diff --git a/llvm/cmake/modules/HandleLLVMStdlib.cmake b/llvm/cmake/modules/HandleLLVMStdlib.cmake
index 7afc10cff74ff..a7e138aa0789b 100644
--- a/llvm/cmake/modules/HandleLLVMStdlib.cmake
+++ b/llvm/cmake/modules/HandleLLVMStdlib.cmake
@@ -13,12 +13,12 @@ if(NOT DEFINED LLVM_STDLIB_HANDLED)
   endfunction()
 
   include(CheckCXXCompilerFlag)
-  include(LLVMCheckLinkerFlag)
+  include(CheckLinkerFlag)
   set(LLVM_LIBCXX_USED 0)
   if(LLVM_ENABLE_LIBCXX)
     if(LLVM_COMPILER_IS_GCC_COMPATIBLE)
       check_cxx_compiler_flag("-stdlib=libc++" CXX_COMPILER_SUPPORTS_STDLIB)
-      llvm_check_linker_flag(CXX "-stdlib=libc++" CXX_LINKER_SUPPORTS_STDLIB)
+      check_linker_flag(CXX "-stdlib=libc++" CXX_LINKER_SUPPORTS_STDLIB)
       if(CXX_COMPILER_SUPPORTS_STDLIB AND CXX_LINKER_SUPPORTS_STDLIB)
         append("-stdlib=libc++"
           CMAKE_CXX_FLAGS CMAKE_EXE_LINKER_FLAGS CMAKE_SHARED_LINKER_FLAGS
@@ -36,7 +36,7 @@ if(NOT DEFINED LLVM_STDLIB_HANDLED)
     if(LLVM_COMPILER_IS_GCC_COMPATIBLE)
       check_cxx_compiler_flag("-static-libstdc++"
                               CXX_COMPILER_SUPPORTS_STATIC_STDLIB)
-      llvm_check_linker_flag(CXX "-static-libstdc++" CXX_LINKER_SUPPORTS_STATIC_STDLIB)
+      check_linker_flag(CXX "-static-libstdc++" CXX_LINKER_SUPPORTS_STATIC_STDLIB)
       if(CXX_COMPILER_SUPPORTS_STATIC_STDLIB AND
         CXX_LINKER_SUPPORTS_STATIC_STDLIB)
         append("-static-libstdc++"
diff --git a/llvm/cmake/modules/LLVMCheckLinkerFlag.cmake b/llvm/cmake/modules/LLVMCheckLinkerFlag.cmake
deleted file mode 100644
index e09bbc66f2d26..0000000000000
--- a/llvm/cmake/modules/LLVMCheckLinkerFlag.cmake
+++ /dev/null
@@ -1,28 +0,0 @@
-include(CheckLinkerFlag OPTIONAL)
-
-if (COMMAND check_linker_flag)
-  macro(llvm_check_linker_flag)
-    check_linker_flag(${ARGN})
-  endmacro()
-else()
-  # Until the minimum CMAKE version is 3.18
-
-  include(CheckCXXCompilerFlag)
-
-  # cmake builtin compatible, except we assume lang is C or CXX
-  function(llvm_check_linker_flag lang flag out_var)
-    cmake_policy(PUSH)
-    cmake_policy(SET CMP0056 NEW)
-    set(_CMAKE_EXE_LINKER_FLAGS_SAVE ${CMAKE_EXE_LINKER_FLAGS})
-    set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${flag}")
-    if("${lang}" STREQUAL "C")
-      check_c_compiler_flag("" ${out_var})
-    elseif("${lang}" STREQUAL "CXX")
-      check_cxx_compiler_flag("" ${out_var})
-    else()
-      message(FATAL_ERROR "\"${lang}\" is not C or CXX")
-    endif()
-    set(CMAKE_EXE_LINKER_FLAGS ${_CMAKE_EXE_LINKER_FLAGS_SAVE})
-    cmake_policy(POP)
-  endfunction()
-endif()
diff --git a/runtimes/CMakeLists.txt b/runtimes/CMakeLists.txt
index 24f4851169591..15f9334776e05 100644
--- a/runtimes/CMakeLists.txt
+++ b/runtimes/CMakeLists.txt
@@ -116,27 +116,7 @@ filter_prefixed("${CMAKE_ASM_IMPLICIT_INCLUDE_DIRECTORIES}" ${LLVM_BINARY_DIR} C
 # brittle. We should ideally move this to runtimes/CMakeLists.txt.
 llvm_check_compiler_linker_flag(C "--unwindlib=none" CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG)
 if (CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG)
-  set(ORIG_CMAKE_REQUIRED_FLAGS "${CMAKE_REQUIRED_FLAGS}")
-  set(CMAKE_REQUIRED_FLAGS "${CMAKE_REQUIRED_FLAGS} --unwindlib=none")
-  # TODO: When we can require CMake 3.14, we should use
-  # CMAKE_REQUIRED_LINK_OPTIONS here. Until then, we need a workaround:
-  # When using CMAKE_REQUIRED_FLAGS, this option gets added both to
-  # compilation and linking commands. That causes warnings in the
-  # compilation commands during cmake tests. This is normally benign, but
-  # when testing whether -Werror works, that test fails (due to the
-  # preexisting warning).
-  #
-  # Therefore, before we can use CMAKE_REQUIRED_LINK_OPTIONS, check if we
-  # can use --start-no-unused-arguments to silence the warnings about
-  # --unwindlib=none during compilation.
-  #
-  # We must first add --unwindlib=none to CMAKE_REQUIRED_FLAGS above, to
-  # allow this subsequent test to succeed, then rewrite CMAKE_REQUIRED_FLAGS
-  # below.
-  check_c_compiler_flag("--start-no-unused-arguments" C_SUPPORTS_START_NO_UNUSED_ARGUMENTS)
-  if (C_SUPPORTS_START_NO_UNUSED_ARGUMENTS)
-    set(CMAKE_REQUIRED_FLAGS "${ORIG_CMAKE_REQUIRED_FLAGS} --start-no-unused-arguments --unwindlib=none --end-no-unused-arguments")
-  endif()
+  set(CMAKE_REQUIRED_LINK_OPTIONS "${CMAKE_REQUIRED_LINK_OPTIONS} --unwindlib=none")
 endif()
 
 # Disable use of the installed C++ standard library when building runtimes.

h-vetinari · 2024-06-20T11:20:21Z

Found some more CMake cruft... Unfortunately, it doesn't improve the situation with libunwind, but at least that means I should be able to break this out into a separate PR.

h-vetinari · 2024-06-20T11:32:58Z

I think the riddle may be explained by @mstorsjo's excellent commit message from 4169b52 (which I just found):

If the CMake requirement is bumped to 3.14, we could use
CMAKE_REQUIRED_LINK_OPTIONS instead, removing the need for the
--{start,end}-no-unused-arguments options. (However, do note that
CMAKE_REQUIRED_LINK_OPTIONS is problematic in combination with
CMAKE_TRY_COMPILE_TARGET_TYPE set to STATIC_LIBRARY; see
https://gitlab.kitware.com/cmake/cmake/-/issues/23454.)

... which is indeed what libunwind specifies:

llvm-project/libunwind/CMakeLists.txt

Line 224 in 6859685

set(CMAKE_TRY_COMPILE_TARGET_TYPE STATIC_LIBRARY)

mstorsjo · 2024-06-20T11:33:11Z

Hm, this seems to affect some feature detection (and constraints) later on, though I don't see why/how removing the specification of --unwindlib=none to the driver (leaving it only to the linker) would cause that:

llvm-project/libunwind/src/CMakeLists.txt

Lines 99 to 106 in 1c046ca

if (LIBUNWIND_ENABLE_SHARED AND

NOT (CXX_SUPPORTS_FNO_EXCEPTIONS_FLAG AND

CXX_SUPPORTS_FUNWIND_TABLES_FLAG))

message(FATAL_ERROR

"Compiler doesn't support generation of unwind tables if exception "

"support is disabled. Building libunwind DSO with runtime dependency "

"on C++ ABI library is not supported.")

endif()

I actually tested doing this cleanup back around when I wrote this - sorry I never got around to amending the comment here explaining what has to be done. The case with libunwind is a bit non-obvious.

I rebased my test change from 2022, see mstorsjo@runtimes-link-options-unwindlib-none. Can you try this out and see if it fixes the issue you're seeing?

…ng of the --unwindlib=none option This avoids passing the option unnecessarily to compilation commands (where it causes warnings). This fails in practice with libunwind, where setting CMAKE_TRY_COMPILE_TARGET_TYPE to STATIC_LIBRARY breaks it, as the option from CMAKE_REQUIRED_LINK_OPTIONS ends up passed to the "ar" tool too.

h-vetinari · 2024-06-20T11:47:27Z

Haha, just when I realized the key to the whole thing - luckily my comment won the race condition at least. ;-)

Thanks a lot for the fix, I've pulled it into the PR (also rebased on main while I'm at it) - and CI didn't fail as previously, so this looks great! Thank you!

mstorsjo · 2024-06-20T12:01:37Z

Thanks a lot for the fix, I've pulled it into the PR (also rebased on main while I'm at it) - and CI didn't fail as previously, so this looks great! Thank you!

Ok, good! Btw, it might be good to add a reference to https://gitlab.kitware.com/cmake/cmake/-/issues/23454 in the new commit message too (or perhaps even in a comment in libunwind/CMakeLists.txt?), to clarify that this is indeed not a bug but expected behaviour.

ldionne

LGTM. The CI failures are flakes.

ldionne

Actually, I think some of the libunwind failures on BuildKite are legitimate. See https://buildkite.com/llvm-project/libcxx-ci/builds/35920.

h-vetinari · 2024-06-21T02:53:19Z

Actually, I think some of the libunwind failures on BuildKite are legitimate. See https://buildkite.com/llvm-project/libcxx-ci/builds/35920.

Yeah, it looks like the picolib-noexceptions build is still trying to actually build the exceptions, and then fails to find the symbols when linking. I've tried looking for anything that would switch arm/pico or the "-mthumb" that's in Armv7Thumb-no-exceptions.cmake which sets the baseline config for that job.

I do not think I'll be able to fix this without some help, or at least pointers.

ldionne · 2024-06-21T13:39:14Z

I looked into it and I don't quite understand. We should definitely be building with -fno-exceptions in that configuration (and I can see that in the logs), so I don't understand why we have things like:

| ld.lld: error: undefined symbol: __cxa_begin_catch
| >>> referenced by cxa_handlers.cpp:63 (/home/tcwg-buildbot/worker/linaro-aarch64-libcxx-02/llvm-project/libcxx-ci/build/armv7m-picolibc-no-exceptions/../../libcxxabi/src/cxa_handlers.cpp:63)
| >>>               cxa_handlers.cpp.obj:(std::__terminate(void (*)())) in archive /home/tcwg-buildbot/worker/linaro-aarch64-libcxx-02/llvm-project/libcxx-ci/build/armv7m-picolibc-no-exceptions/lib/libc++abi.a

There should not be references to these functions since we're building without exceptions.

Pinging the bot owner @mplatings

DavidSpickett · 2024-06-21T14:33:54Z

Linaro will look into this.

Reproducing the build is quite easy if you do want to look yourselves in the meantime:

<llvm project root>$ ./libcxx/utils/ci/run-buildbot armv7m-picolibc-no-exceptions

You'll need a recent qemu-system-arm installed to run the tests, I have 8.1.3 locally. Picolib is downloaded and built for you by the script.

antmox · 2024-06-24T14:59:04Z

I may be missing something, but it looks like the armv7m-picolibc-no-exceptions issue is still related to https://gitlab.kitware.com/cmake/cmake/-/issues/23454

I notice that once the --unwindlib=none flag is added to CMAKE_REQUIRED_LINK_OPTIONS, it is also added to ar commands during the next cmake tests (starting with CXX_SUPPORTS_NOSTDLIBXX_FLAG).
Failing these tests will result in fno-exceptions not being used later

This is probably due to the fact that for Armv7M-picolibc, CMAKE_TRY_COMPILE_TARGET_TYPE is set to STATIC_LIBRARY here: https://github.com/llvm/llvm-project/blob/main/libcxx/cmake/caches/Armv7M-picolibc.cmake#L7C35-L7C49

w.r.t. interference between CMAKE_REQUIRED_LINK_OPTIONS and static libraries

…lags_if_supported

…x.cmake

…lags_if_supported

h-vetinari · 2024-06-27T07:34:44Z

So I've been trying to follow down the rabbit hole of the failing flag checks, and it seems the combination of CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG plus https://gitlab.kitware.com/cmake/cmake/-/issues/23454 has a wider blast radius than anticipated.

I'm not claiming that adding the _previous_CMAKE_{REQUIRED_LINK_OPTIONS,TRY_COMPILE_TARGET_TYPE} dance everywhere is the right approach here, but it was - so far - the obvious path to just try to get things green again. It's conceivable though that it would be easier to simply shift the detection of CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG until after the other flag checks have been performed? 🤔

mstorsjo · 2024-06-27T07:53:05Z

So I've been trying to follow down the rabbit hole of the failing flag checks, and it seems the combination of CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG plus https://gitlab.kitware.com/cmake/cmake/-/issues/23454 has a wider blast radius than anticipated.

I'm not claiming that adding the _previous_CMAKE_{REQUIRED_LINK_OPTIONS,TRY_COMPILE_TARGET_TYPE} dance everywhere is the right approach here, but it was - so far - the obvious path to just try to get things green again. It's conceivable though that it would be easier to simply shift the detection of CXX_SUPPORTS_UNWINDLIB_EQ_NONE_FLAG until after the other flag checks have been performed? 🤔

That's probably not possible...

The point is that when bootstrapping a new sysroot from scratch (i.e. building the initial libunwind etc), in a configuration where libunwind is linked in automatically, every test that tries to do linking will fail (as it implicitly tries to link in libunwind, which does not exist yet). Therefore, we need to add --unwindlib=none as an additional linker flag, as soon as possible, so that all following cmake checks will get the right result.

Also, in general, setting CMAKE_TRY_COMPILE_TARGET_TYPE to STATIC_LIBRARY in too wide a context will also give false positive checks, for cases where we intentionally want to check whether linking some library works and is found. But perhaps the way you do it here, adding it in a narrow context only when doing specific checks, is the right way? I'm not sure...

So that cmake issue seems to be really, really unfortunate here. :-( I wonder if the cure is worse than the disease here - and if it would be better to just keep what we have now - and simplify it only if cmake adds something like CMAKE_REQUIRED_DYNAMIC_LINK_OPTIONS or so.

h-vetinari · 2024-06-27T07:57:05Z

So that cmake issue seems to be really, really unfortunate here. :-( I wonder if the cure is worse than the disease here [...]

Yup, that's a distinct possibility IMO...

[...] and if it would be better to just keep what we have now - and simplify it only if cmake adds something like CMAKE_REQUIRED_DYNAMIC_LINK_OPTIONS or so.

It would probably make sense to report back on the CMake issue how big the fallout from this is? Perhaps the CMake devs would reconsider, or at least take it as an indicator for the necessity of CMAKE_REQUIRED_DYNAMIC_LINK_OPTIONS? I think you understand the problem space much better than me (I'm mostly stumbling around in a dark room TBH), so if you could do that that would be great!

…g-ix.cmake

…_LINK_OPTIONS

Broken out from #93429 Somewhat closing the loop opened by 7017e6c. Co-authored-by: Ryan Prichard <[email protected]>

instead, teach existing llvm_check_compiler_linker_flag to do it

h-vetinari · 2024-08-12T05:44:08Z

I'm trying another approach here (now that #96171 showed we cannot get rid of llvm_check_compiler_linker_flag anyway) -- instead of doing the manual work-arounds for the linker flags & static targets everywhere, teach the existing function to do it (optionally).

There's not so many callers in the code base that this looks infeasible per se, but people might want to design the API differently (e.g. have a dedicated helper function for the case under consideration here). I'm also expecting this to break in various ways until I've figured out the right invocations everywhere. There's also several implementations of {target_,}add_flag_if_supported, which might have to expose similar arguments. Or we unify all those functions as well. Happy to iterate, or have someone else take over - I'll be the first to admit that CMake is not my forte 😅

h-vetinari requested a review from a team as a code owner May 26, 2024 22:43

llvmbot added the cmake Build system in general and CMake in particular label May 26, 2024

ldionne approved these changes Jun 11, 2024

View reviewed changes

llvmbot added the clang Clang issues not falling into any other category label Jun 20, 2024

h-vetinari requested a review from a team as a code owner June 20, 2024 11:00

llvmbot added compiler-rt libc++ libc++ C++ Standard Library. Not GNU libstdc++. Not libc++abi. libunwind openmp:libomp OpenMP host runtime offload labels Jun 20, 2024

h-vetinari force-pushed the unwindlib branch from dc24ea3 to e254f8d Compare June 20, 2024 11:04

h-vetinari mentioned this pull request Jun 20, 2024

[cmake] switch to CMake's native check_{compiler,linker}_flag #96171

Merged

h-vetinari force-pushed the unwindlib branch from e254f8d to 8c1b899 Compare June 20, 2024 11:40

add comment (and CMake issue reference) about incompatible options

816e9e6

ldionne approved these changes Jun 20, 2024

View reviewed changes

ldionne requested changes Jun 20, 2024

View reviewed changes

h-vetinari mentioned this pull request Jun 22, 2024

libcxx v18.1.8 conda-forge/libcxx-feedstock#155

Merged

8 tasks

[cmake] move check for -fno-exceptions to "safe zone"

3f917d2

w.r.t. interference between CMAKE_REQUIRED_LINK_OPTIONS and static libraries

h-vetinari force-pushed the unwindlib branch from 48764a8 to 3f917d2 Compare June 26, 2024 02:31

h-vetinari added 8 commits June 26, 2024 15:40

[cmake] same for -fno-rtti

a8dd830

[cmake] run _all_ flag checks in runtimes/CMakeLists.txt without linker

7444457

[cmake] also do static-vs-linker work-around for cxx_add_warning_flags

216799a

[cmake] also do static-vs-linker work-around for target_add_compile_f…

47e7688

…lags_if_supported

[cmake] also do static-vs-linker work-around in libcxx/cmake/config-i…

11cbf4f

…x.cmake

[cmake] also do static-vs-linker work-around for add_flag_if_supported

6ade702

[cmake] also do static-vs-linker work-around in runtimes/CMakeLists.txt

fcd9671

[cmake] also do static-vs-linker work-around for libcxx{,abi}'s add_f…

f34093a

…lags_if_supported

h-vetinari requested a review from a team as a code owner June 27, 2024 07:32

[cmake] also do static-vs-linker work-around in libcxxabi/cmake/confi…

37bb8e1

…g-ix.cmake

h-vetinari force-pushed the unwindlib branch from 0701fd1 to 37bb8e1 Compare June 27, 2024 22:03

h-vetinari added 2 commits June 28, 2024 09:10

[cmake] also do static-vs-linker work-around in libunwind/cmake/confi…

74b9d3f

…g-ix.cmake

[cmake] fix up order of work-around with definition of CMAKE_REQUIRED…

8c2fda5

…_LINK_OPTIONS

ldionne pushed a commit that referenced this pull request Jul 31, 2024

[cmake] switch to CMake's native check_{compiler,linker}_flag (#96171)

89946bd

Broken out from #93429 Somewhat closing the loop opened by 7017e6c. Co-authored-by: Ryan Prichard <[email protected]>

h-vetinari added 2 commits August 12, 2024 15:16

Merge remote-tracking branch 'upstream/main' into unwindlib

bbc4a0f

avoid resetting context for CMake flag checks manually

9576869

instead, teach existing llvm_check_compiler_linker_flag to do it

h-vetinari marked this pull request as draft August 12, 2024 05:38

[runtimes] remove workaround for old CMake when setting --unwindlib=none #93429

Are you sure you want to change the base?

[runtimes] remove workaround for old CMake when setting --unwindlib=none #93429

Uh oh!

Conversation

h-vetinari commented May 26, 2024

Uh oh!

h-vetinari commented May 26, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ldionne commented Jun 11, 2024

Uh oh!

ldionne left a comment

Choose a reason for hiding this comment

Uh oh!

h-vetinari commented Jun 20, 2024

Uh oh!

llvmbot commented Jun 20, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

h-vetinari commented Jun 20, 2024

Uh oh!

h-vetinari commented Jun 20, 2024

Uh oh!

mstorsjo commented Jun 20, 2024

Uh oh!

h-vetinari commented Jun 20, 2024

Uh oh!

mstorsjo commented Jun 20, 2024

Uh oh!

ldionne left a comment

Choose a reason for hiding this comment

Uh oh!

ldionne left a comment

Choose a reason for hiding this comment

Uh oh!

h-vetinari commented Jun 21, 2024

Uh oh!

ldionne commented Jun 21, 2024

Uh oh!

DavidSpickett commented Jun 21, 2024

Uh oh!

antmox commented Jun 24, 2024

Uh oh!

h-vetinari commented Jun 27, 2024

Uh oh!

mstorsjo commented Jun 27, 2024

Uh oh!

h-vetinari commented Jun 27, 2024

Uh oh!

h-vetinari commented Aug 12, 2024

Uh oh!

Uh oh!

[runtimes] remove workaround for old CMake when setting `--unwindlib=none` #93429

[runtimes] remove workaround for old CMake when setting `--unwindlib=none` #93429

h-vetinari commented May 26, 2024 •

edited

Loading

llvmbot commented Jun 20, 2024 •

edited

Loading