[libc-commits] [libc] Revert "[libc] fix aarch64 linux full build (#95358)" (PR #95419)

[via libc-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-libc

Author: Schrodinger ZHU Yifan (SchrodingerZhu)

<details>
<summary>Changes</summary>

This reverts commit ca05204f9aa258c5324d5675c7987c7e570168a0.

I misinterpreted `__getauxval` with `__getauxval2`. It should be the same as `getauxval` and we should be able to provide it in a more proper way.

I also found the GCC patch that actually cause this problem:
https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=1266778548e20de82983b6446f3cb685068cfb1e

---
Full diff: https://github.com/llvm/llvm-project/pull/95419.diff


3 Files Affected:

- (modified) libc/config/linux/aarch64/entrypoints.txt (-7) 
- (modified) libc/src/__support/threads/linux/CMakeLists.txt (-1) 
- (modified) libc/test/IntegrationTest/test.cpp (-6) 


``````````diff
diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt
index 7ce088689b925..db96a80051a8d 100644
--- a/libc/config/linux/aarch64/entrypoints.txt
+++ b/libc/config/linux/aarch64/entrypoints.txt
@@ -643,12 +643,6 @@ if(LLVM_LIBC_FULL_BUILD)
     libc.src.pthread.pthread_mutexattr_setrobust
     libc.src.pthread.pthread_mutexattr_settype
     libc.src.pthread.pthread_once
-    libc.src.pthread.pthread_rwlockattr_destroy
-    libc.src.pthread.pthread_rwlockattr_getkind_np
-    libc.src.pthread.pthread_rwlockattr_getpshared
-    libc.src.pthread.pthread_rwlockattr_init
-    libc.src.pthread.pthread_rwlockattr_setkind_np
-    libc.src.pthread.pthread_rwlockattr_setpshared
     libc.src.pthread.pthread_setspecific
 
     # sched.h entrypoints
@@ -759,7 +753,6 @@ if(LLVM_LIBC_FULL_BUILD)
     libc.src.unistd._exit
     libc.src.unistd.environ
     libc.src.unistd.execv
-    libc.src.unistd.fork
     libc.src.unistd.getopt
     libc.src.unistd.optarg
     libc.src.unistd.optind
diff --git a/libc/src/__support/threads/linux/CMakeLists.txt b/libc/src/__support/threads/linux/CMakeLists.txt
index 8e6cd7227b2c8..9bf88ccc84557 100644
--- a/libc/src/__support/threads/linux/CMakeLists.txt
+++ b/libc/src/__support/threads/linux/CMakeLists.txt
@@ -64,7 +64,6 @@ add_object_library(
     .futex_utils
     libc.config.linux.app_h
     libc.include.sys_syscall
-    libc.include.fcntl
     libc.src.errno.errno
     libc.src.__support.CPP.atomic
     libc.src.__support.CPP.stringstream
diff --git a/libc/test/IntegrationTest/test.cpp b/libc/test/IntegrationTest/test.cpp
index 27e7f29efa0f1..3bdbe89a3fb62 100644
--- a/libc/test/IntegrationTest/test.cpp
+++ b/libc/test/IntegrationTest/test.cpp
@@ -79,10 +79,4 @@ void *realloc(void *ptr, size_t s) {
 // Integration tests are linked with -nostdlib. BFD linker expects
 // __dso_handle when -nostdlib is used.
 void *__dso_handle = nullptr;
-
-// On some platform (aarch64 fedora tested) full build integration test
-// objects need to link against libgcc, which may expect a __getauxval
-// function. For now, it is fine to provide a weak definition that always
-// returns false.
-[[gnu::weak]] bool __getauxval(uint64_t, uint64_t *) { return false; }
 } // extern "C"

``````````

</details>


https://github.com/llvm/llvm-project/pull/95419