[libc] [search] implement hcreate(_r)/hsearch(_r)/hdestroy(_r) (#73469)

This patch implements `hcreate(_r)/hsearch(_r)/hdestroy(_r)` as
specified in https://man7.org/linux/man-pages/man3/hsearch.3.html.

Notice that `neon/asimd` extension is not yet added in this patch. 

- The implementation is largely simplified from rust's
[`hashbrown`](https://github.com/rust-lang/hashbrown/blob/master/src/raw/mod.rs)
as we only consider fix-sized insertion-only hashtables. Technical
details are provided in code comments.

- This patch also contains a portable string hash function, which is
derived from [`aHash`](https://github.com/tkaitchuck/aHash)'s fallback
routine. Not using any SIMD acceleration, it has a good enough quality
(passing all SMHasher tests) and is not too bad in speed.

- Some general functionalities are added, such as `memory_size`,
`offset_to`(alignment), `next_power_of_two`, `is_power_of_two`.
`ctz/clz` are extended to support shorter integers.

Author: SchrodingerZhu

libc/cmake/modules/LLVMLibCFlagRules.cmake

@@ -132,6 +132,8 @@ endfunction(get_fq_dep_list_without_flag)
 # Special flags
 set(FMA_OPT_FLAG "FMA_OPT")
 set(ROUND_OPT_FLAG "ROUND_OPT")
+# SSE2 is the baseline for x86_64, so we add a negative flag to disable it if needed.
+set(DISABLE_SSE2_OPT_FLAG "DISABLE_SSE2_OPT")
 
 # Skip FMA_OPT flag for targets that don't support fma.
 if(NOT((LIBC_TARGET_ARCHITECTURE_IS_X86 AND (LIBC_CPU_FEATURES MATCHES "FMA")) OR
@@ -143,3 +145,8 @@ endif()
 if(NOT(LIBC_TARGET_ARCHITECTURE_IS_X86 AND (LIBC_CPU_FEATURES MATCHES "SSE4_2")))
   set(SKIP_FLAG_EXPANSION_ROUND_OPT TRUE)
 endif()
+
+# Skip DISABLE_SSE2_OPT flag for targets that don't support SSE2.
+if(NOT(LIBC_TARGET_ARCHITECTURE_IS_X86 AND (LIBC_CPU_FEATURES MATCHES "SSE2")))
+  set(SKIP_FLAG_EXPANSION_DISABLE_SSE2_OPT TRUE)
+endif()

libc/cmake/modules/LLVMLibCObjectRules.cmake

@@ -18,6 +18,14 @@ function(_get_common_compile_options output_var flags)
     set(ADD_SSE4_2_FLAG TRUE)
   endif()
 
+  list(FIND flags ${DISABLE_SSE2_OPT_FLAG} no_sse2)
+  if(${no_sse2} LESS 0)
+    list(FIND flags "${DISABLE_SSE2_OPT_FLAG}__ONLY" no_sse2)
+  endif()
+  if((${no_sse2} GREATER -1) AND (LIBC_CPU_FEATURES MATCHES "SSE2"))
+    set(DISABLE_SSE2_FLAG TRUE)
+  endif()
+
   set(compile_options ${LIBC_COMPILE_OPTIONS_DEFAULT} ${ARGN})
   if(LLVM_COMPILER_IS_GCC_COMPATIBLE)
     list(APPEND compile_options "-fpie")
@@ -58,12 +66,18 @@ function(_get_common_compile_options output_var flags)
     if(ADD_SSE4_2_FLAG)
       list(APPEND compile_options "-msse4.2")
     endif()
+    if(DISABLE_SSE2_FLAG)
+      list(APPEND compile_options "-mno-sse2")
+    endif()
   elseif(MSVC)
     list(APPEND compile_options "/EHs-c-")
     list(APPEND compile_options "/GR-")
     if(ADD_FMA_FLAG)
       list(APPEND compile_options "/arch:AVX2")
     endif()
+    if(DISABLE_SSE2_FLAG)
+      list(APPEND compile_options "/arch:SSE")
+    endif()
   endif()
   if (LIBC_TARGET_ARCHITECTURE_IS_GPU)
     list(APPEND compile_options "-nogpulib")

libc/config/linux/aarch64/entrypoints.txt

@@ -130,6 +130,14 @@ set(TARGET_LIBC_ENTRYPOINTS
     #libc.src.stdio.scanf
     #libc.src.stdio.fscanf
 
+    # search.h entrypoints
+    libc.src.search.hcreate
+    libc.src.search.hcreate_r
+    libc.src.search.hsearch
+    libc.src.search.hsearch_r
+    libc.src.search.hdestroy
+    libc.src.search.hdestroy_r
+
     # sys/mman.h entrypoints
     libc.src.sys.mman.madvise
     libc.src.sys.mman.mmap

libc/config/linux/aarch64/headers.txt

@@ -12,6 +12,7 @@ set(TARGET_PUBLIC_HEADERS
     libc.include.stdlib
     libc.include.string
     libc.include.strings
+    libc.include.search
     libc.include.sys_mman
     libc.include.sys_socket
     libc.include.sys_syscall

libc/config/linux/api.td

@@ -248,3 +248,7 @@ def TermiosAPI : PublicAPI<"termios.h"> {
 def SetJmpAPI : PublicAPI<"setjmp.h"> {
   let Types = ["jmp_buf"];
 }
+
+def SearchAPI : PublicAPI<"search.h"> {
+  let Types = ["ACTION", "ENTRY", "struct hsearch_data"];
+}

libc/config/linux/arm/entrypoints.txt

@@ -89,6 +89,14 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.stdlib.strtoul
     libc.src.stdlib.strtoull
 
+    # search.h entrypoints
+    libc.src.search.hcreate
+    libc.src.search.hcreate_r
+    libc.src.search.hsearch
+    libc.src.search.hsearch_r
+    libc.src.search.hdestroy
+    libc.src.search.hdestroy_r
+
     # sys/mman.h entrypoints
     libc.src.sys.mman.mmap
     libc.src.sys.mman.munmap

libc/config/linux/arm/headers.txt

@@ -7,4 +7,5 @@ set(TARGET_PUBLIC_HEADERS
     libc.include.stdlib
     libc.include.string
     libc.include.strings
+    libc.include.search
 )

libc/config/linux/riscv/entrypoints.txt

@@ -136,6 +136,14 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.stdio.scanf
     libc.src.stdio.fscanf
 
+    # search.h entrypoints
+    libc.src.search.hcreate
+    libc.src.search.hcreate_r
+    libc.src.search.hsearch
+    libc.src.search.hsearch_r
+    libc.src.search.hdestroy
+    libc.src.search.hdestroy_r
+
     # sys/mman.h entrypoints
     libc.src.sys.mman.madvise
     libc.src.sys.mman.mmap

libc/config/linux/riscv/headers.txt

@@ -17,6 +17,7 @@ set(TARGET_PUBLIC_HEADERS
     libc.include.stdlib
     libc.include.string
     libc.include.strings
+    libc.include.search
     libc.include.termios
     libc.include.threads
     libc.include.time

libc/config/linux/x86_64/entrypoints.txt

@@ -497,6 +497,14 @@ if(LLVM_LIBC_FULL_BUILD)
     libc.src.spawn.posix_spawn_file_actions_destroy
     libc.src.spawn.posix_spawn_file_actions_init
 
+    # search.h entrypoints
+    libc.src.search.hcreate
+    libc.src.search.hcreate_r
+    libc.src.search.hsearch
+    libc.src.search.hsearch_r
+    libc.src.search.hdestroy
+    libc.src.search.hdestroy_r
+
     # threads.h entrypoints
     libc.src.threads.call_once
     libc.src.threads.cnd_broadcast