cmake : enable warnings in llama (#10474)

* cmake : enable warnings in llama

ggml-ci

* cmake : add llama_get_flags and respect LLAMA_FATAL_WARNINGS

* cmake : get_flags -> ggml_get_flags

* speculative-simple : fix warnings

* cmake : reuse ggml_get_flags

ggml-ci

* speculative-simple : fix compile warning

ggml-ci

CMakeLists.txt

@@ -82,6 +82,7 @@ option(LLAMA_CURL "llama: use libcurl to download model from an URL" OFF)
 
 # Required for relocatable CMake package
 include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info.cmake)
+include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/common.cmake)
 
 # override ggml options
 set(GGML_SANITIZE_THREAD    ${LLAMA_SANITIZE_THREAD})

cmake/common.cmake

@@ -0,0 +1,33 @@
+function(llama_add_compile_flags)
+    if (LLAMA_FATAL_WARNINGS)
+        if (CMAKE_CXX_COMPILER_ID MATCHES "GNU" OR CMAKE_CXX_COMPILER_ID MATCHES "Clang")
+            list(APPEND C_FLAGS   -Werror)
+            list(APPEND CXX_FLAGS -Werror)
+        elseif (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
+            add_compile_options(/WX)
+        endif()
+    endif()
+
+    if (LLAMA_ALL_WARNINGS)
+        if (NOT MSVC)
+            list(APPEND C_FLAGS -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes
+                                -Werror=implicit-int -Werror=implicit-function-declaration)
+
+            list(APPEND CXX_FLAGS -Wmissing-declarations -Wmissing-noreturn)
+
+            list(APPEND WARNING_FLAGS -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function)
+
+            list(APPEND C_FLAGS   ${WARNING_FLAGS})
+            list(APPEND CXX_FLAGS ${WARNING_FLAGS})
+
+            ggml_get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION})
+
+            add_compile_options("$<$<COMPILE_LANGUAGE:C>:${C_FLAGS};${GF_C_FLAGS}>"
+                                "$<$<COMPILE_LANGUAGE:CXX>:${CXX_FLAGS};${GF_CXX_FLAGS}>")
+        else()
+            # todo : msvc
+            set(C_FLAGS   "" PARENT_SCOPE)
+            set(CXX_FLAGS "" PARENT_SCOPE)
+        endif()
+    endif()
+endfunction()

common/CMakeLists.txt

@@ -2,6 +2,8 @@
 
 find_package(Threads REQUIRED)
 
+llama_add_compile_flags()
+
 # Build info header
 #
 

examples/CMakeLists.txt

@@ -6,6 +6,10 @@ find_package(Threads REQUIRED)
 
 # ...
 
+# flags
+
+llama_add_compile_flags()
+
 # examples
 
 include_directories(${CMAKE_CURRENT_SOURCE_DIR})

examples/speculative-simple/speculative-simple.cpp

@@ -70,13 +70,13 @@ int main(int argc, char ** argv) {
     std::vector<llama_token> inp;
     inp = common_tokenize(ctx_tgt, params.prompt, true, true);
 
-    if (llama_n_ctx(ctx_tgt) < (int) inp.size()) {
+    if (llama_n_ctx(ctx_tgt) < (uint32_t) inp.size()) {
         LOG_ERR("%s: the prompt exceeds the context size (%d tokens, ctx %d)\n", __func__, (int) inp.size(), llama_n_ctx(ctx_tgt));
 
         return 1;
     }
 
-    if (llama_n_batch(ctx_tgt) < (int) inp.size()) {
+    if (llama_n_batch(ctx_tgt) < (uint32_t) inp.size()) {
         LOG_ERR("%s: the prompt exceeds the batch size (%d tokens, batch %d)\n", __func__, (int) inp.size(), llama_n_batch(ctx_tgt));
 
         return 1;
@@ -155,7 +155,7 @@ int main(int argc, char ** argv) {
         // evaluate the target model on [id_last, draft0, draft1, ..., draftN-1]
         {
             // do not waste time on small drafts
-            if (draft.size() < n_draft_min) {
+            if (draft.size() < (size_t) n_draft_min) {
                 draft.clear();
             }
 

ggml/src/CMakeLists.txt

@@ -24,7 +24,7 @@ if (NOT MSVC)
     endif()
 endif()
 
-function(get_flags CCID CCVER)
+function(ggml_get_flags CCID CCVER)
     set(C_FLAGS "")
     set(CXX_FLAGS "")
 
@@ -41,6 +41,7 @@ function(get_flags CCID CCVER)
     elseif (CCID STREQUAL "GNU")
         set(C_FLAGS   -Wdouble-promotion)
         set(CXX_FLAGS -Wno-array-bounds)
+
         if (CCVER VERSION_GREATER_EQUAL 8.1.0)
             list(APPEND CXX_FLAGS -Wextra-semi)
         endif()
@@ -69,7 +70,7 @@ if (GGML_ALL_WARNINGS)
         list(APPEND C_FLAGS   ${WARNING_FLAGS})
         list(APPEND CXX_FLAGS ${WARNING_FLAGS})
 
-        get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION})
+        ggml_get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION})
 
         add_compile_options("$<$<COMPILE_LANGUAGE:C>:${C_FLAGS};${GF_C_FLAGS}>"
                             "$<$<COMPILE_LANGUAGE:CXX>:${CXX_FLAGS};${GF_CXX_FLAGS}>")

ggml/src/ggml-cuda/CMakeLists.txt

@@ -132,7 +132,7 @@ if (CUDAToolkit_FOUND)
 
         message("-- CUDA host compiler is ${CUDA_CCID} ${CUDA_CCVER}")
 
-        get_flags(${CUDA_CCID} ${CUDA_CCVER})
+        ggml_get_flags(${CUDA_CCID} ${CUDA_CCVER})
         list(APPEND CUDA_CXX_FLAGS ${CXX_FLAGS} ${GF_CXX_FLAGS})  # This is passed to -Xcompiler later
     endif()
 

src/CMakeLists.txt

@@ -5,6 +5,8 @@ if (WIN32)
     endif()
 endif()
 
+llama_add_compile_flags()
+
 #
 # libraries
 #