diff --git a/misc/py-vllm/files/patch-cmake_cpu__extension.cmake b/misc/py-vllm/files/patch-cmake_cpu__extension.cmake
index 6a7d97be5543..8ae39b4c76a6 100644
--- a/misc/py-vllm/files/patch-cmake_cpu__extension.cmake
+++ b/misc/py-vllm/files/patch-cmake_cpu__extension.cmake
@@ -1,111 +1,123 @@
---- cmake/cpu_extension.cmake.orig	2026-05-10 07:34:36 UTC
-+++ cmake/cpu_extension.cmake
-@@ -20,6 +20,11 @@ set (ENABLE_NUMA TRUE)
+--- cmake/cpu_extension.cmake.orig	2026-05-14 23:50:01.854145000 +0200
++++ cmake/cpu_extension.cmake	2026-05-14 23:50:01.866986000 +0200
+@@ -20,6 +20,11 @@
  
  set (ENABLE_NUMA TRUE)
  
 +# FreeBSD does not have libnuma
 +if (CMAKE_SYSTEM_NAME STREQUAL "FreeBSD")
 +    set(ENABLE_NUMA OFF)
 +endif()
 +
  #
  # Check the compile flags
  #
-@@ -31,16 +36,23 @@ else()
+@@ -31,16 +36,23 @@
          "-fopenmp"
          "-DVLLM_CPU_EXTENSION")
  
 -    # locate PyTorch's libgomp (e.g. site-packages/torch.libs/libgomp-947d5fa1.so.1.0.0)
 -    # and create a local shim dir with it
 -    vllm_prepare_torch_gomp_shim(VLLM_TORCH_GOMP_SHIM_DIR)
 +    if (CMAKE_SYSTEM_NAME STREQUAL "FreeBSD")
 +        find_library(OPEN_MP
 +            NAMES omp
 +            REQUIRED
 +        )
 +    else()
 +        # locate PyTorch's libgomp (e.g. site-packages/torch.libs/libgomp-947d5fa1.so.1.0.0)
 +        # and create a local shim dir with it
 +        vllm_prepare_torch_gomp_shim(VLLM_TORCH_GOMP_SHIM_DIR)
  
 -    find_library(OPEN_MP
 -        NAMES gomp
 -        PATHS ${VLLM_TORCH_GOMP_SHIM_DIR}
 -        NO_DEFAULT_PATH
 -        REQUIRED
 -    )
 +        find_library(OPEN_MP
 +            NAMES gomp
 +            PATHS ${VLLM_TORCH_GOMP_SHIM_DIR}
 +            NO_DEFAULT_PATH
 +            REQUIRED
 +        )
 +    endif()
      # Set LD_LIBRARY_PATH to include the shim dir at build time to use the same libgomp as PyTorch
      if (OPEN_MP)
          set(ENV{LD_LIBRARY_PATH} "${VLLM_TORCH_GOMP_SHIM_DIR}:$ENV{LD_LIBRARY_PATH}")
-@@ -48,12 +60,25 @@ if (NOT MACOSX_FOUND)
+@@ -48,11 +60,25 @@
  endif()
  
  if (NOT MACOSX_FOUND)
 -    execute_process(COMMAND cat /proc/cpuinfo
 -                    RESULT_VARIABLE CPUINFO_RET
 -                    OUTPUT_VARIABLE CPUINFO)
 -    if (NOT CPUINFO_RET EQUAL 0)
 -        message(FATAL_ERROR "Failed to check CPU features via /proc/cpuinfo")
 +    # Try Linux /proc/cpuinfo first, then the FreeBSD linuxulator path
 +    if (EXISTS "/proc/cpuinfo")
 +        set(_cpuinfo_path "/proc/cpuinfo")
 +    elseif (EXISTS "/compat/linux/proc/cpuinfo")
 +        set(_cpuinfo_path "/compat/linux/proc/cpuinfo")
 +    else()
 +        set(_cpuinfo_path "")
-     endif()
++    endif()
 +    if (_cpuinfo_path)
 +        execute_process(COMMAND cat ${_cpuinfo_path}
 +                        RESULT_VARIABLE CPUINFO_RET
 +                        OUTPUT_VARIABLE CPUINFO)
 +        if (NOT CPUINFO_RET EQUAL 0)
 +            message(FATAL_ERROR "Failed to check CPU features via ${_cpuinfo_path}")
 +        endif()
 +    else()
 +        message(STATUS "No cpuinfo available; relying on CMAKE_SYSTEM_PROCESSOR for ISA detection")
-+        set(CPUINFO "")
-+    endif()
++        set(CPUINFO " ")
++        set(_USE_SYSPROCESSOR_ISA TRUE)
+     endif()
  endif()
  
- 
-@@ -106,9 +131,10 @@ if (CMAKE_SYSTEM_PROCESSOR MATCHES "x86_64|amd64" OR E
+@@ -101,14 +127,22 @@
+     if (ENABLE_ARM_BF16)
+         set(ARM_BF16_FOUND ON)
+         message(STATUS "ARM BF16 support enabled via VLLM_CPU_ARM_BF16 environment variable")
++    endif()
++    # Override ISA based on CMAKE_SYSTEM_PROCESSOR when cpuinfo not available
++    if(_USE_SYSPROCESSOR_ISA)
++        if(CMAKE_SYSTEM_PROCESSOR MATCHES "powerpc64le|powerpc64|powerpc")
++            set(POWER9_FOUND ON)
++            message(STATUS "Detected ${CMAKE_SYSTEM_PROCESSOR}; enabling POWER9 ISA")
++        endif()
+     endif()
+ endif()
  
  if (CMAKE_SYSTEM_PROCESSOR MATCHES "x86_64|amd64" OR ENABLE_X86_ISA)
      set(ENABLE_X86_ISA ON)
 -    if (NOT (CMAKE_CXX_COMPILER_ID STREQUAL "GNU" AND
 -            CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 12.3))
 -        message(FATAL_ERROR "X86 backend requires gcc/g++ >= 12.3")
 +    if (NOT (
 +            (CMAKE_CXX_COMPILER_ID STREQUAL "GNU" AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 12.3) OR
 +            (CMAKE_CXX_COMPILER_ID STREQUAL "Clang" AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 15.0)))
 +        message(FATAL_ERROR "X86 backend requires gcc/g++ >= 12.3 or clang >= 15.0")
      endif()
      list(APPEND CXX_COMPILE_FLAGS "-mf16c")
      list(APPEND CXX_COMPILE_FLAGS_AVX512 ${CXX_COMPILE_FLAGS})
-@@ -446,9 +472,15 @@ if (ENABLE_X86_ISA)
+@@ -446,9 +486,15 @@
      message(STATUS "CPU extension (AVX512F) source files: ${VLLM_EXT_SRC_AVX512}")
      message(STATUS "CPU extension (AVX2) source files: ${VLLM_EXT_SRC_AVX2}")
  
 -    set(_C_LIBS numa dnnl_ext)
 -    set(_C_AVX512_LIBS numa dnnl_ext)
 -    set(_C_AVX2_LIBS numa)
 +    if(ENABLE_NUMA)
 +        set(_C_LIBS numa dnnl_ext)
 +        set(_C_AVX512_LIBS numa dnnl_ext)
 +        set(_C_AVX2_LIBS numa)
 +    else()
 +        set(_C_LIBS dnnl_ext)
 +        set(_C_AVX512_LIBS dnnl_ext)
 +        set(_C_AVX2_LIBS "")
 +    endif()
  
      # AMX + AVX512F + AVX512BF16 + AVX512VNNI
      define_extension_target(
diff --git a/misc/py-vllm/files/patch-csrc_cpu_cpu__types__vsx.hpp b/misc/py-vllm/files/patch-csrc_cpu_cpu__types__vsx.hpp
new file mode 100644
index 000000000000..f0c1f7e7979d
--- /dev/null
+++ b/misc/py-vllm/files/patch-csrc_cpu_cpu__types__vsx.hpp
@@ -0,0 +1,49 @@
+--- csrc/cpu/cpu_types_vsx.hpp.orig	2026-05-15 01:00:58.314737000 +0200
++++ csrc/cpu/cpu_types_vsx.hpp	2026-05-15 01:00:58.314987000 +0200
+@@ -64,7 +64,7 @@
+ } f32x4x4_t;
+ 
+ typedef struct i32x4x4_t {
+-  __vector int32_t val[4];
++  __vector int val[4];
+ } i32x4x4_t;
+ 
+ struct FP32Vec8;
+@@ -76,7 +76,7 @@
+   __vector signed short reg;
+ 
+   explicit BF16Vec8(const void* ptr)
+-      : reg((__vector signed short)vec_xl(0, (__vector signed short*)ptr)) {}
++      : reg((__vector signed short)vec_xl(0, (signed short*)ptr)) {}
+ 
+   explicit BF16Vec8(const FP32Vec8&);
+ 
+@@ -294,20 +294,20 @@
+   i32x4x4_t reg;
+ 
+   explicit INT32Vec16(const void* data_ptr) {
+-    reg.val[0] = vec_xl(0, reinterpret_cast<const __vector int32_t*>(data_ptr));
++    reg.val[0] = vec_xl(0, reinterpret_cast<const int32_t*>(data_ptr));
+     reg.val[1] =
+-        vec_xl(16, reinterpret_cast<const __vector int32_t*>(data_ptr));
++        vec_xl(16, reinterpret_cast<const int32_t*>(data_ptr));
+     reg.val[2] =
+-        vec_xl(32, reinterpret_cast<const __vector int32_t*>(data_ptr));
++        vec_xl(32, reinterpret_cast<const int32_t*>(data_ptr));
+     reg.val[3] =
+-        vec_xl(48, reinterpret_cast<const __vector int32_t*>(data_ptr));
++        vec_xl(48, reinterpret_cast<const int32_t*>(data_ptr));
+   }
+ 
+   void save(int32_t* ptr) const {
+-    vec_xst(reg.val[0], 0, reinterpret_cast<__vector int32_t*>(ptr));
+-    vec_xst(reg.val[1], 16, reinterpret_cast<__vector int32_t*>(ptr));
+-    vec_xst(reg.val[2], 32, reinterpret_cast<__vector int32_t*>(ptr));
+-    vec_xst(reg.val[3], 48, reinterpret_cast<__vector int32_t*>(ptr));
++    vec_xst(reg.val[0], 0, reinterpret_cast<int32_t*>(ptr));
++    vec_xst(reg.val[1], 16, reinterpret_cast<int32_t*>(ptr));
++    vec_xst(reg.val[2], 32, reinterpret_cast<int32_t*>(ptr));
++    vec_xst(reg.val[3], 48, reinterpret_cast<int32_t*>(ptr));
+   }
+ 
+   void save(int32_t* ptr, const int elem_num) const {