[v0.25] Leave cuda/intel backend to decide how to report PCI vendor failure

src/common/pci.cpp

@@ -28,7 +28,6 @@ bool pciutils_is_vendor_id_present(struct pci_dev *devices, int id) {
     }
   }
 
-  printErr("Unable to find a valid device for vendor id 0x%.4X using pciutils", id);
   return false;
 }
 

src/cuda/cuda.cpp

@@ -152,6 +152,7 @@ struct gpu_info* get_gpu_info_cuda(struct pci_dev *devices, int gpu_idx) {
   strcpy(gpu->name, deviceProp.name);
 
   if((gpu->pci = get_pci_from_pciutils(devices, PCI_VENDOR_ID_NVIDIA, gpu_idx)) == NULL) {
+    printErr("Unable to find a valid device for vendor id 0x%.4X using pciutils", PCI_VENDOR_ID_NVIDIA);
     return NULL;
   }
   gpu->arch = get_uarch_from_cuda(gpu);

src/intel/intel.cpp

@@ -26,6 +26,7 @@ struct gpu_info* get_gpu_info_intel(struct pci_dev *devices) {
 
   if(gpu->pci == NULL) {
     // No Intel iGPU found in PCI, which means it is not present
+    printWarn("Unable to find a valid device for vendor id 0x%.4X using pciutils", PCI_VENDOR_ID_INTEL);
     return NULL;
   }