Re: [PATCH] drm/msm/adreno: Remove VLA usage

@@ -91,12 +93,13 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname) ret = qcom_mdt_load(dev, fw, fwname, GPU_PAS_ID, mem_region, mem_phys, mem_size, NULL); } else {

• char newname[strlen("qcom/") + strlen(fwname) + 1];
  

• char *newname;
  
  

• sprintf(newname, "qcom/%s", fwname);
  

• newname = kasprintf(GFP_KERNEL, "qcom/%s", fwname);
  
  

  ret = qcom_mdt_load(dev, fw, newname, GPU_PAS_ID, mem_region, mem_phys, mem_size, NULL);

I have taken another look also at this update suggestion. Now I wonder why the return value is not checked for the added name construction in the way as it is specified for the function “adreno_request_fw”. Will another condition check make sense at this place?

Regards, Markus