Re: [libvirt] [PATCH 1/2] virTypedParamsSerialize: minor fixes

On 04/25/2018 11:55 AM, Marc Hartmayer wrote: > 1. Don't allocate if there is nothing that needs to be > allocated. Especially as the result of calling calloc(0, ...) is > implementation-defined. Following VIR_ALLOC_N one finds : VIR_ALLOC_N(params_val, nparams) which equates to # define VIR_ALLOC_N(ptr, count) \ virAllocN(&(ptr), sizeof(*(ptr)), (count), true, \ VIR_FROM_THIS, __FILE__, __FUNCTION__, __LINE__) or virAllocN(&params_val, sizeof(params_val), nparams, true, ...) and eventually/essentially *params_val = calloc(sizeof(params_val), nparams) If the returned value is NULL then we error w/ OOM (4th param=true). So, unless @params_val had no elements, it won't be calloc(0,...) and thus the question becomes is there a more specific path you are referencing here? FWIW: My f26 man page for calloc says: "The calloc() function allocates memory for an array of nmemb elements of size bytes each and returns a pointer to the allocated memory. The memory is set to zero. If nmemb or size is 0, then calloc() returns either NULL, or a unique pointer value that can later be successfully passed to free()" We have so many places in the code that use VIR_ALLOC_N and do not check if the sizeof() or count is 0 - which makes me wonder even more about the specific case in which you are referencing. I looked through the various remote_protocol.x structures that would use this and it seems they all use remote_typed_param for the structure being returned, so it's not clear how any of them could have a sizeof() returning 0.