missmatch in enum error values

[LuisAyuso]

Hi there,

I am running a program which contains some error, and the Cuda runtime keeps returning MissingConfiguration

I can see in the cuda_sys code that this enum value maps to 2, nevertheless MissingConfiguration in the Cuda runtime maps to 52, while 2 is

    /**
     * The API call failed because it was unable to allocate enough memory to
     * perform the requested operation.
     */
    cudaErrorMemoryAllocation             =      2,

    /**
     * The device function being invoked (usually via ::cudaLaunchKernel()) was not
     * previously configured via the ::cudaConfigureCall() function.
     */
    cudaErrorMissingConfiguration         =      52

In the rust code there are no 52 nor 53 values:

    PeerAccessAlreadyEnabled = 50,
    PeerAccessNotEnabled = 51,
    DeviceAlreadyInUse = 54,
    ProfilerDisabled = 55,

I am using cuda 10.1.

• is this version supported?
• is this a version missmatch?
• Should I generate the bindings myself from the Cuda header?

Cheers