Re: data modified by second NRT write with same memory location

[Philippe Gerum <rpm@xenomai.org>]

yo sang <stlfatboy@hotmail.com> writes:

> Thanks for the reply.
>
> I still have several doubts:
> 1. it's very common to delete data right after calling write(). Data should be completely copied to ensure a successful write() call. 
>

Except that your code is simply not doing that. Please read it
again. Hint: this implementation introduces sharing by mean of a
pointer.

> 2. The fd opened for nrt is blocking by default. If I'm using un-blocking way of write(), it's reasonable to maintain several distinct RTMem instances at write end.
>

This is not the point in discussion. The issue is with the broken
indirection the pointer adds.

> 3. If I modify the RTMem class as below, the result is correct then:
>
> class RTMem
> {
>  public:
>  	RTMem() {}
>
>  	RTMem(const char* data, int size, RTMemType type) {
>  		//Data = Buf;
>  		memcpy(Data, data, size);
>  		Size = size;
>  		Type = type;
>  	}
>
>  	int			Size{};
>  	RTMemType	Type{};
>  	//char*		Data{};	
>  	char		Data[256]{};
>
>  };

Obviously it is, indeed.

-- 
Philippe.