How does Ethernet know how long a frame is?

The Physical Coding Sublayer is responsible for delimiting the frames, and sending them up to the MAC layer.

In Gigabit Ethernet, for example, the 8B/10B encoding scheme uses a 10 bit codegroup to encode an 8-bit byte. The extra two bits tell whether a byte is control information or data. Control information can be Configuration, Start_of_packet, End_of_packet, IDLE, Carrier_extend, Error_propagation.

That is how a NIC knows where a frame start and ends. This also means that the length of the frame is not known before it has fully decoded, analogous to a NULL-terminated string in C.

The article you really want to answer your question is http://en.wikipedia.org/wiki/Ethernet_II_framing; which says:

As this industry-developed standard went through a formal IEEE standardization process, the EtherType field was changed to a (data) length field in the new 802.3 standard. (Original Ethernet packets define their length with the framing that surrounds it, rather than with an explicit length count.) Since the packet recipient still needs to know how to interpret the packet, the standard required an IEEE 802.2 header to follow the length and specify the packet type.