BOAC has already described one form of high speed stall.
Another meaning can be a stall where the flow behaviour has beome (significantly) Mach dependent. The classically simple stall occurs, as has been said, where the angle of attack exceeds a critical value, causing widespread flow separation and lift loss. It can often be assumed that for relatively low speeds the critical value is constant for a given configuration, and so the stall speed can be expressed as a single value of EAS, valid for all altitudes.
However, that's not entirely valid; as the mach number is increased the wing will likely stall at lower angles of attack, such that at some point the stall speed, expressed in EAS, will start to increase. I've heard this called "shock stall", a "Mach stall" or a "high speed stall" - all terms to differentiate it from the more common low speed version.
The flow characteristics are similar at a gross level - separation of the boundary layer and loss of lift - but the mechanism may be different - the boundary layer separation is likely to be triggered by shocks, not by the normal mechanism of an adverse pressure gradient.
The recovery requires, as always with a stall, reduction of the angle of attack in order to get the flow back into the attached flow condition. But depending upon the exact conditions, other techniques which also help in recovery - such as adding power and building speed margin to prevent secondary stalls and regain additional control power for the low speed stall - may be inappropriate at higher speeds.