Say, in the year 925 was there some sort of clever chimney system, perhaps a high ceiling that allowed a space under the smoke, or did they simply deal with a quite smoky house because it was preferable to the cold?
Longhouses are very interesting examples of natural ventilation. Characteristic of Scandinavian longhouses is the lack of chimneys. Instead, ventilation is achieved, as you guess, by a high ceiling and two openings, one at each end called "vindöga/vindue/vindauga" (in Swedish/Danish/Norwegian respectively). This is by the way where the english word "window" comes from.
Since the ceilings are relatively high and the houses not nearly as insulated as a modern one the hot smoke rises towards the ceiling and escapes through (usually) one of these openings, depending on wind direction mostly. Cold air is sucked in through cracks and openings. This is a typical and early example of natural ventilation. Doesn't work very well if the outdoor temperature is higher than the indoor temperature but that's rarely much of a problem in scandinavia.
NB: I can not provide any english sources for this, the best I can do is Danish and Swedish archeology papers. If anyone can help me find a good english source I'd be greatful.
Viking longhouses had a small opening either at both or just one of the gables of the roof, called a vind-ue or vindauga (wind-eye) from which smoke would escape (this is also the source for the English word window).
Chimneys would arrive to the cities of former viking countries (Scandinavia) during the late medieval era and to the countryside in the early 18th century.
Experimental archeology from the Swedish viking age trading town of Birka where archeologists and artisans have reconstructed the houses of the town have shown that for the most part, the wind-eye is capable of venting smoke. However, the hearth in the middle of the room was not very effective, and the amount of labour needed to fell trees, chop firewood and store firewood to burn led the archeologists to estimate that with the animals indoors, temperatures in a longhouse would be 12-14 degrees celsius in winter.
Sources:
There is exceptionally little archaeological evidence for anything higher than ca. 10cm from the floor. In some cases we have some scraps, such as a house in Hedeby where an entire wall collapsed flat (Schietzel, Berichte über die Ausgräbungen 16, 1981), or some single wooden rooftiles in a ditch. When archaeology fails, there are few other sources to reconstruct no-longer standing buildings. One is images. Unfortunately, while Viking Scandinavia is fairly rich in images (the Gotlandic picture stones, or artistic pieces like the Franks Casket, or even coinage), houses are always shown very stylised, if at all, which does not really help much for our purpose. One other iconographic category which is really helpful are a type of stone grave-marker from Northern England called a hogback stone. These stones were shaped in the form of a building, often emphasising the roof. They correspond fairly well with what we expect a high-status hall-building to look like, with curved walls and a curved roof. These stones always show shingles (presumably of wood) as a roof covering, and show no opening at all in the middle for smoke. It is possible some shingles might have been raised slightly to allow for ventilation, but this is not visible. It is also possible that there were openings in the gable ends, but in the stones the gables are obscured by artistic elements, such as standing bears.
Another source might be texts, such as the Beowulf or Icelandic sagas, but I am not familiar enough with them to know whether any refer to ventilation, smoke, or holes in roofs.
Beyond the primary evidence we can also do experimental archaeology, which is simply building a house the way we deem most plausible, and see whether the result is practical. When it is not, the design must be modified, under the assumption that people in the past must have lived in liveable conditions and that there is a fault in our modern reconstruction.
The most rigorous experimental archaeology of a variety of roof constructions has been performed in Lejre, Denmark. Here it was shown that thatched roofs (such as described for the Migration-period hall in Beowulf) are actually fairly permeable to smoke, and no opening is necessary. Yet on a whole, it seemed that people just had to deal with smoky rooms, particularly on damp, windless days. The report on indoor climatic conditions can be read online here.
