You use the verb "to be," which is rather deceptive in this situation (and in questions of optics generally). The room itself would not "be dark" or "be light," it would be a collection of various particles, some of which would be photons in the visible spectrum. It really only makes sense to ask how the room would appear to an observer looking through the viewport.
Even if you had completely reflective surfaces, the way the room appeared would depend significantly on the orentientation of the mirrors, the orientation of the viewing hole, the focus of the light, and perhaps most importantly, the refractive and diffusive properties of the mirrors. Ultimately, the appearance of the room would be a result of the light that falls upon the viewport. While there are an infinite variety of possible arrangements, the two extremes roughly align with the two possibilities you suggest.
For the room to look completely dark, no light paths fall on the viewport. A laser perpendicular to two parallel mirrors with the line of sight also parallel to the mirrors would produce this effect.
For the room to look completely bright, all (or however much you require to meet that definition) light paths fall on the viewport. A room in the shape of a truncated paraboloid with the viewport at the focal point would produce this effect.
I decided to move these up to my answer to avoid a prolonged comment conversation.
Your edits do help, but still do not provide enough specificity. If the viewport is coincident with the wall (that is with no collimating effect), it would be possible to see the lightbulb so it would certainly not be totally dark. Additionally, assuming that the mirrors are perfectly reflective with no effective refraction (no beveling of the mirror surface, etc.) the room would appear mostly bright. However, the exact nature of the appearance still depends on the orientation of the viewport and light source. The most "average" appearance would be a line of bulbs of decreasing size.
It is certainly not the case, however, that "light is getting reflected from every place possible." Stand in between two parallel mirrors and see what happens. Now imagine that the image does not get successivly weaker (due to imperfect reflectivity). Even though there are a large number of reflections (not infinite but a lot), they do not take up all of the mirror's surface. The situation would slightly change with mirrors on all sides, but there would still be dark spots... again depending on the arrangement.
Sadly, however, my response even includes unstated assumptions; I have, for example, assumed a macroscopic viewport. It is, I think, unwise to use words like "infinitesimal" when talking about optics. Optical phenomena are inherently macroscopic. It just doesn't make sense to talk about an infinitesimal viewport and an isotropic light. These notions are helpful for field theories like electromagnetism, but do not apply well to the aggregate behavior of photons.