As background, there are two sets of emissions we need to consider: first are the "criteria" or "toxic" emissions. These are regulated. Second are the CO2 emissions--these are not regulated, at lest, not yet.
1. In general, diesel has two main areas of criteria emissions concern compared to gasoline engines: oxides of nitrogen (NOx--which contribute to smog) and particulate matter (PM, also called soot). Both are demonstrated to have bad health effects. Diesels emit more of each than a petrol engine, due to the nature of the combustion process.
Emissions regulations for cars used to treat each as a separate category. That has changed in the US, however, with the current emissions regulations--diesels and petrol cars have the same "tailpipe out" emissions criteria to meet, regardless of the size of the engine. In other words, regardless of the car size or power, they all have to meet the criteria of their category. (Bigger vehicles and diesels thus end up spending more on their exhaust aftertreatment solutions...) In Europe, however, the two are still treated as separate. While the PM standards will be the same, the diesel standards for NOx are more lenient than the petrol standards. In other words, the same car model will emit more NOx as a diesel than as a petrol car.
So the specific answer to #1 is that its the quantity. NOx is NOx, diesels just emit more of it.
2. The increased fuel efficiency of the diesel doesn't really affect the tailpipe out emissions in that cars are built to meet the standards for their category. Standards are specified in so many grams of a pollutant emitted per kilometer. There are some variations, and some ultra-low emissions categories that automakers can shoot for, but again, for most cars currently, the amount of emissions out the tailpipe is specified by the regulations. Automakers build to that spec. Where diesel's efficiency DOES make a big difference, however, is in CO2 emissions. Because diesels burn less fuel for a comparable power output, they emit less carbon dioxide than current petrol engines.
So, to get specific. The current Euro 4 regulations for passenger cars specify 0.25 g/km NOx for diesels, 0.08 g/km for petrol. Assuming you have a Euro 4 car, then, with, say, 20,000km annual driving, you'd produce 5kg of NOx in a diesel, and 1.6 kg of NOx in a petrol car. But taking your 6 and 8 liter numbers, you'd produce about 3,180 kg of CO2 in the diesel, vs. about 3,800 kg CO2 for petrol.
3. This is a very tough question, because so much depends on the nature of the crude oil going in to the refinery. The more sulfur the oil contains, and heavier it is, the more effort has to go into the refining processes...the more energy required to process it (with associated CO2 emissions, etc.). In general, the easiest (read least energy intensive) process is refining light, sweet (low sulfur) crude to gasoline. Unfortunately, there's not much light sweet crude left. (The chief engineer at Shell US told me recently that while a few years ago they could get 10 times the energy out of a barrel of petrol compared to what they put in to processing it, with the tar sands crude that's now hitting refineries in the US, they're expecting maybe 1-2 times. That's a huge swing, and reflects the increasing effort going into ALL refining. Basically, refining is a process of breaking up the amazing number of complex hydrocarbon chains in crude oil into a variety of products, that essentially vary by weight. The lighter hydrocarbons (kerosene, petrol), the middle distillates (diesel), and the really heavy nasty stuff that ends up in asphalt and coke. You can tweak the relative amounts of output by successively upgrading heavier product into lighter using heat, hydrogen and different catalysts.
However, there is a recent update to the Well-to-Wheels lifecycle analysis for the European context that suggests the following:
"Whereas the total amount of energy (and other resources) used by refineries is well documented, there is no simple, non-controversial way to allocate energy, emissions or cost to a specific product. Distributing the resources used in refining amongst the various products invariably involves the use of arbitrary allocation keys that can have a major influence on the results. More to the point, such a simplistic allocation method ignores the complex interactions, constraints, synergies within a refinery and also between the different refineries in a certain region and is likely to lead to misleading conclusions. From an energy and GHG emissions point of view, this is also likely to give an incomplete picture as it ignores overall changes in energy/carbon content of feeds and products....We thus considered that, in the context of this study, the energy and GHG emissions associated with production and use of conventional fuels should be representative of how the EU refineries would have to adapt to a marginal reduction of demand."
"From this analysis it appears that, in Europe, marginal diesel fuel is more energy-intensive than marginal gasoline. In recent years Europe has seen an unprecedented growth in diesel fuel demand while gasoline has been stagnating or even dropping. According to all forecasts, this trend will continue in future years, driven by increased dieselization of the personal car and the growth of freight transport in line with GDP."
This report is available here
http://ies.jrc.cec.eu.int/media/scr.../H04/Well_to_Wheels/WTT/WTT_Report_010307.pdf