There could be a thousand ways in which you could go about this, but what you eventually do will come down to a matter of both what you desire to achieve and practicality, determined by what equipment you have available, the venue itself, the level of the PA, etc, etc.
Anyway... here's my 2 cents.
To begin with, no matter how you mic it up, you're going to end up with spill from the PA leaking into the mics. This may be what you want ("band in the room sound") or it may be something you wish to minimise. Either way, it is unavoidable that you'll end up with spill in the mics, so perhaps the best way to go about it is to use the spill to your advantage and go for that "band in the room sound." Even if you don't use it (say, you only mix the room mics in when you want crowd noise such as the applause at the end of a tune, and turn them down when you don't) you can still leave the option of getting the band in a room vibe if you want it.
Approaching it from this angle, here's a couple of considerations that come to mind.
If I remember correctly, the human ear needs about 20ms of delay before it is able to detect a difference in timing. What this implies is that if the spill from the PA is being picked up by the mics more than 20ms after the direct sounds you're recording then you're going to hear a slap delay sort of effect on the band as you bring the level of the room mics up. This means that if you stick your room mics any more than about 6 metres from the PA, you're probably going to want to time delay the band mix relative to the room mics. If you're going to be mixing on a DAW or digital console with facilities to delay channels, then I wouldn't worry about it until you're mixing (and simply move the room mic waveforms back a few ms in the DAW until you achieve the desired ambience without getting a slap delay effect or, delay the band mics to match the room mics on the digital console), but if you're recording and mixing on a medium that doesn't allow you to move your sounds in time, then you're probably going to want to keep the mics under 6m from the PA.
Another consideration would be how 'real' you want the final product to be. Of course, you're never going to recreate the live experience in someone's living room, so if you're going for the live vibe, you're probably going to want to try and create a more hyper-real experience - which you will almost certainly create by virtue of the fact that you'll probably derive most of your band mix directly from their individual mics/inputs which will then be listened to at reasonable levels in domestic environments on domestic equipment. One way to go about using your room mics to create a more hyper-real experience would be to try and achieve a sound where the room mics are not only adding ambience to your band mix, but also providing a clear audience sound. Try using cardioid or figure of 8 mics with their dead sides facing the PA. What this will allow you to do is boost the treble in the room mics which will allow the audience more clarity in the mix whilst also partly unmuffling the band sound, which will enchance ambience without having a band mix that bites your ears off because you've recorded a treble heavy band sound in the room.
As for recording the crowd, you're going to want to get your mics into a position that allows them to 'see' as much of the crowd as possible, as evenly as possible. By this I mean you'll probably achieve a much bigger crowd sound using fewer well placed mics elevated well away from the crowd than you will with 40 mics placed to pick up all the different sections individually. Getting the mics so they see as much of the crowd as possible means that no one section is going to stand out any more than the others (which would diminish the size of the other sections of the crowd, which will diminish the perceived size of the crowd as a whole). By doing this you'll also be capturing a natural blending of the crowd in the room in a way which you'll never achieve in your mix room blending multiple mics, once again, making the crowd sound bigger.
As for the surround part, I have no experience or real knowledge of how to go about it, but I can only imagine that in your case simpler would be better. Perhaps just aim for four mics to pick up front left and right, and rear left and right. Perhaps a stereo pair near the front whic you can mix to the front channels, and another stereo pair half way back into the crowd mixed to the rear. Depending on how you go about this you could time align the front mics to the band mix, and have the rear mics so they're delayed a little to the band mix which may help put the listener a little more in the room. Of course, make sure you check the phasing of all your room mics with the band mix, making sure that they're reinforcing the sound of the band and not diminishing it. This includes the rear mics - they may be mixed into the front two channels when listened to in stereo instead of surround.
Oh the possibilities... good luck with it.
