Increase write flash request timeout to 2.5s, fixes #98

[ntamas]

As discussed in #98, this PR raises the timeout for "write flash" requests to 2.5 seconds unconditionally to cater for the increased time needed to erase a page of 128K on an STM32F405.

Note that the timeout is applied unconditionally, even if we are flashing to the nRF51. In my opinion this is not a big deal because the flash packets go through just fine almost all the time, so there is almost never any need for re-sending them. However, re-sending them too early when the packet in fact went through can break the flashing process, as indicated by #98.

[ataffanel]

Thanks a lot, this is going to help a lot of people and save a lot of time!
I merged the PR, lets keep the issue open a little while to make sure the problem is indeed gone.

[ntamas]

I'm fine with keeping the issue open but you just closed it ;)

[ataffanel]

Thanks for the heads-up, it was automatically closed by github because of this PR title :-).