This may well be a bug in Firefox or some other userspace component, however it is serious enough that the kernel developers ought to know about it. I tried several times to download a Linux kernel tarball; each time Firefox reported that the entire file was downloaded, however the file in my filesystem was quite a lot shorter than the size reported at kernel.org. I speculated that Firefox had a caching bug, and so closed all my tabs one by one, then closed the window. The file was still incomplete. I launched a new Firefox, retried the download, still incomplete. I shut down my box, booted it, then was able to download successfully. Perhaps the problem has something to do with Hibernate; I do know there were some problems in the past with hibernating my motherboard, a SuperMicro X7DWA-N. Under a much earlier kernel, with Fedora 7, while I could go into and out of hibernation, after coming out the networking did not work. With my present Linux Mint Cinnamon 17.1 64-bit install, I've been using hibernation quite a bit. I am just now seeing some evidence that I may have a more serious problem, perhaps I have bad memory or something. I'm going to run Memtest86. I'm using an AMCC3ware 9690-SA 4-disk RAID5, with battery backup so as to enable write caching.
Can you attach here content of /proc/mounts, df, and /var/log/messages from around the time when the problem happened? Also what kernel version do you run (uname -a).
The most likely cause is that you're hitting this long-standing Firefox bug: https://bugzilla.mozilla.org/show_bug.cgi?id=237623 (opened 2004, last marked fixed in February, marked as scheduled for FF 39)