Obviously the cops are in on it. / Obviously the cops just don't believe her but she knows the cameras are there. / Obviously the stalker is monitoring her phone calls to the cops. / Aliens
> I understand I sound like a paranoid woman but really I'm not.
Just because you're paranoid doesn't mean there isn't an invisible demon about to eat your face.
My cat is what protects me from demons. I have chosen to believe that every time she goes absolutely batshit insane running around and attacking the air, she's actually defending me from demons which I cannot see. I think Lovecraft was onto something.
> It is said that in Ulthar, which lies beyond the river Skai, no man may kill a cat; and this I can verily believe as I gaze upon him who sitteth purring before the fire. For the cat is cryptic, and close to strange things which men cannot see.
> Throw a stick, and the servile dog wheezes and pants and shambles to bring it to you. Do the same before a cat, and he will eye you with coolly polite and somewhat bored amusement. And just as inferior people prefer the inferior animal which scampers excitedly because somebody else wants something, so do superior people respect the superior animal which lives its own life and knows that the puerile stick-throwings of alien bipeds are none of its business and beneath its notice.
On a totally unrelated note, the one thing that pisses me off about Linux kernel development more than most is the fact that the damn things still doesn't have a revoke() syscall, despite it being so useful for security.
I fought with myself about whether or not I should just locally maintain my own revoke() patch or just wait until, one day, enough people want it that it is finally adopted upstream. Even the BSDs have it...
It's a syscall that is called against a file and tells the kernel to irreversibly revoke any associated file descriptors. So, for example, you could have Xorg open the master DRM node as root, then drop to non-root, and when the VT switches, the master node's file descriptor can be revoked and Xorg won't be able to do anything about it.
Well so it can't abuse its powers (having an open master node, even if it is not root anymore). Because the checks for a file occur during open(), not a subsequent read() or write(). A revoke() call makes it possible to revoke those privileges from outside.
also, I suppose if you wrote the call in, its going to be a pain to ensure its on every computer you use + you need to hope/test it dosen't break on a new kernel?
Adding a new non-trivial syscall is a maintenance nightmare (though revoke() is not that complex, but it still touches the low-level VFS layer). All the patches I've wrote that I maintain myself are considerably simpler or far more modular, e.g. only involving hooks at key places in the kernel.
I think people tried twice to get a revoke() into the upstream kernel, but both times it was shot down because people were bickering over how exactly a memory-mapped file created with mmap() should behave after revocation (whether reads to the memory backed by the revoked FD should return 0s, or instantly trigger a SIGSEGV, or what).
With the proper solution obviously raising a SIGSEGV, because returning 0s without any warning is a ticking time-bomb...
Hi, I'm studying computer science at university and I'd like to learn about cybersecurity and hacking. What are some good ressources to get started (books, websites, etc.) ?
