Messages in this thread | | | Date | Thu, 10 Jul 2008 15:06:51 -0700 (PDT) | From | Linus Torvalds <> | Subject | Re: [PATCH] x86_64: fix delayed signals |
| |
On Thu, 10 Jul 2008, Roland McGrath wrote: > > There are many different scenarios that could hit this bug, most of > them races. The simplest one to demonstrate does not require any > race: when one signal has done handler setup at the check before > returning from a syscall, and there is another signal pending that > should be handled. The second signal's handler should interrupt the > first signal handler before it actually starts (so the interrupted PC > is still at the handler's entry point). Instead, it runs away until > the next kernel entry (next syscall, tick, etc).
I have this dim memory of at least _some_ of this being on purpose.
If you look at old kernels (_really_ old ones - I think it's way before even the historical git archive, but I didn't take a look), we used to set up several stack frames at once, so that we'd nest the stack frames completely.
In other words, the code in do_signal() used to literally be a loop, something like
while ((signr = get_signal_to_deliver(&info, &ka, regs, NULL)) > 0) { .. setup signal frame ..
(No, I don't think that's at all accurate of the actual code we used to have - I just took the current do_signal() code as an example)
And that explicit loop was removed in order for us to have just a single outstanding signal at a time. I forget the exact details why.
But if you really want that behaviour, then re-introducing the loop would likely be the better approach (or should be combined), since I think you effectively just re-introduced it (at a much bigger granularity).
Hmm.
Linus
| |