Re: A true story of a crash.

Andi Kleen (ak@muc.de)
16 Aug 1998 19:36:02 +0200

In article <m0z88Qk-000aNHC@the-village.bc.nu>,
alan@lxorguk.ukuu.org.uk (Alan Cox) writes:
>> What if, instead of immediately sending SIGKILL, it sent a gradually
>> escalating series of signals, accompanied by suitable printk()s, over a
>> period of a few tens of seconds (stopping the process if enough memory
>> were freed), with SIGKILL only as a last resort?

> On both out of memory and stack limit exceeded you can only send SIGKILL
> Think about what happens if you need to allocate a page of user stack..

If the user set an alternate signal stack with sigaltstack() then you can
send something else (except for out of physical memory when there is no
chance to swap the potentially swapped out signal stack page in, but
if that happens there is not much the kernel can do anyways)

-Andi

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html