[Cryptography] Swap space (Re: It's all K&R's fault)

[Jerry Leichter]

On Apr 24, 2014, at 5:13 PM, Nemo <nemo at self-evident.org> wrote:
> ...(Prior
> to Linux, no O/S even considered the "overcommit_memory" approach
> because, let's face it, it's idiotic.)
Actually, AIX implemented overcommitted memory *way* before Linux did.  In fact, it originally *only* implemented overcommitted memory:  When you did a large malloc(), you probably got memory pages that would be committed on first access (or your process would die at that point).

This cause so many problems with existing code that it was, as I recall, eventually made an option.

Elsewhere you mention the Linux OOM killer process.  This, in fact, was also an AIX innovation, many years ago.  It (and a number of other system-management-related features) comes from the mainframe world, where you expect the OS to do its best to keep as much running as possible, even under severe stress.
                                                        -- Jerry

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4813 bytes
Desc: not available
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20140424/dfe9908e/attachment.bin>