-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dianne Hackborn wrote: [...] > If you turn off over-commit, I believe Android won't even boot on a G1, > because Linux would need to assume that RAM is needed for every mmapped() > thing, and we run out of it well before the system is fully up.
Not necessarily --- my intuitive understanding would be that overcommit would only applies to anonymous storage and not to file storage, as mmapped() files are backed by actual *files* and not by the pool of swap pages. (Besides, most of those mmapped() files that Android's using are read-only and their pages can be silently discarded at any point.) [...] > Put another way: if you are really using so much memory that the kernel > needs to kill your app, you are well past the line of being a good citizen Oh, absolutely. But it would be rather nice to have a medium line between happy-happy-fun-joy and the boys coming round saying, "'ere, Mr. Torvalds would like a *word* with you..." (What we've actually got is a portable API implementation for native apps. The memory allocation function is defined to return NULL on allocation failure. It had never occurred to us that *making* it return NULL on allocation failure could be hard.) - -- ┌─── dg@cowlark.com ───── http://www.cowlark.com ───── │ │ "People who think they know everything really annoy those of us who │ know we don't." --- Bjarne Stroustrup -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iD8DBQFKrslpf9E0noFvlzgRAuuOAKCAEARJSEPcHxcrEowsxwcMBvkxGwCcC1Jz kuny8oLQqn1S8mBDmqp9Bwk= =iCwo -----END PGP SIGNATURE----- --~--~---------~--~----~------------~-------~--~----~ unsubscribe: android-porting+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-porting -~----------~----~----~----~------~----~------~--~---