Messages in this thread | | | Date | Mon, 24 Jan 2000 11:34:09 +0000 (GMT) | From | Tigran Aivazian <> | Subject | Re: [patch-2.3.40-pre6] kzalloc() (ala kmem_zalloc() of SVR4) |
| |
(dropped Linus from cc as I believe it is not a very important topic)
On 24 Jan 2000, Jes Sorensen wrote: > Jeff> I proposed something along these lines several months ago, poke > Jeff> around for the thread. While I agree with the patch, because it > Jeff> will save me a few lines of code here and there, an idea that > Jeff> came out of the previous discussion is also worth considering -- > Jeff> zeroing pages during system idle time, and then handing those > Jeff> out from kzalloc() (or kmalloc(GFP_ZERO|GFP_xxx) or whatever) > > Another thing I forgot to bring up earlier in the dicussion, most of > the places doing kmalloc();memset(); are in initialization parts of > the code, ie. they will be marked __init, so the optimization is a > no-win, the only thing you save is to write one line less of code.
There are plenty of cases where it is used without __init. However, now that I hear Jeff's idea of GFP_ZERO flag to kmalloc() I agree that it is superiour. There is no need to invent (and export) a new function (with a funny name as well) where the old one will suffice. Also, it saves an extra function call.
Regards, Tigran.
- 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.tux.org/lkml/
| |