From: Nick Piggin <[EMAIL PROTECTED]> Date: Fri, 18 May 2007 07:12:38 +0200
> The page->virtual thing is just a bonus (although have you seen what > sort of hoops SPARSEMEM has to go through to find page_address?! It > will definitely be a win on those architectures). If you set the bit ranges in asm/sparsemem.h properly, as I have currently on sparc64, it isn't bad at all. It's a single extra dereference from a table that sits in the main kernel image and thus is in a locked TLB entry. SPARSEMEM_EXTREME is pretty much unnecessary and with the virtual mem-map stuff the sparsemem overhead goes away entirely and we're back to "page - mem_map" type simple calculations obviating any dereferencing advantage from page->virtual. > 0.2% of memory, or 2MB per GB. But considering we already use 14MB per > GB for the page structures, it isn't like I'm introducing an order of > magnitude problem. All these little things add up, let's not suck like some other OSs by having that kind of mentality. Show me instead a change that makes page struct 8 bytes smaller :-)))) - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
