From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <452F323D.2040600@yahoo.com.au> Date: Fri, 13 Oct 2006 16:29:17 +1000 From: Nick Piggin MIME-Version: 1.0 Subject: Re: Driver-driven paging? References: <452A68E9.3000707@tungstengraphics.com> <452A7AD3.5050006@yahoo.com.au> <452E8849.8050201@surriel.com> In-Reply-To: <452E8849.8050201@surriel.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org Return-Path: To: Rik van Riel Cc: Thomas Hellstrom , linux-mm@kvack.org List-ID: Rik van Riel wrote: > Nick Piggin wrote: > >> Your best bet might be to have a userspace "memory manager" process, >> which >> allocates pages (anonymous or file backed), and has your device driver >> access them with get_user_pages. The get_user_pages takes care of >> faulting >> the pages back in, and when they are released, the memory manager will >> swap them out on demand. > > > Wouldn't tmpfs be simpler ? It could be... actually having the pages inserted into a tmpfs filesystem by the kernel does sound better than having it try to use the swap code directly. I still don't know that tmpfs could quite handle that yet, but it does sound like an interesting avenue (or maybe making your own filesystem using some tmpfs interfaces). Good idea. For an initial cut, I think having a memory manager process will work today, and should do everything needed. So it might be a good way to quickly evaluate the functionality. -- Send instant messages to your online friends http://au.messenger.yahoo.com -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org