Messages in this thread | | | Date | Thu, 28 Dec 2000 12:44:14 -0500 | From | Chris Mason <> | Subject | Re: innd mmap bug in 2.4.0-test12 |
| |
On Thursday, December 28, 2000 16:15:48 +0100 Daniel Phillips <phillips@innominate.de> wrote:
> On Thu, 28 Dec 2000, Rik van Riel wrote: >> On Thu, 28 Dec 2000, Daniel Phillips wrote: >> >> > It's logical that PageDirty should never be get for ramfs, >> >> No. Not setting PageDirty will cause the system to move the >> page to the inactive_clean list and happily reclaim your data. >> >> We _have to_ use something like PageDirty for this, and >> checking for the ->writepage method will even allow us to >> do stuff like dynamically switching swapping support for >> ramfs on/off (or other funny things). > > You're suggesting using the absence of a method as a kind of flag, but > the code is really too full of obscure stuff like that already. > > How about taking an extra user on the ramfs pages instead. It doesn't > sound right to set PageDirty when you are not requesting IO.
I think a dirty page without a writepage func seems a bit broken. How about we give ramfs a writepage func that just returns 1. That way nobody does any special if (ramfs_page(page)) kinds of tests...
-chris
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |