Messages in this thread | | | Date | Mon, 10 Oct 2005 16:06:06 +0100 (BST) | From | Hugh Dickins <> | Subject | Re: [RFC]should swap-file opened with O_DIRECT? |
| |
On Mon, 10 Oct 2005, Yingchao Zhou wrote: > > The sys_swapon system call open the swap-file through > filp_open(..., O_RDWR|O_LARGEFILE,...). > In this way, I think that the pageout process of anonymous > pages finally will write them out through (swap-file)->f_ops->write,
No, they are written out by swap_writepage. The opening in sys_swapon gets a handle of what is to be used for swap, but it then goes its own way.
> and it will result in caches of swapfile. However, swapping only > happens when memory is tight. So why not set O_DIRECT? Is there any > special reason to keep caches of swapfile?
Caches of swapfile are useful (but yes, can be discarded when memory is tight). There's nothing indirect about them: the page that is mapped into userspace gets written to disk, the page that is read in from disk gets mapped back into userspace.
Hugh - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |