On Wed, Jun 16, 2010 at 12:59:54PM -0400, Rik van Riel wrote: > __GFP_IO can wait for filesystem activity Hmm I think it's for submitting I/O, not about waiting. At some point you may not enter the FS because of the FS locks you already hold (like within writepage itself), but you can still submit I/O through blkdev layer. > __GFP_FS can kick off new filesystem activity Yes that's for dcache/icache/writepage or anything that can reenter the fs locks and deadlock IIRC. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>