On 05/06/2014 11:33 AM, Jerome Glisse wrote: > So how can i solve the issue at hand. A device that has its own page > table and can not mirror the cpu page table, nor can the device page > table be updated atomicly from the cpu. Yes such device will exist TLB invalidation on very large systems can already take essentially forever. Are we OK with extending that "forever" period for heterogeneous memory management with crappy devices, or is this something we could/should look into improving in the general case? -- All rights reversed -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>