On Wed, Sep 05, 2012 at 04:35:49PM +0300, Haggai Eran wrote: > The following short patch series completes the support for allowing clients to > sleep in mmu notifiers (specifically in invalidate_page and > invalidate_range_start/end), adding on the work done by Andrea Arcangeli and > Sagi Grimberg in http://marc.info/?l=linux-mm&m=133113297028676&w=3 > > This patchset is a preliminary step towards on-demand paging design to be > added to the Infiniband stack. Our goal is to avoid pinning pages in > memory regions registered for IB communication, so we need to get > notifications for invalidations on such memory regions, and stop the hardware > from continuing its access to the invalidated pages. The hardware operation > that flushes the page tables can block, so we need to sleep until the hardware > is guaranteed not to access these pages anymore. Since people have been asking about the need for on demand paging in devices: this can be useful for KVM where we sometimes want to let guest directly (bypassing the hypervisor) control a virtual function of a PCI device (prevented by an iommu from accessing host memory). Currently this means host needs to pin all guest memory that *might* be used by this virtual function which breaks setups with memory overcommit; at the moment we can address this by means of ballooning - cooperative memory management - but this has some obvious problems: for example, to get some memory out of a low priority guest it needs to run so the balloon can get inflated. By comparison on demand paging would not require guest cooperation. The problem is not specific to Infiniband; addressing it by on demand paging does require hardware and host driver support though. If Infiniband drivers happen to be the first to implement it more power to them. -- MST -- 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>