On Tue, Sep 11, 2012 at 03:05:54PM +0300, Avi Kivity wrote: > On 09/11/2012 12:57 PM, Gleb Natapov wrote: > > >> > If you get -ENOMEM when allocating a page without GFP_ATOMIC (or > >> > GFP_NOIO etc) then the entire host is dead anyway. The same thing can > >> > happen if the guest (or userspace) touches a yet-unallocated page, or if > >> > the page fault path fails to allocate mmu pages, or any of a thousand > >> > other allocations we have all over. > >> Then it is just simpler to sigkill the guest right away. What's the > >> point in returning error if you believe that userspace can't handle it > >> and will likely not run long enough to even get to userspace due to > >> memory shortage. > >> > > And although this is not the route I will go the question remains. How > > do we return to the kernel after userspace exit in the middle of IO that > > is handled by an in-kernel device. Looks like the kernel will expect > > emulation result from userspace on next ioctl(RUN). > > Option 1 is to rewind everything to before the instruction. Option 2 is > to document that errors are not recoverable. > This will make all error non recoverable since userspace has no way to know that error happened during IO emulation. So we either rewind everything to before the instruction or never return an error to userspace during IO. -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
