On Wed, Sep 20, 2017 at 2:08 AM, David Miller <da...@davemloft.net> wrote: > From: Petar Penkov <peterpenko...@gmail.com> > Date: Tue, 19 Sep 2017 21:26:14 -0700 > >> Furthermore, in a way testing already requires specific kernel >> configuration. In this particular example, syzkaller prefers >> synchronous operation and therefore needs 4KSTACKS disabled. Other >> features that require rebuilding are KASAN and dbx. From this point >> of view, I still think that having the TUN_NAPI flag has value. > > Then I think this path could be enabled/disabled with a runtime flag > just as easily, no?
I think that the compile time option was chosen because of the ns_capable check, so that with user namespaces unprivileged processes can control this path. Perhaps we can require capable() only to set IFF_NAPI_FRAGS. Then we can convert the napi_gro_receive path to be conditional on a new IFF_NAPI flag instead of this compile time option.
