On Fri, 7 Jun 2019, Paul Moore wrote: > On Thu, Jun 6, 2019 at 8:45 PM James Morris <jmorris@xxxxxxxxx> wrote: > > On Wed, 5 Jun 2019, Paul Moore wrote: > > > On Wed, Jun 5, 2019 at 1:05 PM Casey Schaufler <casey@xxxxxxxxxxxxxxxx> wrote: > > > > On 6/5/2019 9:51 AM, Janne Karhunen wrote: > > > > > > > > One hook with an added "bool blocking" argument, if > > > > that's the only difference? > > > > > > I think there is value in keeping a similar convention to the notifier > > > code on which this is based, see include/linux/notifier.h. > > > > Although this doesn't seem to be what other users in the kernel are doing. > > How many of them potentially have the need for both blocking and > non-blocking notifiers? I didn't go through the entire list of > callers, but it seems all that I looked at used only one type. The > simple fact that we started with one type of notifier for the LSM, and > we are now switching to the other (and getting lucky that it is safe > to do so for the existing callers) seems to lend some weight to the > argument we may need both and adding "block"/"blocking"/etc. to the > name has value. Fair enough. -- James Morris <jmorris@xxxxxxxxx>