On Wed 2016-03-09 12:16:47, Torsten Duwe wrote: > On Wed, Mar 09, 2016 at 11:13:05AM +0100, Jiri Kosina wrote: > > On Wed, 9 Mar 2016, Torsten Duwe wrote: > > > was my first choice. Arguments on the stack? I thought we'll deal with them > > > once we get there (e.g. _really_ need to patch a varargs function or one > > > with a silly signature). > > > > Well, the problem is, once such need arises, it's too late already. > > No, not if it's documented. > > > You need to be able to patch the kernels which are already out there, > > running on machines potentially for ages once all of a sudden there is a > > CVE for >8args / varargs function. > > Then you'd need a solution like I sent out yesterday, with a pre-prologue > caller that pops the extra frame, so the replacement can be more straight- > forward. Or you can just deal with the shifted offsets in the replacement. > > I'll try to demonstrate the alternative. That would then be required for > _all_ replacement functions. Or can the live patching framework differentiate > and tell ftrace_caller whether to place a stack frame or not? > > Miroslav? Petr? Can we have 2 sorts of replacement functions? I personally prefer to keep most functions without any special hack, especially when it is needed only for one architecture. If a hack is needed for "corner cases" and it is documented then, IMHO, we could live with it for some time. We test all patches anyway, so. But I could not speak for the LivePatching maintainers whose are Josh and Jiri. Best Regards, Petr -- To unsubscribe from this list: send the line "unsubscribe live-patching" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
