On Tue, Jul 14, 2020 at 10:03 AM Brian Gerst <brge...@gmail.com> wrote:> > On Tue, Jul 14, 2020 at 2:40 AM Christoph Hellwig <h...@lst.de> wrote: > > > > On Tue, Jun 16, 2020 at 10:23:13AM -0400, Brian Gerst wrote: > > > Christoph Hellwig uncovered an issue with how we currently handle X32 > > > syscalls. Currently, we can only use COMPAT_SYS_DEFINEx() for X32 > > > specific syscalls. These changes remove that restriction and allow > > > native syscalls. > > > > Did this go anywhere? > > This approach wasn't well received, so I'd just go with this as the > simplest solution: > https://lore.kernel.org/lkml/CAK8P3a17h782gO65qJ9Mmz0EuiTSKQPEyr_=nvqotnmqzuh...@mail.gmail.com/ >
I'm okay with either approach, although I think the original approach is nicer than the simplified #define approach. In my mind, the __x64_omg_so_many_underscores prefixes really mean "don't think too hard about these -- we just decided to make extra long names", so whatever. We can clean it up more some day. > -- > Brian Gerst
