Re: [PATCH 4.9 00/86] 4.9.88-stable review

From: Greg Kroah-Hartman
Date: Thu Mar 22 2018 - 04:19:35 EST


On Wed, Mar 21, 2018 at 11:20:04PM +0530, Naresh Kamboju wrote:
> On 21 March 2018 at 19:02, Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > On Tue, Mar 20, 2018 at 11:49:19PM +0000, Ben Hutchings wrote:
> >> On Sun, 2018-03-18 at 11:27 +0100, Greg Kroah-Hartman wrote:
> >> > On Sat, Mar 17, 2018 at 03:48:24PM +0530, Naresh Kamboju wrote:
> >> [...]
> >> > > Results from Linaroâs test farm.
> >> > > No regressions on arm64, arm, qemu_x86_64 and x86_64.
> >> > >
> >> > > NOTE:
> >> > > CONFIG_LEGACY_VSYSCALL_EMULATE=y is enabled from config as default.
> >> > > Enabled vsyscall=native and vsyscall=none from kernel command line for testing
> >> > > selftests/x86/test_vsyscall test case on x86_64 and qemu_x86_64.
> >> > >
> >> > > test_vdso_64 failed on x86_64 device when vsyscall=none is enabled.
> >> > > We will investigate this new test failure.
> >> >
> >> > I think vsyscall=none just went away :)
> >>
> >> No, vsyscall=native went away upstream. vsyscall=none should still
> >> work everywhere.
> >
> > Ah, sorry, got that wrong, thanks for the correction.
> >
> >> However, test_vdso_64 isn't able to detect what the current
> >> configuration is and it will crash with vsyscall=none (or the
> >> equivalent default in kconfig).
>
>
> Shall i skip running "test_vdso_64" when vsyscall=none ?
> or
> Shall we fix the test case to work when vsyscall=none ?

Why not start a new email thread with the author of that test to get the
right people involved here?

But first verify that it all works on the 4.15.y tree.

thanks,

greg k-h