On Thu, Sep 06, 2018 at 02:43:33PM -0000, D. J. Bernstein wrote:
> Mika Westerberg writes:
> > The problem here is that the S3 path which may accidentally work may
> > break something else because it was never tested by Lenovo.
>
> No, that isn't what Lenovo says. The BIOS documentation for this laptop
> even mentions S3. What Lenovo says is that they kicked S3 out of DSDT to
> work around a Windows bug. ("Dear customers, When ACPI table was changed
> to support S3, Windows Modern Standby doesn't work. Modern Standby and
> Suspend/Resume functions are working exclusively.")
It does not say "work around BIOS bug" but instead that Windows treats
the two exclusively and I suppose they prefer Modern Standby over S3.
> The links also show Linux users consistently being happier with this
> laptop after enabling S3 through a DSDT override. There are also enough
> tests to confidently say that this kernel patch also enables S3 on this
> laptop, without a DSDT override.
>
> The concerns I was expressing about testing come from the possibility
> that the patch accidentally affects _other_ laptops---that I might have
> screwed up something in the DMI part or the setup part.
I don't doubt the patch does not do what is says, I just expressed my
conserns of going from tested path to something not tested (they
obviously did not test S3 in Windows if they wanted to use Modern
Standby then S3 was not an option).
I actually have this particular system and to me s2idle works pretty OK
but I haven't measured the power consumed nor I have run full distro so
it could be that I missed something.
> > One thing that consumes quite lot power are the two Thunderbolt ports on
> > that system. They are always present (unless you go and switch to legacy
> > Thunderbolt mode from the BIOS)
>
> This is only part of the problem. The links that I provided show tons of
> complaints about Linux s2idle power consumption on this laptop _after_
> people turn on Thunderbolt assist in the BIOS. Note that we're talking
> about a flagship Lenovo laptop here, not something obscure.
>
> Also, do you have any idea why Linux s2idle doesn't wake up on this
> laptop under Xen? For me, adding S3 support made the difference between
>
> * a laptop that I'm using with Linux every day and
> * a useless un-suspendable laptop.
One guess is that because s2idle is handled entirely in the OS so BIOS
and/or possible hypervisor is not involved. But I don't know Xen at all
so could be something totally different.
> Evidently there's still interesting work to do on Linux s2idle even
> after years of development. Perhaps enough s2idle development will
> remove the need for S3 on this laptop. However, I don't see a schedule
> for making this happen. The current situation is that the S3 patch is
> working and s2idle isn't.
Well, s2idle works but it consumes more power ;-)
