Hi,
On Wed, Sep 27, 2023 at 03:43:29PM +0200, Kamil Paral wrote:
> On Wed, Sep 27, 2023 at 7:16 AM Mika Westerberg
> <[1]mika.westerberg@xxxxxxxxxxxxxxx> wrote:
>
> > I would also try to change all the BIOS settings back to defaults, see
> > that it works (it is probably in "user" security level then), then
> > switch back to "secure" (only change this one option) and try if it now
> > works. It could be that some setting just did not get commited properly.
>
> Hello, I have some interesting findings.
> I reset the BIOS ("load BIOS defaults"), however that doesn't reset the TB
> security level option - whatever is set is kept. So I can't really find
> out what the default value is and whether I changed it. However, to my
> surprise, the resume was suddenly fast even with the Secure Connect
> security level. So I went through options one by one and finally
> discovered that the "Wake by Thunderbolt 3" option makes the difference.
> It has the following description:
> "Enable/Disable Wake Feature with Thunderbolt 3 ports. If Enabled, the
> battery life during low power state may become shorter."
> It is enabled by default, I had it disabled before. When I enable it, I
> have a 5 sec resume even with Secure Connect TB security level. The dmesg
> for such a resume is here:
> [2]https://bugzilla-attachments.redhat.com/attachment.cgi?id=1990815
Okay that now looks like what is expected. The firmware re-connects the
tunnels and the resume can continue.
I actually don't know what that option does either but I would expect
that it leaves the PME and small power enabled for the port so that
probably allows the firmware to do this even in "secure" mode or so.
> I'm sorry I haven't spotted and tested such an obviously named option
> earlier. My understanding was that this option either wakes up my laptop
> when I connect the cable (tested now, it doesn't appear to do so), or that
> it wakes it up when I move/click my dock-connected mouse (tested, doesn't
> appear to do so). So originally I wanted this disabled (and I had no
> kernel issues with it disabled, prior to e8b908146d44), but it seems I
> misunderstood what it does. And I'm still unsure what exactly it's
> supposed to do and why it should kill my battery life in the process,
> according to its description.
> Hopefully this helps to bring more light into this?
> However, I also discovered that even though Wake by TB3 + Secure Connect
> gives me a fast resume in a normal case, I still see a 60+ sec resume
> delay when I disconnect and then reconnect the TB cable while the laptop
> is suspended (which is a frequent use case for me). That doesn't happen
> with the User Authorization level (regardless of Wake by TB value). So I'm
> staying with User Authorization for the moment.
If you apply the patch from here:
https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=pm&id=6786c2941fe1788035f99c98c932672138b3fbc5
Does the problem go away with the disconnect case too (and so that you
have "secure" enabled)?
