Thinkpad X1 5th gen TPM chip

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Thinkpad X1 5th gen TPM chip

Edd Barrett-3
Hi,

I was looking at the manual page for tpm(4) and noticed that it has
something to to with suspend/resume:

>      Functionality is limited to instructing the device to save its
>      state before a system suspend.

The X1 5th generation has a MSFT0101 TPM chip which we don't yet attach
a driver to. tpm(4) seems happy to attach to with the following small
change. Then in dmesg I have:

tpm0 at acpi0: TPM_ addr 0xfed40000/0x5000: device 0x001b15d1 rev 0x10

There are no errors relating to tpm in the dmesg.

Sadly having tpm(4) attached doesn't fix suspend/resume on this machine,
but I wonder if the tpm driver should be used on this machine anyway? I
know nothing about tpm (really, I'm just stabbing in the dark here), so
I'm hoping someone else will have more of an idea.

Thanks


Index: tpm.c
===================================================================
RCS file: /cvs/src/sys/dev/acpi/tpm.c,v
retrieving revision 1.3
diff -u -p -r1.3 tpm.c
--- tpm.c 1 Jul 2018 19:40:49 -0000 1.3
+++ tpm.c 11 Mar 2019 10:57:38 -0000
@@ -194,6 +194,7 @@ const char *tpm_hids[] = {
  "BCM0102",
  "NSC1200",
  "ICO0102",
+ "MSFT0101",
  NULL
 };
 

--
Best Regards
Edd Barrett

http://www.theunixzoo.co.uk

Reply | Threaded
Open this post in threaded view
|

Re: Thinkpad X1 5th gen TPM chip

Mike Larkin-2
On Mon, Mar 11, 2019 at 11:24:42AM +0000, Edd Barrett wrote:

> Hi,
>
> I was looking at the manual page for tpm(4) and noticed that it has
> something to to with suspend/resume:
>
> >      Functionality is limited to instructing the device to save its
> >      state before a system suspend.
>
> The X1 5th generation has a MSFT0101 TPM chip which we don't yet attach
> a driver to. tpm(4) seems happy to attach to with the following small
> change. Then in dmesg I have:
>
> tpm0 at acpi0: TPM_ addr 0xfed40000/0x5000: device 0x001b15d1 rev 0x10
>
> There are no errors relating to tpm in the dmesg.
>
> Sadly having tpm(4) attached doesn't fix suspend/resume on this machine,
> but I wonder if the tpm driver should be used on this machine anyway? I
> know nothing about tpm (really, I'm just stabbing in the dark here), so
> I'm hoping someone else will have more of an idea.
>
> Thanks
>

What is broken with suspend/resume on this machine?

>
> Index: tpm.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/acpi/tpm.c,v
> retrieving revision 1.3
> diff -u -p -r1.3 tpm.c
> --- tpm.c 1 Jul 2018 19:40:49 -0000 1.3
> +++ tpm.c 11 Mar 2019 10:57:38 -0000
> @@ -194,6 +194,7 @@ const char *tpm_hids[] = {
>   "BCM0102",
>   "NSC1200",
>   "ICO0102",
> + "MSFT0101",
>   NULL
>  };
>  
>
> --
> Best Regards
> Edd Barrett
>
> http://www.theunixzoo.co.uk
>

Reply | Threaded
Open this post in threaded view
|

Re: Thinkpad X1 5th gen TPM chip

Edd Barrett-3

Hi Mike,

We've actually spoken about this issue before on this thread:
 http://openbsd-archive.7691.n7.nabble.com/Thinkpad-X1-Carbon-5th-Gen-Flaky-Suspend-td334772.html

The symptoms remain the same and a colleague who also runs OpenBSD on an identical machine has the same issue.

Some time ago, I tried your strategy of moving the LED indicator flashing around to see if we could see where it gets stuck.

I got as far as deducing that the problem lies in a device driver (as opposed to stuff in the leadup to waking up devices), before a failed wake caused my disk to corrupt to the point where it wouldn't boot.

I've not had the time or the guts to try again, but Google searches suggest that people have had similar issues with Linux and even Windows on this model, so I'm not even certain it's a software issue. Then again, two (different) colleagues run Arch Linux on their X1 5gs and have no problem suspending.

Cheers!

Edd Barrett

Mon Mar 11 23:47:54 GMT 2019 Mike Larkin <[hidden email]>:

> On Mon, Mar 11, 2019 at 11:24:42AM +0000, Edd Barrett wrote:
> > Hi,
> >
> > I was looking at the manual page for tpm(4) and noticed that it has
> > something to to with suspend/resume:
> >
> > > Functionality is limited to instructing the device to save its
> > > state before a system suspend.
> >
> > The X1 5th generation has a MSFT0101 TPM chip which we don't yet attach
> > a driver to. tpm(4) seems happy to attach to with the following small
> > change. Then in dmesg I have:
> >
> > tpm0 at acpi0: TPM_ addr 0xfed40000/0x5000: device 0x001b15d1 rev 0x10
> >
> > There are no errors relating to tpm in the dmesg.
> >
> > Sadly having tpm(4) attached doesn't fix suspend/resume on this machine,
> > but I wonder if the tpm driver should be used on this machine anyway? I
> > know nothing about tpm (really, I'm just stabbing in the dark here), so
> > I'm hoping someone else will have more of an idea.
> >
> > Thanks
> >
>
> What is broken with suspend/resume on this machine?
>
> >
> > Index: tpm.c
> > ===================================================================
> > RCS file: /cvs/src/sys/dev/acpi/tpm.c,v
> > retrieving revision 1.3
> > diff -u -p -r1.3 tpm.c
> > --- tpm.c 1 Jul 2018 19:40:49 -0000 1.3
> > +++ tpm.c 11 Mar 2019 10:57:38 -0000
> > @@ -194,6 +194,7 @@ const char *tpm_hids[] = {
> > "BCM0102",
> > "NSC1200",
> > "ICO0102",
> > + "MSFT0101",
> > NULL
> > };
> >
> >
> > --
> > Best Regards
> > Edd Barrett
> >
> > http://www.theunixzoo.co.uk
> >
>

Reply | Threaded
Open this post in threaded view
|

Re: Thinkpad X1 5th gen TPM chip

Mike Larkin-2
On Tue, Mar 12, 2019 at 12:35:12AM +0000, Edd Barrett wrote:

>
> Hi Mike,
>
> We've actually spoken about this issue before on this thread:
>  http://openbsd-archive.7691.n7.nabble.com/Thinkpad-X1-Carbon-5th-Gen-Flaky-Suspend-td334772.html
>
> The symptoms remain the same and a colleague who also runs OpenBSD on an identical machine has the same issue.
>
> Some time ago, I tried your strategy of moving the LED indicator flashing around to see if we could see where it gets stuck.
>
> I got as far as deducing that the problem lies in a device driver (as opposed to stuff in the leadup to waking up devices), before a failed wake caused my disk to corrupt to the point where it wouldn't boot.
>
> I've not had the time or the guts to try again, but Google searches suggest that people have had similar issues with Linux and even Windows on this model, so I'm not even certain it's a software issue. Then again, two (different) colleagues run Arch Linux on their X1 5gs and have no problem suspending.
>
> Cheers!
>
> Edd Barrett
>

I see. Good luck in your diagnosis!

-ml



> Mon Mar 11 23:47:54 GMT 2019 Mike Larkin <[hidden email]>:
>
> > On Mon, Mar 11, 2019 at 11:24:42AM +0000, Edd Barrett wrote:
> > > Hi,
> > >
> > > I was looking at the manual page for tpm(4) and noticed that it has
> > > something to to with suspend/resume:
> > >
> > > > Functionality is limited to instructing the device to save its
> > > > state before a system suspend.
> > >
> > > The X1 5th generation has a MSFT0101 TPM chip which we don't yet attach
> > > a driver to. tpm(4) seems happy to attach to with the following small
> > > change. Then in dmesg I have:
> > >
> > > tpm0 at acpi0: TPM_ addr 0xfed40000/0x5000: device 0x001b15d1 rev 0x10
> > >
> > > There are no errors relating to tpm in the dmesg.
> > >
> > > Sadly having tpm(4) attached doesn't fix suspend/resume on this machine,
> > > but I wonder if the tpm driver should be used on this machine anyway? I
> > > know nothing about tpm (really, I'm just stabbing in the dark here), so
> > > I'm hoping someone else will have more of an idea.
> > >
> > > Thanks
> > >
> >
> > What is broken with suspend/resume on this machine?
> >
> > >
> > > Index: tpm.c
> > > ===================================================================
> > > RCS file: /cvs/src/sys/dev/acpi/tpm.c,v
> > > retrieving revision 1.3
> > > diff -u -p -r1.3 tpm.c
> > > --- tpm.c 1 Jul 2018 19:40:49 -0000 1.3
> > > +++ tpm.c 11 Mar 2019 10:57:38 -0000
> > > @@ -194,6 +194,7 @@ const char *tpm_hids[] = {
> > > "BCM0102",
> > > "NSC1200",
> > > "ICO0102",
> > > + "MSFT0101",
> > > NULL
> > > };
> > >
> > >
> > > --
> > > Best Regards
> > > Edd Barrett
> > >
> > > http://www.theunixzoo.co.uk
> > >
> >
>