Bug 199117

Summary: power stays on after poweroff command - N5550 NAS box (Intel Atom D2550)
Product: ACPI Reporter: Dominik Mierzejewski (dominik)
Component: BIOSAssignee: Zhang Rui (rui.zhang)
Status: CLOSED UNREPRODUCIBLE    
Severity: normal CC: lenb, rui.zhang, yu.c.chen
Priority: P1    
Hardware: Intel   
OS: Linux   
Kernel Version: kernel-4.20.15-200.fc29.x86_64 Subsystem:
Regression: No Bisected commit-id:
Attachments: full dmesg from next boot
acpidump output
dmesg from suspend and resume session

Description Dominik Mierzejewski 2018-03-14 14:29:17 UTC
On the Thecus N5550 NAS box (Intel Atom D2550-based) running Fedora 27, power stays on after issuing the poweroff command. The following messages remain visible on the console:

Powering off.
[ 4199.165227] ACPI Error:
[ 4199.165229] ACPI Error: [\_SB_.PCI0,LPCB.TPM_.PTS_] Namespace lookup failure, AE_NOT_FOUND (20170831/psargs-364)
[ 4199.165243] ACPI Error: Method parse/execution failed \_PTS, AE_NOT_FOUND (20170831/psparse-550)
[ 4199.165868] reboot: Power down

reboot command works.
Comment 1 Dominik Mierzejewski 2018-03-14 14:49:16 UTC
Created attachment 274727 [details]
full dmesg from next boot
Comment 2 Zhang Rui 2018-05-08 02:25:32 UTC
please attach the acpidump output for this platform.
Comment 3 Dominik Mierzejewski 2018-05-10 21:23:31 UTC
Created attachment 275911 [details]
acpidump output
Comment 4 Zhang Rui 2018-05-30 05:48:05 UTC
    Method (_PTS, 1, NotSerialized)  // _PTS: Prepare To Sleep
    {
        P80D = Zero
        P8XH (Zero, Arg0)
        G1S3 = Ones
        G1S2 = Ones
        G1S = One
        G1E = One
        G0S = Ones
        If ((Arg0 == 0x03))
        {
            If ((DTSE && MPEN))
            {
                TRAP (TRTD, 0x1E)
            }
        }

        If (CondRefOf (TCGM))
        {
            \_SB.PCI0.LPCB.TPM.PTS (Arg0)
        }
    }
_PTS method invokes  \_SB.PCI0.LPCB.TPM.PTS, while  \_SB.PCI0.LPCB.TPM.PTS is not defined anywhere.
This looks like a BIOS bug to me. can you please check if there is any TPM related options in BIOS and see if changing that BIOS setting helps or not?
Comment 5 Dominik Mierzejewski 2018-06-02 22:18:33 UTC
Thanks for the analysis. Unfortunately, there are no TPM-related options. The only thing I can see is "TPM not detected" entry at the bottom of "Security" screen.

Retrying with kernel-4.16.12-300.fc28.x86_64 (Fedora 28) I get the same error. Is there anything else I can do?
Comment 6 Zhang Rui 2018-06-25 08:18:11 UTC
does S3/S4 works on your platform?
if yes, can you please attach the dmesg output after resume from S3/S4?
Comment 7 Dominik Mierzejewski 2018-06-26 16:51:35 UTC
(In reply to Zhang Rui from comment #6)
> does S3/S4 works on your platform?
> if yes, can you please attach the dmesg output after resume from S3/S4?

The system appears to enter suspend state (fans stop) after running "systemctl suspend", but after I wake it up using the power button, the screen remains blank. That seems to be a separate issue, possibly related to the broken gma500_gfx driver not being loaded because I blacklisted due to https://bugzilla.kernel.org/show_bug.cgi?id=199115 . The system is nevertheless otherwise functional (e.g. Ctrl+Alt+Del from the console reboots it gracefully) and I'm attaching the dmesg from the session.
Comment 8 Dominik Mierzejewski 2018-06-26 16:52:27 UTC
Created attachment 276881 [details]
dmesg from suspend and resume session
Comment 9 Dominik Mierzejewski 2018-11-21 10:27:23 UTC
Still occurring with Fedora 29 kernel-4.19.2-300.fc29.x86_64.
Comment 10 Dominik Mierzejewski 2018-12-26 13:09:41 UTC
Still reproducible with Fedora 29 kernel-4.19.10-300.fc29.x86_64.
Comment 11 Dominik Mierzejewski 2019-01-13 23:33:20 UTC
With kernel-4.19.13-300.fc29.x86_64 I get the following messages:

[ 520.872566] ACPI: Preparing to enter system sleep state S5
[ 520.877255] ACPI Error:
[ 520.877259] No handler or method for GPE 0500-0x000000000000052F conflicts with OpRegion 0x0000000000000500-0x000000000000053F (\GB
[ 520.877262] No handler or method for GPE 00, disabling event (20180810/evgpe-835)
[ 520.881642] Could not resolve [\_SB.PCI0.LPCB.TPM.PTS], AE_NOT_FOUND (20180810/psargs-330)
[ 520.895000] ACPI Error: Method parse/execution failed \_PTS, AE_NOT_FOUND (20180810/psparse-516)
[ 520.899532] reboot: Power down
Comment 12 Dominik Mierzejewski 2019-03-19 10:02:57 UTC
4.20.15-200.fc29.x86_64:
[  87.817882] ACPI: Preparing to enter system sleep state S5
[  87.822587] ACPI Error: No handler or method for GPE 00, disabling event
[  87.822591] ACPI BIOS Error (bug):
[  87.822596]  (20181003/evgpe-835)
[  87.827644] Could not resolve [\_SB.PCI0.LPCB.TPM.PTS], AE_NOT_FOUND (20180810/psargs-330)
[  87.841262] ACPI Error: Method parse/execution failed \_PTS, AE_NOT_FOUND (20180810/psparse-516)
[  87.845857] reboot: Power down

Is there anything I can do to make this board power down correctly?
Comment 13 Dominik Mierzejewski 2019-04-09 14:04:53 UTC
5.0.5-200.fc29.x86_64:

ACPI: Preparing to enter system sleep state S5
ACPI Error:
ACPI Error:
Could not resolve [\_SB.PCI0.LPCB.TPM.PTS], AE_NOT_FOUND
No handler or method for GPE 00, disabling event (20181213/evgpe-835)
 (20180810/psargs-330)
ACPI Error: Method parse/execution failed \_PTS, AE_NOT_FOUND (20181213/psparse-531)
reboot: Power down
Comment 14 Dominik Mierzejewski 2019-10-22 13:46:55 UTC
So, for the record, I was told[0] to flash a different UEFI firmware[1] (seems to be the same as [2]) by tech support and with that firmware the bug is gone. They won't recommend it, because it's for the "Windows NAS" model (W5000+)[3], but it works for me and the hardware specs seem to be the same as the N5550[4].

[0] https://thecus.kayako.com/Tickets/Ticket/View/38798
[1] https://www.dropbox.com/s/85c4yeqyefj2yr3/WCDV_05.zip?dl=1
[2] http://www.thecus.com/Downloads/BIOS/WCDV_05.zip
[3] http://wss.thecus.com/product_W5000.php
[4] http://www.thecus.com/product.php?PROD_ID=76
Comment 15 Zhang Rui 2019-12-31 05:40:28 UTC
Good to know.
so this indeed seems like a BIOS problem and you're luck to have this fixed by BIOS upgrade. :)
Bug closed.