Bug 191751 - ACPI Exception: AE_BAD_PARAMETER occurs on suspend - Cherry Trail
Summary: ACPI Exception: AE_BAD_PARAMETER occurs on suspend - Cherry Trail
Status: NEEDINFO
Alias: None
Product: Platform Specific/Hardware
Classification: Unclassified
Component: x86-64 (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: platform_x86_64@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-01-02 18:39 UTC by Jonas Aaberg
Modified: 2017-11-21 19:57 UTC (History)
2 users (show)

See Also:
Kernel Version: 4.10rc2
Subsystem:
Regression: No
Bisected commit-id:


Attachments
DSDT (45.72 KB, application/x-bzip2)
2017-01-02 18:39 UTC, Jonas Aaberg
Details
kernel config (40.77 KB, application/x-bzip2)
2017-01-02 18:39 UTC, Jonas Aaberg
Details
dmesg of suspend/resume (26.21 KB, application/x-bzip2)
2017-01-02 18:40 UTC, Jonas Aaberg
Details
acpidump (140.43 KB, application/x-bzip2)
2017-01-03 05:26 UTC, Jonas Aaberg
Details

Description Jonas Aaberg 2017-01-02 18:39:10 UTC
Created attachment 249671 [details]
DSDT

When doing:
# echo freeze > /sys/power/state

on my Asus T100HAN - Chery trail based computer, I get the following error upon suspend:
--
ACPI Exception: AE_BAD_PARAMETER, Returned by Handler for [UserDefinedRegion] (20160930/evregion-300)
[ 1016.840647] ACPI Error: Method parse/execution failed [\_SB.P33X._STA] (Node ffff88017b0c1c58), AE_BAD_PARAMETER (20160930/psparse-543)
--
Comment 1 Jonas Aaberg 2017-01-02 18:39:36 UTC
Created attachment 249681 [details]
kernel config
Comment 2 Jonas Aaberg 2017-01-02 18:40:49 UTC
Created attachment 249691 [details]
dmesg of suspend/resume
Comment 3 Lv Zheng 2017-01-03 02:26:48 UTC
You need to upload acpidump output for filing an acpi bug.
Comment 4 Jonas Aaberg 2017-01-03 05:26:24 UTC
Created attachment 249811 [details]
acpidump

sorry for missing it, got interrupted while filing the bug.
Comment 5 Lv Zheng 2017-01-03 05:41:32 UTC
            Method (_STA, 0, NotSerialized)  // _STA: Status
            {
                If (^^PCI0.I2C7.PMIC.AVBG == One)
                {
                    ^^PCI0.I2C7.PMIC.G33X = Zero
                    If (^^PCI0.I2C7.PMIC.AVBL == One)
                    {
                        Return (^^PCI0.I2C7.PMIC.V33X) /* \_SB_.PCI0.I2C7.PMIC.V33X */
                    }
                }

                Return (Zero)
            }
Among AVBG, G33X, AVBL, V33X, V33X is defined as such an operation region field:
                OperationRegion (PMOP, 0x8D, Zero, 0x0100)
                Field (PMOP, DWordAcc, NoLock, Preserve)
                {
                    ...
                    V33X,   32
                }
0x8D is not a standard operation region ID.
A platform specific operation region driver will be required for it.
It's beyond our support scope, let me re-assign to platform specific category.

Thanks
Lv
Comment 6 Lv Zheng 2017-01-03 05:50:58 UTC
Quite a lot such kind of opregion:
The followings are covered by drivers/acpi/pmic/intel-pmic.c:
	Line 17146:                 OperationRegion (PMOP, 0x8D, Zero, 0x0100)
	Line 17171:                 OperationRegion (DPTF, 0x8C, Zero, 0x0100)
	Line 17617:                 OperationRegion (PMOP, 0x8D, Zero, 0x0100)
	Line 17635:                 OperationRegion (DPTF, 0x8C, Zero, 0x0100)
	Line 18094:                 OperationRegion (DPTF, 0x8C, Zero, 0x0100)
	Line 18123:                 OperationRegion (PMOP, 0x8D, Zero, 0x0100)
	Line 18656:                 OperationRegion (REGS, 0x8F, Zero, 0x50)
	Line 19056:                 OperationRegion (DPTF, 0x8C, Zero, 0x0100)
	Line 19085:                 OperationRegion (PMOP, 0x8D, Zero, 0x0100)
So you need to configure CONFIG_PMIC_OPREGON and related hardware drivers for them.

Couldn't find a driver in the current kernel:
	Line 20613:             OperationRegion (REGS, 0x87, Zero, 0x30)
	Line 16378:             OperationRegion (GMMR, 0x91, Zero, 0x8000)
	Line 16543:             OperationRegion (GMMR, 0x92, Zero, 0x8000)
	Line 16657:             OperationRegion (GMMR, 0x94, Zero, 0x8000)
Looks like feature gaps.

Thanks and best regards
Lv
Comment 7 Jonas Aaberg 2017-01-03 07:37:21 UTC
Thanks Lv. Is there a connection between the missing support for PMIC Dollar Cove TI (INT33F5) (Device PMI2 in attached the DSDT) and your findings? Or is it something else?
Comment 8 Lv Zheng 2017-01-03 08:20:41 UTC
I don't have knowledge related to that.
It's quite platform specific.

I think you should ask:
Bin Gao <bin.gao@intel.com>
It seems he is from an Intel team responsible for upstream platform specific PMIC opregion support.
Comment 9 mirh 2017-11-21 19:57:59 UTC
I was to CC Bin, then I noticed good Hans should have fixed this between c06632ea054c49510efacb42c52aab693c45b7ba and 72ebe5a0b3505019d4e7592c0dd9281e0b9e44c7

Note You need to log in before you can comment on or make changes to this bug.