Re: port-amd64/53286: hdaudio(4), iwm(4) and rtsx(4) detections fail in UEFI case

[Frank Kardel <kardel%netbsd.org@localhost>]

I also have such a UEFI system.

The main difference is that UEFI boot does not enable all devices and 
our options to enable devices via ACPI do not improve anything.

The symptom is (one of many examples):

 07:00.0 Ethernet controller: Intel Corporation 82572EI Gigabit 
Ethernet Controller (Copper) (rev 06)
        Subsystem: Intel Corporation PRO/1000 PT Server Adapter
-       Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- 
ParErr- Stepping- SERR- FastB2B- DisINTx-
+       Control: I/O- Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- 
ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- 
<TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 64 bytes
        Interrupt: pin A routed to IRQ 5
-       Region 0: Memory at fd440000 (32-bit, non-prefetchable)
-       Region 1: Memory at fd420000 (32-bit, non-prefetchable)
-       Region 2: I/O ports at c000
+       Region 0: Memory at fd440000 (32-bit, non-prefetchable) [disabled]
+       Region 1: Memory at fd420000 (32-bit, non-prefetchable) [disabled]
+       Region 2: I/O ports at c000 [disabled]
        Expansion ROM at fd400000 [disabled]
        Capabilities: [c8] Power Management version 2
                Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA 
PME(D0+,D1-,D2-,D3hot+,D3cold+)

- = CSM boot

+ = UEFI boot

As I/O and MEM is disabled the drivers fail tho initialize. On my system 
not any combination of the options

options         PCI_INTR_FIXUP          # PCI interrupt routing via ACPI
options         PCI_BUS_FIXUP           # fixup PCI bus numbering
options         PCI_ADDR_FIXUP          # fixup PCI I/O addresses

options         ACPI_ACTIVATE_DEV       # If set, activate inactive devices

did give an improvement. I didn't see an BIOS options/knobs to force 
UEFI boot to enable all devices like they are in the CSM case.

The code for ACPI_ACTIVATE_DEV seems incomplete going by the comments.

pciconf.c seems to contain code for enabling devices, but, if I remember 
right, this is not called in amd64/UEFI boot.

Frank



On 05/13/18 21:50, ryo%tetera.org@localhost wrote:
> > Number:         53286
> > Category:       port-amd64
> > Synopsis:       hdaudio(4), iwm(4) and rtsx(4) detections fail in UEFI case
> > Confidential:   no
> > Severity:       serious
> > Priority:       medium
> > Responsible:    port-amd64-maintainer
> > State:          open
> > Class:          sw-bug
> > Submitter-Id:   net
> > Arrival-Date:   Sun May 13 19:50:00 +0000 2018
> > Originator:     Ryo ONODERA
> > Release:        NetBSD 8.99.17
> > Organization:
> 	
> > Environment:
> 	
> 	
> System: NetBSD brownie 8.99.17 NetBSD 8.99.17 (DTRACE7) #20: Mon May 14 04:13:46 UTC 2018 ryoon@brownie:/usr/world/8.99/amd64/obj/sys/arch/amd64/compile/DTRACE7 amd64
> Architecture: x86_64
> Machine: amd64
> > Description:
> In UEFI case, some devices fail as follows on HP Spectre x360 ae019TU laptop.
> In legacy MBR boot case, they are enabled successfully.
>
> hdaudio0 at pci0 dev 31 function 3: HD Audio Controller
> hdaudio0: interrupting at msi4 vec 0
> hdaudio0: timeout stopping RIRB
> hdaudio0: couldn't reset because RIRB is busy
> hdaudio0: device driver failed to attach
>
> iwm0 at pci2 dev 0 function 0: vendor 8086 product 24fd (rev. 0x78)
> iwm0: interrupting at msi2 vec 0
> iwm0: failed to lock the nic
>
> rtsx0 at pci1 dev 0 function 0: vendor 10ec product 525a (rev. 0x01)
> rtsx0: interrupting at msi1 vec 0
> rtsx0: couldn't initialize chip
>
> For rtsx(4), the following in src/sys/dev/ic/rtsx.c fails.
>
>          /* Enable interrupt write-clear (default is read-clear). */
>          RTSX_CLR(sc, RTSX_NFTS_TX_CTRL, RTSX_INT_READ_CLR);
>
> > How-To-Repeat:
> Boot NetBSD/amd64 8.99.17 on HP Spectre x360 ae019TU laptop in UEFI boot mode.
>
> > Fix:
> I have no idea.
>
> > Unformatted:
>   	
>