[Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Fedor Pchelkin]

Hi,

there are IO_PAGE_FAULT errors occassionally thrown into the log with
rtw89_8852be-compatible device in use:

[    7.135509] rtw89_8852be 0000:03:00.0: loaded firmware rtw89/rtw8852b_fw-1.bin
[    7.135610] rtw89_8852be 0000:03:00.0: enabling device (0000 -> 0003)
[    7.137074] rtw89_8852be 0000:03:00.0: Firmware version 0.29.29.8 (39dbf50f), cmd version 0, type 5
[    7.137079] rtw89_8852be 0000:03:00.0: Firmware version 0.29.29.8 (39dbf50f), cmd version 0, type 3
[    7.423852] rtw89_8852be 0000:03:00.0: chip rfe_type is 1
[    7.452843] rtw89_8852be 0000:03:00.0: rfkill hardware state changed to enable
[    7.478466] rtw89_8852be 0000:03:00.0 wlo1: renamed from wlan0
...
[ 1361.803384] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0 flags=0x0000]
[ 1434.918012] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0 flags=0x0000]
[ 1551.553344] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0 flags=0x0000]
[ 1649.346804] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0 flags=0x0000]


address and flags are always all zeros. It is reproducible on the current
mainline v6.15-rc2 kernel and v0.29.29.8 fw version.


It is most probably related to the beacon filter feature enabled in the
firmware. Bisection in the kernel leads to

commit d56c261e5214e51e2c6d22149f63555039b5601e (HEAD)
Author: Po-Hao Huang <phhuang@realtek.com>
Date:   Thu Nov 28 13:54:29 2024 +0800

    wifi: rtw89: 8852b: add beacon filter and CQM support


and in linux-firmware to

commit 20cace1adf6a33cac73595ea3202eb784dea98a6
Author: Po-Hao Huang <phhuang@realtek.com>
Date:   Thu Sep 19 17:02:28 2024 +0800

    rtw89: 8852b: update fw to v0.29.29.7
    
    Enable beacon filter feature.


Hardware info:

$ lspci -nn -s 03:00.0
03:00.0 Network controller [0280]: Realtek Semiconductor Co., Ltd. RTL8852BE PCIe 802.11ax Wireless Network Controller [10ec:b852]

$ lspci -vvvk
03:00.0 Network controller: Realtek Semiconductor Co., Ltd. RTL8852BE PCIe 802.11ax Wireless Network Controller
        DeviceName: Realtek
        Subsystem: Lenovo Device 4853
        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 93
        IOMMU group: 15
        Region 0: I/O ports at 4000 [size=256]
        Region 2: Memory at 98900000 (64-bit, non-prefetchable) [size=1M]
        Capabilities: <access denied>
        Kernel driver in use: rtw89_8852be
        Kernel modules: rtw89_8852be

$ lspci -x
03:00.0 Network controller: Realtek Semiconductor Co., Ltd. RTL8852BE PCIe 802.11ax Wireless Network Controller
00: ec 10 52 b8 07 04 10 10 00 00 80 02 10 00 00 00
10: 01 40 00 00 00 00 00 00 04 00 90 98 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 aa 17 53 48
30: 00 00 00 00 40 00 00 00 00 00 00 00 ff 01 00 00


--
Thanks,
Fedor

RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Ping-Ke Shih]

Fedor Pchelkin <boddah8794@gmail.com> wrote:
> Hi,
> 
> there are IO_PAGE_FAULT errors occassionally thrown into the log with
> rtw89_8852be-compatible device in use:
> 
> [    7.135509] rtw89_8852be 0000:03:00.0: loaded firmware rtw89/rtw8852b_fw-1.bin
> [    7.135610] rtw89_8852be 0000:03:00.0: enabling device (0000 -> 0003)
> [    7.137074] rtw89_8852be 0000:03:00.0: Firmware version 0.29.29.8 (39dbf50f), cmd version 0, type 5
> [    7.137079] rtw89_8852be 0000:03:00.0: Firmware version 0.29.29.8 (39dbf50f), cmd version 0, type 3
> [    7.423852] rtw89_8852be 0000:03:00.0: chip rfe_type is 1
> [    7.452843] rtw89_8852be 0000:03:00.0: rfkill hardware state changed to enable
> [    7.478466] rtw89_8852be 0000:03:00.0 wlo1: renamed from wlan0
> ...
> [ 1361.803384] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0
> flags=0x0000]
> [ 1434.918012] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0
> flags=0x0000]
> [ 1551.553344] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0
> flags=0x0000]
> [ 1649.346804] rtw89_8852be 0000:03:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x0
> flags=0x0000]
> 
> 
> address and flags are always all zeros. It is reproducible on the current
> mainline v6.15-rc2 kernel and v0.29.29.8 fw version.
> 
> 
> It is most probably related to the beacon filter feature enabled in the
> firmware. Bisection in the kernel leads to
> 
> commit d56c261e5214e51e2c6d22149f63555039b5601e (HEAD)
> Author: Po-Hao Huang <phhuang@realtek.com>
> Date:   Thu Nov 28 13:54:29 2024 +0800
> 
>     wifi: rtw89: 8852b: add beacon filter and CQM support
> 
> 
> and in linux-firmware to
> 
> commit 20cace1adf6a33cac73595ea3202eb784dea98a6
> Author: Po-Hao Huang <phhuang@realtek.com>
> Date:   Thu Sep 19 17:02:28 2024 +0800
> 
>     rtw89: 8852b: update fw to v0.29.29.7
> 
>     Enable beacon filter feature.
> 
> 

Thanks for the report. 

As bisection, can I summarize as below table?

    d56c261e52           firmware             IO_PAGE_FAULT
   (CQM support)
  ---------------      -----------          -----------------
      o                 0.29.29.7                  yes
      o                 0.29.29.5                  no
      x                 0.29.29.7                  no
      x                 0.29.29.5                  no

If this table is correct, we will check the beacon filter feature in firmware. 

Re: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Fedor Pchelkin]

On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> As bisection, can I summarize as below table?
> 
>     d56c261e52           firmware             IO_PAGE_FAULT
>    (CQM support)
>   ---------------      -----------          -----------------
>       o                 0.29.29.7                  yes
>       o                 0.29.29.5                  no
>       x                 0.29.29.7                  no
>       x                 0.29.29.5                  no
> 
> If this table is correct, we will check the beacon filter feature in firmware. 
> 

Yes, it correctly describes the situation. My suspicion currently falls on
the beacon filter feature in firmware, too.

RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Ping-Ke Shih]

Fedor Pchelkin <boddah8794@gmail.com> wrote:
> 
> On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> > As bisection, can I summarize as below table?
> >
> >     d56c261e52           firmware             IO_PAGE_FAULT
> >    (CQM support)
> >   ---------------      -----------          -----------------
> >       o                 0.29.29.7                  yes
> >       o                 0.29.29.5                  no
> >       x                 0.29.29.7                  no
> >       x                 0.29.29.5                  no
> >
> > If this table is correct, we will check the beacon filter feature in firmware.
> >
> 
> Yes, it correctly describes the situation. My suspicion currently falls on
> the beacon filter feature in firmware, too.

Could you help to test when we have updated firmware?

Re: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Fedor Pchelkin]

On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
> Could you help to test when we have updated firmware?

Yes.

Re: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Fedor Pchelkin]

On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
> Fedor Pchelkin <boddah8794@gmail.com> wrote:
> > 
> > On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> > > As bisection, can I summarize as below table?
> > >
> > >     d56c261e52           firmware             IO_PAGE_FAULT
> > >    (CQM support)
> > >   ---------------      -----------          -----------------
> > >       o                 0.29.29.7                  yes
> > >       o                 0.29.29.5                  no
> > >       x                 0.29.29.7                  no
> > >       x                 0.29.29.5                  no
> > >
> > > If this table is correct, we will check the beacon filter feature in firmware.
> > >
> > 
> > Yes, it correctly describes the situation. My suspicion currently falls on
> > the beacon filter feature in firmware, too.
> 
> Could you help to test when we have updated firmware?
> 

Hi,

Were you able to reproduce or track down the problem?

RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Ping-Ke Shih]

Fedor Pchelkin <boddah8794@gmail.com> wrote:
> On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
> > Fedor Pchelkin <boddah8794@gmail.com> wrote:
> > >
> > > On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> > > > As bisection, can I summarize as below table?
> > > >
> > > >     d56c261e52           firmware             IO_PAGE_FAULT
> > > >    (CQM support)
> > > >   ---------------      -----------          -----------------
> > > >       o                 0.29.29.7                  yes
> > > >       o                 0.29.29.5                  no
> > > >       x                 0.29.29.7                  no
> > > >       x                 0.29.29.5                  no
> > > >
> > > > If this table is correct, we will check the beacon filter feature in firmware.
> > > >
> > >
> > > Yes, it correctly describes the situation. My suspicion currently falls on
> > > the beacon filter feature in firmware, too.
> >
> > Could you help to test when we have updated firmware?
> >
> 
> Hi,
> 
> Were you able to reproduce or track down the problem?

Yes, we have reproduced this problem, and we found sending firmware command
during deep PS mode will raise IO_PAGE_FAULT. With below setting to disable
deep PS mode, the symptom can disappear.

   rtw89_core.ko disable_ps_mode=y

Now we are arranging CAT-C equipment to capture signals in PCIE bus to see
what happens at that moment. 

One possible solution (workaround) is to avoid sending firmware commands
during deep PS mode, before we found out the root cause. 

RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Bernie Huang]

[-- Attachment #1: Type: text/plain, Size: 2195 bytes --]

-----Original Message-----
From: Ping-Ke Shih <pkshih@realtek.com> 
Sent: Thursday, May 8, 2025 8:59 AM
To: Fedor Pchelkin <boddah8794@gmail.com>
Cc: Bernie Huang <phhuang@realtek.com>; linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org
Subject: RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

Fedor Pchelkin <boddah8794@gmail.com> wrote:
> > On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
> > > Fedor Pchelkin <boddah8794@gmail.com> wrote:
> > > >
> > > > On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> > > > > As bisection, can I summarize as below table?
> > > > >
> > > > >     d56c261e52           firmware             IO_PAGE_FAULT
> > > > >    (CQM support)
> > > > >   ---------------      -----------          -----------------
> > > > >       o                 0.29.29.7                  yes
> > > > >       o                 0.29.29.5                  no
> > > > >       x                 0.29.29.7                  no
> > > > >       x                 0.29.29.5                  no
> > > > >
> > > > > If this table is correct, we will check the beacon filter feature in firmware.
> > > > >
> > > >
> > > > Yes, it correctly describes the situation. My suspicion currently 
> > > > falls on the beacon filter feature in firmware, too.
> > >
> > > Could you help to test when we have updated firmware?
> > >
> > 
> > Hi,
> > 
> > Were you able to reproduce or track down the problem?

> Yes, we have reproduced this problem, and we found sending firmware command during deep PS mode will raise IO_PAGE_FAULT. With below setting to disable deep PS > mode, the symptom can disappear.

>   rtw89_core.ko disable_ps_mode=y

> Now we are arranging CAT-C equipment to capture signals in PCIE bus to see what happens at that moment. 

> One possible solution (workaround) is to avoid sending firmware commands during deep PS mode, before we found out the root cause. 

Hi Fedor,

Could you please check whether the attached firmware fixes this issue?
The firmware version should be 1.29.29.9.
You could check it via command:
`cat /var/log/kern.log | grep "Firmware version"`

Best Regards,
Bernie



[-- Attachment #2: rtw8852b_fw-1.bin --]
[-- Type: application/octet-stream, Size: 1249344 bytes --]

Re: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Fedor Pchelkin]

On Wed, 14. May 03:29, Bernie Huang wrote:
> Hi Fedor,
> 
> Could you please check whether the attached firmware fixes this issue?
> The firmware version should be 1.29.29.9.
> You could check it via command:
> `cat /var/log/kern.log | grep "Firmware version"`

Thanks Bernie, Ping-Ke.

I may confirm there are no IO page faults nor any other regressions seen
on my side with the updated firmware in use.

If needed

Tested-by: Fedor Pchelkin <boddah8794@gmail.com> 

--
Fedor

Re: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Bitterblue Smith]

On 14/05/2025 06:29, Bernie Huang wrote:
> -----Original Message-----
> From: Ping-Ke Shih <pkshih@realtek.com> 
> Sent: Thursday, May 8, 2025 8:59 AM
> To: Fedor Pchelkin <boddah8794@gmail.com>
> Cc: Bernie Huang <phhuang@realtek.com>; linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org
> Subject: RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled
> 
> Fedor Pchelkin <boddah8794@gmail.com> wrote:
>>> On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
>>>> Fedor Pchelkin <boddah8794@gmail.com> wrote:
>>>>>
>>>>> On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
>>>>>> As bisection, can I summarize as below table?
>>>>>>
>>>>>>     d56c261e52           firmware             IO_PAGE_FAULT
>>>>>>    (CQM support)
>>>>>>   ---------------      -----------          -----------------
>>>>>>       o                 0.29.29.7                  yes
>>>>>>       o                 0.29.29.5                  no
>>>>>>       x                 0.29.29.7                  no
>>>>>>       x                 0.29.29.5                  no
>>>>>>
>>>>>> If this table is correct, we will check the beacon filter feature in firmware.
>>>>>>
>>>>>
>>>>> Yes, it correctly describes the situation. My suspicion currently 
>>>>> falls on the beacon filter feature in firmware, too.
>>>>
>>>> Could you help to test when we have updated firmware?
>>>>
>>>
>>> Hi,
>>>
>>> Were you able to reproduce or track down the problem?
> 
>> Yes, we have reproduced this problem, and we found sending firmware command during deep PS mode will raise IO_PAGE_FAULT. With below setting to disable deep PS > mode, the symptom can disappear.
> 
>>   rtw89_core.ko disable_ps_mode=y
> 
>> Now we are arranging CAT-C equipment to capture signals in PCIE bus to see what happens at that moment. 
> 
>> One possible solution (workaround) is to avoid sending firmware commands during deep PS mode, before we found out the root cause. 
> 
> Hi Fedor,
> 
> Could you please check whether the attached firmware fixes this issue?
> The firmware version should be 1.29.29.9.
By the way, what does the first number mean? The previous version was
0.29.29.8. I would expect only the last number to change for this bug fix.

RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled

[Ping-Ke Shih]

Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote:
> On 14/05/2025 06:29, Bernie Huang wrote:
> > -----Original Message-----
> > From: Ping-Ke Shih <pkshih@realtek.com>
> > Sent: Thursday, May 8, 2025 8:59 AM
> > To: Fedor Pchelkin <boddah8794@gmail.com>
> > Cc: Bernie Huang <phhuang@realtek.com>; linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org
> > Subject: RE: [Bug report] rtw89: 8852b: IO_PAGE_FAULTs with beacon filter feature enabled
> >
> > Fedor Pchelkin <boddah8794@gmail.com> wrote:
> >>> On Thu, 17. Apr 01:05, Ping-Ke Shih wrote:
> >>>> Fedor Pchelkin <boddah8794@gmail.com> wrote:
> >>>>>
> >>>>> On Wed, 16. Apr 00:49, Ping-Ke Shih wrote:
> >>>>>> As bisection, can I summarize as below table?
> >>>>>>
> >>>>>>     d56c261e52           firmware             IO_PAGE_FAULT
> >>>>>>    (CQM support)
> >>>>>>   ---------------      -----------          -----------------
> >>>>>>       o                 0.29.29.7                  yes
> >>>>>>       o                 0.29.29.5                  no
> >>>>>>       x                 0.29.29.7                  no
> >>>>>>       x                 0.29.29.5                  no
> >>>>>>
> >>>>>> If this table is correct, we will check the beacon filter feature in firmware.
> >>>>>>
> >>>>>
> >>>>> Yes, it correctly describes the situation. My suspicion currently
> >>>>> falls on the beacon filter feature in firmware, too.
> >>>>
> >>>> Could you help to test when we have updated firmware?
> >>>>
> >>>
> >>> Hi,
> >>>
> >>> Were you able to reproduce or track down the problem?
> >
> >> Yes, we have reproduced this problem, and we found sending firmware command during deep PS mode will
> raise IO_PAGE_FAULT. With below setting to disable deep PS > mode, the symptom can disappear.
> >
> >>   rtw89_core.ko disable_ps_mode=y
> >
> >> Now we are arranging CAT-C equipment to capture signals in PCIE bus to see what happens at that moment.
> >
> >> One possible solution (workaround) is to avoid sending firmware commands during deep PS mode, before
> we found out the root cause.
> >
> > Hi Fedor,
> >
> > Could you please check whether the attached firmware fixes this issue?
> > The firmware version should be 1.29.29.9.
> By the way, what does the first number mean? The previous version was
> 0.29.29.8. I would expect only the last number to change for this bug fix.

This firmware is for testing only. The formal firmware should be 0.29.29.x.