[BUG] iwlwifi: Microcode SW error detected. Restarting 0x2000000.

[BUG] iwlwifi: Microcode SW error detected. Restarting 0x2000000.

[Mika Westerberg]

Hi,

On my Lenovo X1 Carbon gen 5 system,after I upgraded to v4.17-rc1
iwlwifi started to fail with following splats on dmesg:

  iwlwifi 0000:04:00.0: Microcode SW error detected.  Restarting 0x2000000.
  iwlwifi 0000:04:00.0: Start IWL Error Log Dump:
  iwlwifi 0000:04:00.0: Status: 0x00000100, count: 6
  iwlwifi 0000:04:00.0: Loaded firmware version: 31.560484.0
  iwlwifi 0000:04:00.0: 0x00000038 | BAD_COMMAND                 
  iwlwifi 0000:04:00.0: 0x00A002F1 | trm_hw_status0
  iwlwifi 0000:04:00.0: 0x00000000 | trm_hw_status1
  iwlwifi 0000:04:00.0: 0x0002495C | branchlink2
  iwlwifi 0000:04:00.0: 0x0003962E | interruptlink1
  iwlwifi 0000:04:00.0: 0x00000000 | interruptlink2
  iwlwifi 0000:04:00.0: 0x00000000 | data1
  iwlwifi 0000:04:00.0: 0x00000030 | data2
  iwlwifi 0000:04:00.0: 0x00000030 | data3
  iwlwifi 0000:04:00.0: 0x003B31AF | beacon time
  iwlwifi 0000:04:00.0: 0x0004CE4F | tsf low
  iwlwifi 0000:04:00.0: 0x00000000 | tsf hi
  iwlwifi 0000:04:00.0: 0x00000000 | time gp1
  iwlwifi 0000:04:00.0: 0x0004CE50 | time gp2
  iwlwifi 0000:04:00.0: 0x00000001 | uCode revision type
  iwlwifi 0000:04:00.0: 0x0000001F | uCode version major
  iwlwifi 0000:04:00.0: 0x00088D64 | uCode version minor
  iwlwifi 0000:04:00.0: 0x00000230 | hw version
  iwlwifi 0000:04:00.0: 0x00489000 | board version
  iwlwifi 0000:04:00.0: 0x00000000 | hcmd
  iwlwifi 0000:04:00.0: 0x00122080 | isr0
  iwlwifi 0000:04:00.0: 0x00000000 | isr1
  iwlwifi 0000:04:00.0: 0x28201802 | isr2
  iwlwifi 0000:04:00.0: 0x404001C0 | isr3
  iwlwifi 0000:04:00.0: 0x00000000 | isr4
  iwlwifi 0000:04:00.0: 0x802D0051 | last cmd Id
  iwlwifi 0000:04:00.0: 0x00000000 | wait_event
  iwlwifi 0000:04:00.0: 0x0000DF8E | l2p_control
  iwlwifi 0000:04:00.0: 0x00000820 | l2p_duration
  iwlwifi 0000:04:00.0: 0x00000000 | l2p_mhvalid
  iwlwifi 0000:04:00.0: 0x00000000 | l2p_addr_match
  iwlwifi 0000:04:00.0: 0x0000000D | lmpm_pmg_sel
  iwlwifi 0000:04:00.0: 0x13091828 | timestamp
  iwlwifi 0000:04:00.0: 0x00002838 | flow_handler
  iwlwifi 0000:04:00.0: Start IWL Error Log Dump:
  iwlwifi 0000:04:00.0: Status: 0x00000100, count: 7
  iwlwifi 0000:04:00.0: 0x00000034 | NMI_INTERRUPT_WDG
  iwlwifi 0000:04:00.0: 0x00000000 | umac branchlink1
  iwlwifi 0000:04:00.0: 0xC0086950 | umac branchlink2
  iwlwifi 0000:04:00.0: 0xC00842BC | umac interruptlink1
  iwlwifi 0000:04:00.0: 0xC00840CC | umac interruptlink2
  iwlwifi 0000:04:00.0: 0x00000400 | umac data1
  iwlwifi 0000:04:00.0: 0xC00840CC | umac data2
  iwlwifi 0000:04:00.0: 0xDEADBEEF | umac data3
  iwlwifi 0000:04:00.0: 0x0000001F | umac major
  iwlwifi 0000:04:00.0: 0x00088D64 | umac minor
  iwlwifi 0000:04:00.0: 0xC08861DC | frame pointer
  iwlwifi 0000:04:00.0: 0xC08861DC | stack pointer
  iwlwifi 0000:04:00.0: 0x00000000 | last host cmd
  iwlwifi 0000:04:00.0: 0x00004008 | isr status reg
  ieee80211 phy0: Hardware restart was requested
  iwlwifi 0000:04:00.0: FW Error notification: type 0x00000000 cmd_id 0x00
  iwlwifi 0000:04:00.0: FW Error notification: seq 0x0000 service 0x00000000
  iwlwifi 0000:04:00.0: FW Error notification: timestamp 0x           4CE4B
  iwlwifi 0000:04:00.0: Microcode SW error detected.  Restarting 0x2000000.

The controller in question is:

  04:00.0 Network controller [0280]: Intel Corporation Wireless 8265 / 8275 [8086:24fd] (rev 88)

Bisection pointed to the following commit and after I reverted it,
things seem to work fine again:
  
  66fa2424df16 ("iwlwifi: fw api: support the new scan request FW API version")

Any idea what might be wrong with that commit and how to fix it
properly?

Thanks!

Re: [BUG] iwlwifi: Microcode SW error detected. Restarting 0x2000000.

[Luciano Coelho]

On Mon, 2018-04-23 at 15:05 +0300, Mika Westerberg wrote:
> Hi,

Hi Mika,


> On my Lenovo X1 Carbon gen 5 system,after I upgraded to v4.17-rc1
> iwlwifi started to fail with following splats on dmesg:
> 
>   iwlwifi 0000:04:00.0: Microcode SW error detected.  Restarting
> 0x2000000.
>   iwlwifi 0000:04:00.0: Start IWL Error Log Dump:
>   iwlwifi 0000:04:00.0: Status: 0x00000100, count: 6
>   iwlwifi 0000:04:00.0: Loaded firmware version: 31.560484.0
>   iwlwifi 0000:04:00.0: 0x00000038 | BAD_COMMAND                 

[...]

> The controller in question is:
> 
>   04:00.0 Network controller [0280]: Intel Corporation Wireless 8265
> / 8275 [8086:24fd] (rev 88)
> 
> Bisection pointed to the following commit and after I reverted it,
> things seem to work fine again:
>   
>   66fa2424df16 ("iwlwifi: fw api: support the new scan request FW API
> version")
> 
> Any idea what might be wrong with that commit and how to fix it
> properly?

We had a bug in the scan command size for older firmwares.  I have
already made a fix for it, but I hadn't sent it out for upstream yet. 
I'll send it out now and soon send you a link to it in patchwork.

You have two options: recompile the kernel with my patch or; use the
latest version of the firmware from linux-firmware.git[1].

I recommend upgrading the firmware, because we have a lot of other
fixes and improvements there, but if you want to keep your distro's
original firmware package, you can apply my patch instead.

HTH and thanks for reporting!


[1] https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/plain/iwlwifi-8265-36.ucode


--
Cheers,
Luca.

Re: [BUG] iwlwifi: Microcode SW error detected. Restarting 0x2000000.

[Mika Westerberg]

On Mon, Apr 23, 2018 at 03:39:45PM +0300, Luciano Coelho wrote:
> We had a bug in the scan command size for older firmwares.  I have
> already made a fix for it, but I hadn't sent it out for upstream yet. 
> I'll send it out now and soon send you a link to it in patchwork.
> 
> You have two options: recompile the kernel with my patch or; use the
> latest version of the firmware from linux-firmware.git[1].
> 
> I recommend upgrading the firmware, because we have a lot of other
> fixes and improvements there, but if you want to keep your distro's
> original firmware package, you can apply my patch instead.
> 
> HTH and thanks for reporting!

I upgraded the firmware to iwlwifi-8265-36.ucode and the issue is gone
now. Thanks Luca!

Re: [BUG] iwlwifi: Microcode SW error detected. Restarting 0x2000000.

[Luca Coelho]

On Mon, 2018-04-23 at 15:39 +0300, Luciano Coelho wrote:
> On Mon, 2018-04-23 at 15:05 +0300, Mika Westerberg wrote:
> > Hi,
> 
> Hi Mika,
> 
> 
> > On my Lenovo X1 Carbon gen 5 system,after I upgraded to v4.17-rc1
> > iwlwifi started to fail with following splats on dmesg:
> > 
> >   iwlwifi 0000:04:00.0: Microcode SW error detected.  Restarting
> > 0x2000000.
> >   iwlwifi 0000:04:00.0: Start IWL Error Log Dump:
> >   iwlwifi 0000:04:00.0: Status: 0x00000100, count: 6
> >   iwlwifi 0000:04:00.0: Loaded firmware version: 31.560484.0
> >   iwlwifi 0000:04:00.0: 0x00000038 | BAD_COMMAND                 
> 
> [...]
> 
> > The controller in question is:
> > 
> >   04:00.0 Network controller [0280]: Intel Corporation Wireless
> > 8265
> > / 8275 [8086:24fd] (rev 88)
> > 
> > Bisection pointed to the following commit and after I reverted it,
> > things seem to work fine again:
> >   
> >   66fa2424df16 ("iwlwifi: fw api: support the new scan request FW
> > API
> > version")
> > 
> > Any idea what might be wrong with that commit and how to fix it
> > properly?
> 
> We had a bug in the scan command size for older firmwares.  I have
> already made a fix for it, but I hadn't sent it out for upstream
> yet. 
> I'll send it out now and soon send you a link to it in patchwork.

Here's the patch that solves the problem with older FWs:

https://patchwork.kernel.org/patch/10356987/

--
Cheers,
Luca.