QCA6174 hw2.1 on Clevo P750ZM

[Michal Kazior]

On 23 July 2015 at 10:47, Alexandre Maloteaux <a.maloteaux at gmail.com> wrote:
> Hi Michal
>
>> There's an unofficial image that's lying in a pull request on
>> github[1] you could use. It probably needs skip_otp parameter to be
>> used.
>
> I think this is the sumdog firmwares; i tested them and i made all the
> test with the skip_otp parameter
>
>> This looks interesting. Can you please:
>>
>>  rmmod ath10k_pci
>>  modprobe ath10k_pci irq_mode=1
>>
>> Msi range handling could be broken on this hw. If so we'll need to blacklist it.
>
> Still the same issue :
>
> [  213.889825] ath10k_pci 0000:06:00.0: limiting irq mode to: 1
> [  213.889833] ath10k_pci 0000:06:00.0: pci irq legacy interrupts 0
> irq_mode 1 reset_mode 0
> [  216.907165] ath10k_pci 0000:06:00.0: failed to receive initialized
> event from target: b8600448

Whoa. This value cannot be possibly correct. The register shouldn't
ever take such values.

What is the head commit id (kvalo/master) that you've used? You might
be missing a patches[1][2] that fixed a recent breakage in the dev
tree.

Can you make sure you test df0cf3c1249bc0c3a5a2077d6901eb90c19b7fda, please?

When you do, and it still crashes, can you get traces[3], please?


> [  216.907166] ath10k_pci 0000:06:00.0: failed to wait for target after
> cold reset: -110
> [  216.907167] ath10k_pci 0000:06:00.0: failed to reset chip: -110
> [  216.907220] ath10k_pci: probe of 0000:06:00.0 failed with error -110
>
>
>> Don't just go about renaming firmware filenames like that, please.
>> It's used to prevent older ath10k using newer firmware images which it
>> wouldn't handle anyway.
> upon those instructions :
> https://github.com/kvalo/ath10k-firmware/pull/2
> copying firmware-4.bin to firmware-5.bin seems to help for some users

It matters not. Driver tries to fallback until it finds any
firmware-X. If it doesn't find firmware-5.bin it'll try to load
firmware-4.bin, etc. This will be accompanied by a warning in kernel
log for each fallback attempt but it's harmless. Driver doesn't
distinguish firmware functionality by filename numbering.


[1]: https://github.com/kvalo/ath/commit/a052158aa981ca470673f49c636b289ee16894ea
[2]: https://github.com/kvalo/ath/commit/3c7e256a6de378e01098147527082abae05b146e
[3]: https://wireless.wiki.kernel.org/en/users/drivers/ath10k/debug#tracing


Michał

>
> Best Regards
>
>
> On 07/23/2015 09:21 AM, Michal Kazior wrote:
>> On 23 July 2015 at 10:05, Alexandre Maloteaux <a.maloteaux at gmail.com> wrote:
>>> Hi
>>>
>>> I just received a Clevo P750ZM (eurocom P5 pro) with a QCA6174
>>>
>>> 06:00.0 Network controller: Qualcomm Atheros QCA6174 802.11ac Wireless
>>> Network Adapter (rev 20)
>>>
>>> Hi have tested the dissect/assemble procedure  described here :
>>>
>>> https://www.mail-archive.com/ath10k@lists.infradead.org/msg02283.html
>>> https://askubuntu.com/questions/546813/killer-n1525-with-ubuntu-14-10
>> There's an unofficial image that's lying in a pull request on
>> github[1] you could use. It probably needs skip_otp parameter to be
>> used.
>>
>>
>>> Im on Arch and i have tested with the Eurocom Drivers
>>> (http://downloads.eurocom.com/support/drivers/zip/260/260_KillerWLAN_W864.zip)
>>> and the sumdog blobs  (https://github.com/sumdog/ath10k-firmware.git)
>>> directly on those 3 firmwares
>>> Arch official : 4.1.2-2-ARCH
>>> kvalo master : 4.2.0-rc3
>>> kvalo qca branch : 4.1.0-rc6
>>>
>>> I have a different issue for each kernel and i have tested with all
>>> board.bin available but none worked :
>>>
>>> ARCH official issue :
>> [snip]
>>
>> This won't work. 4.1 doesn't contain the necessary patch[2].
>>
>>
>>> kvalo master :
>>>
>>> [    0.000000] Linux version 4.2.0-rc3-wl-ath-KVALO-ATH-MASTER+
>>> (root at z1kadev) (gcc version 5.1.0 (GCC) ) #2 SMP PREEMPT Tue Jul 21
>>> 15:07:02 WAT 2015
>>> [   12.381604] systemd[1]: cups.path: Cannot add dependency job,
>>> ignoring: Unit cups.path failed to load: No such file or directory.
>>> [   12.595610] ath3k: probe of 3-7:1.0 failed with error -22
>>> [   12.595637] usbcore: registered new interface driver ath3k
>>> [   12.627163] ath10k_pci 0000:06:00.0: enabling device (0000 -> 0002)
>>> [   12.627915] ath10k_pci 0000:06:00.0: pci irq msi-x interrupts 8
>>> irq_mode 0 reset_mode 0
>>> [   12.822155] ath10k_pci 0000:06:00.0: Direct firmware load for
>>> ath10k/cal-pci-0000:06:00.0.bin failed with error -2
>>> [   12.822158] ath10k_pci 0000:06:00.0: Falling back to user helper
>>> [   72.915990] ath10k_pci 0000:06:00.0: Direct firmware load for
>>> ath10k/QCA6174/hw2.1/board-pci-168c:003e:1a56:1525.bin failed with error -2
>>> [   72.915993] ath10k_pci 0000:06:00.0: Falling back to user helper
>>> [  133.010331] ath10k_pci 0000:06:00.0: failed to load spec board file,
>>> falling back to generic: -11
>>> [  133.010647] ath10k_pci 0000:06:00.0: Direct firmware load for
>>> ath10k/QCA6174/hw2.1/firmware-5.bin failed with error -2
>>> [  133.010649] ath10k_pci 0000:06:00.0: Falling back to user helper
>>> [  193.105501] ath10k_pci 0000:06:00.0: could not fetch firmware file
>>> 'ath10k/QCA6174/hw2.1/firmware-5.bin': -11
>>> [  194.276201] ath10k_pci 0000:06:00.0: received unsolicited fw crash
>>> interrupt
>>> [  194.276209] ath10k_pci 0000:06:00.0: received unsolicited fw crash
>>> interrupt
>>> [  195.275581] ath10k_pci 0000:06:00.0: failed to receive control
>>> response completion, polling..
>>> [  195.275699] ath10k_pci 0000:06:00.0: received unsolicited fw crash
>>> interrupt
>> This looks interesting. Can you please:
>>
>>  rmmod ath10k_pci
>>  modprobe ath10k_pci irq_mode=1
>>
>> Msi range handling could be broken on this hw. If so we'll need to blacklist it.
>>
>>
>>> [  196.277176] ath10k_pci 0000:06:00.0: Service connect timeout
>>> [  196.277179] ath10k_pci 0000:06:00.0: failed to connect htt (-110)
>>> [  196.347791] ath10k_pci 0000:06:00.0: could not init core (-110)
>>> [  196.347819] ath10k_pci 0000:06:00.0: could not probe fw (-110)
>>>
>>> kvalo qca branch :
>>>
>>> [  386.639108] ath10k_pci 0000:06:00.0: failed to receive initialized
>>> event from target: b8600448
>>> [  386.639110] ath10k_pci 0000:06:00.0: failed to wait for target after
>>> cold reset: -110
>>> [  386.639110] ath10k_pci 0000:06:00.0: failed to reset chip: -110
>>> [  386.639203] ath10k_pci: probe of 0000:06:00.0 failed with error -110
>>> [  419.547152] ath10k_pci 0000:06:00.0: pci irq msi-x interrupts 8
>>> irq_mode 0 reset_mode 0
>>> [  422.565998] ath10k_pci 0000:06:00.0: failed to receive initialized
>>> event from target: b8600448
>>> [  422.566000] ath10k_pci 0000:06:00.0: failed to wait for target after
>>> cold reset: -110
>>> [  422.566001] ath10k_pci 0000:06:00.0: failed to reset chip: -110
>>> [  422.566096] ath10k_pci: probe of 0000:06:00.0 failed with error -110
>>>
>>> im getting only the cal issue  on the kvalo master if i copy
>>> firmware-4.bin to firmware-5.bin.
>> Don't just go about renaming firmware filenames like that, please.
>> It's used to prevent older ath10k using newer firmware images which it
>> wouldn't handle anyway.
>>
>>
>>> If i copy fw-1.bin from dissect.py to
>>> cal-pci-0000:06:00.0.bin . I m getting the issue with the official arch...
>> Uhm, that's wrong. Don't use cal-pci-xxx unless you know what you're doing.
>>
>> fw-1 and fw-2 from dissect are otp and main program binaries.
>> cal-pci-xxx is just calibration data blob. This is something entirely
>> else.
>>
>> If you see OTP failure you need to either set up adequate fw_feature
>> flag in the firmware-X.bin or pass skip_otp=y parameter while probing
>> ath10k_core module.
>>
>>
>> [1]: https://github.com/kvalo/ath10k-firmware/pull/2
>> [2]: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/net/wireless/ath/ath10k/pci.c?id=11a002efbaa7fbd9f6e616695ab42aa9f1caf060
>>
>>
>> Michał
>>
>