Fwd: Fwd: lspci not working

[Arend van Spriel <aspriel@gmail.com>]

On 17-05-15 18:56, Hauke Mehrtens wrote:
> On 05/17/2015 06:49 PM, Arend van Spriel wrote:
>>
>>
>> On 17-05-15 16:48, Schmirr Wurst wrote:
>>> ---------- Forwarded message ----------
>>> From: Schmirr Wurst <schmirrwurst@gmail.com>
>>> Date: 2015-05-17 16:47 GMT+02:00
>>> Subject: Re: Fwd: lspci not working
>>> To: Arend van Spriel <aspriel@gmail.com>
>>>
>>>
>>> 2015-05-17 16:34 GMT+02:00 Arend van Spriel <aspriel@gmail.com>:
>>>> On 17-05-15 16:08, Schmirr Wurst wrote:
>>>>>
>>>>> ---------- Forwarded message ----------
>>>>> From: Schmirr Wurst <schmirrwurst@gmail.com>
>>>>> Date: 2015-05-17 16:07 GMT+02:00
>>>>> Subject: Re: lspci not working
>>>>> To: Hauke Mehrtens <hauke@hauke-m.de>
>>>>>
>>>>>
>>>>> 2015-05-17 15:17 GMT+02:00 Hauke Mehrtens <hauke@hauke-m.de>:
>>>>>>
>>>>>> On 05/17/2015 03:00 PM, Schmirr Wurst wrote:
>>>>>>>
>>>>>>> 2015-05-17 14:57 GMT+02:00 Schmirr Wurst <schmirrwurst@gmail.com>:
>>>>>>>>
>>>>>>>> I'm not familiar with inline answers and mailing list, tried to put
>>>>>>>> some order in my answer + log file
>>>>>>>>
>>>>>>>> 2015-05-17 13:18 GMT+02:00 Rafa? Mi?ecki <zajec5@gmail.com>:
>>>>>>>>>
>>>>>>>>> On 17 May 2015 at 12:23, Arend van Spriel <arend@broadcom.com>
>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> On 05/17/15 02:21, Schmirr Wurst wrote:
>>>>>>>>>>>
>>>>>>>>>>> I tried as suggested to had a look at
>>>>>>>>>>> /sys/bus/sdio/devices, but the only devices around there a 3
>>>>>>>>>>> mmc1:0001:1 to :3, I guess it is something else...
>>>>>>>>>>>
>>>>>>>>>>> I already try to install brcmfmac_sdio , with some tutorial
>>>>>>>>>>> from the
>>>>>>>>>>> internet, but it didn't work...
>>>>>>>>>>> actually, I see under /sys/bus/sdio/drivers brcmfmac_sdio ...
>>>>>>>>>>>
>>>>>>>>>>> In that directory, I see a directory mmc1:0001:2
>>>>>>>>>>> under device I have 0xa94d
>>>>>>>>>>>
>>>>>>>>>>> I completly lost, maybe you understand that information, sorry..
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> I do. The fact that you see a subdirectory mmc1:0001:2 in
>>>>>>>>>> /sys/bus/sdio/drivers/brcmfmac_sdio means that the driver was
>>>>>>>>>> probed
>>>>>>>>>> for
>>>>>>>>>> this device. Now would be a good time to share your (friends')
>>>>>>>>>> kernel
>>>>>>>>>> log,
>>>>>>>>>> but my guess is you are either missing firmware or nvram data
>>>>>>>>>> or both
>>>>>>>>>> for
>>>>>>>>>> this device.
>>>>>>>>
>>>>>>>> I ve attatched the kernel.log here
>>>>>>>> https://drive.google.com/file/d/0B8gm4mLCCQAgMmNlVVFSYmNjOGs/view?usp=sharing
>>>>>>>>
>>>>>>>> In dmsg I see following linked with the brc driver :
>>>>>>>> dmesg | grep brc
>>>>>>>> [    7.987661] brcmf_sdio_drivestrengthinit: No SDIO Drive strength
>>>>>>>> init done for chip 43340 rev 2 pmurev 20
>>>>>>>> [    7.993487] usbcore: registered new interface driver brcmfmac
>>>>>>>> [    7.996318] brcmfmac_sdio mmc1:0001:1: Direct firmware load for
>>>>>>>> brcm/brcmfmac43340-sdio.bin failed with error -2
>>>>>>>> [    9.011572] brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl
>>>>>>>> 0x50
>>>>>>>> [   10.037365] brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl
>>>>>>>> 0x50
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> It could also mean driver was loaded manually. So please also make
>>>>>>>>> sure the friend uses kernel 4.0 or newer.
>>>>>>>>
>>>>>>>> Kernel : 4.0.0 #3 SMP PREEMPT Wed Apr 22 17:52:53 MSK 2015
>>>>>>>> Distro:
>>>>>>>> Distributor ID:    T100 Ubuntu 15.04
>>>>>>>> Description:    Ubuntu 15.04
>>>>>>>> Release:    15.04
>>>>>>>> Codename:    vivid
>>>>>>>> I ve installed it from the magic stick here
>>>>>>>> https://plus.google.com/communities/117853703024346186936
>>>>>>
>>>>>>
>>>>>> The driver complains about missing firmware and Ubuntu 15.04 does not
>>>>>> contain it.
>>>>>>
>>>>>> Please place this file
>>>>>>
>>>>>> https://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/plain/brcm/brcmfmac43340-sdio.bin
>>>>>>
>>>>>>
>>>>>> to /lib/firmware/brcm/brcmfmac43340-sdio.bin
>>>>>>
>>>>>> Hauke
>>>>>
>>>>>     Before that, the system was complaining about
>>>>> brcm/brcmfmac43340-sdio.bin, and now about brcm/brcmfmac43340-sdio.txt
>>>>> but errors are similar...
>>>>
>>>>
>>>> I guess your system in jinxed because of the hostname you setup :-p
>>>>
>>>> Ok, more seriously now. This is the nvram data I mentioned earlier. This
>>>> system may have that info stored in efi variable. You should be able
>>>> to find
>>>> it in /sys/firmware/efi/efivars/nvram-*.
>>>>
>>>> It may be necessary to run the following commands:
>>>>
>>>> # modprobe efivarfs
>>>> # mount -t efivarfs efivarfs /sys/firmware/efi/efivars
>>>>
>>>> Regards,
>>>> Arend
>>>>
>>> I cant be kind with manufacturer like broadcom and nvidia and those
>>> that are builting in such hardware ;)
>>
>> You are talking to a broadcom employee (in disguise ;-) ).
>>
>>> cat /sys/firmware/efi/efivars/nvram-74b00bd9-805a-4d61-b51f-43268123d113
>>> What am I supposed to do with this ?
>>
>> You are kidding? I'll chew it for you:
>>
>> $ cp /sys/firmware/efi/efivars/nvram-74b00bd9*
>> /lib/firmware/brcm/brcmfmac43340-sdio.txt
>
> Why can't the driver directly access this efi var? I haven't checked if
> this is already done somewhere, but letting a user do this manually does
> not seam nice.

Hi Hauke,

Well, we have been relying on firmware files and nvram being available 
under /lib/firmware. Not that we can not change that, but the manual 
copy is is a one time thing and only applicable for systems that ran 
Win8.1. I suppose there would be a way to get the efivar directly or 
have request_firmware api do it if it is considered firmware. Given that 
it is placed under /sys/firmware seems to imply that.

Regards,
Arend

>>
>> Regards,
>> Arend
>>
>