Re: BCM4313 & brcmsmac & 3.12: only semi-working?

[Arend van Spriel <arend@broadcom.com>]

On 11/19/14 14:46, Arend van Spriel wrote:
> On 11/17/14 23:36, Maximilian Engelhardt wrote:
>> On Thursday 09 October 2014 10:21:12 Rafał Miłecki wrote:
>>> On 9 October 2014 09:52, Arend van Spriel<arend@broadcom.com> wrote:
>>>> I have been staring at differences brcmsmac and our internal driver
>>>> (which
>>>> wl is derived from as well). Would you be able to make an mmiotrace
>>>> loading
>>>> wl and brcmsmac.
>>>
>>> Arend,
>>>
>>> As you noticed earlier, this is a WiFi card with BT. I've checked
>>> siutils.c from SDK 7.14.43.7 (got it from Netgear R8000) and it has
>>> three BCM4313 workarounds:
>>>
>>>
>>> 1) si_doattach
>>> SI_MSG(("Applying 4313 WARs\n"));
>>> si_pmu_chipcontrol(sih, 0, CCTRL_4313_12MA_LED_DRIVE,
>>> CCTRL_4313_12MA_LED_DRIVE);
>>>
>>> It's already implemented in bcma, see driver_chipcommon_pmu.c
>>>
>>>
>>> 2) si_epa_4313war
>>> /* EPA Fix */
>>> W_REG(sii->osh,&cc->gpiocontrol,
>>> R_REG(sii->osh,&cc->gpiocontrol) | GPIO_CTRL_EPA_EN_MASK);
>>>
>>> This is what you already implement in ai_epa_4313war
>>>
>>>
>>> 3) si_pmu_synth_pwrsw_4313_war
>>> I don't think it's implemented.
>>>
>>>
>>> 4) si_btcombo_p250_4313_war
>>> I also don't think it's implemented
>>>
>>>
>>> The last one (si_btcombo_p250_4313_war) looks promising, the comment
>>> above the function says:
>>> /** WL/BT control for 4313 btcombo boards>= P250 */
>>> I guess you need to re-implement this function and call it somewhere
>>> in brcmsmac.
>>
>> Hi,
>>
>> I just wanted to ask if there is any progress on this issue since I
>> haven't
>> heard anything for a month. Please let me know if you need any additional
>> information.
>
> I had a bit of problem with the mmiotrace files. Well, my (virtual)
> linux machine had as it ended up swapping on my disk and went down the
> drain.
>
> I still wanted to look at items 3) and 4) that Rafal mentioned although
> item 4) does not seems to apply to Michael's 4313 it may help yours. I
> submitted a patch to get counter statistics from the 802.11 core on the
> chip [1] as I am curious whether that will show a possible lead.
>
> Regards,
> Arend

+ links

patch [1] depends on patch [2]

[1] 
https://git.kernel.org/cgit/linux/kernel/git/linville/wireless-next.git/commit/drivers/net/wireless/brcm80211?id=3fe33c4ce
[2] 
https://git.kernel.org/cgit/linux/kernel/git/linville/wireless-next.git/commit/drivers/net/wireless/brcm80211?id=9146782b1b