Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-10-19 Thread Robert Marko 
What about all of the custom BDF-s that were upstreamed primarly for
IPQ40XX and lately various QCA99XX and QCA98XX radios?
By disabling ath10k-firmware and using the linux-firmware version we are
bound to have to use ipq-wifi again since firmware and board files are
really rarely updated in linux-firmware for QCA radios.


On 26 September 2018 at 11:49, Stanislaw Gruszka 
wrote:

> On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
> > On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
> > > ...
> > > With that update I am fine with squashing the mac80211 updates and
> > > pushing them to OpenWrt master.
> > >
> > > I checked the removed patches and could not find these two patches in
> > > the upstream kernel:
> > > *
> > > package/kernel/mac80211/patches/600-23-rt2x00-
> rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
> >
> > Yes, this one should be dropped according to Stanislaw Gruszka, we've
> > discussed this earlier, but can't spot the thread right now.
>
> Yes, please drop this one and also apply patches from:
> https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=
> ba8f5f0957e00e79af6ad1dbae7e649b720b2b01
> which changes whay how interrupts are handled.
>
> The patches were tested by versious users with positive feedback,
> what is documented in:
> https://bugzilla.kernel.org/show_bug.cgi?id=82751
>
> Thanks
> Stanislaw
>
> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
>
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-10-04 Thread Sven Schönhoff 
On Thu, Oct 4, 2018 at 11:25 AM Koen Vandeputte
 wrote:
>
>
>
> On 2018-10-03 20:06, Dirk Brenken wrote:
> > On Mi, 2018-10-03 at 17:59 +0200, Koen Vandeputte wrote:
> >> On 2018-08-13 17:14, John Crispin wrote:
> >>> Hi,
> >>>
> >>> as 19.01 will probably use v4.14 as baseline and ath79 wont be a
> >>> full
> >>> replacement for ar71xx by then we decided to bump ar71xx to v4.14.
> >>> This is available for testing inside my staging tree ->
> >>>
> >>> https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
> >>>
> >>>
> >>>
> >>> The tree also holds an update to mac80211, bumping it to the v4.18
> >>> wifi drivers. If anyone would like to test ar71xx and/or mac80211,
> >>> feel free to do so.
> >>>
> >>>  John
> >>>
> >>>
> >>> ___
> >>> openwrt-devel mailing list
> >>> openwrt-devel@lists.openwrt.org
> >>>
> >>> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> >>>
> >> Hi All,
> >>
> >> I just tested today's master state (with 4.14 enabled) on following
> >> ar71xx devices which I have at my disposal:
> >>
> >>
> >> - Mikrotik rb2011
> >> - Mikrotik rb912 (with ath10k radio)
> >> - gl.inet MiFi
> >> - Ubnt Rocket M5
> >> - Ubnt Rocket M3
> >> - Ubnt Rocket M2
> >>
> >> All was OK and functioning properly on these devices.
> >>
> >>
> >> If you can test this on a device other than these, please let me know
> >> if
> >> your devices works properly and specify the device model name as
> >> accurately as possible.
> >> Also please provide a full bootlog of the device!
> >>
> >>
> >> To minimize the effort for building/testing, I have updated the
> >> ar71xx_414 branch in my staging tree.
> >> This branch is a clone of the master, but enables 4.14 for ar71xx.
> >>
> >>
> >> Thanks,
> >>
> >> Koen
> >>
> >>
> >> [1]
> >> https://git.openwrt.org/?p=openwrt/staging/xback.git;a=summary
> >>
> >>
> >> ___
> >> openwrt-devel mailing list
> >> openwrt-devel@lists.openwrt.org
> >>
> >> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> >>
> > Hi Koen,
> >
> > as reported few weeks ago, even with current master based on kernel
> > 4.14 on a GL.iNet GL-AR750 my 5G wireless device get lost/disappeared.
> > I'll send you dmesg bootlogs (4.9 vs. 4.14) and my config via PM.
> >
> > Thanks!
> > Dirk
> >
> >
> Hi Dirk,
>
> Thank you for the extended logs.
> It seems you have the same issue as reported by Sven.
>
>
> [   34.414816] pci :00:00.0: no irq found for pin 1
>
>
> I've added a patch in my staging tree (ar71xx_414 branch).  [1]
> Could you try this one out and provide me the bootlog please?
>
>
> Appreciated,
>
> Koen
>
> [1]
> https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414


Hi Koen,

the fix works for me too. Thank you!

I've successfully tested
- TP-LINK Archer C7 v4
https://pastebin.com/cm35pwBV
- TP-Link CPE210 v2.0
https://pastebin.com/fR554pSz
- TP-LINK TL-WR1043N v5
https://pastebin.com/DNR1vMUc


Best regards
Sven

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-10-04 Thread Dirk Brenken 
On Do, 2018-10-04 at 11:25 +0200, Koen Vandeputte wrote:
> 
> On 2018-10-03 20:06, Dirk Brenken wrote:
> > On Mi, 2018-10-03 at 17:59 +0200, Koen Vandeputte wrote:
> > > On 2018-08-13 17:14, John Crispin wrote:
> > > > Hi,
> > > > 
> > > > as 19.01 will probably use v4.14 as baseline and ath79 wont be
> > > > a
> > > > full
> > > > replacement for ar71xx by then we decided to bump ar71xx to
> > > > v4.14.
> > > > This is available for testing inside my staging tree ->
> > > > 
> > > > https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
> > > > 
> > > >   
> > > > 
> > > > 
> > > > The tree also holds an update to mac80211, bumping it to the
> > > > v4.18
> > > > wifi drivers. If anyone would like to test ar71xx and/or
> > > > mac80211,
> > > > feel free to do so.
> > > > 
> > > >  John
> > > > 
> > > > 
> > > > ___
> > > > openwrt-devel mailing list
> > > > openwrt-devel@lists.openwrt.org
> > > > 
> > > > 
> > > > https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> > > > 
> > > > 
> > > 
> > > Hi All,
> > > 
> > > I just tested today's master state (with 4.14 enabled) on
> > > following
> > > ar71xx devices which I have at my disposal:
> > > 
> > > 
> > > - Mikrotik rb2011
> > > - Mikrotik rb912 (with ath10k radio)
> > > - gl.inet MiFi
> > > - Ubnt Rocket M5
> > > - Ubnt Rocket M3
> > > - Ubnt Rocket M2
> > > 
> > > All was OK and functioning properly on these devices.
> > > 
> > > 
> > > If you can test this on a device other than these, please let me
> > > know
> > > if
> > > your devices works properly and specify the device model name as
> > > accurately as possible.
> > > Also please provide a full bootlog of the device!
> > > 
> > > 
> > > To minimize the effort for building/testing, I have updated the
> > > ar71xx_414 branch in my staging tree.
> > > This branch is a clone of the master, but enables 4.14 for
> > > ar71xx.
> > > 
> > > 
> > > Thanks,
> > > 
> > > Koen
> > > 
> > > 
> > > [1]
> > > https://git.openwrt.org/?p=openwrt/staging/xback.git;a=summary
> > > 
> > > 
> > > 
> > > ___
> > > openwrt-devel mailing list
> > > openwrt-devel@lists.openwrt.org
> > > 
> > > 
> > > https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> > > 
> > > 
> > 
> > Hi Koen,
> > 
> > as reported few weeks ago, even with current master based on kernel
> > 4.14 on a GL.iNet GL-AR750 my 5G wireless device get
> > lost/disappeared.
> > I'll send you dmesg bootlogs (4.9 vs. 4.14) and my config via PM.
> > 
> > Thanks!
> > Dirk
> > 
> > 
> 
> Hi Dirk,
> 
> Thank you for the extended logs.
> It seems you have the same issue as reported by Sven.
> 
> 
> [   34.414816] pci :00:00.0: no irq found for pin 1
> 
> 
> I've added a patch in my staging tree (ar71xx_414 branch).  [1]
> Could you try this one out and provide me the bootlog please?
> 
> 
> Appreciated,
> 
> Koen
> 
> [1] 
> https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414
> 
> 
> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> 
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> 
> 

Hi Koen,

great job - the 5G device is back again with kernel 4.14 ... the
bootlog will arrive in a few seconds!

Many thanks!
Dirk




signature.asc
Description: This is a digitally signed message part
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-10-04 Thread Koen Vandeputte 



On 2018-10-03 20:06, Dirk Brenken wrote:

On Mi, 2018-10-03 at 17:59 +0200, Koen Vandeputte wrote:

On 2018-08-13 17:14, John Crispin wrote:

Hi,

as 19.01 will probably use v4.14 as baseline and ath79 wont be a
full
replacement for ar71xx by then we decided to bump ar71xx to v4.14.
This is available for testing inside my staging tree ->

https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
  



The tree also holds an update to mac80211, bumping it to the v4.18
wifi drivers. If anyone would like to test ar71xx and/or mac80211,
feel free to do so.

 John


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org

https://lists.openwrt.org/mailman/listinfo/openwrt-devel


Hi All,

I just tested today's master state (with 4.14 enabled) on following
ar71xx devices which I have at my disposal:


- Mikrotik rb2011
- Mikrotik rb912 (with ath10k radio)
- gl.inet MiFi
- Ubnt Rocket M5
- Ubnt Rocket M3
- Ubnt Rocket M2

All was OK and functioning properly on these devices.


If you can test this on a device other than these, please let me know
if
your devices works properly and specify the device model name as
accurately as possible.
Also please provide a full bootlog of the device!


To minimize the effort for building/testing, I have updated the
ar71xx_414 branch in my staging tree.
This branch is a clone of the master, but enables 4.14 for ar71xx.


Thanks,

Koen


[1]
https://git.openwrt.org/?p=openwrt/staging/xback.git;a=summary


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org

https://lists.openwrt.org/mailman/listinfo/openwrt-devel


Hi Koen,

as reported few weeks ago, even with current master based on kernel
4.14 on a GL.iNet GL-AR750 my 5G wireless device get lost/disappeared.
I'll send you dmesg bootlogs (4.9 vs. 4.14) and my config via PM.

Thanks!
Dirk



Hi Dirk,

Thank you for the extended logs.
It seems you have the same issue as reported by Sven.


[   34.414816] pci :00:00.0: no irq found for pin 1


I've added a patch in my staging tree (ar71xx_414 branch).  [1]
Could you try this one out and provide me the bootlog please?


Appreciated,

Koen

[1] 
https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-10-03 Thread Koen Vandeputte 



On 2018-08-13 17:14, John Crispin wrote:

Hi,

as 19.01 will probably use v4.14 as baseline and ath79 wont be a full 
replacement for ar71xx by then we decided to bump ar71xx to v4.14. 
This is available for testing inside my staging tree ->


https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging 



The tree also holds an update to mac80211, bumping it to the v4.18 
wifi drivers. If anyone would like to test ar71xx and/or mac80211, 
feel free to do so.


    John


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel


Hi All,

I just tested today's master state (with 4.14 enabled) on following 
ar71xx devices which I have at my disposal:



- Mikrotik rb2011
- Mikrotik rb912 (with ath10k radio)
- gl.inet MiFi
- Ubnt Rocket M5
- Ubnt Rocket M3
- Ubnt Rocket M2

All was OK and functioning properly on these devices.


If you can test this on a device other than these, please let me know if 
your devices works properly and specify the device model name as 
accurately as possible.

Also please provide a full bootlog of the device!


To minimize the effort for building/testing, I have updated the 
ar71xx_414 branch in my staging tree.

This branch is a clone of the master, but enables 4.14 for ar71xx.


Thanks,

Koen


[1] https://git.openwrt.org/?p=openwrt/staging/xback.git;a=summary

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-30 Thread Hauke Mehrtens 
On 09/26/2018 12:01 AM, Daniel Golle wrote:
> On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
>> ...
>> With that update I am fine with squashing the mac80211 updates and
>> pushing them to OpenWrt master.
>>
>> I checked the removed patches and could not find these two patches in
>> the upstream kernel:
>> *
>> package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
> 
> Yes, this one should be dropped according to Stanislaw Gruszka, we've
> discussed this earlier, but can't spot the thread right now.

Ok, I just wanted to make sure this was done by intention and not by
mistake. Now I know that this was done by intention.

Hauke

>> *
>> package/kernel/mac80211/patches/367-Revert-ath10k-disable-wake_tx_queue-for-older-device.patch

For me it looks ok that this one got removed.

Hauke



signature.asc
Description: OpenPGP digital signature
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread Robert Marko 
On Wed, 2018-09-26 at 17:24 +0200, John Crispin wrote:
> 
> > We will have to send patches to linux-firmware to get them to
> > update 
> > the files.
> > 
> > John
> 
> as a temp fix we need to revert
> e6bd568051c7ca77a59783fe50203f6e2a427f19 
> right ?
> 
>  John
> 

Sorry for top posting, Gmail is not making life easy.

Asking Kalle Valo to push new board binaries would be easiest.

Its not just reverting that patch as we have devices like OpenMesh A62
that was added only after upstreaming BDF-s for IPQ4019 and its
third radio that is QCA9888 based.

Can we use linux-firmware for firmware and the old ath10k-firmware for
board files as a temporary solution.

Robert


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread John Crispin 


On 26/09/2018 17:14, John Crispin wrote:


On 26/09/2018 17:12, Robert Marko wrote:

I understand the issues with never firmware breaking driver features,
and I dont have anything against using linux-firmware for the firmware
but board files there are too old.
For example board-2.bin for IPQ4019 was last updated on 15.02.2018.
and it does not contain any of the BDF that were upstreamed for
various boards in ipq40xx target.
I just checked and it only has BDF-s for the OpenMesh-A42 board while
ath10k-firmware one has BDFs for eight more boards that are OpenWrt
supported.
So this switch will effectively break the radios on most of ipq40xx 
boards.


Only way around it is to again revert to using ipq-wifi for every board

Regards
Robert Marko

which part of "please dont top post" was not clear, just for future 
reference ?






On 26 September 2018 at 16:59, John Crispin  wrote:

On 26/09/2018 16:52, Robert Marko wrote:

What about all of the custom BDF-s that were upstreamed primarly for
IPQ40XX and lately various QCA99XX and QCA98XX radios?
By disabling ath10k-firmware and using the linux-firmware version 
we are

bound to have to use ipq-wifi again since firmware and board files are
really rarely updated in linux-firmware for QCA radios.


please dont top post ...


On 26 September 2018 at 11:49, Stanislaw Gruszka mailto:sgrus...@redhat.com>> wrote:

 On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
 > On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
 > > ...
 > > With that update I am fine with squashing the mac80211 
updates and

 > > pushing them to OpenWrt master.
 > >
 > > I checked the removed patches and could not find these two
 patches in
 > > the upstream kernel:
 > > *
 > >

package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch 


 >
 > Yes, this one should be dropped according to Stanislaw Gruszka,
 we've
 > discussed this earlier, but can't spot the thread right now.

 Yes, please drop this one and also apply patches from:

https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=ba8f5f0957e00e79af6ad1dbae7e649b720b2b01 



 


 which changes whay how interrupts are handled.

 The patches were tested by versious users with positive feedback,
 what is documented in:
 https://bugzilla.kernel.org/show_bug.cgi?id=82751


 Thanks
 Stanislaw

 ___
 openwrt-devel mailing list
 openwrt-devel@lists.openwrt.org
 
https://lists.openwrt.org/mailman/listinfo/openwrt-devel



using randomly chosen firmwares from a github repo is no real 
alternative.
the linux-firmware files seem to geed enough for most mayor distros. 
the

htt/wmi abi can change between drivers/firmware and newer FW with older
drivers can cause all sorts of weird issues like the DFS false 
positive one.

the aim of this series is to get much closer to upstream in regards to
wireless allowing us to work closer with upstream.

 John



We will have to send patches to linux-firmware to get them to update 
the files.


    John



as a temp fix we need to revert e6bd568051c7ca77a59783fe50203f6e2a427f19 
right ?


    John



___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread John Crispin 


On 26/09/2018 17:12, Robert Marko wrote:

I understand the issues with never firmware breaking driver features,
and I dont have anything against using linux-firmware for the firmware
but board files there are too old.
For example board-2.bin for IPQ4019 was last updated on 15.02.2018.
and it does not contain any of the BDF that were upstreamed for
various boards in ipq40xx target.
I just checked and it only has BDF-s for the OpenMesh-A42 board while
ath10k-firmware one has BDFs for eight more boards that are OpenWrt
supported.
So this switch will effectively break the radios on most of ipq40xx boards.

Only way around it is to again revert to using ipq-wifi for every board

Regards
Robert Marko

which part of "please dont top post" was not clear, just for future 
reference ?






On 26 September 2018 at 16:59, John Crispin  wrote:

On 26/09/2018 16:52, Robert Marko wrote:

What about all of the custom BDF-s that were upstreamed primarly for
IPQ40XX and lately various QCA99XX and QCA98XX radios?
By disabling ath10k-firmware and using the linux-firmware version we are
bound to have to use ipq-wifi again since firmware and board files are
really rarely updated in linux-firmware for QCA radios.


please dont top post ...


On 26 September 2018 at 11:49, Stanislaw Gruszka mailto:sgrus...@redhat.com>> wrote:

 On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
 > On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
 > > ...
 > > With that update I am fine with squashing the mac80211 updates and
 > > pushing them to OpenWrt master.
 > >
 > > I checked the removed patches and could not find these two
 patches in
 > > the upstream kernel:
 > > *
 > >

package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
 >
 > Yes, this one should be dropped according to Stanislaw Gruszka,
 we've
 > discussed this earlier, but can't spot the thread right now.

 Yes, please drop this one and also apply patches from:

https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=ba8f5f0957e00e79af6ad1dbae7e649b720b2b01


 which changes whay how interrupts are handled.

 The patches were tested by versious users with positive feedback,
 what is documented in:
 https://bugzilla.kernel.org/show_bug.cgi?id=82751
 

 Thanks
 Stanislaw

 ___
 openwrt-devel mailing list
 openwrt-devel@lists.openwrt.org
 
 https://lists.openwrt.org/mailman/listinfo/openwrt-devel
 



using randomly chosen firmwares from a github repo is no real alternative.
the linux-firmware files seem to geed enough for most mayor distros. the
htt/wmi abi can change between drivers/firmware and newer FW with older
drivers can cause all sorts of weird issues like the DFS false positive one.
the aim of this series is to get much closer to upstream in regards to
wireless allowing us to work closer with upstream.

 John



We will have to send patches to linux-firmware to get them to update the 
files.


    John





___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread Robert Marko 
I understand the issues with never firmware breaking driver features,
and I dont have anything against using linux-firmware for the firmware
but board files there are too old.
For example board-2.bin for IPQ4019 was last updated on 15.02.2018.
and it does not contain any of the BDF that were upstreamed for
various boards in ipq40xx target.
I just checked and it only has BDF-s for the OpenMesh-A42 board while
ath10k-firmware one has BDFs for eight more boards that are OpenWrt
supported.
So this switch will effectively break the radios on most of ipq40xx boards.

Only way around it is to again revert to using ipq-wifi for every board

Regards
Robert Marko


On 26 September 2018 at 16:59, John Crispin  wrote:
>
> On 26/09/2018 16:52, Robert Marko wrote:
>>
>> What about all of the custom BDF-s that were upstreamed primarly for
>> IPQ40XX and lately various QCA99XX and QCA98XX radios?
>> By disabling ath10k-firmware and using the linux-firmware version we are
>> bound to have to use ipq-wifi again since firmware and board files are
>> really rarely updated in linux-firmware for QCA radios.
>>
> please dont top post ...
>>
>>
>> On 26 September 2018 at 11:49, Stanislaw Gruszka > > wrote:
>>
>> On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
>> > On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
>> > > ...
>> > > With that update I am fine with squashing the mac80211 updates and
>> > > pushing them to OpenWrt master.
>> > >
>> > > I checked the removed patches and could not find these two
>> patches in
>> > > the upstream kernel:
>> > > *
>> > >
>>
>> package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
>> >
>> > Yes, this one should be dropped according to Stanislaw Gruszka,
>> we've
>> > discussed this earlier, but can't spot the thread right now.
>>
>> Yes, please drop this one and also apply patches from:
>>
>> https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=ba8f5f0957e00e79af6ad1dbae7e649b720b2b01
>>
>> 
>> which changes whay how interrupts are handled.
>>
>> The patches were tested by versious users with positive feedback,
>> what is documented in:
>> https://bugzilla.kernel.org/show_bug.cgi?id=82751
>> 
>>
>> Thanks
>> Stanislaw
>>
>> ___
>> openwrt-devel mailing list
>> openwrt-devel@lists.openwrt.org
>> 
>> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
>> 
>>
>>
> using randomly chosen firmwares from a github repo is no real alternative.
> the linux-firmware files seem to geed enough for most mayor distros. the
> htt/wmi abi can change between drivers/firmware and newer FW with older
> drivers can cause all sorts of weird issues like the DFS false positive one.
> the aim of this series is to get much closer to upstream in regards to
> wireless allowing us to work closer with upstream.
>
> John
>
>
>

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread John Crispin 


On 26/09/2018 16:52, Robert Marko wrote:
What about all of the custom BDF-s that were upstreamed primarly for 
IPQ40XX and lately various QCA99XX and QCA98XX radios?
By disabling ath10k-firmware and using the linux-firmware version we 
are bound to have to use ipq-wifi again since firmware and board files 
are really rarely updated in linux-firmware for QCA radios.



please dont top post ...


On 26 September 2018 at 11:49, Stanislaw Gruszka > wrote:


On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
> On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
> > ...
> > With that update I am fine with squashing the mac80211 updates and
> > pushing them to OpenWrt master.
> >
> > I checked the removed patches and could not find these two
patches in
> > the upstream kernel:
> > *
> >

package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
>
> Yes, this one should be dropped according to Stanislaw Gruszka,
we've
> discussed this earlier, but can't spot the thread right now.

Yes, please drop this one and also apply patches from:

https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=ba8f5f0957e00e79af6ad1dbae7e649b720b2b01


which changes whay how interrupts are handled.

The patches were tested by versious users with positive feedback,
what is documented in:
https://bugzilla.kernel.org/show_bug.cgi?id=82751


Thanks
Stanislaw

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org

https://lists.openwrt.org/mailman/listinfo/openwrt-devel



using randomly chosen firmwares from a github repo is no real 
alternative. the linux-firmware files seem to geed enough for most mayor 
distros. the htt/wmi abi can change between drivers/firmware and newer 
FW with older drivers can cause all sorts of weird issues like the DFS 
false positive one. the aim of this series is to get much closer to 
upstream in regards to wireless allowing us to work closer with upstream.


    John




___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-26 Thread Stanislaw Gruszka 
On Wed, Sep 26, 2018 at 12:01:54AM +0200, Daniel Golle wrote:
> On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
> > ...
> > With that update I am fine with squashing the mac80211 updates and
> > pushing them to OpenWrt master.
> > 
> > I checked the removed patches and could not find these two patches in
> > the upstream kernel:
> > *
> > package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
> 
> Yes, this one should be dropped according to Stanislaw Gruszka, we've
> discussed this earlier, but can't spot the thread right now.

Yes, please drop this one and also apply patches from:
https://git.openwrt.org/?p=openwrt/staging/dangole.git;a=commit;h=ba8f5f0957e00e79af6ad1dbae7e649b720b2b01
which changes whay how interrupts are handled.

The patches were tested by versious users with positive feedback,
what is documented in:
https://bugzilla.kernel.org/show_bug.cgi?id=82751

Thanks
Stanislaw

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-25 Thread Daniel Golle 
On Tue, Sep 25, 2018 at 11:15:14PM +0200, Hauke Mehrtens wrote:
> ...
> With that update I am fine with squashing the mac80211 updates and
> pushing them to OpenWrt master.
> 
> I checked the removed patches and could not find these two patches in
> the upstream kernel:
> *
> package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch

Yes, this one should be dropped according to Stanislaw Gruszka, we've
discussed this earlier, but can't spot the thread right now.


> *
> package/kernel/mac80211/patches/367-Revert-ath10k-disable-wake_tx_queue-for-older-device.patch
> 
> I assume it is not a problem that they got removed.
> 
> Hauke
> 




> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-25 Thread Hauke Mehrtens 
On 09/24/2018 10:47 PM, Hauke Mehrtens wrote:
> On 08/13/2018 05:14 PM, John Crispin wrote:
>> Hi,
>>
>> as 19.01 will probably use v4.14 as baseline and ath79 wont be a full
>> replacement for ar71xx by then we decided to bump ar71xx to v4.14. This
>> is available for testing inside my staging tree ->
>>
>> https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
>>
>>
>> The tree also holds an update to mac80211, bumping it to the v4.18 wifi
>> drivers. If anyone would like to test ar71xx and/or mac80211, feel free
>> to do so.
>>
>>     John
> 
> Hi John,
> 
> most of the stuff we need in backports is now already in the upstream git:
> https://git.kernel.org/pub/scm/linux/kernel/git/backports/backports.git/log/
> 
> Johannes is still waiting for a fix iwlwifi then I will make an
> officially tagged backports release of that. I would prefer to use that
> in OpenWrt as a base for now. Can we wait with the mac80211 update in
> OpenWrt some days. It still makes sense to test the stuff which is your
> staging git tree, it should work with kernel 4.9 and 4.14, it still has
> a problem with kernel 4.19 which is also fixed upstream backports.

Hi John,

Now backports-4.19-rc5-1.tar.xz was released, I added it here:
https://git.openwrt.org/?p=openwrt/staging/hauke.git;a=commitdiff;h=da9c00dcc6401f55cebec2de272f3ab2bd492e1b

With that update I am fine with squashing the mac80211 updates and
pushing them to OpenWrt master.

I checked the removed patches and could not find these two patches in
the upstream kernel:
*
package/kernel/mac80211/patches/600-23-rt2x00-rt2800mmio-add-a-workaround-for-spurious-TX_F.patch
*
package/kernel/mac80211/patches/367-Revert-ath10k-disable-wake_tx_queue-for-older-device.patch

I assume it is not a problem that they got removed.

Hauke



signature.asc
Description: OpenPGP digital signature
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-24 Thread Hauke Mehrtens 
On 08/13/2018 05:14 PM, John Crispin wrote:
> Hi,
> 
> as 19.01 will probably use v4.14 as baseline and ath79 wont be a full
> replacement for ar71xx by then we decided to bump ar71xx to v4.14. This
> is available for testing inside my staging tree ->
> 
> https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
> 
> 
> The tree also holds an update to mac80211, bumping it to the v4.18 wifi
> drivers. If anyone would like to test ar71xx and/or mac80211, feel free
> to do so.
> 
>     John

Hi John,

most of the stuff we need in backports is now already in the upstream git:
https://git.kernel.org/pub/scm/linux/kernel/git/backports/backports.git/log/

Johannes is still waiting for a fix iwlwifi then I will make an
officially tagged backports release of that. I would prefer to use that
in OpenWrt as a base for now. Can we wait with the mac80211 update in
OpenWrt some days. It still makes sense to test the stuff which is your
staging git tree, it should work with kernel 4.9 and 4.14, it still has
a problem with kernel 4.19 which is also fixed upstream backports.

Hauke



signature.asc
Description: OpenPGP digital signature
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-09-01 Thread Koen Vandeputte 



Hi John,

Thanks for the updates!


mac80211:


I've tested mac80211 on a large number of targets in the field:

- ~25 devices
- half of them having up to 4 wlan ifaces
- all running in IBSS
- Distance varying from 20m up to ~50km

For my cases, it seems to work well.


Small nit-pick:
In the bootlog, I'm seeing this:

[   11.33] Loading modules backported from Linux version 
v4.18-rc7-0-gacb1872577b3
[   11.33] Backport generated by backports.git 
v4.14-rc2-1-71-g39f4ec76



Is the 2nd line correct?


erm no, i'll rebase ontop of v4.18 (without RC) the next days and look 
into this. its most likely the last version that backports.git was 
tested against and i bumped the s/patch set to v4.18


thanks for testing !
    John



Hi John,

I've been testing the mac80211 bump (4.18 RC7) in the field for some 
time now.
It's currently running on roughly 40 moving devices (~90 wlan 
interfaces) at various distances.

(The devices are mounted on windmill's and vessels.)

All of them are running in IBSS mode, 802.11n, 5.8G, HT40, Long GI.


Enabling dynack triggers crashes in this mac80211 version. (RCU stalls)
I haven't seen these in the past few months running the current OpenWrt 
4.14-RC2 mac version.


https://pastebin.com/raw/UcHt17yZ


I don't know if this will also occur in AP or MESH mode .. hence I'm 
reporting it here ..



Dear Johannes,

Apologies for dragging you into this thread but, being the lead 
regarding mac80211, maybe you have a clue where this gets triggered.



Dear Lorenzo,

Being the designer of dynack, maybe you have an idea?


Thanks all,

Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-30 Thread Sven Schönhoff 
> > My Archer C7 v4 still bricks with the same error reported earlier and
> > confirmed by Koen: Atheros AR8216/AR8236/AR8316: probe of
> > ag71xx-mdio.0:00 failed with error -22
> >
>
> Hi All,
>
> I created a new branch in my staging tree, which is a replica of the
> latest master tree,
> but adds a fix for the switch issue + enables 4.14 on ar71xx by default.
>
> Feel free to build it and try it out.
> Please provide some feedback if it worked for you.
>
> https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414

Thank you Koen, it fixes the switch issue for my Archer C7 v4 too:

[0.713675] switch0: Atheros AR8337 rev. 2 switch registered on ag71xx-mdio.0
[1.432868] libphy: ag71xx_mdio: probed
[2.066476] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00
[uid=004dd036, driver=Atheros AR8216/AR8236/AR8316]
[2.077962] eth0: Atheros AG71xx at 0xb900, irq 4, mode:SGMII

Now that the switch works I've noticed that the 5GHz wifi on my Archer
C7 v4 doesn't work with Kernel 4.14 enabled:

[   10.005449] pci :00:00.0: no irq found for pin 1
[   10.011893] PCI: Enabling device :00:00.0 ( -> 0002)
[   10.017960] ath10k_pci :00:00.0: pci irq legacy oper_irq_mode 1
irq_mode 0 reset_mode 0
[   10.298735] ath10k_pci :00:00.0: Direct firmware load for
ath10k/pre-cal-pci-:00:00.0.bin failed with error -2
[   10.309822] ath10k_pci :00:00.0: Falling back to user helper
[   10.475669] firmware ath10k!pre-cal-pci-:00:00.0.bin:
firmware_loading_store: map pages failed
[   10.485244] ath10k_pci :00:00.0: Direct firmware load for
ath10k/cal-pci-:00:00.0.bin failed with error -2
[   10.495966] ath10k_pci :00:00.0: Falling back to user helper
[   10.704794] ath10k_pci :00:00.0: Direct firmware load for
ath10k/QCA988X/hw2.0/firmware-6.bin failed with error -2
[   10.715891] ath10k_pci :00:00.0: Falling back to user helper
[   10.811651] firmware ath10k!QCA988X!hw2.0!firmware-6.bin:
firmware_loading_store: map pages failed
[   10.827803] ath10k_pci :00:00.0: qca988x hw2.0 target
0x4100016c chip_id 0x043202ff sub :
[   10.837374] ath10k_pci :00:00.0: kconfig debug 0 debugfs 1
tracing 0 dfs 1 testmode 1
[   10.850379] ath10k_pci :00:00.0: firmware ver 10.2.4-1.0-00037
api 5 features no-p2p,raw-mode,mfp,allows-mesh-bcast crc32 a4a52adb
[   10.948081] ath10k_pci :00:00.0: Direct firmware load for
ath10k/QCA988X/hw2.0/board-2.bin failed with error -2
[   10.958903] ath10k_pci :00:00.0: Falling back to user helper
[   11.033046] firmware ath10k!QCA988X!hw2.0!board-2.bin:
firmware_loading_store: map pages failed
[   11.042531] ath10k_pci :00:00.0: board_file api 1 bmi_id N/A
crc32 bebc7c08
[   13.164858] ath10k_pci :00:00.0: failed to receive control
response completion, polling..
[   14.204850] ath10k_pci :00:00.0: Service connect timeout
[   14.210703] ath10k_pci :00:00.0: failed to connect htt (-145)
[   14.346078] ath10k_pci :00:00.0: could not init core (-145)
[   14.352322] ath10k_pci :00:00.0: could not probe fw (-145)
[   14.40] ieee80211 phy1: Atheros AR9561 Rev:0 mem=0xb810, irq=47

Full log: https://pastebin.com/x6u8LEXN

Best regards
Sven
Am Do., 30. Aug. 2018 um 19:18 Uhr schrieb Koen Vandeputte
:
>
>
>
> On 2018-08-23 19:00, Sven Schönhoff wrote:
>  curious what i missed as my archer-c7-v4 is working fine. TPLINK
>  AC1750 device
> >>> Hi,
> >>> please dont top post
> >>> I have no idea what you are asking.
> >>>
> >> Meaning I don't understand his issue mine works fine, so I'm curious as to
> >> what's different
> > I would guess you didn't change KERNEL_PATCHVER to 4.14 in
> > target/linux/ar71xx/Makefile. Kernel 4.14 is not enabled by default
> > yet for ar71xx.
> >
> > My Archer C7 v4 still bricks with the same error reported earlier and
> > confirmed by Koen: Atheros AR8216/AR8236/AR8316: probe of
> > ag71xx-mdio.0:00 failed with error -22
> >
>
> Hi All,
>
> I created a new branch in my staging tree, which is a replica of the
> latest master tree,
> but adds a fix for the switch issue + enables 4.14 on ar71xx by default.
>
> Feel free to build it and try it out.
> Please provide some feedback if it worked for you.
>
> https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414
>
>
>
> [0.00] Linux version 4.14.67 (kvdp@kvdp-BRIX) (gcc version 7.3.0
> (OpenWrt GCC 7.3.0 r7984+5-555c59230402)) #0 Thu Aug 30 14:32:54 2018
> ...
> [0.138438] MIPS: machine is Mikrotik RouterBOARD 2011UiAS
> ...
> [   11.163357] libphy: Fixed MDIO Bus: probed
> [   11.319898] libphy: ag71xx_mdio: probed
> [   11.360844] switch0: Atheros AR8327 rev. 4 switch registered on
> ag71xx-mdio.0
> [   12.447398] libphy: ag71xx_mdio: probed
> [   13.077402] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00
> [uid=004dd034, driver=Atheros AR8216/AR8236/AR8316]
> [   13.088989] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
>

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-30 Thread Koen Vandeputte 



On 2018-08-23 19:00, Sven Schönhoff wrote:

curious what i missed as my archer-c7-v4 is working fine. TPLINK
AC1750 device

Hi,
please dont top post
I have no idea what you are asking.


Meaning I don't understand his issue mine works fine, so I'm curious as to
what's different

I would guess you didn't change KERNEL_PATCHVER to 4.14 in
target/linux/ar71xx/Makefile. Kernel 4.14 is not enabled by default
yet for ar71xx.

My Archer C7 v4 still bricks with the same error reported earlier and
confirmed by Koen: Atheros AR8216/AR8236/AR8316: probe of
ag71xx-mdio.0:00 failed with error -22



Hi All,

I created a new branch in my staging tree, which is a replica of the 
latest master tree,

but adds a fix for the switch issue + enables 4.14 on ar71xx by default.

Feel free to build it and try it out.
Please provide some feedback if it worked for you.

https://git.openwrt.org/?p=openwrt/staging/xback.git;a=shortlog;h=refs/heads/ar71xx_414



[    0.00] Linux version 4.14.67 (kvdp@kvdp-BRIX) (gcc version 7.3.0 
(OpenWrt GCC 7.3.0 r7984+5-555c59230402)) #0 Thu Aug 30 14:32:54 2018

...
[    0.138438] MIPS: machine is Mikrotik RouterBOARD 2011UiAS
...
[   11.163357] libphy: Fixed MDIO Bus: probed
[   11.319898] libphy: ag71xx_mdio: probed
[   11.360844] switch0: Atheros AR8327 rev. 4 switch registered on 
ag71xx-mdio.0

[   12.447398] libphy: ag71xx_mdio: probed
[   13.077402] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 
[uid=004dd034, driver=Atheros AR8216/AR8236/AR8316]

[   13.088989] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
[   13.717716] ag71xx-mdio.1: Found an AR934X built-in switch
[   13.769990] eth1: Atheros AG71xx at 0xba00, irq 5, mode:GMII




Thanks,

Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-23 Thread Sven Schönhoff 
> > > curious what i missed as my archer-c7-v4 is working fine. TPLINK
> > > AC1750 device
> > Hi,
> > please dont top post
> > I have no idea what you are asking.
> >
>
> Meaning I don't understand his issue mine works fine, so I'm curious as to
> what's different

I would guess you didn't change KERNEL_PATCHVER to 4.14 in
target/linux/ar71xx/Makefile. Kernel 4.14 is not enabled by default
yet for ar71xx.

My Archer C7 v4 still bricks with the same error reported earlier and
confirmed by Koen: Atheros AR8216/AR8236/AR8316: probe of
ag71xx-mdio.0:00 failed with error -22

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-23 Thread Outback Dingo 
On Wed, Aug 22, 2018, 17:10 John Crispin  wrote:

>
>
> On 22/08/18 12:02, Outback Dingo wrote:
> > curious what i missed as my archer-c7-v4 is working fine. TPLINK
> > AC1750 device
> Hi,
> please dont top post
> I have no idea what you are asking.
>

Meaning I don't understand his issue mine works fine, so I'm curious as to
what's different

>  John
>
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-23 Thread Koen Vandeputte 



On 2018-08-22 16:14, Koen Vandeputte wrote:





ar71xx:

I'm now currently testing your ar71xx 4.14 bump (comitted today in 
master) on a few different targets:


- Gl-mifi
- Rocket-m5
- RB912
- RB2011
...

Will let you know the results.

Hi John,

I started with trying on a Mikrotik RB2011 as these have a Serial port.

Ethernet does not function correctly.
I see it transmitting packets,  but receiving fails.



Logs from kernel 4.9:

[    1.052813] switch0: Atheros AR8327 rev. 4 switch registered on 
ag71xx-mdio.0

[    1.766947] libphy: ag71xx_mdio: probed
[    2.400863] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 
[uid=004dd034, driver=Atheros AR8216/AR8236/AR8316]

[    2.412536] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
[    3.041177] ag71xx-mdio.1: Found an AR934X built-in switch
[    3.093580] eth1: Atheros AG71xx at 0xba00, irq 5, mode:GMII
[    3.101108] NET: Registered protocol family 10



Logs from kernel 4.14

[    1.066007] switch0: Atheros AR8327 rev. 4 switch registered on 
ag71xx-mdio.0
[    1.073409] Atheros AR8216/AR8236/AR8316: probe of ag71xx-mdio.0:00 
failed with error -22

[    1.102455] libphy: ag71xx_mdio: probed
[    1.737938] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 
[uid=004dd034, driver=Generic PHY]

[    1.747994] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
[    2.377642] ag71xx-mdio.1: Found an AR934X built-in switch
[    2.429938] eth1: Atheros AG71xx at 0xba00, irq 5, mode:GMII
[    2.437637] NET: Registered protocol family 10



Hi John,

I've digged a little bit deeper into the issue and compared/tested the 
driver code between 4.9 and 4.14


The main problem is that within ar8327_hw_config_pdata()  [1]
pdata is NULL when this function gets called ..


[1] 
https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=target/linux/generic/files/drivers/net/phy/ar8327.c;h=7bfc1875094f1bcb149898b2184035040dfce2fd;hb=HEAD#l486



I can confirm that the board_info data gets loaded in the list though.
No idea currently why platform_data would be NULL currently ..


Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread Koen Vandeputte 





ar71xx:

I'm now currently testing your ar71xx 4.14 bump (comitted today in 
master) on a few different targets:


- Gl-mifi
- Rocket-m5
- RB912
- RB2011
...

Will let you know the results.

Hi John,

I started with trying on a Mikrotik RB2011 as these have a Serial port.

Ethernet does not function correctly.
I see it transmitting packets,  but receiving fails.



Logs from kernel 4.9:

[    1.052813] switch0: Atheros AR8327 rev. 4 switch registered on 
ag71xx-mdio.0

[    1.766947] libphy: ag71xx_mdio: probed
[    2.400863] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 
[uid=004dd034, driver=Atheros AR8216/AR8236/AR8316]

[    2.412536] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
[    3.041177] ag71xx-mdio.1: Found an AR934X built-in switch
[    3.093580] eth1: Atheros AG71xx at 0xba00, irq 5, mode:GMII
[    3.101108] NET: Registered protocol family 10



Logs from kernel 4.14

[    1.066007] switch0: Atheros AR8327 rev. 4 switch registered on 
ag71xx-mdio.0
[    1.073409] Atheros AR8216/AR8236/AR8316: probe of ag71xx-mdio.0:00 
failed with error -22

[    1.102455] libphy: ag71xx_mdio: probed
[    1.737938] ag71xx ag71xx.0: connected to PHY at ag71xx-mdio.0:00 
[uid=004dd034, driver=Generic PHY]

[    1.747994] eth0: Atheros AG71xx at 0xb900, irq 4, mode:RGMII
[    2.377642] ag71xx-mdio.1: Found an AR934X built-in switch
[    2.429938] eth1: Atheros AG71xx at 0xba00, irq 5, mode:GMII
[    2.437637] NET: Registered protocol family 10



This is currently the only culprit seen on this specific device.
I'll try to test some more ar71xx targets tomorrow ...


Thanks,

Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread John Crispin 



On 22/08/18 12:02, Outback Dingo wrote:

curious what i missed as my archer-c7-v4 is working fine. TPLINK
AC1750 device

Hi,
please dont top post
I have no idea what you are asking.
    John

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread Outback Dingo 
curious what i missed as my archer-c7-v4 is working fine. TPLINK
AC1750 device
On Wed, Aug 22, 2018 at 11:59 AM Daniel Golle  wrote:
>
> Hi John,
>
> On Wed, Aug 22, 2018 at 10:43:18AM +0200, John Crispin wrote:
> > On 22/08/18 10:40, Koen Vandeputte wrote:
> > > On 2018-08-13 17:14, John Crispin wrote:
> > > > Hi,
> > > >
> > > > as 19.01 will probably use v4.14 as baseline and ath79 wont be a
> > > > full replacement for ar71xx by then we decided to bump ar71xx to
> > > > v4.14. This is available for testing inside my staging tree ->
> > > >
> > > > https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
> > > >
> > > >
> > > > The tree also holds an update to mac80211, bumping it to the v4.18
> > > > wifi drivers. If anyone would like to test ar71xx and/or mac80211,
> > > > feel free to do so.
> > > ...
> > > Hi John,
> > >
> > > Thanks for the updates!
> > >
> > >
> > > mac80211:
> > >
> > >
> > > I've tested mac80211 on a large number of targets in the field:
> > >
> > > - ~25 devices
> > > - half of them having up to 4 wlan ifaces
> > > - all running in IBSS
> > > - Distance varying from 20m up to ~50km
> > >
> > > For my cases, it seems to work well.
>
> I rebased my staging tree on top of yours and tested a bunch of ramips
> devices and noticed that kmod-rt2x00-lib is marked @BROKEN. Removing
> the broken mark resulted in pretty well working rt2x00 drivers, tested
> on MT7620A and doing the job as well as with previous mac80211
> versions.
>
> Hence:
> diff --git a/package/kernel/mac80211/Makefile 
> b/package/kernel/mac80211/Makefile
> index 61cb27b5dd..70a268c535 100644
> --- a/package/kernel/mac80211/Makefile
> +++ b/package/kernel/mac80211/Makefile
> @@ -1238,7 +1238,7 @@ endef
>
>  define KernelPackage/rt2x00-lib
>  $(call KernelPackage/rt2x00/Default)
> -  DEPENDS+= @BROKEN @(PCI_SUPPORT||USB_SUPPORT||TARGET_ramips) 
> +kmod-mac80211 +kmod-lib-crc-itu-t
> +  DEPENDS+= @(PCI_SUPPORT||USB_SUPPORT||TARGET_ramips) +kmod-mac80211 
> +kmod-lib-crc-itu-t
>TITLE+= (LIB)
>FILES:=$(PKG_BUILD_DIR)/drivers/net/wireless/ralink/rt2x00/rt2x00lib.ko
>MENU:=1
>
>
> Cheers
>
>
> Daniel
>
> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread Daniel Golle 
Hi John,

On Wed, Aug 22, 2018 at 10:43:18AM +0200, John Crispin wrote:
> On 22/08/18 10:40, Koen Vandeputte wrote:
> > On 2018-08-13 17:14, John Crispin wrote:
> > > Hi,
> > > 
> > > as 19.01 will probably use v4.14 as baseline and ath79 wont be a
> > > full replacement for ar71xx by then we decided to bump ar71xx to
> > > v4.14. This is available for testing inside my staging tree ->
> > > 
> > > https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging
> > > 
> > > 
> > > The tree also holds an update to mac80211, bumping it to the v4.18
> > > wifi drivers. If anyone would like to test ar71xx and/or mac80211,
> > > feel free to do so.
> > ...
> > Hi John,
> > 
> > Thanks for the updates!
> > 
> > 
> > mac80211:
> > 
> > 
> > I've tested mac80211 on a large number of targets in the field:
> > 
> > - ~25 devices
> > - half of them having up to 4 wlan ifaces
> > - all running in IBSS
> > - Distance varying from 20m up to ~50km
> > 
> > For my cases, it seems to work well.

I rebased my staging tree on top of yours and tested a bunch of ramips
devices and noticed that kmod-rt2x00-lib is marked @BROKEN. Removing
the broken mark resulted in pretty well working rt2x00 drivers, tested
on MT7620A and doing the job as well as with previous mac80211
versions.

Hence:
diff --git a/package/kernel/mac80211/Makefile b/package/kernel/mac80211/Makefile
index 61cb27b5dd..70a268c535 100644
--- a/package/kernel/mac80211/Makefile
+++ b/package/kernel/mac80211/Makefile
@@ -1238,7 +1238,7 @@ endef
 
 define KernelPackage/rt2x00-lib
 $(call KernelPackage/rt2x00/Default)
-  DEPENDS+= @BROKEN @(PCI_SUPPORT||USB_SUPPORT||TARGET_ramips) +kmod-mac80211 
+kmod-lib-crc-itu-t
+  DEPENDS+= @(PCI_SUPPORT||USB_SUPPORT||TARGET_ramips) +kmod-mac80211 
+kmod-lib-crc-itu-t
   TITLE+= (LIB)
   FILES:=$(PKG_BUILD_DIR)/drivers/net/wireless/ralink/rt2x00/rt2x00lib.ko
   MENU:=1


Cheers


Daniel

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread John Crispin 



On 22/08/18 10:40, Koen Vandeputte wrote:



On 2018-08-13 17:14, John Crispin wrote:

Hi,

as 19.01 will probably use v4.14 as baseline and ath79 wont be a full 
replacement for ar71xx by then we decided to bump ar71xx to v4.14. 
This is available for testing inside my staging tree ->


https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging 



The tree also holds an update to mac80211, bumping it to the v4.18 
wifi drivers. If anyone would like to test ar71xx and/or mac80211, 
feel free to do so.


    John


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel



Hi John,

Thanks for the updates!


mac80211:


I've tested mac80211 on a large number of targets in the field:

- ~25 devices
- half of them having up to 4 wlan ifaces
- all running in IBSS
- Distance varying from 20m up to ~50km

For my cases, it seems to work well.


Small nit-pick:
In the bootlog, I'm seeing this:

[   11.33] Loading modules backported from Linux version 
v4.18-rc7-0-gacb1872577b3
[   11.33] Backport generated by backports.git 
v4.14-rc2-1-71-g39f4ec76



Is the 2nd line correct?


erm no, i'll rebase ontop of v4.18 (without RC) the next days and look 
into this. its most likely the last version that backports.git was 
tested against and i bumped the s/patch set to v4.18


thanks for testing !
    John







ar71xx:

I'm now currently testing your ar71xx 4.14 bump (comitted today in 
master) on a few different targets:


- Gl-mifi
- Rocket-m5
- RB912
- RB2011
...

Will let you know the results.


Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel



___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-22 Thread Koen Vandeputte 



On 2018-08-13 17:14, John Crispin wrote:

Hi,

as 19.01 will probably use v4.14 as baseline and ath79 wont be a full 
replacement for ar71xx by then we decided to bump ar71xx to v4.14. 
This is available for testing inside my staging tree ->


https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shortlog;h=refs/heads/staging 



The tree also holds an update to mac80211, bumping it to the v4.18 
wifi drivers. If anyone would like to test ar71xx and/or mac80211, 
feel free to do so.


    John


___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel



Hi John,

Thanks for the updates!


mac80211:


I've tested mac80211 on a large number of targets in the field:

- ~25 devices
- half of them having up to 4 wlan ifaces
- all running in IBSS
- Distance varying from 20m up to ~50km

For my cases, it seems to work well.


Small nit-pick:
In the bootlog, I'm seeing this:

[   11.33] Loading modules backported from Linux version 
v4.18-rc7-0-gacb1872577b3

[   11.33] Backport generated by backports.git v4.14-rc2-1-71-g39f4ec76


Is the 2nd line correct?



ar71xx:

I'm now currently testing your ar71xx 4.14 bump (comitted today in 
master) on a few different targets:


- Gl-mifi
- Rocket-m5
- RB912
- RB2011
...

Will let you know the results.


Koen

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-20 Thread Tom Psyborg 
Why dropping 4.9 support right away?!? I've already tried 4.14 without
mac80211 update and got approximately 20Mbps less throughput on wifi than
with 4.9. If the 4.18 mac80211 update doesn't bring in some of the
peformance lost, I wouldn't even consider 4.14 kernel as a baseline for
anything.

On 13 August 2018 at 17:14, John Crispin  wrote:

> Hi,
>
> as 19.01 will probably use v4.14 as baseline and ath79 wont be a full
> replacement for ar71xx by then we decided to bump ar71xx to v4.14. This is
> available for testing inside my staging tree ->
>
> https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=shor
> tlog;h=refs/heads/staging
>
> The tree also holds an update to mac80211, bumping it to the v4.18 wifi
> drivers. If anyone would like to test ar71xx and/or mac80211, feel free to
> do so.
>
> John
>
>
> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
>
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] RFT: ar71xx/mac80211 update

2018-08-15 Thread Sven Schönhoff 
Hi John,

I've tested your tree on my Archer C7 v4 but it bricked the device.
Only the power led is shining and I don't get a dhcp lease.

I've added the new "libubox: set RPATH for host build" commit to skip
the base-files error during build, added luci in make menuconfig and
unchecked "keep settings" for the flash.

Here is the serial log: https://pastebin.com/tcM2zTcm


Thank you and best regards
Sven

___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel