[LEDE-DEV] [PATCH 2/4] ipq: Don't force selection of the IPQ4019 firmware.

[Christian Lamparter]

(Readded: Hauke Mehrtens <hauke at hauke-m.de>)
On Saturday, January 20, 2018 5:40:08 PM CET Ben Greear wrote:
> 
> On 01/20/2018 02:15 AM, Christian Lamparter wrote:
> > On Friday, January 19, 2018 10:06:50 PM CET Ben Greear wrote:
> >> On 01/19/2018 01:03 PM, Christian Lamparter wrote:
> >>> On Friday, January 19, 2018 9:12:04 PM CET greearb at candelatech.com wrote:
> >>>> From: Ben Greear <greearb at candelatech.com>
> >>>>
> >>>> This will allow us to select the CT IPQ4019 firmware instead if
> >>>> desired.
> >>>>
> >>>> Signed-off-by: Ben Greear <greearb at candelatech.com>
> >>>> ---
> >>>>  package/firmware/ipq-wifi/Makefile | 2 +-
> >>>>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/package/firmware/ipq-wifi/Makefile b/package/firmware/ipq-wifi/Makefile
> >>>> index 519e8c9..6690248 100644
> >>>> --- a/package/firmware/ipq-wifi/Makefile
> >>>> +++ b/package/firmware/ipq-wifi/Makefile
> >>>> @@ -20,7 +20,7 @@ define Package/ipq-wifi-default
> >>>>    SUBMENU:=ath10k IPQ4019 Boarddata
> >>>>    SECTION:=firmware
> >>>>    CATEGORY:=Firmware
> >>>> -  DEPENDS:=@TARGET_ipq806x +ath10k-firmware-qca4019
> >>>> +  DEPENDS:=@TARGET_ipq806x
> >>>>    TITLE:=Custom Board
> >>>>  endef
> >>>>
> >>>
> >>> Wait! I remember talking about this here in the RFC thread:
> >>> <https://www.mail-archive.com/lede-dev@lists.infradead.org/msg09621.html>
> >>> |Hm, this would break the WIFI in the default configuration for the
> >>> |FritzBox 4040 image. Currently it only has a dependency on the
> >>> |ipq-wifi-fritz4040. (So it will end up without a firmware-5.bin)
> >>>
> >>> What gives? Are you trying to break the AVM FRITZ!Box 4040 image on purpose?
> >>
> >> Of course I'm not trying to break anything.  But, I am not sure how to
> >> fix this properly.
> > I remember writing about this too. It's in the reply.
> > <https://www.mail-archive.com/lede-dev@lists.infradead.org/msg09626.html>
> > |I think there's a another way to do this. But it will require to break with
> > |the existing convention of adding the board-2.bin that comes with the
> > |firmware repository to the ath10k-firmware-qca4019 file.
> > |
> > |This way, the custom board-2.bin will stay in place when you switch/update
> > |the firmware-5.bin.
> > |
> > |(The board-2.bin for the reference boards can simply be packaged just like
> > |one of the ipq-wifi board firmwares). And furhtermore, you could provide a
> > |"easy to use/install" custom ipq-wifi.ipk for the board-2.bin you currently
> > |host on your webside.
> > Of course, if you have a better idea let's hear it too. You could look into
> > making virtual packages (Don't know, if that's a thing with opkg, or not.
> > So watch out!) or go a different route. I'm sure there's plenty of ways to
> > do it. If you come up with something, I'll be happy to test it.
> >
> >> Does each platform need to specifically enable a firmware target instead of
> >> depending on a DEPENDS statement to make it work?
> > From what I know, the platform (ipq806x) does not add any firmware packages to
> > DEFAULT_PACKAGES in the target/linux/ipq806x/Makefile. It's all "per-device".
> >
> > (That said, you might want to talk to Sven Eckelmann too. As he has added
> > the ath10k-firmware-qca4019 package to the OpenMesh a42's DEVICE_PACKAGES.
> > So, if ath10k-ct is selected on a a42 it will also include the (now useless)
> > ath10k-firmware-qca4019, right?)
> >
> >> Is there some other way I can provide an option for two different firmware
> >> binaries?
> > The firmware binaries (i.e. firmware-X.bin) are not the problem. It's the
> > "board-2.bin" files that are shipped by the ath10k-firmware-qca4019/9984/..
> > packages.
> 
> Maybe you could test the 3 -v2 patches I posted on your 4019 device and see if
> you can select -ct firmware?
> 
> The board.bin issue would be the same with -ct or stock 4019 firmware AFAIK,
> so if that is still an issue, it is a separate one.

Done. It looks like the ath10k-ct & ct-firmware used to work but was broken by:
<https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=021b96d7c5c668fbcb5375c65cee90832bb2854f>
|rootfs: remove unnecessary and potentially harmful force flags from opkg call
|
|Especially --force-overwrite and --force-depends will often lead to broken
|images; it's better to fail the build in such cases than to silently ignore
|the errors.
|
|Instead, ignore errors in the per-device rootfs opkg remove command, so
|the build doesn't break when packages can't be removed because of
|dependencies.
|
|Signed-off-by: Matthias Schiffer <mschiffer at universe-factory.net>

Is this correct, if so this should be fixed one way or another.

Regards,
Christian