Hi Tomasz, Paul,
On 01.04.2020 14:03, Tomasz Maciej Nowak wrote:
W dniu 01.04.2020 o 08:55, Piotr Dymacz pisze:
Hi Paul,
On 01.04.2020 01:20, Paul Spooren wrote:
The PROFILE names of mvebu/cortexa9/Linksys devices are based on the
consumer names (like linksys_wrt1200ac) instead of the vendor codenames
(like linksys_caiman) which are however used in the rest of the build
system (plathform.sh, bootcount, 01_leds).
A running device is not able to know the profile used for its firmware
creation as `/tmp/sysinfo/board_name` points to the device tree
identifier (equal to Linksys codename) and `/tmp/sysinfo/model` is not
deterministically *convertible* to the PROFILE name.
The introduction of ALT titles (4ee3cf2b5a) allows to store consumer
names and make them available in the `menuconfig` so the profile name is
transparent to regular users.
This patch changes the mvebu/cortexa9/Linksys PROFILES to use the device
tree identifier aka Linksys codename as PROFILE name and uses the
consumer name as ALT title.
What about firmware filenames? Wouldn't this change confuse people searching
for firmware files by device model name?
Your change switch this from 'manufacturer_model' approach to
'manufacturer_manufacturer-codename'. Does Linksys ever mention that codenames
on their website, in marketing materials, etc.?
+1, not even on the device itself there's mention about these code names,
and that's why from me that's a NAK.
How about patching device's DTSes and include 'manufacturer,model' there
instead (in front of the existing ones)? Scripts in 'basic-files' would
also need to be fixed but this way we save this (in my opinion) misuse
of 'DEVICE_ALT*'.
I'm also not sure how many devices are affected here, only Linksys?
--
Cheers,
Piotr
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
