Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
gpio-cdev: move kmod-leds-uleds dependency to MX100
The inclusion of the kmod-leds-uleds into the userspace nu801 package causes a circular dependency inside the buildsystem... which causes it to be picked regardless of other DEPENDS values. In case of the mx100, this could be solved by moving the kmod-leds-uled dependency to the kmod-meraki-mx100. Bonus: drop @!LINUX_5_4 from kmod-meraki-mx100 Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
- Loading branch information
eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seeing this error, rebased to 8822a8d
This package should not be attempting to build, as the target is
MIPS64
and notx86
Edit: I have had to revert to the 5.4 kernel due to the Octeon memleak issue. I suspect
@!LINUX_5_4
being removed exposed this, however,@TARGET_x86
should still keep it from building.eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@chunkeey
hi
this change would not change anything, it still cause the compile of nu801
I think we should drop depends on nu801 in
target/linux/x86/modules.mk
eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it is not reasonable that kmod depends on app, invert is good
eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@ptpt52 I think I know where the gpio-cdev is coming from. It's pulled in as a compile dependency for the kernel. With
this patch, it should no longer built on anything other than x86.
https://git.openwrt.org/?p=openwrt/staging/chunkeey.git;a=commit;h=9a137477811611f6363fb049b8356457798169a6
eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@chunkeey
but I still do not understand why the
kmod-meraki-mx100
need to depends on thenu801
kmod-meraki-mx100 build would failed if no
nu801
? or what?eeb8fd4
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
reason being that x86 has no per-device target device definitions for the supported devices. There are just the
generic squashfs/ext4- images that are supposed to work on every x86 device. So, the kmod-meraki-mx100 is
used as a stand-in-target for the MX100. I.e. once you select it (in the chef/imagebuilder), it pulls in the necessary
ethernet drivers (tg3) and status-led (nu801) for the mx100 to work.
As for "it is not reasonable that kmod depends on app, invert is good". Yes, this setup is an odd one.
Though the invert way also comes with its own problems:
#4846 (spidev-test pulls in kmod-spi-dev and this breaks SDK builts for spidev-test).
(nu801 also works for the ath79-ized Meraki MR18, though this is out-of-tree because
adding it as-is would cause other, already supported ath79 nand to break.)