On 03.02.21 16:47, David Ahern wrote:
On 2/2/21 10:30 AM, Oliver Hartkopp wrote:
On 02.02.21 16:35, David Ahern wrote:
On 2/2/21 3:48 AM, Oliver Hartkopp wrote:
Are you sure this patch is correctly assigned to iproute2-next?
IMO it has to be applied to iproute2 as the functionality is already in
v5.11 which is in rc6 right now.
new features land in iproute2-next just as they do for the kernel with
net-next.
Patches adding support for kernel features should be sent in the same
development window if you want the iproute2 support to match kernel
version.
Oh, I followed the commits from iproute2 until the new include files
from (in this case) 5.11 pre rc1 had been updated (on 2020-12-24):
https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/commit/?id=2953235e61eb672bbdd2de84eb5b91c388f9a9b5
I thought the uapi updates in iproute2 are *always* pulled from the
kernel and not from iprout2-next which was new to me.
I sync kernel headers for iproute2-next with net-next, not linux-next.
Ok. Got it.
Do you expect patches for iproute2-next when the relevant changes become
available in linux-next then?
Even though I did not know about iproute2-next the patch is needed for
the 5.11 kernel (as written in the subject).
From a cursory look it appears CAN commits do not go through the netdev
tree yet the code is under net/can and the admin tool is through
iproute2 and netdevs. Why is that? If features patches flowed through
net-next, we would not have this problem.
CAN commits go through linux-can-next -> net-next. Same for linux-can ->
net.
The len8_dlc patches also went through linux-can-next -> net-next -> net
-> linux
iproute2 provides the configuration interface for CAN drivers under
driver/net/can only.
My only fault was, that I did not send the patch for iproute2-next at
the time when the len8_dlc patches were in net-next, right?
I was just not aware of iproute2-next.
The former patches I posted for iproute2 were always applied by Stephen
to the iproute2 tree directly.
Regards,
Oliver
