ChanServ changed the topic of #armlinux to: ARM kernel talk [Upstream kernel, find your vendor forums for questions about their kernels] | https://libera.irclog.whitequark.org/armlinux
Bitweasil has quit [Read error: Connection reset by peer]
Bitweasil has joined #armlinux
Sally1 has joined #armlinux
apritzel has quit [Ping timeout: 246 seconds]
ARMed has quit [Quit: Connection closed for inactivity]
<krzk>
wens: I am not reading unrelated patches (there would be like 300 per day)
<krzk>
wens: what's the question? so I won't need to read entire thread?
elastic_dog has quit [Ping timeout: 246 seconds]
<wens>
MT6366 is somewhat compatbile to MT6358, but they are not subsets of one another; they do however have chip ID registers that can be read out
<wens>
Angelo was asking if we could just have mt6358 compatible as the fallback, and not add new compatible strings to the driver
broonie has joined #armlinux
<wens>
however since the regulator sets are different, i.e. the bindings need different compatible strings to match against, I assume this will not work?
Crofton has joined #armlinux
<wens>
simplified: MT6358 has regulators A, B, C ... , MT6366 has regulators A, B, D ...
<wens>
my understanding is we need separate compatible strings for them, without possibility of using fallbacks
<krzk>
If devices are different (in incompatible way), but offer detection mechanism (ID register), then we often use fallback/compatibility.
<wens>
but if the regulator node names are enumerated in the bindings, how would we write the bindings so that they don't give errors?
<wens>
I suppose if the sub-nodes are optional, then it could pass?
<krzk>
You have anyway device specific compatible right? So what is the difference in the bindings?
<krzk>
That part of bindings - regulator nodes, some allOf:if:then whatever - is anyway the same, maybe excep twhat you put to actual if: clause.
<wens>
the fallback, mt6358, would have several possible sub-nodes that shouldn't be possible to use with mt6366
<wens>
oh I see, I can reject the ones not usable with mt6366 in the if clause
<krzk>
wens: in any case you would have nodes only for one variant, what is the difference?
<krzk>
if you use separate compatibles, you also have sub-nodes for one variant and sub-nodes for other variant
<krzk>
Exactly the same problem
elastic_dog has joined #armlinux
<wens>
krzk: the problem is the set of possible sub-nodes under the regulator node is different; and with my regulator name cleanup, they become completely non-overlapping
<wens>
the original names have type prefixes and underscores that I want to get rid of
<krzk>
wens: This I get, but whether you use same fallback or not, you have exactly the same overlap and the same problem.
<krzk>
(or lack of overlap)
<wens>
so it's either separate binding files for different compatibles, or if/else clauses; the latter would be the same with or without fallbacks
<wens>
I believe that's what you're saying?
<krzk>
wens: no, I said you will have the same problem in both ways. We did not speak about separate binding files.
<krzk>
wens: so you have ONE binding with two different comaptibles and sub-nodes which differ between them, right? So using now fallbacj, changes nothing...
<wens>
ack
<krzk>
wens: if you have two separate bindings in the first place... using fallback - again, what changed? nothing
<krzk>
The only difference is in first case - one binding file - where the "if:" should strictly match compatible for mt6358 (so not using "contains:")
<krzk>
and in separate comaptibles this could be "contains" or not - doe snot matter because there is no fallback/common compatible.
hanetzer has quit [Ping timeout: 245 seconds]
<wens>
oh, using a strictly matching "if" would make it easier to write
<wens>
thanks for the help. my main question was whether it was doable.
<wens>
I guess I have some rewriting to do
hanetzer has joined #armlinux
frieder has joined #armlinux
<wens>
krzk: one other question, on the DT and bindings side, we'd still need device-specific compatible plus fallback for the PMIC/mfd ("mt6366", "mt6358"), and not just sharing the base model ("mt6358"), right?
sszy has joined #armlinux
ezulian has joined #armlinux
<pivi>
krzk and all: is there an easy way to run the DT checker only on affected dts files when modifying a dtsi?
<pivi>
krzk: my specific concern is to restrict the check on dts file that are affected by a specific change in the dtsi. I know how to run the checker for a specific yaml, for a specific dts. I am looking for a way to check all the dts files affected by a dtsi change
<pivi>
I guess that given the preprocessor is involved is not going to be trivial, but maybe something exist
<pivi>
s/the dtsi/a dtsi/
jluthra has joined #armlinux
heat has joined #armlinux
rockosov has quit [Ping timeout: 244 seconds]
rockosov has joined #armlinux
rgallaispou has joined #armlinux
luispm has quit [Ping timeout: 246 seconds]
headless has joined #armlinux
luispm has joined #armlinux
apritzel has joined #armlinux
bjoto has quit [*.net *.split]
Norkle has quit [*.net *.split]
lvrp16 has quit [*.net *.split]
MWelchUK has quit [*.net *.split]
_alice has quit [*.net *.split]
mkorpershoek has quit [*.net *.split]
benh_ has quit [*.net *.split]
HdkR has quit [*.net *.split]
houze_ has quit [*.net *.split]
Xogium has quit [*.net *.split]
roxell has quit [*.net *.split]
suihkulokki has quit [*.net *.split]
Amanieu has quit [*.net *.split]
Amanieu has joined #armlinux
bjoto has joined #armlinux
MWelchUK has joined #armlinux
Norkle has joined #armlinux
HdkR has joined #armlinux
lvrp16 has joined #armlinux
roxell has joined #armlinux
mkorpershoek has joined #armlinux
benh_ has joined #armlinux
Xogium has joined #armlinux
suihkulokki has joined #armlinux
_alice has joined #armlinux
houze_ has joined #armlinux
cbeznea_ has joined #armlinux
patersonc has joined #armlinux
headless has quit [Quit: Konversation terminated!]
<lag>
ukleinek: I was OOO last week and next - check again tomorrow
<ukleinek>
lag: great, thanks!
<lag>
ukleinek: NP
<NishanthMenon>
pivi: the trick i use is to restrict it to just the arch that is impacted by modifying the .config. I use a script such as https://gist.github.com/nmenon/73caa3eb7737a7f5185f040ca214645e -> then git clean -fdx arch Documentation/;((make W=1 LLVM=1 ARCH=arm64 -j`nproc` dtbs_check > /dev/null) 2>&1)|grep -v "^\s\s*"|sort -u>next.txt
<pivi>
NishanthMenon: thanks for sharing, that helps!
<NishanthMenon>
pivi: np
Amit_T has joined #armlinux
headless has joined #armlinux
Sally1 has quit [Read error: Connection reset by peer]
Sally2 has joined #armlinux
luispm has quit [Quit: Leaving]
luispm has joined #armlinux
luispm has quit [Quit: Leaving]
luispm has joined #armlinux
headless has quit [Quit: Konversation terminated!]