03:51 <JIaxyga[m]> <konradybcio> "JIaxyga: what's the downstream..." <- yupik

07:13 <Daanct12> can anyone here look at: https://patchwork.kernel.org/project/linux-arm-msm/list/?series=872860 ?

09:31 <lumag> aka_[m], please drop the dev_err and pname. Use devm_ioremap_resource().

09:32 <lumag> Daanct12, either konradybcio or I will take a look. It was posted just yesterday

09:35 <aka_[m]> krzk: "Don't duplicate. It's already required." Is only about model and compatible?

09:35 <aka_[m]> Thats what was in 8916 -if block

09:35 <aka_[m]> and i somehow cannot find any example where there are multiple -ifs and then else

09:36 <aka_[m]> if i drop else at the end and add new -if 8916 example fail to validate

09:37 <krzk> aka_[m]: yeah, only about model and compat. I saw later that it's in other block, but anyway - ifi you rebase on my patch, you will not see them there.

09:37 <aka_[m]> aaa i cannot even explain what i mean

09:37 <aka_[m]> krzk: You had patch on this doc?

09:37 <krzk> aka_[m]: there are no multiple if+if+else, instead we do if+if+if covering all possible cases

09:37 <krzk> aka_[m]: yeah, I cced you

09:37 <aka_[m]> seems my next is few days too old then

09:38 <krzk> aka_[m]: no, I meant, I sent it...

09:38 <aka_[m]> just opened thunderbird now i see it @10:33

09:41 <krzk> aka_[m]: so if you rebase on top, then - top-level reg and reg-names should grow to three items, and you should add "minItems: 2" to each. Then you add if:then, with maxItems: 2, and else: with minItems: 3 (depemnding on variant/compatible_

09:41 <aka_[m]> that sounds better

09:45 <aka_[m]> lumag:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/LmkovftSnNMXFsDUKDOlfkbl>)

09:46 <lumag> aka_[m], ?

09:47 <aka_[m]> please drop the dev_err and pname. Use devm_ioremap_resource().

09:47 <lumag> aka_[m], yes. Users of 8916 shouldn't see the error, so drop the dev_err.

09:48 <aka_[m]> so no erroring or returning anywhere?

09:48 <lumag> also you are calling devm_platform_ioremap_resource_byname(), but you already have the resource, no need to get it again

09:48 <aka_[m]> not like i understand much lol

09:48 <aka_[m]> ah yes

09:48 <aka_[m]> i grabbed resource by name

09:49 <lumag> yes, no need to error, if it's not expected to be present

09:50 <aka_[m]> fast rebasing patch

09:51 <aka_[m]> https://github.com/Kiciuk/linux/commit/28b4d7b4e9e2c3cba3e5f0073822c402c5a8e577

09:51 <aka_[m]> should be it like that?

09:52 <aka_[m]> wait

09:52 <aka_[m]> error handling i guess

09:52 <aka_[m]> what if res is empty?

09:54 <aka_[m]> uh i guess devm_ioremap_resource checks res before?

09:57 <lumag> aka_[m], it doesn't give you anything, devm_ioremap_resource would still print an error internally.

09:58 <aka_[m]> okay

09:58 <lumag> if (!res || resource_type(res) != IORESOURCE_MEM) {

09:58 <lumag> ret = dev_err_probe(dev, -EINVAL, "invalid resource %pR\n", res);

09:58 <lumag> So get resource and if it's not NULL, call devm_ioremap_resource

09:59 <aka_[m]> now onto trying to compile with clang...

09:59 <aka_[m]> it kinda drives me nuts to see kernel test robot throw entire stacks and gcc is silent af

10:02 <aka_[m]> lumag: so by default iomem pointer is initialized as Null?

10:03 <lumag> aka_[m], I guess so. Does memory allocation use something like kzalloc?

10:22 <aka_[m]> make -j16 LLVM=1 ARCH=arm64 seems to work

10:23 <aka_[m]> tho not sure if i need W=1

10:30 <aka_[m]> seems clang catches stuff but errors are bit different

10:31 <aka_[m]> very close tho