On Tue, Feb 13, 2024 at 4:30 AM Dmitry Baryshkov
<dmitry.baryshkov@xxxxxxxxxx> wrote:
>
> On Tue, 13 Feb 2024 at 00:22, Rob Herring <robh@xxxxxxxxxx> wrote:
> >
> > On Sun, Jan 21, 2024 at 05:57:42PM +0100, David Wronek wrote:
> > > Document the compatible for the UFS found on SC7180.
> > >
> > > Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
> > > Signed-off-by: David Wronek <davidwronek@xxxxxxxxx>
> > > ---
> > > Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 2 ++
> > > 1 file changed, 2 insertions(+)
> >
> > Should have been picked up by SCSI/UFS maintainers, but it
> > hasn't, so I applied it.
>
> And it now triggers schema warnings, because sc7180-ufshc has 7 clocks
> and 1 reg entries.
And now dropped... Perhaps the dts changes should be too.
Maybe QCom maintainers should require a report of dtbs_check on new
boards. My comparisons of Linus vs. next warnings often show an
increase in QCom warnings. Like right now:
linus: arch/arm64/boot/dts/qcom:1990:265
next: arch/arm64/boot/dts/qcom:1610:298
First number is total warnings. Second number is unique warnings
(stripping dtb name). Some of this is just mismatch between schemas
and dts changes showing up in next, but it doesn't tend to go to 0 as
the merge window approaches. I've seen this several cycles. All the
data is available from my CI jobs, and I regularly look at the diff
with this:
$ less ~/bin/gl-diff-dtb-warnings
#!/bin/sh
[ -z "$1" ] && { echo "Missing arch!"; exit 1; }
arch="$1"
job="job-dtbs-check"
logfile="platform-warnings.log"
# url <branch> <arch>
url() {
local branch="$1"
local arch="$2"
echo "https://gitlab.com/robherring/linux-dt/-/jobs/artifacts/${branch}/raw/${logfile}?job=${job}%3A+%5B${arch}%5D";
}
curl -Ls -o orig.log $(url linus ${arch})
curl -Ls -o next.log $(url next ${arch})
meld orig.log next.log
