Message ID | b55b7bb8-97b6-4d25-9592-8213ecfe7b6c@app.fastmail.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | soc tree for 6.11 | expand |
On Mon, 15 Jul 2024 at 14:03, Arnd Bergmann <arnd@arndb.de> wrote: > > https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git tags/soc-dt-6.11 WTF, Arnd? This doesn't even *build*: > Jagadeesh Kona (8): > clk: qcom: camcc-sm8650: Add SM8650 camera clock controller driver because it has ret = qcom_cc_really_probe(&pdev->dev, &cam_cc_sm8650_desc, regmap); which is entirely wrong. And no, this is not some kind of merge mistake of mine. It was wrong in your tree. qcom_cc_really_probe() has always taken a struct platform_device *pdev as the first argument, not a "struct device *". So the "&pdev->dev" thing is completely wrong, and always has been. Looking around, this was also reported by the kernel test robot at https://lore.kernel.org/all/202407072331.baglL4Sd-lkp@intel.com/ https://lore.kernel.org/all/202407072212.WptVaUDt-lkp@intel.com/ and still I had to just notice it in my local build test. I did fix it up, but this is *not* ok. Linus
The pull request you sent on Mon, 15 Jul 2024 23:10:48 +0200:
> https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git tags/soc-dt-6.11
has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/e3950967f6e6b74a3606739ec50ed19f3398c7d8
Thank you!
On Tue, Jul 16, 2024, at 20:55, Linus Torvalds wrote: > On Mon, 15 Jul 2024 at 14:03, Arnd Bergmann <arnd@arndb.de> wrote: > And no, this is not some kind of merge mistake of mine. It was wrong > in your tree. > > qcom_cc_really_probe() has always taken a > > struct platform_device *pdev > > as the first argument, not a "struct device *". So the "&pdev->dev" > thing is completely wrong, and always has been. > > Looking around, this was also reported by the kernel test robot at > > https://lore.kernel.org/all/202407072331.baglL4Sd-lkp@intel.com/ > https://lore.kernel.org/all/202407072212.WptVaUDt-lkp@intel.com/ > > and still I had to just notice it in my local build test. > > I did fix it up, but this is *not* ok. I'm very sorry about this. I had missed the extra driver slipping in with the qcom DT patches at first and then failed to build test the dt branch standalone even after I realized it was there. I see that the interface change is in git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git#clk-next, which also contains the commit with the driver. This means you'll probably have to revert your fix once you merge that branch. Apparently the new driver was tested by the authors based on the other changes that are in the clk tree, but committed on a branch that didn't have them. Arnd