On 2/21/24 14:13, Andy Shevchenko wrote: > On Wed, Feb 21, 2024 at 12:01:43PM +0100, Thomas Richard wrote: >> On 2/16/24 16:08, Andy Shevchenko wrote: >>> On Fri, Feb 16, 2024 at 08:59:47AM +0100, Thomas Richard wrote: >>>> On 2/15/24 16:27, Andy Shevchenko wrote: >>>>> On Thu, Feb 15, 2024 at 04:17:47PM +0100, Thomas Richard wrote: >>>>>> No need to check the pointer returned by platform_get_drvdata(), as >>>>>> platform_set_drvdata() is called during the probe. >>>>> >>>>> This patch should go _after_ the next one, otherwise the commit message doesn't >>>>> tell full story and the code change bring a potential regression. >>>> >>>> Hello Andy, >>>> >>>> I'm ok to move this patch after the next one. >>>> But for my understanding, could you explain me why changing the order is >>>> important in this case ? >>> >>> Old PM calls obviously can be called in different circumstances and these >>> checks are important. >>> >>> Just squash these two patches to avoid additional churn and we are done. >> >> You mean invert the order instead of squash. > > Either would work, but see how much churn in terms of changing just changed > lines it adds. OK thanks. I'll squash the two patches. And I'll add a comment which explains that I dropped some dead code. Regards, -- Thomas Richard, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
