* Adrian Hunter <adrian.hunter@xxxxxxxxx> [220617 06:20]: > On 17/06/22 07:03, Tony Lindgren wrote: > > We need hardware enabled early in probe to detect capabilities, but must > > not call sdhci_runtime_resume_host() until sdhci_setup_host() has been > > called. Let's check for an initialized controller like we already do > > for context restore. > > Begs the question: why not prevent runtime pm until after sdhci_setup_host(). > Maybe expand the commit message explanation a bit? Sure will add some comments for why sdhci_omap_set_capabilities() is needed before sdhci_setup_host() and post v2. Thanks, Tony
