On Thu, Jun 07, 2018 at 04:33:07AM +0000, Yoshihiro Shimoda wrote: > Hi Geert-san, > > > From: Behalf Of Geert Uytterhoeven, Sent: Wednesday, June 6, 2018 5:58 PM > > > > Hi Simon, > > > > On Wed, Jun 6, 2018 at 10:52 AM, Simon Horman <[email protected]> wrote: > > > On Tue, Jun 05, 2018 at 05:05:15PM +0200, Geert Uytterhoeven wrote: > <snip> > > > This looks fine to me but I will wait to see if there are other reviews > > > before applying. > > > > Let's wait for feedback from Shimoda-san. Perpaps the quirk did rely on > > the buggy behavior. > > I checked /sys/kernel/debug/pm_genpd/pm_genpd_summary. > > < Before applies the patch > > - "3dg-b's" status is "off-0". > - "3dg-b" doesn't have any slaves. > > < After applied the patch > > - "3dg-b's" status is "on". > - "3dg-b" has "3dg-a" slave. > > So, > > Tested-by: Yoshihiro Shimoda <[email protected]>
On Thu, Jun 07, 2018 at 02:47:29PM +0200, Ulf Hansson wrote: > On 5 June 2018 at 17:05, Geert Uytterhoeven <[email protected]> wrote: > > The quirk for R-Car E3 ES1.0 added in commit 086b399965a7ee7e ("soc: > > renesas: r8a77990-sysc: Add workaround for 3DG-{A,B}") makes the 3DG-A > > PM domain a subdomain of the 3DG-B PM domain. However, registering > > 3DG-A with its parent fails silently, as the 3DG-B PM domain hasn't been > > registered yet, and such failures are never reported. > > > > Fix this by: > > 1. Splitting PM Domain initialization in two steps, so all PM domains > > are registered before any child-parent links are established, > > 2. Reporting any failures in establishing child-parent relations. > > > > Check for and report pm_genpd_init() failures, too, as that function > > gained a return value in commit 7eb231c337e00735 ("PM / Domains: Convert > > pm_genpd_init() to return an error code"). > > > > Fixes: 086b399965a7ee7e ("soc: renesas: r8a77990-sysc: Add workaround for > > 3DG-{A,B}") > > Signed-off-by: Geert Uytterhoeven <[email protected]> > > Reviewed-by: Ulf Hansson <[email protected]> Thanks everyone, applied as a fix for v4.18.
