From: Mark Kettenis Date: Wed, 17 Jun 2020 18:34:21 +0200 (CEST) Subject: [PATCH] regulator: fix: enable gpio when requested In-Reply-To: <20200616131835.GS24893@bill-the-cat> (message from Tom Rini on Tue, 16 Jun 2020 09:18:35 -0400) References: <20200427090947.75951-1-kettenis@openbsd.org> <20200616131835.GS24893@bill-the-cat> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de > Date: Tue, 16 Jun 2020 09:18:35 -0400 > From: Tom Rini > > On Mon, Apr 27, 2020 at 11:09:47AM +0200, Mark Kettenis wrote: > > > The fix in commit b7adcdd073c0 has the side-effect that the regulator > > will be disabled when requesting the relevant gpio in > > regulator_common_ofdata_to_platdata() and enabled in > > regulator_pre_probe() when the regulator was already enabled. > > This leads to a short interruption in the 3.3V power to the PCIe > > slot on the firefly-rk3399 which makes an ADATA SX8000NP NVMe SSD > > unhappy. > > > > Fix this by setting the GPIOD_IS_OUT_ACTIVE flag again when the > > 'regulator-boot-on' property is set, but check for this property > > explicitly instead of relying on the "boot_on" member of > > the uclass platdata. > > > > Signed-off-by: Mark Kettenis > > Tested-by: Patrice Chotard > > So, before this patch, my i.MX6D cuboxi platform with FEC ethernet does > not work (dhcp never gets an IP) nor does it work after this patch, and > I'm going to bisect things as it did work as of commit d8a3f5259a36 > ("Merge tag 'u-boot-imx-20200107' of > https://gitlab.denx.de/u-boot/custodians/u-boot-imx"). As Lukasz's > concern is that this will break things like FEC ethernet, I believe what > I see is enough datapoints to say that the area in question needs more > investigation. A follow-up patch from someone that tests multiple > platforms where we know we have seen oddities (firefly-rk3399, whatever > i.MX platform Lukasz was using, cuboxi) seems needed. With that said: > > Applied to u-boot/master now as it fixes some known issues and is > suspected but not confirmed to not break what b7adcdd073c0 was > addressing. Thaks Tom, It's good to have this fix in 2020.7 since we now have NVMe support on RK3399 and the bug fixed by this commit affected many NVMe drivers. Mark