From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alban Bedel Date: Wed, 30 Oct 2013 12:19:32 +0100 Subject: [U-Boot] =?utf-8?q?=5BPATCH_v3=5D_ARM=3A_tegra=3A_Add_the_Tamonte?= =?utf-8?q?n=E2=84=A2_NG_Evaluation_Carrier_board?= In-Reply-To: <526EDE66.2060402@wwwdotorg.org> References: <5260393A.4000609@wwwdotorg.org> <1382365726-29140-1-git-send-email-alban.bedel@avionic-design.de> <526EDE66.2060402@wwwdotorg.org> Message-ID: <20131030121932.645d0190@avionic-0020> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Mon, 28 Oct 2013 16:00:06 -0600 Stephen Warren wrote: > On 10/21/2013 08:28 AM, Alban Bedel wrote: > > Add support for the new Tamonten? NG platform from Avionic Design. > > Currently only I2C, MMC, USB and ethernet have been tested. > > What changed in v3? There's no changelog here, so I don't know where to > concentrate my re-review of this patch... I'll add one. It was unclear to me how these should be done, and it took me quiet some times to find out about this 3 dashes hidden feature. > > diff --git a/board/avionic-design/common/tamonten-ng.c b/board/avionic-design/common/tamonten-ng.c > > > +#define MAX_I2C_RETRY 3 > > I don't think that's used any more. I'll remove it. > Most of the PMU #defines aren't used either. I'll remove them too, althought I hate having hardcoded values in the code instead of a human readable definition of the registers. > > +#ifdef CONFIG_BOARD_EARLY_INIT_F > > I don't think any of these ifdefs are required, since the config file > hard-codes all those values. Same for #if defined(CONFIG_TEGRA_MMC) > below, and perhaps more? Will do. > > +void gpio_early_init(void) > > +{ > > + /* Turn on the alive signal */ > > + gpio_request(GPIO_PV2, "ALIVE"); > > + gpio_direction_output(GPIO_PV2, 1); > > + > > + /* Remove the reset on the reset on the external periph */ > > "reset on the reset on the" ? > > > diff --git a/board/avionic-design/dts/tegra30-tamonten.dtsi b/board/avionic-design/dts/tegra30-tamonten.dtsi > > > + i2c at 7000c000 { > > + status = "okay"; > > + clock-frequency = <100000>; > > + }; > > + > > + i2c at 7000c400 { > > + status = "okay"; > > + clock-frequency = <100000>; > > + }; > ... > > Do all the carrier boards guarantee that all those I2C and SPI, and even > SDHCI busses are routed somewhere? It may be best to defer adding > status="okay" to the leaf board files, so you know there's actually > something hooked up there. Most of these are on the COM module, but you are right, a few of them shouldn't have status="okay" in this dtsi. Alban