From: Lucas Stach Date: Fri, 25 Jan 2013 23:19:21 +0100 Subject: [U-Boot] [PATCH 10/11] tegra20: remove old pinmux setup In-Reply-To: <5103035B.7090606@nvidia.com> References: <1359046100-19385-1-git-send-email-dev@lynxeye.de> <1359046100-19385-11-git-send-email-dev@lynxeye.de> <5103035B.7090606@nvidia.com> Message-ID: <1359152361.1540.97.camel@tellur> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Am Freitag, den 25.01.2013, 14:12 -0800 schrieb Stephen Warren: > On 01/24/2013 08:48 AM, Lucas Stach wrote: > > All boards are converted to the new tablebased pinmux setup. Get rid of > > the old method. > > > diff --git a/arch/arm/cpu/tegra-common/board.c b/arch/arm/cpu/tegra-common/board.c > > > @@ -145,7 +121,6 @@ static void setup_uarts(int uart_ids) > > if (uart_ids & (1 << i)) { > > enum periph_id id = id_for_uart[i]; > > > > - funcmux_select(id, uart_configs[i]); > > clock_ll_start_uart(id); > > } > > } > > Doesn't the debug UART get set up very early, in the SPL, before any > table-based pinmux could be activated? > > If so, I think we need to leave this one funcmux API call in place, so > that the debug UART always works nice and early. > > If not, how much does this series increase the binary of the SPL? > Ah right, I forgot about SPL debug. If we go for FDT based pinmux, we have to init UART in some explicit way, as DT and SPL don't mix. But even then I would like to get rid of the funcmux style and rather let the boards provide a minimal UART pinmux init table, as funcmux doesn't map too well onto the plethora of config options Tegra30 provides for the pinmux. Regards, Lucas