| Message ID | 1310579091-7424-1-git-send-email-swarren@nvidia.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
* Stephen Warren wrote: > DAP2 is used for a bunch of GPIOs. Not tri-stating the pins means audio > signals get sent out there, and this ends up resetting USB and breaking > SDHCI too. [...] > diff --git a/arch/arm/mach-tegra/board-trimslice-pinmux.c b/arch/arm/mach-tegra/board-trimslice-pinmux.c > index 82646a4..47c596c 100644 > --- a/arch/arm/mach-tegra/board-trimslice-pinmux.c > +++ b/arch/arm/mach-tegra/board-trimslice-pinmux.c > @@ -35,7 +35,7 @@ static __initdata struct tegra_pingroup_config trimslice_pinmux[] = { > {TEGRA_PINGROUP_CSUS, TEGRA_MUX_VI_SENSOR_CLK, TEGRA_PUPD_PULL_DOWN, TEGRA_TRI_TRISTATE}, > {TEGRA_PINGROUP_DAP1, TEGRA_MUX_DAP1, TEGRA_PUPD_NORMAL, TEGRA_TRI_NORMAL}, > {TEGRA_PINGROUP_DAP2, TEGRA_MUX_DAP2, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, > - {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_NORMAL}, > + {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, ^^^^ The patch subject and description say that DAP2 is tristated, not DAP3. Thierry
Thierry Reding wrote at Wednesday, July 13, 2011 12:01 PM: > Stephen Warren wrote: > > DAP2 is used for a bunch of GPIOs. Not tri-stating the pins means audio > > signals get sent out there, and this ends up resetting USB and breaking > > SDHCI too. ... > > - {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_NORMAL}, > > + {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, > > The patch subject and description say that DAP2 is tristated, not DAP3. Oops. I've fixed that locally and will repost once Mike acks it. Thanks!
diff --git a/arch/arm/mach-tegra/board-trimslice-pinmux.c b/arch/arm/mach-tegra/board-trimslice-pinmux.c index 82646a4..47c596c 100644 --- a/arch/arm/mach-tegra/board-trimslice-pinmux.c +++ b/arch/arm/mach-tegra/board-trimslice-pinmux.c @@ -35,7 +35,7 @@ static __initdata struct tegra_pingroup_config trimslice_pinmux[] = { {TEGRA_PINGROUP_CSUS, TEGRA_MUX_VI_SENSOR_CLK, TEGRA_PUPD_PULL_DOWN, TEGRA_TRI_TRISTATE}, {TEGRA_PINGROUP_DAP1, TEGRA_MUX_DAP1, TEGRA_PUPD_NORMAL, TEGRA_TRI_NORMAL}, {TEGRA_PINGROUP_DAP2, TEGRA_MUX_DAP2, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, - {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_NORMAL}, + {TEGRA_PINGROUP_DAP3, TEGRA_MUX_DAP3, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, {TEGRA_PINGROUP_DAP4, TEGRA_MUX_DAP4, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE}, {TEGRA_PINGROUP_DDC, TEGRA_MUX_I2C2, TEGRA_PUPD_PULL_UP, TEGRA_TRI_NORMAL}, {TEGRA_PINGROUP_DTA, TEGRA_MUX_VI, TEGRA_PUPD_NORMAL, TEGRA_TRI_TRISTATE},
DAP2 is used for a bunch of GPIOs. Not tri-stating the pins means audio signals get sent out there, and this ends up resetting USB and breaking SDHCI too. In the long run, the kernel should correctly configure all the relevant GPIO pins using tegra_gpio_enable, set each pin as an output, and set each pin's value. Such a change requires full schematic access, which I don't have. Hence, this change is good enough as a quick fix. Signed-off-by: Stephen Warren <swarren@nvidia.com> --- Mike, recall that when I previously tested your TrimSlice patches, I found a problem where playing audio would cause the USB bus to get reset. I recall having solved it with a pinmux change, and I believe this patch is that change. Could you check if it's still necessary and whether this change works? Thanks. arch/arm/mach-tegra/board-trimslice-pinmux.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)