Message ID | 1303951472-22234-2-git-send-email-festevam@gmail.com |
---|---|
State | Superseded |
Headers | show |
Hello, ah, this patch makes my review comment to your patch 1 nil. On Wed, Apr 27, 2011 at 09:44:32PM -0300, Fabio Estevam wrote: > Instead of using gpio_request followed by gpio_direction_output use gpio_request_one. > > Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> > --- > arch/arm/mach-mx5/board-mx51_babbage.c | 8 ++++---- > 1 files changed, 4 insertions(+), 4 deletions(-) > > diff --git a/arch/arm/mach-mx5/board-mx51_babbage.c b/arch/arm/mach-mx5/board-mx51_babbage.c > index 7b919bc..00203fb 100644 > --- a/arch/arm/mach-mx5/board-mx51_babbage.c > +++ b/arch/arm/mach-mx5/board-mx51_babbage.c > @@ -179,25 +179,25 @@ static int gpio_usbh1_active(void) > > /* Set USBH1_STP to GPIO and toggle it */ > mxc_iomux_v3_setup_pad(usbh1stp_gpio); This cannot go into the global list of pins to mux? > - ret = gpio_request(BABBAGE_USBH1_STP, "usbh1_stp"); > + ret = gpio_request_one(BABBAGE_USBH1_STP, > + GPIOF_OUT_INIT_LOW, "usbh1_stp"); > > if (ret) { > pr_debug("failed to get MX51_PAD_USBH1_STP__GPIO_1_27: %d\n", ret); > return ret; > } > - gpio_direction_output(BABBAGE_USBH1_STP, 0); > gpio_set_value(BABBAGE_USBH1_STP, 1); > msleep(100); Why is the sleep between set_value(1) and free? I would have expected it between request_one(INIT_LOW) and set_value(1)? > gpio_free(BABBAGE_USBH1_STP); > > /* De-assert USB PHY RESETB */ > - ret = gpio_request(BABBAGE_USB_PHY_RESET, "phy_reset"); > + ret = gpio_request_one(BABBAGE_USB_PHY_RESET, > + GPIOF_OUT_INIT_HIGH, "phy_reset"); > > if (ret) { > pr_debug("failed to get MX51_PAD_EIM_D21__GPIO_2_5: %d\n", ret); > return ret; > } > - gpio_direction_output(BABBAGE_USB_PHY_RESET, 1); > return 0; > } > > -- > 1.6.0.4 > >
Hi Uwe, On 4/28/2011 4:55 AM, Uwe Kleine-König wrote: ... >> /* Set USBH1_STP to GPIO and toggle it */ >> mxc_iomux_v3_setup_pad(usbh1stp_gpio); > This cannot go into the global list of pins to mux? This pin alternates its mode(GPIO and USB STP), so I think we can leave it defined here. > >> - ret = gpio_request(BABBAGE_USBH1_STP, "usbh1_stp"); >> + ret = gpio_request_one(BABBAGE_USBH1_STP, >> + GPIOF_OUT_INIT_LOW, "usbh1_stp"); >> >> if (ret) { >> pr_debug("failed to get MX51_PAD_USBH1_STP__GPIO_1_27: %d\n", ret); >> return ret; >> } >> - gpio_direction_output(BABBAGE_USBH1_STP, 0); >> gpio_set_value(BABBAGE_USBH1_STP, 1); >> msleep(100); > Why is the sleep between set_value(1) and free? I would have expected it > between request_one(INIT_LOW) and set_value(1)? Agreed. I haven´t changed the logic on this patch. Will send a third patch for fixing this. Thanks, Fabio Estevam
diff --git a/arch/arm/mach-mx5/board-mx51_babbage.c b/arch/arm/mach-mx5/board-mx51_babbage.c index 7b919bc..00203fb 100644 --- a/arch/arm/mach-mx5/board-mx51_babbage.c +++ b/arch/arm/mach-mx5/board-mx51_babbage.c @@ -179,25 +179,25 @@ static int gpio_usbh1_active(void) /* Set USBH1_STP to GPIO and toggle it */ mxc_iomux_v3_setup_pad(usbh1stp_gpio); - ret = gpio_request(BABBAGE_USBH1_STP, "usbh1_stp"); + ret = gpio_request_one(BABBAGE_USBH1_STP, + GPIOF_OUT_INIT_LOW, "usbh1_stp"); if (ret) { pr_debug("failed to get MX51_PAD_USBH1_STP__GPIO_1_27: %d\n", ret); return ret; } - gpio_direction_output(BABBAGE_USBH1_STP, 0); gpio_set_value(BABBAGE_USBH1_STP, 1); msleep(100); gpio_free(BABBAGE_USBH1_STP); /* De-assert USB PHY RESETB */ - ret = gpio_request(BABBAGE_USB_PHY_RESET, "phy_reset"); + ret = gpio_request_one(BABBAGE_USB_PHY_RESET, + GPIOF_OUT_INIT_HIGH, "phy_reset"); if (ret) { pr_debug("failed to get MX51_PAD_EIM_D21__GPIO_2_5: %d\n", ret); return ret; } - gpio_direction_output(BABBAGE_USB_PHY_RESET, 1); return 0; }
Instead of using gpio_request followed by gpio_direction_output use gpio_request_one. Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> --- arch/arm/mach-mx5/board-mx51_babbage.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-)