Message ID | 405efb21a4600efad10413fcf4c72aacce180125.1538570983.git.baruch@tkos.co.il (mailing list archive) |
---|---|
State | New, archived |
Delegated to: | Bjorn Helgaas |
Headers | show |
Series | PCI: armada8k: add support for gpio controlled reset signal | expand |
On Wed, Oct 03, 2018 at 03:49:43PM +0300, Baruch Siach wrote: > This commit adds support for the gpio reset signal binding as described > in the designware-pcie.txt DT binding document. Both the documented > 'reset-gpio' property name, and the more standard 'reset-gpios' name are > supported. Hi Baruch I don't know this code at all, so maybe a dumb question. Why support the old none-standard binding of reset-gpio on new hardware? I'm assuming reset-gpio is marked a deprecated and reset-gpios is recommended? Andrew
Hi Andrew, On Wed, Oct 03, 2018 at 03:28:11PM +0200, Andrew Lunn wrote: > On Wed, Oct 03, 2018 at 03:49:43PM +0300, Baruch Siach wrote: > > This commit adds support for the gpio reset signal binding as described > > in the designware-pcie.txt DT binding document. Both the documented > > 'reset-gpio' property name, and the more standard 'reset-gpios' name are > > supported. > > I don't know this code at all, so maybe a dumb question. Why support > the old none-standard binding of reset-gpio on new hardware? I'm > assuming reset-gpio is marked a deprecated and reset-gpios is > recommended? This is all hidden behind the devm_gpiod_get_optional() call that only sees the "reset" string. The generic code supports both the new property name and also the older one for backward compatibility. This patch changes nothing in this regard. The designware-pcie.txt document mentions the older 'reset-gpio' name. So I mentioned that name as well in the commit log. baruch
On Wed, Oct 03, 2018 at 04:35:43PM +0300, Baruch Siach wrote: > Hi Andrew, > > On Wed, Oct 03, 2018 at 03:28:11PM +0200, Andrew Lunn wrote: > > On Wed, Oct 03, 2018 at 03:49:43PM +0300, Baruch Siach wrote: > > > This commit adds support for the gpio reset signal binding as described > > > in the designware-pcie.txt DT binding document. Both the documented > > > 'reset-gpio' property name, and the more standard 'reset-gpios' name are > > > supported. > > > > I don't know this code at all, so maybe a dumb question. Why support > > the old none-standard binding of reset-gpio on new hardware? I'm > > assuming reset-gpio is marked a deprecated and reset-gpios is > > recommended? > > This is all hidden behind the devm_gpiod_get_optional() call that only sees > the "reset" string. The generic code supports both the new property name and > also the older one for backward compatibility. This patch changes nothing in > this regard. > > The designware-pcie.txt document mentions the older 'reset-gpio' name. So I > mentioned that name as well in the commit log. I need Thomas' ACK to proceed with this patch. Thanks, Lorenzo
Hi Thomas, As the listed maintainer of this driver, can you review this patch? This patch makes the PCIe slot reset sequence independent from the bootloader. Thanks, baruch Lorenzo Pieralisi writes: > On Wed, Oct 03, 2018 at 04:35:43PM +0300, Baruch Siach wrote: >> On Wed, Oct 03, 2018 at 03:28:11PM +0200, Andrew Lunn wrote: >> > On Wed, Oct 03, 2018 at 03:49:43PM +0300, Baruch Siach wrote: >> > > This commit adds support for the gpio reset signal binding as described >> > > in the designware-pcie.txt DT binding document. Both the documented >> > > 'reset-gpio' property name, and the more standard 'reset-gpios' name are >> > > supported. >> > >> > I don't know this code at all, so maybe a dumb question. Why support >> > the old none-standard binding of reset-gpio on new hardware? I'm >> > assuming reset-gpio is marked a deprecated and reset-gpios is >> > recommended? >> >> This is all hidden behind the devm_gpiod_get_optional() call that only sees >> the "reset" string. The generic code supports both the new property name and >> also the older one for backward compatibility. This patch changes nothing in >> this regard. >> >> The designware-pcie.txt document mentions the older 'reset-gpio' name. So I >> mentioned that name as well in the commit log. > > I need Thomas' ACK to proceed with this patch. > > Thanks, > Lorenzo -- http://baruch.siach.name/blog/ ~. .~ Tk Open Systems =}------------------------------------------------ooO--U--Ooo------------{= - baruch@tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
Hello, On Wed, 21 Nov 2018 09:09:46 +0200, Baruch Siach wrote: > As the listed maintainer of this driver, can you review this patch? This > patch makes the PCIe slot reset sequence independent from the > bootloader. Yes, I'll have a look. Sorry for the delay. Thomas
Hello Baruch, Sorry for the delayed review. On Wed, 3 Oct 2018 15:49:43 +0300, Baruch Siach wrote: > #define PCIE_VENDOR_REGS_OFFSET 0x8000 > @@ -137,6 +139,12 @@ static int armada8k_pcie_host_init(struct pcie_port *pp) > struct dw_pcie *pci = to_dw_pcie_from_pp(pp); > struct armada8k_pcie *pcie = to_armada8k_pcie(pci); > > + if (pcie->reset_gpio) { This should be: if (!IS_ERR(pcie->reset_gpio)) Indeed, in the case of an error, pcie->reset_gpio will be non-NULL, with the error encoded as a ERR_PTR(). > + /* assert and then deassert the reset signal */ > + gpiod_set_value_cansleep(pcie->reset_gpio, 1); > + msleep(100); > + gpiod_set_value_cansleep(pcie->reset_gpio, 0); > + } > dw_pcie_setup_rc(pp); An empty new line here would be good before the dw_pcie_setup_rc() call. If you send a new iteration with those two issues fixed, you can directly add my Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Thanks! Thomas
Hello, On Thu, 22 Nov 2018 15:45:23 +0100, Thomas Petazzoni wrote: > This should be: > > if (!IS_ERR(pcie->reset_gpio)) > > Indeed, in the case of an error, pcie->reset_gpio will be non-NULL, > with the error encoded as a ERR_PTR(). Meh, scrap that. If pcie->reset_gpio was an error, probe() has failed. So by the time we are in armada8k_pcie_host_init(), pcie->reset_gpio is either NULL or a valid GPIO. So my comment was stupid. Since the newline thing is way too minor to require a new iteration: Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Thanks, and sorry for the noise. Thomas
On Wed, Oct 03, 2018 at 03:49:43PM +0300, Baruch Siach wrote: > This commit adds support for the gpio reset signal binding as described > in the designware-pcie.txt DT binding document. Both the documented > 'reset-gpio' property name, and the more standard 'reset-gpios' name are > supported. > > Signed-off-by: Baruch Siach <baruch@tkos.co.il> > --- > drivers/pci/controller/dwc/pcie-armada8k.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) Applied to pci/dwc for v4.21, thanks. Lorenzo > diff --git a/drivers/pci/controller/dwc/pcie-armada8k.c b/drivers/pci/controller/dwc/pcie-armada8k.c > index 0c389a30ef5d..b171b6bc15c8 100644 > --- a/drivers/pci/controller/dwc/pcie-armada8k.c > +++ b/drivers/pci/controller/dwc/pcie-armada8k.c > @@ -22,6 +22,7 @@ > #include <linux/resource.h> > #include <linux/of_pci.h> > #include <linux/of_irq.h> > +#include <linux/gpio/consumer.h> > > #include "pcie-designware.h" > > @@ -29,6 +30,7 @@ struct armada8k_pcie { > struct dw_pcie *pci; > struct clk *clk; > struct clk *clk_reg; > + struct gpio_desc *reset_gpio; > }; > > #define PCIE_VENDOR_REGS_OFFSET 0x8000 > @@ -137,6 +139,12 @@ static int armada8k_pcie_host_init(struct pcie_port *pp) > struct dw_pcie *pci = to_dw_pcie_from_pp(pp); > struct armada8k_pcie *pcie = to_armada8k_pcie(pci); > > + if (pcie->reset_gpio) { > + /* assert and then deassert the reset signal */ > + gpiod_set_value_cansleep(pcie->reset_gpio, 1); > + msleep(100); > + gpiod_set_value_cansleep(pcie->reset_gpio, 0); > + } > dw_pcie_setup_rc(pp); > armada8k_pcie_establish_link(pcie); > > @@ -249,6 +257,14 @@ static int armada8k_pcie_probe(struct platform_device *pdev) > goto fail_clkreg; > } > > + /* Get reset gpio signal and hold asserted (logically high) */ > + pcie->reset_gpio = devm_gpiod_get_optional(dev, "reset", > + GPIOD_OUT_HIGH); > + if (IS_ERR(pcie->reset_gpio)) { > + ret = PTR_ERR(pcie->reset_gpio); > + goto fail_clkreg; > + } > + > platform_set_drvdata(pdev, pcie); > > ret = armada8k_add_pcie_port(pcie, pdev); > -- > 2.19.0 >
diff --git a/drivers/pci/controller/dwc/pcie-armada8k.c b/drivers/pci/controller/dwc/pcie-armada8k.c index 0c389a30ef5d..b171b6bc15c8 100644 --- a/drivers/pci/controller/dwc/pcie-armada8k.c +++ b/drivers/pci/controller/dwc/pcie-armada8k.c @@ -22,6 +22,7 @@ #include <linux/resource.h> #include <linux/of_pci.h> #include <linux/of_irq.h> +#include <linux/gpio/consumer.h> #include "pcie-designware.h" @@ -29,6 +30,7 @@ struct armada8k_pcie { struct dw_pcie *pci; struct clk *clk; struct clk *clk_reg; + struct gpio_desc *reset_gpio; }; #define PCIE_VENDOR_REGS_OFFSET 0x8000 @@ -137,6 +139,12 @@ static int armada8k_pcie_host_init(struct pcie_port *pp) struct dw_pcie *pci = to_dw_pcie_from_pp(pp); struct armada8k_pcie *pcie = to_armada8k_pcie(pci); + if (pcie->reset_gpio) { + /* assert and then deassert the reset signal */ + gpiod_set_value_cansleep(pcie->reset_gpio, 1); + msleep(100); + gpiod_set_value_cansleep(pcie->reset_gpio, 0); + } dw_pcie_setup_rc(pp); armada8k_pcie_establish_link(pcie); @@ -249,6 +257,14 @@ static int armada8k_pcie_probe(struct platform_device *pdev) goto fail_clkreg; } + /* Get reset gpio signal and hold asserted (logically high) */ + pcie->reset_gpio = devm_gpiod_get_optional(dev, "reset", + GPIOD_OUT_HIGH); + if (IS_ERR(pcie->reset_gpio)) { + ret = PTR_ERR(pcie->reset_gpio); + goto fail_clkreg; + } + platform_set_drvdata(pdev, pcie); ret = armada8k_add_pcie_port(pcie, pdev);
This commit adds support for the gpio reset signal binding as described in the designware-pcie.txt DT binding document. Both the documented 'reset-gpio' property name, and the more standard 'reset-gpios' name are supported. Signed-off-by: Baruch Siach <baruch@tkos.co.il> --- drivers/pci/controller/dwc/pcie-armada8k.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)