This series add genpd support for the Tegra xHCI device that requires
more than one power-domain to operate.
This series is making changes to more than one subsystem and at first
glance may look like a maintainers nightmare. However, my proposal is
this, once reviewed and people are happy ...
1. Patches 1-3 should be merged first. Note there is a dependency
between patches 1 and 3 such that 3 can only be merged once 1 has
been accepted.
2. Once patches 1-3 are accepted, then 4 and 5 can be merged via the
Tegra tree.
Jon Hunter (5):
PM / Domains: Export symbol for genpd_dev_pm_attach_by_name
dt-bindings: usb: xhci-tegra: Add power-domain details
usb: xhci: tegra: Add genpd support
soc/tegra: pmc: Don't power-up XUSB power-domains
arm64: dts: tegra210: Add power-domains for xHCI
.../bindings/usb/nvidia,tegra124-xusb.txt | 8 ++
arch/arm64/boot/dts/nvidia/tegra210.dtsi | 2 +
drivers/base/power/domain.c | 1 +
drivers/soc/tegra/pmc.c | 16 ----
drivers/usb/host/xhci-tegra.c | 88 +++++++++++++++++++---
5 files changed, 88 insertions(+), 27 deletions(-)
--
2.7.4
Now that the Tegra xHCI driver manages the XUSB power-domains itself,
remove the code to power-up the power-domains used by the xHCI device
from the PMC driver on boot.
Signed-off-by: Jon Hunter <[email protected]>
---
drivers/soc/tegra/pmc.c | 16 ----------------
1 file changed, 16 deletions(-)
diff --git a/drivers/soc/tegra/pmc.c b/drivers/soc/tegra/pmc.c
index ab719fa90150..a68b4476b4ee 100644
--- a/drivers/soc/tegra/pmc.c
+++ b/drivers/soc/tegra/pmc.c
@@ -847,22 +847,6 @@ static void tegra_powergate_add(struct tegra_pmc *pmc, struct device_node *np)
goto remove_resets;
}
- /*
- * FIXME: If XHCI is enabled for Tegra, then power-up the XUSB
- * host and super-speed partitions. Once the XHCI driver
- * manages the partitions itself this code can be removed. Note
- * that we don't register these partitions with the genpd core
- * to avoid it from powering down the partitions as they appear
- * to be unused.
- */
- if (IS_ENABLED(CONFIG_USB_XHCI_TEGRA) &&
- (id == TEGRA_POWERGATE_XUSBA || id == TEGRA_POWERGATE_XUSBC)) {
- if (off)
- WARN_ON(tegra_powergate_power_up(pg, true));
-
- goto remove_resets;
- }
-
err = pm_genpd_init(&pg->genpd, NULL, off);
if (err < 0) {
pr_err("failed to initialise PM domain %s: %d\n", np->name,
--
2.7.4
Populate the power-domain properties for the xHCI device for Tegra210.
Signed-off-by: Jon Hunter <[email protected]>
---
arch/arm64/boot/dts/nvidia/tegra210.dtsi | 2 ++
1 file changed, 2 insertions(+)
diff --git a/arch/arm64/boot/dts/nvidia/tegra210.dtsi b/arch/arm64/boot/dts/nvidia/tegra210.dtsi
index 8fe47d6445a5..2205d66b0443 100644
--- a/arch/arm64/boot/dts/nvidia/tegra210.dtsi
+++ b/arch/arm64/boot/dts/nvidia/tegra210.dtsi
@@ -879,6 +879,8 @@
resets = <&tegra_car 89>, <&tegra_car 156>,
<&tegra_car 143>;
reset-names = "xusb_host", "xusb_ss", "xusb_src";
+ power-domains = <&pd_xusbhost>, <&pd_xusbss>;
+ power-domain-names = "xusb_host", "xusb_ss";
nvidia,xusb-padctl = <&padctl>;
--
2.7.4
Add details for power-domains to the Tegra xHCI bindings so that
generic power-domains can be used for inconjunction with the xHCI
driver.
Signed-off-by: Jon Hunter <[email protected]>
---
Documentation/devicetree/bindings/usb/nvidia,tegra124-xusb.txt | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra124-xusb.txt b/Documentation/devicetree/bindings/usb/nvidia,tegra124-xusb.txt
index 3eee9e505400..4156c3e181c5 100644
--- a/Documentation/devicetree/bindings/usb/nvidia,tegra124-xusb.txt
+++ b/Documentation/devicetree/bindings/usb/nvidia,tegra124-xusb.txt
@@ -59,6 +59,14 @@ For Tegra210:
- avdd-pll-uerefe-supply: PLLE reference PLL power supply. Must supply 1.05 V.
- dvdd-pex-pll-supply: PCIe/USB3 PLL power supply. Must supply 1.05 V.
- hvdd-pex-pll-e-supply: High-voltage PLLE power supply. Must supply 1.8 V.
+- power-domains: A list of PM domain specifiers that reference each power-domain
+ used by the xHCI controller. This list must comprise of a specifier for the
+ XUSBA and XUSBC power-domains. See ../power/power_domain.txt and
+ ../arm/tegra/nvidia,tegra20-pmc.txt for details.
+- power-domain-names: A list of names that represent each of the specifiers in
+ the 'power-domains' property. Must include 'xusb_ss' and 'xusb_host' which
+ represent the power-domains XUSBA and XUSBC, respectively. See
+ ../power/power_domain.txt for details.
Optional properties:
--------------------
--
2.7.4
The generic power-domain framework has been updated to allow devices
that require more than one power-domain to create a new device for
each power-domain required and then link these new power-domain
devices to the consumer device.
Update the Tegra xHCI driver to use the new APIs provided by the
generic power-domain framework so we can use the generic power-domain
framework for managing the xHCI controllers power-domains. Please
note that to maintain backward compatibility with older device-tree
blobs these new generic power-domain APIs are only used if the
'power-domains' property is present and otherwise we fall back to
using the legacy Tegra APIs for managing the power-domains.
Signed-off-by: Jon Hunter <[email protected]>
---
drivers/usb/host/xhci-tegra.c | 88 +++++++++++++++++++++++++++++++++++++------
1 file changed, 77 insertions(+), 11 deletions(-)
diff --git a/drivers/usb/host/xhci-tegra.c b/drivers/usb/host/xhci-tegra.c
index 4b463e5202a4..335d9d7e515e 100644
--- a/drivers/usb/host/xhci-tegra.c
+++ b/drivers/usb/host/xhci-tegra.c
@@ -18,6 +18,7 @@
#include <linux/phy/tegra/xusb.h>
#include <linux/platform_device.h>
#include <linux/pm.h>
+#include <linux/pm_domain.h>
#include <linux/pm_runtime.h>
#include <linux/regulator/consumer.h>
#include <linux/reset.h>
@@ -194,6 +195,11 @@ struct tegra_xusb {
struct reset_control *host_rst;
struct reset_control *ss_rst;
+ struct device *genpd_dev_host;
+ struct device *genpd_dev_ss;
+ struct device_link *genpd_dl_host;
+ struct device_link *genpd_dl_ss;
+
struct phy **phys;
unsigned int num_phys;
@@ -928,6 +934,60 @@ static int tegra_xusb_load_firmware(struct tegra_xusb *tegra)
return 0;
}
+static void tegra_xusb_powerdomain_remove(struct device *dev,
+ struct tegra_xusb *tegra)
+{
+ if (tegra->genpd_dl_ss)
+ device_link_del(tegra->genpd_dl_ss);
+ if (tegra->genpd_dl_host)
+ device_link_del(tegra->genpd_dl_host);
+ if (tegra->genpd_dev_ss)
+ dev_pm_domain_detach(tegra->genpd_dev_ss, true);
+ if (tegra->genpd_dev_host)
+ dev_pm_domain_detach(tegra->genpd_dev_host, true);
+}
+
+static int tegra_xusb_powerdomain_init(struct device *dev,
+ struct tegra_xusb *tegra)
+{
+ int err;
+
+ tegra->genpd_dev_host = genpd_dev_pm_attach_by_name(dev, "xusb_host");
+ if (IS_ERR(tegra->genpd_dev_host)) {
+ err = PTR_ERR(tegra->genpd_dev_host);
+ dev_err(dev, "failed to get host pm-domain: %d\n", err);
+ return err;
+ }
+
+ tegra->genpd_dev_ss = genpd_dev_pm_attach_by_name(dev, "xusb_ss");
+ if (IS_ERR(tegra->genpd_dev_ss)) {
+ tegra_xusb_powerdomain_remove(dev, tegra);
+ err = PTR_ERR(tegra->genpd_dev_ss);
+ dev_err(dev, "failed to get superspeed pm-domain: %d\n", err);
+ return err;
+ }
+
+ tegra->genpd_dl_host = device_link_add(dev, tegra->genpd_dev_host,
+ DL_FLAG_PM_RUNTIME |
+ DL_FLAG_STATELESS);
+ if (!tegra->genpd_dl_host) {
+ tegra_xusb_powerdomain_remove(dev, tegra);
+ dev_err(dev, "adding host device link failed!\n");
+ return -ENODEV;
+ }
+
+ tegra->genpd_dl_ss = device_link_add(dev, tegra->genpd_dev_ss,
+ DL_FLAG_PM_RUNTIME |
+ DL_FLAG_STATELESS);
+ if (!tegra->genpd_dl_ss) {
+ tegra_xusb_powerdomain_remove(dev, tegra);
+ dev_err(dev, "adding superspeed device link failed!\n");
+ return -ENODEV;
+ }
+
+ return 0;
+}
+
static int tegra_xusb_probe(struct platform_device *pdev)
{
struct tegra_xusb_mbox_msg msg;
@@ -1038,7 +1098,11 @@ static int tegra_xusb_probe(struct platform_device *pdev)
goto put_padctl;
}
- if (!pdev->dev.pm_domain) {
+ if (of_property_read_bool(pdev->dev.of_node, "power-domains")) {
+ err = tegra_xusb_powerdomain_init(&pdev->dev, tegra);
+ if (err)
+ goto put_padctl;
+ } else {
tegra->host_rst = devm_reset_control_get(&pdev->dev,
"xusb_host");
if (IS_ERR(tegra->host_rst)) {
@@ -1069,9 +1133,10 @@ static int tegra_xusb_probe(struct platform_device *pdev)
tegra->host_clk,
tegra->host_rst);
if (err) {
+ tegra_powergate_power_off(TEGRA_POWERGATE_XUSBA);
dev_err(&pdev->dev,
"failed to enable XUSBC domain: %d\n", err);
- goto disable_xusba;
+ goto put_padctl;
}
}
@@ -1079,7 +1144,7 @@ static int tegra_xusb_probe(struct platform_device *pdev)
sizeof(*tegra->supplies), GFP_KERNEL);
if (!tegra->supplies) {
err = -ENOMEM;
- goto disable_xusbc;
+ goto put_powerdomains;
}
for (i = 0; i < tegra->soc->num_supplies; i++)
@@ -1089,7 +1154,7 @@ static int tegra_xusb_probe(struct platform_device *pdev)
tegra->supplies);
if (err) {
dev_err(&pdev->dev, "failed to get regulators: %d\n", err);
- goto disable_xusbc;
+ goto put_powerdomains;
}
for (i = 0; i < tegra->soc->num_types; i++)
@@ -1099,7 +1164,7 @@ static int tegra_xusb_probe(struct platform_device *pdev)
sizeof(*tegra->phys), GFP_KERNEL);
if (!tegra->phys) {
err = -ENOMEM;
- goto disable_xusbc;
+ goto put_powerdomains;
}
for (i = 0, k = 0; i < tegra->soc->num_types; i++) {
@@ -1115,7 +1180,7 @@ static int tegra_xusb_probe(struct platform_device *pdev)
"failed to get PHY %s: %ld\n", prop,
PTR_ERR(phy));
err = PTR_ERR(phy);
- goto disable_xusbc;
+ goto put_powerdomains;
}
tegra->phys[k++] = phy;
@@ -1126,7 +1191,7 @@ static int tegra_xusb_probe(struct platform_device *pdev)
dev_name(&pdev->dev));
if (!tegra->hcd) {
err = -ENOMEM;
- goto disable_xusbc;
+ goto put_powerdomains;
}
/*
@@ -1222,12 +1287,13 @@ static int tegra_xusb_probe(struct platform_device *pdev)
disable_rpm:
pm_runtime_disable(&pdev->dev);
usb_put_hcd(tegra->hcd);
-disable_xusbc:
- if (!pdev->dev.pm_domain)
+put_powerdomains:
+ if (of_property_read_bool(pdev->dev.of_node, "power-domains")) {
+ tegra_xusb_powerdomain_remove(&pdev->dev, tegra);
+ } else {
tegra_powergate_power_off(TEGRA_POWERGATE_XUSBC);
-disable_xusba:
- if (!pdev->dev.pm_domain)
tegra_powergate_power_off(TEGRA_POWERGATE_XUSBA);
+ }
put_padctl:
tegra_xusb_padctl_put(tegra->padctl);
return err;
--
2.7.4
Export the symbol for genpd_dev_pm_attach_by_name() so that drivers
built as modules can call this function.
Signed-off-by: Jon Hunter <[email protected]>
---
drivers/base/power/domain.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
index 4b5714199490..b2bbe0d848d2 100644
--- a/drivers/base/power/domain.c
+++ b/drivers/base/power/domain.c
@@ -2397,6 +2397,7 @@ struct device *genpd_dev_pm_attach_by_name(struct device *dev, char *name)
return genpd_dev_pm_attach_by_id(dev, index);
}
+EXPORT_SYMBOL_GPL(genpd_dev_pm_attach_by_name);
static const struct of_device_id idle_state_match[] = {
{ .compatible = "domain-idle-state", },
--
2.7.4
On 26 September 2018 at 14:27, Jon Hunter <[email protected]> wrote:
> Export the symbol for genpd_dev_pm_attach_by_name() so that drivers
> built as modules can call this function.
>
> Signed-off-by: Jon Hunter <[email protected]>
> ---
> drivers/base/power/domain.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
> index 4b5714199490..b2bbe0d848d2 100644
> --- a/drivers/base/power/domain.c
> +++ b/drivers/base/power/domain.c
> @@ -2397,6 +2397,7 @@ struct device *genpd_dev_pm_attach_by_name(struct device *dev, char *name)
>
> return genpd_dev_pm_attach_by_id(dev, index);
> }
> +EXPORT_SYMBOL_GPL(genpd_dev_pm_attach_by_name);
Driver's shouldn't be calling this function, but instead use
dev_pm_domain_attach_by_name().
Is there any problem with that?
[...]
Kind regards
Uffe
On 26/09/18 15:09, Ulf Hansson wrote:
> On 26 September 2018 at 14:27, Jon Hunter <[email protected]> wrote:
>> Export the symbol for genpd_dev_pm_attach_by_name() so that drivers
>> built as modules can call this function.
>>
>> Signed-off-by: Jon Hunter <[email protected]>
>> ---
>> drivers/base/power/domain.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
>> index 4b5714199490..b2bbe0d848d2 100644
>> --- a/drivers/base/power/domain.c
>> +++ b/drivers/base/power/domain.c
>> @@ -2397,6 +2397,7 @@ struct device *genpd_dev_pm_attach_by_name(struct device *dev, char *name)
>>
>> return genpd_dev_pm_attach_by_id(dev, index);
>> }
>> +EXPORT_SYMBOL_GPL(genpd_dev_pm_attach_by_name);
>
> Driver's shouldn't be calling this function, but instead use
> dev_pm_domain_attach_by_name().
>
> Is there any problem with that?
No not at all. Yes that works, sorry it had been a while since I looked
at this and managed to balls it up. OK, I will drop this and correct the
API used by the driver.
Cheers!
Jon
--
nvpublic