+ devicetree (you dropped them from v2 and v3)
On Thu, Oct 06, 2016 at 10:00:04AM +0800, Shawn Lin wrote:
> 在 2016/10/6 6:46, Brian Norris 写道:
> >Hi,
> >
> >On Wed, Oct 05, 2016 at 04:05:17PM +0800, Shawn Lin wrote:
> >>From: Brian Norris <briannorris@xxxxxxxxxxxx>
> >>
> >>rk3399 supports PCIe 2.x link speeds marginally at best, and on some
> >>boards, the link won't train at 5 GT/s at all. Rather than sacrifice 500
> >>ms waiting for training that will never happen, let's add a property
> >>from devicetree to specify link capability.
> >>
> >>Signed-off-by: Brian Norris <briannorris@xxxxxxxxxxxx>
> >>Signed-off-by: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
> >>
> >>---
> >>
> >>Changes in v3:
> >>- Cast a warning for invalid max link speed and use gen1 for it.
> >> That looks better than v2. (Suggested by Brian)
> >>
> >>Changes in v2:
> >>- rename the property to rockchip,max-link-speed according to
> >> Bjorn's recommendation and take some bits from imx6q-pcie to
> >> make this requirement more consisent.
> >>
> >> .../devicetree/bindings/pci/rockchip-pcie.txt | 2 +
> >> drivers/pci/host/pcie-rockchip.c | 62 ++++++++++++++--------
> >> 2 files changed, 42 insertions(+), 22 deletions(-)
> >>
> >>diff --git a/Documentation/devicetree/bindings/pci/rockchip-pcie.txt b/Documentation/devicetree/bindings/pci/rockchip-pcie.txt
> >>index ba67b39..7bb730e 100644
> >>--- a/Documentation/devicetree/bindings/pci/rockchip-pcie.txt
> >>+++ b/Documentation/devicetree/bindings/pci/rockchip-pcie.txt
> >>@@ -42,6 +42,8 @@ Required properties:
> >> Optional Property:
> >> - ep-gpios: contain the entry for pre-reset gpio
> >> - num-lanes: number of lanes to use
> >>+- rockchip,max-link-speed: Specify PCI gen for link capability. Must
> >>+ be '2' for gen2, and '1' for gen1, otherwise will default to gen1.
> >
> >I'm still not sure why we want incorrect DT entries to default to 1
> >instead of just being rejected as incorrect (i.e., and error). But
> >assuming people are OK with that behavior, I'll review the rest...
>
> yep, personally I don't like to stop driver from running or to
> trigger BUG_ON only if the incorrect DT value. If that happens, it
I would never suggest BUG_ON() for this. But failing to probe with an
invalid DT is fine IMO. But I won't insist; the patch is fine with me
other than the doc vs. code issue below. I'll leave it to the
maintainers on this.
> implies we should deploy proper error handling and cast a warning
> should be enough for people to debug. If the max-link-speed is
> incorrect, I assume the owner of this DT doesn't actually read
> the doc at all. At least he/she doesn't know what this property means.
> So we could use the *nature* capability for sure it could bring up
> the boards, for instance 1 lanes, gen1... etc.
>
> I was told to write code like this and have been insisting on this for
> quite a long time. Does it sound sane to you? :)
I tend to prefer strict DT interpretations, so you don't accidentally
cause yourself backwards compatibility problems in the future when you
want to extend this for chip N+1 that supports new features. But it's
not a big deal.
> >
> >> - vpcie3v3-supply: The phandle to the 3.3v regulator to use for PCIe.
> >> - vpcie1v8-supply: The phandle to the 1.8v regulator to use for PCIe.
> >> - vpcie0v9-supply: The phandle to the 0.9v regulator to use for PCIe.
> >>diff --git a/drivers/pci/host/pcie-rockchip.c b/drivers/pci/host/pcie-rockchip.c
> >>index 35591b1..66476ea 100644
> >>--- a/drivers/pci/host/pcie-rockchip.c
> >>+++ b/drivers/pci/host/pcie-rockchip.c
> >>@@ -53,6 +53,7 @@
> >> #define PCIE_CLIENT_ARI_ENABLE HIWORD_UPDATE_BIT(0x0008)
> >> #define PCIE_CLIENT_CONF_LANE_NUM(x) HIWORD_UPDATE(0x0030, ENCODE_LANES(x))
> >> #define PCIE_CLIENT_MODE_RC HIWORD_UPDATE_BIT(0x0040)
> >>+#define PCIE_CLIENT_GEN_SEL_1 HIWORD_UPDATE(0x0080, 0)
> >> #define PCIE_CLIENT_GEN_SEL_2 HIWORD_UPDATE_BIT(0x0080)
> >> #define PCIE_CLIENT_BASIC_STATUS1 (PCIE_CLIENT_BASE + 0x48)
> >> #define PCIE_CLIENT_LINK_STATUS_UP 0x00300000
> >>@@ -205,6 +206,7 @@ struct rockchip_pcie {
> >> struct gpio_desc *ep_gpio;
> >> u32 lanes;
> >> u8 root_bus_nr;
> >>+ int link_gen;
> >> struct device *dev;
> >> struct irq_domain *irq_domain;
> >> };
> >>@@ -443,13 +445,19 @@ static int rockchip_pcie_init_port(struct rockchip_pcie *rockchip)
> >> return err;
> >> }
> >>
> >>+ if (rockchip->link_gen == 2)
> >>+ rockchip_pcie_write(rockchip, PCIE_CLIENT_GEN_SEL_2,
> >>+ PCIE_CLIENT_CONFIG);
> >>+ else
> >>+ rockchip_pcie_write(rockchip, PCIE_CLIENT_GEN_SEL_1,
> >>+ PCIE_CLIENT_CONFIG);
> >>+
> >> rockchip_pcie_write(rockchip,
> >> PCIE_CLIENT_CONF_ENABLE |
> >> PCIE_CLIENT_LINK_TRAIN_ENABLE |
> >> PCIE_CLIENT_ARI_ENABLE |
> >> PCIE_CLIENT_CONF_LANE_NUM(rockchip->lanes) |
> >>- PCIE_CLIENT_MODE_RC |
> >>- PCIE_CLIENT_GEN_SEL_2,
> >>+ PCIE_CLIENT_MODE_RC,
> >> PCIE_CLIENT_CONFIG);
> >>
> >> err = phy_power_on(rockchip->phy);
> >>@@ -550,29 +558,31 @@ static int rockchip_pcie_init_port(struct rockchip_pcie *rockchip)
> >> msleep(20);
> >> }
> >>
> >>- /*
> >>- * Enable retrain for gen2. This should be configured only after
> >>- * gen1 finished.
> >>- */
> >>- status = rockchip_pcie_read(rockchip, PCIE_RC_CONFIG_LCS);
> >>- status |= PCIE_RC_CONFIG_LCS_RETRAIN_LINK;
> >>- rockchip_pcie_write(rockchip, status, PCIE_RC_CONFIG_LCS);
> >>+ if (rockchip->link_gen == 2) {
> >>+ /*
> >>+ * Enable retrain for gen2. This should be configured only after
> >>+ * gen1 finished.
> >>+ */
> >>+ status = rockchip_pcie_read(rockchip, PCIE_RC_CONFIG_LCS);
> >>+ status |= PCIE_RC_CONFIG_LCS_RETRAIN_LINK;
> >>+ rockchip_pcie_write(rockchip, status, PCIE_RC_CONFIG_LCS);
> >>+
> >>+ timeout = jiffies + msecs_to_jiffies(500);
> >>+ for (;;) {
> >>+ status = rockchip_pcie_read(rockchip, PCIE_CORE_CTRL);
> >>+ if ((status & PCIE_CORE_PL_CONF_SPEED_MASK) ==
> >>+ PCIE_CORE_PL_CONF_SPEED_5G) {
> >>+ dev_dbg(dev, "PCIe link training gen2 pass!\n");
> >>+ break;
> >>+ }
> >>
> >>- timeout = jiffies + msecs_to_jiffies(500);
> >>- for (;;) {
> >>- status = rockchip_pcie_read(rockchip, PCIE_CORE_CTRL);
> >>- if ((status & PCIE_CORE_PL_CONF_SPEED_MASK) ==
> >>- PCIE_CORE_PL_CONF_SPEED_5G) {
> >>- dev_dbg(dev, "PCIe link training gen2 pass!\n");
> >>- break;
> >>- }
> >>+ if (time_after(jiffies, timeout)) {
> >>+ dev_dbg(dev, "PCIe link training gen2 timeout, fall back to gen1!\n");
> >>+ break;
> >>+ }
> >>
> >>- if (time_after(jiffies, timeout)) {
> >>- dev_dbg(dev, "PCIe link training gen2 timeout, fall back to gen1!\n");
> >>- break;
> >>+ msleep(20);
> >> }
> >>-
> >>- msleep(20);
> >> }
> >>
> >> /* Check the final link width from negotiated lane counter from MGMT */
> >>@@ -781,6 +791,14 @@ static int rockchip_pcie_parse_dt(struct rockchip_pcie *rockchip)
> >> rockchip->lanes = 1;
> >> }
> >>
> >>+ rockchip->link_gen = 2;
> >
> >Isn't the above assignment violating the binding doc? Setting to 2 means
> >that if the property isn't found, we default to gen2. That's OK with me,
> >but you should make the binding doc match it.
>
> It should be backward compatible with the old DTBs without this
> propery. Otherwise, they did use gen2 but they cannot achieve it now
> without adding the new propery.
Yes, defaulting to gen2 is probably the best choice.
> But I agreee I should commet the doc more clearly.
Yes, just please update the doc.
Brian
> >
> >>+ err = of_property_read_u32(node, "rockchip,max-link-speed",
> >>+ &rockchip->link_gen);
> >
> >^^ you can get sign errors here, since link_gen is 'int', but this
> >function reads 'u32'. Maybe use a temporary u32.
> >
>
> Good catch.
>
> >Brian
> >
> >>+ if (!err && rockchip->link_gen != 1 && rockchip->link_gen != 2) {
> >>+ dev_warn(dev, "invalid max-link-speed, default to use gen1\n");
> >>+ rockchip->link_gen = 1;
> >>+ }
> >>+
> >> rockchip->core_rst = devm_reset_control_get(dev, "core");
> >> if (IS_ERR(rockchip->core_rst)) {
> >> if (PTR_ERR(rockchip->core_rst) != -EPROBE_DEFER)
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
