[12/12] can: rcar_canfd: Add transceiver support

Commit Message

Geert Uytterhoeven Jan. 23, 2023, 6:56 p.m. UTC

Add support for CAN transceivers described as PHYs.

While simple CAN transceivers can do without, this is needed for CAN
transceivers like NXP TJR1443 that need a configuration step (like
pulling standby or enable lines), and/or impose a bitrate limit.

Signed-off-by: Geert Uytterhoeven <[e-mail]>
---
This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
https://lore.kernel.org/all/f53a1[e-mail]
---
 drivers/net/can/rcar/rcar_canfd.c | 30 +++++++++++++++++++++++++-----
 1 file changed, 25 insertions(+), 5 deletions(-)

Comments

Geert Uytterhoeven Jan. 24, 2023, 6:41 p.m. UTC | #1

On Mon, Jan 23, 2023 at 7:56 PM Geert Uytterhoeven
<[e-mail]> wrote:
> Add support for CAN transceivers described as PHYs.
>
> While simple CAN transceivers can do without, this is needed for CAN
> transceivers like NXP TJR1443 that need a configuration step (like
> pulling standby or enable lines), and/or impose a bitrate limit.
>
> Signed-off-by: Geert Uytterhoeven <[e-mail]>
> ---
> This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
> https://lore.kernel.org/all/f53a1[e-mail]

v2: "[PATCH v2 3/9] phy: Add devm_of_phy_optional_get() helper"
    https://lore.kernel.org/all/4cd00[e-mail]

I'll keep you updated when/if this ends up on an immutable branch.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [e-mail]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Marc Kleine-Budde Feb. 2, 2023, 2:40 p.m. UTC | #2

On 24.01.2023 19:41:03, Geert Uytterhoeven wrote:
> On Mon, Jan 23, 2023 at 7:56 PM Geert Uytterhoeven
> <[e-mail]> wrote:
> > Add support for CAN transceivers described as PHYs.
> >
> > While simple CAN transceivers can do without, this is needed for CAN
> > transceivers like NXP TJR1443 that need a configuration step (like
> > pulling standby or enable lines), and/or impose a bitrate limit.
> >
> > Signed-off-by: Geert Uytterhoeven <[e-mail]>
> > ---
> > This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
> > https://lore.kernel.org/all/f53a1[e-mail]
> 
> v2: "[PATCH v2 3/9] phy: Add devm_of_phy_optional_get() helper"
>     https://lore.kernel.org/all/4cd00[e-mail]
> 
> I'll keep you updated when/if this ends up on an immutable branch.

Should I take the patches 1...11 for can-next/main?

regards,
Marc

Geert Uytterhoeven Feb. 2, 2023, 2:53 p.m. UTC | #3

Hi Marc,

On Thu, Feb 2, 2023 at 3:40 PM Marc Kleine-Budde <[e-mail]> wrote:
> On 24.01.2023 19:41:03, Geert Uytterhoeven wrote:
> > On Mon, Jan 23, 2023 at 7:56 PM Geert Uytterhoeven
> > <[e-mail]> wrote:
> > > Add support for CAN transceivers described as PHYs.
> > >
> > > While simple CAN transceivers can do without, this is needed for CAN
> > > transceivers like NXP TJR1443 that need a configuration step (like
> > > pulling standby or enable lines), and/or impose a bitrate limit.
> > >
> > > Signed-off-by: Geert Uytterhoeven <[e-mail]>
> > > ---
> > > This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
> > > https://lore.kernel.org/all/f53a1[e-mail]
> >
> > v2: "[PATCH v2 3/9] phy: Add devm_of_phy_optional_get() helper"
> >     https://lore.kernel.org/all/4cd00[e-mail]
> >
> > I'll keep you updated when/if this ends up on an immutable branch.
>
> Should I take the patches 1...11 for can-next/main?

That would be great, thanks!

I had hoped Vinod would have applied the dependency, and provided
an immutable branch, as originally planned.

Alternatively, I can send a v2 of 12/12 that does not rely on the
new helper, and convert it later.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [e-mail]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Marc Kleine-Budde Feb. 2, 2023, 3:06 p.m. UTC | #4

On 02.02.2023 15:53:08, Geert Uytterhoeven wrote:
> > > > This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
> > > > https://lore.kernel.org/all/f53a1[e-mail]
> > >
> > > v2: "[PATCH v2 3/9] phy: Add devm_of_phy_optional_get() helper"
> > >     https://lore.kernel.org/all/4cd00[e-mail]
> > >
> > > I'll keep you updated when/if this ends up on an immutable branch.
> >
> > Should I take the patches 1...11 for can-next/main?
> 
> That would be great, thanks!

Done.

Marc

Geert Uytterhoeven Feb. 3, 2023, 10:24 a.m. UTC | #5

Hi Marc,

On Thu, Feb 2, 2023 at 4:06 PM Marc Kleine-Budde <[e-mail]> wrote:
> On 02.02.2023 15:53:08, Geert Uytterhoeven wrote:
> > > > > This depends on "[PATCH 1/7] phy: Add devm_of_phy_optional_get() helper".
> > > > > https://lore.kernel.org/all/f53a1[e-mail]
> > > >
> > > > v2: "[PATCH v2 3/9] phy: Add devm_of_phy_optional_get() helper"
> > > >     https://lore.kernel.org/all/4cd00[e-mail]
> > > >
> > > > I'll keep you updated when/if this ends up on an immutable branch.
> > >
> > > Should I take the patches 1...11 for can-next/main?
> >
> > That would be great, thanks!
>
> Done.

Thank you!
Meanwhile, the dependency for 12/12 is now available as an immutable
branch, cfr. https://lore.kernel.org/all/Y9za4a8qyapi4CWD@matsya

Thanks again!


Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [e-mail]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Marc Kleine-Budde Feb. 7, 2023, 7:20 p.m. UTC | #6

On 03.02.2023 11:24:08, Geert Uytterhoeven wrote:
> > > > > I'll keep you updated when/if this ends up on an immutable branch.
> > > >
> > > > Should I take the patches 1...11 for can-next/main?
> > >
> > > That would be great, thanks!
> >
> > Done.
> 
> Thank you!
> Meanwhile, the dependency for 12/12 is now available as an immutable
> branch, cfr. https://lore.kernel.org/all/Y9za4a8qyapi4CWD@matsya

net-next/main is at v6.2-rc6, but does not include this series. I assume
it will be mainlined in the v6.3 merge window. I think you can resend
this patch against can/main once v6.3-rc1 is released.

Marc

Patch

diff --git a/drivers/net/can/rcar/rcar_canfd.c b/drivers/net/can/rcar/rcar_canfd.c
index ef4e1b9a9e1ee280..6df9a259e5e4f92c 100644
--- a/drivers/net/can/rcar/rcar_canfd.c
+++ b/drivers/net/can/rcar/rcar_canfd.c
@@ -35,6 +35,7 @@ 
 #include <linux/netdevice.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/phy/phy.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
 #include <linux/types.h>
@@ -530,6 +531,7 @@  struct rcar_canfd_channel {
 	struct net_device *ndev;
 	struct rcar_canfd_global *gpriv;	/* Controller reference */
 	void __iomem *base;			/* Register base address */
+	struct phy *transceiver;		/* Optional transceiver */
 	struct napi_struct napi;
 	u32 tx_head;				/* Incremented on xmit */
 	u32 tx_tail;				/* Incremented on xmit done */
@@ -1413,11 +1415,17 @@  static int rcar_canfd_open(struct net_device *ndev)
 	struct rcar_canfd_global *gpriv = priv->gpriv;
 	int err;
 
+	err = phy_power_on(priv->transceiver);
+	if (err) {
+		netdev_err(ndev, "failed to power on PHY, error %d\n", err);
+		return err;
+	}
+
 	/* Peripheral clock is already enabled in probe */
 	err = clk_prepare_enable(gpriv->can_clk);
 	if (err) {
 		netdev_err(ndev, "failed to enable CAN clock, error %d\n", err);
-		goto out_clock;
+		goto out_phy;
 	}
 
 	err = open_candev(ndev);
@@ -1437,7 +1445,8 @@  static int rcar_canfd_open(struct net_device *ndev)
 	close_candev(ndev);
 out_can_clock:
 	clk_disable_unprepare(gpriv->can_clk);
-out_clock:
+out_phy:
+	phy_power_off(priv->transceiver);
 	return err;
 }
 
@@ -1480,6 +1489,7 @@  static int rcar_canfd_close(struct net_device *ndev)
 	napi_disable(&priv->napi);
 	clk_disable_unprepare(gpriv->can_clk);
 	close_candev(ndev);
+	phy_power_off(priv->transceiver);
 	return 0;
 }
 
@@ -1711,7 +1721,7 @@  static const struct ethtool_ops rcar_canfd_ethtool_ops = {
 };
 
 static int rcar_canfd_channel_probe(struct rcar_canfd_global *gpriv, u32 ch,
-				    u32 fcan_freq)
+				    u32 fcan_freq, struct phy *transceiver)
 {
 	const struct rcar_canfd_hw_info *info = gpriv->info;
 	struct platform_device *pdev = gpriv->pdev;
@@ -1732,8 +1742,11 @@  static int rcar_canfd_channel_probe(struct rcar_canfd_global *gpriv, u32 ch,
 	ndev->flags |= IFF_ECHO;
 	priv->ndev = ndev;
 	priv->base = gpriv->base;
+	priv->transceiver = transceiver;
 	priv->channel = ch;
 	priv->gpriv = gpriv;
+	if (transceiver)
+		priv->can.bitrate_max = transceiver->attrs.max_link_rate;
 	priv->can.clock.freq = fcan_freq;
 	dev_info(dev, "can_clk rate is %u\n", priv->can.clock.freq);
 
@@ -1836,6 +1849,7 @@  static void rcar_canfd_channel_remove(struct rcar_canfd_global *gpriv, u32 ch)
 
 static int rcar_canfd_probe(struct platform_device *pdev)
 {
+	struct phy *transceivers[RCANFD_NUM_CHANNELS] = { 0, };
 	const struct rcar_canfd_hw_info *info;
 	struct device *dev = &pdev->dev;
 	void __iomem *addr;
@@ -1857,9 +1871,14 @@  static int rcar_canfd_probe(struct platform_device *pdev)
 	for (i = 0; i < info->max_channels; ++i) {
 		name[7] = '0' + i;
 		of_child = of_get_child_by_name(dev->of_node, name);
-		if (of_child && of_device_is_available(of_child))
+		if (of_child && of_device_is_available(of_child)) {
 			channels_mask |= BIT(i);
+			transceivers[i] = devm_of_phy_optional_get(dev,
+							of_child, NULL);
+		}
 		of_node_put(of_child);
+		if (IS_ERR(transceivers[i]))
+			return PTR_ERR(transceivers[i]);
 	}
 
 	if (info->shared_global_irqs) {
@@ -2035,7 +2054,8 @@  static int rcar_canfd_probe(struct platform_device *pdev)
 	}
 
 	for_each_set_bit(ch, &gpriv->channels_mask, info->max_channels) {
-		err = rcar_canfd_channel_probe(gpriv, ch, fcan_freq);
+		err = rcar_canfd_channel_probe(gpriv, ch, fcan_freq,
+					       transceivers[ch]);
 		if (err)
 			goto fail_channel;
 	}