Hi Jacopo,
Thank you for the patch,
On 27/08/18 12:28, Jacopo Mondi wrote:
> As the driver is now allowed to probe with a single output endpoint,
> power management routines shall now take into account the case a CSI-2 TX
> is not enabled.
>
> Unify the adv748x_tx_power() routine to handle transparently TXA and TXB,
> and enable the CSI-2 outputs conditionally.
Great - I definitely want to see this code refactored. Ideally with a
full removal of the register lists ... but we can get there in stages :D
>
> The AFE and HDMI backends have fixed output routes, so enable the designated
> CSI-2 output according to that.
Ah ... but what happens when the links can be changed dynamically ?
I guess we handle that then ...
>
> Signed-off-by: Jacopo Mondi <jacopo+renesas@xxxxxxxxxx>
> ---
> drivers/media/i2c/adv748x/adv748x-afe.c | 2 +-
> drivers/media/i2c/adv748x/adv748x-core.c | 52 +++++++++++++-------------------
> drivers/media/i2c/adv748x/adv748x-csi2.c | 5 ---
> drivers/media/i2c/adv748x/adv748x-hdmi.c | 2 +-
> drivers/media/i2c/adv748x/adv748x.h | 4 +--
> 5 files changed, 25 insertions(+), 40 deletions(-)
>
> diff --git a/drivers/media/i2c/adv748x/adv748x-afe.c b/drivers/media/i2c/adv748x/adv748x-afe.c
> index edd25e8..6d78105 100644
> --- a/drivers/media/i2c/adv748x/adv748x-afe.c
> +++ b/drivers/media/i2c/adv748x/adv748x-afe.c
> @@ -286,7 +286,7 @@ static int adv748x_afe_s_stream(struct v4l2_subdev *sd, int enable)
> goto unlock;
> }
>
> - ret = adv748x_txb_power(state, enable);
> + ret = adv748x_tx_power(&state->txb, enable);
> if (ret)
> goto unlock;
>
> diff --git a/drivers/media/i2c/adv748x/adv748x-core.c b/drivers/media/i2c/adv748x/adv748x-core.c
> index 78d5996..0adbcb6 100644
> --- a/drivers/media/i2c/adv748x/adv748x-core.c
> +++ b/drivers/media/i2c/adv748x/adv748x-core.c
> @@ -292,33 +292,16 @@ static const struct adv748x_reg_value adv748x_power_down_txb_1lane[] = {
> {ADV748X_PAGE_EOR, 0xff, 0xff} /* End of register table */
> };
>
> -int adv748x_txa_power(struct adv748x_state *state, bool on)
> +int adv748x_tx_power(struct adv748x_csi2 *tx, bool on)
> {
> + struct adv748x_state *state = tx->state;
> + const struct adv748x_reg_value *reglist;
> int val;
>
> - val = txa_read(state, ADV748X_CSI_FS_AS_LS);
> - if (val < 0)
> - return val;
> -
> - /*
> - * This test against BIT(6) is not documented by the datasheet, but was
> - * specified in the downstream driver.
> - * Track with a WARN_ONCE to determine if it is ever set by HW.
> - */
> - WARN_ONCE((on && val & ADV748X_CSI_FS_AS_LS_UNKNOWN),
> - "Enabling with unknown bit set");
I was going to query if this was dropped, but the check is identical in
the other function.
The original BSP driver had this condition as a 'fatal' error from what
I recall, which is why I kept it as a warning check.
I've only seen it fire once, when I had an incorrect power-on-off sequence.
I think the bit actually likely only gets set from one of the reglists
in the driver.
The bit is undocumented, and I didn't get any further detail from
querying analog devices. At somepoint I'd like to drop the WARN_ON with
an equivalent change which removes all references to this bit (whichever
entry in the register lists sets the bit).
But that's not an issue for this patch.
> -
> - if (on)
> - return adv748x_write_regs(state, adv748x_power_up_txa_4lane);
> -
> - return adv748x_write_regs(state, adv748x_power_down_txa_4lane);
> -}
> -
> -int adv748x_txb_power(struct adv748x_state *state, bool on)
> -{
> - int val;
> + if (!is_tx_enabled(tx))
> + return 0;
>
> - val = txb_read(state, ADV748X_CSI_FS_AS_LS);
I think we can remove txa_read(), txb_read() now ?
> + val = tx_read(tx, ADV748X_CSI_FS_AS_LS);
> if (val < 0)
> return val;
>
> @@ -331,9 +314,13 @@ int adv748x_txb_power(struct adv748x_state *state, bool on)
> "Enabling with unknown bit set");
>
> if (on)
> - return adv748x_write_regs(state, adv748x_power_up_txb_1lane);
> + reglist = is_txa(tx) ? adv748x_power_up_txa_4lane :
> + adv748x_power_up_txb_1lane;
> + else
> + reglist = is_txa(tx) ? adv748x_power_down_txa_4lane :
> + adv748x_power_down_txb_1lane;
>
> - return adv748x_write_regs(state, adv748x_power_down_txb_1lane);
> + return adv748x_write_regs(state, reglist);
> }
>
> /* -----------------------------------------------------------------------------
> @@ -482,6 +469,7 @@ static const struct adv748x_reg_value adv748x_init_txb_1lane[] = {
> static int adv748x_reset(struct adv748x_state *state)
> {
> int ret;
> + u8 regval = ADV748X_IO_10_PIX_OUT_EN;
As discussed, I believe this refers to the 8-bit TTL I/O port.
Ideally - this functionality should be broken out to a separate port (or
pair of ports) - but that's not for here.
Keeping this here maintains the existing behaviour of the driver so
that's fine, but perhaps if you re-spin - could you add a comment to
state that the ADV748X_IO_10_PIX_OUT_EN refers to the non-supported TTL
IO port ?
>
> ret = adv748x_write_regs(state, adv748x_sw_reset);
> if (ret < 0)
> @@ -496,22 +484,24 @@ static int adv748x_reset(struct adv748x_state *state)
> if (ret)
> return ret;
>
> - adv748x_txa_power(state, 0);
> + adv748x_tx_power(&state->txa, 0);
>
> /* Init and power down TXB */
> ret = adv748x_write_regs(state, adv748x_init_txb_1lane);
> if (ret)
> return ret;
>
> - adv748x_txb_power(state, 0);
> + adv748x_tx_power(&state->txb, 0);
>
> /* Disable chip powerdown & Enable HDMI Rx block */
> io_write(state, ADV748X_IO_PD, ADV748X_IO_PD_RX_EN);
>
> - /* Enable 4-lane CSI Tx & Pixel Port */
> - io_write(state, ADV748X_IO_10, ADV748X_IO_10_CSI4_EN |
> - ADV748X_IO_10_CSI1_EN |
> - ADV748X_IO_10_PIX_OUT_EN);
> + /* Conditionally enable 4-lane CSI Tx & Pixel Port */
> + if (is_tx_enabled(&state->txa))
> + regval |= ADV748X_IO_10_CSI4_EN;
> + if (is_tx_enabled(&state->txb))
> + regval |= ADV748X_IO_10_CSI1_EN;
> + io_write(state, ADV748X_IO_10, regval);
Could be a bit of powersaving here :) Great.
> /* Use vid_std and v_freq as freerun resolution for CP */
> cp_clrset(state, ADV748X_CP_CLMP_POS, ADV748X_CP_CLMP_POS_DIS_AUTO,
> diff --git a/drivers/media/i2c/adv748x/adv748x-csi2.c b/drivers/media/i2c/adv748x/adv748x-csi2.c
> index 709cdea..36bc786 100644
> --- a/drivers/media/i2c/adv748x/adv748x-csi2.c
> +++ b/drivers/media/i2c/adv748x/adv748x-csi2.c
> @@ -18,11 +18,6 @@
>
> #include "adv748x.h"
>
> -static bool is_txa(struct adv748x_csi2 *tx)
> -{
> - return tx == &tx->state->txa;
> -}
> -
> static int adv748x_csi2_set_virtual_channel(struct adv748x_csi2 *tx,
> unsigned int vc)
> {
> diff --git a/drivers/media/i2c/adv748x/adv748x-hdmi.c b/drivers/media/i2c/adv748x/adv748x-hdmi.c
> index aecc2a8..abb6568 100644
> --- a/drivers/media/i2c/adv748x/adv748x-hdmi.c
> +++ b/drivers/media/i2c/adv748x/adv748x-hdmi.c
> @@ -362,7 +362,7 @@ static int adv748x_hdmi_s_stream(struct v4l2_subdev *sd, int enable)
>
> mutex_lock(&state->mutex);
>
> - ret = adv748x_txa_power(state, enable);
> + ret = adv748x_tx_power(&state->txa, enable);
> if (ret)
> goto done;
>
> diff --git a/drivers/media/i2c/adv748x/adv748x.h b/drivers/media/i2c/adv748x/adv748x.h
> index 1cf46c40..2e8d37a 100644
> --- a/drivers/media/i2c/adv748x/adv748x.h
> +++ b/drivers/media/i2c/adv748x/adv748x.h
> @@ -93,6 +93,7 @@ struct adv748x_csi2 {
> #define notifier_to_csi2(n) container_of(n, struct adv748x_csi2, notifier)
> #define adv748x_sd_to_csi2(sd) container_of(sd, struct adv748x_csi2, sd)
> #define is_tx_enabled(_tx) ((_tx)->state->endpoints[(_tx)->port] != NULL)
> +#define is_txa(_tx) ((_tx) == &(_tx)->state->txa)
>
> enum adv748x_hdmi_pads {
> ADV748X_HDMI_SINK,
> @@ -400,8 +401,7 @@ void adv748x_subdev_init(struct v4l2_subdev *sd, struct adv748x_state *state,
> int adv748x_register_subdevs(struct adv748x_state *state,
> struct v4l2_device *v4l2_dev);
>
> -int adv748x_txa_power(struct adv748x_state *state, bool on);
> -int adv748x_txb_power(struct adv748x_state *state, bool on);
> +int adv748x_tx_power(struct adv748x_csi2 *tx, bool on);
>
> int adv748x_afe_init(struct adv748x_afe *afe);
> void adv748x_afe_cleanup(struct adv748x_afe *afe);
>
