gpu/drm: tegra: add DSI support for Tegra20/Tegra30
Subsystems (6)
v2 cover gpu/drm: tegra: add DSI support for Tegra20/Tegra30
Tegra20/Tegra30 DSI is quite similar to Tegra114+ apart MIPI calibration logic and clocks. With a few minor tweaks, existing tegra DSI driver should work on Tegra20/Tegra30 devices just fine. Tested on Motorola Atrix 4G (T20) and ASUS VivoTab RT TF600T (T30).
This patchset depends on Tegra20/Tegra30 CSI bringup since both share MIPI calibration logic. Ideally these patches should be picked after CSI bringup but they will not break anything even if picked before CSI patches.
---
Changes in v2:
- removed all MIPI calibration, it is handled within CSI bringup
- added per-soc structures into of_match
- added fix for hang caused by register access with uninited hw
---
Svyatoslav Ryhel (4):
clk: tegra20: reparent dsi clock to pll_d_out0 gpu/drm: tegra: dsi: move prepare function at the top of encoder
enable
gpu/drm: tegra: dsi: add support for Tegra20/Tegra30
ARM: tegra: adjust DSI nodes for Tegra20/Tegra30
arch/arm/boot/dts/nvidia/tegra20.dtsi | 4 ++ arch/arm/boot/dts/nvidia/tegra30.dtsi | 8 +++ drivers/clk/tegra/clk-tegra20.c | 6 +- drivers/gpu/drm/tegra/drm.c | 2 + drivers/gpu/drm/tegra/dsi.c | 100 ++++++++++++++++---------- drivers/gpu/drm/tegra/dsi.h | 15 ++++ 6 files changed, 95 insertions(+), 40 deletions(-)
Signature · 1 line
2.48.1
PATCH RESEND v2 1/4 clk: tegra20: reparent dsi clock to pll_d_out0
Reparent DSI clock to PLLD_OUT0 instead of directly descend from PLLD.
--- drivers/clk/tegra/clk-tegra20.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/clk/tegra/clk-tegra20.c +3 -3
@@ -801,9 +801,9 @@ static void __init tegra20_periph_clk_init(void) clks[TEGRA20_CLK_MC ] = clk;/* dsi */ - clk =tegra_clk_register_periph_gate ("dsi" ,"pll_d" ,0 , clk_base,0 , -48 , periph_clk_enb_refcnt); -clk_register_clkdev (clk,NULL ,"dsi" ); + clk =tegra_clk_register_periph_gate ("dsi" ,"pll_d_out0" ,0 , + clk_base,0 ,TEGRA20_CLK_DSI , + periph_clk_enb_refcnt); clks[TEGRA20_CLK_DSI ] = clk;/* csus */
Signature · 1 line
2.48.1
Re: [PATCH v2 1/4 RESEND] clk: tegra20: reparent dsi clock to pll_d_out0
Quoted 16 lines · Svyatoslav Ryhel
Reparent DSI clock to
PLLD_OUT0instead of directly descend from PLLD.Signed-off-by Svyatoslav Ryhel <[email protected]>Acked-by Stephen Boyd <[email protected]>--- drivers/clk/tegra/clk-tegra20.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)drivers/clk/tegra/clk-tegra20.c +0
-0@@ -801,9 +801,9 @@ static void __init tegra20_periph_clk_init(void) clks[TEGRA20_CLK_MC ] = clk;
drivers/clk/tegra/clk-tegra20.c static void __init tegra20_periph_clk_init(void) · line 809
Quoted 10 lines
drivers/clk/tegra/clk-tegra20.c +3
-3/* dsi */ - clk =tegra_clk_register_periph_gate ("dsi" ,"pll_d" ,0 , clk_base,0 , -48 , periph_clk_enb_refcnt); -clk_register_clkdev (clk,NULL ,"dsi" ); + clk =tegra_clk_register_periph_gate ("dsi" ,"pll_d_out0" ,0 , + clk_base,0 ,TEGRA20_CLK_DSI , + periph_clk_enb_refcnt); clks[TEGRA20_CLK_DSI ] = clk;/* csus */
Thierry Reding
AppliedRe: [PATCH v2 1/4 RESEND] clk: tegra20: reparent dsi clock to pll_d_out0
In reply to Svyatoslav RyhelQuoted 7 lines · Svyatoslav Ryhel
Reparent DSI clock to
PLLD_OUT0instead of directly descend from PLLD.Signed-off-by Svyatoslav Ryhel <[email protected]>Acked-by Stephen Boyd <[email protected]>--- drivers/clk/tegra/clk-tegra20.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
Applied, thanks.
Thierry
PATCH RESEND v2 2/4 gpu/drm: tegra: dsi: move prepare function to the top of encoder enable
In reply to Svyatoslav RyhelThe tegra_dsi_prepare function performs hardware setup and should be called before any register readings or there will be a risk of device hangup on register access. To avoid this situation, tegra_dsi_prepare must be called at the beginning of tegra_dsi_encoder_enable.
--- drivers/gpu/drm/tegra/dsi.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/gpu/drm/tegra/dsi.c +6 -6
@@ -914,6 +914,12 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)u32 value;int err; + err =tegra_dsi_prepare (dsi); +if (err <0 ) { +dev_err (dsi->dev,"failed to prepare: %d\n" , err); +return ; + } +/* If the bootloader enabled DSI it needs to be disabled * in order for the panel initialization commands to be * properly sent.
@@ -923,12 +929,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)if (value &DSI_POWER_CONTROL_ENABLE )tegra_dsi_disable (dsi); - err =tegra_dsi_prepare (dsi); -if (err <0 ) { -dev_err (dsi->dev,"failed to prepare: %d\n" , err); -return ; - } - state =tegra_dsi_get_state (dsi);tegra_dsi_set_timeout (dsi, state->bclk, state->vrefresh);
Signature · 1 line
2.48.1
Mikko Perttunen
maintainerRe: [PATCH v2 2/4 RESEND] gpu/drm: tegra: dsi: move prepare function to the top of encoder enable
Quoted 31 lines · Svyatoslav Ryhel
The
tegra_dsi_preparefunction performs hardware setup and should be called before any register readings or there will be a risk of device hangup on register access. To avoid this situation,tegra_dsi_preparemust be called at the beginning oftegra_dsi_encoder_enable.Signed-off-by Svyatoslav Ryhel <[email protected]>--- drivers/gpu/drm/tegra/dsi.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)drivers/gpu/drm/tegra/dsi.c +6
-0@@ -914,6 +914,12 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)u32 value;int err; + err =tegra_dsi_prepare (dsi); +if (err <0 ) { +dev_err (dsi->dev,"failed to prepare: %d\n" , err); +return ; + } +/* If the bootloader enabled DSI it needs to be disabled * in order for the panel initialization commands to be * properly sent. @@ -923,12 +929,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)if (value &DSI_POWER_CONTROL_ENABLE )tegra_dsi_disable (dsi);
drivers/gpu/drm/tegra/dsi.c static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) · line 934
Quoted 10 lines
drivers/gpu/drm/tegra/dsi.c +0
-6- err =tegra_dsi_prepare (dsi); -if (err <0 ) { -dev_err (dsi->dev,"failed to prepare: %d\n" , err); -return ; - } - state =tegra_dsi_get_state (dsi);tegra_dsi_set_timeout (dsi, state->bclk, state->vrefresh);
The section of code before the tegra_dsi_prepare call was removed in 'Revert "drm/tegra: dsi: Clear enable register if powered by bootloader"', so this patch should no longer be necessary.
Mikko
Re: [PATCH v2 2/4 RESEND] gpu/drm: tegra: dsi: move prepare function to the top of encoder enable
чт, 15 січ. 2026 р. о 07:54 Mikko Perttunen <[email protected]> пише:
Quoted 41 lines · Svyatoslav Ryhel
Quoted 39 lines
The
tegra_dsi_preparefunction performs hardware setup and should be called before any register readings or there will be a risk of device hangup on register access. To avoid this situation,tegra_dsi_preparemust be called at the beginning oftegra_dsi_encoder_enable.Signed-off-by Svyatoslav Ryhel <[email protected]>--- drivers/gpu/drm/tegra/dsi.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)drivers/gpu/drm/tegra/dsi.c +6
-6@@ -914,6 +914,12 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)u32 value;int err; + err =tegra_dsi_prepare (dsi); +if (err <0 ) { +dev_err (dsi->dev,"failed to prepare: %d\n" , err); +return ; + } +/* If the bootloader enabled DSI it needs to be disabled * in order for the panel initialization commands to be * properly sent. @@ -923,12 +929,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)if (value &DSI_POWER_CONTROL_ENABLE )tegra_dsi_disable (dsi); - err =tegra_dsi_prepare (dsi); -if (err <0 ) { -dev_err (dsi->dev,"failed to prepare: %d\n" , err); -return ; - } - state =tegra_dsi_get_state (dsi);
Quoted 7 lines
drivers/gpu/drm/tegra/dsi.c static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) · line 934
Quoted 2 lines
drivers/gpu/drm/tegra/dsi.c +0
-0tegra_dsi_set_timeout (dsi, state->bclk, state->vrefresh);The section of code before the
tegra_dsi_preparecall was removed in 'Revert "drm/tegra: dsi: Clear enable register if powered by bootloader"', so this patch should no longer be necessary.Mikko
You are correct. I have found this when rebasing onto v6.18 which was much later then this series was resent. Obviously, this patch would be dropped on the next resend/v3. Sorry for inconvenience.
Quoted 2 lines
Re: [PATCH v2 2/4 RESEND] gpu/drm: tegra: dsi: move prepare function to the top of encoder enable
чт, 15 січ. 2026 р. о 09:49 Svyatoslav Ryhel <[email protected]> пише:
Quoted 48 lines
чт, 15 січ. 2026 р. о 07:54 Mikko Perttunen <[email protected]> пише:
Quoted 46 lines · Svyatoslav Ryhel
Quoted 31 lines
The
tegra_dsi_preparefunction performs hardware setup and should be called before any register readings or there will be a risk of device hangup on register access. To avoid this situation,tegra_dsi_preparemust be called at the beginning oftegra_dsi_encoder_enable.Signed-off-by Svyatoslav Ryhel <[email protected]>--- drivers/gpu/drm/tegra/dsi.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)drivers/gpu/drm/tegra/dsi.c +6
-0@@ -914,6 +914,12 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)u32 value;int err; + err =tegra_dsi_prepare (dsi); +if (err <0 ) { +dev_err (dsi->dev,"failed to prepare: %d\n" , err); +return ; + } +/* If the bootloader enabled DSI it needs to be disabled * in order for the panel initialization commands to be * properly sent. @@ -923,12 +929,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)if (value &DSI_POWER_CONTROL_ENABLE )tegra_dsi_disable (dsi);drivers/gpu/drm/tegra/dsi.c static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) · line 934
Quoted 10 lines
drivers/gpu/drm/tegra/dsi.c +0
-6- err =tegra_dsi_prepare (dsi); -if (err <0 ) { -dev_err (dsi->dev,"failed to prepare: %d\n" , err); -return ; - } - state =tegra_dsi_get_state (dsi);tegra_dsi_set_timeout (dsi, state->bclk, state->vrefresh);The section of code before the
tegra_dsi_preparecall was removed in 'Revert "drm/tegra: dsi: Clear enable register if powered by bootloader"', so this patch should no longer be necessary.
Quoted 7 lines
Quoted 2 lines
Mikko
You are correct. I have found this when rebasing onto v6.18 which was much later then this series was resent. Obviously, this patch would be dropped on the next resend/v3. Sorry for inconvenience.
Actually, "Revert "drm/tegra: dsi: Clear enable register if powered by bootloader"" causes a regression for me (DSI on Tegra20 device, Motorola Atrix 4G). DSI initiated by bootloader will be stuck if not disabled pre reconfiguration, Dmitry was right by bringing that change, but he did not take into account that if DSI is accessed from cold state (not configured by bootloader), device will hang. This patch addresses this behavior of Dmitry's patch and prevents regression introduced by reverting it.
Quoted 2 lines
Quoted 2 lines
PATCH RESEND v2 3/4 gpu/drm: tegra: dsi: add support for Tegra20/Tegra30
In reply to Svyatoslav RyhelTegra20 and Tegra30 are fully compatible with existing tegra DSI driver apart from clock configuration and PAD calibration which are addressed by this patch.
--- drivers/gpu/drm/tegra/drm.c | 2 + drivers/gpu/drm/tegra/dsi.c | 88 ++++++++++++++++++++++++------------- drivers/gpu/drm/tegra/dsi.h | 15 +++++++ 3 files changed, 74 insertions(+), 31 deletions(-)
drivers/gpu/drm/tegra/drm.c +2 -0
@@ -1359,10 +1359,12 @@ static SIMPLE_DEV_PM_OPS(host1x_drm_pm_ops, host1x_drm_suspend,static const struct of_device_id host1x_drm_subdevs[] = { { .compatible ="nvidia,tegra20-dc" , }, + { .compatible ="nvidia,tegra20-dsi" , }, { .compatible ="nvidia,tegra20-hdmi" , }, { .compatible ="nvidia,tegra20-gr2d" , }, { .compatible ="nvidia,tegra20-gr3d" , }, { .compatible ="nvidia,tegra30-dc" , }, + { .compatible ="nvidia,tegra30-dsi" , }, { .compatible ="nvidia,tegra30-hdmi" , }, { .compatible ="nvidia,tegra30-gr2d" , }, { .compatible ="nvidia,tegra30-gr3d" , },
drivers/gpu/drm/tegra/dsi.c +57 -31
142 lines not on this page · raw message
drivers/gpu/drm/tegra/dsi.h +15 -0
@@ -95,6 +95,16 @@#define DSI_TALLY_LRX (x) (((x) & 0xff) << 8)#define DSI_TALLY_HTX (x) (((x) & 0xff) << 0)#define DSI_PAD_CONTROL_0 0x4b +/* DSI V0 */ +#define DSI_PAD_CONTROL_PULLDN_ENAB (x) (((x) & 0x1) << 28) +#define DSI_PAD_CONTROL_SLEWUPADJ (x) (((x) & 0x7) << 24) +#define DSI_PAD_CONTROL_SLEWDNADJ (x) (((x) & 0x7) << 20) +#define DSI_PAD_CONTROL_PREEMP_EN (x) (((x) & 0x1) << 19) +#define DSI_PAD_CONTROL_PDIO_CLK (x) (((x) & 0x1) << 18) +#define DSI_PAD_CONTROL_PDIO (x) (((x) & 0x3) << 16) +#define DSI_PAD_CONTROL_LPUPADJ (x) (((x) & 0x3) << 14) +#define DSI_PAD_CONTROL_LPDNADJ (x) (((x) & 0x3) << 12) +/* DSI V1 */ #define DSI_PAD_CONTROL_VS1_PDIO (x) (((x) & 0xf) << 0)#define DSI_PAD_CONTROL_VS1_PDIO_CLK (1 << 8)#define DSI_PAD_CONTROL_VS1_PULLDN (x) (((x) & 0xf) << 16)
@@ -140,4 +150,9 @@ enum tegra_dsi_format { TEGRA_DSI_FORMAT_24P, }; +enum tegra_dsi_version { +TEGRA_DSI_V0 , +TEGRA_DSI_V1 , +}; +#endif
Signature · 1 line
2.48.1
Mikko Perttunen
maintainerRe: [PATCH v2 3/4 RESEND] gpu/drm: tegra: dsi: add support for Tegra20/Tegra30
Quoted 33 lines · Svyatoslav Ryhel
Tegra20 and Tegra30 are fully compatible with existing tegra DSI driver apart from clock configuration and PAD calibration which are addressed by this patch.
Signed-off-by Svyatoslav Ryhel <[email protected]>--- drivers/gpu/drm/tegra/drm.c | 2 + drivers/gpu/drm/tegra/dsi.c | 88 ++++++++++++++++++++++++------------- drivers/gpu/drm/tegra/dsi.h | 15 +++++++ 3 files changed, 74 insertions(+), 31 deletions(-)drivers/gpu/drm/tegra/drm.c +2
-0@@ -1359,10 +1359,12 @@ static SIMPLE_DEV_PM_OPS(host1x_drm_pm_ops, host1x_drm_suspend,static const struct of_device_id host1x_drm_subdevs[] = { { .compatible ="nvidia,tegra20-dc" , }, + { .compatible ="nvidia,tegra20-dsi" , }, { .compatible ="nvidia,tegra20-hdmi" , }, { .compatible ="nvidia,tegra20-gr2d" , }, { .compatible ="nvidia,tegra20-gr3d" , }, { .compatible ="nvidia,tegra30-dc" , }, + { .compatible ="nvidia,tegra30-dsi" , }, { .compatible ="nvidia,tegra30-hdmi" , }, { .compatible ="nvidia,tegra30-gr2d" , }, { .compatible ="nvidia,tegra30-gr3d" , },drivers/gpu/drm/tegra/dsi.c +0
-0
drivers/gpu/drm/tegra/dsi.c to_dsi_state(struct drm_connector_state *state) · line 57
Quoted 5 lines
drivers/gpu/drm/tegra/dsi.c +2
-0@@ -53,6 +53,10 @@ to_dsi_state(struct drm_connector_state *state) return container_of(state, struct tegra_dsi_state, base); } +struct tegra_dsi_config { +u32 dsi_version;
Rather than a dsi_version field, we should have something that describes the difference. e.g. 'bool has_multiple_pad_controls' being true for the "V1" hardware.
Quoted 101 lines
drivers/gpu/drm/tegra/dsi.c +35
-23+}; +struct tegra_dsi {struct host1x_client client;struct tegra_output output;@@ -82,6 +86,8 @@ struct tegra_dsi {/* for ganged-mode support */ struct tegra_dsi *master;struct tegra_dsi *slave; + +const struct tegra_dsi_config *config; }; static inline struct tegra_dsi *@@ -663,39 +669,46 @@ static int tegra_dsi_pad_enable(struct tegra_dsi *dsi) {u32 value; - value =DSI_PAD_CONTROL_VS1_PULLDN (0 ) |DSI_PAD_CONTROL_VS1_PDIO (0 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); +if (dsi->config->dsi_version ==TEGRA_DSI_V1 ) { +/* +* XXX Is this still needed? The module reset is deasserted right +* before this function is called. +*/ +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_0 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_1 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_2 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_3 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_4 ); + + value =DSI_PAD_CONTROL_VS1_PULLDN (0 ) |DSI_PAD_CONTROL_VS1_PDIO (0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); + + value =DSI_PAD_SLEW_UP (0x7 ) |DSI_PAD_SLEW_DN (0x7 ) | +DSI_PAD_LP_UP (0x1 ) |DSI_PAD_LP_DN (0x1 ) | +DSI_PAD_OUT_CLK (0x0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_2 ); + + value =DSI_PAD_PREEMP_PD_CLK (0x3 ) |DSI_PAD_PREEMP_PU_CLK (0x3 ) | +DSI_PAD_PREEMP_PD (0x03 ) |DSI_PAD_PREEMP_PU (0x3 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_3 ); + }else { + value =DSI_PAD_CONTROL_LPUPADJ (0x1 ) |DSI_PAD_CONTROL_LPDNADJ (0x1 ) | +DSI_PAD_CONTROL_PREEMP_EN (0x1 ) |DSI_PAD_CONTROL_SLEWDNADJ (0x6 ) | +DSI_PAD_CONTROL_SLEWUPADJ (0x6 ) |DSI_PAD_CONTROL_PDIO (0 ) | +DSI_PAD_CONTROL_PDIO_CLK (0 ) |DSI_PAD_CONTROL_PULLDN_ENAB (0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); + }return 0 ; }static int tegra_dsi_pad_calibrate (struct tegra_dsi *dsi) { -u32 value;int err; -/* -* XXX Is this still needed? The module reset is deasserted right -* before this function is called. -*/ -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_0 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_1 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_2 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_3 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_4 ); -/* start calibration */ tegra_dsi_pad_enable (dsi); - value =DSI_PAD_SLEW_UP (0x7 ) |DSI_PAD_SLEW_DN (0x7 ) | -DSI_PAD_LP_UP (0x1 ) |DSI_PAD_LP_DN (0x1 ) | -DSI_PAD_OUT_CLK (0x0 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_2 ); - - value =DSI_PAD_PREEMP_PD_CLK (0x3 ) |DSI_PAD_PREEMP_PU_CLK (0x3 ) | -DSI_PAD_PREEMP_PD (0x03 ) |DSI_PAD_PREEMP_PU (0x3 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_3 ); - err =tegra_mipi_start_calibration (dsi->mipi);if (err <0 )return err;@@ -1577,6 +1590,7 @@ static int tegra_dsi_probe(struct platform_device *pdev)if (!dsi)return -ENOMEM ; + dsi->config =of_device_get_match_data (&pdev->dev); dsi->output.dev = dsi->dev = &pdev->dev; dsi->video_fifo_depth =1920 ; dsi->host_fifo_depth =64 ;@@ -1615,7 +1629,7 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto remove; } - dsi->clk_lp =devm_clk_get (&pdev->dev,"lp" ); + dsi->clk_lp =devm_clk_get_optional (&pdev->dev,"lp" );if (IS_ERR (dsi->clk_lp)) { err =dev_err_probe (&pdev->dev,PTR_ERR (dsi->clk_lp),"cannot get low-power clock\n" );
drivers/gpu/drm/tegra/dsi.c static int tegra_dsi_probe(struct platform_device *pdev) · line 1654
Quoted 9 lines
drivers/gpu/drm/tegra/dsi.c +1
-4@@ -1636,10 +1650,12 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto remove; } - err =tegra_dsi_setup_clocks (dsi); -if (err <0 ) { -dev_err (&pdev->dev,"cannot setup clocks\n" ); - goto remove; +if (dsi->config->dsi_version ==TEGRA_DSI_V1 ) {
And 'has_mux_parent_clk', perhaps? Not a very good name, if you come up with something better feel free to use it.
I checked, and looks like Tegra30 in fact allows DSIA/DSIB to be muxed to either PLL_D or PLL_D2, it is just not modelled in the clock driver. The same applies to all of Tegra114, 124, and 210, but only Tegra114 has it presently modelled as a mux whereas 124 and 210 have a gate. The TRMs for 124 and 210 specify that only PLL_D can be used in the mux bit documentation so the bit is probably dysfunctional on those chips.
The correct solution, hence, I think is to have that config field, and it be false for Tegra20, 124, and 210; and true for Tegra30, and 114. Then implement the mux clock in the Tegra30 clock driver.
However, I would settle for setting it to false for both Tegra20 and Tegra30 and leaving a comment next to the Tegra30 entry describing the situation (the mux clock is not currently modelled).
Cheers,
Mikko
Quoted 66 lines
drivers/gpu/drm/tegra/dsi.c +5
-0+ err =tegra_dsi_setup_clocks (dsi); +if (err <0 ) { +dev_err (&pdev->dev,"cannot setup clocks\n" ); + goto remove; + } }drivers/gpu/drm/tegra/dsi.c +14
-4dsi->regs =devm_platform_ioremap_resource (pdev,0 );@@ -1703,11 +1719,21 @@ static void tegra_dsi_remove(struct platform_device *pdev)tegra_mipi_free (dsi->mipi); } +static const struct tegra_dsi_config tegra20_dsi_config = { + .dsi_version =TEGRA_DSI_V0 , +}; + +static const struct tegra_dsi_config tegra114_dsi_config = { + .dsi_version =TEGRA_DSI_V1 , +}; +static const struct of_device_id tegra_dsi_of_match[] = { - { .compatible ="nvidia,tegra210-dsi" , }, - { .compatible ="nvidia,tegra132-dsi" , }, - { .compatible ="nvidia,tegra124-dsi" , }, - { .compatible ="nvidia,tegra114-dsi" , }, + { .compatible ="nvidia,tegra210-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra132-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra124-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra114-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra30-dsi" , .data = &tegra20_dsi_config }, + { .compatible ="nvidia,tegra20-dsi" , .data = &tegra20_dsi_config }, { }, };MODULE_DEVICE_TABLE (of, tegra_dsi_of_match);drivers/gpu/drm/tegra/dsi.h +15
-0@@ -95,6 +95,16 @@#define DSI_TALLY_LRX (x) (((x) & 0xff) << 8)#define DSI_TALLY_HTX (x) (((x) & 0xff) << 0)#define DSI_PAD_CONTROL_0 0x4b +/* DSI V0 */ +#define DSI_PAD_CONTROL_PULLDN_ENAB (x) (((x) & 0x1) << 28) +#define DSI_PAD_CONTROL_SLEWUPADJ (x) (((x) & 0x7) << 24) +#define DSI_PAD_CONTROL_SLEWDNADJ (x) (((x) & 0x7) << 20) +#define DSI_PAD_CONTROL_PREEMP_EN (x) (((x) & 0x1) << 19) +#define DSI_PAD_CONTROL_PDIO_CLK (x) (((x) & 0x1) << 18) +#define DSI_PAD_CONTROL_PDIO (x) (((x) & 0x3) << 16) +#define DSI_PAD_CONTROL_LPUPADJ (x) (((x) & 0x3) << 14) +#define DSI_PAD_CONTROL_LPDNADJ (x) (((x) & 0x3) << 12) +/* DSI V1 */ #define DSI_PAD_CONTROL_VS1_PDIO (x) (((x) & 0xf) << 0)#define DSI_PAD_CONTROL_VS1_PDIO_CLK (1 << 8)#define DSI_PAD_CONTROL_VS1_PULLDN (x) (((x) & 0xf) << 16)@@ -140,4 +150,9 @@ enum tegra_dsi_format { TEGRA_DSI_FORMAT_24P, }; +enum tegra_dsi_version { +TEGRA_DSI_V0 , +TEGRA_DSI_V1 , +}; +#endif
Re: [PATCH v2 3/4 RESEND] gpu/drm: tegra: dsi: add support for Tegra20/Tegra30
чт, 15 січ. 2026 р. о 09:02 Mikko Perttunen <[email protected]> пише:
Quoted 38 lines · Svyatoslav Ryhel
Quoted 36 lines
Tegra20 and Tegra30 are fully compatible with existing tegra DSI driver apart from clock configuration and PAD calibration which are addressed by this patch.
Signed-off-by Svyatoslav Ryhel <[email protected]>--- drivers/gpu/drm/tegra/drm.c | 2 + drivers/gpu/drm/tegra/dsi.c | 88 ++++++++++++++++++++++++------------- drivers/gpu/drm/tegra/dsi.h | 15 +++++++ 3 files changed, 74 insertions(+), 31 deletions(-)drivers/gpu/drm/tegra/drm.c +2
-0@@ -1359,10 +1359,12 @@ static SIMPLE_DEV_PM_OPS(host1x_drm_pm_ops, host1x_drm_suspend,static const struct of_device_id host1x_drm_subdevs[] = { { .compatible ="nvidia,tegra20-dc" , }, + { .compatible ="nvidia,tegra20-dsi" , }, { .compatible ="nvidia,tegra20-hdmi" , }, { .compatible ="nvidia,tegra20-gr2d" , }, { .compatible ="nvidia,tegra20-gr3d" , }, { .compatible ="nvidia,tegra30-dc" , }, + { .compatible ="nvidia,tegra30-dsi" , }, { .compatible ="nvidia,tegra30-hdmi" , }, { .compatible ="nvidia,tegra30-gr2d" , }, { .compatible ="nvidia,tegra30-gr3d" , },drivers/gpu/drm/tegra/dsi.c +0
-0@@ -53,6 +53,10 @@ to_dsi_state(struct drm_connector_state *state) return container_of(state, struct tegra_dsi_state, base); }
Quoted 5 lines
drivers/gpu/drm/tegra/dsi.c to_dsi_state(struct drm_connector_state *state) · line 57
Quoted 2 lines
drivers/gpu/drm/tegra/dsi.c +2
-0+struct tegra_dsi_config { +u32 dsi_version;Rather than a
dsi_versionfield, we should have something that describes the difference. e.g. 'boolhas_multiple_pad_controls' being true for the "V1" hardware.
I used versioning from downstream dsi sources, where with addition of Tegra114 support older DSI controller got v0 and newer v1, but if you find 'has_multiple_pad_controls' more suitable I have no objections.
Quoted 104 lines
Quoted 104 lines
drivers/gpu/drm/tegra/dsi.c +35
-23+}; +struct tegra_dsi {struct host1x_client client;struct tegra_output output;@@ -82,6 +86,8 @@ struct tegra_dsi {/* for ganged-mode support */ struct tegra_dsi *master;struct tegra_dsi *slave; + +const struct tegra_dsi_config *config; }; static inline struct tegra_dsi *@@ -663,39 +669,46 @@ static int tegra_dsi_pad_enable(struct tegra_dsi *dsi) {u32 value; - value =DSI_PAD_CONTROL_VS1_PULLDN (0 ) |DSI_PAD_CONTROL_VS1_PDIO (0 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); +if (dsi->config->dsi_version ==TEGRA_DSI_V1 ) { +/* +* XXX Is this still needed? The module reset is deasserted right +* before this function is called. +*/ +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_0 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_1 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_2 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_3 ); +tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_4 ); + + value =DSI_PAD_CONTROL_VS1_PULLDN (0 ) |DSI_PAD_CONTROL_VS1_PDIO (0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); + + value =DSI_PAD_SLEW_UP (0x7 ) |DSI_PAD_SLEW_DN (0x7 ) | +DSI_PAD_LP_UP (0x1 ) |DSI_PAD_LP_DN (0x1 ) | +DSI_PAD_OUT_CLK (0x0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_2 ); + + value =DSI_PAD_PREEMP_PD_CLK (0x3 ) |DSI_PAD_PREEMP_PU_CLK (0x3 ) | +DSI_PAD_PREEMP_PD (0x03 ) |DSI_PAD_PREEMP_PU (0x3 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_3 ); + }else { + value =DSI_PAD_CONTROL_LPUPADJ (0x1 ) |DSI_PAD_CONTROL_LPDNADJ (0x1 ) | +DSI_PAD_CONTROL_PREEMP_EN (0x1 ) |DSI_PAD_CONTROL_SLEWDNADJ (0x6 ) | +DSI_PAD_CONTROL_SLEWUPADJ (0x6 ) |DSI_PAD_CONTROL_PDIO (0 ) | +DSI_PAD_CONTROL_PDIO_CLK (0 ) |DSI_PAD_CONTROL_PULLDN_ENAB (0 ); +tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_0 ); + }return 0 ; }static int tegra_dsi_pad_calibrate (struct tegra_dsi *dsi) { -u32 value;int err; -/* -* XXX Is this still needed? The module reset is deasserted right -* before this function is called. -*/ -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_0 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_1 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_2 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_3 ); -tegra_dsi_writel (dsi,0 ,DSI_PAD_CONTROL_4 ); -/* start calibration */ tegra_dsi_pad_enable (dsi); - value =DSI_PAD_SLEW_UP (0x7 ) |DSI_PAD_SLEW_DN (0x7 ) | -DSI_PAD_LP_UP (0x1 ) |DSI_PAD_LP_DN (0x1 ) | -DSI_PAD_OUT_CLK (0x0 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_2 ); - - value =DSI_PAD_PREEMP_PD_CLK (0x3 ) |DSI_PAD_PREEMP_PU_CLK (0x3 ) | -DSI_PAD_PREEMP_PD (0x03 ) |DSI_PAD_PREEMP_PU (0x3 ); -tegra_dsi_writel (dsi, value,DSI_PAD_CONTROL_3 ); - err =tegra_mipi_start_calibration (dsi->mipi);if (err <0 )return err;@@ -1577,6 +1590,7 @@ static int tegra_dsi_probe(struct platform_device *pdev)if (!dsi)return -ENOMEM ; + dsi->config =of_device_get_match_data (&pdev->dev); dsi->output.dev = dsi->dev = &pdev->dev; dsi->video_fifo_depth =1920 ; dsi->host_fifo_depth =64 ;@@ -1615,7 +1629,7 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto remove; } - dsi->clk_lp =devm_clk_get (&pdev->dev,"lp" ); + dsi->clk_lp =devm_clk_get_optional (&pdev->dev,"lp" );if (IS_ERR (dsi->clk_lp)) { err =dev_err_probe (&pdev->dev,PTR_ERR (dsi->clk_lp),"cannot get low-power clock\n" );@@ -1636,10 +1650,12 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto remove; }
Quoted 8 lines
drivers/gpu/drm/tegra/dsi.c static int tegra_dsi_probe(struct platform_device *pdev) · line 1653
Quoted 6 lines
drivers/gpu/drm/tegra/dsi.c +1
-4- err =tegra_dsi_setup_clocks (dsi); -if (err <0 ) { -dev_err (&pdev->dev,"cannot setup clocks\n" ); - goto remove; +if (dsi->config->dsi_version ==TEGRA_DSI_V1 ) {And '
has_mux_parent_clk', perhaps? Not a very good name, if you come up with something better feel free to use it.
Noted
Quoted 7 lines
I checked, and looks like Tegra30 in fact allows DSIA/DSIB to be muxed to either
PLL_DorPLL_D2, it is just not modelled in the clock driver. The same applies to all of Tegra114, 124, and 210, but only Tegra114 has it presently modelled as a mux whereas 124 and 210 have a gate. The TRMs for 124 and 210 specify that onlyPLL_Dcan be used in the mux bit documentation so the bit is probably dysfunctional on those chips.The correct solution, hence, I think is to have that config field, and it be false for Tegra20, 124, and 210; and true for Tegra30, and 114. Then implement the mux clock in the Tegra30 clock driver.
However, I would settle for setting it to false for both Tegra20 and Tegra30 and leaving a comment next to the Tegra30 entry describing the situation (the mux clock is not currently modelled).
I clearly understand your point. I agree regarding setting it to false for both Tegra20 and Tegra30 and leaving a comment for Tegra30, it seems to be the optimal solution. Later on if there would be such need, Tegra30 can get its mux too.
A little note regarding mux on Tegra124/210, it seems that bit is not dysfunctional but behavior is intentional to diverge PLLD and D2 as "internal" and "external" main display clock. At least this is what I have concluded while working with Tegra display controller stuff.
Quoted 3 lines
Cheers,
Mikko
Thank you for your review!
Best regards,
Svyatoslav R.
Quoted 70 lines
Quoted 66 lines
drivers/gpu/drm/tegra/dsi.c +5
-0+ err =tegra_dsi_setup_clocks (dsi); +if (err <0 ) { +dev_err (&pdev->dev,"cannot setup clocks\n" ); + goto remove; + } }dsi->regs = devm_platform_ioremap_resource(pdev, 0);drivers/gpu/drm/tegra/dsi.c +14
-4@@ -1703,11 +1719,21 @@ static void tegra_dsi_remove(struct platform_device *pdev)tegra_mipi_free (dsi->mipi); } +static const struct tegra_dsi_config tegra20_dsi_config = { + .dsi_version =TEGRA_DSI_V0 , +}; + +static const struct tegra_dsi_config tegra114_dsi_config = { + .dsi_version =TEGRA_DSI_V1 , +}; +static const struct of_device_id tegra_dsi_of_match[] = { - { .compatible ="nvidia,tegra210-dsi" , }, - { .compatible ="nvidia,tegra132-dsi" , }, - { .compatible ="nvidia,tegra124-dsi" , }, - { .compatible ="nvidia,tegra114-dsi" , }, + { .compatible ="nvidia,tegra210-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra132-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra124-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra114-dsi" , .data = &tegra114_dsi_config }, + { .compatible ="nvidia,tegra30-dsi" , .data = &tegra20_dsi_config }, + { .compatible ="nvidia,tegra20-dsi" , .data = &tegra20_dsi_config }, { }, };MODULE_DEVICE_TABLE (of, tegra_dsi_of_match);drivers/gpu/drm/tegra/dsi.h +15
-0@@ -95,6 +95,16 @@#define DSI_TALLY_LRX (x) (((x) & 0xff) << 8)#define DSI_TALLY_HTX (x) (((x) & 0xff) << 0)#define DSI_PAD_CONTROL_0 0x4b +/* DSI V0 */ +#define DSI_PAD_CONTROL_PULLDN_ENAB (x) (((x) & 0x1) << 28) +#define DSI_PAD_CONTROL_SLEWUPADJ (x) (((x) & 0x7) << 24) +#define DSI_PAD_CONTROL_SLEWDNADJ (x) (((x) & 0x7) << 20) +#define DSI_PAD_CONTROL_PREEMP_EN (x) (((x) & 0x1) << 19) +#define DSI_PAD_CONTROL_PDIO_CLK (x) (((x) & 0x1) << 18) +#define DSI_PAD_CONTROL_PDIO (x) (((x) & 0x3) << 16) +#define DSI_PAD_CONTROL_LPUPADJ (x) (((x) & 0x3) << 14) +#define DSI_PAD_CONTROL_LPDNADJ (x) (((x) & 0x3) << 12) +/* DSI V1 */ #define DSI_PAD_CONTROL_VS1_PDIO (x) (((x) & 0xf) << 0)#define DSI_PAD_CONTROL_VS1_PDIO_CLK (1 << 8)#define DSI_PAD_CONTROL_VS1_PULLDN (x) (((x) & 0xf) << 16)@@ -140,4 +150,9 @@ enum tegra_dsi_format { TEGRA_DSI_FORMAT_24P, }; +enum tegra_dsi_version { +TEGRA_DSI_V0 , +TEGRA_DSI_V1 , +}; +#endif
PATCH RESEND v2 4/4 ARM: tegra: adjust DSI nodes for Tegra20/Tegra30
In reply to Svyatoslav RyhelAdd missing nvidia,mipi-calibrate and cells properties to DSI nodes.
--- arch/arm/boot/dts/nvidia/tegra20.dtsi | 4 ++++ arch/arm/boot/dts/nvidia/tegra30.dtsi | 8 ++++++++ 2 files changed, 12 insertions(+)
arch/arm/boot/dts/nvidia/tegra20.dtsi +4 -0
@@ -238,7 +238,11 @@ dsi@54300000 { reset-names = "dsi"; power-domains = <&pd_core>; operating-points-v2 = <&dsi_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 3>; /* DSI pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; };
arch/arm/boot/dts/nvidia/tegra30.dtsi +8 -0
@@ -343,7 +343,11 @@ dsi@54300000 { reset-names = "dsi"; power-domains = <&pd_core>; operating-points-v2 = <&dsia_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 3>; /* DSIA pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; dsi@54400000 {
@@ -356,7 +360,11 @@ dsi@54400000 { reset-names = "dsi"; power-domains = <&pd_core>; operating-points-v2 = <&dsib_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 4>; /* DSIB pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; };
Signature · 1 line
2.48.1
Re: [PATCH v2 4/4 RESEND] ARM: tegra: adjust DSI nodes for Tegra20/Tegra30
Quoted 43 lines · Svyatoslav Ryhel
Add missing nvidia,mipi-calibrate and cells properties to DSI nodes.
Signed-off-by Svyatoslav Ryhel <[email protected]>--- arch/arm/boot/dts/nvidia/tegra20.dtsi | 4 ++++ arch/arm/boot/dts/nvidia/tegra30.dtsi | 8 ++++++++ 2 files changed, 12 insertions(+)arch/arm/boot/dts/nvidia/tegra20.dtsi +4
-0@@ -238,7 +238,11 @@ dsi@54300000 { reset-names = "dsi"; power-domains = <&pd_core>; operating-points-v2 = <&dsi_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 3>; /* DSI pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; };arch/arm/boot/dts/nvidia/tegra30.dtsi +4
-0@@ -343,7 +343,11 @@ dsi@54300000 { reset-names = "dsi"; power-domains = <&pd_core>; operating-points-v2 = <&dsia_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 3>; /* DSIA pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; dsi@54400000 {@@ -356,7 +360,11 @@ dsi@54400000 { reset-names = "dsi";
arch/arm/boot/dts/nvidia/tegra30.dtsi dsi@54400000 { · line 370
Quoted 10 lines
arch/arm/boot/dts/nvidia/tegra30.dtsi +4
-0power-domains = <&pd_core>; operating-points-v2 = <&dsib_dvfs_opp_table>; + nvidia,mipi-calibrate = <&csi 4>; /* DSIB pad */ status = "disabled"; + + #address-cells = <1>; + #size-cells = <0>; }; };
Thierry Reding
AppliedRe: [PATCH v2 4/4 RESEND] ARM: tegra: adjust DSI nodes for Tegra20/Tegra30
In reply to Svyatoslav RyhelQuoted 7 lines · Svyatoslav Ryhel
Add missing nvidia,mipi-calibrate and cells properties to DSI nodes.
Signed-off-by Svyatoslav Ryhel <[email protected]>--- arch/arm/boot/dts/nvidia/tegra20.dtsi | 4 ++++ arch/arm/boot/dts/nvidia/tegra30.dtsi | 8 ++++++++ 2 files changed, 12 insertions(+)
Applied, thanks.
Thierry
Thierry Reding
AppliedRe: (subset) [PATCH v2 0/4 RESEND] gpu/drm: tegra: add DSI support for Tegra20/Tegra30
In reply to Svyatoslav RyhelFrom: Thierry Reding <[email protected]>
Quoted 11 lines · Svyatoslav Ryhel
Tegra20/Tegra30 DSI is quite similar to Tegra114+ apart MIPI calibration logic and clocks. With a few minor tweaks, existing tegra DSI driver should work on Tegra20/Tegra30 devices just fine. Tested on Motorola Atrix 4G (T20) and ASUS VivoTab RT TF600T (T30).
This patchset depends on Tegra20/Tegra30 CSI bringup since both share MIPI calibration logic. Ideally these patches should be picked after CSI bringup but they will not break anything even if picked before CSI patches.
[...]
Applied, thanks!
[4/4] ARM: tegra: adjust DSI nodes for Tegra20/Tegra30
(no commit info)
Best regards,
Signature · 1 line
Thierry Reding <[email protected]>