* [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support @ 2015-11-17 3:18 Magnus Damm 2015-11-17 3:18 ` [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups Magnus Damm ` (4 more replies) 0 siblings, 5 replies; 12+ messages in thread From: Magnus Damm @ 2015-11-17 3:18 UTC (permalink / raw) To: linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart, Magnus Damm pinctrl: sh-pfc: r8a7794: DU support [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync These patches take the r8a7794 PFC DU support code from the BSP and reworks it to fit the r8a7794 ALT board. Tested with the ALT VGA port - by default PFC is not used however enabling PFC using an incremental (yet to be posted) patch works well. It is worth noting that patch 2-4 modifies the pin groups. This means that the upstream DT ABI for PFC DU will differ compared to the unreviewed BSP code. In general it is not considered good practice to change the pin groups and break compatibility since they are part of the DT ABI. For this particular case upstream never have had PFC DU support for r8a7794, so treating the BSP bindings as experimental and migrate away seems reasonable. Signed-off-by: Magnus Damm <damm+renesas@opensource.se> --- Written against v4.4-rc1 drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 282 ++++++++++++++++++++++++++++++---- 1 file changed, 252 insertions(+), 30 deletions(-) ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm @ 2015-11-17 3:18 ` Magnus Damm 2015-11-29 21:25 ` Linus Walleij 2015-11-17 3:18 ` [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP Magnus Damm ` (3 subsequent siblings) 4 siblings, 1 reply; 12+ messages in thread From: Magnus Damm @ 2015-11-17 3:18 UTC (permalink / raw) To: linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart, Magnus Damm From: Koji Matsuoka <koji.matsuoka.xm@renesas.com> r8a7794 PFC DU support from the R-Car Gen2 v1.9.4 BSP Signed-off-by: Koji Matsuoka <koji.matsuoka.xm@renesas.com> Signed-off-by: Magnus Damm <damm+renesas@opensource.se> --- drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 170 +++++++++++++++++++++++++++++++++++ 1 file changed, 170 insertions(+) diff --git a/drivers/pinctrl/sh-pfc/pfc-r8a7794.c b/drivers/pinctrl/sh-pfc/pfc-r8a7794.c index b25851c..0af7c12 100644 --- a/drivers/pinctrl/sh-pfc/pfc-r8a7794.c +++ b/drivers/pinctrl/sh-pfc/pfc-r8a7794.c @@ -1306,6 +1306,140 @@ static const struct sh_pfc_pin pinmux_pins[] = { PINMUX_GPIO_GP_ALL(), }; +/* - DU --------------------------------------------------------------------- */ +static const unsigned int du0_rgb666_pins[] = { + /* R[7:2], G[7:2], B[7:2] */ + RCAR_GP_PIN(2, 7), RCAR_GP_PIN(2, 6), RCAR_GP_PIN(2, 5), + RCAR_GP_PIN(2, 4), RCAR_GP_PIN(2, 3), RCAR_GP_PIN(2, 2), + RCAR_GP_PIN(2, 15), RCAR_GP_PIN(2, 14), RCAR_GP_PIN(2, 13), + RCAR_GP_PIN(2, 12), RCAR_GP_PIN(2, 11), RCAR_GP_PIN(2, 10), + RCAR_GP_PIN(2, 23), RCAR_GP_PIN(2, 22), RCAR_GP_PIN(2, 21), + RCAR_GP_PIN(2, 20), RCAR_GP_PIN(2, 19), RCAR_GP_PIN(2, 18), +}; +static const unsigned int du0_rgb666_mux[] = { + DU0_DR7_MARK, DU0_DR6_MARK, DU0_DR5_MARK, DU0_DR4_MARK, + DU0_DR3_MARK, DU0_DR2_MARK, + DU0_DG7_MARK, DU0_DG6_MARK, DU0_DG5_MARK, DU0_DG4_MARK, + DU0_DG3_MARK, DU0_DG2_MARK, + DU0_DB7_MARK, DU0_DB6_MARK, DU0_DB5_MARK, DU0_DB4_MARK, + DU0_DB3_MARK, DU0_DB2_MARK, +}; +static const unsigned int du0_rgb888_pins[] = { + /* R[7:0], G[7:0], B[7:0] */ + RCAR_GP_PIN(2, 7), RCAR_GP_PIN(2, 6), RCAR_GP_PIN(2, 5), + RCAR_GP_PIN(2, 4), RCAR_GP_PIN(2, 3), RCAR_GP_PIN(2, 2), + RCAR_GP_PIN(2, 1), RCAR_GP_PIN(2, 0), + RCAR_GP_PIN(2, 15), RCAR_GP_PIN(2, 14), RCAR_GP_PIN(2, 13), + RCAR_GP_PIN(2, 12), RCAR_GP_PIN(2, 11), RCAR_GP_PIN(2, 10), + RCAR_GP_PIN(2, 9), RCAR_GP_PIN(2, 8), + RCAR_GP_PIN(2, 23), RCAR_GP_PIN(2, 22), RCAR_GP_PIN(2, 21), + RCAR_GP_PIN(2, 20), RCAR_GP_PIN(2, 19), RCAR_GP_PIN(2, 18), + RCAR_GP_PIN(2, 17), RCAR_GP_PIN(2, 16), +}; +static const unsigned int du0_rgb888_mux[] = { + DU0_DR7_MARK, DU0_DR6_MARK, DU0_DR5_MARK, DU0_DR4_MARK, + DU0_DR3_MARK, DU0_DR2_MARK, DU0_DR1_MARK, DU0_DR0_MARK, + DU0_DG7_MARK, DU0_DG6_MARK, DU0_DG5_MARK, DU0_DG4_MARK, + DU0_DG3_MARK, DU0_DG2_MARK, DU0_DG1_MARK, DU0_DG0_MARK, + DU0_DB7_MARK, DU0_DB6_MARK, DU0_DB5_MARK, DU0_DB4_MARK, + DU0_DB3_MARK, DU0_DB2_MARK, DU0_DB1_MARK, DU0_DB0_MARK, +}; +static const unsigned int du1_rgb666_pins[] = { + /* R[7:2], G[7:2], B[7:2] */ + RCAR_GP_PIN(4, 7), RCAR_GP_PIN(4, 6), RCAR_GP_PIN(4, 5), + RCAR_GP_PIN(4, 4), RCAR_GP_PIN(4, 3), RCAR_GP_PIN(4, 2), + RCAR_GP_PIN(4, 15), RCAR_GP_PIN(4, 14), RCAR_GP_PIN(4, 13), + RCAR_GP_PIN(4, 12), RCAR_GP_PIN(4, 11), RCAR_GP_PIN(4, 10), + RCAR_GP_PIN(4, 23), RCAR_GP_PIN(4, 22), RCAR_GP_PIN(4, 21), + RCAR_GP_PIN(4, 20), RCAR_GP_PIN(4, 19), RCAR_GP_PIN(4, 18), +}; +static const unsigned int du1_rgb666_mux[] = { + DU1_DR7_MARK, DU1_DR6_MARK, DU1_DR5_MARK, DU1_DR4_MARK, + DU1_DR3_MARK, DU1_DR2_MARK, + DU1_DG7_MARK, DU1_DG6_MARK, DU1_DG5_MARK, DU1_DG4_MARK, + DU1_DG3_MARK, DU1_DG2_MARK, + DU1_DB7_MARK, DU1_DB6_MARK, DU1_DB5_MARK, DU1_DB4_MARK, + DU1_DB3_MARK, DU1_DB2_MARK, +}; +static const unsigned int du1_rgb888_pins[] = { + /* R[7:0], G[7:0], B[7:0] */ + RCAR_GP_PIN(4, 7), RCAR_GP_PIN(4, 6), RCAR_GP_PIN(4, 5), + RCAR_GP_PIN(4, 4), RCAR_GP_PIN(4, 3), RCAR_GP_PIN(4, 2), + RCAR_GP_PIN(4, 1), RCAR_GP_PIN(4, 0), + RCAR_GP_PIN(4, 15), RCAR_GP_PIN(4, 14), RCAR_GP_PIN(4, 13), + RCAR_GP_PIN(4, 12), RCAR_GP_PIN(4, 11), RCAR_GP_PIN(4, 10), + RCAR_GP_PIN(4, 9), RCAR_GP_PIN(4, 8), + RCAR_GP_PIN(4, 23), RCAR_GP_PIN(4, 22), RCAR_GP_PIN(4, 21), + RCAR_GP_PIN(4, 20), RCAR_GP_PIN(4, 19), RCAR_GP_PIN(4, 18), + RCAR_GP_PIN(4, 17), RCAR_GP_PIN(4, 16), +}; +static const unsigned int du1_rgb888_mux[] = { + DU1_DR7_MARK, DU1_DR6_MARK, DU1_DR5_MARK, DU1_DR4_MARK, + DU1_DR3_MARK, DU1_DR2_MARK, DU1_DR1_MARK, DU1_DR0_MARK, + DU1_DG7_MARK, DU1_DG6_MARK, DU1_DG5_MARK, DU1_DG4_MARK, + DU1_DG3_MARK, DU1_DG2_MARK, DU1_DG1_MARK, DU1_DG0_MARK, + DU1_DB7_MARK, DU1_DB6_MARK, DU1_DB5_MARK, DU1_DB4_MARK, + DU1_DB3_MARK, DU1_DB2_MARK, DU1_DB1_MARK, DU1_DB0_MARK, +}; +static const unsigned int du0_clk_out_pins[] = { + /* CLKOUT */ + RCAR_GP_PIN(2, 25), +}; +static const unsigned int du0_clk_out_mux[] = { + DU0_DOTCLKOUT0_MARK +}; +static const unsigned int du1_clk_out_pins[] = { + /* CLKOUT */ + RCAR_GP_PIN(4, 26), +}; +static const unsigned int du1_clk_out_mux[] = { + DU1_DOTCLKOUT1_MARK +}; +static const unsigned int du0_sync_pins[] = { + /* EXVSYNC/VSYNC, EXHSYNC/HSYNC, EXDISP/EXODDF/EXCDE */ + RCAR_GP_PIN(2, 29), RCAR_GP_PIN(2, 28), RCAR_GP_PIN(2, 27), +}; +static const unsigned int du0_sync_mux[] = { + DU0_EXODDF_DU0_ODDF_DISP_CDE_MARK, + DU0_EXVSYNC_DU0_VSYNC_MARK, DU0_EXHSYNC_DU0_HSYNC_MARK +}; +static const unsigned int du1_sync_pins[] = { + /* EXVSYNC/VSYNC, EXHSYNC/HSYNC, EXDISP/EXODDF/EXCDE */ + RCAR_GP_PIN(4, 29), RCAR_GP_PIN(4, 28), RCAR_GP_PIN(4, 27), +}; +static const unsigned int du1_sync_mux[] = { + DU1_EXODDF_DU1_ODDF_DISP_CDE_MARK, + DU1_EXVSYNC_DU1_VSYNC_MARK, DU1_EXHSYNC_DU1_HSYNC_MARK +}; +static const unsigned int du0_cde_disp_pins[] = { + /* CDE DISP */ + RCAR_GP_PIN(2, 31), RCAR_GP_PIN(2, 30), +}; +static const unsigned int du0_cde_disp_mux[] = { + DU0_CDE_MARK, DU0_DISP_MARK +}; +static const unsigned int du1_cde_disp_pins[] = { + /* CDE DISP */ + RCAR_GP_PIN(4, 31), RCAR_GP_PIN(4, 30), +}; +static const unsigned int du1_cde_disp_mux[] = { + DU1_CDE_MARK, DU1_DISP_MARK +}; +static const unsigned int du0_clk_in_pins[] = { + /* CLKIN */ + RCAR_GP_PIN(2, 24), +}; +static const unsigned int du0_clk_in_mux[] = { + DU0_DOTCLKIN_MARK +}; +static const unsigned int du1_clk_in_pins[] = { + /* CLKIN */ + RCAR_GP_PIN(4, 24), +}; +static const unsigned int du1_clk_in_mux[] = { + DU1_DOTCLKIN_MARK +}; + /* - ETH -------------------------------------------------------------------- */ static const unsigned int eth_link_pins[] = { /* LINK */ @@ -2543,6 +2677,18 @@ static const unsigned int usb1_mux[] = { }; static const struct sh_pfc_pin_group pinmux_groups[] = { + SH_PFC_PIN_GROUP(du0_rgb666), + SH_PFC_PIN_GROUP(du0_rgb888), + SH_PFC_PIN_GROUP(du1_rgb666), + SH_PFC_PIN_GROUP(du1_rgb888), + SH_PFC_PIN_GROUP(du0_clk_out), + SH_PFC_PIN_GROUP(du1_clk_out), + SH_PFC_PIN_GROUP(du0_sync), + SH_PFC_PIN_GROUP(du1_sync), + SH_PFC_PIN_GROUP(du0_cde_disp), + SH_PFC_PIN_GROUP(du1_cde_disp), + SH_PFC_PIN_GROUP(du0_clk_in), + SH_PFC_PIN_GROUP(du1_clk_in), SH_PFC_PIN_GROUP(eth_link), SH_PFC_PIN_GROUP(eth_magic), SH_PFC_PIN_GROUP(eth_mdio), @@ -2714,6 +2860,27 @@ static const struct sh_pfc_pin_group pinmux_groups[] = { SH_PFC_PIN_GROUP(usb1), }; +static const char * const du_groups[] = { + "du0_rgb666", + "du0_rgb888", + "du1_rgb666", + "du1_rgb888", + "du0_clk_out", + "du1_clk_out", + "du0_sync", + "du1_sync", + "du0_cde_disp", + "du1_cde_disp", +}; + +static const char * const du0_groups[] = { + "du0_clk_in", +}; + +static const char * const du1_groups[] = { + "du1_clk_in", +}; + static const char * const eth_groups[] = { "eth_link", "eth_magic", @@ -2985,6 +3152,9 @@ static const char * const usb1_groups[] = { }; static const struct sh_pfc_function pinmux_functions[] = { + SH_PFC_FUNCTION(du), + SH_PFC_FUNCTION(du0), + SH_PFC_FUNCTION(du1), SH_PFC_FUNCTION(eth), SH_PFC_FUNCTION(hscif0), SH_PFC_FUNCTION(hscif1), ^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups 2015-11-17 3:18 ` [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups Magnus Damm @ 2015-11-29 21:25 ` Linus Walleij 2015-11-30 8:53 ` Geert Uytterhoeven 0 siblings, 1 reply; 12+ messages in thread From: Linus Walleij @ 2015-11-29 21:25 UTC (permalink / raw) To: Magnus Damm, geert Cc: linux-sh@vger.kernel.org, linux-kernel@vger.kernel.org, koji.matsuoka.xm, linux-gpio@vger.kernel.org, Simon Horman, Laurent Pinchart On Tue, Nov 17, 2015 at 4:18 AM, Magnus Damm <magnus.damm@gmail.com> wrote: > From: Koji Matsuoka <koji.matsuoka.xm@renesas.com> > > r8a7794 PFC DU support from the R-Car Gen2 v1.9.4 BSP > > Signed-off-by: Koji Matsuoka <koji.matsuoka.xm@renesas.com> > Signed-off-by: Magnus Damm <damm+renesas@opensource.se> Acked-by for all from my side, Geert, I expect that you queue these with the rest of the SH PFC stuff. Yours, Linus Walleij ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups 2015-11-29 21:25 ` Linus Walleij @ 2015-11-30 8:53 ` Geert Uytterhoeven 0 siblings, 0 replies; 12+ messages in thread From: Geert Uytterhoeven @ 2015-11-30 8:53 UTC (permalink / raw) To: Linus Walleij Cc: Magnus Damm, linux-sh@vger.kernel.org, linux-kernel@vger.kernel.org, Koji Matsuoka, linux-gpio@vger.kernel.org, Simon Horman, Laurent Pinchart Hi Linus, On Sun, Nov 29, 2015 at 10:25 PM, Linus Walleij <linus.walleij@linaro.org> wrote: >> From: Koji Matsuoka <koji.matsuoka.xm@renesas.com> >> >> r8a7794 PFC DU support from the R-Car Gen2 v1.9.4 BSP >> >> Signed-off-by: Koji Matsuoka <koji.matsuoka.xm@renesas.com> >> Signed-off-by: Magnus Damm <damm+renesas@opensource.se> > > Acked-by for all from my side, Geert, I expect that you > queue these with the rest of the SH PFC stuff. Thank you. I will postpone until HDMI has been tested on ALT, cfr. the discussion in response to the cover letter. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org 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 ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm 2015-11-17 3:18 ` [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups Magnus Damm @ 2015-11-17 3:18 ` Magnus Damm 2015-11-17 3:19 ` [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals Magnus Damm ` (2 subsequent siblings) 4 siblings, 0 replies; 12+ messages in thread From: Magnus Damm @ 2015-11-17 3:18 UTC (permalink / raw) To: linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart, Magnus Damm From: Magnus Damm <damm+renesas@opensource.se> Separate DU CDE and DISP signals to let the r8a7794 ALT board that lacks CDE handle the DISP signal by itself. The groups "du0_cde_disp" and "du1_cde_disp" are replaced by "du0_cde", "du1_cde", "du0_disp" and "du1_disp". Signed-off-by: Magnus Damm <damm+renesas@opensource.se> --- drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 50 +++++++++++++++++++++++----------- 1 file changed, 34 insertions(+), 16 deletions(-) --- 0008/drivers/pinctrl/sh-pfc/pfc-r8a7794.c +++ work/drivers/pinctrl/sh-pfc/pfc-r8a7794.c 2015-11-16 15:04:00.760513000 +0900 @@ -1616,19 +1616,33 @@ static const unsigned int du1_sync_mux[] DU1_EXODDF_DU1_ODDF_DISP_CDE_MARK, DU1_EXVSYNC_DU1_VSYNC_MARK, DU1_EXHSYNC_DU1_HSYNC_MARK }; -static const unsigned int du0_cde_disp_pins[] = { - /* CDE DISP */ - RCAR_GP_PIN(2, 31), RCAR_GP_PIN(2, 30), -}; -static const unsigned int du0_cde_disp_mux[] = { - DU0_CDE_MARK, DU0_DISP_MARK -}; -static const unsigned int du1_cde_disp_pins[] = { - /* CDE DISP */ - RCAR_GP_PIN(4, 31), RCAR_GP_PIN(4, 30), +static const unsigned int du0_cde_pins[] = { + /* CDE */ + RCAR_GP_PIN(2, 31), }; -static const unsigned int du1_cde_disp_mux[] = { - DU1_CDE_MARK, DU1_DISP_MARK +static const unsigned int du0_cde_mux[] = { + DU0_CDE_MARK, +}; +static const unsigned int du1_cde_pins[] = { + /* CDE */ + RCAR_GP_PIN(4, 31), +}; +static const unsigned int du1_cde_mux[] = { + DU1_CDE_MARK +}; +static const unsigned int du0_disp_pins[] = { + /* DISP */ + RCAR_GP_PIN(2, 30), +}; +static const unsigned int du0_disp_mux[] = { + DU0_DISP_MARK +}; +static const unsigned int du1_disp_pins[] = { + /* DISP */ + RCAR_GP_PIN(4, 30), +}; +static const unsigned int du1_disp_mux[] = { + DU1_DISP_MARK }; static const unsigned int du0_clk_in_pins[] = { /* CLKIN */ @@ -2932,8 +2946,10 @@ static const struct sh_pfc_pin_group pin SH_PFC_PIN_GROUP(du1_clk_out), SH_PFC_PIN_GROUP(du0_sync), SH_PFC_PIN_GROUP(du1_sync), - SH_PFC_PIN_GROUP(du0_cde_disp), - SH_PFC_PIN_GROUP(du1_cde_disp), + SH_PFC_PIN_GROUP(du0_cde), + SH_PFC_PIN_GROUP(du1_cde), + SH_PFC_PIN_GROUP(du0_disp), + SH_PFC_PIN_GROUP(du1_disp), SH_PFC_PIN_GROUP(du0_clk_in), SH_PFC_PIN_GROUP(du1_clk_in), SH_PFC_PIN_GROUP(eth_link), @@ -3122,8 +3138,10 @@ static const char * const du_groups[] = "du1_clk_out", "du0_sync", "du1_sync", - "du0_cde_disp", - "du1_cde_disp", + "du0_cde", + "du1_cde", + "du0_disp", + "du1_disp", }; static const char * const du0_groups[] = { ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm 2015-11-17 3:18 ` [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups Magnus Damm 2015-11-17 3:18 ` [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP Magnus Damm @ 2015-11-17 3:19 ` Magnus Damm 2015-12-21 22:01 ` Sergei Shtylyov 2015-11-17 3:19 ` [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync Magnus Damm 2015-11-20 2:46 ` [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Laurent Pinchart 4 siblings, 1 reply; 12+ messages in thread From: Magnus Damm @ 2015-11-17 3:19 UTC (permalink / raw) To: linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart, Magnus Damm From: Magnus Damm <damm+renesas@opensource.se> Add missing r8a7794 DU dot clock output signals and in particular the DU1_DOTCLKOUT0 signal on GP4_25 which is needed by DU1 on the r8a7794 ALT board. The groups "du0_clk_out" and "du1_clk_out" are replaced by "du0_dotclkout0", "du0_dotclkout1", "du1_dotclkout0" and "du1_dotclkout1". Signed-off-by: Magnus Damm <damm+renesas@opensource.se> --- drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 34 ++++++++++++++++++++++++++-------- 1 file changed, 26 insertions(+), 8 deletions(-) --- 0014/drivers/pinctrl/sh-pfc/pfc-r8a7794.c +++ work/drivers/pinctrl/sh-pfc/pfc-r8a7794.c 2015-11-16 22:08:28.230513000 +0900 @@ -1586,18 +1586,32 @@ static const unsigned int du1_rgb888_mux DU1_DB7_MARK, DU1_DB6_MARK, DU1_DB5_MARK, DU1_DB4_MARK, DU1_DB3_MARK, DU1_DB2_MARK, DU1_DB1_MARK, DU1_DB0_MARK, }; -static const unsigned int du0_clk_out_pins[] = { +static const unsigned int du0_clk0_out_pins[] = { /* CLKOUT */ RCAR_GP_PIN(2, 25), }; -static const unsigned int du0_clk_out_mux[] = { +static const unsigned int du0_clk0_out_mux[] = { DU0_DOTCLKOUT0_MARK }; -static const unsigned int du1_clk_out_pins[] = { +static const unsigned int du0_clk1_out_pins[] = { + /* CLKOUT */ + RCAR_GP_PIN(2, 26), +}; +static const unsigned int du0_clk1_out_mux[] = { + DU0_DOTCLKOUT1_MARK +}; +static const unsigned int du1_clk0_out_pins[] = { + /* CLKOUT */ + RCAR_GP_PIN(4, 25), +}; +static const unsigned int du1_clk0_out_mux[] = { + DU1_DOTCLKOUT0_MARK +}; +static const unsigned int du1_clk1_out_pins[] = { /* CLKOUT */ RCAR_GP_PIN(4, 26), }; -static const unsigned int du1_clk_out_mux[] = { +static const unsigned int du1_clk1_out_mux[] = { DU1_DOTCLKOUT1_MARK }; static const unsigned int du0_sync_pins[] = { @@ -2942,8 +2956,10 @@ static const struct sh_pfc_pin_group pin SH_PFC_PIN_GROUP(du0_rgb888), SH_PFC_PIN_GROUP(du1_rgb666), SH_PFC_PIN_GROUP(du1_rgb888), - SH_PFC_PIN_GROUP(du0_clk_out), - SH_PFC_PIN_GROUP(du1_clk_out), + SH_PFC_PIN_GROUP(du0_clk0_out), + SH_PFC_PIN_GROUP(du0_clk1_out), + SH_PFC_PIN_GROUP(du1_clk0_out), + SH_PFC_PIN_GROUP(du1_clk1_out), SH_PFC_PIN_GROUP(du0_sync), SH_PFC_PIN_GROUP(du1_sync), SH_PFC_PIN_GROUP(du0_cde), @@ -3134,8 +3150,10 @@ static const char * const du_groups[] = "du0_rgb888", "du1_rgb666", "du1_rgb888", - "du0_clk_out", - "du1_clk_out", + "du0_dotclkout0", + "du0_dotclkout1", + "du1_dotclkout0", + "du1_dotclkout1", "du0_sync", "du1_sync", "du0_cde", ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals 2015-11-17 3:19 ` [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals Magnus Damm @ 2015-12-21 22:01 ` Sergei Shtylyov 0 siblings, 0 replies; 12+ messages in thread From: Sergei Shtylyov @ 2015-12-21 22:01 UTC (permalink / raw) To: Magnus Damm, linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart Hello. On 11/17/2015 06:19 AM, Magnus Damm wrote: > From: Magnus Damm <damm+renesas@opensource.se> > > Add missing r8a7794 DU dot clock output signals and in > particular the DU1_DOTCLKOUT0 signal on GP4_25 which is > needed by DU1 on the r8a7794 ALT board. > > The groups "du0_clk_out" and "du1_clk_out" are replaced by > "du0_dotclkout0", "du0_dotclkout1", "du1_dotclkout0" and > "du1_dotclkout1". > > Signed-off-by: Magnus Damm <damm+renesas@opensource.se> > --- > > drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 34 ++++++++++++++++++++++++++-------- > 1 file changed, 26 insertions(+), 8 deletions(-) > > --- 0014/drivers/pinctrl/sh-pfc/pfc-r8a7794.c > +++ work/drivers/pinctrl/sh-pfc/pfc-r8a7794.c 2015-11-16 22:08:28.230513000 +0900 [...] > @@ -2942,8 +2956,10 @@ static const struct sh_pfc_pin_group pin > SH_PFC_PIN_GROUP(du0_rgb888), > SH_PFC_PIN_GROUP(du1_rgb666), > SH_PFC_PIN_GROUP(du1_rgb888), > - SH_PFC_PIN_GROUP(du0_clk_out), > - SH_PFC_PIN_GROUP(du1_clk_out), > + SH_PFC_PIN_GROUP(du0_clk0_out), > + SH_PFC_PIN_GROUP(du0_clk1_out), > + SH_PFC_PIN_GROUP(du1_clk0_out), > + SH_PFC_PIN_GROUP(du1_clk1_out), Hm, this kinda contradicts your chngnelog, no? > SH_PFC_PIN_GROUP(du0_sync), > SH_PFC_PIN_GROUP(du1_sync), > SH_PFC_PIN_GROUP(du0_cde), > @@ -3134,8 +3150,10 @@ static const char * const du_groups[] = > "du0_rgb888", > "du1_rgb666", > "du1_rgb888", > - "du0_clk_out", > - "du1_clk_out", > + "du0_dotclkout0", > + "du0_dotclkout1", > + "du1_dotclkout0", > + "du1_dotclkout1", > "du0_sync", > "du1_sync", > "du0_cde", I'm getting the following on the SILK board with this patch: [drm] Initialized drm 1.1.0 20060810 sh-pfc e6060000.pin-controller: does not have pin group du1_dotclkout0 sh-pfc e6060000.pin-controller: invalid group du1_dotclkout0 in map table [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [drm] No driver support for vblank timestamp query. rcar-du feb00000.display: fb0: frame buffer device [drm] Initialized rcar-du 1.0.0 20130110 on minor 0 MBR, Sergei ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm ` (2 preceding siblings ...) 2015-11-17 3:19 ` [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals Magnus Damm @ 2015-11-17 3:19 ` Magnus Damm 2015-11-20 2:46 ` [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Laurent Pinchart 4 siblings, 0 replies; 12+ messages in thread From: Magnus Damm @ 2015-11-17 3:19 UTC (permalink / raw) To: linux-sh Cc: linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert, laurent.pinchart, Magnus Damm From: Magnus Damm <damm+renesas@opensource.se> Break out the r8a7794 DU ODDF (EXDISP/EXODDF/EXDE) signal from the sync group into a separate unit. This makes the sync group fit with the r8a7794 ALT board that only uses HSYNC and VSYNC signals. This makes the r8a7794 PFC similar to the existing r8a7791 PFC DU implementation that also groups together HSYNC and VSYNC without ODDF. The groups "du0_oddf" and "du1_oddf" are added and the "du0_sync" and "du1_sync" groups are modified. Signed-off-by: Magnus Damm <damm+renesas@opensource.se> --- drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) --- 0016/drivers/pinctrl/sh-pfc/pfc-r8a7794.c +++ work/drivers/pinctrl/sh-pfc/pfc-r8a7794.c 2015-11-16 22:27:10.190513000 +0900 @@ -1615,21 +1615,33 @@ static const unsigned int du1_clk1_out_m DU1_DOTCLKOUT1_MARK }; static const unsigned int du0_sync_pins[] = { - /* EXVSYNC/VSYNC, EXHSYNC/HSYNC, EXDISP/EXODDF/EXCDE */ - RCAR_GP_PIN(2, 29), RCAR_GP_PIN(2, 28), RCAR_GP_PIN(2, 27), + /* EXVSYNC/VSYNC, EXHSYNC/HSYNC */ + RCAR_GP_PIN(2, 28), RCAR_GP_PIN(2, 27), }; static const unsigned int du0_sync_mux[] = { - DU0_EXODDF_DU0_ODDF_DISP_CDE_MARK, DU0_EXVSYNC_DU0_VSYNC_MARK, DU0_EXHSYNC_DU0_HSYNC_MARK }; static const unsigned int du1_sync_pins[] = { - /* EXVSYNC/VSYNC, EXHSYNC/HSYNC, EXDISP/EXODDF/EXCDE */ - RCAR_GP_PIN(4, 29), RCAR_GP_PIN(4, 28), RCAR_GP_PIN(4, 27), + /* EXVSYNC/VSYNC, EXHSYNC/HSYNC */ + RCAR_GP_PIN(4, 28), RCAR_GP_PIN(4, 27), }; static const unsigned int du1_sync_mux[] = { - DU1_EXODDF_DU1_ODDF_DISP_CDE_MARK, DU1_EXVSYNC_DU1_VSYNC_MARK, DU1_EXHSYNC_DU1_HSYNC_MARK }; +static const unsigned int du0_oddf_pins[] = { + /* EXDISP/EXODDF/EXCDE */ + RCAR_GP_PIN(2, 29), +}; +static const unsigned int du0_oddf_mux[] = { + DU0_EXODDF_DU0_ODDF_DISP_CDE_MARK, +}; +static const unsigned int du1_oddf_pins[] = { + /* EXDISP/EXODDF/EXCDE */ + RCAR_GP_PIN(4, 29), +}; +static const unsigned int du1_oddf_mux[] = { + DU1_EXODDF_DU1_ODDF_DISP_CDE_MARK, +}; static const unsigned int du0_cde_pins[] = { /* CDE */ RCAR_GP_PIN(2, 31), @@ -2962,6 +2974,8 @@ static const struct sh_pfc_pin_group pin SH_PFC_PIN_GROUP(du1_clk1_out), SH_PFC_PIN_GROUP(du0_sync), SH_PFC_PIN_GROUP(du1_sync), + SH_PFC_PIN_GROUP(du0_oddf), + SH_PFC_PIN_GROUP(du1_oddf), SH_PFC_PIN_GROUP(du0_cde), SH_PFC_PIN_GROUP(du1_cde), SH_PFC_PIN_GROUP(du0_disp), @@ -3156,6 +3170,8 @@ static const char * const du_groups[] = "du1_dotclkout1", "du0_sync", "du1_sync", + "du0_oddf", + "du1_oddf", "du0_cde", "du1_cde", "du0_disp", ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm ` (3 preceding siblings ...) 2015-11-17 3:19 ` [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync Magnus Damm @ 2015-11-20 2:46 ` Laurent Pinchart 2015-11-20 7:16 ` Magnus Damm 2015-11-25 0:16 ` Simon Horman 4 siblings, 2 replies; 12+ messages in thread From: Laurent Pinchart @ 2015-11-20 2:46 UTC (permalink / raw) To: Magnus Damm Cc: linux-sh, linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, horms, geert Hi Magnus, Thank you for the patch. On Tuesday 17 November 2015 12:18:32 Magnus Damm wrote: > pinctrl: sh-pfc: r8a7794: DU support > > [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups > [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP > [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals > [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync > > These patches take the r8a7794 PFC DU support code from the BSP > and reworks it to fit the r8a7794 ALT board. Tested with the ALT > VGA port - by default PFC is not used however enabling PFC using > an incremental (yet to be posted) patch works well. > > It is worth noting that patch 2-4 modifies the pin groups. This > means that the upstream DT ABI for PFC DU will differ compared > to the unreviewed BSP code. > > In general it is not considered good practice to change the pin > groups and break compatibility since they are part of the DT ABI. > > For this particular case upstream never have had PFC DU support > for r8a7794, so treating the BSP bindings as experimental and > migrate away seems reasonable. If we start considering DT bindings that never went upstream as stable we'll have a big problem. I mean even bigger than the upstream DT bindings stability problem :-) > Signed-off-by: Magnus Damm <damm+renesas@opensource.se> Wouldn't it make sense to merge the 4 patches together ? > --- > > Written against v4.4-rc1 > > drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 282 +++++++++++++++++++++++++++---- > 1 file changed, 252 insertions(+), 30 deletions(-) -- Regards, Laurent Pinchart ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support 2015-11-20 2:46 ` [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Laurent Pinchart @ 2015-11-20 7:16 ` Magnus Damm 2015-11-20 14:53 ` Laurent Pinchart 2015-11-25 0:16 ` Simon Horman 1 sibling, 1 reply; 12+ messages in thread From: Magnus Damm @ 2015-11-20 7:16 UTC (permalink / raw) To: Laurent Pinchart Cc: SH-Linux, Linus Walleij, linux-kernel, Koji Matsuoka, linux-gpio, Simon Horman [Horms], geert Hi Laurent, On Fri, Nov 20, 2015 at 11:46 AM, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote: > Hi Magnus, > > Thank you for the patch. > > On Tuesday 17 November 2015 12:18:32 Magnus Damm wrote: >> pinctrl: sh-pfc: r8a7794: DU support >> >> [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups >> [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP >> [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals >> [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync >> >> These patches take the r8a7794 PFC DU support code from the BSP >> and reworks it to fit the r8a7794 ALT board. Tested with the ALT >> VGA port - by default PFC is not used however enabling PFC using >> an incremental (yet to be posted) patch works well. >> >> It is worth noting that patch 2-4 modifies the pin groups. This >> means that the upstream DT ABI for PFC DU will differ compared >> to the unreviewed BSP code. >> >> In general it is not considered good practice to change the pin >> groups and break compatibility since they are part of the DT ABI. >> >> For this particular case upstream never have had PFC DU support >> for r8a7794, so treating the BSP bindings as experimental and >> migrate away seems reasonable. > > If we start considering DT bindings that never went upstream as stable we'll > have a big problem. I mean even bigger than the upstream DT bindings stability > problem :-) I'm not saying that local DT hacks should be considered stable, more that it as usual makes sense to follow upstream first with proper DT review process early on. >> Signed-off-by: Magnus Damm <damm+renesas@opensource.se> > > Wouldn't it make sense to merge the 4 patches together ? Yeah, I guess so. My feeling is also that it would be good to verify HDMI on ALT before commiting to DT bindings. Right now only one DU channel is tested. Cheers, / magnus ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support 2015-11-20 7:16 ` Magnus Damm @ 2015-11-20 14:53 ` Laurent Pinchart 0 siblings, 0 replies; 12+ messages in thread From: Laurent Pinchart @ 2015-11-20 14:53 UTC (permalink / raw) To: Magnus Damm Cc: SH-Linux, Linus Walleij, linux-kernel, Koji Matsuoka, linux-gpio, Simon Horman [Horms], geert Hi Magnus, On Friday 20 November 2015 16:16:15 Magnus Damm wrote: > On Fri, Nov 20, 2015 at 11:46 AM, Laurent Pinchart wrote: > > On Tuesday 17 November 2015 12:18:32 Magnus Damm wrote: > >> pinctrl: sh-pfc: r8a7794: DU support > >> > >> [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups > >> [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP > >> [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals > >> [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync > >> > >> These patches take the r8a7794 PFC DU support code from the BSP > >> and reworks it to fit the r8a7794 ALT board. Tested with the ALT > >> VGA port - by default PFC is not used however enabling PFC using > >> an incremental (yet to be posted) patch works well. > >> > >> It is worth noting that patch 2-4 modifies the pin groups. This > >> means that the upstream DT ABI for PFC DU will differ compared > >> to the unreviewed BSP code. > >> > >> In general it is not considered good practice to change the pin > >> groups and break compatibility since they are part of the DT ABI. > >> > >> For this particular case upstream never have had PFC DU support > >> for r8a7794, so treating the BSP bindings as experimental and > >> migrate away seems reasonable. > > > > If we start considering DT bindings that never went upstream as stable > > we'll have a big problem. I mean even bigger than the upstream DT > > bindings stability problem :-) > > I'm not saying that local DT hacks should be considered stable, more > that it as usual makes sense to follow upstream first with proper DT > review process early on. > > >> Signed-off-by: Magnus Damm <damm+renesas@opensource.se> > > > > Wouldn't it make sense to merge the 4 patches together ? > > Yeah, I guess so. My feeling is also that it would be good to verify > HDMI on ALT before commiting to DT bindings. Right now only one DU > channel is tested. I agree with that. -- Regards, Laurent Pinchart ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support 2015-11-20 2:46 ` [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Laurent Pinchart 2015-11-20 7:16 ` Magnus Damm @ 2015-11-25 0:16 ` Simon Horman 1 sibling, 0 replies; 12+ messages in thread From: Simon Horman @ 2015-11-25 0:16 UTC (permalink / raw) To: Laurent Pinchart Cc: Magnus Damm, linux-sh, linus.walleij, linux-kernel, koji.matsuoka.xm, linux-gpio, geert On Fri, Nov 20, 2015 at 04:46:50AM +0200, Laurent Pinchart wrote: > Hi Magnus, > > Thank you for the patch. > > On Tuesday 17 November 2015 12:18:32 Magnus Damm wrote: > > pinctrl: sh-pfc: r8a7794: DU support > > > > [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups > > [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP > > [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals > > [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync > > > > These patches take the r8a7794 PFC DU support code from the BSP > > and reworks it to fit the r8a7794 ALT board. Tested with the ALT > > VGA port - by default PFC is not used however enabling PFC using > > an incremental (yet to be posted) patch works well. > > > > It is worth noting that patch 2-4 modifies the pin groups. This > > means that the upstream DT ABI for PFC DU will differ compared > > to the unreviewed BSP code. > > > > In general it is not considered good practice to change the pin > > groups and break compatibility since they are part of the DT ABI. > > > > For this particular case upstream never have had PFC DU support > > for r8a7794, so treating the BSP bindings as experimental and > > migrate away seems reasonable. > > If we start considering DT bindings that never went upstream as stable > we'll have a big problem. I mean even bigger than the upstream DT > bindings stability problem :-) FWIW, I agree. > > Signed-off-by: Magnus Damm <damm+renesas@opensource.se> > > Wouldn't it make sense to merge the 4 patches together ? > > > --- > > > > Written against v4.4-rc1 > > > > drivers/pinctrl/sh-pfc/pfc-r8a7794.c | 282 +++++++++++++++++++++++++++---- > > 1 file changed, 252 insertions(+), 30 deletions(-) > > -- > Regards, > > Laurent Pinchart > ^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2015-12-21 22:01 UTC | newest] Thread overview: 12+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2015-11-17 3:18 [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Magnus Damm 2015-11-17 3:18 ` [PATCH 01/04] pinctrl: sh-pfc: r8a7794: Add DU pin groups Magnus Damm 2015-11-29 21:25 ` Linus Walleij 2015-11-30 8:53 ` Geert Uytterhoeven 2015-11-17 3:18 ` [PATCH 02/04] pinctrl: sh-pfc: r8a7794: Separate DU CDE and DISP Magnus Damm 2015-11-17 3:19 ` [PATCH 03/04] pinctrl: sh-pfc: r8a7794: Add missing dot clock signals Magnus Damm 2015-12-21 22:01 ` Sergei Shtylyov 2015-11-17 3:19 ` [PATCH 04/04] pinctrl: sh-pfc: r8a7794: Break out ODDF from sync Magnus Damm 2015-11-20 2:46 ` [PATCH 00/04] pinctrl: sh-pfc: r8a7794: DU support Laurent Pinchart 2015-11-20 7:16 ` Magnus Damm 2015-11-20 14:53 ` Laurent Pinchart 2015-11-25 0:16 ` Simon Horman
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).