arm: socfpga: gen5: DM improvements
Subsystems (4)
cover arm: socfpga: gen5: DM improvements
This series ports more ad-hoc code to DM drivers (reset, clk, timer).
Simon Goldschmidt (6):
ddr: socfpga: gen5: constify altera_gen5_sdram_ops
arm: socfpga: gen5: increase SPL_SYS_MALLOC_F_LEN
timer: dw-apb: add reset handling
arm: socfpga: gen5: move initial reset handling to reset driver
arm: socfpga: gen5: add readonly clk driver
arm: socfpga: gen5: use DM_TIMER for systick
MAINTAINERS | 1 + arch/arm/dts/socfpga-common-u-boot.dtsi | 78 ++++ arch/arm/dts/socfpga.dtsi | 5 + .../dts/socfpga_cyclone5_socrates-u-boot.dtsi | 1 + .../socfpga_cyclone5_socrates_handoff.dtsi | 26 ++ arch/arm/mach-socfpga/Kconfig | 7 +- arch/arm/mach-socfpga/Makefile | 1 - arch/arm/mach-socfpga/reset_manager_gen5.c | 13 - arch/arm/mach-socfpga/spl_gen5.c | 24 +- arch/arm/mach-socfpga/timer.c | 23 -- drivers/clk/altera/Makefile | 1 + drivers/clk/altera/clk-gen5.c | 338 ++++++++++++++++++ drivers/ddr/altera/sdram_gen5.c | 2 +- drivers/timer/dw-apb-timer.c | 18 +- 14 files changed, 483 insertions(+), 55 deletions(-) create mode 100644 arch/arm/dts/socfpga_cyclone5_socrates_handoff.dtsi delete mode 100644 arch/arm/mach-socfpga/timer.c create mode 100644 drivers/clk/altera/clk-gen5.c
Signature · 1 line
2.20.1
PATCH 1/6 ddr: socfpga: gen5: constify altera_gen5_sdram_ops
Make the function pointer struct const, as it does not need to be writable. This doesn't really change anything other than moving this variable to a different section. No functional change.
---
drivers/ddr/altera/sdram_gen5.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ddr/altera/sdram_gen5.c +1 -1
@@ -626,7 +626,7 @@ static int altera_gen5_sdram_get_info(struct udevice *dev,return 0 ; } -static struct ram_ops altera_gen5_sdram_ops = { +static const struct ram_ops altera_gen5_sdram_ops = { .get_info = altera_gen5_sdram_get_info, };
Signature · 1 line
2.20.1
PATCH 2/6 arm: socfpga: gen5: increase SPL_SYS_MALLOC_F_LEN
In reply to Simon GoldschmidtIn preparation to adding more DM based drivers, increase the SPL pre-relocation heap just enough to allow those new drivers to run.
---
arch/arm/mach-socfpga/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/mach-socfpga/Kconfig +1 -1
@@ -13,7 +13,7 @@ config SPL_STACK_R_ADDR default 0x00800000 if TARGET_SOCFPGA_GEN5 config SPL_SYS_MALLOC_F_LEN - default 0x800 if TARGET_SOCFPGA_GEN5 + default 0x1100 if TARGET_SOCFPGA_GEN5 config SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION_TYPE default 0xa2
Signature · 1 line
2.20.1
PATCH 3/6 timer: dw-apb: add reset handling
In reply to Simon GoldschmidtTo use this timer on socfpga as system tick, it needs to take itself out of reset.
---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)
drivers/timer/dw-apb-timer.c +17 -1
@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h>
@@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)
@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret; + ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret); +else +reset_deassert_bulk (&priv->resets); + ret =clk_get_by_index (dev,0 , &clk);if (ret)return ret;
@@ -67,6 +75,13 @@ static int dw_apb_timer_ofdata_to_platdata(struct udevice *dev)return 0 ; } +static int dw_apb_timer_remove (struct udevice *dev) +{ +struct dw_apb_timer_priv *priv =dev_get_priv (dev); + +return reset_release_bulk (&priv->resets); +} +static const struct timer_ops dw_apb_timer_ops = { .get_count = dw_apb_timer_get_count, };
@@ -83,5 +98,6 @@ U_BOOT_DRIVER(dw_apb_timer) = { .probe = dw_apb_timer_probe, .of_match = dw_apb_timer_ids, .ofdata_to_platdata = dw_apb_timer_ofdata_to_platdata, + .remove = dw_apb_timer_remove, .priv_auto_alloc_size = sizeof(struct dw_apb_timer_priv), };
Signature · 1 line
2.20.1
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 32 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)
drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 6 lines
drivers/timer/dw-apb-timer.c +3
-0@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret; + ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);
Shouldn't this be printed by the subsystem ?
[...]
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 37 lines · Marek Vasut
Quoted 35 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret;
Quoted 5 lines
drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 3 lines
drivers/timer/dw-apb-timer.c +3
-0+ ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);Shouldn't this be printed by the subsystem ?
I don't think it's printed by the subsystem. Plus I don't know if it's a warning for all drivers? I also haven't really thought about that, as I just copied the reset handling code from another driver...
Regards,
Simon
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 44 lines · Simon Goldschmidt
Quoted 37 lines · Marek Vasut
Quoted 35 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret;Quoted 5 lines
drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 3 lines
drivers/timer/dw-apb-timer.c +3
-0+ ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);Shouldn't this be printed by the subsystem ?
Quoted 3 lines
I don't think it's printed by the subsystem. Plus I don't know if it's a warning for all drivers? I also haven't really thought about that, as I just copied the reset handling code from another driver...
Hmmmm, I guess we cannot operate without clock anyway, so it should be dev_err(). And I guess it's indeed not a subsystem print afterall.
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 46 lines · Marek Vasut
Quoted 44 lines · Simon Goldschmidt
Quoted 42 lines · Marek Vasut
Quoted 32 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 6 lines
drivers/timer/dw-apb-timer.c +3
-0@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret; + ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);Shouldn't this be printed by the subsystem ?
Quoted 6 lines
Quoted 3 lines
I don't think it's printed by the subsystem. Plus I don't know if it's a warning for all drivers? I also haven't really thought about that, as I just copied the reset handling code from another driver...
Hmmmm, I guess we cannot operate without clock anyway, so it should be
dev_err(). And I guess it's indeed not a subsystem print afterall.
Well, for gen5, it would be dev_err. I'm not sure about arria10 and stratix10 though? I think at least stratix10 doesn't have a clock driver, yet.
Regards,
Simon
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 51 lines · Simon Goldschmidt
Quoted 50 lines · Marek Vasut
Quoted 47 lines · Simon Goldschmidt
Quoted 37 lines · Marek Vasut
Quoted 35 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret;Quoted 5 lines
drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 3 lines
drivers/timer/dw-apb-timer.c +3
-0+ ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);Shouldn't this be printed by the subsystem ?
I don't think it's printed by the subsystem. Plus I don't know if it's a warning for all drivers? I also haven't really thought about that, as I just copied the reset handling code from another driver...
Quoted 5 lines
Quoted 2 lines
Hmmmm, I guess we cannot operate without clock anyway, so it should be
dev_err(). And I guess it's indeed not a subsystem print afterall.Well, for gen5, it would be
dev_err. I'm not sure about arria10 and stratix10 though? I think at least stratix10 doesn't have a clock driver, yet.
But it does have reset driver, doesn't it ?
PATCH 3/6 timer: dw-apb: add reset handling
Quoted 53 lines · Marek Vasut
Quoted 51 lines · Simon Goldschmidt
Quoted 50 lines · Marek Vasut
Quoted 47 lines · Simon Goldschmidt
Quoted 37 lines · Marek Vasut
Quoted 35 lines · Simon Goldschmidt
To use this timer on socfpga as system tick, it needs to take itself out of reset.
Signed-off-by Simon Goldschmidt <[email protected]>---
drivers/timer/dw-apb-timer.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-)drivers/timer/dw-apb-timer.c +3
-1@@ -8,6 +8,7 @@#include <common.h> #include <dm.h> #include <clk.h> +#include <reset.h> #include <timer.h> #include <asm/io.h> @@ -18,7 +19,8 @@#define DW_APB_CTRL 0x8struct dw_apb_timer_priv { -fdt_addr_t regs; +fdt_addr_t regs; +struct reset_ctl_bulk resets; };static int dw_apb_timer_get_count (struct udevice *dev,u64 *count)@@ -42,6 +44,12 @@ static int dw_apb_timer_probe(struct udevice *dev)struct clk clk;int ret;Quoted 5 lines
drivers/timer/dw-apb-timer.c static int dw_apb_timer_probe(struct udevice *dev) · line 49
Quoted 3 lines
drivers/timer/dw-apb-timer.c +3
-0+ ret =reset_get_bulk (dev, &priv->resets); +if (ret) +dev_warn (dev,"Can't get reset: %d\n" , ret);Shouldn't this be printed by the subsystem ?
I don't think it's printed by the subsystem. Plus I don't know if it's a warning for all drivers? I also haven't really thought about that, as I just copied the reset handling code from another driver...
Quoted 7 lines
Quoted 5 lines
Quoted 2 lines
Hmmmm, I guess we cannot operate without clock anyway, so it should be
dev_err(). And I guess it's indeed not a subsystem print afterall.Well, for gen5, it would be
dev_err. I'm not sure about arria10 and stratix10 though? I think at least stratix10 doesn't have a clock driver, yet.But it does have reset driver, doesn't it ?
Yes it does. You confused me with saying "we cannot operate without clock anyway", but this is about reset. So yes, you're right, dev_err() is better.
Regards,
Simon
PATCH 4/6 arm: socfpga: gen5: move initial reset handling to reset driver
In reply to Simon GoldschmidtThis moves disabling all peripherals from ad-hoc code in arch/arm to the socfpga reset driver.
To do this, DM initialization and UCLASS_RESET probing has to be done earlier in the SPL.
---
arch/arm/mach-socfpga/reset_manager_gen5.c | 13 ------------- arch/arm/mach-socfpga/spl_gen5.c | 21 +++++++++------------ 2 files changed, 9 insertions(+), 25 deletions(-)
arch/arm/mach-socfpga/reset_manager_gen5.c +0 -13
@@ -48,19 +48,6 @@ void socfpga_per_reset(u32 reset, int set)clrbits_le32 (reg,1 <<RSTMGR_RESET (reset)); } -/* -* Assert reset on every peripheral but L4WD0. -* Watchdog must be kept intact to prevent glitches -* and/or hangs. -*/ -void socfpga_per_reset_all (void ) -{ -const u32 l4wd0 =1 <<RSTMGR_RESET (SOCFPGA_RESET (L4WD0 )); - -writel (~l4wd0, &reset_manager_base->per_mod_reset); -writel (0xffffffff , &reset_manager_base->per2_mod_reset); -} -#define L3REGS_REMAP_LWHPS2FPGA_MASK 0x10#define L3REGS_REMAP_HPS2FPGA_MASK 0x08#define L3REGS_REMAP_OCRAM_MASK 0x01
arch/arm/mach-socfpga/spl_gen5.c +9 -12
@@ -84,12 +84,19 @@ void board_init_f(ulong dummy)socfpga_sdram_remap_zero ();socfpga_pl310_clear (); + ret =spl_early_init (); +if (ret) { +debug ("spl_early_init() failed: %d\n" , ret); +hang (); + } +debug ("Freezing all I/O banks\n" );/* freeze all IO banks */ sys_mgr_frzctrl_freeze_req (); -/* Put everything into reset but L4WD0. */ -socfpga_per_reset_all (); + ret =uclass_get_device (UCLASS_RESET ,0 , &dev); +if (ret) +debug ("Reset init failed: %d\n" , ret);if (!socfpga_is_booting_from_fpga ()) {/* Put FPGA bridges into reset too. */
@@ -130,16 +137,6 @@ void board_init_f(ulong dummy)debug_uart_init ();#endif - ret =spl_early_init (); -if (ret) { -debug ("spl_early_init() failed: %d\n" , ret); -hang (); - } - - ret =uclass_get_device (UCLASS_RESET ,0 , &dev); -if (ret) -debug ("Reset init failed: %d\n" , ret); -/* enable console uart printing */ preloader_console_init ();
Signature · 1 line
2.20.1
PATCH 4/6 arm: socfpga: gen5: move initial reset handling to reset driver
Quoted 2 lines · Simon Goldschmidt
This moves disabling all peripherals from ad-hoc code in
arch/armto the socfpga reset driver.
That doesn't seem like what the patch does.
Also, the WDT has to be kept enabled, is that functionality retained ?
Quoted 4 lines
To do this, DM initialization and
UCLASS_RESETprobing has to be done earlier in the SPL.Signed-off-by Simon Goldschmidt <[email protected]>
[...]
PATCH 4/6 arm: socfpga: gen5: move initial reset handling to reset driver
Quoted 6 lines · Marek Vasut
Quoted 2 lines · Simon Goldschmidt
This moves disabling all peripherals from ad-hoc code in
arch/armto the socfpga reset driver.That doesn't seem like what the patch does.
D'oh, you're right. I messed that up during development. In effect, this patch leaves periph reset like it was set by the boot rom.
I'll have to work on v2 for that. Moving this code up is nevertheless required to have the reset driver available for the clock manager etc.
Regards,
Simon
Quoted 8 lines
Also, the WDT has to be kept enabled, is that functionality retained ?
Quoted 4 lines
To do this, DM initialization and
UCLASS_RESETprobing has to be done earlier in the SPL.Signed-off-by Simon Goldschmidt <[email protected]>[...]
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
In reply to Simon GoldschmidtThis adds clk-gen5 as a readonly DM_CLK driver that can return clocks for the peripherals.
Further changes are:
- select DM_CLK for gen5 U-Boot and SPL
- add SPL tags to clock nodes in socfpga-common-u-boot.dtsi
- adjust socfpga.dtsi to provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)
---
MAINTAINERS | 1 + arch/arm/dts/socfpga-common-u-boot.dtsi | 70 ++++ arch/arm/dts/socfpga.dtsi | 5 + .../dts/socfpga_cyclone5_socrates-u-boot.dtsi | 1 + .../socfpga_cyclone5_socrates_handoff.dtsi | 26 ++ arch/arm/mach-socfpga/Kconfig | 2 + drivers/clk/altera/Makefile | 1 + drivers/clk/altera/clk-gen5.c | 338 ++++++++++++++++++ 8 files changed, 444 insertions(+) create mode 100644 arch/arm/dts/socfpga_cyclone5_socrates_handoff.dtsi create mode 100644 drivers/clk/altera/clk-gen5.c
MAINTAINERS +1 -0
@@ -94,6 +94,7 @@ M: Simon Goldschmidt <[email protected]> S: Maintainted T: git https://gitlab.denx.de/u-boot/custodians/u-boot-socfpga.git F: arch/arm/mach-socfpga/ +F: drivers/clk/altera/clk-gen5.c ARM AMLOGIC SOC SUPPORT M: Neil Armstrong <[email protected]>
arch/arm/dts/socfpga-common-u-boot.dtsi +70 -0
79 lines not on this page · raw message
arch/arm/dts/socfpga.dtsi +5 -0
@@ -332,6 +332,7 @@ compatible = "altr,socfpga-gate-clk"; clocks = <&mainclk>, <&per_base_clk>; div-reg = <0x64 4 3>; + src-reg = <0x70 0 1>; clk-gate = <0x60 2>; };
@@ -340,6 +341,7 @@ compatible = "altr,socfpga-gate-clk"; clocks = <&mainclk>, <&per_base_clk>; div-reg = <0x64 7 3>; + src-reg = <0x70 1 1>; clk-gate = <0x60 3>; };
@@ -453,6 +455,7 @@ #clock-cells = <0>; compatible = "altr,socfpga-gate-clk"; clocks = <&f2s_periph_ref_clk>, <&main_nand_sdmmc_clk>, <&per_nand_mmc_clk>; + src-reg = <0xac 0 2>; clk-gate = <0xa0 8>; clk-phase = <0 135>; };
@@ -469,6 +472,7 @@ #clock-cells = <0>; compatible = "altr,socfpga-gate-clk"; clocks = <&f2s_periph_ref_clk>, <&main_nand_sdmmc_clk>, <&per_nand_mmc_clk>; + src-reg = <0xac 2 2>; clk-gate = <0xa0 9>; };
@@ -491,6 +495,7 @@ #clock-cells = <0>; compatible = "altr,socfpga-gate-clk"; clocks = <&f2s_periph_ref_clk>, <&main_qspi_clk>, <&per_qspi_clk>; + src-reg = <0xac 4 2>; clk-gate = <0xa0 11>; };
arch/arm/dts/socfpga_cyclone5_socrates-u-boot.dtsi +1 -0
@@ -7,6 +7,7 @@ */ #include "socfpga-common-u-boot.dtsi" +#include "socfpga_cyclone5_socrates_handoff.dtsi" /{ aliases {
arch/arm/dts/socfpga_cyclone5_socrates_handoff.dtsi new file +26 -0
@@ -0,0 +1,26 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + *<auto-generated> + * This code was generated by a tool based on + * handoffs from both Qsys and Quartus. + * + * Changes to this file may be lost if + * the code is regenerated. + *</auto-generated> + */ + +&osc1 { + clock-frequency = <25000000>; +}; + +&osc2 { + clock-frequency = <25000000>; +}; + +&f2s_periph_ref_clk { + clock-frequency = <0>; +}; + +&f2s_sdram_ref_clk { + clock-frequency = <0>; +};
arch/arm/mach-socfpga/Kconfig +2 -0
@@ -53,7 +53,9 @@ config TARGET_SOCFPGA_CYCLONE5 config TARGET_SOCFPGA_GEN5 bool + select CLK select SPL_ALTERA_SDRAM + select SPL_CLK if SPL imply FPGA_SOCFPGA imply SPL_SIZE_LIMIT_SUBTRACT_GD imply SPL_SIZE_LIMIT_SUBTRACT_MALLOC
drivers/clk/altera/Makefile +1 -0
@@ -4,3 +4,4 @@ # obj-$(CONFIG_TARGET_SOCFPGA_ARRIA10) += clk-arria10.o +obj-$(CONFIG_TARGET_SOCFPGA_GEN5) += clk-gen5.o
drivers/clk/altera/clk-gen5.c new file +338 -0
338 lines not on this page · raw message
Signature · 1 line
2.20.1
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 8 lines · Simon Goldschmidt
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)
These should likely be separate patches then ?
[...]
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 5 lines · Marek Vasut
Quoted 3 lines · Simon Goldschmidt
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.
Quoted 7 lines
Quoted 5 lines
Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)These should likely be separate patches then ?
Well, in the end, yes. Especially the handoff.dtsi is required for *all* socfpga_gen5 boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.
I'll change that script and separate these patches in v2.
Regards,
Simon
Quoted 2 lines
[...]
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 12 lines · Simon Goldschmidt
Quoted 11 lines · Marek Vasut
Quoted 8 lines · Simon Goldschmidt
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)
Quoted 7 lines
Quoted 1 line
These should likely be separate patches then ?
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
You probably also want to add DM_FLAG_PRE_RELOC and possibly DM_FLAG_OS_PREPARE to tear down some clock before booting Linux.
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 16 lines · Marek Vasut
Quoted 14 lines · Simon Goldschmidt
Quoted 12 lines · Marek Vasut
Quoted 8 lines · Simon Goldschmidt
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)These should likely be separate patches then ?
Quoted 8 lines
Quoted 5 lines
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
You probably also want to add
DM_FLAG_PRE_RELOCand possiblyDM_FLAG_OS_PREPAREto tear down some clock before booting Linux.
Ack for DM_FLAG_PRE_RELOC, but I don't know about DM_FLAG_OS_PREPARE, as that could at least bring portability issues with Linux versions that can't control
the pripheral clocks (I don't know if that's implemented at all in Linux, yet).
Plus this is currently a readonly driver, so it returns the rate only and doesn't
set/enable/disable anything. I've planned that for the future, but it's not there, yet.
Regards,
Simon
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 23 lines · Simon Goldschmidt
Quoted 22 lines · Marek Vasut
Quoted 19 lines · Simon Goldschmidt
Quoted 5 lines · Marek Vasut
Quoted 3 lines · Simon Goldschmidt
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.Quoted 7 lines
Quoted 5 lines
Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)These should likely be separate patches then ?
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
Quoted 7 lines
Quoted 2 lines
You probably also want to add
DM_FLAG_PRE_RELOCand possiblyDM_FLAG_OS_PREPAREto tear down some clock before booting Linux.Ack for
DM_FLAG_PRE_RELOC, but I don't know aboutDM_FLAG_OS_PREPARE, as that could at least bring portability issues with Linux versions that can't control
the pripheral clocks (I don't know if that's implemented at all in Linux, yet).
Probably the same issues we see with reset, right ?
Quoted 4 lines
Plus this is currently a readonly driver, so it returns the rate only and doesn't
set/enable/disable anything. I've planned that for the future, but it's not there, yet.
OK
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
In reply to Simon GoldschmidtHi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 12 lines · Marek Vasut
Quoted 11 lines · Simon Goldschmidt
Quoted 8 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)
Quoted 7 lines
Quoted 1 line
These should likely be separate patches then ?
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree handoff.dtsi files. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.
So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a UCLASS_PINCTRL driver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to do it that way.
Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a UCLASS_SYSCON driver?
I'd appreciate your input on this, given the insight to the Altera system you have!
Also, this (and holidays) are delaying this series (I want to put it all together in once series), so this will have to wait until after v2019.10, I guess?
Regards,
Simon
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 29 lines · Simon Goldschmidt
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 14 lines · Marek Vasut
Quoted 13 lines · Simon Goldschmidt
Quoted 10 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)Quoted 7 lines
Quoted 1 line
These should likely be separate patches then ?
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.
Quoted 8 lines
So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to do it that way.Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?
What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Quoted 6 lines
I'd appreciate your input on this, given the insight to the Altera system you have!
Also, this (and holidays) are delaying this series (I want to put it all together in once series), so this will have to wait until after v2019.10, I guess?
We're already in RC2, so I think so.
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Am 15.08.2019 um 18:34 schrieb Marek Vasut:
Quoted 36 lines · Simon Goldschmidt
Quoted 35 lines
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 14 lines · Marek Vasut
Quoted 13 lines · Simon Goldschmidt
Quoted 10 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)Quoted 7 lines
Quoted 1 line
These should likely be separate patches then ?
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to do it that way.
Quoted 5 lines
Quoted 2 lines
Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Well, I don't know. I just need a driver that can parse its dts subtree for the config instead of loading from the qts wrapper file. Then this driver needs to be probed at the appropriate place in SPL so that the pins are initialized.
My future plan is then that such a driver could be re-probed after loading some kind of dts overlay matching an FPGA image to be programmed (as that FPGA image can contain different pin settings, e.g. when using loan IO).
I'm just not familiar enough with pinctrl drivers or syscon drivers and could need some input on which direction to take this...
Does a syscon driver for that purpose sound better?
Thanks,
Simon
Quoted 10 lines
Quoted 7 lines
I'd appreciate your input on this, given the insight to the Altera system you have!
Also, this (and holidays) are delaying this series (I want to put it all together in once series), so this will have to wait until after v2019.10, I guess?
We're already in RC2, so I think so.
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 42 lines · Simon Goldschmidt
Am 15.08.2019 um 18:34 schrieb Marek Vasut:
Quoted 41 lines · Simon Goldschmidt
Quoted 39 lines
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 16 lines · Marek Vasut
Quoted 6 lines · Simon Goldschmidt
Quoted 4 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Quoted 8 lines
Quoted 6 lines
Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)These should likely be separate patches then ?
Quoted 7 lines
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to
generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to do it that way.Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?
Quoted 7 lines
Quoted 2 lines
What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Well, I don't know. I just need a driver that can parse its dts subtree for the config instead of loading from the qts wrapper file. Then this driver needs to be probed at the appropriate place in SPL so that the pins are initialized.
Sounds more like misc driver or something along those lines.
Quoted 4 lines
My future plan is then that such a driver could be re-probed after loading some kind of dts overlay matching an FPGA image to be programmed (as that FPGA image can contain different pin settings, e.g. when using loan IO).
But then it becomes a PINMUX driver.
Quoted 4 lines
I'm just not familiar enough with pinctrl drivers or syscon drivers and could need some input on which direction to take this...
Does a syscon driver for that purpose sound better?
Isn't syscon driver for system controllers, which provide e.g. regmaps to subdevices ? I think the altera pinmux shouldn't be syscon.
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Am 15.08.2019 um 19:07 schrieb Marek Vasut:
Quoted 46 lines · Simon Goldschmidt
Quoted 45 lines
Am 15.08.2019 um 18:34 schrieb Marek Vasut:
Quoted 43 lines · Simon Goldschmidt
Quoted 31 lines
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 16 lines · Marek Vasut
Quoted 6 lines · Simon Goldschmidt
Quoted 4 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Quoted 8 lines
Quoted 6 lines
Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for now)These should likely be separate patches then ?
Quoted 7 lines
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to
generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.Quoted 8 lines
So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to do it that way.Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Quoted 6 lines
Quoted 4 lines
Well, I don't know. I just need a driver that can parse its dts subtree for the config instead of loading from the qts wrapper file. Then this driver needs to be probed at the appropriate place in SPL so that the pins are initialized.
Sounds more like misc driver or something along those lines.
Hmm, probing UCLASS_MISC in SPL to get the pinmux initializes sounds a bit strange, but that's probably OK.
Quoted 7 lines
Quoted 5 lines
My future plan is then that such a driver could be re-probed after loading some kind of dts overlay matching an FPGA image to be programmed (as that FPGA image can contain different pin settings, e.g. when using loan IO).
But then it becomes a PINMUX driver.
Well, what I'm writing *is* a writeonly driver controlling the pins. However, since we don't know anything about the iocsr part, we can't implement all the functions in 'struct pinctrl_ops' (just write the given scan chain and that's it). I think PINMUX would fit best, but see below...
Quoted 8 lines
Quoted 5 lines
I'm just not familiar enough with pinctrl drivers or syscon drivers and could need some input on which direction to take this...
Does a syscon driver for that purpose sound better?
Isn't syscon driver for system controllers, which provide e.g. regmaps to subdevices ? I think the altera pinmux shouldn't be syscon.
The thing is that 'sysmgr' already *is* a syscon driver: it provides access to various control bits (e.g. used by the ethernet driver in Linux) but also pinmux registers.
Of course, I could add "scanmgr at 0xfff02000" as a new node in the devicetree that would be the PINMUX driver which accesses the pinmux registers in sysmgr via sysycon... That would keep sysmgr's syscon behaviour working (for other drivers).
Regards,
Simon
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Quoted 56 lines · Simon Goldschmidt
Am 15.08.2019 um 19:07 schrieb Marek Vasut:
Quoted 55 lines · Simon Goldschmidt
Quoted 53 lines
Am 15.08.2019 um 18:34 schrieb Marek Vasut:
Quoted 41 lines · Simon Goldschmidt
Quoted 40 lines
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 17 lines · Marek Vasut
Quoted 14 lines · Simon Goldschmidt
Quoted 11 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for
now)Quoted 1 line
These should likely be separate patches then ?
Quoted 7 lines
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to
generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to
do it that way.Quoted 6 lines
Quoted 3 lines
Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Well, I don't know. I just need a driver that can parse its dts subtree for the config instead of loading from the qts wrapper file. Then this driver needs to be probed at the appropriate place in SPL so that the pins are initialized.
Quoted 4 lines
Quoted 1 line
Sounds more like misc driver or something along those lines.
Hmm, probing
UCLASS_MISCin SPL to get the pinmux initializes sounds a bit strange, but that's probably OK.
Well it's kinda pinmux, but not really. It's not like you can specify, in DT, how to program a pin or a range of pins either. But maybe PINCTRL uclass with some really rudimentary driver is OK.
I am not sure myself to be honest.
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
My future plan is then that such a driver could be re-probed after loading some kind of dts overlay matching an FPGA image to be programmed (as that FPGA image can contain different pin settings, e.g. when using loan IO).
Quoted 7 lines
Quoted 1 line
But then it becomes a PINMUX driver.
Well, what I'm writing *is* a writeonly driver controlling the pins. However, since we don't know anything about the iocsr part, we can't implement all the functions in 'struct
pinctrl_ops' (just write the given scan chain and that's it). I think PINMUX would fit best, but see below...
Go for it then :)
Quoted 11 lines
Quoted 7 lines
Quoted 4 lines
I'm just not familiar enough with pinctrl drivers or syscon drivers and could need some input on which direction to take this...
Does a syscon driver for that purpose sound better?
Isn't syscon driver for system controllers, which provide e.g. regmaps to subdevices ? I think the altera pinmux shouldn't be syscon.
The thing is that 'sysmgr' already *is* a syscon driver: it provides access to various control bits (e.g. used by the ethernet driver in Linux) but also pinmux registers.
Maybe you can have a pinctrl driver that is a consumer of the syscon interface ?
Quoted 8 lines
Of course, I could add "scanmgr at
0xfff02000" as a new node in the devicetree that would be the PINMUX driver which accesses the pinmux registers in sysmgr via sysycon... That would keep sysmgr's syscon behaviour working (for other drivers).Regards,
Simon
Signature · 2 lines
Best regards, Marek Vasut
PATCH 5/6 arm: socfpga: gen5: add readonly clk driver
Marek Vasut <[email protected]> schrieb am Fr., 16. Aug. 2019, 13:09:
Quoted 92 lines · Simon Goldschmidt
Quoted 56 lines
Am 15.08.2019 um 19:07 schrieb Marek Vasut:
Quoted 55 lines · Simon Goldschmidt
Quoted 53 lines
Am 15.08.2019 um 18:34 schrieb Marek Vasut:
Quoted 41 lines · Simon Goldschmidt
Quoted 40 lines
Hi Marek,
Am 24.07.2019 um 09:45 schrieb Simon Goldschmidt:
Quoted 17 lines · Marek Vasut
Quoted 14 lines · Simon Goldschmidt
Quoted 11 lines
This adds clk-gen5 as a readonly
DM_CLKdriver that can return clocks for
the peripherals.Further changes are:
- selectDM_CLKfor gen5 U-Boot and SPL
- add SPL tags to clock nodes insocfpga-common-u-boot.dtsi
- adjustsocfpga.dtsito provide missing src selection registers
- start 'handoff.dtsi' file for socrates (conatains clock speeds for
now)Quoted 1 line
These should likely be separate patches then ?
Quoted 7 lines
Well, in the end, yes. Especially the
handoff.dtsiis required for *all*socfpga_gen5boards, so I'll need to adapt the 'qts-filter.sh' script to
generate it.I'll change that script and separate these patches in v2.
I'm in the process of moving all of the 'qts' header files to devicetree
handoff.dtsifiles. CLK and DDR are already working (pinmux/iocsr not yet) - but I need to work a bit on DM memory consumption.So now I'm faced with a question: in which driver do I implement the pinmux control? From a DM point of view, it could be a
UCLASS_PINCTRLdriver in 'drivers/pinctrl', but since it's more or less read-only unless we'd get more details about the hardware, I'm a bit hesistant to
do it that way.Quoted 6 lines
Quoted 3 lines
Also, the registers are in 'sysmgr', which is handled by the standard "syscon" driver right now, so it could well get a
UCLASS_SYSCONdriver?What do we need read-only pinmux driver for ? I would expect pinmux to be more write-only, from the hardware perspective that is.
Well, I don't know. I just need a driver that can parse its dts subtree for the config instead of loading from the qts wrapper file. Then this driver needs to be probed at the appropriate place in SPL so that the pins are initialized.
Quoted 4 lines
Quoted 1 line
Sounds more like misc driver or something along those lines.
Hmm, probing
UCLASS_MISCin SPL to get the pinmux initializes sounds a bit strange, but that's probably OK.Well it's kinda pinmux, but not really. It's not like you can specify, in DT, how to program a pin or a range of pins either. But maybe PINCTRL uclass with some really rudimentary driver is OK.
I am not sure myself to be honest.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
My future plan is then that such a driver could be re-probed after loading some kind of dts overlay matching an FPGA image to be
programmed
Quoted 10 lines
Quoted 4 lines
Quoted 2 lines
(as that FPGA image can contain different pin settings, e.g. when using loan IO).
But then it becomes a PINMUX driver.
Well, what I'm writing *is* a writeonly driver controlling the pins. However, since we don't know anything about the iocsr part, we can't implement all the functions in 'struct
pinctrl_ops' (just write the given scan chain and that's it). I think PINMUX would fit best, but see below...Go for it then :)
Quoted 8 lines
Quoted 7 lines
Quoted 4 lines
I'm just not familiar enough with pinctrl drivers or syscon drivers and could need some input on which direction to take this...
Does a syscon driver for that purpose sound better?
Isn't syscon driver for system controllers, which provide e.g. regmaps to subdevices ? I think the altera pinmux shouldn't be syscon.
Quoted 7 lines
Quoted 3 lines
The thing is that 'sysmgr' already *is* a syscon driver: it provides access to various control bits (e.g. used by the ethernet driver in Linux) but also pinmux registers.
Maybe you can have a pinctrl driver that is a consumer of the syscon interface ?
Right, I'll add such a pinctrl as a driver for the scanmgr (which is not yet present in the devicetree) which will use the sysmgr pins via syscon.
Thanks for sharing your thoughts on this!
Regards,
Simon
Quoted 14 lines
Quoted 8 lines
Of course, I could add "scanmgr at
0xfff02000" as a new node in the devicetree that would be the PINMUX driver which accesses the pinmux registers in sysmgr via sysycon... That would keep sysmgr's syscon behaviour working (for other drivers).Regards,
Simon--
Best regards,
Marek Vasut
PATCH 6/6 arm: socfpga: gen5: use DM_TIMER for systick
In reply to Simon GoldschmidtThis removes the ad-hoc timer code in arch.
---
arch/arm/dts/socfpga-common-u-boot.dtsi | 8 ++++++++ arch/arm/mach-socfpga/Kconfig | 3 +++ arch/arm/mach-socfpga/Makefile | 1 - arch/arm/mach-socfpga/spl_gen5.c | 3 --- arch/arm/mach-socfpga/timer.c | 23 ----------------------- 5 files changed, 11 insertions(+), 27 deletions(-) delete mode 100644 arch/arm/mach-socfpga/timer.c
arch/arm/dts/socfpga-common-u-boot.dtsi +8 -0
@@ -5,6 +5,10 @@ * Copyright (c) 2019 Simon Goldschmidt */ /{ + chosen { + tick-timer = &timer2; + }; + soc { u-boot,dm-pre-reloc; clkmgr at ffd04000 {
@@ -16,6 +20,10 @@ }; }; +&timer2 { + u-boot,dm-pre-reloc; +}; + &rst { u-boot,dm-pre-reloc; };
arch/arm/mach-socfpga/Kconfig +3 -0
@@ -56,6 +56,9 @@ config TARGET_SOCFPGA_GEN5 select CLK select SPL_ALTERA_SDRAM select SPL_CLK if SPL + select SPL_TIMER if SPL + select TIMER + imply DESIGNWARE_APB_TIMER imply FPGA_SOCFPGA imply SPL_SIZE_LIMIT_SUBTRACT_GD imply SPL_SIZE_LIMIT_SUBTRACT_MALLOC
arch/arm/mach-socfpga/Makefile +0 -1
@@ -16,7 +16,6 @@ obj-y += misc_gen5.o obj-y += reset_manager_gen5.o obj-y += scan_manager.o obj-y += system_manager_gen5.o -obj-y += timer.o obj-y += wrap_pll_config.o obj-y += fpga_manager.o endif
arch/arm/mach-socfpga/spl_gen5.c +0 -3
@@ -103,9 +103,6 @@ void board_init_f(ulong dummy)socfpga_bridges_reset (1 ); } -socfpga_per_reset (SOCFPGA_RESET (OSC1TIMER0 ),0 ); -timer_init (); -debug ("Reconfigure Clock Manager\n" );/* reconfigure the PLLs */ if (cm_basic_init (cm_default_cfg))
arch/arm/mach-socfpga/timer.c deleted +0 -23
@@ -1,23 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0+ -/* -* Copyright (C) 2012 Altera Corporation <www.altera.com> -*/ - -#include <common.h> -#include <asm/io.h> -#include <asm/arch/timer.h> - -#define TIMER_LOAD_VAL 0xFFFFFFFF - -static const struct socfpga_timer *timer_base = (void *)CONFIG_SYS_TIMERBASE ; - -/* -* Timer initialization -*/ -int timer_init (void ) -{ -writel (TIMER_LOAD_VAL , &timer_base->load_val); -writel (TIMER_LOAD_VAL , &timer_base->curr_val); -writel (readl (&timer_base->ctrl) |0x3 , &timer_base->ctrl); -return 0 ; -}
Signature · 1 line
2.20.1
cover arm: socfpga: gen5: DM improvements
In reply to Simon GoldschmidtAm 23.07.2019 um 22:27 schrieb Simon Goldschmidt:
Quoted 1 line
This series ports more ad-hoc code to DM drivers (reset, clk, timer).
This is how far I got with DM/DTS "conversion" for now. Converting from "board/qts/*.h" to DTS is still pending since I yet failed to find a solution that allows FPGA images to define different clock or I/O requirements without requiring Linux drivers for clock or pinmux...
Converting from "board/qts/*.h" to U-Boot devicetree is probably not too hard, but reconfiguration after configuring the FPGA (e.g. from FIT) requires more work.
BTW, we're now quite short on SRAM in SPL. Next steps on this platform might require to separate "boot from MMC" and "boot from QSPI" to reduce SRAM usage.
Regards,
Simon
Quoted 29 lines
Simon Goldschmidt (6):
ddr: socfpga: gen5: constifyaltera_gen5_sdram_ops
arm: socfpga: gen5: increaseSPL_SYS_MALLOC_F_LEN
timer: dw-apb: add reset handling
arm: socfpga: gen5: move initial reset handling to reset driver
arm: socfpga: gen5: add readonly clk driver
arm: socfpga: gen5: useDM_TIMERfor systickMAINTAINERS | 1 + arch/arm/dts/socfpga-common-u-boot.dtsi | 78 ++++ arch/arm/dts/socfpga.dtsi | 5 + .../dts/socfpga_cyclone5_socrates-u-boot.dtsi | 1 + .../socfpga_cyclone5_socrates_handoff.dtsi | 26 ++ arch/arm/mach-socfpga/Kconfig | 7 +- arch/arm/mach-socfpga/Makefile | 1 - arch/arm/mach-socfpga/reset_manager_gen5.c | 13 - arch/arm/mach-socfpga/spl_gen5.c | 24 +- arch/arm/mach-socfpga/timer.c | 23 -- drivers/clk/altera/Makefile | 1 + drivers/clk/altera/clk-gen5.c | 338 ++++++++++++++++++ drivers/ddr/altera/sdram_gen5.c | 2 +- drivers/timer/dw-apb-timer.c | 18 +- 14 files changed, 483 insertions(+), 55 deletions(-)create mode 100644
arch/arm/dts/socfpga_cyclone5_socrates_handoff.dtsidelete mode 100644arch/arm/mach-socfpga/timer.c
create mode 100644drivers/clk/altera/clk-gen5.c