sunxi: Allwinner H5 and OrangePi PC2 support
Subsystems (8)
v3 cover sunxi: Allwinner H5 and OrangePi PC2 support
Hi,
some small changes as requested by reviewers (many thanks for that!)
Maxime, Jagan,
given I get some ACKs or RBs on the remaining patches, is this good to go for 2017.03 still?
Cheers,
Andre.
--------------------------
This series introduces support for the Allwinner H5 SoC with four Cortex-A53 cores. The SoC's peripherals are very similar to the H3, although the cores and the BROM/SRAM layout resembles the A64. The first 8 patches contain some fixes and refactoring, to make code sharing between the three mentioned SoCs easier.
Patch 09/13 adds support for the H5 DRAM controller, by extending the already existing combined H3/A64 DRAM code.
Patch 10/13 renames the existing CONFIG_MACH_SUN8I_H3 config symbol to let it be used by all peripheral code that can be shared between the H3 and H5. Patch 11/13 introduces the H5 SoC config option into Kconfig, which defines this shared symbol as well.
Patch 12/13 adds an easy device tree, which actually uses the H3 .dtsi and overwrites nodes which are different. This is good enough for U-Boot, the DT will be changed anyway once we get the DT merged into the Linux kernel.
The final patch then adds the defconfig for the OrangePi PC2 board. Since this board comes with soldered SPI flash, we enable support for it in the SPL. This has been tested by writing the SPI flash with some special sunxi-fel version. The BROM loaded and executed the SPL, which in turn loaded and executed U-Boot proper. Both parts are 64-bit only for now.
Ethernet support is enabled, but fails at the moment since the EMAC driver does not support setting a GPIO to enable the external Gigabit PHY.
At the moment this build suffers from the same problem as the A64: the ATF is missing, so Linux won't boot easily. With the RFC version of the SPL FIT extension series on the list, applied on top of this one this should now be solved.
This series is on top of origin/master.
Please have a look and let me know your opinion!
Cheers,
Andre.
Changelog v2..v3:
- Adding ACKs and RBs
- remove not needed CONFIG_MACH_SUN50I_H5_64
- move ARM_CORTEX_CPU_IS_UP to arch/arm/Kconfig
Changelog v1..v2:
- Adding Maxime's ACKs
- new patch 2 and 3 to rename CONFIG_TIMER_CLK_FREQ to COUNTER_FREQUENCY
- renaming CONFIG symbol name for single core SoCs
- fixing one checkpatch issue (some are left, though)
- mentioning methodology and reason for missing data in DRAM driver
- renaming shared config symbol to CONFIG_MACH_SUNXI_H3_H5
- updated to lastest origin/master
Andre Przywara (13):
sunxi: fix ACTLR.SMP assembly routine
ARM: rename CONFIG_TIMER_CLK_FREQ to COUNTER_FREQUENCY
fsl: ls102x: remove redundant GENERIC_TIMER_CLK
sunxi: simplify ACTLR.SMP bit set #ifdef
sunxi: configs: merge sun9i and sun50i SPL memory definitions
sunxi: Kconfig: introduce CONFIG_SUNXI_HIGH_SRAM
sunxi: provide ARMv8 mem_map for every ARM64 board
SPI: SPL: sunxi: fix 64-bit build
sunxi: DRAM: add Allwinner H5 support
sunxi: prepare for sharing MACH_SUN8I_H3 config symbol
sunxi: introduce Allwinner H5 config option
sunxi: dts: add basic OrangePi PC 2 device tree file
sunxi: configs: add basic OrangePi PC 2 defconfig
arch/arm/Kconfig | 4 + arch/arm/cpu/armv7/ls102xa/psci.S | 2 +- arch/arm/cpu/armv7/ls102xa/timer.c | 2 +- arch/arm/cpu/armv7/nonsec_virt.S | 4 +- arch/arm/cpu/armv7/sunxi/psci.c | 2 +- arch/arm/dts/Makefile | 2 + arch/arm/dts/sun50i-h5-orangepi-pc2.dts | 147 ++++++++++++++++++++++++++ arch/arm/include/asm/arch-sunxi/clock_sun6i.h | 4 +- arch/arm/include/asm/arch-sunxi/cpu.h | 1 + arch/arm/include/asm/arch-sunxi/cpu_sun4i.h | 4 +- arch/arm/include/asm/arch-sunxi/dram.h | 2 +- arch/arm/include/asm/arch-sunxi/spl.h | 2 +- arch/arm/mach-sunxi/Makefile | 2 +- arch/arm/mach-sunxi/board.c | 12 +-- arch/arm/mach-sunxi/clock_sun6i.c | 6 +- arch/arm/mach-sunxi/cpu_info.c | 2 + arch/arm/mach-sunxi/dram_sun8i_h3.c | 97 ++++++++++++++--- arch/arm/mach-sunxi/usb_phy.c | 4 +- board/sunxi/Kconfig | 28 ++++- board/sunxi/MAINTAINERS | 5 + board/sunxi/board.c | 6 +- configs/orangepi_pc2_defconfig | 16 +++ drivers/mtd/spi/Kconfig | 2 +- drivers/mtd/spi/sunxi_spi_spl.c | 16 +-- drivers/net/sun8i_emac.c | 2 +- drivers/power/Kconfig | 4 +- drivers/usb/host/ehci-sunxi.c | 2 +- include/configs/exynos-common.h | 2 +- include/configs/ls1021aiot.h | 7 +- include/configs/ls1021aqds.h | 7 +- include/configs/ls1021atwr.h | 7 +- include/configs/mx7_common.h | 2 +- include/configs/sun50i.h | 1 - include/configs/sunxi-common.h | 24 ++--- scripts/config_whitelist.txt | 1 - 35 files changed, 332 insertions(+), 99 deletions(-) create mode 100644 arch/arm/dts/sun50i-h5-orangepi-pc2.dts create mode 100644 configs/orangepi_pc2_defconfig
Signature · 1 line
2.8.2
PATCH v3 1/13 sunxi: fix ACTLR.SMP assembly routine
If we take the liberty to use register r0 to perform our bit set, we should be nice enough to tell the compiler about it.
Add r0 to the clobber list to avoid potential mayhem.
--- arch/arm/mach-sunxi/board.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
arch/arm/mach-sunxi/board.c +2 -1
@@ -188,7 +188,8 @@ void s_init(void) asmvolatile ("mrc p15, 0, r0, c1, c0, 1\n" "orr r0, r0, #1 << 6\n" -"mcr p15, 0, r0, c1, c0, 1\n" ); +"mcr p15, 0, r0, c1, c0, 1\n" + :::"r0" );#endif #if definedCONFIG_MACH_SUN6I || definedCONFIG_MACH_SUN8I_H3 /* Enable non-secure access to some peripherals */
Signature · 1 line
2.8.2
PATCH v3 1/13 sunxi: fix ACTLR.SMP assembly routine
Quoted 6 lines · Andre Przywara
If we take the liberty to use register r0 to perform our bit set, we should be nice enough to tell the compiler about it.
Add r0 to the clobber list to avoid potential mayhem.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 2/13 ARM: rename CONFIG_TIMER_CLK_FREQ to COUNTER_FREQUENCY
In reply to Andre PrzywaraMany ARMv8 boards define a constant COUNTER_FREQUENCY to specify the frequency of the ARM Generic Timer (aka. arch timer).
ARMv7 boards traditionally used CONFIG_TIMER_CLK_FREQ for the same purpose. It seems useful to unify them.
Since there are less occurences of the latter version, lets convert all users over to COUNTER_FREQUENCY.
--- arch/arm/cpu/armv7/nonsec_virt.S | 4 ++-- arch/arm/cpu/armv7/sunxi/psci.c | 2 +- board/sunxi/board.c | 6 +++--- include/configs/exynos-common.h | 2 +- include/configs/ls1021aiot.h | 2 +- include/configs/ls1021aqds.h | 2 +- include/configs/ls1021atwr.h | 2 +- include/configs/mx7_common.h | 2 +- include/configs/sun50i.h | 1 - include/configs/sunxi-common.h | 2 +- scripts/config_whitelist.txt | 1 - 11 files changed, 12 insertions(+), 14 deletions(-)
arch/arm/cpu/armv7/nonsec_virt.S +2 -2
@@ -188,11 +188,11 @@ ENTRY(_nonsec_init) * we do this here instead. * But first check if we have the generic timer. */ -#ifdef CONFIG_TIMER_CLK_FREQ +#ifdef COUNTER_FREQUENCY mrc p15, 0, r0, c0, c1, 1 @ read ID_PFR1 and r0, r0, #CPUID_ARM_GENTIMER_MASK @ mask arch timer bits cmp r0, #(1 << CPUID_ARM_GENTIMER_SHIFT) - ldreq r1, =CONFIG_TIMER_CLK_FREQ + ldreq r1, =COUNTER_FREQUENCY mcreq p15, 0, r1, c14, c0, 0 @ write CNTFRQ #endif
arch/arm/cpu/armv7/sunxi/psci.c +1 -1
@@ -46,7 +46,7 @@ static u32 __secure cp15_read_cntp_ctl(void)return val; } -#define ONE_MS (CONFIG_TIMER_CLK_FREQ / 1000) +#define ONE_MS (COUNTER_FREQUENCY / 1000) static void __secure __mdelay(u32 ms) {
board/sunxi/board.c +3 -3
@@ -100,14 +100,14 @@ int board_init(void)* we avoid the risk of writing to it. */ asmvolatile ("mrc p15, 0, %0, c14, c0, 0" :"=r" (freq)); -if (freq !=CONFIG_TIMER_CLK_FREQ ) { +if (freq !=COUNTER_FREQUENCY ) {debug ("arch timer frequency is %d Hz, should be %d, fixing ...\n" , - freq,CONFIG_TIMER_CLK_FREQ ); + freq,COUNTER_FREQUENCY );#ifdef CONFIG_NON_SECURE printf ("arch timer frequency is wrong, but cannot adjust it\n" );#else asmvolatile ("mcr p15, 0, %0, c14, c0, 0" - : :"r" (CONFIG_TIMER_CLK_FREQ )); + : :"r" (COUNTER_FREQUENCY ));#endif } }
include/configs/exynos-common.h +1 -1
@@ -23,7 +23,7 @@/* input clock of PLL: 24MHz input clock */ #define CONFIG_SYS_CLK_FREQ 24000000 -#define CONFIG_TIMER_CLK_FREQ CONFIG_SYS_CLK_FREQ +#define COUNTER_FREQUENCY CONFIG_SYS_CLK_FREQ#define CONFIG_SETUP_MEMORY_TAGS #define CONFIG_CMDLINE_TAG
include/configs/ls1021aiot.h +1 -1
@@ -252,7 +252,7 @@#define CONFIG_PEN_ADDR_BIG_ENDIAN #define CONFIG_LAYERSCAPE_NS_ACCESS #define CONFIG_SMP_PEN_ADDR 0x01ee0200 -#define CONFIG_TIMER_CLK_FREQ 12500000 +#define COUNTER_FREQUENCY 12500000#define CONFIG_HWCONFIG #define HWCONFIG_BUFFER_SIZE 256
include/configs/ls1021aqds.h +1 -1
@@ -509,7 +509,7 @@ unsigned long get_board_ddr_clk(void);#define CONFIG_PEN_ADDR_BIG_ENDIAN #define CONFIG_LAYERSCAPE_NS_ACCESS #define CONFIG_SMP_PEN_ADDR 0x01ee0200 -#define CONFIG_TIMER_CLK_FREQ 12500000 +#define COUNTER_FREQUENCY 12500000#define CONFIG_HWCONFIG #define HWCONFIG_BUFFER_SIZE 256
include/configs/ls1021atwr.h +1 -1
@@ -379,7 +379,7 @@#define CONFIG_PEN_ADDR_BIG_ENDIAN #define CONFIG_LAYERSCAPE_NS_ACCESS #define CONFIG_SMP_PEN_ADDR 0x01ee0200 -#define CONFIG_TIMER_CLK_FREQ 12500000 +#define COUNTER_FREQUENCY 12500000#define CONFIG_HWCONFIG #define HWCONFIG_BUFFER_SIZE 256
include/configs/mx7_common.h +1 -1
@@ -21,7 +21,7 @@#define CONFIG_MXC_GPT_HCLK #define CONFIG_SYSCOUNTER_TIMER #define CONFIG_SC_TIMER_CLK 8000000/* 8Mhz */ -#define CONFIG_TIMER_CLK_FREQ CONFIG_SC_TIMER_CLK +#define COUNTER_FREQUENCY CONFIG_SC_TIMER_CLK#define CONFIG_SYS_FSL_CLK #define CONFIG_SYS_BOOTM_LEN 0x1000000
include/configs/sun50i.h +0 -1
@@ -18,7 +18,6 @@#define CONFIG_SUNXI_USB_PHYS 1 -#define COUNTER_FREQUENCY CONFIG_TIMER_CLK_FREQ#define GICD_BASE 0x1c81000#define GICC_BASE 0x1c82000
include/configs/sunxi-common.h +1 -1
@@ -53,7 +53,7 @@#endif /* CPU */ -#define CONFIG_TIMER_CLK_FREQ 24000000 +#define COUNTER_FREQUENCY 24000000/* * The DRAM Base differs between some models. We cannot use macros for the
scripts/config_whitelist.txt +0 -1
@@ -6541,7 +6541,6 @@ CONFIG_TI816X_EVM_DDR3 CONFIG_TI816X_USE_EMIF0 CONFIG_TI816X_USE_EMIF1 CONFIG_TI81XX -CONFIG_TIMER_CLK_FREQ CONFIG_TIMESTAMP CONFIG_TIZEN CONFIG_TI_KEYSTONE_SERDES
Signature · 1 line
2.8.2
PATCH v3 2/13 ARM: rename CONFIG_TIMER_CLK_FREQ to COUNTER_FREQUENCY
Quoted 9 lines · Andre Przywara
Many ARMv8 boards define a constant
COUNTER_FREQUENCYto specify the frequency of the ARM Generic Timer (aka. arch timer).
ARMv7 boards traditionally usedCONFIG_TIMER_CLK_FREQfor the same purpose. It seems useful to unify them.
Since there are less occurences of the latter version, lets convert all users over toCOUNTER_FREQUENCY.Signed-off-by Andre Przywara <[email protected]>Reviewed-by York Sun <[email protected]>
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 3/13 fsl: ls102x: remove redundant GENERIC_TIMER_CLK
In reply to Andre PrzywaraSome Freescale boards used an extra version of the constant to hold the Generic Timer frequency. This can easily be covered by the now unified COUNTER_FREQUENCY constant, so remove this extra variable from those boards.
--- arch/arm/cpu/armv7/ls102xa/psci.S | 2 +- arch/arm/cpu/armv7/ls102xa/timer.c | 2 +- include/configs/ls1021aiot.h | 5 ----- include/configs/ls1021aqds.h | 5 ----- include/configs/ls1021atwr.h | 5 ----- 5 files changed, 2 insertions(+), 17 deletions(-)
arch/arm/cpu/armv7/ls102xa/psci.S +1 -1
@@ -37,7 +37,7 @@ .align 5 -#define ONE_MS (GENERIC_TIMER_CLK / 1000) +#define ONE_MS (COUNTER_FREQUENCY / 1000) #define RESET_WAIT (30 * ONE_MS) .globl psci_version
arch/arm/cpu/armv7/ls102xa/timer.c +1 -1
@@ -62,7 +62,7 @@ int timer_init(void)/* Enable System Counter */ writel (SYS_COUNTER_CTRL_ENABLE , &sctr->cntcr); - freq =GENERIC_TIMER_CLK ; + freq =COUNTER_FREQUENCY ; asm("mcr p15, 0, %0, c14, c0, 0" : :"r" (freq));/* Set PL1 Physical Timer Ctrl */
include/configs/ls1021aiot.h +0 -5
@@ -36,11 +36,6 @@#define CONFIG_CMD_EXT2 #endif -/* -* Generic Timer Definitions -*/ -#define GENERIC_TIMER_CLK 12500000 -#define CONFIG_SYS_CLK_FREQ 100000000#define CONFIG_DDR_CLK_FREQ 100000000
include/configs/ls1021aqds.h +0 -5
@@ -27,11 +27,6 @@#define CONFIG_SYS_INIT_RAM_ADDR OCRAM_BASE_ADDR#define CONFIG_SYS_INIT_RAM_SIZE OCRAM_SIZE -/* -* Generic Timer Definitions -*/ -#define GENERIC_TIMER_CLK 12500000 -#ifndef __ASSEMBLY__unsigned long get_board_sys_clk (void );unsigned long get_board_ddr_clk (void );
include/configs/ls1021atwr.h +0 -5
@@ -56,11 +56,6 @@#define CONFIG_SYS_USB_XHCI_MAX_ROOT_PORTS 2#endif -/* -* Generic Timer Definitions -*/ -#define GENERIC_TIMER_CLK 12500000 -#define CONFIG_SYS_CLK_FREQ 100000000#define CONFIG_DDR_CLK_FREQ 100000000
Signature · 1 line
2.8.2
PATCH v3 3/13 fsl: ls102x: remove redundant GENERIC_TIMER_CLK
Quoted 7 lines · Andre Przywara
Some Freescale boards used an extra version of the constant to hold the Generic Timer frequency. This can easily be covered by the now unified
COUNTER_FREQUENCYconstant, so remove this extra variable from those boards.Signed-off-by Andre Przywara <[email protected]>Reviewed-by York Sun <[email protected]>
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 4/13 sunxi: simplify ACTLR.SMP bit set #ifdef
In reply to Andre PrzywaraInstead of enumerating all SoC families that need that bit set, let's just express this more clearly: The SMP bits needs to be set on SMP capable ARMv7 CPUs. It's much easier in Kconfig to express it the other way round, so we use ! CPU_IS_UP and ! ARM64.
--- arch/arm/Kconfig | 4 ++++ arch/arm/mach-sunxi/board.c | 5 +---- board/sunxi/Kconfig | 2 ++ 3 files changed, 7 insertions(+), 4 deletions(-)
arch/arm/Kconfig +4 -0
@@ -126,6 +126,10 @@ config ENABLE_ARM_SOC_BOOT0_HOOK ARM_SOC_BOOT0_HOOK which contains the required assembler preprocessor code. +config ARM_CORTEX_CPU_IS_UP + bool + default n + config USE_ARCH_MEMCPY bool "Use an assembly optimized implementation of memcpy" default y
arch/arm/mach-sunxi/board.c +1 -4
@@ -180,10 +180,7 @@ void s_init(void)/* No H3 BSP, boot0 seems to not modify SUNXI_SRAMC_BASE + 0x44 */ #endif -#if definedCONFIG_MACH_SUN6I || \ - definedCONFIG_MACH_SUN7I || \ - definedCONFIG_MACH_SUN8I || \ - definedCONFIG_MACH_SUN9I +#if !defined(CONFIG_ARM_CORTEX_CPU_IS_UP ) && !defined(CONFIG_ARM64 )/* Enable SMP mode for CPU0, by setting bit 6 of Auxiliary Ctl reg */ asmvolatile ( "mrc p15, 0, r0, c1, c0, 1\n"
board/sunxi/Kconfig +2 -0
@@ -50,12 +50,14 @@ choice config MACH_SUN4I bool "sun4i (Allwinner A10)" select CPU_V7 + select ARM_CORTEX_CPU_IS_UP select SUNXI_GEN_SUN4I select SUPPORT_SPL config MACH_SUN5I bool "sun5i (Allwinner A13)" select CPU_V7 + select ARM_CORTEX_CPU_IS_UP select SUNXI_GEN_SUN4I select SUPPORT_SPL
Signature · 1 line
2.8.2
PATCH v3 4/13 sunxi: simplify ACTLR.SMP bit set #ifdef
Quoted 18 lines · Andre Przywara
Instead of enumerating all SoC families that need that bit set, let's just express this more clearly: The SMP bits needs to be set on SMP capable ARMv7 CPUs. It's much easier in
Kconfigto express it the other way round, so we use !CPU_IS_UPand ! ARM64.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/Kconfig | 4 ++++ arch/arm/mach-sunxi/board.c | 5 +---- board/sunxi/Kconfig | 2 ++ 3 files changed, 7 insertions(+), 4 deletions(-)arch/arm/Kconfig +0
-0
arch/arm/Kconfig config ENABLE_ARM_SOC_BOOT0_HOOK · line 131
Quoted 6 lines
arch/arm/Kconfig +3
-0@@ -126,6 +126,10 @@ config ENABLE_ARM_SOC_BOOT0_HOOK ARM_SOC_BOOT0_HOOK which contains the required assembler preprocessor code. +config ARM_CORTEX_CPU_IS_UP + bool + default n
Better to place this in sunxi, since no other code using this expect sunxi and the name CORTEX may also refer arm64 use something 32 related.
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 4/13 sunxi: simplify ACTLR.SMP bit set #ifdef
Hi,
Quoted 22 lines · Jagan Teki
Quoted 21 lines · Andre Przywara
Instead of enumerating all SoC families that need that bit set, let's just express this more clearly: The SMP bits needs to be set on SMP capable ARMv7 CPUs. It's much easier in
Kconfigto express it the other way round, so we use !CPU_IS_UPand ! ARM64.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/Kconfig | 4 ++++ arch/arm/mach-sunxi/board.c | 5 +---- board/sunxi/Kconfig | 2 ++ 3 files changed, 7 insertions(+), 4 deletions(-)arch/arm/Kconfig +0
-0@@ -126,6 +126,10 @@ config ENABLE_ARM_SOC_BOOT0_HOOK ARM_SOC_BOOT0_HOOK which contains the required assembler preprocessor code.
Quoted 7 lines
arch/arm/Kconfig config ENABLE_ARM_SOC_BOOT0_HOOK · line 131
Quoted 3 lines
arch/arm/Kconfig +3
-0+config ARM_CORTEX_CPU_IS_UP + bool + default nBetter to place this in sunxi, since no other code using this expect sunxi and the name CORTEX may also refer arm64 use something 32 related.
Sigh, can you please check back with Maxime on what's the right thing here?
http://lists.denx.de/pipermail/u-boot/2017-January/279417.html
If it's about the name, shall we use ARM_CORTEX_V7_CPU_IS_UP?
I was briefly tempted to unify all ACTLR.SMP bit sets from all over the ARM code, but this looks like a can of worms to me, so I'd rather keep this one closed.
Cheers,
Andre.
PATCH v3 4/13 sunxi: simplify ACTLR.SMP bit set #ifdef
Quoted 33 lines · Andre Przywara
Hi,
Quoted 22 lines · Jagan Teki
Quoted 21 lines · Andre Przywara
Instead of enumerating all SoC families that need that bit set, let's just express this more clearly: The SMP bits needs to be set on SMP capable ARMv7 CPUs. It's much easier in
Kconfigto express it the other way round, so we use !CPU_IS_UPand ! ARM64.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/Kconfig | 4 ++++ arch/arm/mach-sunxi/board.c | 5 +---- board/sunxi/Kconfig | 2 ++ 3 files changed, 7 insertions(+), 4 deletions(-)arch/arm/Kconfig +0
-0@@ -126,6 +126,10 @@ config ENABLE_ARM_SOC_BOOT0_HOOK ARM_SOC_BOOT0_HOOK which contains the required assembler preprocessor code.Quoted 7 lines
arch/arm/Kconfig config ENABLE_ARM_SOC_BOOT0_HOOK · line 131
Quoted 3 lines
arch/arm/Kconfig +3
-0+config ARM_CORTEX_CPU_IS_UP + bool + default nBetter to place this in sunxi, since no other code using this expect sunxi and the name CORTEX may also refer arm64 use something 32 related.
Quoted 8 lines
Sigh, can you please check back with Maxime on what's the right thing here?
http://lists.denx.de/pipermail/u-boot/2017-January/279417.htmlIf it's about the name, shall we use
ARM_CORTEX_V7_CPU_IS_UP?I was briefly tempted to unify all ACTLR.SMP bit sets from all over the ARM code, but this looks like a can of worms to me, so I'd rather keep this one closed.
There's at least one other SoC in a similar situation (imx7 solo), so I think it should really be made generic. I don't really care about the name though, but keeping it in sunxi is just creating yet another candidate for a duplicated and redundant option.
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170207/259e4157/attachment.sig>
PATCH v3 5/13 sunxi: configs: merge sun9i and sun50i SPL memory definitions
In reply to Andre PrzywaraFor some reason we were pretty conservative when defining the maximum SPL size for the Allwinner A80(sun9i) SoC.
According to the manual the SRAM A1 is even 40KB, but the BROM probably still has the 32 KiB load limit. For the sake of simplicity, merge the SPL memory definitions for the A64 and A80 SoCs, since both SoC share the BROM/SRAM A1 memory layout.
This helps to further simplify this is in the next patch.
--- include/configs/sunxi-common.h | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-)
include/configs/sunxi-common.h +2 -5
@@ -186,12 +186,9 @@#define CONFIG_SPL_BOARD_LOAD_IMAGE #endif -#if defined(CONFIG_MACH_SUN9I ) -#define CONFIG_SPL_TEXT_BASE 0x10040/* sram start+header */ -#define CONFIG_SPL_MAX_SIZE 0x5fc0/* ? KiB on sun9i */ -#elif defined(CONFIG_MACH_SUN50I ) +#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I )#define CONFIG_SPL_TEXT_BASE 0x10040/* sram start+header */ -#define CONFIG_SPL_MAX_SIZE 0x7fc0/* 32 KiB on sun50i */ +#define CONFIG_SPL_MAX_SIZE 0x7fc0/* 32 KiB on sun9/50i */ #else #define CONFIG_SPL_TEXT_BASE 0x40/* sram start+header */ #define CONFIG_SPL_MAX_SIZE 0x5fc0/* 24KB on sun4i/sun7i */
Signature · 1 line
2.8.2
PATCH v3 5/13 sunxi: configs: merge sun9i and sun50i SPL memory definitions
Quoted 9 lines · Andre Przywara
For some reason we were pretty conservative when defining the maximum SPL size for the Allwinner A80(sun9i) SoC.
According to the manual the SRAM A1 is even 40KB, but the BROM probably still has the 32 KiB load limit. For the sake of simplicity, merge the SPL memory definitions for the A64 and A80 SoCs, since both SoC share the BROM/SRAM A1 memory layout.
This helps to further simplify this is in the next patch.Signed-off-by Andre Przywara <[email protected]>
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170202/295ee844/attachment.sig>
PATCH v3 5/13 sunxi: configs: merge sun9i and sun50i SPL memory definitions
Quoted 9 lines · Maxime Ripard
Quoted 8 lines · Andre Przywara
For some reason we were pretty conservative when defining the maximum SPL size for the Allwinner A80(sun9i) SoC.
According to the manual the SRAM A1 is even 40KB, but the BROM probably still has the 32 KiB load limit. For the sake of simplicity, merge the SPL memory definitions for the A64 and A80 SoCs, since both SoC share the BROM/SRAM A1 memory layout.
This helps to further simplify this is in the next patch.
Quoted 3 lines
Quoted 1 line
Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 6/13 sunxi: Kconfig: introduce CONFIG_SUNXI_HIGH_SRAM
In reply to Andre PrzywaraTraditionally Allwinner SoCs have their boot ROM mapped just below 4GB, while the first SRAM region is mapped at address 0.
With the extended physical memory support of the A80 this was changed, so the BROM is now at address 0 and the SRAM region starts right behind this at 64KB. This configuration seems to be called "high SRAM". Instead of enumerating the SoCs which have copied this configuration, let's call a spade a spade and introduce a Kconfig option for this setup. SoCs implementing this (A80, A64 and H5, so far), can then select this configuration.
Simplify the config header definition on the way.
--- arch/arm/include/asm/arch-sunxi/spl.h | 2 +- board/sunxi/Kconfig | 6 ++++++ include/configs/sunxi-common.h | 19 +++++++------------ 3 files changed, 14 insertions(+), 13 deletions(-)
arch/arm/include/asm/arch-sunxi/spl.h +1 -1
@@ -12,7 +12,7 @@#define SPL_SIGNATURE "SPL"/* marks "sunxi" SPL header */ #define SPL_HEADER_VERSION 1 -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) +#ifdef CONFIG_SUNXI_HIGH_SRAM #define SPL_ADDR 0x10000#else #define SPL_ADDR 0x0
board/sunxi/Kconfig +6 -0
@@ -27,6 +27,10 @@ config SPL_POWER_SUPPORT config SPL_SERIAL_SUPPORT default y +config SUNXI_HIGH_SRAM + bool + default n + # Note only one of these may be selected at a time! But hidden choices are # not supported by Kconfig config SUNXI_GEN_SUN4I
@@ -120,6 +124,7 @@ config MACH_SUN8I_H3 config MACH_SUN9I bool "sun9i (Allwinner A80)" select CPU_V7 + select SUNXI_HIGH_SRAM select SUNXI_GEN_SUN6I select SUPPORT_SPL
@@ -127,6 +132,7 @@ config MACH_SUN50I bool "sun50i (Allwinner A64)" select ARM64 select SUNXI_GEN_SUN6I + select SUNXI_HIGH_SRAM select SUPPORT_SPL endchoice
include/configs/sunxi-common.h +7 -12
@@ -86,7 +86,7 @@#define CONFIG_SPL_BSS_MAX_SIZE 0x00080000/* 512 KiB */ -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) +#ifdef CONFIG_SUNXI_HIGH_SRAM /* * The A80's A1 sram starts at 0x00010000 rather then at 0x00000000 and is * slightly bigger. Note that it is possible to map the first 32 KiB of the
@@ -186,29 +186,24 @@#define CONFIG_SPL_BOARD_LOAD_IMAGE #endif -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) +#ifdef CONFIG_SUNXI_HIGH_SRAM #define CONFIG_SPL_TEXT_BASE 0x10040/* sram start+header */ -#define CONFIG_SPL_MAX_SIZE 0x7fc0/* 32 KiB on sun9/50i */ +#define CONFIG_SPL_MAX_SIZE 0x7fc0/* 32 KiB */ +#define LOW_LEVEL_SRAM_STACK 0x00018000#else #define CONFIG_SPL_TEXT_BASE 0x40/* sram start+header */ #define CONFIG_SPL_MAX_SIZE 0x5fc0/* 24KB on sun4i/sun7i */ +#define LOW_LEVEL_SRAM_STACK 0x00008000/* End of sram */ #endif +#define CONFIG_SPL_STACK LOW_LEVEL_SRAM_STACK +#ifndef CONFIG_ARM64 #define CONFIG_SPL_LDSCRIPT "arch/arm/cpu/armv7/sunxi/u-boot-spl.lds"#endif #define CONFIG_SPL_PAD_TO 32768/* decimal for 'dd' */ -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) -/* FIXME: 40 KiB instead of 32 KiB ? */ -#define LOW_LEVEL_SRAM_STACK 0x00018000 -#define CONFIG_SPL_STACK LOW_LEVEL_SRAM_STACK -#else -/* end of 32 KiB in sram */ -#define LOW_LEVEL_SRAM_STACK 0x00008000/* End of sram */ -#define CONFIG_SPL_STACK LOW_LEVEL_SRAM_STACK -#endif /* I2C */ #if defined CONFIG_AXP152_POWER || defined CONFIG_AXP209_POWER || \
Signature · 1 line
2.8.2
PATCH v3 6/13 sunxi: Kconfig: introduce CONFIG_SUNXI_HIGH_SRAM
Quoted 37 lines · Andre Przywara
Traditionally Allwinner SoCs have their boot ROM mapped just below 4GB, while the first SRAM region is mapped at address 0.
With the extended physical memory support of the A80 this was changed, so the BROM is now at address 0 and the SRAM region starts right behind this at 64KB. This configuration seems to be called "high SRAM". Instead of enumerating the SoCs which have copied this configuration, let's call a spade a spade and introduce aKconfigoption for this setup. SoCs implementing this (A80, A64 and H5, so far), can then select this configuration.
Simplify the config header definition on the way.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/include/asm/arch-sunxi/spl.h | 2 +- board/sunxi/Kconfig | 6 ++++++ include/configs/sunxi-common.h | 19 +++++++------------ 3 files changed, 14 insertions(+), 13 deletions(-)arch/arm/include/asm/arch-sunxi/spl.h +1
-1@@ -12,7 +12,7 @@#define SPL_SIGNATURE "SPL"/* marks "sunxi" SPL header */ #define SPL_HEADER_VERSION 1 -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) +#ifdef CONFIG_SUNXI_HIGH_SRAM #define SPL_ADDR 0x10000#else #define SPL_ADDR 0x0board/sunxi/Kconfig +0
-0
board/sunxi/Kconfig config SPL_POWER_SUPPORT · line 32
Quoted 6 lines
board/sunxi/Kconfig +3
-0@@ -27,6 +27,10 @@ config SPL_POWER_SUPPORT config SPL_SERIAL_SUPPORT default y +config SUNXI_HIGH_SRAM + bool + default n
Write some help text, once written
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 6/13 sunxi: Kconfig: introduce CONFIG_SUNXI_HIGH_SRAM
Hi Jagan,
thanks for the review!
Quoted 41 lines · Jagan Teki
Quoted 40 lines · Andre Przywara
Traditionally Allwinner SoCs have their boot ROM mapped just below 4GB, while the first SRAM region is mapped at address 0.
With the extended physical memory support of the A80 this was changed, so the BROM is now at address 0 and the SRAM region starts right behind this at 64KB. This configuration seems to be called "high SRAM". Instead of enumerating the SoCs which have copied this configuration, let's call a spade a spade and introduce aKconfigoption for this setup. SoCs implementing this (A80, A64 and H5, so far), can then select this configuration.
Simplify the config header definition on the way.Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/include/asm/arch-sunxi/spl.h | 2 +- board/sunxi/Kconfig | 6 ++++++ include/configs/sunxi-common.h | 19 +++++++------------ 3 files changed, 14 insertions(+), 13 deletions(-)arch/arm/include/asm/arch-sunxi/spl.h +1
-1@@ -12,7 +12,7 @@#define SPL_SIGNATURE "SPL"/* marks "sunxi" SPL header */ #define SPL_HEADER_VERSION 1 -#if defined(CONFIG_MACH_SUN9I ) || defined(CONFIG_MACH_SUN50I ) +#ifdef CONFIG_SUNXI_HIGH_SRAM #define SPL_ADDR 0x10000#else #define SPL_ADDR 0x0board/sunxi/Kconfig +0
-0@@ -27,6 +27,10 @@ config SPL_POWER_SUPPORT config SPL_SERIAL_SUPPORT default y
Quoted 5 lines
board/sunxi/Kconfig config SPL_POWER_SUPPORT · line 32
Quoted 3 lines
board/sunxi/Kconfig +3
-0+config SUNXI_HIGH_SRAM + bool + default nWrite some help text, once written
I wanted originally, but then saw that none of the invisible options (no text after bool) have a help text, mainly because it will probably never be shown.
I can still add some help text here for documentation purposes, or in some other place. Or leave it up for the commit message to cover that.
Cheers,
Andre.
Quoted 5 lines
Reviewed-by Jagan Teki <[email protected]>thanks!
PATCH v3 7/13 sunxi: provide ARMv8 mem_map for every ARM64 board
In reply to Andre PrzywaraEvery armv8 board needs the memory map, so change the #ifdef to ARM64 to avoid enumerating every single board or SoC.
--- arch/arm/mach-sunxi/board.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/mach-sunxi/board.c +1 -1
@@ -40,7 +40,7 @@ struct fel_stash { struct fel_stash fel_stash __attribute__((section(".data"))); -#ifdef CONFIG_MACH_SUN50I +#ifdef CONFIG_ARM64 #include <asm/armv8/mmu.h> static struct mm_region sunxi_mem_map[] = {
Signature · 1 line
2.8.2
PATCH v3 7/13 sunxi: provide ARMv8 mem_map for every ARM64 board
Quoted 8 lines · Andre Przywara
Every armv8 board needs the memory map, so change the #ifdef to ARM64 to avoid enumerating every single board or SoC.
Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/mach-sunxi/board.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 8/13 SPI: SPL: sunxi: fix 64-bit build
In reply to Andre PrzywaraAddresses passed on to readl and writel are expected to be of the same size as a pointer. Change the parameter types of sunxi_spi0_read_data() to make the compiler happy and allow a warning-free aarch64 compile.
--- drivers/mtd/spi/sunxi_spi_spl.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/mtd/spi/sunxi_spi_spl.c +8 -8
@@ -185,14 +185,14 @@ static void spi0_deinit(void)#define SPI_READ_MAX_SIZE 60/* FIFO size, minus 4 bytes of the header */ static void sunxi_spi0_read_data (u8 *buf,u32 addr,u32 bufsize, -u32 spi_ctl_reg, -u32 spi_ctl_xch_bitmask, -u32 spi_fifo_reg, -u32 spi_tx_reg, -u32 spi_rx_reg, -u32 spi_bc_reg, -u32 spi_tc_reg, -u32 spi_bcc_reg) +ulong spi_ctl_reg, +ulong spi_ctl_xch_bitmask, +ulong spi_fifo_reg, +ulong spi_tx_reg, +ulong spi_rx_reg, +ulong spi_bc_reg, +ulong spi_tc_reg, +ulong spi_bcc_reg) {writel (4 + bufsize, spi_bc_reg);/* Burst counter (total bytes) */ writel (4 , spi_tc_reg);/* Transfer counter (bytes to send) */
Signature · 1 line
2.8.2
PATCH v3 8/13 SPI: SPL: sunxi: fix 64-bit build
Quoted 7 lines · Andre Przywara
Addresses passed on to readl and writel are expected to be of the same size as a pointer. Change the parameter types of
sunxi_spi0_read_data()to make the compiler happy and allow a warning-free aarch64 compile.Signed-off-by Andre Przywara <[email protected]>Reviewed-by Simon Glass <[email protected]>Reviewed-by Maxime Ripard <[email protected]>
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 9/13 sunxi: DRAM: add Allwinner H5 support
In reply to Andre PrzywaraThe DRAM controller in the Allwinner H5 SoC is again very similar to the one in the H3 and A64.
Based on the existing socid parameter, add support for this controller by reusing the bulk of the code and only deviating where needed. These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. Also add the delay line parameters taken from the boot0 and libdram disassembly.
Register setup differences between H5 and H3 are courtesy of Jens Kuske.
--- arch/arm/include/asm/arch-sunxi/cpu.h | 1 + arch/arm/mach-sunxi/dram_sun8i_h3.c | 97 +++++++++++++++++++++++++++++------ 2 files changed, 82 insertions(+), 16 deletions(-)
arch/arm/include/asm/arch-sunxi/cpu.h +1 -0
@@ -15,5 +15,6 @@#define SOCID_A64 0x1689#define SOCID_H3 0x1680 +#define SOCID_H5 0x1718#endif /* _SUNXI_CPU_H */
arch/arm/mach-sunxi/dram_sun8i_h3.c +81 -16
187 lines not on this page · raw message
Signature · 1 line
2.8.2
PATCH v3 9/13 sunxi: DRAM: add Allwinner H5 support
On Feb 1, 2017 2:37 AM, "Andre Przywara" <[email protected]> wrote:
The DRAM controller in the Allwinner H5 SoC is again very similar to the one in the H3 and A64.
Based on the existing socid parameter, add support for this controller by reusing the bulk of the code and only deviating where needed. These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. Also add the delay line parameters taken from the boot0 and libdram disassembly.
Can you split this patch with delay line params as separate patch.
Register setup differences between H5 and H3 are courtesy of Jens Kuske.
--- arch/arm/include/asm/arch-sunxi/cpu.h | 1 + arch/arm/mach-sunxi/dram_sun8i_h3.c | 97 +++++++++++++++++++++++++++++- ----- 2 files changed, 82 insertions(+), 16 deletions(-)
(quoted diff) +0 -0
b/arch/arm/include/asm/arch-sunxi/cpu.h
index 6f96a97..e8e670e 100644
arch/arm/include/asm/arch-sunxi/cpu.h +1 -0
@@ -15,5 +15,6 @@#define SOCID_A64 0x1689#define SOCID_H3 0x1680 +#define SOCID_H5 0x1718#endif /* _SUNXI_CPU_H */
arch/arm/mach-sunxi/dram_ renamed from arch/arm/mach-sunxi/dram_sun8i_h3.c +0 -0
sun8i_h3.c
index 9f7cc7f..d681a9d 100644
arch/arm/mach-sunxi/dram_sun8i_h3.c +8 -0
@@ -177,6 +177,34 @@ static void mctl_set_master_priority_a64(void)writel (0x81000004 , &mctl_com->mdfs_bwlr[2 ]); } +static void mctl_set_master_priority_h5 (void ) +{ +struct sunxi_mctl_com_reg *const mctl_com = + (struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE ; + +/* enable bandwidth limit windows and set windows size 1us */ +writel (399 , &mctl_com->tmr); +writel ((1 <<16 ), &mctl_com->bwcr);
I'm not fond of using direct numerical values make code unhealthy please use macros with bitops. Note that this comment will apply rest of the code where it applies.
+
+ /* set cpu high priority */
+ writel(0x00000001, &mctl_com->mapr);
+
+ /* Port 2 is reserved per Allwinner's linux-3.10 source, yet
+ * they initialise it */
+ MBUS_CONF( CPU, true, HIGHEST, 0, 300, 260, 150);
+ MBUS_CONF( GPU, true, HIGHEST, 0, 600, 400, 200);
+ MBUS_CONF(UNUSED, true, HIGHEST, 0, 512, 256, 96);
+ MBUS_CONF( DMA, true, HIGHEST, 0, 256, 128, 32);
+ MBUS_CONF( VE, true, HIGHEST, 0, 1900, 1500, 1000);
+ MBUS_CONF( CSI, true, HIGHEST, 0, 150, 120, 100);
+ MBUS_CONF( NAND, true, HIGH, 0, 256, 128, 64);
+ MBUS_CONF( SS, true, HIGHEST, 0, 256, 128, 64);
+ MBUS_CONF( TS, true, HIGHEST, 0, 256, 128, 64);
+ MBUS_CONF( DI, true, HIGH, 0, 1024, 256, 64);
+ MBUS_CONF( DE, true, HIGHEST, 3, 3400, 2400, 1024);
+ MBUS_CONF(DE_CFD, true, HIGHEST, 0, 600, 400, 200);
+}
+
static void mctl_set_master_priority(uint16_t socid)
{
switch (socid) {
@@ -186,6 +214,9 @@ static void mctl_set_master_priority(uint16_t socid)
case SOCID_A64:
mctl_set_master_priority_a64();
return;
+ case SOCID_H5:
+ mctl_set_master_priority_h5();
+ return;
}}
@@ -256,7 +287,7 @@ static void mctl_set_timing_params(uint16_t socid, struct dram_para *para)
/* set two rank timing */
clrsetbits_le32(&mctl_ctl->dramtmg[8], (0xff << 8) | (0xff << 0),- (0x66 << 8) | (0x10 << 0));
+ ((socid == SOCID_H5 ? 0x33 : 0x66) << 8) | (0x10 <<
0));
/* set PHY interface timing, write latency and read latency
configure */
writel((0x2 << 24) | (t_rdata_en << 16) | (0x1 << 8) |
@@ -391,7 +422,7 @@ static void mctl_sys_init(uint16_t socid, struct
dram_para *para)
CCM_DRAMCLK_CFG_DIV(1) |
CCM_DRAMCLK_CFG_SRC_PLL11 |
CCM_DRAMCLK_CFG_UPD);
- } else if (socid == SOCID_H3) {
+ } else if (socid == SOCID_H3 || socid == SOCID_H5) {
clock_set_pll5(CONFIG_DRAM_CLK * 2 * 1000000, false);
clrsetbits_le32(&ccm->dram_clk_cfg,
CCM_DRAMCLK_CFG_DIV_MASK |
@@ -410,7 +441,7 @@ static void mctl_sys_init(uint16_t socid, struct
dram_para *para)
setbits_le32(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_RST);
udelay(10);- writel(0xc00e, &mctl_ctl->clken);
+ writel(socid == SOCID_H5 ? 0x8000 : 0xc00e, &mctl_ctl->clken);
udelay(500);}
@@ -434,7 +465,10 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para)
/* setting VTC, default disable all VT */
clrbits_le32(&mctl_ctl->pgcr[0], (1 << 30) | 0x3f);- clrsetbits_le32(&mctl_ctl->pgcr[1], 1 << 24, 1 << 26);
+ if (socid == SOCID_H5)
+ setbits_le32(&mctl_ctl->pgcr[1], (1 << 24) | (1 << 26));
+ else
+ clrsetbits_le32(&mctl_ctl->pgcr[1], 1 << 24, 1 << 26);
/* increase DFI_PHY_UPD clock */
writel(PROTECT_MAGIC, &mctl_com->protect);
@@ -444,15 +478,22 @@ static int mctl_channel_init(uint16_t socid, struct
dram_para *para)
udelay(100);
/* set dramc odt */
- for (i = 0; i < 4; i++)
- clrsetbits_le32(&mctl_ctl->dx[i].gcr, (0x3 << 4) |
- (0x1 << 1) | (0x3 << 2) | (0x3 << 12) |
- (0x3 << 14),
- IS_ENABLED(CONFIG_DRAM_ODT_EN) ?- DX_GCR_ODT_DYNAMIC :DX_GCR_ODT_OFF);
+ for (i = 0; i < 4; i++) {
+ u32 clearmask = (0x3 << 4) | (0x1 << 1) | (0x3 << 2) |
+ (0x3 << 12) | (0x3 << 14);
+ u32 setmask = IS_ENABLED(CONFIG_DRAM_ODT_EN) ?
+ DX_GCR_ODT_DYNAMIC : DX_GCR_ODT_OFF;
+
+ if (socid == SOCID_H5) {
+ clearmask |= 0x2 << 8;
+ setmask |= 0x4 << 8;
+ }
+ clrsetbits_le32(&mctl_ctl->dx[i].gcr, clearmask, setmask);
+ }
/* AC PDR should always ON */
- setbits_le32(&mctl_ctl->aciocr, 0x1 << 1);
+ clrsetbits_le32(&mctl_ctl->aciocr, socid == SOCID_H5 ? (0x1 << 11)
: 0,
+ 0x1 << 1);
/* set DQS auto gating PD mode */
setbits_le32(&mctl_ctl->pgcr[2], 0x3 << 6);
@@ -464,7 +505,7 @@ static int mctl_channel_init(uint16_t socid, struct
dram_para *para)
/* dphy & aphy phase select 270 degree */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 <<
8),
(0x1 << 10) | (0x2 << 8));
- } else if (socid == SOCID_A64) {
+ } else if (socid == SOCID_A64 || socid == SOCID_H5) {
/* dphy & aphy phase select ? */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 <<
8),
(0x0 << 10) | (0x3 << 8));@@ -488,11 +529,12 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para)
mctl_phy_init(PIR_PLLINIT | PIR_DCAL | PIR_PHYRST |
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);
- } else if (socid == SOCID_A64) {
+ } else if (socid == SOCID_A64 || socid == SOCID_H5) {
clrsetbits_le32(&mctl_ctl->zqcr, 0xffffff, CONFIG_DRAM_ZQ);
mctl_phy_init(PIR_ZCAL | PIR_PLLINIT | PIR_DCAL |PIR_PHYRST |
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);
+ /* no PIR_QSGATE for H5 ???? */
}
/* detect ranks and bus width */
@@ -533,7 +575,7 @@ static int mctl_channel_init(uint16_t socid, struct
dram_para *para)
/* set PGCR3, CKE polarity */
if (socid == SOCID_H3)
writel(0x00aa0060, &mctl_ctl->pgcr[3]);
- else if (socid == SOCID_A64)
+ else if (socid == SOCID_A64 || socid == SOCID_H5)
writel(0xc0aa0060, &mctl_ctl->pgcr[3]);
/* power down zq calibration module for power save */
@@ -604,6 +646,22 @@ static void mctl_auto_detect_dram_size(struct
dram_para *para)
3, 4, 0, 3, 4, 1, 4, 0, \
1, 1, 0, 1, 13, 5, 4 }+#define SUN8I_H5_DX_READ_DELAYS \
+ {{ 14, 15, 17, 17, 17, 17, 17, 18, 17, 3, 3 }, \
+ { 21, 21, 12, 22, 21, 21, 21, 21, 21, 3, 3 }, \
+ { 16, 19, 19, 17, 22, 22, 21, 22, 19, 3, 3 }, \
+ { 21, 21, 22, 22, 20, 21, 19, 19, 19, 3, 3 } }
+#define SUN8I_H5_DX_WRITE_DELAYS \
+ {{ 1, 2, 3, 4, 3, 4, 4, 4, 6, 6, 6 }, \
+ { 6, 6, 6, 5, 5, 5, 5, 5, 6, 6, 6 }, \
+ { 0, 2, 4, 2, 6, 5, 5, 5, 6, 6, 6 }, \
+ { 3, 3, 3, 2, 2, 1, 1, 1, 4, 4, 4 } }
+#define SUN8I_H5_AC_DELAYS \
+ { 0, 0, 5, 5, 0, 0, 0, 0, \
+ 0, 0, 0, 0, 3, 3, 3, 3, \
+ 3, 3, 3, 3, 3, 3, 3, 3, \+ 3, 3, 3, 3, 2, 0, 0 }
+
unsigned long sunxi_dram_init(void)
{
struct sunxi_mctl_com_reg * const mctl_com =
@@ -625,6 +683,10 @@ unsigned long sunxi_dram_init(void)
.dx_read_delays = SUN50I_A64_DX_READ_DELAYS,
.dx_write_delays = SUN50I_A64_DX_WRITE_DELAYS,
.ac_delays = SUN50I_A64_AC_DELAYS,
+#elif defined(CONFIG_MACH_SUN50I_H5)
+ .dx_read_delays = SUN8I_H5_DX_READ_DELAYS,
+ .dx_write_delays = SUN8I_H5_DX_WRITE_DELAYS,
+ .ac_delays = SUN8I_H5_AC_DELAYS,
#endif
};
/*
@@ -636,6 +698,8 @@ unsigned long sunxi_dram_init(void)
uint16_t socid = SOCID_H3;
#elif defined(CONFIG_MACH_SUN50I)
uint16_t socid = SOCID_A64;+#elif defined(CONFIG_MACH_SUN50I_H5)
+ uint16_t socid = SOCID_H5;
#endif
mctl_sys_init(socid, ¶);
@@ -652,8 +716,9 @@ unsigned long sunxi_dram_init(void)
if (socid ==[U-Boot] [linux-sunxi] Re: [PATCH v3 09/13] sunxi: DRAM: add Allwinner H5 support
Hi,
Quoted 14 lines · Jagan Teki
On Feb 1, 2017 2:37 AM, "Andre Przywara" <[email protected]> wrote:
The DRAM controller in the Allwinner H5 SoC is again very similar to the one in the H3 and A64.
Based on the existing socid parameter, add support for this controller by reusing the bulk of the code and only deviating where needed. These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. Also add the delay line parameters taken from the boot0 and libdram disassembly.
Quoted 1 line
Can you split this patch with delay line params as separate patch.
It looks like the delay lines are for the H5, merely taken from different sources. They are and should be part of the same patch adding support for DRAM on the H5.
Quoted 42 lines
Register setup differences between H5 and H3 are courtesy of Jens Kuske.
Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/include/asm/arch-sunxi/cpu.h | 1 + arch/arm/mach-sunxi/dram_sun8i_h3.c | 97 +++++++++++++++++++++++++++++------ 2 files changed, 82 insertions(+), 16 deletions(-)arch/arm/include/asm/arch-sunxi/cpu.h +1
-0@@ -15,5 +15,6 @@#define SOCID_A64 0x1689#define SOCID_H3 0x1680 +#define SOCID_H5 0x1718#endif /* _SUNXI_CPU_H */ arch/arm/mach-sunxi/dram_sun8i_h3.c +8
-0@@ -177,6 +177,34 @@ static void mctl_set_master_priority_a64(void)writel (0x81000004 , &mctl_com->mdfs_bwlr[2 ]); } +static void mctl_set_master_priority_h5 (void ) +{ +struct sunxi_mctl_com_reg *const mctl_com = + (struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE ; + +/* enable bandwidth limit windows and set windows size 1us */ +writel (399 , &mctl_com->tmr); +writel ((1 <<16 ), &mctl_com->bwcr);
Quoted 3 lines
I'm not fond of using direct numerical values make code unhealthy please use macros with bitops. Note that this comment will apply rest of the code where it applies.
I think you are being unreasonable. The commit message clearly states that the added code either comes from register dumps, disassembled blobs, or comparison of released code:
"""
These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. """
For this particular instance, see
which gives the exact same comment, and no named macros. Adding a macro for it and calling it H5_DRAM_BW_UNKNOWN_MAGIC is not an improvement. Same goes for the next few lines.
Allwinner has _never_ released documents for the DRAM controllers or DRAM PHYs, and only sometimes releases code for DRAM init for some SoCs, sometimes with questionable licenses (or lack of), of which some don't match what is actually seen in provided blobs. Considering what the community has access to. This patch seems to be quite good.
Regards
ChenYu
Quoted 204 lines
arch/arm/mach-sunxi/dram_sun8i_h3.c +24
-1+ +/* set cpu high priority */ +writel (0x00000001 , &mctl_com->mapr); + +/* Port 2 is reserved per Allwinner's linux-3.10 source, yet +* they initialise it */ +MBUS_CONF (CPU , true,HIGHEST ,0 ,300 ,260 ,150 ); +MBUS_CONF (GPU , true,HIGHEST ,0 ,600 ,400 ,200 ); +MBUS_CONF (UNUSED , true,HIGHEST ,0 ,512 ,256 ,96 ); +MBUS_CONF (DMA , true,HIGHEST ,0 ,256 ,128 ,32 ); +MBUS_CONF (VE , true,HIGHEST ,0 ,1900 ,1500 ,1000 ); +MBUS_CONF (CSI , true,HIGHEST ,0 ,150 ,120 ,100 ); +MBUS_CONF (NAND , true,HIGH ,0 ,256 ,128 ,64 ); +MBUS_CONF (SS , true,HIGHEST ,0 ,256 ,128 ,64 ); +MBUS_CONF (TS , true,HIGHEST ,0 ,256 ,128 ,64 ); +MBUS_CONF (DI , true,HIGH ,0 ,1024 ,256 ,64 ); +MBUS_CONF (DE , true,HIGHEST ,3 ,3400 ,2400 ,1024 ); +MBUS_CONF (DE_CFD , true,HIGHEST ,0 ,600 ,400 ,200 ); +} +static void mctl_set_master_priority (uint16_t socid) {switch (socid) {@@ -186,6 +214,9 @@ static void mctl_set_master_priority(uint16_t socid)case SOCID_A64 :mctl_set_master_priority_a64 ();return ; +case SOCID_H5 : +mctl_set_master_priority_h5 (); +return ; } }@@ -256,7 +287,7 @@ static void mctl_set_timing_params(uint16_t socid, struct dram_para *para) /* set two rank timing */ clrsetbits_le32(&mctl_ctl->dramtmg[8], (0xff << 8) | (0xff << 0), - (0x66 << 8) | (0x10 << 0)); + ((socid ==SOCID_H5 ?0x33 :0x66 ) <<8 ) | (0x10 << 0));/* set PHY interface timing, write latency and read latency configure */ writel((0x2 << 24) | (t_rdata_en << 16) | (0x1 << 8) |arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -391,7 +422,7 @@ static void mctl_sys_init(uint16_t socid, struct dram_para *para) CCM_DRAMCLK_CFG_DIV(1) | CCM_DRAMCLK_CFG_SRC_PLL11 | CCM_DRAMCLK_CFG_UPD); - }else if (socid ==SOCID_H3 ) { + }else if (socid ==SOCID_H3 || socid ==SOCID_H5 ) {clock_set_pll5 (CONFIG_DRAM_CLK *2 *1000000 , false); clrsetbits_le32(&ccm->dram_clk_cfg,
CCM_DRAMCLK_CFG_DIV_MASK|arch/arm/mach-sunxi/dram_sun8i_h3.c +5
-2@@ -410,7 +441,7 @@ static void mctl_sys_init(uint16_t socid, struct dram_para *para) setbits_le32(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_RST);udelay (10 ); -writel (0xc00e , &mctl_ctl->clken); +writel (socid ==SOCID_H5 ?0x8000 :0xc00e , &mctl_ctl->clken);udelay (500 ); }@@ -434,7 +465,10 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) /* setting VTC, default disable all VT */ clrbits_le32(&mctl_ctl->pgcr[0], (1 << 30) | 0x3f); -clrsetbits_le32 (&mctl_ctl->pgcr[1 ],1 <<24 ,1 <<26 ); +if (socid ==SOCID_H5 ) +setbits_le32 (&mctl_ctl->pgcr[1 ], (1 <<24 ) | (1 <<26 )); +else +clrsetbits_le32 (&mctl_ctl->pgcr[1 ],1 <<24 ,1 <<26 );/* increase DFI_PHY_UPD clock */ writel(PROTECT_MAGIC, &mctl_com->protect);arch/arm/mach-sunxi/dram_sun8i_h3.c +14
-7@@ -444,15 +478,22 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) udelay(100);/* set dramc odt */ -for (i =0 ; i <4 ; i++) -clrsetbits_le32 (&mctl_ctl->dx[i].gcr, (0x3 <<4 ) | - (0x1 <<1 ) | (0x3 <<2 ) | (0x3 <<12 ) | - (0x3 <<14 ), -IS_ENABLED (CONFIG_DRAM_ODT_EN ) ? -DX_GCR_ODT_DYNAMIC : DX_GCR_ODT_OFF); +for (i =0 ; i <4 ; i++) { +u32 clearmask = (0x3 <<4 ) | (0x1 <<1 ) | (0x3 <<2 ) | + (0x3 <<12 ) | (0x3 <<14 ); +u32 setmask =IS_ENABLED (CONFIG_DRAM_ODT_EN ) ? +DX_GCR_ODT_DYNAMIC :DX_GCR_ODT_OFF ; + +if (socid ==SOCID_H5 ) { + clearmask |=0x2 <<8 ; + setmask |=0x4 <<8 ; + } +clrsetbits_le32 (&mctl_ctl->dx[i].gcr, clearmask, setmask); + }/* AC PDR should always ON */ -setbits_le32 (&mctl_ctl->aciocr,0x1 <<1 ); +clrsetbits_le32 (&mctl_ctl->aciocr, socid ==SOCID_H5 ? (0x1 <<11 ) :0 , +0x1 <<1 );/* set DQS auto gating PD mode */ setbits_le32(&mctl_ctl->pgcr[2], 0x3 << 6);arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -464,7 +505,7 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) /* dphy & aphy phase select 270 degree */ clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 <<8 ), (0x1 << 10) | (0x2 << 8)); - }else if (socid ==SOCID_A64 ) { + }else if (socid ==SOCID_A64 || socid ==SOCID_H5 ) {/* dphy & aphy phase select ? */ clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8), (0x0 << 10) | (0x3 << 8));arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -488,11 +529,12 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) mctl_phy_init(PIR_PLLINIT | PIR_DCAL | PIR_PHYRST | PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE); - }else if (socid ==SOCID_A64 ) { + }else if (socid ==SOCID_A64 || socid ==SOCID_H5 ) {clrsetbits_le32 (&mctl_ctl->zqcr,0xffffff ,CONFIG_DRAM_ZQ ); mctl_phy_init(PIR_ZCAL | PIR_PLLINIT | PIR_DCAL | PIR_PHYRST|
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);arch/arm/mach-sunxi/dram_sun8i_h3.c +2
-1+/* no PIR_QSGATE for H5 ???? */ }/* detect ranks and bus width */ @@ -533,7 +575,7 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para)/* set PGCR3, CKE polarity */ if (socid ==SOCID_H3 )writel (0x00aa0060 , &mctl_ctl->pgcr[3 ]); -else if (socid ==SOCID_A64 ) +else if (socid ==SOCID_A64 || socid ==SOCID_H5 )writel (0xc0aa0060 , &mctl_ctl->pgcr[3 ]);/* power down zq calibration module for power save */arch/arm/mach-sunxi/dram_sun8i_h3.c +16
-0@@ -604,6 +646,22 @@ static void mctl_auto_detect_dram_size(struct dram_para *para) 3, 4, 0, 3, 4, 1, 4, 0, \ 1, 1, 0, 1, 13, 5, 4 } +#define SUN8I_H5_DX_READ_DELAYS \ + {{ 14, 15, 17, 17, 17, 17, 17, 18, 17, 3, 3 }, \ + { 21, 21, 12, 22, 21, 21, 21, 21, 21, 3, 3 }, \ + { 16, 19, 19, 17, 22, 22, 21, 22, 19, 3, 3 }, \ + { 21, 21, 22, 22, 20, 21, 19, 19, 19, 3, 3 } } +#define SUN8I_H5_DX_WRITE_DELAYS \ + {{ 1, 2, 3, 4, 3, 4, 4, 4, 6, 6, 6 }, \ + { 6, 6, 6, 5, 5, 5, 5, 5, 6, 6, 6 }, \ + { 0, 2, 4, 2, 6, 5, 5, 5, 6, 6, 6 }, \ + { 3, 3, 3, 2, 2, 1, 1, 1, 4, 4, 4 } } +#define SUN8I_H5_AC_DELAYS \ + { 0, 0, 5, 5, 0, 0, 0, 0, \ + 0, 0, 0, 0, 3, 3, 3, 3, \ + 3, 3, 3, 3, 3, 3, 3, 3, \ + 3, 3, 3, 3, 2, 0, 0 } +unsigned long sunxi_dram_init (void ) {struct
sunxi_mctl_com_reg* constmctl_com=arch/arm/mach-sunxi/dram_sun8i_h3.c +6
-1@@ -625,6 +683,10 @@ unsigned long sunxi_dram_init(void) .dx_read_delays = SUN50I_A64_DX_READ_DELAYS, .dx_write_delays = SUN50I_A64_DX_WRITE_DELAYS, .ac_delays = SUN50I_A64_AC_DELAYS, +#elif defined(CONFIG_MACH_SUN50I_H5 ) + .dx_read_delays = SUN8I_H5_DX_READ_DELAYS, + .dx_write_delays = SUN8I_H5_DX_WRITE_DELAYS, + .ac_delays = SUN8I_H5_AC_DELAYS,#endif };/* @@ -636,6 +698,8 @@ unsigned long sunxi_dram_init(void)uint16_t socid =SOCID_H3 ;#elif defined(CONFIG_MACH_SUN50I )uint16_t socid =SOCID_A64 ; +#elif defined(CONFIG_MACH_SUN50I_H5 ) +uint16_t socid =SOCID_H5 ;#endif mctl_sys_init (socid, ¶);@@ -652,8 +716,9 @@ unsigned long sunxi_dram_init(void) if (socid == --You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe at googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
[U-Boot] [linux-sunxi] Re: [PATCH v3 09/13] sunxi: DRAM: add Allwinner H5 support
Hi,
Chen-Yu, thanks for your comments.
Quoted 17 lines · Chen-Yu Tsai
Hi,
Quoted 14 lines · Jagan Teki
On Feb 1, 2017 2:37 AM, "Andre Przywara" <[email protected]> wrote:
The DRAM controller in the Allwinner H5 SoC is again very similar to the one in the H3 and A64.
Based on the existing socid parameter, add support for this controller by reusing the bulk of the code and only deviating where needed. These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. Also add the delay line parameters taken from the boot0 and libdram disassembly.
Quoted 5 lines
Quoted 1 line
Can you split this patch with delay line params as separate patch.
It looks like the delay lines are for the H5, merely taken from different sources. They are and should be part of the same patch adding support for DRAM on the H5.
Yeah, I think they really belong into this patch. I don't see how separating them would help, short of creating bisectability problems.
Quoted 65 lines
Quoted 42 lines
Register setup differences between H5 and H3 are courtesy of Jens Kuske.
Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>--- arch/arm/include/asm/arch-sunxi/cpu.h | 1 + arch/arm/mach-sunxi/dram_sun8i_h3.c | 97 +++++++++++++++++++++++++++++------ 2 files changed, 82 insertions(+), 16 deletions(-)arch/arm/include/asm/arch-sunxi/cpu.h +1
-0@@ -15,5 +15,6 @@#define SOCID_A64 0x1689#define SOCID_H3 0x1680 +#define SOCID_H5 0x1718#endif /* _SUNXI_CPU_H */ arch/arm/mach-sunxi/dram_sun8i_h3.c +8
-0@@ -177,6 +177,34 @@ static void mctl_set_master_priority_a64(void)writel (0x81000004 , &mctl_com->mdfs_bwlr[2 ]); } +static void mctl_set_master_priority_h5 (void ) +{ +struct sunxi_mctl_com_reg *const mctl_com = + (struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE ; + +/* enable bandwidth limit windows and set windows size 1us */ +writel (399 , &mctl_com->tmr); +writel ((1 <<16 ), &mctl_com->bwcr);Quoted 3 lines
I'm not fond of using direct numerical values make code unhealthy please use macros with bitops. Note that this comment will apply rest of the code where it applies.
I think you are being unreasonable. The commit message clearly states that the added code either comes from register dumps, disassembled blobs, or comparison of released code:
"""
These new bits set or cleared here and there have been mostly found by looking at DRAM register dumps after using the H5 boot0 and comparing them to what we set in the code. So for now it's mostly unclear what those bits actually mean - hence the missing names and comments. """For this particular instance, see
which gives the exact same comment, and no named macros. Adding a macro for it and calling it
H5_DRAM_BW_UNKNOWN_MAGICis not an improvement. Same goes for the next few lines.
Quoted 5 lines
Allwinner has _never_ released documents for the DRAM controllers or DRAM PHYs, and only sometimes releases code for DRAM init for some SoCs, sometimes with questionable licenses (or lack of), of which some don't match what is actually seen in provided blobs. Considering what the community has access to. This patch seems to be quite good.
I think we have some sort of rewrite of this code ahead of us anyway, hopefully we can address some of these points then in a reasonable manner. But until then I'd like to keep it at "399" instead of using WINDOW_SIZE_1US or _guessing_ how this is computed from some frequency.
Cheers,
Andre.
Quoted 204 lines
Quoted 204 lines
arch/arm/mach-sunxi/dram_sun8i_h3.c +24
-1+ +/* set cpu high priority */ +writel (0x00000001 , &mctl_com->mapr); + +/* Port 2 is reserved per Allwinner's linux-3.10 source, yet +* they initialise it */ +MBUS_CONF (CPU , true,HIGHEST ,0 ,300 ,260 ,150 ); +MBUS_CONF (GPU , true,HIGHEST ,0 ,600 ,400 ,200 ); +MBUS_CONF (UNUSED , true,HIGHEST ,0 ,512 ,256 ,96 ); +MBUS_CONF (DMA , true,HIGHEST ,0 ,256 ,128 ,32 ); +MBUS_CONF (VE , true,HIGHEST ,0 ,1900 ,1500 ,1000 ); +MBUS_CONF (CSI , true,HIGHEST ,0 ,150 ,120 ,100 ); +MBUS_CONF (NAND , true,HIGH ,0 ,256 ,128 ,64 ); +MBUS_CONF (SS , true,HIGHEST ,0 ,256 ,128 ,64 ); +MBUS_CONF (TS , true,HIGHEST ,0 ,256 ,128 ,64 ); +MBUS_CONF (DI , true,HIGH ,0 ,1024 ,256 ,64 ); +MBUS_CONF (DE , true,HIGHEST ,3 ,3400 ,2400 ,1024 ); +MBUS_CONF (DE_CFD , true,HIGHEST ,0 ,600 ,400 ,200 ); +} +static void mctl_set_master_priority (uint16_t socid) {switch (socid) {@@ -186,6 +214,9 @@ static void mctl_set_master_priority(uint16_t socid)case SOCID_A64 :mctl_set_master_priority_a64 ();return ; +case SOCID_H5 : +mctl_set_master_priority_h5 (); +return ; } }@@ -256,7 +287,7 @@ static void mctl_set_timing_params(uint16_t socid, struct dram_para *para) /* set two rank timing */ clrsetbits_le32(&mctl_ctl->dramtmg[8], (0xff << 8) | (0xff << 0), - (0x66 << 8) | (0x10 << 0)); + ((socid ==SOCID_H5 ?0x33 :0x66 ) <<8 ) | (0x10 << 0));/* set PHY interface timing, write latency and read latency configure */ writel((0x2 << 24) | (t_rdata_en << 16) | (0x1 << 8) |arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -391,7 +422,7 @@ static void mctl_sys_init(uint16_t socid, struct dram_para *para) CCM_DRAMCLK_CFG_DIV(1) | CCM_DRAMCLK_CFG_SRC_PLL11 | CCM_DRAMCLK_CFG_UPD); - }else if (socid ==SOCID_H3 ) { + }else if (socid ==SOCID_H3 || socid ==SOCID_H5 ) {clock_set_pll5 (CONFIG_DRAM_CLK *2 *1000000 , false); clrsetbits_le32(&ccm->dram_clk_cfg,
CCM_DRAMCLK_CFG_DIV_MASK|arch/arm/mach-sunxi/dram_sun8i_h3.c +5
-2@@ -410,7 +441,7 @@ static void mctl_sys_init(uint16_t socid, struct dram_para *para) setbits_le32(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_RST);udelay (10 ); -writel (0xc00e , &mctl_ctl->clken); +writel (socid ==SOCID_H5 ?0x8000 :0xc00e , &mctl_ctl->clken);udelay (500 ); }@@ -434,7 +465,10 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) /* setting VTC, default disable all VT */ clrbits_le32(&mctl_ctl->pgcr[0], (1 << 30) | 0x3f); -clrsetbits_le32 (&mctl_ctl->pgcr[1 ],1 <<24 ,1 <<26 ); +if (socid ==SOCID_H5 ) +setbits_le32 (&mctl_ctl->pgcr[1 ], (1 <<24 ) | (1 <<26 )); +else +clrsetbits_le32 (&mctl_ctl->pgcr[1 ],1 <<24 ,1 <<26 );/* increase DFI_PHY_UPD clock */ writel(PROTECT_MAGIC, &mctl_com->protect);arch/arm/mach-sunxi/dram_sun8i_h3.c +14
-7@@ -444,15 +478,22 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) udelay(100);/* set dramc odt */ -for (i =0 ; i <4 ; i++) -clrsetbits_le32 (&mctl_ctl->dx[i].gcr, (0x3 <<4 ) | - (0x1 <<1 ) | (0x3 <<2 ) | (0x3 <<12 ) | - (0x3 <<14 ), -IS_ENABLED (CONFIG_DRAM_ODT_EN ) ? -DX_GCR_ODT_DYNAMIC : DX_GCR_ODT_OFF); +for (i =0 ; i <4 ; i++) { +u32 clearmask = (0x3 <<4 ) | (0x1 <<1 ) | (0x3 <<2 ) | + (0x3 <<12 ) | (0x3 <<14 ); +u32 setmask =IS_ENABLED (CONFIG_DRAM_ODT_EN ) ? +DX_GCR_ODT_DYNAMIC :DX_GCR_ODT_OFF ; + +if (socid ==SOCID_H5 ) { + clearmask |=0x2 <<8 ; + setmask |=0x4 <<8 ; + } +clrsetbits_le32 (&mctl_ctl->dx[i].gcr, clearmask, setmask); + }/* AC PDR should always ON */ -setbits_le32 (&mctl_ctl->aciocr,0x1 <<1 ); +clrsetbits_le32 (&mctl_ctl->aciocr, socid ==SOCID_H5 ? (0x1 <<11 ) :0 , +0x1 <<1 );/* set DQS auto gating PD mode */ setbits_le32(&mctl_ctl->pgcr[2], 0x3 << 6);arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -464,7 +505,7 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) /* dphy & aphy phase select 270 degree */ clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 <<8 ), (0x1 << 10) | (0x2 << 8)); - }else if (socid ==SOCID_A64 ) { + }else if (socid ==SOCID_A64 || socid ==SOCID_H5 ) {/* dphy & aphy phase select ? */ clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8), (0x0 << 10) | (0x3 << 8));arch/arm/mach-sunxi/dram_sun8i_h3.c +1
-1@@ -488,11 +529,12 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para) mctl_phy_init(PIR_PLLINIT | PIR_DCAL | PIR_PHYRST | PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE); - }else if (socid ==SOCID_A64 ) { + }else if (socid ==SOCID_A64 || socid ==SOCID_H5 ) {clrsetbits_le32 (&mctl_ctl->zqcr,0xffffff ,CONFIG_DRAM_ZQ ); mctl_phy_init(PIR_ZCAL | PIR_PLLINIT | PIR_DCAL | PIR_PHYRST|
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);arch/arm/mach-sunxi/dram_sun8i_h3.c +2
-1+/* no PIR_QSGATE for H5 ???? */ }/* detect ranks and bus width */ @@ -533,7 +575,7 @@ static int mctl_channel_init(uint16_t socid, struct dram_para *para)/* set PGCR3, CKE polarity */ if (socid ==SOCID_H3 )writel (0x00aa0060 , &mctl_ctl->pgcr[3 ]); -else if (socid ==SOCID_A64 ) +else if (socid ==SOCID_A64 || socid ==SOCID_H5 )writel (0xc0aa0060 , &mctl_ctl->pgcr[3 ]);/* power down zq calibration module for power save */arch/arm/mach-sunxi/dram_sun8i_h3.c +16
-0@@ -604,6 +646,22 @@ static void mctl_auto_detect_dram_size(struct dram_para *para) 3, 4, 0, 3, 4, 1, 4, 0, \ 1, 1, 0, 1, 13, 5, 4 } +#define SUN8I_H5_DX_READ_DELAYS \ + {{ 14, 15, 17, 17, 17, 17, 17, 18, 17, 3, 3 }, \ + { 21, 21, 12, 22, 21, 21, 21, 21, 21, 3, 3 }, \ + { 16, 19, 19, 17, 22, 22, 21, 22, 19, 3, 3 }, \ + { 21, 21, 22, 22, 20, 21, 19, 19, 19, 3, 3 } } +#define SUN8I_H5_DX_WRITE_DELAYS \ + {{ 1, 2, 3, 4, 3, 4, 4, 4, 6, 6, 6 }, \ + { 6, 6, 6, 5, 5, 5, 5, 5, 6, 6, 6 }, \ + { 0, 2, 4, 2, 6, 5, 5, 5, 6, 6, 6 }, \ + { 3, 3, 3, 2, 2, 1, 1, 1, 4, 4, 4 } } +#define SUN8I_H5_AC_DELAYS \ + { 0, 0, 5, 5, 0, 0, 0, 0, \ + 0, 0, 0, 0, 3, 3, 3, 3, \ + 3, 3, 3, 3, 3, 3, 3, 3, \ + 3, 3, 3, 3, 2, 0, 0 } +unsigned long sunxi_dram_init (void ) {struct
sunxi_mctl_com_reg* constmctl_com=arch/arm/mach-sunxi/dram_sun8i_h3.c +6
-1@@ -625,6 +683,10 @@ unsigned long sunxi_dram_init(void) .dx_read_delays = SUN50I_A64_DX_READ_DELAYS, .dx_write_delays = SUN50I_A64_DX_WRITE_DELAYS, .ac_delays = SUN50I_A64_AC_DELAYS, +#elif defined(CONFIG_MACH_SUN50I_H5 ) + .dx_read_delays = SUN8I_H5_DX_READ_DELAYS, + .dx_write_delays = SUN8I_H5_DX_WRITE_DELAYS, + .ac_delays = SUN8I_H5_AC_DELAYS,#endif };/* @@ -636,6 +698,8 @@ unsigned long sunxi_dram_init(void)uint16_t socid =SOCID_H3 ;#elif defined(CONFIG_MACH_SUN50I )uint16_t socid =SOCID_A64 ; +#elif defined(CONFIG_MACH_SUN50I_H5 ) +uint16_t socid =SOCID_H5 ;#endif mctl_sys_init (socid, ¶);@@ -652,8 +716,9 @@ unsigned long sunxi_dram_init(void) if (socid == --You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe at googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
PATCH v3 10/13 sunxi: prepare for sharing MACH_SUN8I_H3 config symbol
In reply to Andre PrzywaraThe Allwinner H5 is very close to the H3 SoC, but has ARMv8 cores. To allow sharing the clocks, GPIO and driver code easily, create an architecture agnostic MACH_SUNXI_H3_H5 Kconfig symbol. Rename the existing symbol to MACH_SUNXI_H3_H5 where code is shared and let it be selected by a new shared Kconfig option.
--- arch/arm/include/asm/arch-sunxi/clock_sun6i.h | 4 ++-- arch/arm/include/asm/arch-sunxi/cpu_sun4i.h | 4 ++-- arch/arm/include/asm/arch-sunxi/dram.h | 2 +- arch/arm/mach-sunxi/Makefile | 2 +- arch/arm/mach-sunxi/board.c | 2 +- arch/arm/mach-sunxi/clock_sun6i.c | 6 +++--- arch/arm/mach-sunxi/usb_phy.c | 4 ++-- board/sunxi/Kconfig | 14 +++++++++----- drivers/mtd/spi/Kconfig | 2 +- drivers/net/sun8i_emac.c | 2 +- drivers/power/Kconfig | 4 ++-- drivers/usb/host/ehci-sunxi.c | 2 +- 12 files changed, 26 insertions(+), 22 deletions(-)
arch/arm/include/asm/arch-sunxi/clock_sun6i.h +2 -2
@@ -242,7 +242,7 @@ struct sunxi_ccm_reg {/* ahb_gate0 offsets */ #define AHB_GATE_OFFSET_USB_OHCI1 30#define AHB_GATE_OFFSET_USB_OHCI0 29 -#ifdef CONFIG_MACH_SUN8I_H3 +#ifdef CONFIG_MACH_SUNXI_H3_H5 /* * These are EHCI1 - EHCI3 in the datasheet (EHCI0 is for the OTG) we call * them 0 - 2 like they were called on older SoCs.
@@ -293,7 +293,7 @@ struct sunxi_ccm_reg {#define CCM_USB_CTRL_PHY1_CLK (0x1 << 9)#define CCM_USB_CTRL_PHY2_CLK (0x1 << 10)#define CCM_USB_CTRL_PHY3_CLK (0x1 << 11) -#ifdef CONFIG_MACH_SUN8I_H3 +#ifdef CONFIG_MACH_SUNXI_H3_H5 /* * These are OHCI1 - OHCI3 in the datasheet (OHCI0 is for the OTG) we call * them 0 - 2 like they were called on older SoCs.
arch/arm/include/asm/arch-sunxi/cpu_sun4i.h +2 -2
@@ -56,7 +56,7 @@#define SUNXI_USB2_BASE 0x01c1c000#endif #ifdef CONFIG_SUNXI_GEN_SUN6I -#if defined(CONFIG_MACH_SUN8I_H3 ) || defined(CONFIG_MACH_SUN50I ) +#if defined(CONFIG_MACH_SUNXI_H3_H5 ) || defined(CONFIG_MACH_SUN50I )#define SUNXI_USBPHY_BASE 0x01c19000#define SUNXI_USB0_BASE 0x01c1a000#define SUNXI_USB1_BASE 0x01c1b000
@@ -94,7 +94,7 @@#define SUNXI_KEYPAD_BASE 0x01c23000#define SUNXI_TZPC_BASE 0x01c23400 -#if defined(CONFIG_MACH_SUN8I_A83T ) || defined(CONFIG_MACH_SUN8I_H3 ) || \ +#if defined(CONFIG_MACH_SUN8I_A83T ) || defined(CONFIG_MACH_SUNXI_H3_H5 ) || \ defined(CONFIG_MACH_SUN50I )/* SID address space starts at 0x01c1400, but e-fuse is@offset 0x200 */ #define SUNXI_SIDC_BASE 0x01c14000
arch/arm/include/asm/arch-sunxi/dram.h +1 -1
@@ -24,7 +24,7 @@#include <asm/arch/dram_sun8i_a33.h> #elif defined(CONFIG_MACH_SUN8I_A83T )#include <asm/arch/dram_sun8i_a83t.h> -#elif defined(CONFIG_MACH_SUN8I_H3 ) || defined(CONFIG_MACH_SUN50I ) +#elif defined(CONFIG_MACH_SUNXI_H3_H5 ) || defined(CONFIG_MACH_SUN50I )#include <asm/arch/dram_sun8i_h3.h> #elif defined(CONFIG_MACH_SUN9I )#include <asm/arch/dram_sun9i.h>
arch/arm/mach-sunxi/Makefile +1 -1
@@ -48,7 +48,7 @@ obj-$(CONFIG_MACH_SUN7I) += dram_sun4i.o obj-$(CONFIG_MACH_SUN8I_A23) += dram_sun8i_a23.o obj-$(CONFIG_MACH_SUN8I_A33) += dram_sun8i_a33.o obj-$(CONFIG_MACH_SUN8I_A83T) += dram_sun8i_a83t.o -obj-$(CONFIG_MACH_SUN8I_H3) += dram_sun8i_h3.o +obj-$(CONFIG_MACH_SUNXI_H3_H5) += dram_sun8i_h3.o obj-$(CONFIG_MACH_SUN9I) += dram_sun9i.o obj-$(CONFIG_MACH_SUN50I) += dram_sun8i_h3.o endif
arch/arm/mach-sunxi/board.c +1 -1
@@ -98,7 +98,7 @@ static int gpio_init(void)sunxi_gpio_set_cfgpin (SUNXI_GPB (0 ),SUN8I_A33_GPB_UART0 );sunxi_gpio_set_cfgpin (SUNXI_GPB (1 ),SUN8I_A33_GPB_UART0 );sunxi_gpio_set_pull (SUNXI_GPB (1 ),SUNXI_GPIO_PULL_UP ); -#elif CONFIG_CONS_INDEX ==1 && defined(CONFIG_MACH_SUN8I_H3 ) +#elif CONFIG_CONS_INDEX ==1 && defined(CONFIG_MACH_SUNXI_H3_H5 )sunxi_gpio_set_cfgpin (SUNXI_GPA (4 ),SUN8I_H3_GPA_UART0 );sunxi_gpio_set_cfgpin (SUNXI_GPA (5 ),SUN8I_H3_GPA_UART0 );sunxi_gpio_set_pull (SUNXI_GPA (5 ),SUNXI_GPIO_PULL_UP );
arch/arm/mach-sunxi/clock_sun6i.c +3 -3
@@ -22,7 +22,7 @@ void clock_init_safe(void)struct sunxi_ccm_reg *const ccm = (struct sunxi_ccm_reg *)SUNXI_CCM_BASE ; -#if !defined(CONFIG_MACH_SUN8I_H3 ) && !defined(CONFIG_MACH_SUN50I ) +#if !defined(CONFIG_MACH_SUNXI_H3_H5 ) && !defined(CONFIG_MACH_SUN50I )struct sunxi_prcm_reg *const prcm = (struct sunxi_prcm_reg *)SUNXI_PRCM_BASE ;
@@ -51,7 +51,7 @@ void clock_init_safe(void)void clock_init_sec (void ) { -#ifdef CONFIG_MACH_SUN8I_H3 +#ifdef CONFIG_MACH_SUNXI_H3_H5 struct sunxi_ccm_reg *const ccm = (struct sunxi_ccm_reg *)SUNXI_CCM_BASE ;
@@ -152,7 +152,7 @@ void clock_set_pll5(unsigned int clk, bool sigma_delta_enable)const int max_n =32 ;int k =1 , m =2 ; -#ifdef CONFIG_MACH_SUN8I_H3 +#ifdef CONFIG_MACH_SUNXI_H3_H5 clrsetbits_le32 (&ccm->pll5_tuning_cfg,CCM_PLL5_TUN_LOCK_TIME_MASK |CCM_PLL5_TUN_INIT_FREQ_MASK ,CCM_PLL5_TUN_LOCK_TIME (2 ) |CCM_PLL5_TUN_INIT_FREQ (16 ));
arch/arm/mach-sunxi/usb_phy.c +2 -2
@@ -146,10 +146,10 @@ __maybe_unused static void usb_phy_write(struct sunxi_usb_phy *phy, int addr, } } -#if defined(CONFIG_MACH_SUN8I_H3 ) || defined(CONFIG_MACH_SUN50I ) +#if defined(CONFIG_MACH_SUNXI_H3_H5 ) || defined(CONFIG_MACH_SUN50I )static void sunxi_usb_phy_config (struct sunxi_usb_phy *phy) { -#if definedCONFIG_MACH_SUN8I_H3 +#if definedCONFIG_MACH_SUNXI_H3_H5 if (phy->id ==0 )clrbits_le32 (SUNXI_USBPHY_BASE +REG_PHY_UNK_H3 ,0x01 );#endif
board/sunxi/Kconfig +9 -5
@@ -47,6 +47,11 @@ config SUNXI_GEN_SUN6I watchdog, etc. +config MACH_SUNXI_H3_H5 + bool + select SUNXI_GEN_SUN6I + select SUPPORT_SPL + choice prompt "Sunxi SoC Variant" optional
@@ -117,8 +122,7 @@ config MACH_SUN8I_H3 select CPU_V7_HAS_NONSEC select CPU_V7_HAS_VIRT select ARCH_SUPPORT_PSCI - select SUNXI_GEN_SUN6I - select SUPPORT_SPL + select MACH_SUNXI_H3_H5 select ARMV7_BOOT_SEC_DEFAULT if OLD_SUNXI_KERNEL_COMPAT config MACH_SUN9I
@@ -140,7 +144,7 @@ endchoice # The sun8i SoCs share a lot, this helps to avoid a lot of "if A23 || A33" config MACH_SUN8I bool - default y if MACH_SUN8I_A23 || MACH_SUN8I_A33 || MACH_SUN8I_H3 || MACH_SUN8I_A83T + default y if MACH_SUN8I_A23 || MACH_SUN8I_A33 || MACH_SUNXI_H3_H5 || MACH_SUN8I_A83T config RESERVE_ALLWINNER_BOOT0_HEADER bool "reserve space for Allwinner boot0 header"
@@ -328,7 +332,7 @@ config OLD_SUNXI_KERNEL_COMPAT config MMC0_CD_PIN string "Card detect pin for mmc0" - default "PF6" if MACH_SUN8I_A83T || MACH_SUN8I_H3 || MACH_SUN50I + default "PF6" if MACH_SUN8I_A83T || MACH_SUNXI_H3_H5 || MACH_SUN50I default "" ---help--- Set the card detect pin for mmc0, leave empty to not use cd. This
@@ -493,7 +497,7 @@ config AXP_GPIO config VIDEO bool "Enable graphical uboot console on HDMI, LCD or VGA" - depends on !MACH_SUN8I_A83T && !MACH_SUN8I_H3 && !MACH_SUN9I && !MACH_SUN50I + depends on !MACH_SUN8I_A83T && !MACH_SUNXI_H3_H5 && !MACH_SUN9I && !MACH_SUN50I default y ---help--- Say Y here to add support for using a cfb console on the HDMI, LCD
drivers/mtd/spi/Kconfig +1 -1
@@ -132,7 +132,7 @@ if SPL config SPL_SPI_SUNXI bool "Support for SPI Flash on Allwinner SoCs in SPL" - depends on MACH_SUN4I || MACH_SUN5I || MACH_SUN7I || MACH_SUN8I_H3 || MACH_SUN50I + depends on MACH_SUN4I || MACH_SUN5I || MACH_SUN7I || MACH_SUNXI_H3_H5 || MACH_SUN50I ---help--- Enable support for SPI Flash. This option allows SPL to read from sunxi SPI Flash. It uses the same method as the boot ROM, so does
drivers/net/sun8i_emac.c +1 -1
@@ -62,7 +62,7 @@#define AHB_GATE_OFFSET_EPHY 0 -#if defined(CONFIG_MACH_SUN8I_H3 ) +#if defined(CONFIG_MACH_SUNXI_H3_H5 )#define SUN8I_GPD8_GMAC 2#else #define SUN8I_GPD8_GMAC 4
drivers/power/Kconfig +2 -2
@@ -12,7 +12,7 @@ choice default AXP209_POWER if MACH_SUN4I || MACH_SUN5I || MACH_SUN7I default AXP221_POWER if MACH_SUN6I || MACH_SUN8I_A23 || MACH_SUN8I_A33 default AXP818_POWER if MACH_SUN8I_A83T - default SUNXI_NO_PMIC if MACH_SUN8I_H3 || MACH_SUN50I + default SUNXI_NO_PMIC if MACH_SUNXI_H3_H5 || MACH_SUN50I config SUNXI_NO_PMIC bool "board without a pmic"
@@ -60,7 +60,7 @@ config AXP818_POWER config SY8106A_POWER bool "SY8106A pmic support" - depends on MACH_SUN8I_H3 + depends on MACH_SUNXI_H3_H5 ---help--- Select this to enable support for the SY8106A pmic found on some H3 boards.
drivers/usb/host/ehci-sunxi.c +1 -1
@@ -45,7 +45,7 @@ static int ehci_usb_probe(struct udevice *dev)* clocks resp. phys. */ priv->ahb_gate_mask =1 <<AHB_GATE_OFFSET_USB_EHCI0 ; -#if defined(CONFIG_MACH_SUN8I_H3 ) || defined(CONFIG_MACH_SUN50I ) +#if defined(CONFIG_MACH_SUNXI_H3_H5 ) || defined(CONFIG_MACH_SUN50I ) extra_ahb_gate_mask =1 <<AHB_GATE_OFFSET_USB_OHCI0 ;#endif priv->phy_index = ((uintptr_t )hccr -SUNXI_USB1_BASE ) /BASE_DIST ;
Signature · 1 line
2.8.2
PATCH v3 11/13 sunxi: introduce Allwinner H5 config option
In reply to Andre PrzywaraThe Allwinner H5 Soc is bascially an H3 with high SRAM and ARMv8 cores. As the peripherals and the pinmuxing are almost identical, we piggy back on the shared MACH_SUN8I_H3_H5 config symbol.
--- arch/arm/mach-sunxi/cpu_info.c | 2 ++ board/sunxi/Kconfig | 6 ++++++ 2 files changed, 8 insertions(+)
arch/arm/mach-sunxi/cpu_info.c +2 -0
@@ -91,6 +91,8 @@ int print_cpuinfo(void)puts ("CPU: Allwinner A80 (SUN9I)\n" );#elif definedCONFIG_MACH_SUN50I puts ("CPU: Allwinner A64 (SUN50I)\n" ); +#elif definedCONFIG_MACH_SUN50I_H5 +puts ("CPU: Allwinner H5 (SUN50I)\n" );#else #warning Please update cpu_info.c with correct CPU informationputs ("CPU: SUNXI Family\n" );
board/sunxi/Kconfig +6 -0
@@ -139,6 +139,12 @@ config MACH_SUN50I select SUNXI_HIGH_SRAM select SUPPORT_SPL +config MACH_SUN50I_H5 + bool "sun50i (Allwinner H5)" + select ARM64 + select MACH_SUNXI_H3_H5 + select SUNXI_HIGH_SRAM + endchoice # The sun8i SoCs share a lot, this helps to avoid a lot of "if A23 || A33"
Signature · 1 line
2.8.2
PATCH v3 11/13 sunxi: introduce Allwinner H5 config option
Quoted 5 lines · Andre Przywara
The Allwinner H5 Soc is bascially an H3 with high SRAM and ARMv8 cores. As the peripherals and the pinmuxing are almost identical, we piggy back on the shared
MACH_SUN8I_H3_H5config symbol.Signed-off-by Andre Przywara <[email protected]>
Thanks!
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170202/646a504f/attachment.sig>
PATCH v3 12/13 sunxi: dts: add basic OrangePi PC 2 device tree file
In reply to Andre PrzywaraThe OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Create a new .dts file for it by including the (32-bit) H3 SoC .dtsi and changing the differing components accordingly.
This is a preliminary device tree mostly for U-Boot's own sake, it is expected to be updated once the official DT gets accepted upstream.
--- arch/arm/dts/Makefile | 2 + arch/arm/dts/sun50i-h5-orangepi-pc2.dts | 147 ++++++++++++++++++++++++++++++++ 2 files changed, 149 insertions(+) create mode 100644 arch/arm/dts/sun50i-h5-orangepi-pc2.dts
arch/arm/dts/Makefile +2 -0
@@ -294,6 +294,8 @@ dtb-$(CONFIG_MACH_SUN8I_H3) += \ sun8i-h3-orangepi-plus.dtb \ sun8i-h3-orangepi-plus2e.dtb \ sun8i-h3-nanopi-neo.dtb +dtb-$(CONFIG_MACH_SUN50I_H5) += \ + sun50i-h5-orangepi-pc2.dtb dtb-$(CONFIG_MACH_SUN50I) += \ sun50i-a64-pine64-plus.dtb \ sun50i-a64-pine64.dtb
arch/arm/dts/sun50i-h5-orangepi-pc2.dts new file +147 -0
147 lines not on this page · raw message
Signature · 1 line
2.8.2
PATCH v3 12/13 sunxi: dts: add basic OrangePi PC 2 device tree file
Quoted 7 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Create a new .dts file for it by including the (32-bit) H3 SoC .dtsi and changing the differing components accordingly.
This is a preliminary device tree mostly for U-Boot's own sake, it is expected to be updated once the official DT gets accepted upstream.Signed-off-by Andre Przywara <[email protected]>
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170202/b8923a67/attachment.sig>
PATCH v3 12/13 sunxi: dts: add basic OrangePi PC 2 device tree file
In reply to Andre PrzywaraQuoted 10 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Create a new .dts file for it by including the (32-bit) H3 SoC .dtsi and changing the differing components accordingly.
This is a preliminary device tree mostly for U-Boot's own sake, it is expected to be updated once the official DT gets accepted upstream.Signed-off-by Andre Przywara <[email protected]>--- arch/arm/dts/Makefile | 2 + arch/arm/dts/sun50i-h5-orangepi-pc2.dts | 147 ++++++++++++++++++++++++++++++++
Please squash 13/13 with this, I would see a single patch for initial support.
thanks!
Signature · 4 lines
Jagan Teki Free Software Engineer | www.openedev.com U-Boot, Linux | Upstream Maintainer Hyderabad, India.
PATCH v3 12/13 sunxi: dts: add basic OrangePi PC 2 device tree file
Hi,
Quoted 7 lines · Jagan Teki
Quoted 6 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Create a new .dts file for it by including the (32-bit) H3 SoC .dtsi and changing the differing components accordingly.
This is a preliminary device tree mostly for U-Boot's own sake, it is expected to be updated once the official DT gets accepted upstream.
Quoted 6 lines
Quoted 4 lines
Signed-off-by Andre Przywara <[email protected]>--- arch/arm/dts/Makefile | 2 + arch/arm/dts/sun50i-h5-orangepi-pc2.dts | 147 ++++++++++++++++++++++++++++++++Please squash 13/13 with this, I would see a single patch for initial support.
How comes?
I think those two are really separate topics, and having a DT file in this directory really doesn't hurt anything, until it actually gets referenced in the next patch.
I'd keep DT patches separate, really, and in general always would prefer more, but smaller patches to fewer, but bigger ones.
Cheers,
Andre.
PATCH v3 12/13 sunxi: dts: add basic OrangePi PC 2 device tree file
Quoted 15 lines · Andre Przywara
Hi,
Quoted 12 lines · Jagan Teki
Quoted 10 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Create a new .dts file for it by including the (32-bit) H3 SoC .dtsi and changing the differing components accordingly.
This is a preliminary device tree mostly for U-Boot's own sake, it is expected to be updated once the official DT gets accepted upstream.Signed-off-by Andre Przywara <[email protected]>--- arch/arm/dts/Makefile | 2 + arch/arm/dts/sun50i-h5-orangepi-pc2.dts | 147 ++++++++++++++++++++++++++++++++
Quoted 8 lines
Quoted 1 line
Please squash 13/13 with this, I would see a single patch for initial support.
How comes?
I think those two are really separate topics, and having a DT file in this directory really doesn't hurt anything, until it actually gets referenced in the next patch.
I'd keep DT patches separate, really, and in general always would prefer more, but smaller patches to fewer, but bigger ones.
I concur.
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170207/d06d9de0/attachment.sig>
PATCH v3 13/13 sunxi: configs: add basic OrangePi PC 2 defconfig
In reply to Andre PrzywaraThe OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Add a (64-bit only) defconfig defining the required options to build the U-Boot proper.
--- board/sunxi/MAINTAINERS | 5 +++++ configs/orangepi_pc2_defconfig | 16 ++++++++++++++++ 2 files changed, 21 insertions(+) create mode 100644 configs/orangepi_pc2_defconfig
board/sunxi/MAINTAINERS +5 -0
@@ -242,6 +242,11 @@ M: Icenowy Zheng <[email protected]> S: Maintained F: configs/orangepi_zero_defconfig +ORANGEPI PC 2 BOARD +M: Andre Przywara <[email protected]> +S: Maintained +F: configs/orangepi_pc2_defconfig + R16 EVB PARROT BOARD M: Quentin Schulz <[email protected]> S: Maintained
configs/orangepi_pc2_defconfig new file +16 -0
@@ -0,0 +1,16 @@ +CONFIG_ARM=y +CONFIG_ARCH_SUNXI=y +CONFIG_MACH_SUN50I_H5=y +CONFIG_SPL=y +CONFIG_DRAM_CLK=672 +CONFIG_DRAM_ZQ=3881977 +CONFIG_DEFAULT_DEVICE_TREE="sun50i-h5-orangepi-pc2" +# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set +CONFIG_CONSOLE_MUX=y +# CONFIG_CMD_IMLS is not set +# CONFIG_CMD_FLASH is not set +# CONFIG_CMD_FPGA is not set +CONFIG_SUN8I_EMAC=y +CONFIG_USB_EHCI_HCD=y +CONFIG_SPL_SPI_SUNXI=y +CONFIG_SPL_SPI_FLASH_SUPPORT=y
Signature · 1 line
2.8.2
PATCH v3 13/13 sunxi: configs: add basic OrangePi PC 2 defconfig
Quoted 5 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Add a (64-bit only) defconfig defining the required options to build the U-Boot proper.
Signed-off-by Andre Przywara <[email protected]>
Maxime
Signature · 10 lines
Maxime Ripard, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170202/95497315/attachment.sig>
PATCH v3 13/13 sunxi: configs: add basic OrangePi PC 2 defconfig
Quoted 8 lines · Maxime Ripard
Quoted 5 lines · Andre Przywara
The OrangePi PC 2 is a typical SBC with the 64-bit Allwinner H5 SoC. Add a (64-bit only) defconfig defining the required options to build the U-Boot proper.
Signed-off-by Andre Przywara <[email protected]>Acked-by Maxime Ripard <[email protected]>
Thanks a lot!
I realised yesterday that rebasing against current master requires three additional not-defines to be added - for the SPL partitions stuff. So we need:
# CONFIG_SPL_DOS_PARTITION is not set
# CONFIG_SPL_ISO_PARTITION is not set
# CONFIG_SPL_EFI_PARTITION is not set
in the _defconfig to avoid pulling that code in and blowing up the SPL size.
I would like to wait how the fix for the SPL stack issue will look like and then decide whether I send a v4 or a fixup patch.
Cheers,
Andre.