Mirrors/u-boot
2
0
Fork 0
mirror of https://github.com/AsahiLinux/u-boot synced 2024-11-15 01:17:39 +00:00
Code Issues Projects Releases Packages Wiki Activity
u-boot/arch/arm/mach-sunxi/dram_sunxi_dw.c

949 lines
30 KiB
C
Raw Normal View History

SPDX: Convert all of our single license tags to Linux Kernel style When U-Boot started using SPDX tags we were among the early adopters and there weren't a lot of other examples to borrow from. So we picked the area of the file that usually had a full license text and replaced it with an appropriate SPDX-License-Identifier: entry. Since then, the Linux Kernel has adopted SPDX tags and they place it as the very first line in a file (except where shebangs are used, then it's second line) and with slightly different comment styles than us. In part due to community overlap, in part due to better tag visibility and in part for other minor reasons, switch over to that style. This commit changes all instances where we have a single declared license in the tag as both the before and after are identical in tag contents. There's also a few places where I found we did not have a tag and have introduced one. Signed-off-by: Tom Rini <trini@konsulko.com>
2018-05-06 21:58:06 +00:00
// SPDX-License-Identifier: GPL-2.0+
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/*
* sun8i H3 platform dram controller init
*
* (C) Copyright 2007-2015 Allwinner Technology Co.
* Jerry Wang <wangflord@allwinnertech.com>
* (C) Copyright 2015 Vishnu Patekar <vishnupatekar0510@gmail.com>
* (C) Copyright 2015 Hans de Goede <hdegoede@redhat.com>
* (C) Copyright 2015 Jens Kuske <jenskuske@gmail.com>
*/
#include <common.h>
common: Drop init.h from common header Move this uncommon header out of the common header. Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-10 17:40:02 +00:00
#include <init.h>
common: Drop log.h from common header Move this header out of the common header. Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-10 17:40:05 +00:00
#include <log.h>
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
#include <asm/io.h>
#include <asm/arch/clock.h>
#include <asm/arch/dram.h>
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#include <asm/arch/cpu.h>
common: Drop linux/delay.h from common header Move this uncommon header out of the common header. Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-10 17:40:11 +00:00
#include <linux/delay.h>
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
#include <linux/kconfig.h>
static void mctl_phy_init(u32 val)
{
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
writel(val | PIR_INIT, &mctl_ctl->pir);
mctl_await_completion(&mctl_ctl->pgsr[0], PGSR_INIT_DONE, 0x1);
}
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
static void mctl_set_bit_delays(struct dram_para *para)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
int i, j;
clrbits_le32(&mctl_ctl->pgcr[0], 1 << 26);
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
for (i = 0; i < NR_OF_BYTE_LANES; i++)
for (j = 0; j < LINES_PER_BYTE_LANE; j++)
writel(DXBDLR_WRITE_DELAY(para->dx_write_delays[i][j]) |
DXBDLR_READ_DELAY(para->dx_read_delays[i][j]),
&mctl_ctl->dx[i].bdlr[j]);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
for (i = 0; i < 31; i++)
writel(ACBDLR_WRITE_DELAY(para->ac_delays[i]),
&mctl_ctl->acbdlr[i]);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
#ifdef CONFIG_MACH_SUN8I_R40
/* DQSn, DMn, DQn output enable bit delay */
for (i = 0; i < 4; i++)
writel(0x6 << 24, &mctl_ctl->dx[i].sdlr);
#endif
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
setbits_le32(&mctl_ctl->pgcr[0], 1 << 26);
}
sunxi: H3: Rework MBUS priority setup So far the MBUS priority setup was done by writing "magic" values taken from a DRAM controller register dump after a boot0 run. By peeking at the Linux (sic!) MBUS driver [1] from the Allwinner BSP kernel, we learned more about the actual meaning of those bits. Add macros and refactor the setup function to make the MBUS setup much more readable and meaningful. The actual values used now are a transformation of the values used before, which are assembled by the new code to result in the same register writes. So this rework does not change any settings, also the code size stays the same. The respective source files in the BSP kernel had a proper GPL header, so lifting this code and information into U-Boot is legal. [Andre: provide a convenience macro to fit definitions on one line] [1] https://github.com/longsleep/linux-pine64/blob/lichee-dev-v3.10.65/drivers/bus/sunxi_mbus.c Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:38 +00:00
enum {
MBUS_PORT_CPU = 0,
MBUS_PORT_GPU = 1,
MBUS_PORT_UNUSED = 2,
MBUS_PORT_DMA = 3,
MBUS_PORT_VE = 4,
MBUS_PORT_CSI = 5,
MBUS_PORT_NAND = 6,
MBUS_PORT_SS = 7,
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
MBUS_PORT_DE_V3S = 8,
MBUS_PORT_DE_CFD_V3S = 9,
sunxi: H3: Rework MBUS priority setup So far the MBUS priority setup was done by writing "magic" values taken from a DRAM controller register dump after a boot0 run. By peeking at the Linux (sic!) MBUS driver [1] from the Allwinner BSP kernel, we learned more about the actual meaning of those bits. Add macros and refactor the setup function to make the MBUS setup much more readable and meaningful. The actual values used now are a transformation of the values used before, which are assembled by the new code to result in the same register writes. So this rework does not change any settings, also the code size stays the same. The respective source files in the BSP kernel had a proper GPL header, so lifting this code and information into U-Boot is legal. [Andre: provide a convenience macro to fit definitions on one line] [1] https://github.com/longsleep/linux-pine64/blob/lichee-dev-v3.10.65/drivers/bus/sunxi_mbus.c Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:38 +00:00
MBUS_PORT_TS = 8,
MBUS_PORT_DI = 9,
MBUS_PORT_DE = 10,
MBUS_PORT_DE_CFD = 11,
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
MBUS_PORT_UNKNOWN1 = 12,
MBUS_PORT_UNKNOWN2 = 13,
MBUS_PORT_UNKNOWN3 = 14,
sunxi: H3: Rework MBUS priority setup So far the MBUS priority setup was done by writing "magic" values taken from a DRAM controller register dump after a boot0 run. By peeking at the Linux (sic!) MBUS driver [1] from the Allwinner BSP kernel, we learned more about the actual meaning of those bits. Add macros and refactor the setup function to make the MBUS setup much more readable and meaningful. The actual values used now are a transformation of the values used before, which are assembled by the new code to result in the same register writes. So this rework does not change any settings, also the code size stays the same. The respective source files in the BSP kernel had a proper GPL header, so lifting this code and information into U-Boot is legal. [Andre: provide a convenience macro to fit definitions on one line] [1] https://github.com/longsleep/linux-pine64/blob/lichee-dev-v3.10.65/drivers/bus/sunxi_mbus.c Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:38 +00:00
};
enum {
MBUS_QOS_LOWEST = 0,
MBUS_QOS_LOW,
MBUS_QOS_HIGH,
MBUS_QOS_HIGHEST
};
sunxi: Silence warning about non-static inline function When compiling with CONFIG_SPL_SERIAL=n, gcc warns about mbus_configure_port not being marked as static: In file included from include/common.h:34, from arch/arm/mach-sunxi/dram_sunxi_dw.c:11: include/log.h:185:4: warning: 'printf' is static but used in inline function 'mbus_configure_port' which is not static 185 | printf(pr_fmt(fmt), ##args); \ | ^~~~~~ include/log.h:192:2: note: in expansion of macro 'debug_cond' 192 | debug_cond(_DEBUG, fmt, ##args) | ^~~~~~~~~~ arch/arm/mach-sunxi/dram_sunxi_dw.c:100:2: note: in expansion of macro 'debug' 100 | debug("MBUS port %d cfg0 %08x cfg1 %08x\n", port, cfg0, cfg1); | ^~~~~ Fix this by updating the function accordingly. Signed-off-by: Samuel Holland <samuel@sholland.org> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
2020-05-07 23:02:33 +00:00
static inline void mbus_configure_port(u8 port,
bool bwlimit,
bool priority,
u8 qos, /* MBUS_QOS_LOWEST .. MBUS_QOS_HIGEST */
u8 waittime, /* 0 .. 0xf */
u8 acs, /* 0 .. 0xff */
u16 bwl0, /* 0 .. 0xffff, bandwidth limit in MB/s */
u16 bwl1,
u16 bwl2)
sunxi: H3: Rework MBUS priority setup So far the MBUS priority setup was done by writing "magic" values taken from a DRAM controller register dump after a boot0 run. By peeking at the Linux (sic!) MBUS driver [1] from the Allwinner BSP kernel, we learned more about the actual meaning of those bits. Add macros and refactor the setup function to make the MBUS setup much more readable and meaningful. The actual values used now are a transformation of the values used before, which are assembled by the new code to result in the same register writes. So this rework does not change any settings, also the code size stays the same. The respective source files in the BSP kernel had a proper GPL header, so lifting this code and information into U-Boot is legal. [Andre: provide a convenience macro to fit definitions on one line] [1] https://github.com/longsleep/linux-pine64/blob/lichee-dev-v3.10.65/drivers/bus/sunxi_mbus.c Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:38 +00:00
{
struct sunxi_mctl_com_reg * const mctl_com =
(struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE;
const u32 cfg0 = ( (bwlimit ? (1 << 0) : 0)
| (priority ? (1 << 1) : 0)
| ((qos & 0x3) << 2)
| ((waittime & 0xf) << 4)
| ((acs & 0xff) << 8)
| (bwl0 << 16) );
const u32 cfg1 = ((u32)bwl2 << 16) | (bwl1 & 0xffff);
debug("MBUS port %d cfg0 %08x cfg1 %08x\n", port, cfg0, cfg1);
writel(cfg0, &mctl_com->mcr[port][0]);
writel(cfg1, &mctl_com->mcr[port][1]);
}
#define MBUS_CONF(port, bwlimit, qos, acs, bwl0, bwl1, bwl2) \
mbus_configure_port(MBUS_PORT_ ## port, bwlimit, false, \
MBUS_QOS_ ## qos, 0, acs, bwl0, bwl1, bwl2)
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
static void mctl_set_master_priority_h3(void)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
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 */
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
writel((1 << 16) | (400 << 0), &mctl_com->bwcr);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* set cpu high priority */
writel(0x00000001, &mctl_com->mapr);
sunxi: H3: Rework MBUS priority setup So far the MBUS priority setup was done by writing "magic" values taken from a DRAM controller register dump after a boot0 run. By peeking at the Linux (sic!) MBUS driver [1] from the Allwinner BSP kernel, we learned more about the actual meaning of those bits. Add macros and refactor the setup function to make the MBUS setup much more readable and meaningful. The actual values used now are a transformation of the values used before, which are assembled by the new code to result in the same register writes. So this rework does not change any settings, also the code size stays the same. The respective source files in the BSP kernel had a proper GPL header, so lifting this code and information into U-Boot is legal. [Andre: provide a convenience macro to fit definitions on one line] [1] https://github.com/longsleep/linux-pine64/blob/lichee-dev-v3.10.65/drivers/bus/sunxi_mbus.c Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:38 +00:00
MBUS_CONF( CPU, true, HIGHEST, 0, 512, 256, 128);
MBUS_CONF( GPU, true, HIGH, 0, 1536, 1024, 256);
MBUS_CONF(UNUSED, true, HIGHEST, 0, 512, 256, 96);
MBUS_CONF( DMA, true, HIGHEST, 0, 256, 128, 32);
MBUS_CONF( VE, true, HIGH, 0, 1792, 1600, 256);
MBUS_CONF( CSI, true, HIGHEST, 0, 256, 128, 32);
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, 8192, 6120, 1024);
MBUS_CONF(DE_CFD, true, HIGH, 0, 1024, 288, 64);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
static void mctl_set_master_priority_v3s(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((1 << 16) | (400 << 0), &mctl_com->bwcr);
/* set cpu high priority */
writel(0x00000001, &mctl_com->mapr);
MBUS_CONF( CPU, true, HIGHEST, 0, 160, 100, 80);
MBUS_CONF( GPU, true, HIGH, 0, 1792, 1536, 0);
MBUS_CONF( UNUSED, true, HIGHEST, 0, 256, 128, 80);
MBUS_CONF( DMA, true, HIGH, 0, 256, 100, 0);
MBUS_CONF( VE, true, HIGH, 0, 2048, 1600, 0);
MBUS_CONF( CSI, true, HIGHEST, 0, 384, 256, 0);
MBUS_CONF( NAND, true, HIGH, 0, 100, 50, 0);
MBUS_CONF( SS, true, HIGH, 0, 384, 256, 0);
MBUS_CONF( DE_V3S, false, HIGH, 0, 8192, 4096, 0);
MBUS_CONF(DE_CFD_V3S, true, HIGH, 0, 640, 256, 0);
}
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
static void mctl_set_master_priority_a64(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);
/* Port 2 is reserved per Allwinner's linux-3.10 source, yet they
* initialise it */
MBUS_CONF( CPU, true, HIGHEST, 0, 160, 100, 80);
MBUS_CONF( GPU, false, HIGH, 0, 1536, 1400, 256);
MBUS_CONF(UNUSED, true, HIGHEST, 0, 512, 256, 96);
MBUS_CONF( DMA, true, HIGH, 0, 256, 80, 100);
MBUS_CONF( VE, true, HIGH, 0, 1792, 1600, 256);
MBUS_CONF( CSI, true, HIGH, 0, 256, 128, 0);
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, HIGH, 2, 8192, 6144, 2048);
MBUS_CONF(DE_CFD, true, HIGH, 0, 1280, 144, 64);
writel(0x81000004, &mctl_com->mdfs_bwlr[2]);
}
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
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);
/* 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);
}
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
static void mctl_set_master_priority_r40(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);
/* 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);
/*
* The port names are probably wrong, but no correct sources
* are available.
*/
MBUS_CONF( DE, true, HIGH, 0, 128, 48, 0);
MBUS_CONF( DE_CFD, true, HIGH, 0, 384, 256, 0);
MBUS_CONF(UNKNOWN1, true, HIGHEST, 0, 512, 384, 256);
MBUS_CONF(UNKNOWN2, true, HIGHEST, 2, 8192, 6144, 1024);
MBUS_CONF(UNKNOWN3, true, HIGH, 0, 1280, 144, 64);
}
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
static void mctl_set_master_priority(uint16_t socid)
{
switch (socid) {
case SOCID_H3:
mctl_set_master_priority_h3();
return;
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
case SOCID_V3S:
mctl_set_master_priority_v3s();
return;
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
case SOCID_A64:
mctl_set_master_priority_a64();
return;
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
case SOCID_H5:
mctl_set_master_priority_h5();
return;
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
case SOCID_R40:
mctl_set_master_priority_r40();
return;
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
}
}
static u32 bin_to_mgray(int val)
{
static const u8 lookup_table[32] = {
0x00, 0x01, 0x02, 0x03, 0x06, 0x07, 0x04, 0x05,
0x0c, 0x0d, 0x0e, 0x0f, 0x0a, 0x0b, 0x08, 0x09,
0x18, 0x19, 0x1a, 0x1b, 0x1e, 0x1f, 0x1c, 0x1d,
0x14, 0x15, 0x16, 0x17, 0x12, 0x13, 0x10, 0x11,
};
return lookup_table[clamp(val, 0, 31)];
}
static int mgray_to_bin(u32 val)
{
static const u8 lookup_table[32] = {
0x00, 0x01, 0x02, 0x03, 0x06, 0x07, 0x04, 0x05,
0x0e, 0x0f, 0x0c, 0x0d, 0x08, 0x09, 0x0a, 0x0b,
0x1e, 0x1f, 0x1c, 0x1d, 0x18, 0x19, 0x1a, 0x1b,
0x10, 0x11, 0x12, 0x13, 0x16, 0x17, 0x14, 0x15,
};
return lookup_table[val & 0x1f];
}
static void mctl_h3_zq_calibration_quirk(struct dram_para *para)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
int zq_count;
#if defined CONFIG_SUNXI_DRAM_DW_16BIT
zq_count = 4;
#else
zq_count = 6;
#endif
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
if ((readl(SUNXI_SRAMC_BASE + 0x24) & 0xff) == 0 &&
(readl(SUNXI_SRAMC_BASE + 0xf0) & 0x1) == 0) {
u32 reg_val;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
clrsetbits_le32(&mctl_ctl->zqcr, 0xffff,
CONFIG_DRAM_ZQ & 0xffff);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
writel(PIR_CLRSR, &mctl_ctl->pir);
mctl_phy_init(PIR_ZCAL);
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
reg_val = readl(&mctl_ctl->zqdr[0]);
reg_val &= (0x1f << 16) | (0x1f << 0);
reg_val |= reg_val << 8;
writel(reg_val, &mctl_ctl->zqdr[0]);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
reg_val = readl(&mctl_ctl->zqdr[1]);
reg_val &= (0x1f << 16) | (0x1f << 0);
reg_val |= reg_val << 8;
writel(reg_val, &mctl_ctl->zqdr[1]);
writel(reg_val, &mctl_ctl->zqdr[2]);
} else {
int i;
u16 zq_val[6];
u8 val;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
writel(0x0a0a0a0a, &mctl_ctl->zqdr[2]);
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
for (i = 0; i < zq_count; i++) {
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
u8 zq = (CONFIG_DRAM_ZQ >> (i * 4)) & 0xf;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
writel((zq << 20) | (zq << 16) | (zq << 12) |
(zq << 8) | (zq << 4) | (zq << 0),
&mctl_ctl->zqcr);
writel(PIR_CLRSR, &mctl_ctl->pir);
mctl_phy_init(PIR_ZCAL);
zq_val[i] = readl(&mctl_ctl->zqdr[0]) & 0xff;
writel(REPEAT_BYTE(zq_val[i]), &mctl_ctl->zqdr[2]);
writel(PIR_CLRSR, &mctl_ctl->pir);
mctl_phy_init(PIR_ZCAL);
val = readl(&mctl_ctl->zqdr[0]) >> 24;
zq_val[i] |= bin_to_mgray(mgray_to_bin(val) - 1) << 8;
}
writel((zq_val[1] << 16) | zq_val[0], &mctl_ctl->zqdr[0]);
writel((zq_val[3] << 16) | zq_val[2], &mctl_ctl->zqdr[1]);
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
if (zq_count > 4)
writel((zq_val[5] << 16) | zq_val[4],
&mctl_ctl->zqdr[2]);
sunxi: Fix H3 DRAM impedance calibration on rev. A chips H3 seems to have a silicon bug breaking the impedance calibration. This is currently worked around in software by multiple steps combining the results to replace the wrong values. Revision A chips need a different workaround, which is present in the vendor bootloader too, but got overlooked in lack of information and affected boards till now. This commit adds a simplified version without correction factor, which would be 1.00 for all known boards anyway. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Jagan Teki <jteki@openedev.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2016-09-21 18:08:30 +00:00
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
static void mctl_v3s_zq_calibration_quirk(struct dram_para *para)
{
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
u32 reg_val;
clrsetbits_le32(&mctl_ctl->zqcr, 0xffffff,
CONFIG_DRAM_ZQ & 0xffffff);
mctl_phy_init(PIR_ZCAL);
reg_val = readl(&mctl_ctl->zqdr[0]);
reg_val &= (0x1f << 16) | (0x1f << 0);
reg_val |= reg_val << 8;
writel(reg_val, &mctl_ctl->zqdr[0]);
reg_val = readl(&mctl_ctl->zqdr[1]);
reg_val &= (0x1f << 16) | (0x1f << 0);
reg_val |= reg_val << 8;
writel(reg_val, &mctl_ctl->zqdr[1]);
}
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
static void mctl_set_cr(uint16_t socid, struct dram_para *para)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
struct sunxi_mctl_com_reg * const mctl_com =
(struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE;
sunxi: Add selective DRAM type and timing DRAM chip varies, and one code cannot satisfy all DRAMs. Add options to select a timing set. Currently only DDR3-1333 (the original set) is added into it. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:18 +00:00
writel(MCTL_CR_BL8 | MCTL_CR_INTERLEAVED |
#if defined CONFIG_SUNXI_DRAM_DDR3
MCTL_CR_DDR3 | MCTL_CR_2T |
sunxi: add support for the DDR2 in V3s SoC Allwinner V3s SoC features a co-packaged DDR2 DRAM chip, which needs its timing param. Add support for it. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Acked-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:20 +00:00
#elif defined CONFIG_SUNXI_DRAM_DDR2
MCTL_CR_DDR2 | MCTL_CR_2T |
sunxi: add LPDDR3 DRAM type support for DesignWare-like DRAM controller Some A64 boards (SoPine and Pinebook production batch) use LPDDR3 DRAM chips. Add support for LPDDR3 DRAM in the DesignWare-like DRAM controller code. Real LPDDR3 chips' support is not added yet in this commit. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:23 +00:00
#elif defined CONFIG_SUNXI_DRAM_LPDDR3
MCTL_CR_LPDDR3 | MCTL_CR_1T |
sunxi: Add selective DRAM type and timing DRAM chip varies, and one code cannot satisfy all DRAMs. Add options to select a timing set. Currently only DDR3-1333 (the original set) is added into it. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:18 +00:00
#else
#error Unsupported DRAM type!
#endif
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
(para->ranks[0].bank_bits == 3 ? MCTL_CR_EIGHT_BANKS : MCTL_CR_FOUR_BANKS) |
sunxi: Rename bus-width related macros in H3 DRAM code The DesignWare DRAM controller used by H3 and newer SoCs use a bit to identify whether the DRAM is half-width. As H3 itself come with 32-bit DRAM, the two modes of the bit used to be named "MCTL_CR_32BIT" and "MCTL_CR_16BIT", but for SoCs with 16-bit DRAM they're really 8-bit and 16-bit. Rename the bit's macro, and also rename the variable name in dram_sun8i_h3.c. This commit do not add 16-bit DRAM controller support, but the support will be introduced in next commit. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:15 +00:00
MCTL_CR_BUS_FULL_WIDTH(para->bus_full_width) |
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
(para->dual_rank ? MCTL_CR_DUAL_RANK : MCTL_CR_SINGLE_RANK) |
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
MCTL_CR_PAGE_SIZE(para->ranks[0].page_size) |
MCTL_CR_ROW_BITS(para->ranks[0].row_bits), &mctl_com->cr);
if (para->dual_rank && (socid == SOCID_A64 || socid == SOCID_R40)) {
writel((para->ranks[1].bank_bits == 3 ? MCTL_CR_EIGHT_BANKS : MCTL_CR_FOUR_BANKS) |
MCTL_CR_BUS_FULL_WIDTH(para->bus_full_width) |
MCTL_CR_DUAL_RANK |
MCTL_CR_PAGE_SIZE(para->ranks[1].page_size) |
MCTL_CR_ROW_BITS(para->ranks[1].row_bits), &mctl_com->cr_r1);
}
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
if (socid == SOCID_R40) {
/* Mux pin to A15 address line for single rank memory. */
sunxi: enable dual rank memory on R40 Previously we do not have proper dual rank memory detection on R40 (because we omitted PIR_QSGATE, which does not work on R40 with our configuration), and dual rank memory is just simply disabled as early R40 boards available (Banana Pi M2 Ultra and Berry) have single rank memory. As a board with dual rank memory (Forlinx OKA40i-C) is now known to us, we need to have a way to do memory rank detection to support that board. Add some routine to detect memory rank by trying to access the memory in rank 1 and check for error status of the memory controller, and then enable dual rank memory on R40. Similar routine can be used to detect half DQ width (which is also detected by PIR_QSGATE on other SoCs), but it's left unimplemented because there's no known R40 board with half DQ width now. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> [Andre: Move R40 detect code call into sunxi_dram_init()] Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:25 +00:00
if (!para->dual_rank)
setbits_le32(&mctl_com->cr_r1, MCTL_CR_R1_MUX_A15);
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
static void mctl_sys_init(uint16_t socid, struct dram_para *para)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
struct sunxi_ccm_reg * const ccm =
(struct sunxi_ccm_reg *)SUNXI_CCM_BASE;
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
clrbits_le32(&ccm->mbus0_clk_cfg, MBUS_CLK_GATE);
clrbits_le32(&ccm->mbus_reset, CCM_MBUS_RESET_RESET);
clrbits_le32(&ccm->ahb_gate0, 1 << AHB_GATE_OFFSET_MCTL);
clrbits_le32(&ccm->ahb_reset0_cfg, 1 << AHB_RESET_OFFSET_MCTL);
clrbits_le32(&ccm->pll5_cfg, CCM_PLL5_CTRL_EN);
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
if (socid == SOCID_A64 || socid == SOCID_R40)
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
clrbits_le32(&ccm->pll11_cfg, CCM_PLL11_CTRL_EN);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
udelay(10);
clrbits_le32(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_RST);
udelay(1000);
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
if (socid == SOCID_A64 || socid == SOCID_R40) {
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
clock_set_pll11(CONFIG_DRAM_CLK * 2 * 1000000, false);
clrsetbits_le32(&ccm->dram_clk_cfg,
CCM_DRAMCLK_CFG_DIV_MASK |
CCM_DRAMCLK_CFG_SRC_MASK,
CCM_DRAMCLK_CFG_DIV(1) |
CCM_DRAMCLK_CFG_SRC_PLL11 |
CCM_DRAMCLK_CFG_UPD);
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
} else if (socid == SOCID_H3 || socid == SOCID_H5 || socid == SOCID_V3S) {
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
clock_set_pll5(CONFIG_DRAM_CLK * 2 * 1000000, false);
clrsetbits_le32(&ccm->dram_clk_cfg,
CCM_DRAMCLK_CFG_DIV_MASK |
CCM_DRAMCLK_CFG_SRC_MASK,
CCM_DRAMCLK_CFG_DIV(1) |
CCM_DRAMCLK_CFG_SRC_PLL5 |
CCM_DRAMCLK_CFG_UPD);
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
mctl_await_completion(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_UPD, 0);
setbits_le32(&ccm->ahb_reset0_cfg, 1 << AHB_RESET_OFFSET_MCTL);
setbits_le32(&ccm->ahb_gate0, 1 << AHB_GATE_OFFSET_MCTL);
setbits_le32(&ccm->mbus_reset, CCM_MBUS_RESET_RESET);
setbits_le32(&ccm->mbus0_clk_cfg, MBUS_CLK_GATE);
setbits_le32(&ccm->dram_clk_cfg, CCM_DRAMCLK_CFG_RST);
udelay(10);
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
writel(socid == SOCID_H5 ? 0x8000 : 0xc00e, &mctl_ctl->clken);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
udelay(500);
}
sunxi: H3/A64: fix non-ODT setting According to Jens disabling the on-die-termination should set bit 5, not bit 1 in the respective register. Fix this. Reported-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:43 +00:00
/* These are more guessed based on some Allwinner code. */
#define DX_GCR_ODT_DYNAMIC (0x0 << 4)
#define DX_GCR_ODT_ALWAYS_ON (0x1 << 4)
#define DX_GCR_ODT_OFF (0x2 << 4)
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
static int mctl_channel_init(uint16_t socid, struct dram_para *para)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
struct sunxi_mctl_com_reg * const mctl_com =
(struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE;
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
unsigned int i;
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
mctl_set_cr(socid, para);
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
mctl_set_timing_params(socid, para);
mctl_set_master_priority(socid);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* setting VTC, default disable all VT */
clrbits_le32(&mctl_ctl->pgcr[0], (1 << 30) | 0x3f);
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
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);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* increase DFI_PHY_UPD clock */
writel(PROTECT_MAGIC, &mctl_com->protect);
udelay(100);
clrsetbits_le32(&mctl_ctl->upd2, 0xfff << 16, 0x50 << 16);
writel(0x0, &mctl_com->protect);
udelay(100);
/* set dramc odt */
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
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);
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* AC PDR should always ON */
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
clrsetbits_le32(&mctl_ctl->aciocr, socid == SOCID_H5 ? (0x1 << 11) : 0,
0x1 << 1);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* set DQS auto gating PD mode */
setbits_le32(&mctl_ctl->pgcr[2], 0x3 << 6);
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
if (socid == SOCID_H3) {
/* dx ddr_clk & hdr_clk dynamic mode */
clrbits_le32(&mctl_ctl->pgcr[0], (0x3 << 14) | (0x3 << 12));
/* dphy & aphy phase select 270 degree */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8),
(0x1 << 10) | (0x2 << 8));
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
} else if (socid == SOCID_V3S) {
/* dx ddr_clk & hdr_clk dynamic mode */
clrbits_le32(&mctl_ctl->pgcr[0], (0x3 << 14) | (0x3 << 12));
/* dphy & aphy phase select 270 degree */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8),
(0x1 << 10) | (0x1 << 8));
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
} else if (socid == SOCID_A64 || socid == SOCID_H5) {
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
/* dphy & aphy phase select ? */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8),
(0x0 << 10) | (0x3 << 8));
} else if (socid == SOCID_R40) {
/* dx ddr_clk & hdr_clk dynamic mode (tpr13[9] == 0) */
clrbits_le32(&mctl_ctl->pgcr[0], (0x3 << 14) | (0x3 << 12));
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
/* dphy & aphy phase select ? */
clrsetbits_le32(&mctl_ctl->pgcr[2], (0x3 << 10) | (0x3 << 8),
(0x0 << 10) | (0x3 << 8));
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* set half DQ */
sunxi: Rename bus-width related macros in H3 DRAM code The DesignWare DRAM controller used by H3 and newer SoCs use a bit to identify whether the DRAM is half-width. As H3 itself come with 32-bit DRAM, the two modes of the bit used to be named "MCTL_CR_32BIT" and "MCTL_CR_16BIT", but for SoCs with 16-bit DRAM they're really 8-bit and 16-bit. Rename the bit's macro, and also rename the variable name in dram_sun8i_h3.c. This commit do not add 16-bit DRAM controller support, but the support will be introduced in next commit. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:15 +00:00
if (!para->bus_full_width) {
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
#if defined CONFIG_SUNXI_DRAM_DW_32BIT
sunxi: H3: add and rename some DRAM contoller registers The IOCR registers got renamed to BDLR to match the public documentation of similar controllers. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:39 +00:00
writel(0x0, &mctl_ctl->dx[2].gcr);
writel(0x0, &mctl_ctl->dx[3].gcr);
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
#elif defined CONFIG_SUNXI_DRAM_DW_16BIT
writel(0x0, &mctl_ctl->dx[1].gcr);
#else
#error Unsupported DRAM bus width!
#endif
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
/* data training configuration */
clrsetbits_le32(&mctl_ctl->dtcr, 0xf << 24,
(para->dual_rank ? 0x3 : 0x1) << 24);
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
mctl_set_bit_delays(para);
udelay(50);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
if (socid == SOCID_V3S) {
mctl_v3s_zq_calibration_quirk(para);
mctl_phy_init(PIR_PLLINIT | PIR_DCAL | PIR_PHYRST |
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);
} else if (socid == SOCID_H3) {
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
mctl_h3_zq_calibration_quirk(para);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
mctl_phy_init(PIR_PLLINIT | PIR_DCAL | PIR_PHYRST |
PIR_DRAMRST | PIR_DRAMINIT | PIR_QSGATE);
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
} else if (socid == SOCID_A64 || socid == SOCID_H5) {
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
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);
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
/* no PIR_QSGATE for H5 ???? */
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
} else if (socid == SOCID_R40) {
clrsetbits_le32(&mctl_ctl->zqcr, 0xffffff, CONFIG_DRAM_ZQ);
mctl_phy_init(PIR_ZCAL | PIR_PLLINIT | PIR_DCAL | PIR_PHYRST |
PIR_DRAMRST | PIR_DRAMINIT);
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* detect ranks and bus width */
if (readl(&mctl_ctl->pgsr[0]) & (0xfe << 20)) {
/* only one rank */
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
if (((readl(&mctl_ctl->dx[0].gsr[0]) >> 24) & 0x2)
#if defined CONFIG_SUNXI_DRAM_DW_32BIT
|| ((readl(&mctl_ctl->dx[1].gsr[0]) >> 24) & 0x2)
#endif
) {
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
clrsetbits_le32(&mctl_ctl->dtcr, 0xf << 24, 0x1 << 24);
para->dual_rank = 0;
}
/* only half DQ width */
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
#if defined CONFIG_SUNXI_DRAM_DW_32BIT
sunxi: H3: add and rename some DRAM contoller registers The IOCR registers got renamed to BDLR to match the public documentation of similar controllers. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:39 +00:00
if (((readl(&mctl_ctl->dx[2].gsr[0]) >> 24) & 0x1) ||
((readl(&mctl_ctl->dx[3].gsr[0]) >> 24) & 0x1)) {
writel(0x0, &mctl_ctl->dx[2].gcr);
writel(0x0, &mctl_ctl->dx[3].gcr);
sunxi: Rename bus-width related macros in H3 DRAM code The DesignWare DRAM controller used by H3 and newer SoCs use a bit to identify whether the DRAM is half-width. As H3 itself come with 32-bit DRAM, the two modes of the bit used to be named "MCTL_CR_32BIT" and "MCTL_CR_16BIT", but for SoCs with 16-bit DRAM they're really 8-bit and 16-bit. Rename the bit's macro, and also rename the variable name in dram_sun8i_h3.c. This commit do not add 16-bit DRAM controller support, but the support will be introduced in next commit. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:15 +00:00
para->bus_full_width = 0;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
sunxi: add option for 16-bit DW DRAM controller Some Allwinner SoCs features a DesignWare-like controller with only 16 bit bus width. Add support for them. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:16 +00:00
#elif defined CONFIG_SUNXI_DRAM_DW_16BIT
if ((readl(&mctl_ctl->dx[1].gsr[0]) >> 24) & 0x1) {
writel(0x0, &mctl_ctl->dx[1].gcr);
para->bus_full_width = 0;
}
#endif
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
mctl_set_cr(socid, para);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
udelay(20);
/* re-train */
mctl_phy_init(PIR_QSGATE);
if (readl(&mctl_ctl->pgsr[0]) & (0xfe << 20))
return 1;
}
/* check the dramc status */
mctl_await_completion(&mctl_ctl->statr, 0x1, 0x1);
/* liuke added for refresh debug */
setbits_le32(&mctl_ctl->rfshctl0, 0x1 << 31);
udelay(10);
clrbits_le32(&mctl_ctl->rfshctl0, 0x1 << 31);
udelay(10);
/* set PGCR3, CKE polarity */
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
if (socid == SOCID_H3 || socid == SOCID_V3S)
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
writel(0x00aa0060, &mctl_ctl->pgcr[3]);
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
else if (socid == SOCID_A64 || socid == SOCID_H5 || socid == SOCID_R40)
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
writel(0xc0aa0060, &mctl_ctl->pgcr[3]);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* power down zq calibration module for power save */
setbits_le32(&mctl_ctl->zqcr, ZQCR_PWRDOWN);
/* enable master access */
writel(0xffffffff, &mctl_com->maer);
return 0;
}
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
/*
* Test if memory at offset offset matches memory at a certain base
*/
static bool mctl_mem_matches_base(u32 offset, ulong base)
{
/* Try to write different values to RAM at two addresses */
writel(0, base);
writel(0xaa55aa55, base + offset);
dsb();
/* Check if the same value is actually observed when reading back */
return readl(base) ==
readl(base + offset);
}
static void mctl_auto_detect_dram_size_rank(uint16_t socid, struct dram_para *para, ulong base, struct rank_para *rank)
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
{
/* detect row address bits */
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
rank->page_size = 512;
rank->row_bits = 16;
rank->bank_bits = 2;
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
mctl_set_cr(socid, para);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
for (rank->row_bits = 11; rank->row_bits < 16; rank->row_bits++)
if (mctl_mem_matches_base((1 << (rank->row_bits + rank->bank_bits)) * rank->page_size, base))
sunxi: add bank detection code to H3 DRAM initialization code Some DDR2 DRAM have only four banks, not eight. Add code to detect this situation. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:17 +00:00
break;
/* detect bank address bits */
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
rank->bank_bits = 3;
sunxi: add bank detection code to H3 DRAM initialization code Some DDR2 DRAM have only four banks, not eight. Add code to detect this situation. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:17 +00:00
mctl_set_cr(socid, para);
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
for (rank->bank_bits = 2; rank->bank_bits < 3; rank->bank_bits++)
if (mctl_mem_matches_base((1 << rank->bank_bits) * rank->page_size, base))
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
break;
/* detect page size */
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
rank->page_size = 8192;
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
mctl_set_cr(socid, para);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
for (rank->page_size = 512; rank->page_size < 8192; rank->page_size *= 2)
if (mctl_mem_matches_base(rank->page_size, base))
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
break;
}
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
static unsigned long mctl_calc_rank_size(struct rank_para *rank)
{
return (1UL << (rank->row_bits + rank->bank_bits)) * rank->page_size;
}
sunxi: enable dual rank memory on R40 Previously we do not have proper dual rank memory detection on R40 (because we omitted PIR_QSGATE, which does not work on R40 with our configuration), and dual rank memory is just simply disabled as early R40 boards available (Banana Pi M2 Ultra and Berry) have single rank memory. As a board with dual rank memory (Forlinx OKA40i-C) is now known to us, we need to have a way to do memory rank detection to support that board. Add some routine to detect memory rank by trying to access the memory in rank 1 and check for error status of the memory controller, and then enable dual rank memory on R40. Similar routine can be used to detect half DQ width (which is also detected by PIR_QSGATE on other SoCs), but it's left unimplemented because there's no known R40 board with half DQ width now. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> [Andre: Move R40 detect code call into sunxi_dram_init()] Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:25 +00:00
/*
* Because we cannot do mctl_phy_init(PIR_QSGATE) on R40 now (which leads
* to failure), it's needed to detect the rank count of R40 in another way.
*
* The code here is modelled after time_out_detect() in BSP, which tries to
* access the memory and check for error code.
*
* TODO: auto detect half DQ width here
*/
static void mctl_r40_detect_rank_count(struct dram_para *para)
{
global: Move remaining CONFIG_SYS_SDRAM_* to CFG_SYS_SDRAM_* The rest of the unmigrated CONFIG symbols in the CONFIG_SYS_SDRAM namespace do not easily transition to Kconfig. In many cases they likely should come from the device tree instead. Move these out of CONFIG namespace and in to CFG namespace. Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
2022-11-16 18:10:37 +00:00
ulong rank1_base = (ulong) CFG_SYS_SDRAM_BASE +
sunxi: enable dual rank memory on R40 Previously we do not have proper dual rank memory detection on R40 (because we omitted PIR_QSGATE, which does not work on R40 with our configuration), and dual rank memory is just simply disabled as early R40 boards available (Banana Pi M2 Ultra and Berry) have single rank memory. As a board with dual rank memory (Forlinx OKA40i-C) is now known to us, we need to have a way to do memory rank detection to support that board. Add some routine to detect memory rank by trying to access the memory in rank 1 and check for error status of the memory controller, and then enable dual rank memory on R40. Similar routine can be used to detect half DQ width (which is also detected by PIR_QSGATE on other SoCs), but it's left unimplemented because there's no known R40 board with half DQ width now. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> [Andre: Move R40 detect code call into sunxi_dram_init()] Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:25 +00:00
mctl_calc_rank_size(&para->ranks[0]);
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
/* Enable read time out */
setbits_le32(&mctl_ctl->pgcr[0], 0x1 << 25);
(void) readl((void *) rank1_base);
udelay(10);
if (readl(&mctl_ctl->pgsr[0]) & (0x1 << 13)) {
clrsetbits_le32(&mctl_ctl->dtcr, 0xf << 24, 0x1 << 24);
para->dual_rank = 0;
}
/* Reset PHY FIFO to clear it */
clrbits_le32(&mctl_ctl->pgcr[0], 0x1 << 26);
udelay(100);
setbits_le32(&mctl_ctl->pgcr[0], 0x1 << 26);
/* Clear error status */
setbits_le32(&mctl_ctl->pgcr[0], 0x1 << 24);
/* Clear time out flag */
clrbits_le32(&mctl_ctl->pgsr[0], 0x1 << 13);
/* Disable read time out */
clrbits_le32(&mctl_ctl->pgcr[0], 0x1 << 25);
}
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
static void mctl_auto_detect_dram_size(uint16_t socid, struct dram_para *para)
{
global: Move remaining CONFIG_SYS_SDRAM_* to CFG_SYS_SDRAM_* The rest of the unmigrated CONFIG symbols in the CONFIG_SYS_SDRAM namespace do not easily transition to Kconfig. In many cases they likely should come from the device tree instead. Move these out of CONFIG namespace and in to CFG namespace. Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
2022-11-16 18:10:37 +00:00
mctl_auto_detect_dram_size_rank(socid, para, (ulong)CFG_SYS_SDRAM_BASE, &para->ranks[0]);
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
if ((socid == SOCID_A64 || socid == SOCID_R40) && para->dual_rank) {
global: Move remaining CONFIG_SYS_SDRAM_* to CFG_SYS_SDRAM_* The rest of the unmigrated CONFIG symbols in the CONFIG_SYS_SDRAM namespace do not easily transition to Kconfig. In many cases they likely should come from the device tree instead. Move these out of CONFIG namespace and in to CFG namespace. Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
2022-11-16 18:10:37 +00:00
mctl_auto_detect_dram_size_rank(socid, para, (ulong)CFG_SYS_SDRAM_BASE + mctl_calc_rank_size(&para->ranks[0]), &para->ranks[1]);
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
}
}
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
/*
* The actual values used here are taken from Allwinner provided boot0
* binaries, though they are probably board specific, so would likely benefit
* from invidual tuning for each board. Apparently a lot of boards copy from
* some Allwinner reference design, so we go with those generic values for now
* in the hope that they are reasonable for most (all?) boards.
*/
#define SUN8I_H3_DX_READ_DELAYS \
{{ 18, 18, 18, 18, 18, 18, 18, 18, 18, 0, 0 }, \
{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 }, \
{ 18, 18, 18, 18, 18, 18, 18, 18, 18, 0, 0 }, \
{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 }}
#define SUN8I_H3_DX_WRITE_DELAYS \
{{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 10, 10 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 10, 10 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 10, 10 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 6 }}
#define SUN8I_H3_AC_DELAYS \
{ 0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0 }
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
#define SUN8I_V3S_DX_READ_DELAYS \
{{ 8, 8, 8, 8, 8, 8, 8, 8, 8, 0, 0 }, \
{ 7, 7, 7, 7, 7, 7, 7, 7, 7, 0, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }}
#define SUN8I_V3S_DX_WRITE_DELAYS \
{{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 4 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }}
#define SUN8I_V3S_AC_DELAYS \
{ 0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0 }
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
#define SUN8I_R40_DX_READ_DELAYS \
{{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 }, \
{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 }, \
{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 }, \
{ 14, 14, 14, 14, 14, 14, 14, 14, 14, 0, 0 } }
#define SUN8I_R40_DX_WRITE_DELAYS \
{{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0 }, \
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0 } }
#define SUN8I_R40_AC_DELAYS \
{ 0, 0, 3, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0, 0, \
0, 0, 0, 0, 0, 0, 0 }
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#define SUN50I_A64_DX_READ_DELAYS \
{{ 16, 16, 16, 16, 17, 16, 16, 17, 16, 1, 0 }, \
{ 17, 17, 17, 17, 17, 17, 17, 17, 17, 1, 0 }, \
{ 16, 17, 17, 16, 16, 16, 16, 16, 16, 0, 0 }, \
{ 17, 17, 17, 17, 17, 17, 17, 17, 17, 1, 0 }}
#define SUN50I_A64_DX_WRITE_DELAYS \
{{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 15, 15 }, \
{ 0, 0, 0, 0, 1, 1, 1, 1, 0, 10, 10 }, \
{ 1, 0, 1, 1, 1, 1, 1, 1, 0, 11, 11 }, \
{ 1, 0, 0, 1, 1, 1, 1, 1, 0, 12, 12 }}
#define SUN50I_A64_AC_DELAYS \
{ 5, 5, 13, 10, 2, 5, 3, 3, \
0, 3, 3, 3, 1, 0, 0, 0, \
3, 4, 0, 3, 4, 1, 4, 0, \
1, 1, 0, 1, 13, 5, 4 }
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
#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 }
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
unsigned long sunxi_dram_init(void)
{
struct sunxi_mctl_com_reg * const mctl_com =
(struct sunxi_mctl_com_reg *)SUNXI_DRAM_COM_BASE;
struct sunxi_mctl_ctl_reg * const mctl_ctl =
(struct sunxi_mctl_ctl_reg *)SUNXI_DRAM_CTL0_BASE;
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
unsigned long size;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
struct dram_para para = {
sunxi: enable dual rank detection in DesignWare-like DRAM code The DesignWare-like DRAM code used to set the controller defaultly to single rank mode, which makes it not able to detect the second rank. Set the default value to dual rank, thus the rank detection code can work and finally the rank setting will be the correct value. Currently we know little about the dual-rank on R40, and the usage of A15 address line seems to be breaking dual-rank support. The only R40 board currently available (Sinovoip Banana Pi M2 Ultra) uses A15 rather than dual-rank, thus we cannot do research for it. So dual rank detection is temporarily disabled on R40. This change is tested on a Orange Pi One (H3, single rank), a Pine64+ 2GiB version (A64, single rank) , a Pinebook early prototype with DDR3 (A64, dual rank) and a SoPine with some LPDDR3 patch (A64, dual CS pins on one chip). Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:19 +00:00
.dual_rank = 1,
sunxi: Rename bus-width related macros in H3 DRAM code The DesignWare DRAM controller used by H3 and newer SoCs use a bit to identify whether the DRAM is half-width. As H3 itself come with 32-bit DRAM, the two modes of the bit used to be named "MCTL_CR_32BIT" and "MCTL_CR_16BIT", but for SoCs with 16-bit DRAM they're really 8-bit and 16-bit. Rename the bit's macro, and also rename the variable name in dram_sun8i_h3.c. This commit do not add 16-bit DRAM controller support, but the support will be introduced in next commit. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:15 +00:00
.bus_full_width = 1,
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
.ranks = {
{
.row_bits = 15,
.bank_bits = 3,
.page_size = 4096,
},
{
.row_bits = 15,
.bank_bits = 3,
.page_size = 4096,
}
},
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#if defined(CONFIG_MACH_SUN8I_H3)
sunxi: H3: add DRAM controller single bit delay support So far the DRAM driver for the H3 SoC (and apparently boot0/libdram as well) only applied coarse delay line settings, with one delay value for all the data lines in each byte lane and one value for the control lines. Instead of setting the delays for whole bytes only allow setting it for each individual bit. Also add support for address/command lane delays. For the purpose of this patch the rules for the existing coarse settings were just applied to the new scheme, so the actual register writes don't change for the H3. Other SoCs will utilize this feature later properly. With a stock GCC 5.3.0 this increases the dram_sun8i_h3.o code size from 2296 to 2344 Bytes. [Andre: move delay parameters into macros to ease later sharing, use defines for numbers of delay registers, extend commit message] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:40 +00:00
.dx_read_delays = SUN8I_H3_DX_READ_DELAYS,
.dx_write_delays = SUN8I_H3_DX_WRITE_DELAYS,
.ac_delays = SUN8I_H3_AC_DELAYS,
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
#elif defined(CONFIG_MACH_SUN8I_V3S)
.dx_read_delays = SUN8I_V3S_DX_READ_DELAYS,
.dx_write_delays = SUN8I_V3S_DX_WRITE_DELAYS,
.ac_delays = SUN8I_V3S_AC_DELAYS,
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
#elif defined(CONFIG_MACH_SUN8I_R40)
.dx_read_delays = SUN8I_R40_DX_READ_DELAYS,
.dx_write_delays = SUN8I_R40_DX_WRITE_DELAYS,
.ac_delays = SUN8I_R40_AC_DELAYS,
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#elif defined(CONFIG_MACH_SUN50I)
.dx_read_delays = SUN50I_A64_DX_READ_DELAYS,
.dx_write_delays = SUN50I_A64_DX_WRITE_DELAYS,
.ac_delays = SUN50I_A64_AC_DELAYS,
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
#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,
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#endif
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
};
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
/*
* Let the compiler optimize alternatives away by passing this value into
* the static functions. This saves us #ifdefs, but still keeps the binary
* small.
*/
#if defined(CONFIG_MACH_SUN8I_H3)
uint16_t socid = SOCID_H3;
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
#elif defined(CONFIG_MACH_SUN8I_R40)
uint16_t socid = SOCID_R40;
sunxi: add support for V3s DRAM controller Allwinner V3s features a DRAM controller like the on in H3, but with a DDR2 DRAM. Add support for it. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Tested-by: Jagan Teki <jagan@amarulasolutions.com>
2017-06-03 09:10:21 +00:00
#elif defined(CONFIG_MACH_SUN8I_V3S)
sunxi: make V3s DRAM initialization more proper Previously, because we have no source code about the DRAM initialization of V3s and missing some configurations (delays and MBUS QoS info), our V3s DRAM initialization sequence is hacked from the H3 one. As the SDK shipped with PineCube contains source code for V3s libdram, we can retrieve these information from it and tweak some other magic bits. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
2020-10-16 09:33:08 +00:00
uint16_t socid = SOCID_V3S;
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#elif defined(CONFIG_MACH_SUN50I)
uint16_t socid = SOCID_A64;
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
#elif defined(CONFIG_MACH_SUN50I_H5)
uint16_t socid = SOCID_H5;
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
#endif
mctl_sys_init(socid, &para);
if (mctl_channel_init(socid, &para))
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
return 0;
if (para.dual_rank)
writel(0x00000303, &mctl_ctl->odtmap);
else
writel(0x00000201, &mctl_ctl->odtmap);
udelay(1);
/* odt delay */
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
if (socid == SOCID_H3)
writel(0x0c000400, &mctl_ctl->odtcfg);
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
if (socid == SOCID_A64 || socid == SOCID_H5 || socid == SOCID_R40) {
/* VTF enable (tpr13[8] == 1) */
sunxi: DRAM: add Allwinner H5 support 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. Register setup differences between H5 and H3 are courtesy of Jens Kuske. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2017-02-16 01:20:26 +00:00
setbits_le32(&mctl_ctl->vtfcr,
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
(socid != SOCID_A64 ? 3 : 2) << 8);
/* DQ hold disable (tpr13[26] == 1) */
sunxi: A64: use H3 DRAM initialization code for A64 as well The A64 DRAM controller is very similar to the H3 one, so the code can be reused with some small changes. This refactoring does not change the code size for the existing H3 part. [Andre: rework from #ifdefs to using socid parameters in static functions, minor fixes, merging in fixes from Jens] Signed-off-by: Jens Kuske <jenskuske@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com> Reviewed-by: Jagan Teki <jagan@openedev.com>
2017-01-02 11:48:42 +00:00
clrbits_le32(&mctl_ctl->pgcr[2], (1 << 13));
}
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
/* clear credit value */
setbits_le32(&mctl_com->cccr, 1 << 31);
udelay(10);
sunxi: enable dual rank memory on R40 Previously we do not have proper dual rank memory detection on R40 (because we omitted PIR_QSGATE, which does not work on R40 with our configuration), and dual rank memory is just simply disabled as early R40 boards available (Banana Pi M2 Ultra and Berry) have single rank memory. As a board with dual rank memory (Forlinx OKA40i-C) is now known to us, we need to have a way to do memory rank detection to support that board. Add some routine to detect memory rank by trying to access the memory in rank 1 and check for error status of the memory controller, and then enable dual rank memory on R40. Similar routine can be used to detect half DQ width (which is also detected by PIR_QSGATE on other SoCs), but it's left unimplemented because there's no known R40 board with half DQ width now. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> [Andre: Move R40 detect code call into sunxi_dram_init()] Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:25 +00:00
if (socid == SOCID_R40) {
mctl_r40_detect_rank_count(&para);
mctl_set_cr(SOCID_R40, &para);
}
sunxi: Use H3/A64 DRAM initialization code for R40 The R40 seems to have a variant of the memory controller found in the H3 and A64 SoCs. Adapt the code for use on the R40. The changes are based on released DRAM code and comparing register dumps from boot0. Signed-off-by: Chen-Yu Tsai <wens@csie.org> Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
2016-12-01 11:09:57 +00:00
mctl_auto_detect_dram_size(socid, &para);
mctl_set_cr(socid, &para);
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
sunxi: support asymmetric dual rank DRAM on A64/R40 Previously we have known that R40 has a configuration register for its rank 1, which allows different configuration than rank 0. Reverse engineering of newest libdram of A64 from Allwinner shows that A64 has this register too. It's bit 0 (which enables dual rank in rank 0 configuration register) means a dedicated rank size setup is used for rank 1. Now, Pine64 scheduled to use a 3GiB LPDDR3 DRAM chip (which has 2GiB rank 0 and 1GiB rank 1) on PinePhone, that makes asymmetric dual rank DRAM support necessary. Add this support. The code could support both A64 and R40, but because dual rank detection is broken on R40 now, we cannot really use it on R40 currently. Signed-off-by: Icenowy Zheng <icenowy@aosc.io> Reviewed-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Peter Robinson <pbrobinson@gmail.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
2021-02-25 16:13:24 +00:00
size = mctl_calc_rank_size(&para.ranks[0]);
if (socid == SOCID_A64 || socid == SOCID_R40) {
if (para.dual_rank)
size += mctl_calc_rank_size(&para.ranks[1]);
} else if (para.dual_rank) {
size *= 2;
}
return size;
sunxi: Add H3 DRAM initialization support Based on existing A23/A33 code and the original H3 boot0. Signed-off-by: Jens Kuske <jenskuske@gmail.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
2015-11-17 14:12:59 +00:00
}
Reference in a new issue Copy permalink