[U-Boot] [PATCH] rockchip: rk3288: correct sdram setting

[Chris Zhong <zyw@rock-chips.com>]

Hi Simon

On 03/01/2016 10:04 AM, Simon Glass wrote:
> Hi Chris,
>
> On 29 February 2016 at 05:16, Chris Zhong <zyw@rock-chips.com> wrote:
>> The DMC driver in v3.14 kernel[0] get the ddr setting from PMU_SYS_REG2,
>> and it expects uboot to store the value using a same protocol. But now
>> the ddr setting value is different with DMC, so if you enable the DMC,
>> system would crash in kernel. Correct the sdram setting here, according
>> to the requirements of kernel.
>>
>> [0]
>> https://chromium.googlesource.com/chromiumos/third_party/kernel/+/
>> chromeos-3.14/drivers/clk/rockchip/clk-rk3288-dmc.c
>>
>> Signed-off-by: Chris Zhong <zyw@rock-chips.com>
>> ---
>>
>>   arch/arm/include/asm/arch-rockchip/ddr_rk3288.h | 42 ++++++++---------
>>   arch/arm/mach-rockchip/rk3288/sdram_rk3288.c    | 61 +++++++++++--------------
>>   2 files changed, 48 insertions(+), 55 deletions(-)
>>
>> diff --git a/arch/arm/include/asm/arch-rockchip/ddr_rk3288.h b/arch/arm/include/asm/arch-rockchip/ddr_rk3288.h
>> index fccabcd..f2e3130 100644
>> --- a/arch/arm/include/asm/arch-rockchip/ddr_rk3288.h
>> +++ b/arch/arm/include/asm/arch-rockchip/ddr_rk3288.h
>> @@ -459,26 +459,26 @@ enum {
>>    * [3:2] bw_ch0
>>    * [1:0] dbw_ch0
>>   */
>> -#define SYS_REG_DDRTYPE_SHIFT          13
>> -#define SYS_REG_DDRTYPE_MASK           7
>> -#define SYS_REG_NUM_CH_SHIFT           12
>> -#define SYS_REG_NUM_CH_MASK            1
>> -#define SYS_REG_ROW_3_4_SHIFT(ch)      (30 + (ch))
>> -#define SYS_REG_ROW_3_4_MASK           1
>> -#define SYS_REG_CHINFO_SHIFT(ch)       (28 + (ch))
>> -#define SYS_REG_RANK_SHIFT(ch)         (11 + (ch) * 16)
>> -#define SYS_REG_RANK_MASK              1
>> -#define SYS_REG_COL_SHIFT(ch)          (9 + (ch) * 16)
>> -#define SYS_REG_COL_MASK               3
>> -#define SYS_REG_BK_SHIFT(ch)           (8 + (ch) * 16)
>> -#define SYS_REG_BK_MASK                        1
>> -#define SYS_REG_CS0_ROW_SHIFT(ch)      (6 + (ch) * 16)
>> -#define SYS_REG_CS0_ROW_MASK           3
>> -#define SYS_REG_CS1_ROW_SHIFT(ch)      (4 + (ch) * 16)
>> -#define SYS_REG_CS1_ROW_MASK           3
>> -#define SYS_REG_BW_SHIFT(ch)           (2 + (ch) * 16)
>> -#define SYS_REG_BW_MASK                        3
>> -#define SYS_REG_DBW_SHIFT(ch)          ((ch) * 16)
>> -#define SYS_REG_DBW_MASK               3
>> +#define SYS_REG_ENC_ROW_3_4(n, ch)     ((n) << (30 + (ch)))
>> +#define SYS_REG_DEC_ROW_3_4(n, ch)     ((n >> (30 + ch)) & 0x1)
>> +#define SYS_REG_ENC_CHINFO(ch)         (1 << (28 + (ch)))
>> +#define SYS_REG_ENC_DDRTYPE(n)         ((n) << 13)
>> +#define SYS_REG_ENC_NUM_CH(n)          (((n) - 1) << 12)
>> +#define SYS_REG_DEC_NUM_CH(n)          (1 + ((n >> 12) & 0x1))
>> +#define SYS_REG_ENC_RANK(n, ch)                (((n) - 1) << (11 + ((ch) * 16)))
>> +#define SYS_REG_DEC_RANK(n, ch)                (1 + ((n >> (11 + 16 * ch)) & 0x1))
>> +#define SYS_REG_ENC_COL(n, ch)         (((n) - 9) << (9 + ((ch) * 16)))
>> +#define SYS_REG_DEC_COL(n, ch)         (9 + ((n >> (9 + 16 * ch)) & 0x3))
>> +#define SYS_REG_ENC_BK(n, ch)          (((n) == 3 ? 0 : 1) \
>> +                                       << (8 + ((ch) * 16)))
>> +#define SYS_REG_DEC_BK(n, ch)          (3 - ((n >> (8 + 16 * ch)) & 0x1))
>> +#define SYS_REG_ENC_CS0_ROW(n, ch)     (((n) - 13) << (6 + ((ch) * 16)))
>> +#define SYS_REG_DEC_CS0_ROW(n, ch)     (13 + ((n >> (6 + 16 * ch)) & 0x3))
>> +#define SYS_REG_ENC_CS1_ROW(n, ch)     (((n) - 13) << (4 + ((ch) * 16)))
>> +#define SYS_REG_DEC_CS1_ROW(n, ch)     (13 + ((n >> (4 + 16 * ch)) & 0x3))
>> +#define SYS_REG_ENC_BW(n, ch)          ((2 >> (n)) << (2 + ((ch) * 16)))
>> +#define SYS_REG_DEC_BW(n, ch)          (2 >> ((n >> (2 + 16 * ch)) & 0x3))
>> +#define SYS_REG_ENC_DBW(n, ch)         ((2 >> (n)) << (0 + ((ch) * 16)))
>> +#define SYS_REG_DEC_DBW(n, ch)         (2 >> ((n >> (0 + 16 * ch)) & 0x3))
> Are the above shift/masks actually wrong? I'm not keen on this style
> of packing and unpacking registers since it is really hard to read and
> it's not obvious that pack and unpack work the same way.
Actually, I copy these code from coreboot[0], can we just keep the code 
consistent with it?

[0]https://chromium.googlesource.com/chromiumos/third_party/coreboot/+/
firmware-veyron-6588.B/src/soc/rockchip/rk3288/sdram.c

>
>>   #endif
>> diff --git a/arch/arm/mach-rockchip/rk3288/sdram_rk3288.c b/arch/arm/mach-rockchip/rk3288/sdram_rk3288.c
>> index e9e2211..4db39ec 100644
>> --- a/arch/arm/mach-rockchip/rk3288/sdram_rk3288.c
>> +++ b/arch/arm/mach-rockchip/rk3288/sdram_rk3288.c
>> @@ -551,27 +551,27 @@ static void dram_cfg_rbc(const struct chan_info *chan, u32 chnum,
>>   static void dram_all_config(const struct dram_info *dram,
>>                              const struct rk3288_sdram_params *sdram_params)
>>   {
>> -       unsigned int chan;
>> +       unsigned int channel;
>>          u32 sys_reg = 0;
>>
>> -       sys_reg |= sdram_params->base.dramtype << SYS_REG_DDRTYPE_SHIFT;
>> -       sys_reg |= (sdram_params->num_channels - 1) << SYS_REG_NUM_CH_SHIFT;
>> -       for (chan = 0; chan < sdram_params->num_channels; chan++) {
>> +       sys_reg |= SYS_REG_ENC_DDRTYPE(sdram_params->base.dramtype);
>> +       sys_reg |= SYS_REG_ENC_NUM_CH(sdram_params->num_channels);
>> +       for (channel = 0; channel < sdram_params->num_channels; channel++) {
>>                  const struct rk3288_sdram_channel *info =
>> -                       &sdram_params->ch[chan];
>> -
>> -               sys_reg |= info->row_3_4 << SYS_REG_ROW_3_4_SHIFT(chan);
>> -               sys_reg |= chan << SYS_REG_CHINFO_SHIFT(chan);
>> -               sys_reg |= (info->rank - 1) << SYS_REG_RANK_SHIFT(chan);
>> -               sys_reg |= (info->col - 9) << SYS_REG_COL_SHIFT(chan);
>> -               sys_reg |= info->bk == 3 ? 1 << SYS_REG_BK_SHIFT(chan) : 0;
>> -               sys_reg |= (info->cs0_row - 13) << SYS_REG_CS0_ROW_SHIFT(chan);
>> -               sys_reg |= (info->cs1_row - 13) << SYS_REG_CS1_ROW_SHIFT(chan);
>> -               sys_reg |= info->bw << SYS_REG_BW_SHIFT(chan);
>> -               sys_reg |= info->dbw << SYS_REG_DBW_SHIFT(chan);
>> -
>> -               dram_cfg_rbc(&dram->chan[chan], chan, sdram_params);
>> +                       &(sdram_params->ch[channel]);
>> +               sys_reg |= SYS_REG_ENC_ROW_3_4(info->row_3_4, channel);
>> +               sys_reg |= SYS_REG_ENC_CHINFO(channel);
>> +               sys_reg |= SYS_REG_ENC_RANK(info->rank, channel);
>> +               sys_reg |= SYS_REG_ENC_COL(info->col, channel);
>> +               sys_reg |= SYS_REG_ENC_BK(info->bk, channel);
>> +               sys_reg |= SYS_REG_ENC_CS0_ROW(info->cs0_row, channel);
>> +               sys_reg |= SYS_REG_ENC_CS1_ROW(info->cs1_row, channel);
>> +               sys_reg |= SYS_REG_ENC_BW(info->bw, channel);
>> +               sys_reg |= SYS_REG_ENC_DBW(info->dbw, channel);
> Can you fix this bug while still using the original #defines?
>
>> +
>> +               dram_cfg_rbc(&dram->chan[channel], channel, sdram_params);
>>          }
>> +
>>          writel(sys_reg, &dram->pmu->sys_reg[2]);
>>          rk_clrsetreg(&dram->sgrf->soc_con2, 0x1f, sdram_params->base.stride);
>>   }
>> @@ -712,23 +712,16 @@ size_t sdram_size_mb(struct rk3288_pmu *pmu)
>>          size_t size_mb = 0;
>>          u32 ch;
>>          u32 sys_reg = readl(&pmu->sys_reg[2]);
>> -       u32 chans;
>> -
>> -       chans = 1 + ((sys_reg >> SYS_REG_NUM_CH_SHIFT) & SYS_REG_NUM_CH_MASK);
>> -
>> -       for (ch = 0; ch < chans; ch++) {
>> -               rank = 1 + (sys_reg >> SYS_REG_RANK_SHIFT(ch) &
>> -                       SYS_REG_RANK_MASK);
>> -               col = 9 + (sys_reg >> SYS_REG_COL_SHIFT(ch) & SYS_REG_COL_MASK);
>> -               bk = sys_reg & (1 << SYS_REG_BK_SHIFT(ch)) ? 3 : 0;
>> -               cs0_row = 13 + (sys_reg >> SYS_REG_CS0_ROW_SHIFT(ch) &
>> -                               SYS_REG_CS0_ROW_MASK);
>> -               cs1_row = 13 + (sys_reg >> SYS_REG_CS1_ROW_SHIFT(ch) &
>> -                               SYS_REG_CS1_ROW_MASK);
>> -               bw = (sys_reg >> SYS_REG_BW_SHIFT(ch)) &
>> -                       SYS_REG_BW_MASK;
>> -               row_3_4 = sys_reg >> SYS_REG_ROW_3_4_SHIFT(ch) &
>> -                       SYS_REG_ROW_3_4_MASK;
>> +       u32 ch_num = SYS_REG_DEC_NUM_CH(sys_reg);
>> +
>> +       for (ch = 0; ch < ch_num; ch++) {
>> +               rank = SYS_REG_DEC_RANK(sys_reg, ch);
>> +               col = SYS_REG_DEC_COL(sys_reg, ch);
>> +               bk = SYS_REG_DEC_BK(sys_reg, ch);
>> +               cs0_row = SYS_REG_DEC_CS0_ROW(sys_reg, ch);
>> +               cs1_row = SYS_REG_DEC_CS1_ROW(sys_reg, ch);
>> +               bw = SYS_REG_DEC_BW(sys_reg, ch);
>> +               row_3_4 = SYS_REG_DEC_ROW_3_4(sys_reg, ch);
>>
>>                  chipsize_mb = (1 << (cs0_row + col + bk + bw - 20));
>>
>> --
>> 2.6.3
>>
> Regards,
> Simon
>
>
>