Re: [PATCH 3/9] hw/arm/aspeed: qcom-dc-scm-v1: add block backed FRU device

[Joel Stanley <joel@jms.id.au>]

On Wed, 22 Jun 2022 at 17:29, Jae Hyun Yoo <quic_jaehyoo@quicinc.com> wrote:
>
> From: Graeme Gregory <quic_ggregory@quicinc.com>
>
> The FRU device uses the index 0 device on bus IF_NONE.
>
> -drive file=$FRU,format=raw,if=none
>
> file must match FRU size of 128k
>
> Signed-off-by: Graeme Gregory <quic_ggregory@quicinc.com>
> ---
>  hw/arm/aspeed.c | 22 +++++++++++++++++-----
>  1 file changed, 17 insertions(+), 5 deletions(-)
>
> diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c
> index 785cc543d046..36d6b2c33e48 100644
> --- a/hw/arm/aspeed.c
> +++ b/hw/arm/aspeed.c
> @@ -992,17 +992,29 @@ static void fby35_i2c_init(AspeedMachineState *bmc)
>       */
>  }
>
> +static void qcom_dc_scm_fru_init(I2CBus *bus, uint8_t addr, uint32_t rsize)
> +{
> +    I2CSlave *i2c_dev = i2c_slave_new("at24c-eeprom", addr);
> +    DeviceState *dev = DEVICE(i2c_dev);
> +    /* Use First Index for DC-SCM FRU */
> +    DriveInfo *dinfo = drive_get(IF_NONE, 0, 0);
> +
> +    qdev_prop_set_uint32(dev, "rom-size", rsize);
> +
> +    if (dinfo) {
> +        qdev_prop_set_drive(dev, "drive", blk_by_legacy_dinfo(dinfo));
> +    }
> +
> +    i2c_slave_realize_and_unref(i2c_dev, bus, &error_abort);
> +}
> +
>  static void qcom_dc_scm_bmc_i2c_init(AspeedMachineState *bmc)
>  {
>      AspeedSoCState *soc = &bmc->soc;
>
>      i2c_slave_create_simple(aspeed_i2c_get_bus(&soc->i2c, 15), "tmp105", 0x4d);
>
> -    uint8_t *eeprom_buf = g_malloc0(128 * 1024);
> -    if (eeprom_buf) {
> -        smbus_eeprom_init_one(aspeed_i2c_get_bus(&soc->i2c, 15), 0x53,
> -                              eeprom_buf);
> -    }

Again, it's strange to see code that was just added being removed. If
you want the FRU to be in its own patch then remove the eeprom from
the previous patch.

> +    qcom_dc_scm_fru_init(aspeed_i2c_get_bus(&soc->i2c, 15), 0x53, 128 * 1024);
>  }
>
>  static bool aspeed_get_mmio_exec(Object *obj, Error **errp)
> --
> 2.25.1
>