Re: [PATCH 1/2] hw/sd/pxa2xx_mmci: Do not create SD card within the SDHCI controller

[Peter Maydell]

On Sun, 5 Jul 2020 at 22:33, Philippe Mathieu-Daudé <f4bug@amsat.org> wrote:
>
> SDHCI controllers provide a SD Bus to plug SD cards, but don't
> come with SD card plugged in :) Let the machine/board object
> create and plug the SD cards when required.

This too is not an SDHCI controller.

> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> ---
>  include/hw/arm/pxa.h |  3 +--
>  hw/arm/pxa2xx.c      | 39 +++++++++++++++++++++++++++++----------
>  hw/sd/pxa2xx_mmci.c  | 11 ++---------
>  3 files changed, 32 insertions(+), 21 deletions(-)
> @@ -2135,15 +2136,24 @@ PXA2xxState *pxa270_init(MemoryRegion *address_space,
>
>      s->gpio = pxa2xx_gpio_init(0x40e00000, s->cpu, s->pic, 121);
>
> -    dinfo = drive_get(IF_SD, 0, 0);
> -    if (!dinfo && !qtest_enabled()) {
> -        warn_report("missing SecureDigital device");
> -    }
>      s->mmc = pxa2xx_mmci_init(address_space, 0x41100000,
> -                    dinfo ? blk_by_legacy_dinfo(dinfo) : NULL,
>                      qdev_get_gpio_in(s->pic, PXA2XX_PIC_MMC),
>                      qdev_get_gpio_in(s->dma, PXA2XX_RX_RQ_MMCI),
>                      qdev_get_gpio_in(s->dma, PXA2XX_TX_RQ_MMCI));
> +    dinfo = drive_get(IF_SD, 0, 0);
> +    if (dinfo) {
> +        DeviceState *carddev;
> +
> +        /* Create and plug in the sd card */
> +        carddev = qdev_new(TYPE_SD_CARD);
> +        qdev_prop_set_drive_err(carddev, "drive",
> +                                blk_by_legacy_dinfo(dinfo), &error_fatal);
> +        qdev_realize_and_unref(carddev, qdev_get_child_bus(DEVICE(s->mmc),
> +                                                           "sd-bus"),
> +                               &error_fatal);
> +    } else if (!qtest_enabled()) {
> +        warn_report("missing SecureDigital device");
> +    }

When we finally get round to qdevifying the PXA2xx this will
have to move up another layer, since pxa270_init() is
effectively the SoC, not the board model. But for the moment
this is as far as we can sensibly lift the code.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM