[PATCH 3/3] riscv: dts: spacemit: k3-deepcomputing-fml13v05: Enable SD card support
Sandie Cao
sandie.cao at deepcomputing.io
Wed Sep 23 02:51:09 PDT 2026
Hi, Troy:
> From: "Troy Mitchell"<troy.mitchell at linux.spacemit.com>
> Date: Wed, Sep 23, 2026, 17:14
> On Wed Sep 23, 2026 at 4:01 PM +08, Sandie Cao wrote:
> > Hi, Troy:
> >
> >> From: "Troy Mitchell"<troy.mitchell at linux.spacemit.com>
> >> Date: Wed, Sep 23, 2026, 15:02
> >> On Wed, Sep 23, 2026 at 02:35:20PM +0800, Sandie Cao wrote:
> >> > FML13V05 SD card slot is inside the board, it doesn't need cd-gpio,
> >> > FML13V05 Keeps GPIO88 to low, Then SD card is always on for SD Card
> >> > driver.
> >>
> >> An internal slot does not necessarily mean the card is non-removable.
> >> I don't think its location alone justifies keeping card detect asserted:
> >> GPIO88 would report a card present even when the slot is empty.
> >
> > Yes. The FML13V05 hardware is designed for this. We can't change it.
>
> I may not have been clear: an internal card slot does not necessarily
> mean the card is non-removable. If GPIO88 is connected to the
> card-detect switch, it still needs a pull-up. Is GPIO88 tied low on
> the board, or does it reflect the state of that switch?
The FML13V05 SD card slot doesn't have cd-gpio. But the K3 SD card driver need it, so the GPIO88 is tied low on board.
>
> >
> >>
> >> > We check the vmmc-supply is p3v3. We have the same SDK DTS config with
> >> > k3_com260[1] , so, this patch follows the SD card patch of k3_com260
> >> > dts[2].
> >> >
> >> > If this is not correct, we will change it.
> >>
> >> Could you confirm from the schematics that the SD power wiring matches
> >> CoM260 and that the card's VDD is actually connected to ALDO4? If so,
> >> this supply reference is fine. The same SDK configuration alone does
> >> not establish the wiring.
> >>
> >> If the card instead uses a fixed, always-on supply with no software power
> >> control, vmmc-supply is optional for the MMC core; it should not point to
> >> an unrelated ALDO4 rail.
> >
> > We check the schematics both of COM260 and FML13V05, both use p3v3 (a fixed, always-on supply with no software power control), not ALDO4. ALD04 is only for USB.
> > We will correct it in next patch.
> I'll send a separate patch to fix the CoM260 vmmc-supply reference.
> You can focus on FML13V05 in this series.
Got it. Thanks
Best regards
Sandie
More information about the linux-riscv
mailing list