Hi,
Having messed with this for the past couple hours, I'm not convinced
this solution is any better than the previous patch. What is basically
happening is that dw_mmc is utilizing mmc_add_host, which pulls it
into the initrd, but it has "dependencies" on other drivers
registering dw_mc_probe routines before it will register a mmc host.
The platform driver registers a dw_mmc_probe routine, but not before
the k3 driver registers via pltfm_register. Since the dw_mmc_k3 driver
get triggered via its OF/DT device description, it pulls in the rest
of the stack in order to resolve all the missing symbols. So, maybe
the dw_mmc driver needs to be more monolithic, but good luck with
that...
So, in the end I think the previous version might be a more generic solution..
On Wed, Apr 26, 2017 at 9:59 PM, Jeremy Linton <lintonrjeremy@xxxxxxxxx> wrote:
> It is common for arm64 devices to boot from mmc & sd devices
> we need to assure that those drivers make it to the initrd.
>
> Signed-off-by: Jeremy Linton <lintonrjeremy@xxxxxxxxx>
> ---
> modules.d/90kernel-modules/module-setup.sh | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/modules.d/90kernel-modules/module-setup.sh b/modules.d/90kernel-modules/module-setup.sh
> index b1bced69..863ce932 100755
> --- a/modules.d/90kernel-modules/module-setup.sh
> +++ b/modules.d/90kernel-modules/module-setup.sh
> @@ -25,10 +25,12 @@ installkernel() {
>
> if [[ "$(uname -m)" == arm* || "$(uname -m)" == aarch64 ]]; then
> # arm/aarch64 specific modules
> + _blockfuncs+='|dw_mc_probe|dw_mci_pltfm_register'
> instmods \
> "=drivers/clk" \
> "=drivers/dma" \
> "=drivers/i2c/busses" \
> + "=drivers/mfd" \
> "=drivers/phy" \
> "=drivers/power" \
> "=drivers/regulator" \
> --
> 2.12.2
>
--
To unsubscribe from this list: send the line "unsubscribe initramfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
