PATCH mmc: sunxi: fix MACH_SUN9I typo in get_mclk_offset()

[Omar Ivan Fardjoume]

>
> Hi Andre,
>
> Thanks for the review and for offering to fix up the patch formatting —
> much appreciated. And also thanks for working on Ethernet for the A80
> boards — I'd be happy to test those patches once you post them.
>
> Regarding the 8-bit bus width condition, I agree that an exclusion-based
> approach (!SUN4I && !SUN5I && !SUN6I && !SUN7I) seems cleaner long-term
> than having to add every newer SoC to an allow-list.
>
> For the get_mclk_offset() issue, I also grep'ed the tree and confirmed
> that CONFIG_MACH_SUN9I_A80 only appears in that location, which was
> actually what caught my attention while investigating why my eMMC was not
> working.
>
> Regarding eMMC boot partitions, I haven't tested them. My setup currently
> boots from the regular eMMC user area. I'm not building a complete
> distribution from scratch; I'm using JohanG's Debian Bookworm images and
> U-Boot as a base while working on restoring functionality on the
> Cubieboard4 A80. Debian Bookworm is currently running reliably on my board
> after applying several device tree (DTS) fixes and additions.
>
> So far I have:
>
>    -
>
>    eMMC boot without any SD card present.
>    -
>
>    all four USB Type-A ports.
>    -
>
>    AP6330 WiFi.
>    -
>
>    Ethernet (already working from JohanG's images).
>
> The two fixes I posted were required to get reliable eMMC boot on my
> board. Before fixing them, MMC2 was effectively unusable during boot. After
> applying both changes, SPL successfully loads U-Boot proper from eMMC and
> Debian 12 boots normally without an SD card installed. At least until you
> run apt upgrade, the kernel changes, and it refuses to boot again — which
> I suppose is the next mystery waiting to be solved.
>
> For my convenience, I use a script that clones the running Debian
> installation from SD to eMMC and installs U-Boot there. Since this setup
> uses the normal eMMC user area and I am currently using JohanG's pre-built
> U-Boot binaries, I have not had a need for CONFIG_SUPPORT_EMMC_BOOT in my
> workflow. However, if support for the dedicated eMMC boot partitions is
> considered useful for future users, enabling it in the defconfig sounds
> reasonable to me.
>
> I have documented the process and supporting scripts in a public
> repository in case it is useful for anyone else trying to keep A80 boards
> alive.
>
> If useful, I'd also be happy to test future A80-related patches on real
> hardware. There do not seem to be many active A80 users left, so I'm glad
> to help with testing whenever needed.
>
> Thanks again for taking the time to review and test the fixes.
>
> Best regards,
>
> Omar Ivan Fardjoume
>