Re: One less thing blocking UEFI+FDT support for HoneyComb
Dmitry Salychev <[email protected]>
| Newsgroups | gmane.os.freebsd.devel.arm |
|---|---|
| Message-ID | <[email protected]> |
"Bjoern A. Zeeb" <[email protected]> writes: > On Thu, 30 May 2024, Mark Millard wrote: > >> (B) could well involve things not involving dpaa2 usage >> via FDT. > > Is there's a problem with that please email dsl@ and bz@ and we can have > a look. In general dpaa2 workes with both ACPI and FDT so... I asked about UEFI images on the unofficial solidrun channel [1] recently: dsl: btw, https://images.solid-run.com/LX2k/lx2160a_uefi seems abandoned. Does anyone know why? jnettlet: There is no business use case for it. Since we were unable to get patches into mainline linux for the final devices, and fixes there was no real reason to move forward with edk2 and ACPI since u-boot and device-tree didn't have these limitations but when I tried one of the FTD-based images from [2], it didn't work because of a missing PHY for 1 Gbps port. It looks like our "glue" code in memacphy_fdt/memac_mdio_fdt didn't work in this specific case. FDT overlay for "fsl_mc: fsl-mc@80c000000" to provide correct values for the DPMAC corresponding to the 1 Gbps port (dpmac17 IIRC) would be a workaround for now, but I'll probably have to debug more. [1] https://discord.com/channels/620838168794497044/665456384971767818 [2] https://images.solid-run.com/LX2k/lx2160a_build Regards, Dmitry -- https://wiki.freebsd.org/DmitrySalychev