On 14.03.2018 19:10, Stephen Warren wrote: > On 03/14/2018 06:11 AM, Peter Geis wrote: >> Good Morning, >> >> Thank you, this is definitely part of my problem. >> Is there a way, or should there be a way to denote this via the device tree, >> instead of patching the kernel to accept a command line function? >> (Forgive me if this is a stupid question, this is my first true foray into >> kernel development) > > I am not aware of any way to represent storage device partition layout in device > tree (although for raw MTD devices I think there is a way). > > It would be possible in theory to develop such a feature, but since the current > firmware on the device doesn't support it, and hence there's no good way to get > the information into the DT, I don't expect doing so would solve any problems here. You can set the cmdline in DT, see "2.3 Runtime configuration" of [0]. See [1] for how to specify partition in the cmdline, basically you need to specify partition size and its start offset. But first check if bootloader passes the address of GPT in cmdline, that would be the easiest variant. [0] https://www.kernel.org/doc/Documentation/devicetree/usage-model.txt [1] https://www.kernel.org/doc/Documentation/block/cmdline-partition.txt >> On 03/13/2018 04:58 PM, Dmitry Osipenko wrote: >>> On 13.03.2018 22:24, Stephen Warren wrote: >>>> On 03/13/2018 01:15 PM, Dmitry Osipenko wrote: >>>>> On 13.03.2018 21:08, Peter Geis wrote: >>>>>> Good Afternoon, >>>>>> >>>>>> I have successfully booted 4.14 and 4.16 on the Ouya, with L2 cache support. >>>>> >>>>> Awesome :) >>>>> >>>>>> I am having three major hangups right now. >>>>>> >>>>>> I can't seem to find a valid frambuffer driver for the fbconsole. >>>>> >>>>> Please show the dmesg log, it should contain some errors. Add drm.debug=0xe to >>>>> the kernels cmdline. >>>>> >>>>>> The old kernel eEMMC support seems to be a hack job. >>>>> >>>>> Most likely eMMC has a custom downstream partitioning that is not supported by >>>>> upstream. You should be able to specify partitions manually using >>>>> CONFIG_CMDLINE_PARTITION, though I never used it myself. If that's your >>>>> problem >>>>> with eMMC. >>>> >>>> Yes, older Android devices tended to use a "TegraPT" (Tegra Partition Table) >>>> format rather than GPT. They also tend to consider the eMMC boot0, boot1, and >>>> user HW partitions as a single logical device that gets partitioned by TegraPT, >>>> rather than as completely separate devices. At least that's true for the >>>> flashing utilties; I can't remember whether TegraPT only exists within the user >>>> HW partition or not. The best reference for this is one of the NVIDIA >>>> downstream >>>> kernels (such as the original Ouya kernel source). >>> >>> Or if bootloader passes something like "gpt gpt_sector=61079551" in cmdline, >>> then apply this patch [0] and it should just work (if eMMC HW is up and running >>> fine). >>> >>> [0] >>> https://github.com/digetx/picasso_upstream_support/commit/9ced0e6b27545953bd68b9b042d221c6faade203 >>> >>> >> > -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
