Aw, I thought the entire machine should work at the first patch. Thank you for the detailed clarification! I think I've got the idea. I'll split it up.
Could you please tell me what size is appropriate for a single patch? ________________________________ От: Peter Maydell <peter.mayd...@linaro.org> Отправлено: 22 мая 2023 г. 13:58:42 Кому: Kambalin, Sergey Копия: Sergey Kambalin; qemu-...@nongnu.org; qemu-devel@nongnu.org Тема: Re: [PATCH] [rpi4b] Make bootable rpi4b model On Mon, 22 May 2023 at 11:42, Kambalin, Sergey <sergey.kamba...@auriga.com> wrote: > > Hi! > > Unfortunately it can't be split without losing a functionality. It is a > minimal amount of code to make it able to boot the kernel (and therefore > confirm that it works). No, it absolutely can. Each individual patch should be a coherent chunk of work, and needs to compile cleanly, but it doesn't have to be immediately useful on its own. The usual setup is that a patchseries adding a new board gradually adds pieces like new devices or bugfixes to existing code, and it's only in a patch fairly late in the series that the new board proper is added and enabled. In a 5 minute scan of this patch I saw at least one cleanup patch that should be separate (changing hard-coded numbers in the switch cases in the bcm2835_property.c file). Anything where you're touching the existing bcm2835/2836 code because you need to refactor it to be a better base for the bcm2838 work should be a separate patch (this is particularly important so we can review that the changes don't break the existing boards). And the usual approach with a new board is that you have a patch per new device being added (you have several here) and then a patch at the end for the board changes. New test cases can be their own patch. Documentation (which seems to be missing here) can be its own patch. I would estimate that this will end up being at least 6 patches, probably more. thanks -- PMM
