On 2020-07-19, Jonas Smedegaard wrote: > Quoting Jonas Smedegaard (2020-04-27 09:00:49) >> Quoting Vagrant Cascadian (2020-04-20 19:25:15) >> > Control: tags 935035 moreinfo >> > >> > On 2019-08-18, Jonas Smedegaard wrote: >> > > U-Boot supports the DIY laptop Olimex Teres-I, except the builtin >> > > keyboard is not detected. >> > ... >> > > A working patchset is pushed to git branch wip-teres-i-keyboard: >> > > https://salsa.debian.org/debian/u-boot/tree/wip/teres-i-keyboard >> > ... >> > > This patchset is not in mainline U-Boot, but I intent to propose it. >> > >> > Any status update on mainline support? Is it still an issue with newer >> > u-boot packages in Debian? >> >> Last I checked it was still needed. >>
>> I noticed a recent change to the USB driver (commit 31232de) that >> *might* render all or half of my patch unnecessary, but that's just a >> guess based on the commit message - needs to be actually tested. >> >> If still needed, then status of the patch is that U-boot developers >> wants a more general cleanup of the related options, won't accept my >> minimal patch as-is. Again, I have not yet taken the time to do that. > > Patch is still needed, and I have now updated it upstream: > https://patchwork.ozlabs.org/project/uboot/patch/20200719135632.681954-2...@jones.dk/ This link gets both patches: https://patchwork.ozlabs.org/project/uboot/list/?series=190758&state=* It looks like they've gotten a few reviews already, and presuming no objections surface, I could apply them in the next u-boot 2020.07 upload, and hopefully will be applied upstream for 2020.10. > The newest upstream patch is larger, because developers requested I do a > more general change, where earlier proposed patches generalized only > across sunxi devices. > > There should be no functional difference between the older more minimal > patches and the newer one, and I therefore propose to apply the older > one for Buster. The wip-teres-i-keyboard branch on salsa? That includes a bunch of changes for platforms (unifier) that we don't support. I think could probably be trimmed down to a more minimal patch. e.g. It doesn't look like CONFIG_USE_PREBOOT is needed at all; I *think* you could just use CONFIG_PREBOOT without adding Kconfig support (needs testing) and then use ifdef/ifndef directly where the preboot command is added. I'd normally be hesitant to add non-upstreamed patches, and will want to wait to see how that plays out, but once the functionality lands in experimental or unstable, I'd be willing to consider a more minimal patch for buster. > Please tell me if you disagree, then I will prepare a backport of the > newest more general patchset. I think the full patchsets are too invasive for buster, so would prefer to avoid it. > Or, obviously, please tell me if you consider it unlikely that this > change can be accepted in Buster, so we stop spend more time on that. I think it can be done, but we'll have to be careful not to break other platforms. I have several other platforms (pine64+, pinebook, others) that might be impacted by these changes, so can check for breakage at least. live well, vagrant
signature.asc
Description: PGP signature
