16:16 <Hypfer> Are there any special things to consider if I'd want to load and execute toc0 for an A133 via FEL or is it like with the previous eGON format where you just load it at address xyz and then jump into that?

16:16 <Hypfer> Also, does what I just wrote make any sense or have I misunderstood something?

16:17 <Jookia> Has anyone managed to get CAN working on the T113 or D1?

16:28 <Hypfer> What I can find via google is that 2017, a ToDo item about booting toc0 via FEL was added to the wiki https://linux-sunxi.org/index.php?title=TOC0&oldid=19608

17:08 <apritzel> Hypfer: TOC0 is different in that you can't just load and jump to it

17:08 <apritzel> I don't think anyone ever attempted to provide TOC0 support for sunxi-fel

17:09 <apritzel> as in: parse the TOC0 on the host, and just upload the SPL/boot0 payload to SRAM

17:09 <apritzel> however there is code that converts between TOC0 and eGON: https://github.com/ssvb/sunxi-tools/commits/toc0

17:10 <apritzel> that might allow you make a standard eGON image out of the TOC0, and just use that

17:49 <Hypfer> apritzel will that converted image also work if secure boot is enabled or would that un-sign the image?

18:07 <Hypfer> hmmh that ruby toc2egon file doesn't like this toc0 file for some reason. chunk1_size is too large by 4 byte

18:10 <Hypfer> if I edit that check out, it still complains about the missing eGON header but maybe that's because the script expects a toc0 built by egon2toc.rb?

18:18 <apritzel> Hypfer: with FEL booting you actually don't need any eGON or TOC0 format, and any signatures are irrelevant, because in FEL mode you are already unlocked

18:19 <apritzel> IIUC in TOC0 the payload is just signed, not encrypted, so you can just ignore the signature and extract the SPL/boot0

18:20 <apritzel> Hypfer: ah, good point about toc2egon just being the reverse, it might indeed not be universal

18:21 <apritzel> Hypfer: if you don't mind the extra mile, check out https://source.denx.de/u-boot/u-boot/-/blob/master/tools/sunxi_toc0.c

18:21 <Hypfer> does the SPL/boot0 have some magic value as a header so that I could conceivably peel that out of the toc0 without having to parse the format?

18:22 <apritzel> or the explanation here: https://linux-sunxi.org/TOC0

18:23 <apritzel> I think the SPL payload embedded in TOC0 does not have a special header, but with the information above and a hex editor you should be able to figure out the offset

18:24 <Hypfer> oh I saw that you were the person who had some comments on the now-closed A133 PR: https://github.com/linux-sunxi/sunxi-tools/pull/192

18:25 <Hypfer> do you think that apart from code quality there are any major issues that might make this patch unsuitable for a quick test run?

18:25 <Hypfer> I seem to be able to read and write memory and also jump to adresses so I guess it should be enough?

18:27 <apritzel> A133 support in sunxi-fel? That should be good to go, maybe observing the comments

18:27 <apritzel> I had my version working on my tablet

18:27 <Hypfer> Do you still have that somewhere?

18:27 <apritzel> sure, let me check

18:29 <apritzel> https://github.com/apritzel/sunxi-tools/commits/a133-wip

18:29 <apritzel> Hypfer: just pushed it there

18:30 <Hypfer> thanks a lot!

18:30 <apritzel> my tablet didn't use TOC0, IIRC, so I just uplodaded boot0 and let it initialise the DRAM

18:31 <Hypfer> i suppose those parameters are hw-specific or would it also be possible to push any boot0 as long as it's the same dram type?

18:33 <apritzel> yeah, the DRAM parameters might differ, and it would outright fail if you have a different DRAM type (LPDDR3 vs. DDR3 vs. LPDDR4, for instance)

18:34 <apritzel> Hypfer: my tablet is the Teclast P80, what device do you have?

18:36 <apritzel> Hypfer: try this one: https://paste.c-net.org/JokerComrade

18:36 <apritzel> if I can trust my notes, this should be the version that returns to FEL after init'ing the DRAM

18:41 <apritzel> that blob is expecting DDR3 at 624 MHz

18:58 <Hypfer> ah okay that probably won't work here then. It's DDR4 iirc

18:58 <Hypfer> thanks though!

19:18 <Hypfer> okay I've finished a rudimentary parser for my toc0 based on the wiki but what confuses me about it is that it only has two item headers for two items of which both are type 0x010101 certificate item

19:19 <Hypfer> it is way larger than those 2x 764 byte + headers though. a full 96kb

19:20 <Hypfer> heh. No nvm I'm just stupid. Forgot to increment the offset :)