12:08 <grawity> ukky: in this case? mostly just to get at some kind of interactive CLI that's serial-friendly (and would let me e.g. manually specify vmlinuz options) -- neither grub nor syslinux play well with this particular device's console/serial redirection

12:09 <grawity> but anyway I ended up pxe-booting openbsd in the end

12:13 <grawity> would still be handy if the higher level "bootloader" part of u-boot was usable without the "system BIOS replacement" part though

14:25 <marex> grawity: https://readme.phys.ethz.ch/documentation/how_to_setup_the_serial_console/

14:25 <marex> grawity: seems grub does support serial console

15:33 <Forty-Bot> what's the name for when the linker combines two static variables with identical contents

15:33 <Forty-Bot> or something like that

15:37 <f_> 7win 14

15:37 <f_> Oops.

15:41 <ukky> Forty-Bot: '-fmerge-constants' and '-fmerge-all-constants'? https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html

15:42 <Forty-Bot> no, it's something else

15:42 <Forty-Bot> iirc it had to do with zero initialization

15:42 <grawity> marex: it does but things get weird on this particular platform, since it has no VGA output whatsoever, so both syslinux and grub end up thinking that the console is 0 columns wide... and they both try to mirror their console output to serial output instead of having them independent, so grub's line editing is all kinds of wonky, for example. In contrast both u-boot and OpenBSD bootloader just take

15:42 <grawity> regular line input so they work fine.

15:56 <marex> grawity: so fix grub, send patch, and problem solved ?

15:57 <grawity> there's only so many things I can "fix" in my free time

15:58 <ukky> grawity: Check this patch for BIOS>GRUB>U-Boot boot sequence: https://lists.denx.de/pipermail/u-boot/2011-November/108932.html

16:15 <marex> ukky: 12 years ago :)

16:15 <marex> grawity: its probably a matter of hacking in some constant, so, shrug

16:21 <ukky> marex: What grawity wants is not that simple to achieve. But maybe following that patch would be the easiest way to inject U-Boot in-between Legacy X86 BIOS and Linux.

16:21 <grawity> I was hoping to make the process easier, not harder, but thanks

17:30 <marex> ukky: "syslinux and grub end up thinking that the console is 0 columns wide" -> should be easy to patch , no ?

17:36 <ukky> marex: yes. But they should be willing to patch. Most are just looking for what is already available.

17:53 <marex> ukky: scratch your own itch or something

17:53 <marex> seems nothing is readily available

19:16 <Forty-Bot> marex: form imx8, is the ROM page size `rom_api_query_boot_infor(QUERY_PAGE_SZ, ...)` 4k?

19:16 <Forty-Bot> *for

19:35 <GrantM11235> I have hit this assert at __fsg_is_set a few times on sunxi https://source.denx.de/u-boot/u-boot/-/blob/25edd247a84a31298c22a34ec5cf2851cbf61f70/drivers/usb/gadget/f_mass_storage.c#L397

19:35 <marex> Forty-Bot: I dont know, but I could try it at some point, why ?

19:36 <Forty-Bot> I'm wondering if it's actually dynamic or if it's something we can hard-code a max size for and panic if we assumed something too small

19:36 <GrantM11235> Is this a know issue? Am I doing something wrong? It is an intermittent problem, most of the time ums works fine

19:38 <Forty-Bot> GrantM11235: that looks like a bug; you can send a report to the list

21:26 <marex> paulbarker: well, RZ/G2 is at 95% and the UART patches also work, nice