03:19 <m4t> i ended up needing jtag anyways ;( because i flashed .bin and not .kwb

03:19 <m4t> consistently seeing a hang when running: mtdparts; mtdparts delall; mtdparts

04:41 <m4t> seems better behaved after setting CONFIG_MTDIDS_DEFAULT and CONFIG_MTDPARTS_DEFAULT

11:50 <alex88_> Hi everyne, is there a list of issues for uboot?

16:18 <ldts> agraf: quick question, some years ago - yeah, sorry- you enabled support for zynqmpbif, any reason why you never integrated with Makefile.spl?

19:25 <marex> ldts: ask monstr (CEST) tomorrow

19:26 <marex> ldts: or send patch

19:35 <ldts> yeah, before sending a patch I need to get it to work :)

19:36 <ldts> but sure, once I have SPL + u-boot.itb (FIT with u-boot, ATF, OP-TEE etc) will post something....

19:36 <ldts> would be nice to have some info though...anyway, testing now

19:44 <ldts> marex: ah, michal is monstr. ok, thanks

20:00 <agraf> ldts: what exactly are you wondering?

20:01 <agraf> ldts: The bif file format is typically used with the proprietary Xilinx tooling. Are you asking why I didn't make use of it with SPL?

20:01 <agraf> ldts: I really don't understand the question :)

20:01 <agraf> ldts: with SPL, your whole flow is typically different. You run SPL first and that then directly loads all subcomponents

20:01 <agraf> (Not sure how the PMU is handled in that case though - monstr would know)

20:06 <ldts> agraf: hey, thanks for coming back to me...so with BIF the user can specify the image layout to enable secure boot (just trying to sign SPL and have it checked with the bootrom). ...ATM, with the zynqmpimage I can of course boot SPL and then a u-boot.itb (all the other components) from QSPI. Now I just want to make sure SPL is validated by rom hence the need to move to a different image layout...I thought you used bif for this,

20:06 <ldts> otherwise, what was the original intention?

20:07 <ldts> https://www.xilinx.com/support/documentation/sw_manuals/xilinx2018_2/ug1283-bootgen-user-guide.pdf page 22 for the layout

20:07 <agraf> ldts: I didn't use secure boot with bif back then :)

20:08 <agraf> ldts: IIRC I had 2 reasons for bif

20:08 <agraf> ldts: One was that I wanted to work with the Vivado provided non-SPL early boot loader and have U-Boot loaded from there. But I wanted to have all of that built with open source tools

20:09 <agraf> ldts: the other reason was that I was building a generic version of U-Boot that just adjusted itself to the target config using a DT you provide in the bif

20:09 <ldts> ah I see...yes this is documented (I read this way of booting somewhere); ok

20:09 <agraf> ldts: that way I could have a single "distro" U-Boot build and the only thing you would need to do per board was to provide your own DT in the bif file and generate a blob

20:10 <agraf> ldts: But I never really got around to toy with the secure boot parts of it. Happy to hear it's useful for you under that aspect though! :)

20:11 <ldts> I hope it is - not working yet....but if it is not, it will be eventually (Vivado is eating my hard disk big time :) )

20:11 <ldts> so we want to move away from vivado and have just mkimage (instead of bootgen)

20:12 <ldts> thanks for the background..I understand a bit better now!

20:12 <agraf> ldts: Yeah, in some intermediate version of my endeavour I did pull in bootgen into the open build service as blob and ran it through qemu-x86_64 (the build was on aarch64!). But it was icky on all levels. Licensing as well as stability.

20:13 <agraf> ldts: I really don't understand why Xilinx doesn't just release bootgen under some OSS license. It's really hurting themselves the most

20:13 <ldts> I put it all in a container - so I can use jtag xsdb as well...it is about 10GB....crazy

20:15 <agraf> Sure, you can always add another level of abstraction ;)

20:15 <agraf> (Not that it would make it any safer / legal to use in an automated build environment ...)

20:15 <ldts> true. .ok, I'll keep at it, I mean, is "just" and image layout, it has to work one way or another

20:16 <ldts> (I am not using encryption btw, something that bootgen aparently can do)

20:18 <agraf> Yup :). Most of the bif and target boot file logic is publicly specified too

20:18 <agraf> Happy hacking then! :D

20:18 <ldts> thanks for doing that commit :)