08:01 <apritzel> warpme (or anyone else): do you know who the original author of the Trusted Firmware "T507 patches" is? I need a Signed-off-by: for the upstreaming

08:02 <warpme> it was yunari i think

08:13 <apritzel> junari__: did you write the T507 TF-A patches? I reworked and fixed ;-) them, but want to keep your authorship, of course

10:01 <apritzel> junari: did you write the T507 TF-A patches? I reworked and fixed ;-) them, but want to keep your authorship, of course

10:03 <junari> apritzel: Yes, iuncuim is my nickname on github

10:03 <apritzel> junari: ah, great, thanks!

10:05 <junari> I don't care about autorship

10:05 <junari> But it would be interesting to see your work

10:08 <apritzel> junari: but you did quite some work on figuring the registers out, so you deserve the credit

10:09 <apritzel> and legally I cannot just take your code from github and submit it, because I need your Signed-off-by: for that

10:43 <apritzel> junari: warpme: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/20334

10:44 <apritzel> please also mind that there are four patches before that, which prepare the stage

10:50 <apritzel> warpme: junari_: I tried to extend the power up/down sequence to match the original one, please let me know if that breaks anything

11:08 <warpme> apritzel : on mainline 2.8 atf i'm getting like this: https://controlc.com/e05d8d3e

11:09 <junari> warpme: You lost one of the patches

11:10 <junari> feat(allwinner): add oracle function to detect H616 die variant

11:11 <junari> apritzel: Thanks for the proper patches

11:12 <warpme> hmm - so commit https://review.trustedfirmware.org/plugins/gitiles/TF-A/trusted-firmware-a/+/ca78e4096caf98b54c745341770602a4d422e880 in not enough?

11:12 <apritzel> warpme: no, you need all five of them

11:12 <warpme> ok https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/20333 thisi also needed

11:15 <warpme> ok clear!

11:39 <warpme> apritzel : just to confirm: for both varinats of h313 i need to use bl31 from mainline atf 2.8 + 5 patches built for PLAT=sun50i_h616 right?

12:42 <apritzel> warpme: yes, the detection is automatic

12:43 <apritzel> I based the patches on current mainline, but I guess it works with 2.8 as well

12:44 <warpme> asking as on "t507-like" box i'm getting famous "only 1 core booted" : https://controlc.com/b200fcc5

12:51 <apritzel> warpme: thanks, will have a look

12:58 <apritzel> warpme: btw, on a booted system, can you check if /sys/devices/soc0/revision differs between the two boxes?

14:21 <warpme> apritzel : here it is https://controlc.com/acfdfe9e6

14:24 <warpme> i see in form atf: BL31: Detected Allwinner H616 SoC (1823) maybe it is worth to add "rev00[02]" in this log line?

14:25 <warpme> in form -> from

14:50 <apritzel> warpme: somewhat interesting that it doesn't work, since I didn't really change the power on sequence

14:56 <warpme> apritzel : i personally learn that when playing with HW variants/quirks it is really useful to add logging/verbosity. Current code behaviour seems the same like mainline without t507 additions. So maybe soc rev. isn't correctly detected or is detected but logic is different or.... Pls add debug log entries and I will test it i.e. on my hw....

15:24 <apritzel> warpme: you can easily add INFO("..."); lines into the two branches of sunxi_cpu_on(), but you won't see them at Linux runtime, unless you enable TF-A output at "runtime" as well

15:24 <apritzel> which you should be able to do by commenting the call to console_switch_state() in plat/allwinner/common/sunxi_bl31_setup.c:bl31_plat_runtime_setup

15:59 <warpme> apritzel : i added INFO in conditionals; commented console_switch_state(). rebuild. I see my INFO strings in binary. But on boot can't find any strings in console nor uart

16:00 <apritzel> I then might need to look up again how to exactly enable runtime logging, will report later today