10:11 <f_> Kwiboo: A311D is g12b I thought?

10:11 <f_> g12a is S905X2 IIRC?

10:18 <Kwiboo> f_: you are correct, my initial though was that my vim3 was g12a, I trusted the output from BL31 that mentions G12A :-)

10:19 <f_> BL31 always says G12A I think

10:19 <f_> bl2 does say g12b on g12b though iirc.

10:21 <Kwiboo> yes I noticed that, my vim3 is in fact a g12b with a a311d soc after a closer look, and the radxa zero is a s905y2 (g12a) and the radxa zero2 is a a311d (g12b)

10:23 <Kwiboo> my testing on all those three have resulted in three different results :-S, vim3 could freeze due to an irq in early resume and also froze later in resume process, the zero2 (same soc) did not have an irq freeze and looked to come much longer, but still froze just after it reported resume is complete

10:23 <Kwiboo> and on the zero (g12a) everything works fine :-)

10:26 <Kwiboo> I will likely just adjust the patch message to be more accurate and still include it when I send out the series that also fixes/cleanup other parts

10:30 <f_> I'll have to give 7.1 a try when I can

10:36 <Kwiboo> f_: thanks for assisting and testing!, I will try to get the meson dw-hdmi patches out later, found a new use-after-free issue with dw-hdmi-cec so I will re-spin the initial dw-hdmi series first

12:52 <narmstrong> I got both socs in CI, I'll try to give it a spin and try to do a suspend/resume

21:25 <josch> 👀 my irssii highlight on suspend/resume triggered -- i have a few bananapi cm4 with a311d here which would like to be able to resume (i can mail hardware to somebody if that would help)

21:27 <f_> Resume doesn't work for you?

21:27 <f_> I'll have to try on my CM4 too then

21:30 <josch> f_: yes, last time in #mnt-reform you said that the problem is that the SCP firmware that is bl30 has no idea what the reform keyboard is so it cannot be resumed via the keyboard

21:30 <f_> oh that

22:03 <Kwiboo> I have now posted the drm/meson dw-hdmi series and the last patch is related to suspend/resume, I was not fully sure it that patch is needed/wanted: https://lore.kernel.org/dri-devel/20260518194744.2483580-14-jonas@kwiboo.se/

22:04 <Kwiboo> even the existing call in meson_dw_hdmi_pm_suspend() to 'Reset TOP' looks wrong, like it is taking the block out of reset instead of putting it in reset :-S