Is openH264 being used? Seems not but unsure.
If it is being used, it needs both SECTION .text and SECTION .rodata replaced by macros similar to the others. Guess best added to asm_inc.asm.
This should be in the development thread, but replying here cos it's easier

@mods perhaps you can move the dev comments to the other thread?
Just got home from a weekend away, just confirming a couple of things.
1) re: ffmpeg - confirming that obj\third_party\ffmpeg only contains:
15/02/23 16:32 735 124 ffmpeg.ninja
8/01/23 19:36 0 124 ffmpeg.stamp
26/01/23 17:35 0 124 ffmpeg_features.stamp
15/02/23 16:32 1,151 124 ffmpeg_internal.ninja
8/01/23 19:36 1,536 124 libffmpeg_internal.a
2) re: openh264, confirming that obj\third_party\openh264 contains:
12/02/23 16:24 <DIR> 124 common
12/02/23 16:24 <DIR> 124 encoder
12/02/23 16:24 <DIR> 124 processing
15/02/23 16:32 4,867 124 common.ninja
12/02/23 16:24 0 124 common.stamp
15/02/23 16:32 9,129 124 encoder.ninja
12/02/23 16:24 0 124 encoder.stamp
15/02/23 16:32 6,793 124 processing.ninja
12/02/23 16:24 0 124 processing.stamp
and there are 65 .o files in the 3x directories above, so I think the asm is being used. I'll try look into making the necessary changes during the week.
Edit: openh264 isn't currently using asm on os/2:
use_assembler = (is_win || is_mac || is_linux || is_chromeos) &&
(current_cpu == "x86" || current_cpu == "x64") && !is_msan
Edit2: I updated yuv's row_x86.asm but it didn't trigger any source to be rebuilt. Grepping the libaom I don't see anything that triggers row_x86.asm to be built

Edit3: I'm going to rebuild *without* proprietary codecs being enbaled - I don't believe that switch is enabled in the bww chromium build, I just enabled it to try and increase the html5test score
