04:08 <Dark-Show> I seem to be one of those affected GLXBadFBConfig users, RPi4 64bit kernel and userland (Fresh install using newest PiOS image). Box86 + Wine32 won't launch anything that translates a need for OpenGL. Box64 + Wine64 untested. Making matters more confusing, commits that I personally had marked as "mostly stable" now give the same error. Leading me to believe (guess) its related to the 5.15 -> 6.1 kernel update that happened recently. The

04:08 <Dark-Show> distro packages seem to not suffer from GLXBadFBConfig but have numerous rendering issues I haven't seen before. I'd chock it up to it being a RPi4 but I (googled) seen many other mesa drivers are hitting this issue currently as well.

04:11 <Dark-Show> As a side note, when cross compiling on the RPi4 itself, I use '-Dllvm=disabled' for the armhf side.

12:07 <zmike> you can try creating an issue and post some info there

12:20 <Dark-Show> Been meaning to make an account, might be the push I needed

13:15 <neobrain> Am I right to assume that Zink doesn't typically call vkBindBufferMemory with an address like 0x1c100000000? (i.e. perfectly aligned)

13:15 <neobrain> (For context, I'm trying to get guest-side zink running in FEX and running into some issues)

14:05 <zmike> it's whatever the driver supplies

14:15 <neobrain> oh, that might actually be a potential bug then. Do you know off the top of your head which API zink pulls the offset from?

14:16 <neobrain> ... I guess it depends on the specific context, which I sadly don't have since I'm looking at an undebuggable guest black box D:

14:22 <neobrain> Also looks like bo_create_internal is getting called with size=0, which I'm guessing isn't right either

14:43 <zmike> yea that sounds bad