| + my_pc_is_amiga (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 03:18 | |
| my_pc_is_amiga | I was just trying 1.10 -- don't have AX yet. I still see this issue and so assume this wasn't addressed in the 1.10 release: https://www.gregdonner.org/zz9000/troubleshoot10.mov | 03:21 |
|---|---|---|
| my_pc_is_amiga | this is on an A4000T... | 03:55 |
| - my_pc_is_amiga (QUIT: Quit: Client exiting) (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 05:21 | |
| + my_pc_is_amiga (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 05:57 | |
| my_pc_is_amiga | A totally seperate thing I found and posted here: https://community.mnt.re/t/zz9000-p96-driver-mu-hit-with-displaychain-no/880 | 05:57 |
| my_pc_is_amiga | this is only happening when I set DISPLAYCHAIN=NO...with DISPLAYCHAIN=YES no hit... | 05:58 |
| my_pc_is_amiga | I was trying DISPLAYCHAIN=NO, just to see if any issue going between graphic cards, etc. and this looks like it esentially turns off flickerfixer (Amiga output on HDMI) and so no longer see any split screen/corrupt screen issues that I see with DISPLAYCHAIN=YES. | 05:59 |
| - my_pc_is_amiga (QUIT: Quit: Client exiting) (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 06:09 | |
| + my_pc_is_amiga (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 07:26 | |
| my_pc_is_amiga | I don't know how it is really working but there is a MoniSwitch member of the BoardInfo structure that is not be utilzed in SetSwitch...not sure if this is at all related to ZZ9000 switching issues between Amiga output and RTG or the Mu hit | 07:30 |
| - my_pc_is_amiga (QUIT: Quit: Client exiting) (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 07:47 | |
| _Bnu | So, some code that gets executed when displaychain is set to no, but not otherwise? I assume it's the call to "SetGC" that's messing with it, but I don't actually know what that part is supposed to do. | 10:47 |
| _Bnu | (I haven't touched that one since I started working on the driver.) | 10:47 |
| minute | hum tee dum | 10:52 |
| minute | unrelated fun fact: there's no Z2 release yet because AX audio is very distorted in Z2 mode but fine in Z3 mode | 10:53 |
| _Bnu | Even in regular scandoubler mode? | 10:57 |
| minute | hmm i did not try that, good call | 10:58 |
| minute | my assumption is that maybe there is some scribbling over buffers | 10:58 |
| _Bnu | Ah, I haven't touched the scratch buffer stuff in Z2 mode at all, so it should be the same as before. | 12:56 |
| _Bnu | (Mostly because there was no additional memory to be had there.) | 12:56 |
| minute | mhmmm mhmm | 14:04 |
| _Bnu | That's why there are those checks for when the scratch area is used in the driver, where it doesn't do it in case it's too big for Z2. | 14:25 |
| _Bnu | Since it's only 128KB or whatever it was. | 14:25 |
| _Bnu | 64KB, maybe? | 15:04 |
| _Bnu | 128KB was perhaps the size of that creepy FPGA SRAM. | 15:05 |
| + my_pc_is_amiga (~my_pc_is_@c-71-197-66-36.hsd1.ca.comcast.net) | 16:11 | |
| my_pc_is_amiga | @_Bnu: Thomas on EAB made a comment on the hit...based on that I used "Findhunkoffset"findhunkoffset libs:Picasso96/ZZ9000.card 0 0x00000770 | 19:15 |
| my_pc_is_amiga | Section 0, offset 0x00000770 is at | 19:15 |
| my_pc_is_amiga | _SetGC + 0x0000000e | 19:15 |
| my_pc_is_amiga | and so yeah seems to be coming from SetGC() | 19:16 |
| my_pc_is_amiga | still the bigger issue for me is the switching of amiga video and RTG when DISPLAYCHAIN=YES (not related to hit)... | 19:16 |
| - xet7 (QUIT: Remote host closed the connection) (~xet7@user/xet7) | 19:36 | |
| + xet7 (~xet7@user/xet7) | 19:38 | |
| Chain|Q | my_pc_is_amiga: based on the hit, and the offset, I think SetGC() is called with NULL ModeInfo in this case, so when the driver tries to read mode_info->Width, it hits. (0x12 offset is Width in the ModeInfo struct, and that's the first read there, also A1 is NULL in your hit regdump) Now, in turn SetSwitch() (gonna assume only called with DISPLAYCHAIN=NO) in the driver calls SetGC(), with b->ModeInfo as | 21:11 |
| Chain|Q | argument, but in turn SetGC() b->ModeInfo is set inside SetGC() first... So if that's some first "init" call, and SetSwitch is called first, ModeInfo could be NULL there, and there's your hit. | 21:11 |
| Chain|Q | this is entirely just by reading the code, but I'm pretty confident something like this happens. | 21:12 |
| Chain|Q | cc: _Bnu & minute | 21:12 |
| forcie | https://cdn.c3voc.de/hls/gerpdemo/native_hd.m3u8 | 21:14 |
| forcie | compos at the GERP 2022 party are live | 21:14 |
| my_pc_is_amiga | @ChainIQ -- thanks for checking --- this seems like that is it... | 22:03 |
| my_pc_is_amiga | from what you are saying... | 22:03 |
| _Bnu | Yes, that was what I figured as well, but I don't actually know what the code is supposed to do. | 23:57 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!