| _Bnu | Which could possibly be solved by using some sort of really complicated DMA "command queue" in memory, but making sure that stays in sync and doesn't overflow or something without communicating with the ARM through AXI would also be a challenge. | 00:01 |
|---|---|---|
| mntmn | that's what axi is supposed to do... | 00:01 |
| mntmn | i mean the interconnect and stuff | 00:02 |
| mntmn | it doesn't really make sense to pile workarounds on top of that without first looking at how the axi stuff itself can go wrong perhaps | 00:02 |
| mntmn | M_AXI_GP0 has a bunch of config flags for example | 00:02 |
| mntmn | HAS_CACHE for example | 00:02 |
| _Bnu | Lack of Nibbles support, surely... | 00:02 |
| _Bnu | And yeah, I didn't dare touch any of those. | 00:03 |
| _Bnu | The only thing I've messed with is the unaligned transfers for the VDMA, haha. | 00:03 |
| mntmn | also, i see that for some reason all that usb buffer select stuff is going through multiple stages | 00:04 |
| _Bnu | I tried doing the usual Goole searches for XiLcAcHeFlUsH and stuff, but didn't find anything useful. | 00:05 |
| mntmn | axi_reg5 <= slv_reg5; | 00:06 |
| mntmn | usbb <= axi_reg5; | 00:06 |
| _Bnu | Greetings from Germany and usbb. | 00:06 |
| mntmn | that's a few clocks later | 00:06 |
| mntmn | if the amiga was quick enough maybe it could write to buffer memory via dma before usbb was set... | 00:07 |
| mntmn | or if the axi transaction is delayed enough... | 00:07 |
| mntmn | you know, such things | 00:07 |
| mntmn | and the axi stuff runs only at 100mhz\ | 00:08 |
| mntmn | naja the zorro state machine also at 100mhz | 00:08 |
| mntmn | so amiga can't go faster than that | 00:08 |
| _Bnu | Time to overclock the FPGA fabric to 175MHz so we have to rework all the video timings to 1.75x. | 00:08 |
| mntmn | mhm | 00:09 |
| mntmn | anyway i should go home && sleep | 00:10 |
| _Bnu | Me too... except I'm already home... | 00:11 |
| mntmn | !!! | 00:12 |
| - RobDangerous (QUIT: Ping timeout: 245 seconds) (~Thunderbi@p200300ec8f3e1c00946d7ef7cfdde9d0.dip0.t-ipconnect.de) | 00:18 | |
| mntmn | _Bnu: in the care home i thought about this more and i think the next logical step would be to move usb buffer sel to an fpga register (without going through arm) and or getting rid of multiple buffers/paging and introduce one bigger buffer | 01:07 |
| - bfarmilo (QUIT: Read error: Connection reset by peer) (~bfarmilo@202.56.40.240) | 01:12 | |
| + bfarmilo (~bfarmilo@202.56.40.240) | 01:14 | |
| - bfarmilo (QUIT: Ping timeout: 258 seconds) (~bfarmilo@202.56.40.240) | 02:26 | |
| + bfarmilo (~bfarmilo@202.56.40.240) | 02:31 | |
| CottonUndies | Q1: For the upcoming Audio Add-on for the ZZ, is there already there, or possible to consider a SPDIF digital out? | 02:39 |
| CottonUndies | Q: For the upcoming bootable ZZ USB: Is there any amazing possibility of having an additional USB storage stick work at the same time off of a hub? I only ask as if I can boot off of one USB, the 2nd USB would still remain useful to transfer files. Or maybe a solution already is available that I'm not thinking of? | 02:41 |
| - bfarmilo (QUIT: Quit: Leaving) (~bfarmilo@202.56.40.240) | 05:44 | |
| + bfarmilo (~bfarmilo@202.56.40.240) | 05:50 | |
| _Bnu | ACTION bibbers | 08:25 |
| _Bnu | CottonUndies: There's some way to use smb/samba/whatever it's actually called network shares on the Amiga, but unfortunately I don't know exactly how it works. I've never really had any working network shares on my home network, so... | 09:09 |
| _Bnu | But you can probably use that for reasonably fast file sharing between computrons. | 09:09 |
| Jope | smbfs works nicely | 09:22 |
| Jope | the latest one from github | 09:22 |
| + RobDangerous (~Thunderbi@p200300ec8f3e1c002d0ccd643d9ce569.dip0.t-ipconnect.de) | 09:49 | |
| _Bnu | ACTION commits some codes | 11:10 |
| mntmn | :0 | 11:17 |
| _Bnu | 49.51Hz should be enough for anyone, right... | 11:45 |
| pasik | no! we want 49.92 Hz! | 11:45 |
| pasik | :) | 11:45 |
| _Bnu | I mean, if MiniMig does NTSC Amiga at 59.84 Hz... | 11:45 |
| pasik | heh | 11:46 |
| mntmn | haha | 11:52 |
| mntmn | the pushenings | 11:52 |
| _Bnu | I really wish I could remember how I got the BALLSTER PROGRAM CABLE to work in Windows the last time around, lol. | 12:14 |
| _Bnu | A4000 pls. Just load the Pinball... | 12:17 |
| mntmn | oop | 12:29 |
| _Bnu | https://cdn.discordapp.com/attachments/789807768554831902/873152899105062922/unknown.png | 12:38 |
| _Bnu | Do you think it's close enough... maybe I can get it a tiny bit closer, though. | 12:38 |
| _Bnu | Abusing my custom RGBEasy applications to check the refresh rates... | 12:38 |
| _Bnu | Maybe I can get a tiny bit closer from the other side... let's see. | 12:41 |
| mntmn | haha what is this | 12:41 |
| _Bnu | Just using the RGBEasy SDK for the capture cards to monitor settings changes on inputs. | 12:41 |
| _Bnu | Input 1 being the ZZ9000 HDMI and Input 2 being the analog RGB. | 12:42 |
| _Bnu | Hm, next step down is 49.918. Technically closer to 49.920 than 49.926, so that's probably the one to go with. | 12:44 |
| pasik | sounds like datapath card ? | 12:44 |
| _Bnu | Yeah, I have a bunch of Datapath cards. | 12:45 |
| pasik | i have one aswell | 12:45 |
| pasik | visionrgb e2s or smth | 12:45 |
| _Bnu | Yeah, I have three of those, I think. Maybe two and one E2. | 12:45 |
| _Bnu | Two VisionAV with two inputs, and some E1 cards. | 12:46 |
| pasik | nice | 12:46 |
| _Bnu | Ah yeah, I have two E2S and one E2, because I sent an E2S to mntmn, haha. | 12:46 |
| mntmn | i need an arm64 driver for it... | 12:47 |
| mntmn | 49.918 rounded up is already 49.92 :3 | 12:48 |
| _Bnu | I would recommend running it in a Windows VM or something, since the SDK for Linux kind of sucks and is V4L-based. | 12:48 |
| Jope | nice and round <3 | 12:48 |
| _Bnu | So it keeps breaking every single time they update the V4L crap in the kernel. | 12:49 |
| _Bnu | Which is, like... every single time. | 12:49 |
| _Bnu | King of sucks. | 12:50 |
| _Bnu | https://cdn.discordapp.com/attachments/789807768554831902/873156338421739520/unknown.png | 12:51 |
| _Bnu | Also behold my super-sophisticated way of interfacing with the RGBEasyThing command line app... | 12:52 |
| pasik | 0.00 Hz | 12:52 |
| _Bnu | Yeah, no signal. I don't bother checking for no signal, though. | 12:52 |
| pasik | sorry, more precise 0.000 Hz | 12:52 |
| _Bnu | I turned off the Amyga because it's time for lunch... | 12:52 |
| mntmn | ok i am in the office | 13:21 |
| mntmn | _Bnu: do you think it would run on windows on arm... | 13:21 |
| _Bnu | Doubtful... maybe if you could get Datapath to compile a driver package for it, haha. | 13:23 |
| mntmn | get me datapath on the phone... | 13:26 |
| _Bnu | Let's see, they email me now and then... | 13:29 |
| _Bnu | +44 (0)1332 294 441 | 13:29 |
| mntmn | that's... the united kingdom | 13:31 |
| mntmn | ok i wanted to debug some usb buffer stuff >:| | 13:31 |
| mntmn | _Bnu: i can try to explain my theory | 13:32 |
| mntmn | usbb <= axi_reg5; | 13:32 |
| mntmn | this is in the always block, right | 13:32 |
| mntmn | so it is always happening, each clock | 13:32 |
| mntmn | and it comes from axi_reg5 <= slv_reg5; | 13:33 |
| mntmn | and this slv_reg5 is set in the AXI4-lite state machine from xilinx that is in the same file but different process | 13:33 |
| mntmn | we do not know for sure when a thing arrives there from AXI interconnect and those things | 13:33 |
| mntmn | so imagine amiga says, please select buffer 2 | 13:33 |
| mntmn | it goes through ARM, ARM says ok lets tell the FPGA about it | 13:34 |
| mntmn | via AXI4-lite interconnect | 13:34 |
| mntmn | so the little car with buffer 2 select in it goes on a roadtrip through AXI buses and interconnects | 13:34 |
| mntmn | while the ARM goes on with its business and acknowledges the zorro write for this | 13:34 |
| mntmn | (running at 666mhz, while the little axi car drives with 100mhz) | 13:34 |
| mntmn | so the next transaction is amiga writing some bytes directly to memory through the FPGA via AXI HP bus | 13:35 |
| mntmn | we don't know if the bufsel request actually arrives before that and if usbb is toggled on time | 13:36 |
| mntmn | did the little car arrive... | 13:36 |
| mntmn | it could even switch in the middle | 13:36 |
| mntmn | i.e. first byte goes to buffer 1, next byte goes to buffer 2... | 13:36 |
| mntmn | i will try to falsify my theory and prove myself wrong | 13:37 |
| mntmn | by introducing a register for usbb in the fpga zorro side of things. | 13:37 |
| _Bnu | I hate that little car... AXI should buy a new car... | 13:38 |
| mntmn | and setting that instead of going through zorro->arm->axi4lite->zorro | 13:38 |
| mntmn | i hope my theory is wrong but i also hope it is not wrong | 13:38 |
| mntmn | and regarding the cache flush: i imagine that axi transactions are halted for a while | 13:39 |
| mntmn | so this could maybe introduce a traffic jam | 13:39 |
| mntmn | while AXI HP is maybe still going through... but i have not confirmed any of this empirically | 13:40 |
| mntmn | we could also read back the usbb set before writing to it... | 13:41 |
| mntmn | naja. | 13:41 |
| mntmn | generate beepstream... | 13:52 |
| mntmn | it is really convenient to have device drivers inside of the firmware btw | 13:52 |
| mntmn | just one giant blorb... | 13:52 |
| _Bnu | ACTION generates blobs | 13:56 |
| _Bnu | Yeah, it's just somewhat inconvenient to update them, haha. That's why I keep them as separate external files on the PiStorm. | 13:56 |
| _Bnu | But without guaranteed file system access to a USB stick or something to load them from, that would be difficult. | 13:56 |
| mntmn | ZZetPatch | 14:00 |
| mntmn | lollll | 14:01 |
| mntmn | everything was fine until a buffer > 0 was needed... | 14:01 |
| mntmn | you should've seen the glitches | 14:01 |
| + isoriano (~isoriano@5.147.209.166) | 14:02 | |
| + isoriano_ (~isoriano@5.147.209.166) | 14:02 | |
| isoriano_ | Good afternoon. | 14:02 |
| RobDangerous | Oooh, Bnu blobs | 14:03 |
| mntmn | hi | 14:03 |
| isoriano_ | mntmn: you are the one who motivates people :-) https://ibb.co/g3HFTLv | 14:09 |
| mntmn | isoriano_: haha pretty cool, and i didn't have to do anything! | 14:10 |
| isoriano_ | I really like that idea so much with your ZZ!! | 14:11 |
| mntmn | _Bnu: LOL our USB buffer DMA area is 0xa000-0x10000 right | 14:12 |
| mntmn | 0xffff that is | 14:12 |
| mntmn | that's 24kB or 48 buffers | 14:13 |
| mntmn | the whole buffer toggling is nonsense | 14:13 |
| mntmn | because there are only 32 | 14:13 |
| _Bnu | I just need buffers... | 14:13 |
| _Bnu | Actually, I could really use someone to confirm to me what the refresh rates are on an NTSC Amiga. | 14:14 |
| _Bnu | It should be 59.928 and 50.35X Hz, but I have no way of checking myself. | 14:15 |
| mntmn | i also cannot check this... we need local experts | 14:16 |
| mntmn | ok bizarre | 14:28 |
| mntmn | if you do memcpy(a,b,512) it is ok for gcc to not include a real memcpy | 14:28 |
| mntmn | but if you do memcpy(a,b,512*x), that is a no-no, needs a real implementation | 14:28 |
| mntmn | the forbidden function http://amigadev.elowar.com/read/ADCD_2.1/Includes_and_Autodocs_3._guide/node01FA.html | 14:30 |
| - WoC (QUIT: Remote host closed the connection) (woc@2603-8080-cd40-00c6-eaaf-8a29-4225-dc1f.res6.spectrum.com) | 14:35 | |
| + WoC (woc@2603-8080-cd40-00c6-eaaf-8a29-4225-dc1f.res6.spectrum.com) | 14:35 | |
| _Bnu | CopyMemQuick is dangerous. | 14:39 |
| mntmn | it looks very dangerous | 14:44 |
| mntmn | ok writing now goes always wrong in predictable way. | 14:44 |
| mntmn | all the md5sums of my copies are wrong in the same way | 14:44 |
| mntmn | lets see | 14:44 |
| mntmn | maybe it would be smorter to make a file that has some regular pattern in it... | 14:45 |
| mntmn | they start to be wrong at 0x4000. | 14:46 |
| mntmn | that's the same like yesterday | 14:46 |
| mntmn | 0x9000-0xcfff is ok again | 14:47 |
| mntmn | 0xd000-0x21800 or so is wrong again' | 14:47 |
| mntmn | >>> 0x4000/512 | 14:48 |
| mntmn | 32 | 14:48 |
| mntmn | that's the number of buffers | 14:48 |
| mntmn | > #define BLOCKS_AT_ONCE 32 | 14:48 |
| mntmn | ok so it wraps around | 14:57 |
| mntmn | at 0x4000 the same content from 0x0000 starts again! | 14:57 |
| _Bnu | Power squats. | 14:57 |
| mntmn | ah i think because i made a bug | 14:59 |
| mntmn | ja | 14:59 |
| mntmn | this: memcpy(registers-0xd0+0xa000+offset+(j<<SD_SECTOR_SHIFT), data+offset+(j<<SD_SECTOR_SHIFT), 512); | 14:59 |
| mntmn | should be: memcpy(registers-0xd0+0xa000+(j<<SD_SECTOR_SHIFT), data+offset+(j<<SD_SECTOR_SHIFT), 512); | 14:59 |
| mntmn | ok workbench boots again :D | 15:03 |
| mntmn | wow id_data_abort: arm_exception_handler()! | 15:04 |
| _Bnu | It's probably poking in some memory area where it's not allowed... | 15:09 |
| _Bnu | Like the ETH buffers or something. I had this happen with the RTG when they were in the middle of the framebuffer memory somewhere. | 15:10 |
| mntmn | maybe | 15:10 |
| _Bnu | Also, I finally figured out the intended use of those micro SD->SD card adapters. | 15:10 |
| mntmn | ok now it works fine, but i also had to put a bit of m68k cache flushing after the amiga-side memcpy. because of 68030. | 15:10 |
| mntmn | is there a amiga disk speed testing thing... | 15:10 |
| _Bnu | Oh yeah, in case you missed it, I pushed a commit to drivers earlier. So you don't orphan it or something, haha. | 15:11 |
| _Bnu | There's SysInfo's "drives" thing. | 15:11 |
| mntmn | ah yeah | 15:11 |
| _Bnu | Should be around 5-10MB/sec? | 15:11 |
| mntmn | sysinfo crashes the amooga | 15:12 |
| mntmn | maybe because of my frankenworkbench | 15:12 |
| mntmn | (guru on start) | 15:12 |
| _Bnu | I think it comes with everything it needs on the floppy image, but not sure. | 15:13 |
| _Bnu | Also, the intended use of the SD card adapters is to push the micro SD card in/out of the ZZ9000... | 15:13 |
| _Bnu | https://cdn.discordapp.com/attachments/789807768554831902/873193485451276398/unknown.png | 15:19 |
| _Bnu | Luckily there seems to have been a perfect fit for NTSC on a PAL Amiga. | 15:19 |
| mntmn | awesome | 15:20 |
| mntmn | what did you tweak? | 15:20 |
| _Bnu | { 720, 480, 720, 752, 800, 490, 492, 525, 0, 25, 25175000, 60, 0, 61, 5, 49 }, // 720x480 non-standard VSync (PAL Amiga) | 15:20 |
| _Bnu | Clock wizard numbors. | 15:20 |
| _Bnu | // 61 / 05 / 049 = 24.8980 MHz | 15:21 |
| _Bnu | From my ever so useful findmhz application. | 15:21 |
| _Bnu | (It's literally only useful for this.) | 15:21 |
| _Bnu | 50.349... feels close enough to approximately 50.35. | 15:37 |
| _Bnu | It's also as close as I can get without going to 50.36 | 15:39 |
| _Bnu | https://cdn.discordapp.com/attachments/789807768554831902/873201343161065502/unknown.png | 15:50 |
| mntmn | cool cool! | 15:54 |
| mntmn | my coworkers forced me to make lunch, so i don't know where i was before doing that | 15:57 |
| _Bnu | I'm going to push another commit to the drivers... | 15:58 |
| _Bnu | Is it okay if I push some stuff to firmware wip19a, or do you have a zillion changes to main.c queued up? | 16:01 |
| mntmn | zillion | 16:01 |
| _Bnu | Or a trillion... | 16:01 |
| mntmn | i do have a bunch of stuff yeah... but it depends which parts you are touching | 16:01 |
| _Bnu | Just videocap stuff. | 16:01 |
| mntmn | should be fine | 16:02 |
| _Bnu | That's what they all say... and then the world ends... | 16:02 |
| mntmn | amiga is borked | 16:02 |
| mntmn | what to tell him https://twitter.com/Styrbjorn/status/1423645473450123264?s=20 | 16:03 |
| mntmn | dme driver or something? | 16:03 |
| _Bnu | Uhhh. | 16:06 |
| _Bnu | Never really seen that happen. | 16:06 |
| _Bnu | It looks like it's trying to display a requester. | 16:06 |
| mntmn | yeah | 16:06 |
| mntmn | this lha is illegal https://aminet.net/package/util/moni/SysInfo | 16:07 |
| mntmn | i cannot extract it fully on linux... | 16:07 |
| _Bnu | Is that AmigaOS 3.2? ScreenMode prefs look so weird. | 16:08 |
| _Bnu | Well, extracting Amiga LHA files on anything but Amiga is not recommended (TM). | 16:08 |
| mntmn | yeah everything looks weird there | 16:09 |
| mntmn | yeah i just wanted to compare the md5sum | 16:09 |
| mntmn | because SysInfo always crashes on the amiga | 16:09 |
| mntmn | all this corruption... | 16:09 |
| _Bnu | Deleting user data. | 16:10 |
| mntmn | ah i still have write corruption. now it's just in 10% of cases. | 16:10 |
| _Bnu | Also, Jope, RobDangerous, anyone else who feels like testing it, the .card files and the boot.bin in http://www.apehead.se/snakes9000 should now allow using the non-standard refresh rates. | 16:11 |
| mntmn | :O | 16:11 |
| _Bnu | Requires a file in ENV called ZZ9000-NS-VSYNC, or ZZ9000-NS-VSYNC-NTSC on an NTSC Amiga. | 16:11 |
| RobDangerous | Super cool, will spend the evening with this. | 16:11 |
| mntmn | ok this time my one borked file starts having errors at 0xd000. | 16:12 |
| mntmn | exactly 0xd000-0x10fff | 16:13 |
| _Bnu | Good, good... | 16:13 |
| mntmn | that's the same content repeated from 0x9000. | 16:14 |
| RobDangerous | What do all those different card files mean? For different CPUs? What's DME? | 16:15 |
| mntmn | >>> hex(0xd000-0x9000) | 16:16 |
| mntmn | '0x4000' | 16:16 |
| mntmn | that looks like the same 32*512 buffer is written again, i.e. not updated for the next set of writes. | 16:16 |
| _Bnu | RobDangerous: "DMA" is typically the one you want, the numbers are just compiled for different processors. | 16:17 |
| _Bnu | DME has the cache-tricking dummy read enabled, should only be used if you have issues with graphics corruption. | 16:17 |
| RobDangerous | Ah, all right. | 16:18 |
| mntmn | lets flush those caches | 16:18 |
| mntmn | down the toilet | 16:18 |
| _Bnu | Well, DMA is the one you want on Zorro III. Not on Zorro II... | 16:18 |
| RobDangerous | Going Zorro 3, my 90s hardware is up to date. | 16:19 |
| _Bnu | I realized I hadn't actually tested it with the file in Env-Archive... but now I have, and it does work... | 16:21 |
| _Bnu | But yeah, just create an empty file in Prefs:Env-Archive and name it ZZ9000-NS-VSYNC | 16:21 |
| _Bnu | Then you have to reboot for it to end up in ENV:, I believe. | 16:22 |
| RobDangerous | Aye aye | 16:22 |
| mntmn | hmm. how can this be so broken. | 16:24 |
| _Bnu | I don't know... but I'll fry some locally produced hamburgers, maybe that'll fix it... | 16:26 |
| mntmn | do you want to relauch | 16:31 |
| _Bnu | No, no re-lunch for me... dinner time... | 16:35 |
| mntmn | ok i might have fixed the problem and also sysinfo is ok now | 16:47 |
| mntmn | it says ~1 MB/s read test ;) | 16:48 |
| mntmn | but that's with a bunch of debugging going on | 16:48 |
| mntmn | yeah it is fixed | 16:52 |
| mntmn | _Bnu: the solution was to add a MMU TLB entry STRONG_ORDERED for the 1MB page that contains the usb buffer. | 16:52 |
| mntmn | we had the same for the ethernet buffer | 16:52 |
| mntmn | no explicit cache flushing needed then also | 16:53 |
| mntmn | we might wanna make sure that the same is done for any param/scratch buffers for graphics commands. | 16:53 |
| _Bnu | But I have no MMU config at all. ('A`) | 16:55 |
| _Bnu | Which means that you may still have the USB stuff lock up at some unspecified point in time. | 16:56 |
| _Bnu | I never had any trouble with reads or writes until it completely locked up. | 16:56 |
| mntmn | what do you mean you have no mmu config | 16:57 |
| mntmn | i'm talking about ARM MMU | 16:57 |
| mntmn | i.e. in ZZ9000OS | 16:57 |
| mntmn | it's about the correct data being in the buffer before sending it off to the usb drive... | 16:57 |
| mntmn | that was not ensured before | 16:58 |
| mntmn | because we wrote it by FPGA->DDR and didn't tell the ARM about it | 16:58 |
| mntmn | so the ARM wrote stale cache contents to disk sometimes | 16:58 |
| mntmn | there was a very similar problem in the past with ethernet. packets containing duplicate blocks | 16:59 |
| mntmn | so JPEGs were corrupted and stuff | 16:59 |
| mntmn | (long time ago) | 17:00 |
| mntmn | i need a replacement for the cpu tool, to enable all those bursts. sysinfo showed me they make a lot of perf on 030 | 17:00 |
| _Bnu | Oh. | 17:00 |
| mntmn | also i want to move all that OS stuff to fastram... but i don't wanna use mutools i think (???) | 17:01 |
| _Bnu | Setting the VRAM to be strongly ordered would probably completely murder performance, though. It should be keeping track of it with ACP, but I guess it doesn't, or something. | 17:02 |
| mntmn | i'm not talking about the VRAM | 17:02 |
| mntmn | only about the parameter area | 17:02 |
| mntmn | i.e. were you poke your x y z values for commands | 17:02 |
| _Bnu | Ah, okay. | 17:02 |
| mntmn | where | 17:02 |
| _Bnu | The only problem I had with that was the command writes were getting lost, never any of the values. | 17:02 |
| mntmn | that's probably some different problem (that i don't experience right now, but we will see) | 17:03 |
| mntmn | does anyone know why setpatch is actually needed? | 17:07 |
| mntmn | so far it doesn't seem to matter at all that it is missing... | 17:07 |
| _Bnu | SetPatch does things like load the updated scsi.device on WB 3.1.4 if you don't have WB 3.1.4 ROMs. | 17:10 |
| _Bnu | It also loads some of the Boing Bag patch stuff in 3.9/3.5. | 17:10 |
| _Bnu | Basically if your Amiga reboots once on booting to Workbench, SetPatch does something. | 17:10 |
| _Bnu | Otherwise there's nothing for it to do. | 17:10 |
| mntmn | hm. | 17:11 |
| _Bnu | But I think they come with their own SetPatch anyway, which may not be compatible with anything else. | 17:11 |
| mntmn | i destroyed my shell... | 17:12 |
| mntmn | i am missing the L directory in ZZombie. | 17:14 |
| mntmn | nobody noticed | 17:14 |
| mntmn | so you don't like L that much i guess... | 17:15 |
| mntmn | nice, RTG still works | 17:15 |
| _Bnu | For file systems mostly, I think. | 17:16 |
| mntmn | forgot how that tool for pushing around windows with the keyboard was called. | 17:17 |
| mntmn | WinZoom. | 17:18 |
| _Bnu | Windows 10 for Amiga. | 17:19 |
| _Bnu | Also, Claude has taught me about the tiny Hitlers that German tourists bring along with them on their travels. The Reisenführer. | 17:20 |
| _Bnu | He is teaching me about German customs so I'll fit in better when I visit and don't get arrested instantly. | 17:21 |
| _Bnu | For instance I didn't know that it was illegal to eat Schinkensalat on Thursdays. | 17:24 |
| mntmn | how offensive!! | 17:29 |
| mntmn | i am sorry to say that my amiga system is stable now | 17:30 |
| _Bnu | Poke around in it a bit with a screwdriver. | 17:36 |
| mntmn | haha | 17:38 |
| mntmn | 1280x720 is smol... https://pbs.twimg.com/media/E8HkIrvWQAMJeUs?format=png&name=4096x4096 | 17:40 |
| mntmn | enhancing to 1920x1080 | 17:43 |
| _Bnu | Yeah, I remember Nibbles/QB64 being even smoler, and I couldn't figure out how to scale it up, haha. | 17:43 |
| _Bnu | I'm sure there's some keyboard shortcut I don't know... | 17:43 |
| mntmn | cmd+F | 17:44 |
| mntmn | but in this case, i want 1:1 pixelz | 17:44 |
| mntmn | whatever happened to Scalos and where did that come from in the first place? | 17:47 |
| mntmn | lulz it is still around https://en.wikipedia.org/wiki/Scalos | 17:47 |
| mntmn | > Satanic Dreams Software Team | 17:48 |
| mntmn | nice, open source under BSD | 17:48 |
| _Bnu | Cmd+Hyper | 17:48 |
| _Bnu | I seem to recall there being some problem with Scalos on 68k, though. | 17:50 |
| mntmn | amirc is the first thing that complains about missing setpatch. during install | 17:52 |
| mntmn | crazy how accetable that website looks https://twitter.com/mntmn/status/1423673811237543949?s=20 | 17:54 |
| _Bnu | Why would an IRC client complain about SetPatch. ('A`) | 18:13 |
| _Bnu | That's weird though, are random parts of Scalos not open source, or what? | 18:14 |
| + mntamiga (~root@p54a43f94.dip0.t-ipconnect.de) | 18:19 | |
| mntamiga | ha! | 18:19 |
| mntamiga | extremely online | 18:19 |
| mntamiga | text entry into mui text fields is very slow for some reason | 18:19 |
| _Bnu | Why not, I say... | 18:22 |
| mntamiga | :3 | 18:23 |
| _Bnu | And you're not running the firmware with the millions of debug messages on rect_pattern? | 18:23 |
| mntamiga | also i needed to setup socket.library for amirc for some reason. | 18:23 |
| mntamiga | no, text output is fine | 18:23 |
| mntamiga | it really only MUI text fields. | 18:23 |
| _Bnu | Ah, weird. I wonder what it uses to draw text. | 18:23 |
| mntamiga | text entry i mean | 18:24 |
| _Bnu | Maybe bobs, drawing them 600 times per second when you type text. | 18:24 |
| mntamiga | the log window with all your chat messages is fine | 18:24 |
| mntamiga | just typing is super delayed | 18:24 |
| mntamiga | and only in MUI | 18:24 |
| mntamiga | how can i check which p96 version this is... | 18:25 |
| _Bnu | You never can... | 18:27 |
| mntamiga | damn | 18:27 |
| _Bnu | If it's the one from Aminet, it's probably 2.0. | 18:27 |
| mntamiga | i don't remember | 18:28 |
| Jope | amigaaa | 18:28 |
| _Bnu | Because if it was 1.4, it would probably crash, lol. | 18:28 |
| mntamiga | i will patch in the latest beta from thor | 18:28 |
| _Bnu | Yeah, that's what I'm running since I don't feel like paying for P96 again right now... | 18:28 |
| mntamiga | rtg.library 42.681 | 18:29 |
| _Bnu | Because they of course released 3.10 just after my one year license ran out, lolb. | 18:29 |
| _Bnu | "Is _Bnu's P96 license about to run out?" "Yes." "Okay! Release a new version!" | 18:29 |
| - mntamiga (QUIT: Quit: System rebooting) (~root@p54a43f94.dip0.t-ipconnect.de) | 18:31 | |
| _Bnu | System is rebooting. | 18:31 |
| _Bnu | Download AmiTetris for entertainment. | 18:32 |
| Jope | all kinds of tetraminoes | 18:34 |
| Jope | all finnish sceners played this http://www.nic.funet.fi/pub/amiga/games/tetris/Zetris-1.0.lha | 18:35 |
| RobDangerous | Same thing happened to me and I sent Jens a mail because I was confused and unaware that the license even runs out at all and he just offered me to sent me the latest version because it released only shortly after my license expired. | 18:35 |
| _Bnu | The first Tetris I ever played was the Game Boy one, so I really disliked all the computer variants, haha. | 18:36 |
| RobDangerous | But I said noooo, take some money, I want you to have food on the table. | 18:36 |
| Jope | _Bnu, the gameboy one was a good one tho | 18:36 |
| _Bnu | Yes, but I really dislike the computer ones. | 18:37 |
| Jope | :-) | 18:37 |
| RobDangerous | And that's why Jens is now rich enough to buy a Game Boy Tetris cartridge on eBay. | 18:37 |
| _Bnu | Because they all had (to me) really fucked up controls in comparison. | 18:37 |
| forcie | is P96 a one year license now? last time i ordered it was valid for like 10 downloads | 18:37 |
| _Bnu | 10 downloads/one year. | 18:37 |
| _Bnu | Whichever comes first. | 18:37 |
| forcie | i have only downloaded it once and i still have the archive here, lolz | 18:38 |
| forcie | better wait for some major updates before i use up another precious download. | 18:38 |
| _Bnu | Well, 3.10 is the latest major update as far as I know. | 18:39 |
| _Bnu | And it's out now. | 18:39 |
| _Bnu | Has the secondary palette thing on screen split, which AFAIK you can only use in WinUAE and on the ZZ9000? | 18:40 |
| _Bnu | Also fixes some other things and things. | 18:42 |
| _Bnu | * HIRES mouse pointers are now again available on Os 3.2 due to a new interface between RTG graphics and the intuition pointer class. | 18:44 |
| _Bnu | Little do they know that the ZZ9000 supports up to 48x48 256 color mouse pointers with their very own palette... | 18:45 |
| mntmn | haha | 18:53 |
| mntmn | this amiga stuff is making me mad... | 19:00 |
| mntmn | the mount command from aros is subtly different so it cannot mount the DI2 file from diskimagecontrol. | 19:01 |
| mntmn | give me UnADF or give me death | 19:01 |
| mntmn | i unadfd workbench2.1 extras to run Font prefs just once... | 19:09 |
| mntmn | and pressed save to generate the missing font.preffs | 19:09 |
| mntmn | font.prefs | 19:09 |
| mntmn | that fixes all the shitness of topaz 9 | 19:09 |
| _Bnu | I always just use GoADF for ADF stuff... but that costs monies. | 19:17 |
| _Bnu | Waiting for fat commands. | 19:19 |
| _Bnu | Maybe we can make our own Workbench and call all the settings Prefferences. | 19:20 |
| _Bnu | Pfefferences. | 19:21 |
| _Bnu | Wokbench. | 19:23 |
| _Bnu | Intuton.libary | 19:25 |
| RobDangerous | New boot.bin messed up the rtg output a little: https://imgbox.com/W3LAeIcg | 19:26 |
| _Bnu | Uhh, did you make sure you install the .card file first? | 19:26 |
| _Bnu | And/or what display mode is that? | 19:26 |
| RobDangerous | Ah. Of course not. | 19:26 |
| RobDangerous | 16 bit 1920x1080 | 19:27 |
| _Bnu | Ah yeah, while the driver with panning supports is backward compatible with the old boot.bin, the new boot.bin is not backward compatible with the drivers that don't have panning support. | 19:28 |
| _Bnu | Which is why the card driver should always be removed or replaced before switching to a new firmware. | 19:28 |
| RobDangerous | Okidoki, I can fix it! | 19:28 |
| + mntamiga (~root@p54a43f94.dip0.t-ipconnect.de) | 19:29 | |
| _Bnu | You'll probably also have to reattach the P96 settings to the board using Picasso96Mode. | 19:33 |
| mntmn | _Bnu: what do i need to use 68060.library? is it possible with stock 3.1 rom? | 19:33 |
| _Bnu | I had to change the chip/board type back to the old one because otherwise it completely broke Picasso96 2.0 support. | 19:33 |
| mntmn | i mean what do i need to use 68060 | 19:33 |
| _Bnu | You just need 68060.library, for emulation of missing opcodes, yeah. | 19:34 |
| - mntamiga (QUIT: Ping timeout: 272 seconds) (~root@p54a43f94.dip0.t-ipconnect.de) | 19:35 | |
| _Bnu | mntamiga died. ('A`) | 19:35 |
| mntmn | ok cool will try | 19:35 |
| _Bnu | At least that's all I ever do to get my CS MkII to work... copy 68060.library to LIBS:, haha. | 19:36 |
| mntmn | who loads 68060.library? | 19:43 |
| mntmn | maybe setpatch? ;) | 19:43 |
| _Bnu | Not that I know of... | 19:45 |
| mntmn | the warpengine does not work. at all. | 19:50 |
| mntmn | with these 3.1 stock roms | 19:51 |
| mntmn | perhaps | 19:51 |
| _Bnu | Does it like not boot at all? | 19:51 |
| _Bnu | It should definitely boot and guru. | 19:51 |
| mntmn | nope just black screen | 19:52 |
| mntmn | (zz9000 not inserted) | 19:52 |
| mntmn | sad story | 19:53 |
| _Bnu | ('A`) | 19:53 |
| RobDangerous | Ah, reattach settings to the board, I wonder what I have to click for that. But kicking out all ZZ9000 modes also worked for fixing Workbench. Aaaand... https://imgbox.com/skv0hkYz | 19:54 |
| _Bnu | Kickstarts should definitely work, it's just Workbench itself that can't boot. | 19:54 |
| RobDangerous | The stuttering is gone. That's why more important to me than RTG modes anyway :-) And even my TV agrees... it's FIXED. | 19:54 |
| RobDangerous | *way more important | 19:54 |
| _Bnu | Hahah, nice | 19:54 |
| _Bnu | Yeah, it will still repeat a single frame every few minutes or whatever 0.001.5Hz adds up to. | 19:55 |
| mntmn | ok kickstart doesn't work. so something is wrong with the warpengine probably | 19:55 |
| _Bnu | But to me at least it's not noticeable at all compared to when it goes at 50+ Hz. | 19:56 |
| RobDangerous | Fixing that also would of course be perfect. mntmn, will you add perfect framerate sync if I write a Warp3D driver for the ZZ9000 in return? | 19:56 |
| RobDangerous | But the current fix is already super great and changes things from unplayable to perfectly fine for me personally. | 19:56 |
| mntmn | ?? | 19:57 |
| RobDangerous | !! | 19:57 |
| mntmn | i will not | 19:57 |
| mntmn | _Bnu's fix must suffice | 19:57 |
| mntmn | also i'm unhappy because no 060 for me | 19:57 |
| RobDangerous | Ah, pity. Maybe I'll try my luck with some Warp3D anyway. Already did it a few centuries ago for WinUAE. | 19:57 |
| mntmn | i will try the csmk1 that i should send back really soon now | 19:59 |
| _Bnu | I'm sure "sync to sync" will be a consideration in the future technology for the theoretical MNTCX device. | 20:00 |
| RobDangerous | If we first manage to make Mr mntmn happy again. | 20:01 |
| mntmn | how do i need to set the 3 clock jumpers in a3000 for the cs mk1 | 20:01 |
| mntmn | the manual only talks about the a4000 | 20:01 |
| _Bnu | I don't know... I think my clock jumper is like under the CPU board for some stupid reason. | 20:01 |
| _Bnu | So I can't see what it's set to. | 20:01 |
| mntmn | there are 3 clock jumpers | 20:01 |
| mntmn | on the a4000 only one i think | 20:02 |
| _Bnu | Ah, hm | 20:02 |
| mntmn | i have brdclock, cpuclock and quadclock | 20:02 |
| _Bnu | Claude might know, if he's around. | 20:02 |
| mntmn | oh! | 20:02 |
| mntmn | something is working. | 20:02 |
| mntmn | quad+brd to ext, cpu to int. | 20:02 |
| _Bnu | !!! | 20:03 |
| _Bnu | It wouldn't surprise me if there's some jumper in an Amiga that just has !!! silkscreened next to it. | 20:05 |
| mntmn | haha | 20:05 |
| mntmn | autoboot did not crash... | 20:05 |
| mntmn | setcpu immediately crashes | 20:05 |
| mntmn | highly unstable | 20:06 |
| mntmn | i guess my power supply is too weak for 68060s | 20:07 |
| mntmn | ok dinner time | 20:07 |
| _Bnu | RobDangerous: To reattach the modes to the board, you usually just open Picasso96Mode and go to the right mouse button menu and select Attach settings to board and then choosing the ZZ9000 menu item. | 20:09 |
| RobDangerous | Thanks, I think I can manage to do that. | 20:11 |
| _Bnu | https://cdn.discordapp.com/attachments/870965415000539146/873267219767902278/unknown.png | 20:12 |
| _Bnu | I am suddenly reminded that my ZZ9000 modes are wrong, and I'm missing 1280x720. | 20:12 |
| mntmn | and you have secret modez | 20:14 |
| mntmn | that should go into the release as well | 20:14 |
| _Bnu | Technically the only secret mode I have is 1600x1200, haha. | 20:14 |
| _Bnu | The other ones don't work, they're from the CV64, I believe. | 20:14 |
| _Bnu | Oh, I do have the 30Hz 2560x1440 as well. | 20:14 |
| Jope | 1600x1200 <3 | 20:20 |
| Jope | a fine mode if there ever was one | 20:20 |
| Jope | the a3k's clock jumpers definitely deserve a webpage of their own.. correct settings for each board | 20:21 |
| Jope | super ambiguous | 20:21 |
| RobDangerous | Totally different topic - I'm having a look at the zz9000 sdk. Apparently ARM programs can access the framebuffer which is brilliant. But is there also a way to communicate with Amiga programs? Somehow via those event-something functions maybe? | 20:23 |
| _Bnu | The SDK is uhh... kind of in a state of functioning, maybe not functioning. | 20:23 |
| RobDangerous | Hmm. Is there interest in pimping that? I'd like to make Quake run faster. | 20:24 |
| _Bnu | You'd be trying to run Quake bare metal on the ARM in that case, haha. | 20:25 |
| RobDangerous | I'd like to route Warp3D calls to the ARM. | 20:25 |
| _Bnu | There's already a graphics acceleration SDK of sorts. | 20:25 |
| _Bnu | It can be most readily observed in action in the ZZ9000 accelerated Frontier Intro. | 20:26 |
| _Bnu | And of course the ZZ9000-accelerated build of ScummVM... | 20:26 |
| RobDangerous | Oh, googling... | 20:26 |
| mntmn | yeah if you need textured triangles or something the best way would be to clean up and get this into the main graphics functions | 20:26 |
| _Bnu | They're both on github on my fork, builds arein snakes9000, haha. | 20:26 |
| _Bnu | By pure chance I was also already running ScumMVM and watching the Full Throttle intro again... | 20:27 |
| _Bnu | https://cdn.discordapp.com/attachments/870965415000539146/873271210270654484/2021-08-06_20-25-51.mp4 | 20:28 |
| RobDangerous | Seems familiar. | 20:28 |
| _Bnu | The standard build of ScummVM runs like absolute balls on anything worse than a Vampire, haha. | 20:29 |
| _Bnu | Honestly, really looking forward to when I can abuse the ZZ9000 audio module for the AHI stuff. | 20:30 |
| _Bnu | No more throwing 68k CPU time at audio mixing or bad resampling. | 20:30 |
| _Bnu | You can almost play Curse of Monkey Island in ZZ9000 ScummVM, but the music is screwing it up big time. | 20:31 |
| _Bnu | The ScummVM build is in buld.zip, by the way. | 20:31 |
| _Bnu | And the Frontier into is fintro-gcc. | 20:31 |
| RobDangerous | Aha, found some code by some beeanyew. | 20:33 |
| RobDangerous | ACC_OP_FILL_CIRCLE - ok, so more of those is the idea. | 20:33 |
| _Bnu | Yes. There is technically already a textured polygon thing, but only the flat polygons are hooked up. | 20:34 |
| _Bnu | And the flat polygons need a tiny bug fix. | 20:34 |
| _Bnu | Which, I have the github link from mntmn open somewhere in my Firefox tab list... | 20:34 |
| mntmn | ah i remember wot you mean | 20:34 |
| mntmn | why and where did i fix it again... | 20:34 |
| mntmn | ah yeah | 20:34 |
| mntmn | maze | 20:35 |
| mntmn | https://github.com/mntmn/maze/blob/main/draw.c#L56 | 20:35 |
| RobDangerous | What are non-flat polygons? | 20:35 |
| mntmn | ok, i will flash some keyboards and then wrap up all my changes so far, i think usb stuff is reasonably stable now | 20:36 |
| mntmn | but couldn't really test with 060, i only saw that the partitions where in early startup. | 20:37 |
| pasik | mntmn: about 68060.library .. isn't it the updated 68040.library that loads it? | 20:38 |
| pasik | if 060 is detected | 20:38 |
| mntmn | oh? | 20:38 |
| pasik | which is why you need to install both the 68040.library and 68060.library | 20:38 |
| mntmn | but who loads 68040.library then? | 20:38 |
| _Bnu | Non-flat polygons would be polygons that have some sort of shading between colors and stuff. | 20:39 |
| _Bnu | While flat ones are just a single color. | 20:39 |
| RobDangerous | Ah! | 20:39 |
| RobDangerous | Flat shaded polygons. | 20:39 |
| _Bnu | Yeah, it was enough for the Frontier intro, basically. | 20:40 |
| _Bnu | Since it's not based on the PC version, which has textures. | 20:40 |
| _Bnu | The main reason the Frontier intro ever drops below 60 FPS is also because it does its own polygon clipping in software on the 68k. | 20:41 |
| _Bnu | If it could be told to just send the full polygons to the ZZ9000, it'd be a lot faster. | 20:41 |
| forcie | mntmn: exec.library loads 68040.library | 20:42 |
| forcie | if i remember correctly | 20:42 |
| pasik | or setpatch loads 68040.library ? | 20:42 |
| pasik | i'm not totally sure how it works | 20:42 |
| forcie | setpatch probably reloads it, yeah | 20:42 |
| pasik | jope maybe can describe the magic | 20:42 |
| _Bnu | I'm pretty sure you don't need 68040.library with Kick 3.1. | 20:43 |
| _Bnu | Because I do not have a "dummy" 68040 library installed. | 20:44 |
| RobDangerous | But if it doesn't do the clipping it ignores the frontiers and then its name wouldn't fit anymore. | 20:45 |
| pasik | "older systems (such as 3.1) do not recognise the 68060 - instead they mistake it for a 68040 and so load 68040.library for it. When you actually use a 68060, you need the 68040.library as well as this acts as a wrapper, pointing the system at the correct 68060.library instead." | 20:47 |
| pasik | that's what i've done with my blizzard 1260 with amigaos 3.1; install the dummy wrapper 68040.library and the actual 68060.library | 20:48 |
| pasik | newer amigaos versions probably can load the 68060.library directly | 20:51 |
| RobDangerous | ok, I think I kind of understand the process now. Add stuff to main.c and gfx.c in ZZ9000OS. Compile a new boot.bin. Call new stuff like the Frontier intro does. Play Quake with broken graphics. Repeat until graphics are not broken. | 20:51 |
| mntmn | yes | 20:52 |
| mntmn | but at what point is it loaded? what triggers it? | 20:53 |
| mntmn | (680x0.library) | 20:53 |
| pasik | I bet jope knows the answer.. | 20:54 |
| _Bnu | So maybe the CS MkII boot ROM loads 68060.library for me? | 20:59 |
| _Bnu | That would be nice of it. | 20:59 |
| _Bnu | RobDangerous: Kind of. At least building a new boot.bin only takes a few seconds. | 21:00 |
| RobDangerous | All right. I'll play around with that a bit. There will be more questions. | 21:01 |
| _Bnu | There are a lot of applications and games that could be accelerated using it, problem is that many of them are close to impossible to build for me. | 21:04 |
| _Bnu | Either because they need Docker (I can't use Docker, because it requires Hyper-V to be enabled on Windows for any kind of performance) or because they only build with some combination of ancent GCC and some assembler that I can't set up the environment for to save my life. | 21:05 |
| RobDangerous | Well, just accelerate Warp3D and that automatically accelerates... like... five games or something. | 21:07 |
| _Bnu | Yes, I also considered SDL, but that's also magically impossible to build for AmigaOS 3. | 21:08 |
| RobDangerous | And it slows down Payback because that's the game with the most draw-calls I've ever seen. | 21:08 |
| _Bnu | But yeah if you had the source code for a game with clearly defined drawing calls and stuff, you could just add optional ZZ9000 acceleration calls like I did with ScummVM. | 21:09 |
| mntmn | ok i will now check in some code | 21:09 |
| _Bnu | And speed it up infinitely, because Zorro is in no way suitable for use with transferring frame buffers repeatedly. | 21:09 |
| _Bnu | Damn, not code... | 21:10 |
| RobDangerous | Nice nice. | 21:10 |
| - TadeusTaD (QUIT: Quit: Stay Cheeki Breeki) (tadeustad@psifactor.pl) | 21:11 | |
| _Bnu | I also added some tiny similar stuff to DoomAttack for the PiStorm, like I did for the ZZ9000, so now it actually runs acceptable on the emulated 68030, haha. | 21:11 |
| RobDangerous | Ah, Doom, I heard about that. | 21:16 |
| _Bnu | ACTION refreshes the zz9000 drivers and firmware repos all day | 21:20 |
| mntmn | _Bnu: i merged and pushed to firmware | 21:42 |
| mntmn | _Bnu: i will now update driverz | 21:42 |
| _Bnu | ACTION downloads the commits over and over | 21:44 |
| _Bnu | DDoSing git gitlab... | 21:47 |
| _Bnu | And then gitlab was angry because I logged in from the same IP as last time, but the IP geolocation had changed, lol. | 21:48 |
| mntmn | _Bnu: pushed to gcc-test branch | 21:52 |
| mntmn | i *think* i pushed all the things | 21:52 |
| mntmn | i should maybe do a quick test of the complete package... | 21:53 |
| mntmn | yeah i have to push the firmware again | 21:53 |
| mntmn | because of rebuilt zzusb.device | 21:53 |
| - RobDangerous (QUIT: Ping timeout: 256 seconds) (~Thunderbi@p200300ec8f3e1c002d0ccd643d9ce569.dip0.t-ipconnect.de) | 21:55 | |
| mntmn | pushed | 21:56 |
| + RobDangerous (~Thunderbi@p200300ec8f3e1c002d0ccd643d9ce569.dip0.t-ipconnect.de) | 21:58 | |
| _Bnu | Push, push, push, pull... | 22:00 |
| _Bnu | It's just like air cooling. | 22:00 |
| _Bnu | And yeah, sorry, I can't properly set the flags for bash scripts on Windows, haha. | 22:00 |
| mntmn | haha no problem | 22:03 |
| _Bnu | And MSYS2 of course completely ignores it and just runs the bash script even if it isn't executable... | 22:03 |
| mntmn | but i revenged and put this copy command in the zzusb.device build script that only works on my system... | 22:03 |
| _Bnu | I will port it to M1 Mac... | 22:03 |
| _Bnu | Or maybe G4 Mac Mini. | 22:04 |
| mntmn | i will push one more time soon... | 22:06 |
| mntmn | because i made a bug in the zzusb.device debughex/str functions | 22:07 |
| mntmn | btw don't forget to remove your old ZZUSBStorage.device if you have it on the amiga | 22:07 |
| _Bnu | But I love my old ZZ9000USBStorage.device... | 22:08 |
| mntmn | it has to go! | 22:08 |
| _Bnu | Maybe I can rename it to 20mips.device | 22:08 |
| _Bnu | Why do I have a file called "MEPPO" in DEVS: | 22:09 |
| mntmn | :O | 22:09 |
| mntmn | i pushed again to both repos | 22:09 |
| mntmn | and testing now. | 22:09 |
| _Bnu | Where is my stupid USB stick... | 22:11 |
| _Bnu | Oh, on the floor. Of course. | 22:12 |
| mntmn | do i need to set up special things for the split palette to work? | 22:13 |
| _Bnu | No, just the driver with the flag set. | 22:14 |
| _Bnu | And the latest P96 beta thing you sent to me. | 22:14 |
| _Bnu | Well, the latest driver with the flag set and the set_feature write... | 22:15 |
| _Bnu | Otherwise it won't be enabled, for backwards compatibility reasons. | 22:15 |
| _Bnu | Sorry about that, many RTG updat today. | 22:17 |
| _Bnu | Well, just two... | 22:17 |
| _Bnu | But many compared to zero... | 22:17 |
| mntmn | hmm it's like split screen is not there at all for me | 22:17 |
| mntmn | like, i drag on a screen and stuff freezes for a moment but there is always one screen shown, no split | 22:17 |
| _Bnu | ? ('A`) | 22:18 |
| _Bnu | It worked earlier at least... | 22:18 |
| mntmn | maybe it's just borked for me | 22:18 |
| mntmn | i couldn't test it so far at all on this new setup | 22:18 |
| mntmn | hard to say if i have the correct rtg.library | 22:18 |
| mntmn | apparently it came in p96.zip... | 22:19 |
| mntmn | hmm but the mouse behaves as if the screen was moved | 22:20 |
| mntmn | i will debug the setting of the split | 22:22 |
| _Bnu | I have a feeling your boot ROM might be doing something naughty to the list of DosNodes or something... | 22:22 |
| mntmn | i hope not? | 22:24 |
| mntmn | it should just AddBootNode | 22:24 |
| mntmn | that happens in zzusb.device | 22:24 |
| _Bnu | Hm. | 22:24 |
| _Bnu | Both the palette and the split works fine here. | 22:24 |
| mntmn | oh! | 22:24 |
| mntmn | what did i mess up | 22:24 |
| _Bnu | Maybe you forgot the other P96 zip with the actual installer? | 22:25 |
| _Bnu | There was some zip file later that just had the header file and rtg.library. | 22:25 |
| _Bnu | But the original one came in an LHA file. | 22:25 |
| mntmn | is p96 more than the rtg.library? | 22:26 |
| _Bnu | I think so... | 22:26 |
| mntmn | ok it's maybe borked on the amiga side, split position is never written | 22:26 |
| mntmn | what else is p96? | 22:26 |
| mntmn | afaik it is only rtg.library | 22:26 |
| _Bnu | Well, maybe. Not sure. | 22:27 |
| _Bnu | But the only reason the split position wouldn't be written is either because: | 22:27 |
| _Bnu | A) You're using an old .card file. | 22:27 |
| _Bnu | B) You're using an old P96 that doesn't actually call that function. | 22:27 |
| mntmn | yeah. | 22:27 |
| mntmn | i agree | 22:27 |
| mntmn | or C) my system is fuxxed beyond recognition | 22:28 |
| _Bnu | https://cdn.discordapp.com/attachments/870965415000539146/873301439974367282/unknown.png | 22:28 |
| _Bnu | Not proper resolution, but the secondary palette is definitely kicking in. | 22:28 |
| _Bnu | Else PPaint would have borked colors. | 22:28 |
| mntmn | nice!! | 22:28 |
| mntmn | extremely odd. | 22:29 |
| _Bnu | But yeah, with your changes compiled into the firmware I don't actually get any DosNodes from the FastATA 4000, haha. | 22:30 |
| mntmn | this p96 is def. trying to handle the split but it does not write the split. the front/back screen changes if i drag the screen all the way up or not | 22:30 |
| _Bnu | Does your boot ROM get angry if there's no Amiga partition on the USB stick? | 22:30 |
| mntmn | _Bnu: maybe only one autoboot thing is used? | 22:30 |
| _Bnu | Not typically, no. | 22:30 |
| mntmn | hm | 22:30 |
| _Bnu | I can have both IDE and FastATA, and several boot ROMs is nothing too uncommon. | 22:30 |
| mntmn | ok | 22:31 |
| mntmn | you can turn off the bootrom stuff with 1 flag in the top of mntzorro.v | 22:31 |
| _Bnu | I refuse... disabling features is for noobs... | 22:31 |
| mntmn | ok | 22:31 |
| _Bnu | Let's see here. | 22:32 |
| mntmn | my rtg.library hash is 932af260380d48907143073de861a8cf. both on the amiga and in that weird zip i got. | 22:32 |
| mntmn | -rw------- 1 mntmn mntmn 123336 Aug 5 22:25 p96.zip | 22:32 |
| _Bnu | I can't actually tell if you're returning 1 at the end of DiagEntry, because I can't follow the code... sorry, I'm really, really bad at reading 68k assembler, haha. | 22:32 |
| _Bnu | But if you don't return 1 at the end of DiagEntry, Kickstart will abort everything from that point. | 22:33 |
| mntmn | ah! | 22:33 |
| mntmn | the 1 should go into d0, and i think i just leave the result from whatever the last call was in there | 22:33 |
| mntmn | jsr -102(a6) // InitResident | 22:34 |
| mntmn | // return code in d0 | 22:34 |
| mntmn | rts | 22:34 |
| mntmn | so i leave the result of InitResident in there | 22:34 |
| mntmn | object - Return value from the init code, usually the library | 22:34 |
| mntmn | or device base. NULL for failure. | 22:34 |
| mntmn | ah so that is the return value from zzusb.device's init function then | 22:35 |
| mntmn | that returns the configdev for some reason... | 22:35 |
| mntmn | _Bnu: i can put a return of 1 in the rom | 22:36 |
| _Bnu | Yeah, I remember having to add it to the PiSCSI boot ROM for it to work properly when there were other PICs in line after it. | 22:37 |
| mntmn | _Bnu: pushed new bootrom.h to firmware. you need to refresh that file in the sdk so eclipse picks it up. | 22:37 |
| _Bnu | bootorm.h... | 22:38 |
| _Bnu | Orm is Swedish for snake... | 22:38 |
| mntmn | BootStorm | 22:38 |
| mntmn | :D | 22:38 |
| mntmn | so it is PistOrm | 22:38 |
| mntmn | not PiStorm | 22:38 |
| - RobDangerous (QUIT: Ping timeout: 268 seconds) (~Thunderbi@p200300ec8f3e1c002d0ccd643d9ce569.dip0.t-ipconnect.de) | 22:42 | |
| _Bnu | Well, it's definitely doing something else now... I wonder where the butts I put my UART kabul. | 22:43 |
| _Bnu | I "cleaned up" the stuff on the table recently and probably put it somewhere nearby. | 22:43 |
| pasik | to an obvious place where you find it easily and quickly.. | 22:43 |
| _Bnu | Yes. | 22:45 |
| _Bnu | Next to the 5.25" floppy box, of course! | 22:47 |
| mntmn | ok so my rtg.library or monitor file is broked | 22:47 |
| mntmn | the driver never pokes the split pos | 22:47 |
| mntmn | but i can poke it by hand | 22:47 |
| mntmn | if i poke to 0x5e | 22:47 |
| mntmn | _Bnu: you built ZZ9000.card from the same code that i have yes? | 22:48 |
| _Bnu | Yeah. | 22:49 |
| mntmn | ach scheisse | 22:49 |
| _Bnu | Did you accidentally run build.sh... | 22:49 |
| mntmn | i made useless changes in the old mntgfx.c but this should be whatever | 22:51 |
| mntmn | because i built the other one... | 22:51 |
| _Bnu | RDB not found...! Which isn't unexpected, but then it just sits there... :D | 22:52 |
| mntmn | haha | 22:52 |
| mntmn | oops | 22:52 |
| mntmn | it should not sit, just abort... | 22:53 |
| _Bnu | Yeah, no more output after that. | 22:53 |
| _Bnu | Not sure if there should be. | 22:53 |
| mntmn | no, no more output required | 22:53 |
| mntmn | computer hangs then? | 22:54 |
| _Bnu | Yeah. | 22:54 |
| _Bnu | Without the return 1, it instantly skipped everything and dumped me to the Kickstart screen with only DF0. | 22:54 |
| mntmn | did you try a reset? | 22:55 |
| _Bnu | Yeah, a bunch, gets stuck in the same place. | 22:59 |
| mntmn | btw little bug: split_position is not reset with handle_amiga_reset. | 22:59 |
| _Bnu | Ah yeah, I forgot about that. | 23:00 |
| mntmn | i just put it in | 23:01 |
| _Bnu | Also, I see you're only saving a few registers, not sure if that matters though. | 23:01 |
| mntmn | hmhm | 23:01 |
| mntmn | could matter | 23:02 |
| mntmn | can the zzusb.device work if 68060.library is not loaded? maybe it has 68060-incompatible code? | 23:02 |
| mntmn | hmm but i had partitions added briefly on 68060. | 23:02 |
| mntmn | so it seemed to work | 23:02 |
| _Bnu | Actually, your DiagEntry is not the same as mine, so maybe it shouldn't return 1. I'm really not sure. | 23:04 |
| mntmn | ;_; | 23:04 |
| _Bnu | Like, I don't do the InitResident in DiagEntry at all, DiagEntry is just the reloc in the PiSCSI boot ROM. | 23:05 |
| mntmn | where do you do the initresident then? | 23:05 |
| _Bnu | In InitEntry. | 23:05 |
| mntmn | what's that? | 23:05 |
| mntmn | maybe something i don't have | 23:05 |
| _Bnu | To quote the example boot ROM . . . | 23:05 |
| mntmn | i have only DiagEntry and BootEntry | 23:05 |
| mntmn | or, DiagPt and BootPt | 23:06 |
| _Bnu | "Beginning of ROM driver code and data that is accessed only in the ROM space. This must all be position-independent." | 23:06 |
| _Bnu | But yeah, really sorry about that. I honestly don't know how the boot ROMs work outside of modifying the example boot ROM to run the code I want to run... | 23:07 |
| _Bnu | https://github.com/beeanyew/pistorm/blob/wip-crap/platforms/amiga/piscsi/device_driver_amiga/bootrom.s#L251 | 23:08 |
| mntmn | ok, i will look at this again on sunday or so | 23:09 |
| mntmn | wtf just happened... | 23:09 |
| _Bnu | But on Sunday we have to go to church... | 23:09 |
| mntmn | ah i know wtf just happened | 23:09 |
| _Bnu | It's 11:11 PM... | 23:11 |
| _Bnu | ACTION beeps and begins to explode | 23:12 |
| mntmn | some kind of insane shit is going on... | 23:12 |
| mntmn | ok so the driver goes into SetSplitPosition | 23:14 |
| mntmn | i put a debug thing in there, like *((uint16_t*)0x400000f2) = pos; | 23:14 |
| _Bnu | I'll disable the small autoboot for now and re-enable it in church on Sunday or whenever you get a chance to take a look at it. | 23:14 |
| mntmn | and i get positions over uart! | 23:14 |
| mntmn | _Bnu: yep | 23:14 |
| mntmn | but... the driver does not write to register 0x5e | 23:15 |
| _Bnu | Are you using the non-DMA driver or something? It's entirely possible that I haven't tested that. | 23:15 |
| mntmn | yes, non-DMA | 23:15 |
| _Bnu | Yeah in that case the non-DMA code is probably the problem. | 23:16 |
| mntmn | ;_; | 23:16 |
| mntmn | but! | 23:16 |
| mntmn | the code says: | 23:16 |
| mntmn | zzwrite32(®isters->blitter_src_hi, offset); | 23:16 |
| mntmn | zzwrite16(®isters->blitter_set_split_pos, pos); | 23:16 |
| mntmn | and i think blitter_set_split_pos should be 0x5e | 23:16 |
| _Bnu | Yeah, looks like it based on the list. | 23:17 |
| mntmn | i will replace it with zzwrite16(0x4000005e, pos); for the lulz | 23:17 |
| _Bnu | Oh no, I know what it is, lol. | 23:17 |
| mntmn | tell! | 23:18 |
| _Bnu | MNTZZ9KRegs* registers = (MNTZZ9KRegs*)b->RegisterBase; | 23:18 |
| _Bnu | Add that after #else | 23:18 |
| mntmn | ok? :D | 23:18 |
| _Bnu | Yeah, it'll work, haha. | 23:18 |
| _Bnu | I did a fail, you see. | 23:19 |
| _Bnu | There's a global "registers" that's only supposed to be there if DMARTG is defined. | 23:19 |
| mntmn | oop | 23:19 |
| mntmn | ok | 23:19 |
| _Bnu | But I forgot to delete it, so it's always there. | 23:19 |
| _Bnu | I'll push a fax. | 23:19 |
| _Bnu | Fax poshed. | 23:20 |
| mntmn | :O | 23:20 |
| mntmn | yes, it works, goddamit!!! | 23:21 |
| mntmn | this split stuff is very slow on my computer though. | 23:22 |
| mntmn | like, dragging on the screen bar always makes the computer think very hard | 23:22 |
| mntmn | and it pauses for like half a second | 23:22 |
| mntmn | as if it is copying big buffers back and forth in the background... | 23:23 |
| mntmn | maybe screen dragging can only be fast on 68060 | 23:23 |
| mntmn | the mouse pointer offset of the front screen is also off | 23:23 |
| mntmn | the back screen's is correct | 23:24 |
| mntmn | ah, it's not always off. | 23:24 |
| _Bnu | Yeah, it depends on the resolution. | 23:25 |
| mntmn | which part? the speed? | 23:25 |
| mntmn | or the offness? | 23:25 |
| _Bnu | That too. | 23:25 |
| _Bnu | But the mouse pointer as well. | 23:25 |
| _Bnu | At 320x240 it's spot on, at 640x480 it's one pixel off. Possibly because 640x480 is 59.4Hz or something. | 23:26 |
| _Bnu | But I haven't messed with the refresh rates for the other modes at all. | 23:26 |
| mntmn | ah no for me its like up to 8 or more pixels off | 23:26 |
| mntmn | and in both directions | 23:26 |
| _Bnu | Uhhh | 23:26 |
| mntmn | also it is fantastically slow | 23:26 |
| mntmn | also in low resolutions | 23:26 |
| _Bnu | That's very strange, it definitely shouldn't be. | 23:27 |
| mntmn | but it might be my illegal workbench setup | 23:27 |
| mntmn | like, fastiprefs and so on | 23:27 |
| _Bnu | And it's not puking out debug info or something? | 23:27 |
| mntmn | amiga? | 23:27 |
| mntmn | ah, i don't have an amiga serial debug cable | 23:27 |
| mntmn | at the moment | 23:27 |
| mntmn | attached | 23:27 |
| _Bnu | If the Amiga does, it would be very slow. | 23:27 |
| mntmn | naja it is extremely slow | 23:27 |
| _Bnu | It would probably also be slow if the ZZ9000 did it. | 23:27 |
| mntmn | no, i mean it is REALLY slow | 23:27 |
| mntmn | like 1fps slow | 23:28 |
| _Bnu | Yeah, it might be Amiga serial debug output. | 23:28 |
| _Bnu | Which defaults to 9600 bps or something. | 23:28 |
| mntmn | so maybe thor's version just has a lot of debug stuff in there? | 23:28 |
| _Bnu | I'm using Thor's version, though... | 23:28 |
| _Bnu | So most likely something else. | 23:28 |
| mntmn | >:|||| | 23:28 |
| _Bnu | Maybe an AROS library? | 23:28 |
| mntmn | could be | 23:28 |
| mntmn | VerbotenOS | 23:28 |
| _Bnu | Hard to say without being able to see the Amiga serial output, which I can only see on the PiStorm anyway, haha. | 23:29 |
| _Bnu | I don't have any cable suitable to hook up Amiga serial to my PC. | 23:29 |
| mntmn | yeah i have a amiga serial to usb cable contraption but that is for another day | 23:29 |
| _Bnu | I forgot to click Generate Bipstream... | 23:29 |
| _Bnu | "Why is it still not working... oh." | 23:29 |
| mntmn | i will try the iprefs from wb2.1 | 23:30 |
| mntmn | something is wrong anyway that is hiding my menu bar all the time | 23:30 |
| mntmn | yeah, fastiprefs is at fault | 23:31 |
| mntmn | the original commodore iprefs doesn't kill my menu bar | 23:31 |
| mntmn | but that's not the reason for slow screen dragging. | 23:32 |
| _Bnu | The menu bar is satan, after all... | 23:33 |
| mntmn | schnicasso schneinty schnix | 23:33 |
| _Bnu | But yeah, it seems like it's not calling BlitRect for some reason. AROS library feels like a likely culprit. Or some missing Intuition stuff. | 23:34 |
| _Bnu | Since AROS loves to redraw everything pixel by pixel... | 23:34 |
| - xet7 (QUIT: Read error: Connection reset by peer) (~xet7@user/xet7) | 23:34 | |
| mntmn | hmm no i am not using AROS rom. | 23:34 |
| mntmn | i am using Kickstart 3.1 | 23:34 |
| + xet7 (~xet7@user/xet7) | 23:34 | |
| mntmn | and i have no libraries that interfere with rom things | 23:35 |
| mntmn | maybe it's because i don't run setpatch | 23:35 |
| _Bnu | Is Intuition in ROM... maybe it is... | 23:35 |
| mntmn | or it's just slow on 68030/25 | 23:35 |
| mntmn | _Bnu: intuition is totally in rom | 23:35 |
| _Bnu | Untuition. | 23:36 |
| mntmn | early startup uses it | 23:36 |
| mntmn | and boot with no startup sequence... | 23:36 |
| mntmn | also the whole workbench is in ROM | 23:36 |
| mntmn | LoadWB is really only 1 rom call to open it | 23:36 |
| mntmn | except if you have 3.1.x+ that uses workbench.library outside of rom. | 23:37 |
| mntmn | exec, dos, graphics, intuition, workbench, all in R O M :O | 23:39 |
| _Bnu | I am Workbench... | 23:40 |
| _Bnu | Maybe you can run that Sashimi thing and check if there's serial output? | 23:41 |
| mntmn | that's a good idea! | 23:41 |
| _Bnu | Fish to the rescue... | 23:42 |
| _Bnu | I still don't understand why the Xilinx SDK deletes the path to the .bif file after the first time you build the boot image, haha. | 23:45 |
| _Bnu | And yeah, disabling the bot orm makes the FastATA boot again. | 23:46 |
| mntmn | :o | 23:47 |
| mntmn | so it's not debug output. | 23:47 |
| mntmn | sashimi shows nothing | 23:47 |
| mntmn | ok i will make feierabend | 23:47 |
| _Bnu | ('A`) | 23:47 |
| mntmn | i need to bring the dog to my home... | 23:47 |
| _Bnu | Freitagsmann. | 23:47 |
| mntmn | naja we are on the way to a release... just some more testing and fixing | 23:47 |
| _Bnu | I'll push the split_pos fix to the firmware as well(s). | 23:48 |
| mntmn | the firmware needs a fix for it? :3 | 23:50 |
| _Bnu | Yeah, it needs to be set to 0 in handle_amyga_recet. | 23:51 |
| _Bnu | You said that . . . | 23:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
