| - S0rin (QUIT: Ping timeout: 252 seconds) (~S0rin@user/s0rin) | 01:12 | |
| + S0rin (~S0rin@user/s0rin) | 01:15 | |
| - RobDangerous (QUIT: Ping timeout: 252 seconds) (~Thunderbi@p200300ec8f12b6006412f528d5658683.dip0.t-ipconnect.de) | 02:20 | |
| - KK (QUIT: Ping timeout: 256 seconds) (~KK@93-139-149-249.adsl.net.t-com.hr) | 06:49 | |
| + Mike (~Mike@102.165.16.90) | 08:36 | |
| - Mike (QUIT: Client Quit) (~Mike@102.165.16.90) | 08:37 | |
| + RobDangerous (~Thunderbi@p200300ec8f12b600104d828bf1654761.dip0.t-ipconnect.de) | 10:01 | |
| mntmn | hey i'm mailing with thomas who apparently developed the Prelude | 10:31 |
|---|---|---|
| mntmn | > Ich habe ja damals die Prelude entwickelt, die ihren Datenstrom über einen 1024-Byte FIFO erhält. Immer, wenn der halb leer läuft, gibt’s einen INT6, der spätestens in ca. 2,9 ms abgearbeitet werden muss, sonst reißt der Datenstrom ab. | 10:31 |
| mntmn | he wrote that some drivers do not handle the Z flag correctly when exiting their intservers | 10:32 |
| mntmn | that's the explanation for int6 conflicts with some other cards | 10:32 |
| mntmn | but i also only recently discovered this and put that in the eth&audio drivers | 10:33 |
| - RobDangerous (QUIT: Ping timeout: 245 seconds) (~Thunderbi@p200300ec8f12b600104d828bf1654761.dip0.t-ipconnect.de) | 11:31 | |
| + RobDangerous (~Thunderbi@p200300ec8f24f800104d828bf1654761.dip0.t-ipconnect.de) | 11:57 | |
| - RobDangerous (QUIT: Client Quit) (~Thunderbi@p200300ec8f24f800104d828bf1654761.dip0.t-ipconnect.de) | 11:57 | |
| + RobDangerous (~Thunderbi@p200300ec8f24f800104d828bf1654761.dip0.t-ipconnect.de) | 11:58 | |
| _Bnu | Ah yeah, but isn't that just returning 1/0 when exiting the interrupt server? Handled/unhandled. | 12:29 |
| _Bnu | At least that's what it seemed like based on the ancient Commodore documentation... | 12:32 |
| _Bnu | This function must preserve no registers. | 12:38 |
| _Bnu | mntmn, help me fix the IRQ handling on the PiStorm... | 12:44 |
| mntmn | _Bnu: well it's not directly about the value but if it clears or sets Z as a side effect, so 0=1 and not 0=0 | 13:04 |
| mntmn | but yeah, returning something works | 13:04 |
| mntmn | _Bnu: unfortunately i have to process 100 emails | 13:04 |
| _Bnu | Can't you let the VIC-20 handle that... | 13:29 |
| _Bnu | But yeah, hm. I'm pretty sure the Commodore docs said you should "return" 0 or 1, and there's a note about the bug if you set your IRQ server priority really high and it gets mixed up with the custom chips. | 13:30 |
| _Bnu | Then you actually have to return the value already in d0 even if you don't handle it, else it gets messed up. | 13:30 |
| _Bnu | But I don't know why you'd set the interrupt server priority that high. | 13:30 |
| mntmn | hmm, would be cool to use commodores for some aspects of this business, but would require additional p r o g r a m m i n g | 13:49 |
| mntmn | also, old computers crash all the time, really annoying... | 13:49 |
| mntmn | _Bnu: http://amigadev.elowar.com/read/ADCD_2.1/Includes_and_Autodocs_3._guide/node01DA.html | 13:49 |
| mntmn | _Bnu: > (Take care with High Level Language servers, the language may not | 13:50 |
| mntmn | have a mechanism for reliably setting the Z flag on exit). | 13:50 |
| _Bnu | Yeah, but hm. I'm pretty sure any "return" with for instance bebbo-gcc just sets d0 to that value. Unless you super wrangle it. | 14:02 |
| - RobDangerous (QUIT: Ping timeout: 245 seconds) (~Thunderbi@p200300ec8f24f800104d828bf1654761.dip0.t-ipconnect.de) | 14:34 | |
| mntmn | _Bnu: yeah. i know. just saying it was not obvious for me at first that "setting Z" means returning 0. i had it the wrong way around at first. | 14:35 |
| _Bnu | To me, setting Z usually means going to sleep... | 14:39 |
| mntmn | true, true | 14:59 |
| _Bnu | I added some poop guarding and possibly reduced the number of fraudulent IRQs, but I still don't know if this is correct. It probably isn't. | 15:19 |
| apolkosnik[m] | IRQ handler can be interrupted by a higher level IRQ | 17:00 |
| _Bnu | Yes. | 17:05 |
| _Bnu | But applying the "only go to a higher IRQ" logic doesn't work when using Musashi. | 17:06 |
| _Bnu | And because seemingly no one knows how it's supposed to work, or they don't feel like sharing, I have no idea what the fuck. | 17:06 |
| apolkosnik[m] | It would be interesting to try to teach AI emulation | 18:03 |
| + KK (~KK@93-139-156-89.adsl.net.t-com.hr) | 18:51 | |
| - KK (QUIT: Quit: Client closed) (~KK@93-139-156-89.adsl.net.t-com.hr) | 19:15 | |
| + KK (~KK@93-139-156-89.adsl.net.t-com.hr) | 21:12 | |
| + RobDangerous (~Thunderbi@p200300ec8f35fa001191f1d2311b0cd1.dip0.t-ipconnect.de) | 21:27 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!