CPU interrupts: Revision history

26 February 2024

• curprev 21:4121:41, 26 February 2024‎ Tepples talk contribs‎ 8,480 bytes +491‎ add a bit more meat to the lead section, linking to IRQ and NMI early

23 October 2023

• curprev 10:3910:39, 23 October 2023‎ Orl talk contribs‎ m 7,989 bytes +46‎ Replace deadline to "MOS 6502 programming manual" to a wayback machine link

16 February 2021

• curprev 06:3506:35, 16 February 2021‎ Rainwarrior talk contribs‎ 7,943 bytes +99‎ →‎Interrupt hijacking: note that the B bit is still set with the NMI hijack of BRK, according to V6502, and it can be dealt with

24 October 2018

• curprev 22:5522:55, 24 October 2018‎ Rainwarrior talk contribs‎ 7,844 bytes +8‎ link IRQ

8 September 2018

• curprev 06:1506:15, 8 September 2018‎ Rainwarrior talk contribs‎ 7,836 bytes +342‎ →‎Interrupt hijacking: explain that this bug mostly only affects BRK
• curprev 05:5205:52, 8 September 2018‎ Rainwarrior talk contribs‎ 7,494 bytes +54‎ →‎Branch instructions and interrupts: clarify that it happens on cycle 2 and 4 but not 3.

12 March 2017

• curprev 20:2920:29, 12 March 2017‎ Tepples talk contribs‎ 7,440 bytes +16‎ →‎Notes: Additional article about NMI
• curprev 20:2620:26, 12 March 2017‎ Tepples talk contribs‎ 7,424 bytes +267‎ →‎Interrupt hijacking: Conspicuous by its absence was the address of each of the three vectors. Reported by DementedPurple at https://forums.nesdev.com/viewtopic.php?p=191016#p191016

12 September 2013

• curprev 23:0123:01, 12 September 2013‎ Ulfalizer talk contribs‎ 7,157 bytes +259‎ Add short intro

29 August 2013

• curprev 11:3911:39, 29 August 2013‎ Ulfalizer talk contribs‎ m 6,898 bytes +1‎ No edit summary
• curprev 11:3611:36, 29 August 2013‎ Ulfalizer talk contribs‎ 6,897 bytes +102‎ Try to be clearer about the output from the level detector
• curprev 11:3311:33, 29 August 2013‎ Ulfalizer talk contribs‎ m 6,795 bytes 0‎ No edit summary
• curprev 11:3211:32, 29 August 2013‎ Ulfalizer talk contribs‎ 6,795 bytes +136‎ Rephrase detailed behavior in terms of edge and level detectors and fix CLI inaccuracy (the flag really changes during the "third" cycle, overlapping the opcode fetch)

26 August 2013

• curprev 07:1907:19, 26 August 2013‎ Ulfalizer talk contribs‎ m 6,659 bytes −6‎ No edit summary
• curprev 07:1607:16, 26 August 2013‎ Ulfalizer talk contribs‎ 6,665 bytes +143‎ All two-cycle instructions poll interrupts at the end of the first cycle (as a side note, this is why the branch instructions do too, since they can be two-cycle)
• curprev 06:5306:53, 26 August 2013‎ Ulfalizer talk contribs‎ 6,522 bytes +201‎ NMI priority over IRQ
• curprev 06:2706:27, 26 August 2013‎ Ulfalizer talk contribs‎ 6,321 bytes +346‎ Give possible explanation for why interrupts are sometimes said to be polled during the final cycle
• curprev 06:1106:11, 26 August 2013‎ Ulfalizer talk contribs‎ m 5,975 bytes +1‎ No edit summary
• curprev 06:1006:10, 26 August 2013‎ Ulfalizer talk contribs‎ 5,974 bytes +148‎ Be clearer about how IRQ and NMI are similar
• curprev 06:0306:03, 26 August 2013‎ Ulfalizer talk contribs‎ 5,826 bytes +185‎ A reset is basically just another interrupt type
• curprev 05:5605:56, 26 August 2013‎ Ulfalizer talk contribs‎ m 5,641 bytes +16‎ No edit summary
• curprev 05:5405:54, 26 August 2013‎ Ulfalizer talk contribs‎ 5,625 bytes −7‎ s/during before/before/
• curprev 05:5305:53, 26 August 2013‎ Ulfalizer talk contribs‎ 5,632 bytes −5‎ Branch instructions poll interrupts just before the second cycle (like all two-cycle instructions do)
• curprev 05:4905:49, 26 August 2013‎ Ulfalizer talk contribs‎ 5,637 bytes +70‎ Interrupts are polled right before the PCH fixup cycle for branches. Looks like the interrupt hijacking timing was already correct.
• curprev 05:3805:38, 26 August 2013‎ Ulfalizer talk contribs‎ 5,567 bytes −56‎ Interrupts are actually polled at the end of the second-to-last cycle (verified in Visual 6502, and doing it like that makes blargg's cpu_interrupts_v2 pass) (more changes coming)

24 August 2013

• curprev 07:1207:12, 24 August 2013‎ Koitsu talk contribs‎ 5,623 bytes +5‎ →‎Delayed IRQ response after CLI, SEI, and PLP
• curprev 07:1007:10, 24 August 2013‎ Koitsu talk contribs‎ 5,618 bytes +22‎ →‎Interrupt hijacking
• curprev 07:0907:09, 24 August 2013‎ Koitsu talk contribs‎ 5,596 bytes +92‎ →‎Delayed IRQ response after CLI, SEI, and PLP

11 June 2013

• curprev 00:5100:51, 11 June 2013‎ Ulfalizer talk contribs‎ 5,504 bytes −11‎ Interrupt polling actually happens on the falling edge of φ2

22 May 2013

• curprev 13:4113:41, 22 May 2013‎ Ulfalizer talk contribs‎ m 5,515 bytes 0‎ Ulfalizer moved page CPU interrupt quirks to CPU interrupts: It is about more than just "quirks"
• curprev 12:4612:46, 22 May 2013‎ Ulfalizer talk contribs‎ 5,515 bytes +110‎ Interrupt polling occurs during φ1 (confirmed in Visual 6502)

14 April 2013

• curprev 04:5304:53, 14 April 2013‎ Ulfalizer talk contribs‎ 5,405 bytes −16‎ Rearrange sections to put the hijacking quirk last and reword CLI/SEI/PLP explanation a bit
• curprev 04:4604:46, 14 April 2013‎ Ulfalizer talk contribs‎ 5,421 bytes −48‎ No PC increments at all for IRQ/NMI, so reword BRK comparison

12 April 2013

• curprev 14:2614:26, 12 April 2013‎ Ulfalizer talk contribs‎ 5,469 bytes −14‎ No PC increments at all for IRQ/NMI
• curprev 11:3511:35, 12 April 2013‎ Ulfalizer talk contribs‎ 5,483 bytes +24‎ Add I flag set points to tick-by-tick breakdowns
• curprev 05:4105:41, 12 April 2013‎ Ulfalizer talk contribs‎ 5,459 bytes −85‎ Remove link to thread as it isn't really necessary with the added information and also inaccurate about the polling
• curprev 05:1005:10, 12 April 2013‎ Ulfalizer talk contribs‎ 5,544 bytes +878‎ Add a primer on low-level interrupt behavior and details on the branch instructions
• curprev 03:2503:25, 12 April 2013‎ Ulfalizer talk contribs‎ 4,666 bytes +206‎ No edit summary
• curprev 02:1702:17, 12 April 2013‎ Ulfalizer talk contribs‎ 4,460 bytes +48‎ Update explanation for CLI/SEI/PLP IRQ delay and link to a more relevant post

11 April 2013

• curprev 18:5018:50, 11 April 2013‎ Lidnariq talk contribs‎ m 4,412 bytes +21‎ →‎Notes: make link to visual6502 wiki slightly nicer
• curprev 13:4113:41, 11 April 2013‎ Ulfalizer talk contribs‎ 4,391 bytes +247‎ Add explanation for the delayed IRQ behavior
• curprev 13:2513:25, 11 April 2013‎ Ulfalizer talk contribs‎ m 4,144 bytes 0‎ Might as well use A for both addresses
• curprev 13:2313:23, 11 April 2013‎ Ulfalizer talk contribs‎ m 4,144 bytes −24‎ Not sure if the signal status is actually queried at that point, so just say "determines", as that has been verified down to the tick
• curprev 13:2113:21, 11 April 2013‎ Ulfalizer talk contribs‎ 4,168 bytes +143‎ Add link to page on visual6502.org that covers the interrupt hijacking quirk
• curprev 13:1913:19, 11 April 2013‎ Ulfalizer talk contribs‎ 4,025 bytes −154‎ Remove some redundancy
• curprev 13:1013:10, 11 April 2013‎ Ulfalizer talk contribs‎ 4,179 bytes +1,018‎ Add tick-by-tick execution profiles for IRQ and NMI
• curprev 10:0510:05, 11 April 2013‎ Ulfalizer talk contribs‎ 3,161 bytes +166‎ Add note about the B flag
• curprev 09:5009:50, 11 April 2013‎ Ulfalizer talk contribs‎ m 2,995 bytes +6‎ s/IRQ/interrupt/ to be clearer
• curprev 09:4809:48, 11 April 2013‎ Ulfalizer talk contribs‎ m 2,989 bytes +31‎ No edit summary
• curprev 09:4709:47, 11 April 2013‎ Ulfalizer talk contribs‎ m 2,958 bytes −5‎ Reword a bit