525
edits
Changes
→struct DebugEventInfo
|-
|-
| u32
| u32
| Flags. Bit0 means that svcContinueDebugEvent needs to be called for this event (except for EXIT PROCESS events, for which you need to call svcContinueDebugEvent even if this bit is clear)
| Flags. Bit0 means that svcContinueDebugEvent needs to be called for this event
| u8[4]
| u8[4]
| Remnants of the corresponding flags in [[KEventInfo]], always 0 here
| Remnants of the corresponding flags in [[KEventInfo]], always 0 here
| 2
| 2
|-
|-
| EXIT PROCESS
| EXIT PROCESS (1)
| 3
| 3
|-
|-
| 4
| 4
|-
|-
| DLL LOAD *
| DLL LOAD (3)
| 5
| 5
|-
|-
| DLL UNLOAD *
| DLL UNLOAD (3)
| 6
| 6
|-
|-
| SCHEDULE IN **
| SCHEDULE IN (1) (2)
| 7
| 7
|-
|-
| SCHEDULE OUT *
| SCHEDULE OUT (1) (2)
| 8
| 8
|-
|-
| SYSCALL IN *
| SYSCALL IN (1) (2)
| 9
| 9
|-
|-
| SYSCALL OUT *
| SYSCALL OUT (1) (2)
| 10
| 10
|-
|-
| 11
| 11
|-
|-
| MAP *
| MAP (1) (2)
| 12
| 12
|}
|}
<nowiki>*</nowiki> Unused
<nowiki>(1)</nowiki> Non-blocking: all other events preempt and block all the threads of their process until they are continued.
<nowiki>(2)</nowiki> There is handling code in the kernel but nothing signal those events.
<nowiki>**</nowiki> Referenced but never used in practise
<nowiki>(3)</nowiki> Completely removed from the kernel, but referenced in DMNT. Stubbed relocation code (e.g., in Process9 and in PXI sysmodule) and even whole libraries (e.g., in PXI sysmodule's .rodata section) seem to indicate that Nintendo used dynamic libraries early in system development.
When calling svcDebugActiveProcess, an ATTACH PROCESS debug event is signaled, then ATTACH THREAD for each of its opened threads, then finally ATTACH BREAK.
When calling svcDebugActiveProcess, an ATTACH PROCESS debug event is signaled, then ATTACH THREAD for each of its opened threads, then finally ATTACH BREAK.
| 1
| 1
|-
|-
| UNHANDLED EXCEPTION
| DEBUG TERMINATE
| 2
| 2
|}
|}