Macintosh-SW/powermac-rom
1
0
Fork 0
mirror of https://github.com/elliotnunn/powermac-rom.git synced 2024-06-07 08:30:20 +00:00
Code Issues Projects Releases Wiki Activity
933bd48f01
powermac-rom/NanoKernel/NKTasks.s

1565 lines
30 KiB
ArmAsm
Raw Normal View History

Auto-generate per-file function lists
2018-06-30 08:00:35 +00:00
; AUTO-GENERATED SYMBOL LIST
; IMPORTS:
; NKAddressSpaces
; FindAreaAbove
; NKIndex
; DeleteID
; LookupID
; MakeID
; NKMPCalls
; CommonMPCallReturnPath
; MPCall_6_0x78
; ReleaseAndReturnMPCall
; ReleaseAndReturnMPCallInvalidIDErr
; ReleaseAndReturnMPCallOOM
; ReleaseAndReturnZeroFromMPCall
; ReleaseAndScrambleMPCall
; ReturnMPCallInvalidIDErr
; ReturnMPCallOOM
; NKPoolAllocator
; PoolAlloc
; PoolAllocClear
; PoolFree
; NKScheduler
; CalculateTimeslice
; FlagSchEval
; FlagSchEvaluationIfTaskRequires
; SchRdyTaskNow
; SchTaskUnrdy
; NKSync
; EnqueueMessage
; NKThud
; panic
; NKTimers
; DequeueTimer
; GetTime
; EXPORTS:
; CreateTask (=> NKInit, NKMPCalls)
; TasksFuncThatIsNotAMPCall (=> NKScheduler)
; ThrowTaskToDebugger (=> NKInterrupts)
Initial commit
2017-11-19 04:11:07 +00:00
; This file mostly provides MPCall implementations related to multitasking.
; We won't understand this very well until someone disassembles MPLibrary.
Local_Panic set *
b panic
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; ARG ProcessID r3
; RET OSStatus r3, TaskID r10
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
; kMPCreateTaskSuspendedMask is ignored?
Initial commit
2017-11-19 04:11:07 +00:00
DeclareMPCall 7, MPCall_7
MPCall_7 ; OUTSIDE REFERER
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
rlwinm. r8, r5, 0, ~0x00000006 ; kMPCreateTaskValidOptionsMask minus kMPCreateTaskSuspendedMask
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
bl LookupID
cmpwi r9, Process.kIDClass
mr r30, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r16, Process.Flags(r30)
rlwinm. r17, r16, 0, 30, 30
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; ARG CPUFlags r7, Process *r8
Initial commit
2017-11-19 04:11:07 +00:00
bl CreateTask
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; RET Task *r8
Initial commit
2017-11-19 04:11:07 +00:00
mr. r31, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndScrambleMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
Initial commit
2017-11-19 04:11:07 +00:00
mfsprg r15, 0
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r17, Task.ID(r31)
stw r17, ContextBlock.r10(r6)
lhz r16, EWA.CPUIndex(r15)
sth r16, Task.CPUIndex(r31)
addi r16, r31, Task.ContextBlock
lwz r17, ContextBlock.r7(r6)
stw r17, ContextBlock.r12(r16)
lwz r17, ContextBlock.r8(r6)
stw r17, ContextBlock.CodePtr(r16)
lwz r17, ContextBlock.r9(r6)
stw r17, ContextBlock.r2(r16)
Initial commit
2017-11-19 04:11:07 +00:00
stw r4, 0x0098(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
Reverse power management functions This is part of the work to get the mini running well. The kcPowerDispatch and kcCacheDispatch entry points were explored. The dead code implementing a Timer "Heartbeat" was also used to label some structures better.
2018-03-09 11:46:39 +00:00
lwz r17, ContextBlock.r6(r6)
stw r17, ContextBlock.LR(r31)
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r16, Task.Flags(r28)
Initial commit
2017-11-19 04:11:07 +00:00
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
rlwinm. r8, r5, 0, kMPCreateTaskTakesAllExceptionsMask
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @noflag
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
_bset r16, r16, Task.kFlagTakesAllExceptions
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
@noflag
Initial commit
2017-11-19 04:11:07 +00:00
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
rlwinm. r8, r5, 0, kMPCreateTaskNotDebuggableMask
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @noflag2
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
_bset r16, r16, Task.kFlagNotDebuggable
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
@noflag2
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
stw r16, Task.Flags(r28)
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
b ReleaseAndReturnZeroFromMPCall
Initial commit
2017-11-19 04:11:07 +00:00
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
; ARG Flags r7, Process *r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; RET Task *r8
Initial commit
2017-11-19 04:11:07 +00:00
CreateTask
; Save arg and lr for later
mr r27, r8
mflr r29
; Create the 1k TASK struct in the pool and give it an ID, leave ptr in r28
li r8, 0x400 ;Task.Size
Reverse the NanoKernel pool allocator
2018-03-23 09:22:56 +00:00
bl PoolAllocClear
Initial commit
2017-11-19 04:11:07 +00:00
mr. r28, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @fail_oom
Initial commit
2017-11-19 04:11:07 +00:00
; Allocate an opaque ID for it
li r9, Task.kIDClass
bl MakeID
cmpwi r8, 0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @fail_no_id
Initial commit
2017-11-19 04:11:07 +00:00
; ID and sign it
stw r8, Task.ID(r28)
lisori r8, Task.kSignature
stw r8, Task.Signature(r28)
; Untitled. Usually set to creator code of owning MacOS process
lisori r8, '----'
stw r8, Task.Name(r28)
; Create a subordinate notification struct -- NOPENOPENOPE
li r8, 0x1c ;Notification.Size
Reverse the NanoKernel pool allocator
2018-03-23 09:22:56 +00:00
bl PoolAllocClear
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r8, 0
stw r8, Task.NotificationPtr(r28)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @fail_note_oom
Initial commit
2017-11-19 04:11:07 +00:00
lisori r9, 'note'
stw r9, 4(r8)
; Create a semaphore struct inside the task
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
addi r16, r28, Task.PageFaultSema
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
_lstart r17, Semaphore.kSignature
stw r16, Semaphore.BlockedTasks + LLL.Next(r16)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
_lfinish
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
stw r16, Semaphore.BlockedTasks + LLL.Prev(r16)
stw r17, Semaphore.BlockedTasks + LLL.Signature(r16)
Initial commit
2017-11-19 04:11:07 +00:00
li r16, 1
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
stw r16, Task.PageFaultSema + Semaphore.MaxValue(r28)
Initial commit
2017-11-19 04:11:07 +00:00
li r16, 0
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
stw r16, Task.PageFaultSema + Semaphore.Value(r28)
Initial commit
2017-11-19 04:11:07 +00:00
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
addi r8, r28, Task.PageFaultSema
Initial commit
2017-11-19 04:11:07 +00:00
li r9, Semaphore.kIDClass
bl MakeID
cmpwi r8, 0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @fail_semq_no_id
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
stw r8, Task.PageFaultSema + Semaphore.BlockedTasks + LLL.Freeform(r28)
Initial commit
2017-11-19 04:11:07 +00:00
; Allocate a vector (i.e. AltiVec) save area
; Conditionally, that is
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
rlwinm. r8, r7, 0, 1 << (31 - EWA.kFlagVec)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @non_altivec_task
Initial commit
2017-11-19 04:11:07 +00:00
; Allocate and check
li r8, 0x214 ;VectorSaveArea.Size ; room for v registers plus 20 bytes
Reverse the NanoKernel pool allocator
2018-03-23 09:22:56 +00:00
bl PoolAllocClear
Initial commit
2017-11-19 04:11:07 +00:00
andi. r9, r8, 16-1 ; Sanity check: aligned to size of vector register?
cmpwi cr1, r8, 0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne Local_Panic
beq cr1, @fail_altivec_oom
Initial commit
2017-11-19 04:11:07 +00:00
; Point to it (from inside and outside the ECB-like area)
stw r8, Task.VectorSaveArea(r28)
stw r8, Task.ContextBlock + ContextBlock.VectorSaveArea(r28)
; Fill the actual register parts with 0x7fffffff
li r16, 0x80 ;VectorSaveArea.RegisterAreaSize / 4
subi r8, r8, 4
lwz r17, PSA.VectorRegInitWord(r1)
@vectorarea_copyloop
subi r16, r16, 1
stwu r17, 4(r8)
cmpwi r16, 0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bgt @vectorarea_copyloop
Initial commit
2017-11-19 04:11:07 +00:00
@non_altivec_task
; Some unexplored DLYQ stuff
addi r16, r1, PSA.DelayQueue
addi r17, r28, 0x08
stw r16, 0x0000(r17)
stw r16, 0x0008(r17)
lwz r18, 0x000c(r16)
stw r18, 0x000c(r17)
stw r17, 0x0008(r18)
stw r17, 0x000c(r16)
li r16, 0
Reverse power management functions This is part of the work to get the mini running well. The kcPowerDispatch and kcCacheDispatch entry points were explored. The dead code implementing a Timer "Heartbeat" was also used to label some structures better.
2018-03-09 11:46:39 +00:00
stb r16, Task.State(r28)
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r16, 9 ; (Z>>Task.kFlag28) | (Z>>Task.kFlag31)
stw r16, Task.Flags(r28)
Initial commit
2017-11-19 04:11:07 +00:00
lisori r16, 'time'
stw r16, 0x0024(r28)
li r16, 1
stb r16, 0x0036(r28)
li r16, 100
stw r16, Task.Weight(r28)
li r16, Task.kNominalPriority
stb r16, Task.Priority(r28)
addi r16, r28, Task.ContextBlock
stw r16, Task.ContextBlockPtr(r28) ; overridden to real ECB on blue
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
lwz r16, PSA.FlagsTemplate(r1)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
stw r16, Task.ContextBlock + ContextBlock.Flags(r28)
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, PSA.UserModeMSR(r1)
stw r16, Task.ContextBlock + ContextBlock.MSR(r28)
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
addi r16, r1, KDP.VecBaseSystem
stw r16, Task.VecBase(r28)
Initial commit
2017-11-19 04:11:07 +00:00
li r16, 0
lwz r17, Task.NotificationPtr(r28)
stw r16, 0x0010(r17)
stw r16, 0x0014(r17)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r16, kMPTaskAbortedErr
Initial commit
2017-11-19 04:11:07 +00:00
stw r16, 0x0018(r17)
li r16, 0
stw r16, Task.Zero1(r28)
stw r16, Task.Zero2(r28)
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
stw r16, Task.CodeFaultCtr(r28)
stw r16, Task.DataFaultCtr(r28)
stw r16, Task.PreemptCtr(r28)
Initial commit
2017-11-19 04:11:07 +00:00
; Who knows that these are for
bl GetTime
stw r8, Task.CreateTime1(r28)
stw r9, Task.CreateTime1 + 4(r28)
stw r8, Task.CreateTime2(r28)
stw r9, Task.CreateTime2 + 4(r28)
stw r8, Task.CreateTime3(r28)
stw r9, Task.CreateTime3 + 4(r28)
lwz r16, KDP.NanoKernelInfo + NKNanoKernelInfo.TaskCount(r1)
addi r16, r16, 1
stw r16, KDP.NanoKernelInfo + NKNanoKernelInfo.TaskCount(r1)
; Squeeze some info (including my owning process) out of the passed PROC ptr
stw r27, Task.OwningProcessPtr(r28)
lwz r16, Process.ID(r27)
stw r16, Task.ProcessID(r28)
lwz r17, Process.SystemAddressSpacePtr(r27)
stw r17, Task.AddressSpacePtr(r28)
lwz r16, AddressSpace.TaskCount(r17)
addi r16, r16, 1
stw r16, AddressSpace.TaskCount(r17)
lwz r16, Process.TaskCount(r27)
addi r16, r16, 1
stw r16, Process.TaskCount(r27)
; Restore and return
mtlr r29
mr r8, r28
blr
@fail_altivec_oom
lwz r8, 0x00a0(r28)
bl DeleteID
@fail_semq_no_id:
lwz r8, 0x009c(r28)
bl PoolFree
@fail_note_oom
lwz r8, 0x0000(r28)
bl DeleteID
@fail_no_id
mr r8, r28
bl PoolFree
@fail_oom
mtlr r29
li r8, 0
blr
DeclareMPCall 8, MPCall_8
MPCall_8 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
mr r31, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
lbz r16, 0x0018(r31)
cmpwi r16, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwz r8, 0x0060(r31)
; r8 = id
bl LookupID
cmpwi r9, Process.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne Local_Panic
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, 0x0008(r8)
rlwinm. r17, r16, 0, 30, 30
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwz r30, 0x0088(r31)
stw r4, 0x0074(r31)
stw r5, 0x011c(r30)
lwz r18, 0x009c(r31)
lwz r16, 0x0134(r6)
lwz r17, 0x013c(r6)
stw r16, 0x0010(r18)
stw r17, 0x0014(r18)
lwz r16, 0x0144(r6)
lwz r17, 0x014c(r6)
stw r16, 0x010c(r30)
stw r16, 0x0090(r31)
stw r17, 0x0094(r31)
addi r16, r31, 0x08
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
RemoveFromList r16, scratch1=r17, scratch2=r18
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl SchRdyTaskNow
Initial commit
2017-11-19 04:11:07 +00:00
bl CalculateTimeslice
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl FlagSchEvaluationIfTaskRequires
Initial commit
2017-11-19 04:11:07 +00:00
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
DeclareMPCall 9, MPCall_9
MPCall_9 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
; r8 = something not sure what
; r9 = 0:inval, 1:proc, 2:task, 3:timer, 4:q, 5:sema, 6:cr, 7:cpu, 8:addrspc, 9:evtg, 10:cohg, 11:area, 12:not, 13:log
mr r31, r8
cmpwi r9, 0x02
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, 0x0064(r31)
lbz r17, 0x0018(r31)
rlwinm. r18, r16, 0, 30, 30
cmpwi cr1, r17, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
beq cr1, MPCall_9_0xb4
Initial commit
2017-11-19 04:11:07 +00:00
mfsprg r15, 0
lhz r18, 0x001a(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lhz r17, EWA.CPUIndex(r15)
Initial commit
2017-11-19 04:11:07 +00:00
cmpw r18, r17
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_9_0xe0
Initial commit
2017-11-19 04:11:07 +00:00
ori r16, r16, 0x400
stw r16, 0x0064(r31)
li r17, 0x01
stb r17, 0x0019(r31)
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl FlagSchEval
Use new _AssertAndRelease macro This is the usual way to release a lock. Most MP calls employ it indirectly, through their return path.
2017-12-12 07:54:19 +00:00
_AssertAndRelease PSA.SchLock, scratch=r16
Initial commit
2017-11-19 04:11:07 +00:00
subi r10, r10, 4
b MPCall_6_0x78
MPCall_9_0x98 ; OUTSIDE REFERER
lwz r16, 0x0064(r31)
ori r16, r16, 0x02
stw r16, 0x0064(r31)
lwz r17, 0x009c(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r16, kMPTaskAbortedErr
Initial commit
2017-11-19 04:11:07 +00:00
stw r16, 0x0018(r17)
b MPCall_9_0xfc
MPCall_9_0xb4
ori r16, r16, 0x02
stw r16, 0x0064(r31)
addi r16, r31, 0x08
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
RemoveFromList r16, scratch1=r17, scratch2=r18
Initial commit
2017-11-19 04:11:07 +00:00
b MPCall_9_0xf0
MPCall_9_0xe0
ori r16, r16, 0x02
stw r16, 0x0064(r31)
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl SchTaskUnrdy
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_9_0xf0
lwz r17, 0x009c(r31)
li r3, 0x00
stw r4, 0x0018(r17)
MPCall_9_0xfc
Identify all accesses to the Primary System Area Whenever a field was added to the end of the Primary System Area (PSA) record, the offsets of all other fields would change. When reversing NK version 02.27, I noticed that 32 bytes had been added to the end of the PSA between 02.27 and 02.28. To build a byte-perfect 02.27, all references to PSA fields must therefore use the record definition in NKPublic.s instead of a numeric offset. In this commit, all PSA references are identified by field name (_FFF if the field has not yet been reversed).
2018-04-22 09:13:30 +00:00
addi r16, r1, PSA.DelayQueue
Initial commit
2017-11-19 04:11:07 +00:00
addi r17, r31, 0x08
stw r16, 0x0000(r17)
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
InsertAsPrev r17, r16, scratch=r18
Initial commit
2017-11-19 04:11:07 +00:00
lbz r8, 0x0037(r31)
cmpwi r8, 0x01
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_9_0x130
Initial commit
2017-11-19 04:11:07 +00:00
addi r8, r31, 0x20
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl DequeueTimer
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_9_0x130
lwz r8, 0x0098(r31)
; r8 = id
bl LookupID
cmpwi r9, Queue.kIDClass
li r18, 0x00
lwz r17, 0x009c(r31)
stw r18, 0x009c(r31)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_9_0x15c
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
mr r8, r17
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl EnqueueMessage ; Message *r8, Queue *r31
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCall
MPCall_9_0x15c
mr r8, r17
bl PoolFree
b ReleaseAndReturnMPCall
DeclareMPCall 10, MPCall_10
MPCall_10 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
lbz r16, 0x0018(r31)
cmpwi r16, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, 0x0064(r31)
rlwinm. r16, r16, 0, 30, 30
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r31
bl TasksFuncThatIsNotAMPCall
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
TasksFuncThatIsNotAMPCall
mflr r27
mr r26, r8
addi r16, r26, 0x08
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
RemoveFromList r16, scratch1=r17, scratch2=r18
Initial commit
2017-11-19 04:11:07 +00:00
lwz r8, 0x0000(r26)
bl DeleteID
lwz r8, 0x00a0(r26)
bl DeleteID
lwz r8, 0x009c(r26)
cmpwi r8, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @_0x98
Initial commit
2017-11-19 04:11:07 +00:00
bl PoolFree
@_0x98
lwz r8, 0x008c(r26)
cmpwi r8, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @_0xa8
Initial commit
2017-11-19 04:11:07 +00:00
bl PoolFree
@_0xa8
lwz r17, 0x006c(r26)
lwz r16, 0x0010(r17)
addi r16, r16, -0x01
stw r16, 0x0010(r17)
lwz r17, 0x0070(r26)
lwz r16, 0x000c(r17)
addi r16, r16, -0x01
stw r16, 0x000c(r17)
mr r8, r26
bl PoolFree
lwz r16, 0x0ecc(r1)
addi r16, r16, -0x01
stw r16, 0x0ecc(r1)
mtlr r27
blr
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
; int MPIsTaskBlue(TaskID)
; Returns true if the Task ID is the same as the blue Task.
; If the Task ID sent is 0, it uses the calling Task's ID.
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
DeclareMPCall 11, MPIsTaskBlue
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
MPIsTaskBlue ; OUTSIDE REFERER
Initial commit
2017-11-19 04:11:07 +00:00
mfsprg r16, 0
cmpwi r3, 0x00
Reverse power management functions This is part of the work to get the mini running well. The kcPowerDispatch and kcCacheDispatch entry points were explored. The dead code implementing a Timer "Heartbeat" was also used to label some structures better.
2018-03-09 11:46:39 +00:00
lwz r17, PSA.PA_BlueTask(r1)
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r18, EWA.PA_CurTask(r16)
lwz r19, Task.ID(r17)
bne @ID_Provided
lwz r3, Task.ID(r18);if r3 is 0, use calling Task's ID
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
@ID_Provided
Initial commit
2017-11-19 04:11:07 +00:00
cmpw r3, r19
li r3, 0x01
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq CommonMPCallReturnPath
Initial commit
2017-11-19 04:11:07 +00:00
li r3, 0x00
b CommonMPCallReturnPath
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
; returns the ID and SomeLabelField of the blue Task.
; Needs a name. MPGetBlueTaskIDAndSomeLabelField() isn't good enough
Initial commit
2017-11-19 04:11:07 +00:00
DeclareMPCall 12, MPCall_12
MPCall_12 ; OUTSIDE REFERER
mfsprg r14, 0
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r15, EWA.PA_CurTask(r14)
lwz r3, Task.ID(r15)
lwz r4, Task.SomeLabelField(r15)
Initial commit
2017-11-19 04:11:07 +00:00
b CommonMPCallReturnPath
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
; OSStatus MPSetTaskWeight(TaskID, weight)
; sets the weight of a Task. Recalculates scheduler stuff if needed.
; Name is just a guess. Change it if you find out what this is really called.
DeclareMPCall 14, MPSetTaskWeight
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
MPSetTaskWeight ; OUTSIDE REFERER
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r4, 0x01
cmpwi cr1, r4, 10000
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
blt ReturnMPCallInvalidIDErr
bgt cr1, ReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lbz r16, Task.State(r31)
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r16, 0x01
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bne @Set_Weight;if task isn't running, you can change its weight freely
lwz r16, Task.QueueMember(r31)
lwz r17, Task.Weight(r31)
lwz r18, ReadyQueue.TotalWeight(r16)
subf r17, r17, r4 ;get the change in weight
add r18, r17, r18;add the change to ReadyQueue.TotalWeight
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r17, 0x00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
stw r18, ReadyQueue.TotalWeight(r16)
beq @Set_Weight;don't mess with scheduler if weight is unchanged
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl FlagSchEval
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
@Set_Weight
stw r4, Task.Weight(r31)
Initial commit
2017-11-19 04:11:07 +00:00
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; MPLibrary passthrough
; When an exception occurs, the message to the queue will be:
; (32 bits) task ID
; (32 bits) exception type a la MachineExceptions.h
; (32 bits) 0 (reserved)
; ARG TaskID r3, QueueID r4
; RET OSStatus r3
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
DeclareMPCall 56, MPSetExceptionHandler
MPSetExceptionHandler
Initial commit
2017-11-19 04:11:07 +00:00
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
mr r8, r4
Initial commit
2017-11-19 04:11:07 +00:00
bl LookupID
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
cmpwi r9, 0
cmpwi cr1, r9, Queue.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq @isnil
bne cr1, ReleaseAndReturnMPCallInvalidIDErr
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
@isnil
Initial commit
2017-11-19 04:11:07 +00:00
mr r30, r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
stw r4, Task.ExceptionHandlerID(r31)
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnZeroFromMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; MPLibrary passthrough
Initial commit
2017-11-19 04:11:07 +00:00
; Throws an exception to a specified task.
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; ARG TaskID r3, ExceptionKind r4
; RET OSStatus r3
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
DeclareMPCall 57, MPThrowException
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
MPThrowException
Initial commit
2017-11-19 04:11:07 +00:00
mfsprg r15, 0
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; This is gold!
lwz r16, Task.Flags(r31)
mtcr r16
li r3, kMPTaskAbortedErr
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
bc BO_IF, Task.kFlagAborted, ReleaseAndReturnMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskStoppedErr
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
bc BO_IF, Task.kFlagStopped, ReleaseAndReturnMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bc BO_IF, 14, ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lbz r17, 0x0018(r31)
lhz r18, 0x001a(r31)
cmpwi cr1, r17, 0x00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bc BO_IF_NOT, 6, KCThrowException_0x70
Initial commit
2017-11-19 04:11:07 +00:00
ori r16, r16, 0x600
stw r4, 0x00f8(r31)
stw r16, 0x0064(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskBlockedErr
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCall
KCThrowException_0x70
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lhz r19, EWA.CPUIndex(r15)
Initial commit
2017-11-19 04:11:07 +00:00
cmpw r19, r18
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne KCThrowException_0xb8
Initial commit
2017-11-19 04:11:07 +00:00
ori r16, r16, 0x200
stw r4, 0x00f8(r31)
stw r16, 0x0064(r31)
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl SchTaskUnrdy
Identify all accesses to the Primary System Area Whenever a field was added to the end of the Primary System Area (PSA) record, the offsets of all other fields would change. When reversing NK version 02.27, I noticed that 32 bytes had been added to the end of the PSA between 02.27 and 02.28. To build a byte-perfect 02.27, all references to PSA fields must therefore use the record definition in NKPublic.s instead of a numeric offset. In this commit, all PSA references are identified by field name (_FFF if the field has not yet been reversed).
2018-04-22 09:13:30 +00:00
addi r16, r1, PSA.DbugQueue
Initial commit
2017-11-19 04:11:07 +00:00
addi r17, r31, 0x08
stw r16, 0x0000(r17)
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
InsertAsPrev r17, r16, scratch=r18
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskStoppedErr
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCall
KCThrowException_0xb8
lwz r3, 0x0000(r31)
ori r16, r16, 0x400
stw r16, 0x0064(r31)
li r17, 0x01
stb r17, 0x0019(r31)
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl FlagSchEval
Use new _AssertAndRelease macro This is the usual way to release a lock. Most MP calls employ it indirectly, through their return path.
2017-12-12 07:54:19 +00:00
_AssertAndRelease PSA.SchLock, scratch=r16
Initial commit
2017-11-19 04:11:07 +00:00
subi r10, r10, 4
b MPCall_6_0x78
DeclareMPCall 58, MPCall_58
MPCall_58 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r29, Task.Flags(r31)
Initial commit
2017-11-19 04:11:07 +00:00
mtcr r29
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskAbortedErr
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bc BO_IF, Task.kFlagAborted, ReleaseAndReturnMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bc BO_IF, Task.kFlagPageFaulted, MPCall_58_0x44
bc BO_IF_NOT, Task.kFlagStopped, ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0x44
mtcr r4
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r30, Task.ContextBlockPtr(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bc BO_IF_NOT, 31, MPCall_58_0x68
Initial commit
2017-11-19 04:11:07 +00:00
li r8, 0x1c
Reverse the NanoKernel pool allocator
2018-03-23 09:22:56 +00:00
bl PoolAlloc
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r8, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndScrambleMPCall
Initial commit
2017-11-19 04:11:07 +00:00
li r3, 0x00
b MPCall_58_0x114
MPCall_58_0x68
li r17, 0x3800
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
rlwinm. r8, r29, 0, 18, 18;Task.kFlagPageFaulted, but empw complains when I call it that
andc r29, r29, r17 ;clear Task.kFlagPageFaulted, Task.kFlag19, and Task.kFlag20
Initial commit
2017-11-19 04:11:07 +00:00
li r17, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne cr7, MPCall_58_0x80
Initial commit
2017-11-19 04:11:07 +00:00
ori r17, r17, 0x400
MPCall_58_0x80
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
ble cr7, MPCall_58_0x88
Initial commit
2017-11-19 04:11:07 +00:00
ori r17, r17, 0x200
MPCall_58_0x88
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r18, ContextBlock.MSR(r30)
rlwimi r18, r17, 0, MSR_SEbit, MSR_BEbit
Initial commit
2017-11-19 04:11:07 +00:00
stw r18, 0x00a4(r30)
li r19, 0x600
lwz r17, 0x0008(r31)
Identify all accesses to the Primary System Area Whenever a field was added to the end of the Primary System Area (PSA) record, the offsets of all other fields would change. When reversing NK version 02.27, I noticed that 32 bytes had been added to the end of the PSA between 02.27 and 02.28. To build a byte-perfect 02.27, all references to PSA fields must therefore use the record definition in NKPublic.s instead of a numeric offset. In this commit, all PSA references are identified by field name (_FFF if the field has not yet been reversed).
2018-04-22 09:13:30 +00:00
addi r18, r1, PSA.DbugQueue
Initial commit
2017-11-19 04:11:07 +00:00
andc r29, r29, r19
cmpw cr1, r17, r18
stw r29, 0x0064(r31)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_58_0xb4
bne cr1, MPCall_58_0xe0
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0xb4
addi r16, r31, 0x08
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
RemoveFromList r16, scratch1=r17, scratch2=r18
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r31
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl SchRdyTaskNow
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl FlagSchEvaluationIfTaskRequires
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0xe0
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
ThrowTaskToDebugger ; OUTSIDE REFERER
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
addi r16, r1, PSA.DbugQueue
addi r17, r31, Task.QueueMember
stw r16, LLL.Freeform(r17)
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
InsertAsPrev r17, r16, scratch=r18
Initial commit
2017-11-19 04:11:07 +00:00
li r8, 0x1c
Reverse the NanoKernel pool allocator
2018-03-23 09:22:56 +00:00
bl PoolAlloc
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r29, Task.Flags(r31)
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
_bset r29, r29, Task.kFlagStopped
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0x114
mtcr r29
mr r28, r8
Lots of work on Address Spaces
2018-04-23 15:22:55 +00:00
bc BO_IF, Task.kFlagNotDebuggable, MPCall_58_0x13c
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
bc BO_IF, Task.kFlag20, MPCall_58_0x13c
Identify all accesses to the Primary System Area Whenever a field was added to the end of the Primary System Area (PSA) record, the offsets of all other fields would change. When reversing NK version 02.27, I noticed that 32 bytes had been added to the end of the PSA between 02.27 and 02.28. To build a byte-perfect 02.27, all references to PSA fields must therefore use the record definition in NKPublic.s instead of a numeric offset. In this commit, all PSA references are identified by field name (_FFF if the field has not yet been reversed).
2018-04-22 09:13:30 +00:00
lwz r8, PSA._8e8(r1)
Initial commit
2017-11-19 04:11:07 +00:00
; r8 = id
bl LookupID
cmpwi r9, Queue.kIDClass
mr r30, r8
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
_bset r29, r29, Task.kFlag20
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_58_0x184
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0x13c
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bc BO_IF, 19, MPCall_58_0x158
Initial commit
2017-11-19 04:11:07 +00:00
lwz r8, 0x00f4(r31)
; r8 = id
bl LookupID
cmpwi r9, Queue.kIDClass
mr r30, r8
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
_bset r29, r29, Task.kFlag19
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_58_0x184
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_58_0x158
mr. r8, r28
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bnel PoolFree
Initial commit
2017-11-19 04:11:07 +00:00
addi r16, r31, 0x08
Insert calls to known linked-list macros The presence of InitList, InsertAsNext, InsertAsPrev and RemoveFromList calls should make some hitherto unexplored code easier to read.
2017-12-13 04:10:54 +00:00
RemoveFromList r16, scratch1=r17, scratch2=r18
Initial commit
2017-11-19 04:11:07 +00:00
b MPCall_9_0x98
MPCall_58_0x184
mr. r8, r28
stw r29, 0x0064(r31)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_58_0x1a4
Initial commit
2017-11-19 04:11:07 +00:00
lwz r8, 0x0028(r30)
cmpwi r8, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_58_0x114
Initial commit
2017-11-19 04:11:07 +00:00
lwz r17, 0x0008(r8)
stw r17, 0x0028(r30)
MPCall_58_0x1a4
bl LoadSomeData
lwz r16, 0x0088(r31)
lwz r17, 0x0000(r31)
mflr r18
stw r17, 0x0010(r8)
lwz r17, 0x0074(r16)
lbz r19, 0x0040(r16)
lbzx r18, r18, r19
stw r18, 0x0014(r8)
stw r17, 0x0018(r8)
stw r18, 0x00f8(r31)
mr r31, r30
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl EnqueueMessage ; Message *r8, Queue *r31
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCall
LoadSomeData ; OUTSIDE REFERER
blrl
dc.l 0x0002020d
dc.l 0x01080003
dc.l 0x090a0403
dc.l 0x07000500
dc.l 0x0b0b0403
dc.l 0x07060505
dc.l 0x11000000
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
; Used to extract task state. This will be tricky.
Initial commit
2017-11-19 04:11:07 +00:00
DeclareMPCall 59, MPCall_59
MPCall_59 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr. r8, r3
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_59_0x30
Initial commit
2017-11-19 04:11:07 +00:00
; r8 = id
bl LookupID
cmpwi r9, Queue.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
MPCall_59_0x30
Identify all accesses to the Primary System Area Whenever a field was added to the end of the Primary System Area (PSA) record, the offsets of all other fields would change. When reversing NK version 02.27, I noticed that 32 bytes had been added to the end of the PSA between 02.27 and 02.28. To build a byte-perfect 02.27, all references to PSA fields must therefore use the record definition in NKPublic.s instead of a numeric offset. In this commit, all PSA references are identified by field name (_FFF if the field has not yet been reversed).
2018-04-22 09:13:30 +00:00
stw r3, PSA._8e8(r1)
Initial commit
2017-11-19 04:11:07 +00:00
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
DeclareMPCall 60, MPCall_60
MPCall_60 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
cmpwi r4, 0x05
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x288
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r16, Task.Flags(r31)
Initial commit
2017-11-19 04:11:07 +00:00
mtcr r16
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskAbortedErr
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bc BO_IF, Task.kFlagAborted, ReleaseAndReturnMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bc BO_IF, Task.kFlagPageFaulted, MPCall_60_0x4c
bc BO_IF_NOT, Task.kFlagStopped, ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_60_0x4c
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lbz r16, Task.State(r31)
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r16, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r4, 0x00
cmpwi cr1, r4, 0x01
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0xf8
beq cr1, MPCall_60_0x10c
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r4, 0x02
cmpwi cr1, r4, 0x03
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x150
beq cr1, MPCall_60_0x1c0
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r4, 0x04
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r16, Task.ContextBlockPtr(r31)
Initial commit
2017-11-19 04:11:07 +00:00
li r17, 0x00
cmplwi r5, 0x00
cmplwi cr1, r5, 0x04
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
beq MPCall_60_0xac;gets ID of Area SRR0 is in
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq cr1, MPCall_60_0xc0
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x08
cmplwi cr1, r5, 0x0c
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0xc8
beq cr1, MPCall_60_0xd0
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCallOOM
MPCall_60_0xac
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r8, Task.AddressSpacePtr(r31)
lwz r9, ContextBlock.SRR0(r16)
Initial commit
2017-11-19 04:11:07 +00:00
bl FindAreaAbove
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r17, Area.ID(r8)
Initial commit
2017-11-19 04:11:07 +00:00
b MPCall_60_0x36c
MPCall_60_0xc0
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r17, ContextBlock.SRR0(r16)
Initial commit
2017-11-19 04:11:07 +00:00
b MPCall_60_0x36c
MPCall_60_0xc8
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r17, Task.ErrToReturnIfIDie(r31)
Initial commit
2017-11-19 04:11:07 +00:00
b MPCall_60_0x36c
MPCall_60_0xd0
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r17, ContextBlock.SavedFlags(r16)
lwz r18, 0x0064(r16);some unknown ContextBlock value
rlwinm. r8, r17, 0, EWA.kFlagLowSaves, EWA.kFlagLowSaves
Initial commit
2017-11-19 04:11:07 +00:00
li r17, 0x02
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x36c
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
rlwinm. r8, r18, 0, EWA.kFlag1, EWA.kFlag1
Initial commit
2017-11-19 04:11:07 +00:00
li r17, 0x01
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_60_0x36c
Initial commit
2017-11-19 04:11:07 +00:00
li r17, 0x00
b MPCall_60_0x36c
MPCall_60_0xf8
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r16, Task.ContextBlockPtr(r31)
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi cr1, r5, 0xf8
andi. r17, r5, 0x07
addi r16, r16, 0xfc
b MPCall_60_0x124
MPCall_60_0x10c
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r16, Task.ContextBlockPtr(r31)
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x100
cmplwi cr1, r5, 0xf8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x144
Initial commit
2017-11-19 04:11:07 +00:00
andi. r17, r5, 0x07
addi r16, r16, 0x1fc
MPCall_60_0x124
add r16, r16, r5
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bgt cr1, ReleaseAndReturnMPCallOOM
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwzu r17, 0x0004(r16)
lwzu r18, 0x0004(r16)
lwzu r19, 0x0004(r16)
lwzu r20, 0x0004(r16)
b MPCall_60_0x3a8
MPCall_60_0x144
lwz r17, 0x00e4(r16)
li r18, 0x00
b MPCall_60_0x37c
MPCall_60_0x150
lwz r16, 0x0088(r31)
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
rlwinm. r8, r7, 0, EWA.kFlagVec, EWA.kFlagVec
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, 0x00d8(r16)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi cr3, r16, 0x00
cmplwi r5, 0x200
cmplwi cr2, r5, 0x210
cmplwi cr1, r5, 0x1f0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beql cr3, Local_Panic
beq MPCall_60_0x1a4
beq cr2, MPCall_60_0x1b8
Initial commit
2017-11-19 04:11:07 +00:00
andi. r8, r5, 0x0f
add r16, r16, r5
subi r16, r16, 4
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bgt cr1, ReleaseAndReturnMPCallOOM
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwzu r17, 0x0004(r16)
lwzu r18, 0x0004(r16)
lwzu r19, 0x0004(r16)
lwzu r20, 0x0004(r16)
b MPCall_60_0x3a8
MPCall_60_0x1a4
lwz r17, 0x0200(r16)
lwz r18, 0x0204(r16)
lwz r19, 0x0208(r16)
lwz r20, 0x020c(r16)
b MPCall_60_0x3a8
MPCall_60_0x1b8
lwz r17, 0x0210(r16)
b MPCall_60_0x36c
MPCall_60_0x1c0
lwz r16, 0x0088(r31)
li r17, 0x00
cmplwi r5, 0x00
cmplwi cr1, r5, 0x08
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x21c
beq cr1, MPCall_60_0x228
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x10
cmplwi cr1, r5, 0x30
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x234
beq cr1, MPCall_60_0x240
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x1c
cmplwi cr1, r5, 0x20
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x24c
beq cr1, MPCall_60_0x254
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x24
cmplwi cr1, r5, 0x28
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x25c
beq cr1, MPCall_60_0x264
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x2c
cmplwi cr1, r5, 0x18
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x278
beq cr1, MPCall_60_0x280
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCallOOM
MPCall_60_0x21c
lwz r17, 0x00f0(r16)
lwz r18, 0x00f4(r16)
b MPCall_60_0x37c
MPCall_60_0x228
lwz r17, 0x00e8(r16)
lwz r18, 0x00ec(r16)
b MPCall_60_0x37c
MPCall_60_0x234
lwz r17, 0x00f8(r16)
lwz r18, 0x00fc(r16)
b MPCall_60_0x37c
MPCall_60_0x240
lwz r17, 0x0070(r16)
lwz r18, 0x0074(r16)
b MPCall_60_0x37c
MPCall_60_0x24c
lwz r17, 0x00d4(r16)
b MPCall_60_0x36c
MPCall_60_0x254
lwz r17, 0x00a4(r16)
b MPCall_60_0x36c
MPCall_60_0x25c
lwz r17, 0x00c4(r16)
b MPCall_60_0x36c
MPCall_60_0x264
lbz r17, 0x0040(r16)
bl LoadSomeData
mflr r18
lbzx r17, r18, r17
b MPCall_60_0x36c
MPCall_60_0x278
li r17, 0x00
b MPCall_60_0x36c
MPCall_60_0x280
lwz r17, 0x00dc(r16)
b MPCall_60_0x36c
MPCall_60_0x288
cmplwi cr1, r5, 0x04
cmplwi r5, 0x14
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq cr1, MPCall_60_0x2c4
beq MPCall_60_0x2e4
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi cr1, r5, 0x20
cmplwi r5, 0x30
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq cr1, MPCall_60_0x2f4
beq MPCall_60_0x308
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi cr1, r5, 0x40
cmplwi r5, 0x3c
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq cr1, MPCall_60_0x320
beq MPCall_60_0x318
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi cr1, r5, 0x50
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq cr1, MPCall_60_0x34c
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCallOOM
MPCall_60_0x2c4
lwz r17, 0x0074(r31)
lwz r18, 0x0008(r31)
lwz r18, 0x0004(r18)
lhz r19, 0x001a(r31)
lbz r20, 0x0018(r31)
rlwimi r19, r20, 16, 8, 15
lwz r20, 0x001c(r31)
b MPCall_60_0x3a8
MPCall_60_0x2e4
lwz r17, 0x0060(r31)
lwz r18, 0x00c0(r31)
lwz r19, 0x00c4(r31)
b MPCall_60_0x390
MPCall_60_0x2f4
lwz r17, 0x00c8(r31)
lwz r18, 0x00cc(r31)
lwz r19, 0x00d0(r31)
lwz r20, 0x00d4(r31)
b MPCall_60_0x3a8
MPCall_60_0x308
lwz r17, 0x00e0(r31)
lwz r18, 0x00e4(r31)
lwz r19, 0x00e8(r31)
b MPCall_60_0x390
MPCall_60_0x318
lwz r17, 0x0078(r31)
b MPCall_60_0x36c
MPCall_60_0x320
lbz r20, 0x0018(r31)
li r17, 0x00
lwz r16, 0x0008(r31)
lwz r18, 0x0070(r31)
cmpwi r20, 0x00
lwz r19, 0x0094(r31)
lwz r20, 0x0090(r31)
lwz r18, 0x0000(r18)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne MPCall_60_0x3a8
Initial commit
2017-11-19 04:11:07 +00:00
lwz r17, 0x0000(r16)
b MPCall_60_0x3a8
MPCall_60_0x34c
mfsprg r18, 0
lwz r20, 0x0088(r31)
lwz r19, -0x0008(r18)
cmpw r19, r31
lwz r17, 0x0004(r18)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_60_0x36c
Initial commit
2017-11-19 04:11:07 +00:00
lwz r17, 0x010c(r20)
b MPCall_60_0x36c
MPCall_60_0x36c
li r21, 0x04
stw r17, 0x0134(r6)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_60_0x37c
li r21, 0x08
stw r17, 0x0134(r6)
stw r18, 0x013c(r6)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_60_0x390
li r21, 0x0c
stw r17, 0x0134(r6)
stw r18, 0x013c(r6)
stw r19, 0x0144(r6)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_60_0x3a8
li r21, 0x10
stw r17, 0x0134(r6)
stw r18, 0x013c(r6)
stw r19, 0x0144(r6)
stw r20, 0x014c(r6)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
DeclareMPCall 61, MPCall_61
MPCall_61 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r31, r8
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lwz r16, Task.Flags(r31)
Initial commit
2017-11-19 04:11:07 +00:00
mtcr r16
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
li r3, kMPTaskAbortedErr
bc BO_IF, 30, ReleaseAndReturnMPCall
bc BO_IF, 18, MPCall_61_0x44
bc BO_IF_NOT, 22, ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
MPCall_61_0x44
lbz r16, 0x0018(r31)
cmpwi r16, 0x00
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
lwz r17, 0x0134(r6)
lwz r18, 0x013c(r6)
lwz r19, 0x0144(r6)
lwz r20, 0x014c(r6)
cmpwi r4, 0x00
cmpwi cr1, r4, 0x01
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0x84
beq cr1, MPCall_61_0x98
Initial commit
2017-11-19 04:11:07 +00:00
cmpwi r4, 0x02
cmpwi cr1, r4, 0x03
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0xe8
beq cr1, MPCall_61_0x170
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCallOOM
MPCall_61_0x84
lwz r16, 0x0088(r31)
cmplwi cr1, r5, 0xf8
andi. r8, r5, 0x07
addi r16, r16, 0xfc
b MPCall_61_0xb0
MPCall_61_0x98
lwz r16, 0x0088(r31)
cmplwi r5, 0x100
cmplwi cr1, r5, 0xf8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0xd8
Initial commit
2017-11-19 04:11:07 +00:00
andi. r8, r5, 0x07
addi r16, r16, 0x1fc
MPCall_61_0xb0
add r16, r16, r5
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bgt cr1, ReleaseAndReturnMPCallOOM
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
li r21, 0x10
stwu r17, 0x0004(r16)
stwu r18, 0x0004(r16)
stwu r19, 0x0004(r16)
stwu r20, 0x0004(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0xd8
li r21, 0x04
stw r17, 0x00e4(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0xe8
lwz r16, 0x0088(r31)
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
rlwinm. r8, r7, 0, EWA.kFlagVec, EWA.kFlagVec
Initial commit
2017-11-19 04:11:07 +00:00
lwz r16, 0x00d8(r16)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi cr3, r16, 0x00
cmplwi r5, 0x200
cmplwi cr2, r5, 0x210
cmplwi cr1, r5, 0x1f0
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beql cr3, Local_Panic
beq MPCall_61_0x144
beq cr2, MPCall_61_0x160
Initial commit
2017-11-19 04:11:07 +00:00
andi. r8, r5, 0x0f
add r16, r16, r5
subi r16, r16, 4
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bgt cr1, ReleaseAndReturnMPCallOOM
bne ReleaseAndReturnMPCallOOM
Initial commit
2017-11-19 04:11:07 +00:00
li r21, 0x10
stwu r17, 0x0004(r16)
stwu r18, 0x0004(r16)
stwu r19, 0x0004(r16)
stwu r20, 0x0004(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x144
li r21, 0x10
stw r17, 0x0200(r16)
stw r18, 0x0204(r16)
stw r19, 0x0208(r16)
stw r20, 0x020c(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x160
li r21, 0x04
stw r17, 0x0210(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x170
lwz r16, 0x0088(r31)
cmplwi r5, 0x00
cmplwi cr1, r5, 0x08
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0x1b0
beq cr1, MPCall_61_0x1c4
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x10
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0x1d8
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x1c
cmplwi cr1, r5, 0x20
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0x1ec
beq cr1, MPCall_61_0x1fc
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi r5, 0x24
cmplwi cr1, r5, 0x18
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq MPCall_61_0x218
beq cr1, MPCall_61_0x228
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnMPCallOOM
MPCall_61_0x1b0
li r21, 0x08
stw r17, 0x00f0(r16)
stw r18, 0x00f4(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x1c4
li r21, 0x08
stw r17, 0x00e8(r16)
stw r18, 0x00ec(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x1d8
li r21, 0x08
stw r17, 0x00f8(r16)
stw r18, 0x00fc(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x1ec
li r21, 0x04
stw r17, 0x00d4(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x1fc
li r21, 0x04
lwz r18, 0x00a4(r16)
Reverse PPC-68k context switching (incomplete)
2018-03-31 08:17:23 +00:00
rlwimi r18, r17, 0, 20, 23 ; MSR[FE0/SE/BE/FE1]
Initial commit
2017-11-19 04:11:07 +00:00
rlwimi r18, r17, 0, 31, 31
stw r18, 0x00a4(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x218
li r21, 0x04
stw r17, 0x00c4(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
MPCall_61_0x228
li r21, 0x04
stw r17, 0x00dc(r16)
stw r21, 0x0154(r6)
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
DeclareMPCall 63, MPCall_63
MPCall_63 ; OUTSIDE REFERER
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
stw r4, Task.SomeLabelField(r8)
Initial commit
2017-11-19 04:11:07 +00:00
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
; OSStatus MPSetTaskCPU(TaskID, CPUID)
; Makes it so that a Task can only run on a single CPU.
DeclareMPCall 114, MPSetTaskCPU
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
MPSetTaskCPU ; OUTSIDE REFERER
Initial commit
2017-11-19 04:11:07 +00:00
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
; r8 = id
bl LookupID
cmpwi r9, Task.kIDClass
mr r31, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r4
; r8 = id
bl LookupID
cmpwi r9, CPU.kIDClass
mr r30, r8
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r16, Task.Flags(r31)
lwz r17, Task.PreemptCtr(r31)
rlwinm. r8, r16, 0, Task.kFlagAborted, Task.kFlagAborted
Initial commit
2017-11-19 04:11:07 +00:00
cmplw cr1, r17, r5
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r18, CPU.Flags(r30)
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallOOM
bne cr1, ReleaseAndReturnMPCallOOM
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
rlwinm. r8, r18, 0, CPU.kFlagScheduled, CPU.kFlagScheduled
Initial commit
2017-11-19 04:11:07 +00:00
cmplwi cr1, r17, 0x04
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
beq ReleaseAndReturnMPCallOOM
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
lwz r16, Task.Flags(r31)
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
lhz r17, CPU.EWA + EWA.CPUIndex(r30)
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
ori r16, r16, 0x40;Task.kFlag25
stw r16, Task.Flags(r31)
sth r17, Task.CPUIndex(r31)
rlwinm. r8, r16, 0, Task.kFlag26, Task.kFlag26
Initial commit
2017-11-19 04:11:07 +00:00
mr r8, r31
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
bne @No_Requeueing
Work on the scheduler and interrupt handling code
2018-03-23 09:23:06 +00:00
bl SchTaskUnrdy
bl SchRdyTaskNow
Initial commit
2017-11-19 04:11:07 +00:00
Look at a bunch of Task code
2018-04-29 22:39:50 +00:00
@No_Requeueing
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
bl FlagSchEvaluationIfTaskRequires
Initial commit
2017-11-19 04:11:07 +00:00
; r1 = kdp
b ReleaseAndReturnZeroFromMPCall
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
; ARG TaskID r3, OSType r4
; RET OSStatus r3
Initial commit
2017-11-19 04:11:07 +00:00
DeclareMPCall 126, KCSetTaskType
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
KCSetTaskType
Initial commit
2017-11-19 04:11:07 +00:00
_Lock PSA.SchLock, scratch1=r16, scratch2=r17
mr r8, r3
bl LookupID
cmpwi r9, Task.kIDClass
Remove unnecessary static branch hints These + and - characters in a branch mnemonic can cause the assembler to produce conditional branch instructions with that hint the branch predictor. The default for forward branches is -, and for backward branches is +. If a mnemonic is issued with the opposite sign, then bit 10 of the instruction (the LSB of the BO field) is set. My long-retired "ppcdisasm" script inserted these hints unconditionally, despite 98% of them not being required. The code is much cleaner now. I read in some old MPW release notes that PPCAsm and PPCLink together exhibit a quirk when linking conditional branches to imported symbols. PPCAsm always assembles these conditional branches as if they were forward branches (that is, a + hint will always set the hint bit, and a - hint will never). I hoped to use this property to divine whether the NanoKernel was linked from one or many assembly files, but I was frustrated by the lack of conditional branches between files.
2018-04-15 07:05:02 +00:00
bne ReleaseAndReturnMPCallInvalidIDErr
Initial commit
2017-11-19 04:11:07 +00:00
Reverse userspace sync primitives Namely queues, semaphores, critical regions, event groups and "notifications". The MP calls implementing these services have been named after their MPLibrary wrapper functions. This convention will be followed in the future (no more NKCreateEvent).
2018-02-17 16:57:01 +00:00
stw r4, Task.Name(r8)
Initial commit
2017-11-19 04:11:07 +00:00
b ReleaseAndReturnZeroFromMPCall
Reference in New Issue Copy Permalink