Recorded on hardware (pspautotests threads/tls/timeout), a Tlspl
allocation follows the same timeout rule as the other waits, including
failing at once for 0 and 1us without writing the timeout back, which the
shared rule it moved to in the last commits didn't give it yet. Before
that it waited the raw timeout, ~30us short.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Priority-ordered waiting lists were sorted with a comparator wrapper per
object (msgpipe, fpl, vpl), or searched with a copy of the same function
(mutex, mbx). HLEKernel::SortWaitingThreadsByPriority() and
FindBestPriorityWaiter() now do both for any waiting list, of thread ids
or of structs with a threadID.
HLEKernel::ClearWaitingThreads() replaces the identical cancel/delete
loops in semaphores, event flags, fpl and vpl.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Semaphores, event flags, mutexes, lwmutexes, mbx, msgpipes, fpl, vpl,
tlspl and WaitThreadEnd each had their own CoreTiming event, handler
registration and savestate entry for wait timeouts, and their own function
to schedule one. Now one event (WaitThreadEnd's, renamed) times out all of
them, keyed by thread, and dispatches on the thread's wait type to a
timeoutFunc registered alongside the begin/end callback functions.
__KernelWaitCurThreadWithTimeout() starts such a wait, and the HLEKernel
helpers have overloads that use the shared event.
Old savestates still load: each object's section reads its old event id
and points it at the shared handler, so a timeout pending in the state
goes off as before. Checked with a state saved mid-wait by the previous
build, and with four games.
The one behaviour change: tlspl timeouts now follow the same hardware
rule as the others, where they used the raw timeout.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Every wait with a timeout behaves the same on hardware (pspautotests
threads/scheduling/waittimeouts). The deadline is taken, and the alarm set
up a moment later. If the deadline has passed by then, the wait fails with
WAIT_TIMEOUT at once, without yielding or writing the timeout back. That's
usual for 0us, half the time for 1us, and rare after; AllocateVpl does more
first. Otherwise it ends max(t, 205us) + ~35us after the call. Each object
had its own guess (24/245, 25/250, 20/250 and so on), and only MsgPipe had
the immediate case.
__KernelWaitTimesOutAtOnce() and __KernelWaitTimeoutUs() now do it for
semaphores, event flags, mutexes, lwmutexes, mbx, msgpipes, fpl, vpl and
WaitThreadEnd. The latency past the deadline isn't counted in the time
left written back.
Outcomes that hardware decides by the clock's phase (these, and
sceKernelDelayThread returning at once) go with the likelier one. Ones
between 50% and certain are instead spread evenly over calls, so a polling
loop can't lock into never yielding (sceKernelThread section version 7).
This replaces the pseudo-random choice for delays.
Also adds threads/scheduling/readyqueue, which already passes.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Homebrew written for a PSP-2000+ under custom firmware can use the top
32MB of RAM directly without setting MEMSIZE, which leaves the user
partition at its normal size. NJEMU's slim builds do this (#8925).
Previously we didn't map that memory at all for PBPs; MEMSIZE=1 isn't a
workaround either, since it grows the partition and the heap and stacks
land where the program writes.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Reject sizes past the end of the state before allocating (FPL, PGF,
achievements, SAS grain, savedata list, the memory fast path), fail
instead of desyncing on a SAS voice count mismatch, and free what old
states' paths and shrinking pointer containers dropped.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
The third argument is a timeout pointer, as threadman.prx shows. A kernel
address from user mode is ILLEGAL_ADDR there; we used to write through it.
Also, no lookup by index: the syscall requires the exact uid.
Adds the threads/tls/allocate test, recorded on hardware.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
This is the syscall usersystemlib's sceKernelGetTlsAddr makes when the
thread's cached TLS address is null, as (uid, &addr, 0). Code that has to
run before usersystemlib.prx is loaded (like plugins built with a Rust SDK)
inlines sceKernelGetTlsAddr and imports this directly.
Shares the allocation with sceKernelGetTlsAddr. A thread waiting on a full
pool stores its address pointer as the wait value, so no state changes.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
The new sysmem tests run the same partition sweep from both privilege levels, which settles
several things that were guesses:
The valid range is 1-6, not 1-9-except-7. sceKernelCreateVpl, CreateFpl, CreateMsgPipe and
AllocPartitionMemory all let 8 and 9 through to the permission check, so a caller asking for
partition 8 got ILLEGAL_PERM where hardware says ILLEGAL_ARGUMENT. Privilege changes the
permission check, not the range - 1, 3 and 4 are refused from user mode and work from kernel mode
in every one of these APIs, which is the evidence the earlier BlockAllocatorFromID change was
missing.
sceKernelAllocPartitionMemory reports an out-of-range partition differently depending on which
entry point was used - ILLEGAL_ARGUMENT through SysMemUserForUser, ILLEGAL_PARTITION through
SysMemForKernel. Both NIDs land on the same function here, and hleIsKernelMode() is precisely
"came in through the kernel NID", so it picks the right one.
sceKernelCreateHeap had four "TODO: Validate error code" comments and no test at all - it's
kernel-only, which is why. All four are now recorded: out-of-range partitions are
ILLEGAL_PARTITION, a size of zero or less is HEAPBLOCK_ALLOC_FAILED before anything is allocated,
a NULL name is refused with ERROR, and flags really are ignored. sceKernelAllocHeapMemoryWithOption
had its validation backwards: the option struct's size field isn't checked at all, while the
alignment must be a power of two from 4 to 0x80.
Not fixed, and split into sysmem/kernel/heapgrow in tests_next: a real heap will hand out a block
larger than the heap itself, so the size isn't a cap. Ours is a fixed allocator over the reserved
block. Worth establishing how far the real one grows before implementing that.
Risk: the range change makes partitions 8 and 9 fail earlier and with a different code than
before. Nothing in tests_good depended on the old behaviour except two expectations that had
drifted from hardware, corrected in the submodule.
PPSSPP decided whether a caller was privileged with hleIsKernelMode(), which reports whether the
syscall being executed is itself a kernel-only export. That's a different question from the one
the hardware answers: on a PSP the privilege belongs to the calling module, and a kernel module
reaches sceKernelCreateTlspl through the ordinary ThreadManForUser NID like anything else. So a
kernel module asking for partition 1, 3 or 4 got ILLEGAL_PERM where a real PSP hands it over,
which the new threads/tls/kernel/partition test shows directly.
BlockAllocatorFromID now also accepts a caller whose thread belongs to a kernel module, via a new
__KernelCurThreadIsKernelMode(). It checks the thread's own attribute first and then the owning
module, because a kernel module's main thread isn't necessarily flagged kernel - the attribute
comes from PSP_MAIN_THREAD_ATTR, which needn't set it. That mirrors how sceKernelCreateThread
already works out allowKernel.
This only ever widens access, and only for threads belonging to kernel modules, so games are
unaffected - they run in user modules and see exactly what they saw before.
threads/tls/kernel/partition now records the sweep from a kernel module, which settles the range
question the user-mode recording couldn't: privilege changes the permission check, not the range.
Partitions 1, 3 and 4 are ILLEGAL_PERM from user mode and fine from kernel mode, while 7 and up
are ILLEGAL_ARGUMENT either way. So the check goes back to a plain 1-6 for both, and the
kernel-mode carve-out from the last commit - which would have let 8 and 9 through - is gone.
The hardware doc gains a section on kernel-mode tests: what COMMON_KERNEL does, why the stock
crt0 makes a kernel PRX unloadable, which libraries can't be imported, and how much room there
actually is in the kernel partition.
sceKernelUtilsSha1Block* had the same single global context that MD5 did, so it gets the same
treatment: state, counters and block buffer now live at ctxAddr in the layout hash/sha1ctx
records off hardware. Unlike MD5, SHA-1 does not stream whole blocks through buf, which happens
to be what our sha1_update already does - so no fill-in step is needed there.
The Tlspl partition range from the last commit was too broad a cut. Hardware says only 1-6 exist,
but that recording is from user mode, and BlockAllocatorFromID deliberately maps 8 and 10 to the
user partition for a kernel-mode caller - rejecting them outright would have taken that away.
The tightened range now applies to user mode only and kernel mode keeps what it had.
threads/tls/partition also shows the answer doesn't depend on the compiled SDK version, checked
across 1.00 through 6.06, and that partition 5 is accepted - which no test had covered.
Three fixes, all of them things the new hardware tests turned up.
sceMd5Block* and sceKernelUtilsMd5Block* shared one static md5_context and ignored the context
pointer the caller passed in, with a TODO saying it would do "unless games do several MD5
concurrently". hash/md5ctx shows a real PSP keeps everything in the caller's 96 bytes and happily
runs two digests at once, so do that instead: the state, the counters and the block buffer now
live at ctxAddr in the game's own memory, in the layout the test pins down. Two interleaved
digests come out right, and a context that gets copied mid-digest carries on correctly. As a
side effect the state is now covered by savestates, which a file-static never was.
MersenneTwister masked both halves with 0x80000000 where the low half needs 0x7FFFFFFF, so
sceMt19937UInt and sceKernelUtilsMt19937UInt were returning a sequence that isn't MT19937 at
all - every number differed from hardware from the first draw. hash/mt19937ctx computes the
reference sequence itself and confirms the PSP is plain MT19937; with the mask fixed we match it
for both seeds tested. Init also twists the array immediately, as hardware does, so a context
that has been seeded but not drawn from now holds what a real one would.
sceKernelCreateTlspl accepted partitions up to 9 before falling through to the permission check.
Hardware draws the line at 6 - threads/tls/create records 7, 8, 9 and 10 all returning
ILLEGAL_ARGUMENT - so 8 and 9 were coming back ILLEGAL_PERM. Note this is genuinely different
from sceKernelCreateVpl right above it, which does let 8 and 9 through to ILLEGAL_PERM; the two
had been sharing a check that was only ever right for Vpl.
Risk worth naming: the MT19937 change alters the numbers any game gets from these calls. That's
the point - they were wrong - but a savestate taken mid-sequence will resume with a generator
that behaves differently from the one that made it.
sceKernelCreateFpl and sceKernelCreateTlspl validate blockSize * count with the
formula from hardware, but that works in 4 byte alignment while the actual
allocation uses the alignment from the options struct, which the caller picks.
A large alignment inflates each block enough that the aligned total can wrap:
we'd then allocate a small block but keep the full block count, and
sceKernelAllocateTlspl / the Fpl allocate paths hand out
address + block * alignedSize well outside the allocation.
Also give PartitionMemoryBlock::address a default. Only the savestate
constructor leaves it unset, and DoState returns early when the section is
missing, which left the destructor freeing an uninitialized address.
Initialize the wokeThreads locals that were passed by reference uninitialized in
the event flag, VPL and semaphore timeout handlers. Harmless today since the
callee only ever assigns to them, but every other use in the same files starts
at false.
Do the semaphore overflow check in 64-bit, so a large signal value can't wrap
past it into currentCount.
CoreTiming::UnscheduleEvent returns the scheduled time minus the current time,
which is negative when the event is overdue but hasn't been processed yet - the
exact situation when a wait is satisfied right around its own timeout. Only the
semaphore clamped it; everywhere else we wrote (u32)cyclesToUs(negative) into
the game's timeout variable, i.e. a huge bogus "remaining time".
Pulled the shared shape into HLEKernel::WriteRemainingTimeout so it can't drift
apart again - event flags, mbx, fpl, vpl, msgpipe, mutex, lwmutex and semaphore
all go through it now. The two thread-end sites keep their own copy since they
unschedule even when the game passed no timeout pointer, and sceUsb just gets
the clamp.
For some reason, a pointer used to allocate the heap for scePaf is not
initialized. This hacks aroung that.
Additionally zero out the specific 4-byte "category 1 alarm count" address in vsh_module.
This gets us much further.
__KernelLoadELFFromPtr creates its PSPModule and inserts it into loadedModules
before it knows whether the file is loadable, so every failure exit has to
delete the decrypt buffer, Cleanup() the module and Destroy() it. Five of the
seven did. The "unreasonable decrypted size" exit and the decompression-failure
exit just returned - leaking the buffer, and leaving a live kernel object with
its UID stuck in loadedModules for the rest of the session.
While tracing that: the fake-module path frees newptr and then runs for another
sixty lines with ptr still pointing into it. Nothing reads it today - the exits
below use head, which points into the original input rather than the copy - so
there's no use-after-free and no double free, but that's a property of the
current code rather than anything enforced. Both pointers are nulled after the
delete so a future mistake there crashes instead of reading freed heap.
And the function read the magic, and in the ~SCE branch a second word after it,
before anything established the input was that big. The non-PBP caller
guarantees it, but the PBP path computes elfSize from two offsets in the file
and passes whatever comes out, including zero. Checked at the top, before the
module object exists, so that exit needs no cleanup of its own.
pspautotests 314/314 with --graphics=software, and an EBOOT.PBP still boots.
Co-Authored-By: Claude Opus 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01GZq8ZtJmFY7bkX5FVkr3P9
* Rename LogType to Log
* Explicitly use the Log:: enum when logging. Allows for autocomplete when editing.
* Mac/ARM64 buildfix
* Do the same with the hle result log macros
* Rename the log names to mixed case while at it.
* iOS buildfix
* Qt buildfix attempt, ARM32 buildfix
Henrik RydgårdandClaude Opus 5.5
Nemoumbra
Herman Semenov