Repository navigation
feat(libkernel): implement sceKernelGetProcessAbiType - #11939
ionut27001 wants to merge 1 commit into
Conversation
|
Independent reviewer checks on exact head 3e3726d: |
MarcZX10
left a comment
MarcZX10
left a comment
There was a problem hiding this comment.
The implementation is supported by the measured own-pid, -1, and 0 cases; my local syntax, conventions, and NID checks pass. I found one contribution-rule blocker before approval: #9326 is open and changes the same docs/dev/debt/libkernel.md region. I fetched both exact heads and ran git merge-tree against their common base; the merge conflicts in that debt file (the shared CMake/API files merge cleanly). Please coordinate the agreed landing order with #9326 and record it in Depends on/the overlap note, or rebase and update the report after one lands. The hosted Windows test and policy checks are still running; this request is specifically for the confirmed documentation conflict.
|
Other open pull requests touch the same code. Merging one will break the other, or both implement the same thing. If this one builds on them, list them in Depends on.
Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.
|
Returns 1 for the own pid and for -1, and 0 for pid 0, as a payload process on a PS5 (firmware 12.70) read (p86). Any other pid throws: another payload process got 0x80020003 for a stack pointer and an unspecified first argument (p85), and other existing pids were not tried. Adds a guest test and cuts the debt entry to what is still open.
3e3726d to
4fddc50
Compare
|
macOS run of head
AI-assisted: yes |
What
Implements
sceKernelGetProcessAbiType(int pid)incore/libs/prx/libkernel/System/src/Process.cppand removes the throwing declaration fromUnimplemented.cpp. The own pid and -1 return 1, pid 0 returns 0, and any other pid throws as unmodelled; the second argument is ignored here, and whether the export uses one is unknown. The debt entry says what is still open: other existing pids, which process pid 0 names, the meaning of the values, and a title process.Measured in two payload process runs on a PS5 (firmware 12.70), p85 and p86 (
E:/CLI/ps5-oracle/results/20261027T103808Z_p85_libkernel/result.log,E:/CLI/ps5-oracle/results/20261027T104816Z_p86_libkernel/result.log), result inrax:getpid()(108): 1 with a 16-byte0xa5buffer and with null as the second argument (p86 linescase abi_out pid=108 rax=0x1 buf=a5a5a5a5a5a5a5a5a5a5a5a5a5a5a5a5,case abi_null pid=108 rax=0x1)case abi_out pid=-1 rax=0x1 buf=a5a5...,case abi_null pid=-1 rax=0x1)case abi_out pid=0 rax=0 buf=a5a5...,case abi_null pid=0 rax=0)0x80020003(p85 v2case abi_buf rax=0x80020003 buf=0000...,case abi_noarg rax=0x80020003)abi_outcases the 16-byte0xa5buffer passed as the second argument was not written.Other existing pids were not tried, so they throw here instead of returning
SCE_KERNEL_ERROR_ESRCH, which was seen only for a stack pointer and an unspecified first argument (p85). PS5_MihawkOS (PPSA99170) imports this export.Tested
Windows 11, WinLibs GCC 15.2.0 (Release, Ninja), commit 4fddc50 on main 75f3702 (the revert check below ran on ea35db574d, the same code on main 431a515):
ctest -R ^guest_process_abi_type$passed twice.sceKernelGetProcessAbiType not implemented).Checklist
main; no other open PR implements the same functionsNotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file.mdfiles or images added to the repository (attach them to this PR)docs/dev/debt/libkernel.mdconflict; this lands after it and rebases)