Stefanos Kornilios Mitsis Poiitidis
Stefanos Kornilios Mitsis Poiitidis
Currently the Visual Debugger is non functional, and it depends on debug interfaces that are no longer implemented and miss features like multiple thread support. A large part of those...
Currently we'll happily translate pages without PROT_EXEC or not mapped at all. We need to only permit code to run in guest mapped pages, and only ones with PROT_EXEC, or...
Follow up from #1842, We can use much finer grained mman / translation locks, vs the lock the world approach taken there. - Use an inverval lock tree - Lock...
An internal backlog of tests (mostly for me), so the tracker is not spammed with too many of these tickets. If you have any corner cases that you want tested,...
Some commonly used anti-patterns that make our lifes harder while working with FEX - Re-use of class names with different namespaces (makes navigation, search and replace, debugging hard) - Overly,...
`GetReg` for i32 `GetReg` for i64 `GetSrc` for FPR `GetSrcPair` for FPR pairs? We need a uniform way, across all backends, for RA to work
Dump shows like ``` %ssa18 i64 = EntrypointOffset #0x32, u8:Tmp:RegisterSize (%ssa19 i64) Break %ssa18 i64, {0.11.0.128} ``` The paser can't parse this
- Inspect/Visualise index fixes - Inspect code caches - Show IR disasm - Show OBJ disasm
- Missmatching argument vs env values - Two layers of mappings that we don't really need to go through - Factor common code out of FEXLoader, IRLoader, TestHarness, etc -...
Recently, a PR removed this, and broke useful backtraces for debugging ``` Switching to thread 2 (Thread 0x7fff5eac7640 (LWP 1501682))] #0 syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38 38 ../sysdeps/unix/sysv/linux/x86_64/syscall.S: No such file...