docs: clarify runtime code generation limits for --to-intel (#4257)
This commit is contained in:
parent
98c07d8e77
commit
656c81ac8b
|
|
@ -33,7 +33,7 @@ macOS output (x86-64 Mach-O, which runs under Rosetta on Apple silicon):
|
|||
relinker --macos source/input.elf eboot
|
||||
```
|
||||
|
||||
Add `--to-intel` for Intel hosts, or `--to-rosetta` for Rosetta; a title that runs under Rosetta on Apple silicon is relinked with `--macos --to-rosetta`. The output format defaults to Linux ELF regardless of the filename; `.exe` alone does not select Windows.
|
||||
Add `--to-intel` for Intel hosts, or `--to-rosetta` for Rosetta; a title that runs under Rosetta on Apple silicon is relinked with `--macos --to-rosetta`. `--to-intel` rewrites AMD-specific instructions found in the executable and bundled modules; it cannot translate AMD-specific instructions generated by a title at runtime, so such titles require an AMD host. The output format defaults to Linux ELF regardless of the filename; `.exe` alone does not select Windows.
|
||||
|
||||
The executable, registry, icon and desktop entry output files must not refer to the input executable or a bundled module being converted, including through a hard link or symbolic link. These collisions fail before any output is written. An existing output file can be replaced if it is separate from those inputs.
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue