See the "Redox OS/Support" channel. This commit temporarily fixes the
issue by rolling back the changes to preload the executable.
Signed-off-by: Anhad Singh <andypython@protonmail.com>
Previously, this was implemented incorrectly because `tcb.mspace` was
set just *before* jumping to the program's entry point. However,
`alloc_init` is called when the init functions are run for `libc.so`, so
`tcb.mspace` must be initialized *before* the init functions are
run.
Signed-off-by: Anhad Singh <andypython@protonmail.com>
There are a few functions where I allowed the lint. The lower level and
FFI nature of relibc means that there are areas with a lot of unsafe.
Some functions are basically long blocks of unsafe or contain lots of
small blocks of unsafe. unsafe_op_in_unsafe_fn doesn't add clarity to
these functions. Instead, it reduces readability by adding an indent or
small "unsafe { .. }" clutter.
Uses a `spin::Mutex` for the global static `_r_debug` instance to remove
warning about a global mutable static.
Guarantees exclusive mutable access to the `_r_debug` variable.
Currently, how ld.so errors are handled by functions in dlfcn is wrong.
dlopen calls `linker.load_library` which returns a
`Result<ObjectHandler, goblin::Error>`. Now `goblin::Error` may have
variants that are heap allocated. For example:
`Error::Malformed(format!("invalid path: '{}': {}", path, err))`. The
error string would be allocated by ld.so's allocator but will be dropped
inside libc. This is UB.
After this patch, we now return a custom `DlError` instead. To get more
information about the error, `LD_DEBUG=all` can be set.
Signed-off-by: Anhad Singh <andypython@protonmail.com>
In the next big refactor (next PR), all of the platform functionality
used by both relibc and ld.so will be moved into a `platform`/`sysdeps`
crate and then ld.so would be moved out of relibc and not link with it.
I think doing it in a seperate PR would make it more managable, as when
I did half of it, the diff was pretty huge and that way it would be
easier to review too :)
Signed-off-by: Anhad Singh <andypython@protonmail.com>
Wildan M