vasilito/RedBear-OS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
2,603 Commits 12 Branches 0 Tags
d3073fdad402b81ee831fe845928166b495a88c3
Commit Graph

206 Commits

Author SHA1 Message Date
Anhad Singh d3073fdad4 feat(ld.so): more debug info for scopes 
When scopes is set for `LD_DEBUG`, more debug information is displayed
now,

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-19 00:11:51 +11:00
Anhad Singh 99449094d4 chore: formatting 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-19 00:03:07 +11:00
Anhad Singh 4a08220664 fix(ld.so): copy relocations 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-18 23:51:53 +11:00
Anhad Singh e3dc5d0617 chore(dso): remove unused members 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-18 21:20:10 +11:00
Anhad Singh 3440985729 fix(dso): 32-bit 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-15 21:32:40 +11:00
Anhad Singh 07ff79069e feat(ld.so): add support for DT_RELR 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-15 13:49:27 +11:00
Jeremy Soller a3c12e27e4 Fix build for i686 2025-01-14 18:01:26 -07:00
Anhad Singh b9e0505715 fix(ld.so): less locks 
Acquire the global scope lock at the start of relocate.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-13 23:48:03 +11:00
Jeremy Soller af50d18d16 Return error instead of panic when load_library cannot find root DSO 2025-01-09 13:33:29 -07:00
Jeremy Soller 02520b194e Fix compilation with trace feature 2025-01-09 12:23:23 -07:00
Anhad Singh aae27c5e0c feat(ld.so): switch to object crate 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-08 16:50:47 +11:00
Anhad Singh e49072655c feat(ld.so): partially switch to object 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-08 16:50:10 +11:00
Anhad Singh e219f9768e feat(ld.so): GNU hash table 
Note that the ELF specification mandates the presence of either the GNU
hash table or SysV hash table. See Figure 2-7. Dynamic Array Tags.

Also both can be present, though that is not currently handled.

BLAZINGLY FAST!

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2025-01-08 16:48:26 +11:00
Anhad Singh bfe1c0b545 fix(ld.so): redox build 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 19:30:37 +11:00
Anhad Singh 01e6eda258 misc(dlfcn): cleanup 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 19:03:05 +11:00
Anhad Singh 42d884c6a3 fix(ld.so): errors UB 
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>
 2024-12-31 18:35:20 +11:00
Anhad Singh 9cf72fd0b2 feat(ld.so): add all option for LD_DEBUG 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 18:00:45 +11:00
Anhad Singh 60f700fb50 chore(linker): remove todo 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 15:26:23 +11:00
Anhad Singh c25292ae21 misc(ld.so): clippy fixes 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 15:25:40 +11:00
Anhad Singh cfe89f828a feat: rwlock instead of mutex for global scope 
* Rwlock -> InnerRwLock
* RwLock makes use of InnerRwLock.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-31 00:54:52 +11:00
Anhad Singh 2be4316aec feat(dlfcn): return DSO as the handle 
Fine because:
> The value of this handle should not be interpreted in any way by the
> caller.
From: https://pubs.opengroup.org/onlinepubs/009696599/functions/dlopen.html

* Add tests for unload
* Implement NOLOAD so we can test that the loaded object has been
  destroyed.
* Minor cleanup.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-25 16:17:51 +11:00
Anhad Singh 1612c78774 fix(ld.so): disable lazy binding on non-x86_64 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 023a005688 fix(ld.so): non-pie lazy 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 1db68ab978 feat(ld.so): scopes 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 13d819ac89 misc(ld.so: __plt_resolve_inner): println -> trace 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh e3587017f6 fix(ld.so): stub plt stub for non-x86_64 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 0a0ecb018f fix(ld.so: adjust_stack): remove dead code 
Does not do anything.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh c6fd0517ac feat(ld.so: config): handle LD_BIND_NOW 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 5ae1b8e974 feat(ld.so): lazy binding 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 30970e111a feat(ld.so): enable warnings just for ld.so 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 959982b8a4 fix(ld.so): use Strtab::get_at() 
> use of deprecated method `goblin::strtab::Strtab::<'a>::get`: Bad performance, use get_at() instead

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh cc5e0ceebc fix(ld.so): function signature 
No longer expects a self TLS.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh eb82f9afeb misc(lint): unused variables and imports 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 500848bc76 fix(ld.so): remove unreachable loop {} 
_exit() is noreturn

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-24 20:51:53 +11:00
Anhad Singh 39c8cc864d fix(ci): formatting 
Was my fault.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-17 03:14:33 +11:00
Anhad Singh de9e8a6177 feat(ld.so): config 
Currently supported options are `LD_DEBUG` and `LD_LIBRARY_PATH`.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-17 01:07:17 +11:00
Anhad Singh 9928091aad fix(ci) 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-12 16:11:42 +11:00
Anhad Singh d3bab6eeee fix(redox: ld.so): unmap old TCB 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-12 16:00:11 +11:00
Anhad Singh 8c1722d583 fix(ld.so): redox 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-11 16:51:51 +11:00
Anhad Singh de00d6c6b9 fix(ld.so): temporary fix 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-05 23:34:16 +11:00
Anhad Singh 32077ee341 fix(ld.so): set tls module id to 0 if no tls 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-05 23:33:41 +11:00
Anhad Singh c286ad2868 fix(ld.so): do not depend on TLS at all 
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>
 2024-12-05 23:11:59 +11:00
Anhad Singh 55ba8d8f6b misc(load_objs_recursive): do not add empty masters 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-05 15:00:35 +11:00
Anhad Singh 1c846f8013 fix comment 
x86 and x86_64 are below TP and others are above TP

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-12-05 15:00:35 +11:00
Anhad Singh f6fa4e8246 feat(ld_so): setup sighandler 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-11-30 00:02:33 +11:00
Anhad Singh e9f3147e3b fix(tcb): do not use Vec::from_raw_parts 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-11-29 20:55:02 +11:00
Anhad Singh 1a7726541a fix TCB and TLS 
Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-11-29 18:40:16 +11:00
Anhad Singh c88eb3d5d0 Revert "fix(linker): deref uninit TCB" 
This reverts commit 37dbf5cbb2.
 2024-11-29 00:03:09 +11:00
Anhad Singh 37dbf5cbb2 fix(linker): deref uninit TCB 
The bug is described for `x86_64` and Linux but is the same for other architectures and on Redox.

`Tcb::current()` is used to retrieve the current TCB, which is done by by reading `fs:[0x10]`. The TCB layout describes that at offset `0x10`, there is `tcb_ptr: *mut GenericTcb<...>`, which is nothing more but a pointer to itself.

This is fine as otherwise a system call would be required to get the TCB (`arch_prctl(ARCH_GET_FS)` on Linux).

However, this is problematic as the function may be called when the FS base is not set, and in that case the expected output of the function should be [`None`], but we don't currently handle that.

To fix this, any code paths that maybe call this function on an uninitialized TCB are be switched to call `current_slow()`. Which just uses `arch_prctl(ARCH_GET_FS)` to get the FS base and check if it's non-zero.

Signed-off-by: Anhad Singh <andypython@protonmail.com>
 2024-11-25 16:09:06 +11:00
Peter Limkilde Svendsen 06fec0b843 Add various basic docstrings 2024-10-27 23:25:57 +01:00