Skip to content

fix/futex: forget to .await in futex_exit - #63

Closed
greatbridf wants to merge 153 commits into
refactor/mm/pte-mapfrom
fix/pthread-test
Closed

fix/futex: forget to .await in futex_exit#63
greatbridf wants to merge 153 commits into
refactor/mm/pte-mapfrom
fix/pthread-test

Conversation

@greatbridf

@greatbridf greatbridf commented Oct 19, 2026

Copy link
Copy Markdown
Owner

Summary

Testcase /mnt/pthread_test hangs after the threads are joined (issue #61): the process reaches the end of the test but never exits, freezing forever in the clear-tid futex wait.

Root cause: futex_exit() in src/kernel/task/futex.rs called do_futex_exit(thread) without awaiting the returned future:

let _ = do_futex_exit(thread);

Because let _ = <future> merely drops the future, do_futex_exit() was never executed at all. As a result the thread was never detached from its futex state and the clear-tid futex wakeups were never performed, so the joiner kept waiting forever.

Fix

One-line fix: actually await the future so do_futex_exit() runs.

 pub async fn futex_exit(thread: &Thread) {
     // We don't care about any error happened inside.
     // If they've set up a wrong pointer, good luck to them...
-    let _ = do_futex_exit(thread);
+    let _ = do_futex_exit(thread).await;
 }

Fixes a regression from b6d54d6 ("mm, proc: add an exited thread reaper").

Verification

Reproduced and verified with make run MODE=release, running /mnt/pthread_test in the guest:

  • Before: the test hangs forever after pthread_join.
  • After: the test prints Final counter value: 20000 and the process exits normally.

Fixes #61

@greatbridf
vfs: rework the vfs subsystem with async
19f9ef3 
Inode and superblock rework:

Remove old Inode trait as it used to undertake too much responsibility.
The new method use three new traits: InodeOps is used to acquire generic
inode attributes. InodeFileOps and InodeDirOps handle file and directory
requests respectively. All the three have async fn trait methods and
don't need to be ?Sized. Then, we implement Inode, InodeFile and
InodeDir for the implementors of the three "Ops" traits, erasing their
actual type and provide generic dyn interface by wrapping the futures in
boxes. We should provide an io worker? or some io context with an
allocator for futures to reduce the overhead of io requests, or come up
with some better ideas?

For inode usage, we introduce InodeRef and InodeUse. InodeRef is a
simple wrapper for Weak<impl Inode> and InodeUse for Arc<impl Inode>.
This helps us use them better as we can't define impls for Arc<dyn
Inode>'s as they are foreign types. We also provide some more helper
methods for them.

After the change, we don't impose ANY structural restriction except for
the spinlock wrapped InodeInfo. The InodeInfo struct design might need
rethinking but the current implementation seems to be fine aside from
unnecessary locks when we try to access some of its fields but this
shouldn't be a VERY big problem and very urgent...

Similar changes are also made to superblock traits and types. But for
the superblock objects, we use a SuperBlockComplex struct to store
common fields such as whether the superblock is read only, their device
id and so on. Also the structs have a superblock rwsem inside. But we
haven't decided how to use that (such as whether we should acquire the
lock and pass it to the inode methods) and even whether they should
exist and be there. This will need further thinking so we put this off
for now...

Filesystem rework:

Rework tmpfs, fatfs and procfs with the new technology mentioned above,
leaving the old ext4 unchanged. The current implementation of ext4 uses
some "random" library from the "camp". Its code hasn't been fully
reviewed for time reasons but seems to be rather "problematic"... We
might rewrite the whole module later and the page cache subsystem
requires fully reworking as well. So we put this work off as well.

Block device and other parts rework:

Wraps PageCacheBackend, MountCreator and BlockRequestQueue with
async_trait to provide dyn compatible async functions. Dentry walking
functions are also put to the heap since they are recursive functions...
This has similar problems to the inode traits, ugly solution. Further
optimization is required.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
vfs, rcu: rework path walking with new rcu syntax
ea21223 
The old path walking algorithm requires recursion, which is not
supported in async rust. So we boxed them all as a temporary solution in
previous commits. This would introduce mass overhead even for fast path
walks just because we might sleep in `readlink()` and `lookup()` calls.

The new proposed method is to break the walk into several phases similar
to that in Linux: RCU walk and REF walk. The RCU walk will hold the RCU
lock and never blocks so the function itself can be non-async. If we hit
non-present dentries, we will fallback to REF walk. In REF walks, we
clone the Arcs and consult to the VFS layer for an accurate answer.

Note that in both the two methods mentioned above, symlinks are not
handled and will be returned directly with all path components left
untouched. We have a dedicated async function to follow the symlinks by
recursive calling the walk function. This can be slow and won't be
called frequently. So we wrapped the function with `Box::pin()` to break
the recursion chain. After the symlink resolution is done, we return to
the original position and continue the walk.

We found that the association of an inode to a dentry is one way. So the
`data` RCUPointer field is actually unnecessary and we can use the atomic
dentry type to sync readers with the writer. This way we can eliminate
`DentryData` allocations and improve performance.

We also introduced a new RCU read lock syntax. In the RCU walk mentioned
above, we need to store dentry references protected by some RCU read lock.
With the old syntax, we can't express the lifetime associtated by the
common RCU read lock. The new syntax provides a `rcu_read_lock()` method
to acquire the RCU read lock. The lock returned has a associated lifetime
so we can use it in the RCU session.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style, vfs: remove unused imports and InodeRef
49a4903 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
vfs: fix debug print of Mode structs
632f1c7 
If we have invalid format, we should print a None instead of panicking.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mem, paging: introduce page locks and exclusive pages
210a669 
Introduce the new page locking mechanism to ensure exclusiveness of
pages when we access them. The underlying locks are not implemented yet
for now because we will change the paging structs in the following few
patches.

Introduce a new `PageExcl` struct representing a page that conforms with
Rust's ownership rules. The page owned exclusively can be accessed
without taking page locks.

Remove the `MemoryBlock` structs as they are not easy to use and
represent barely no semantic meanings.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
tls: rework of arch's UserTLS design
9223243 
Separate old UserTLS into UserTLS and UserTLSDescriptor.

UserTLS is for threads to hold infomation about its storage. Descriptors
are used in clone syscalls.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mem, slab: rework the slab system
841bb37 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mem, buddy: rework the buddy system
f17236f 
Introduce `Zone`s: a Zone is a region of physical memory that all share
the same NUMA node. The Zone will hold all RawPage structs. Buddy
allocator will now store a reference to the zone that hold all its
pages. Thus, we make the buddy allocator independent of underlying
physical page frame management framework.

Remove unnecessary page flags and structs.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
vfs: rework of inode and page cache system
45268e7 
- Bump rust compiler version to nightly-2026-01-09.
- Inode rework: add a generic Inode struct.
- Add a macro to help function tweaks.
- PageCache rework: reduce complexity and try to decouple.
- Adapt fat32, tmpfs to the new page cache system.
- Change the way we process mapped pages and load ELF executables.
- Refine handling flags in `MMArea::handle_mmap`.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mem: introduce new Folio abstraction
913f71f 
- Remove struct `Page` and add `Folio`s to represent adjacent pages.
- Introduce `Zone`s similar to that in Linux. Looking forward to
  removing all occurrence of `ArchPhysAccess` and so on.
- Adapt existing code to new `Folio` interface in a dirty and rough way.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
script: add a script to help translate stacktraces
cfc959e 
- Add script/backtrace to translate backtraces.
- Add a cut sign in the kernel panic routine to indicate the start of
  stack backtrace.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style: fix or suppress warnings
b0c8ef4 
No functional changes

- Add `{extern_,}symbol_addr` macro to retrieve symbol address.
- Remove manual impl Send and Sync for RawPage
- Make elided '_ lifetimes in return types explicit
- Suppress unused warnings by allowing them
- Remove really unused functions
- Refactor `println_trace` macro to suppress unused variable warnings

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
user, init: update riscv64 init script
bc84f0b 
Make print message prettier. Also panic when we have an error.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
chore: remove and ignore vscode settings from svc
47d8906 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm, proc: add an exited thread reaper
b6d54d6 
- thd.exit() will set thd.dead and send it to the reaper
- delay the release of process mm until we reap it
- extract futex logic out of exit and exec routine

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
proc: rewrite process list organization
997edb0 
- Use intrusive lists to store and organize the process hierarchy.
- Remove `FileArray::open_console()`. Do it in the init script instead.
- Fix open logic: acquire controlling terminals only if O_NOCTTY is not
  set. Put this into TerminalFile::open().
- Send SIGHUP and then SIGCONT to foreground pgroup procs when the
  controlling terminal is dropped.
- Set the controlling terminal of sessions in Terminal.
- Limit max line width to 80. Format some codes.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style: reformat the files related to next patches
2392ac1 
Reformat the files with new the format style to make the real changes
clearer.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
riscv64, hal: simplify ArchMemory::free_ram implementation
6a00786 
We already have `FDT.present_ram().free_ram()`. Remove the impl in
`ArchMemory` to avoid confusion.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style: add helper macros to retrieve symbol constants
53ae185 
- Add `extern_symbol_value` to retrieve far relative symbol values.
- Get `BSS_LENGTH` and `__kernel_end` using `extern_symbol_addr`.
- Get `_ap_start` using `extern_symbol_value`.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style: reformat files
a117be1 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
riscv64, linker: make sure vdso lies inside .data
4d272fe 
With current ldscript, linkers will put vdso data after `__kernel_end`,
which is buggy since we use the symbol to indicate the end of our kernel
image and newly allocated pages may overwrite those positions.

Change by place the vdso inside REGION_DATA. Remove old VDSO memory
region. Align .data section end to page size border.

Add a helper macro to retrieve .vdso section symbol addresses.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
riscv64: rewrite FDT and present free memory parsing
5814c00 
Strip out the memory used by the kernel and FDT data out of free memory
block returned by FDT.

Closes: #54 ("Random kernel freezing on process creation / exiting")
Fixes: 4351cf5 ("partial work: fix riscv64 bootstrap")
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
sysinit: pointee type should be u8 when using ptr::write_bytes
20b12df 
Using *const (), no bytes are written to the position, which might
result in uninitialized memory access.

Fixes: ebd3d12 ("change(x86): optimize bootstrap code, remove kinit.cpp")
Fixes: 191877a ("feat(hal): impl basic single hart bootstrap for riscv64")
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style: reformat file
74c7e0a 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
hal, mm: alloc basic folios from low to high addr
6a0f703 
This can shorten qemu memory map.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
script: refactor the build image script
3842078 
Rearrange code to pass args instead of by envs, no functionality
changes.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
test: rework test script and github workflow
c7c5cda 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
ci: install expect and update printf usage
bcc4158 
- printf -v is non-standard, use var=$(printf) instead.
- expect is not installed in the runner. install it manually

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
script, build-img: add script folder for each platforms
38a3272 
Add PLATFORM in Makefile to indicate the actual system setup. Put build
image script of each platforms to dedicated folders. This makes creating
test setups easier.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style, hal: rename x86 trap.rs to trap/mod.rs
d4380e0 
No functional change.

Signed-off-by: greatbridf <greatbridf@icloud.com>
greatbridf added 27 commits March 17, 2026 23:34
@greatbridf
mm: rename insert_new() to insert() and remove the old one
a654701 
There used to be two `insert()` impls due to historical reasons that we
need to operate on raw `MemArea`s. We hope to remove raw MemArea
handles, so do the cleanups.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: add helper function to create anon mapping
b8af819 
Add a helper to create `AnonMapping` struct.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: update mappings' offsets when splitting areas
74d6122 
I notice that when we isolate the areas for protecting or unmapping, we
forgot to update the underlying mappings' offsets. This may cause wrong
offset passed to read() in fault handlers, despite we haven't observed
the problem yet.

Fix this by refactoring the isolate() logic: we recreate every area
within the isolate range, call newly defined split functions and handle
the offset correctly.

Fixes: 9b18fc8 ("WIP: mem: remove unused MemArea member functions")
Fixes: f37b288 ("WIP: mem: reimpl `MMList::protect` using `isolate()`")
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: create MemAreas using new() func
f53d61c 
Replace all MemArea construction with `Self::new()` calls so we can
modify it more easily.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: make handle_cow() private
0303cf3 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: add a helper to create new MMLists
5ff19c9 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: rework MMList creation while doing clones
543d912 
I extract the part cloning the area list and processing cloned areas (
particularly creating new PTEs in the new pgtable and modifying the old
ones in the old table) into a separate function where shared areas and
private areas are handled differently. Then, creation of new MMLists can
be put off till the area list and pgtable are ready.

`MemArea::deep_clone()` exists only for creating cloned MMLists, thus
can be removed.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: add a macro to impl Deref for transparent wrappers
6f8e43c 
For repr(transparent) types that need to impl Deref trait and returns
&self.0, we introduce a macro called TransparentDeref to derive Deref
for them and use the macro.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: don't expose AnonMapping outside of mem module
c4dc884 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: move PageOffset definition to mod types
9c6626e 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: don't expose FileMapping outside of mem module
68a0200 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: use PageOffset when creating file mappings
b9c4706 
Add a helper function to ensure alignment and create PageOffset. Then
use PageOffset and the helper internally while creating file mappings.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: use PageOffset inside FileMapping and fault handlers
750b909 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: rework Mapping to divide it into 3 classes
8552663 
There are in fact 3 types of mapping:
- File private: fault in from underlying files and modifications are not
  carried back to the file. After faulting in, the folios are associated
  to some anonymous mapping and CoW on forks.
- File shared: fault in just as the previous one but all folios from the
  same mapping are shared from the page cache.
- Anonymous: Always private. Shared anonymous mappings should use tmpfs
  + shared file mappings.

Now add a variant to the enum Mapping, and unify the fault handling
logic related to the mapping. Introduce new helpers to create the
mappings.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: don't read the PTE twice setting CoW
0c992f5 
The PTE is read to get the pfn and attr. We can reuse it if the PTE is
not present and we need to copy the PTE as is.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: associate newly alloced anon folios with the anon mapping
c7b2600 
When handling missing anon or writings to private file PTEs, we need to
alloc a new anon folio and add a mapping. Pass the anon mapping to the
function and alloc the anon folio with the mapping.

It's worth noticing that handling CoW PTEs also needs this
theoretically. This should be done in next few patches.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: use the anon mapping to alloc folio in CoW handling
d56bf6f 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm, page_cache: remove alloc_zeroed()
92563ea 
This function was used by some fs to allocate blank pages. Remove this
since it's no longer used.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: introduce the AnonFolio
42c3291 
For anonymous pages mapped to user space, introduce a new wrapper around
the struct Folio named AnonFolio, which is also a mappable folio type
similar to folios for some PageCache. Before mapping a page to some
area, associate the folio with the AnonMapping using `add_folio()`.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
style, vfs: reformat fat32.rs, file.rs and inode.rs
34d60c0 
No functional change.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
vfs, inode: remove {read,write}_direct
f9aa5aa 
The two functions are unused for long so remove them. If they are needed
later, add them back manually then.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm, vfs: change read_page() arg type to FolioOwned
99598a2 
It doesn't make sense for read_page() to accept a &mut CachePage. All it
needs to do is to read and copy data from the underlying filesystem to
the given folio. So change that to FolioOwned, and allocate and pass a
FolioOwned to the function.

After the change, we don't need to create CachePages outsides the page
cache module. Change CachePage::new() to be private.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
hal, riscv: support user bit of table PTEs
3d5ef32 
The standard says that PA_U bit should be cleared and reserved for
future use, so we used to ignore the USER flag for table attributes.
This results in memory leaks of page table folios as they're never
released (we only release page table folios that is both present and
for userspace usage, but USER bit is ignored when set).

The workaround is to reuse the PA_COW bit to indicate the entry is for
userspace usage. An entry can be CoW only if it's a leaf entry. So both
PA_COW and PA_MMAP can be reused. We'll take the CoW bit as it has a
smaller number...

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: refactor set PTE CoW logic, removing PageTable helpers
cf4805a 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
mm: introduce MapFolio for folios mapped to page tables
6feabcf 
Introduce MapFolio, a wrapper around Folio for folios whose contents can
be mapped into page tables, along with the per-folio data backing it.
That data, currently a mapcount counting how many page-table entries
reference the folio, is stored inline in RawPage's PageData union and is
guarded by the new MAPPED page flag. A folio is turned mappable by
FolioOwned::into_mappable(), which initializes the data and publishes
the flag with release semantics; Folio::drop() tears the data down again
on the last reference.

AnonFolio and CachePage now wrap MapFolio instead of a bare Folio, so
anonymous and file-cache folios become mappable at creation time. The
free add_mapping()/duplicate_mapping()/remove_mapping() helpers of
mm_list::mapping are folded into MapFolio methods that bump or drop the
mapcount while handing a refcounted PFN to (or retrieving one from) the
page tables.

The CoW path in MemArea::handle_cow() now decides whether the faulted
page must be copied by checking whether its mapcount is one
(MapFolio::is_exclusive()) instead of reading the shared refcount after
removing the mapping. The refcount conflated page-table mappings with
the folio's other owners, and the old check had to remove the mapping,
inspect the refcount, and re-add it; mapcount tracks mappings directly,
so an exclusive page can simply be made writable in place without
refcount churn. Checking is not yet race-free against concurrent faults
on the same page -- locking the folio is left as an XXX comment.

Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
chore: update gitignore for vim coc settings
59041aa 
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf
fix/futex: forget to .await in futex_exit
bf9f278 
Testcase `pthread_test` freezes when waiting for clear tid futex
wakeups. It turns out to be that we forget to call .await in
`futex_exit()` so `do_futex_exit()` is never called.

Fix by appending the missing `.await`.

Fixes: b6d54d6 ("mm, proc: add an exited thread reaper")
Signed-off-by: greatbridf <greatbridf@icloud.com>
@greatbridf greatbridf closed this Oct 27, 2026
@greatbridf
greatbridf deleted the fix/pthread-test branch October 27, 2026 16:20
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@greatbridf