From 4fc87c240b8f30e22b7ebaae29d57105589e1c0b Mon Sep 17 00:00:00 2001 From: Carlos Llamas Date: Tue, 10 Feb 2026 23:28:20 +0000 Subject: rust_binder: fix oneway spam detection The spam detection logic in TreeRange was executed before the current request was inserted into the tree. So the new request was not being factored in the spam calculation. Fix this by moving the logic after the new range has been inserted. Also, the detection logic for ArrayRange was missing altogether which meant large spamming transactions could get away without being detected. Fix this by implementing an equivalent low_oneway_space() in ArrayRange. Note that I looked into centralizing this logic in RangeAllocator but iterating through 'state' and 'size' got a bit too complicated (for me) and I abandoned this effort. Cc: stable Cc: Alice Ryhl Fixes: eafedbc7c050 ("rust_binder: add Rust Binder driver") Signed-off-by: Carlos Llamas Reviewed-by: Alice Ryhl Link: https://patch.msgid.link/20260210232949.3770644-1-cmllamas@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/range_alloc/array.rs | 35 +++++++++++++++++++++++++++-- drivers/android/binder/range_alloc/mod.rs | 4 ++-- drivers/android/binder/range_alloc/tree.rs | 18 +++++++-------- 3 files changed, 44 insertions(+), 13 deletions(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/range_alloc/array.rs b/drivers/android/binder/range_alloc/array.rs index 07e1dec2ce63..ada1d1b4302e 100644 --- a/drivers/android/binder/range_alloc/array.rs +++ b/drivers/android/binder/range_alloc/array.rs @@ -118,7 +118,7 @@ impl ArrayRangeAllocator { size: usize, is_oneway: bool, pid: Pid, - ) -> Result { + ) -> Result<(usize, bool)> { // Compute new value of free_oneway_space, which is set only on success. let new_oneway_space = if is_oneway { match self.free_oneway_space.checked_sub(size) { @@ -146,7 +146,38 @@ impl ArrayRangeAllocator { .ok() .unwrap(); - Ok(insert_at_offset) + // Start detecting spammers once we have less than 20% + // of async space left (which is less than 10% of total + // buffer size). + // + // (This will short-circuit, so `low_oneway_space` is + // only called when necessary.) + let oneway_spam_detected = + is_oneway && new_oneway_space < self.size / 10 && self.low_oneway_space(pid); + + Ok((insert_at_offset, oneway_spam_detected)) + } + + /// Find the amount and size of buffers allocated by the current caller. + /// + /// The idea is that once we cross the threshold, whoever is responsible + /// for the low async space is likely to try to send another async transaction, + /// and at some point we'll catch them in the act. This is more efficient + /// than keeping a map per pid. + fn low_oneway_space(&self, calling_pid: Pid) -> bool { + let mut total_alloc_size = 0; + let mut num_buffers = 0; + + // Warn if this pid has more than 50 transactions, or more than 50% of + // async space (which is 25% of total buffer size). Oneway spam is only + // detected when the threshold is exceeded. + for range in &self.ranges { + if range.state.is_oneway() && range.state.pid() == calling_pid { + total_alloc_size += range.size; + num_buffers += 1; + } + } + num_buffers > 50 || total_alloc_size > self.size / 4 } pub(crate) fn reservation_abort(&mut self, offset: usize) -> Result { diff --git a/drivers/android/binder/range_alloc/mod.rs b/drivers/android/binder/range_alloc/mod.rs index 2301e2bc1a1f..1f4734468ff1 100644 --- a/drivers/android/binder/range_alloc/mod.rs +++ b/drivers/android/binder/range_alloc/mod.rs @@ -188,11 +188,11 @@ impl RangeAllocator { self.reserve_new(args) } Impl::Array(array) => { - let offset = + let (offset, oneway_spam_detected) = array.reserve_new(args.debug_id, args.size, args.is_oneway, args.pid)?; Ok(ReserveNew::Success(ReserveNewSuccess { offset, - oneway_spam_detected: false, + oneway_spam_detected, _empty_array_alloc: args.empty_array_alloc, _new_tree_alloc: args.new_tree_alloc, _tree_alloc: args.tree_alloc, diff --git a/drivers/android/binder/range_alloc/tree.rs b/drivers/android/binder/range_alloc/tree.rs index 838fdd2b47ea..48796fcdb362 100644 --- a/drivers/android/binder/range_alloc/tree.rs +++ b/drivers/android/binder/range_alloc/tree.rs @@ -164,15 +164,6 @@ impl TreeRangeAllocator { self.free_oneway_space }; - // Start detecting spammers once we have less than 20% - // of async space left (which is less than 10% of total - // buffer size). - // - // (This will short-circut, so `low_oneway_space` is - // only called when necessary.) - let oneway_spam_detected = - is_oneway && new_oneway_space < self.size / 10 && self.low_oneway_space(pid); - let (found_size, found_off, tree_node, free_tree_node) = match self.find_best_match(size) { None => { pr_warn!("ENOSPC from range_alloc.reserve_new - size: {}", size); @@ -203,6 +194,15 @@ impl TreeRangeAllocator { self.free_tree.insert(free_tree_node); } + // Start detecting spammers once we have less than 20% + // of async space left (which is less than 10% of total + // buffer size). + // + // (This will short-circuit, so `low_oneway_space` is + // only called when necessary.) + let oneway_spam_detected = + is_oneway && new_oneway_space < self.size / 10 && self.low_oneway_space(pid); + Ok((found_off, oneway_spam_detected)) } -- cgit v1.2.3 From 8ef2c15aeae07647f530d30f6daaf79eb801bcd1 Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Wed, 18 Feb 2026 11:53:26 +0000 Subject: rust_binder: check ownership before using vma When installing missing pages (or zapping them), Rust Binder will look up the vma in the mm by address, and then call vm_insert_page (or zap_page_range_single). However, if the vma is closed and replaced with a different vma at the same address, this can lead to Rust Binder installing pages into the wrong vma. By installing the page into a writable vma, it becomes possible to write to your own binder pages, which are normally read-only. Although you're not supposed to be able to write to those pages, the intent behind the design of Rust Binder is that even if you get that ability, it should not lead to anything bad. Unfortunately, due to another bug, that is not the case. To fix this, store a pointer in vm_private_data and check that the vma returned by vma_lookup() has the right vm_ops and vm_private_data before trying to use the vma. This should ensure that Rust Binder will refuse to interact with any other VMA. The plan is to introduce more vma abstractions to avoid this unsafe access to vm_ops and vm_private_data, but for now let's start with the simplest possible fix. C Binder performs the same check in a slightly different way: it provides a vm_ops->close that sets a boolean to true, then checks that boolean after calling vma_lookup(), but this is more fragile than the solution in this patch. (We probably still want to do both, but the vm_ops->close callback will be added later as part of the follow-up vma API changes.) It's still possible to remap the vma so that pages appear in the right vma, but at the wrong offset, but this is a separate issue and will be fixed when Rust Binder gets a vm_ops->close callback. Cc: stable Fixes: eafedbc7c050 ("rust_binder: add Rust Binder driver") Reported-by: Jann Horn Reviewed-by: Jann Horn Signed-off-by: Alice Ryhl Acked-by: Danilo Krummrich Acked-by: Liam R. Howlett Link: https://patch.msgid.link/20260218-binder-vma-check-v2-1-60f9d695a990@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/page_range.rs | 83 +++++++++++++++++++++++++++--------- 1 file changed, 63 insertions(+), 20 deletions(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/page_range.rs b/drivers/android/binder/page_range.rs index fdd97112ef5c..67aae783e8b8 100644 --- a/drivers/android/binder/page_range.rs +++ b/drivers/android/binder/page_range.rs @@ -142,6 +142,30 @@ pub(crate) struct ShrinkablePageRange { _pin: PhantomPinned, } +// We do not define any ops. For now, used only to check identity of vmas. +static BINDER_VM_OPS: bindings::vm_operations_struct = pin_init::zeroed(); + +// To ensure that we do not accidentally install pages into or zap pages from the wrong vma, we +// check its vm_ops and private data before using it. +fn check_vma(vma: &virt::VmaRef, owner: *const ShrinkablePageRange) -> Option<&virt::VmaMixedMap> { + // SAFETY: Just reading the vm_ops pointer of any active vma is safe. + let vm_ops = unsafe { (*vma.as_ptr()).vm_ops }; + if !ptr::eq(vm_ops, &BINDER_VM_OPS) { + return None; + } + + // SAFETY: Reading the vm_private_data pointer of a binder-owned vma is safe. + let vm_private_data = unsafe { (*vma.as_ptr()).vm_private_data }; + // The ShrinkablePageRange is only dropped when the Process is dropped, which only happens once + // the file's ->release handler is invoked, which means the ShrinkablePageRange outlives any + // VMA associated with it, so there can't be any false positives due to pointer reuse here. + if !ptr::eq(vm_private_data, owner.cast()) { + return None; + } + + vma.as_mixedmap_vma() +} + struct Inner { /// Array of pages. /// @@ -308,6 +332,18 @@ impl ShrinkablePageRange { inner.size = num_pages; inner.vma_addr = vma.start(); + // This pointer is only used for comparison - it's not dereferenced. + // + // SAFETY: We own the vma, and we don't use any methods on VmaNew that rely on + // `vm_private_data`. + unsafe { + (*vma.as_ptr()).vm_private_data = ptr::from_ref(self).cast_mut().cast::() + }; + + // SAFETY: We own the vma, and we don't use any methods on VmaNew that rely on + // `vm_ops`. + unsafe { (*vma.as_ptr()).vm_ops = &BINDER_VM_OPS }; + Ok(num_pages) } @@ -399,22 +435,24 @@ impl ShrinkablePageRange { // // Using `mmput_async` avoids this, because then the `mm` cleanup is instead queued to a // workqueue. - MmWithUser::into_mmput_async(self.mm.mmget_not_zero().ok_or(ESRCH)?) - .mmap_read_lock() - .vma_lookup(vma_addr) - .ok_or(ESRCH)? - .as_mixedmap_vma() - .ok_or(ESRCH)? - .vm_insert_page(user_page_addr, &new_page) - .inspect_err(|err| { - pr_warn!( - "Failed to vm_insert_page({}): vma_addr:{} i:{} err:{:?}", - user_page_addr, - vma_addr, - i, - err - ) - })?; + check_vma( + MmWithUser::into_mmput_async(self.mm.mmget_not_zero().ok_or(ESRCH)?) + .mmap_read_lock() + .vma_lookup(vma_addr) + .ok_or(ESRCH)?, + self, + ) + .ok_or(ESRCH)? + .vm_insert_page(user_page_addr, &new_page) + .inspect_err(|err| { + pr_warn!( + "Failed to vm_insert_page({}): vma_addr:{} i:{} err:{:?}", + user_page_addr, + vma_addr, + i, + err + ) + })?; let inner = self.lock.lock(); @@ -667,12 +705,15 @@ unsafe extern "C" fn rust_shrink_free_page( let mmap_read; let mm_mutex; let vma_addr; + let range_ptr; { // CAST: The `list_head` field is first in `PageInfo`. let info = item as *mut PageInfo; // SAFETY: The `range` field of `PageInfo` is immutable. - let range = unsafe { &*((*info).range) }; + range_ptr = unsafe { (*info).range }; + // SAFETY: The `range` outlives its `PageInfo` values. + let range = unsafe { &*range_ptr }; mm = match range.mm.mmget_not_zero() { Some(mm) => MmWithUser::into_mmput_async(mm), @@ -717,9 +758,11 @@ unsafe extern "C" fn rust_shrink_free_page( // SAFETY: The lru lock is locked when this method is called. unsafe { bindings::spin_unlock(&raw mut (*lru).lock) }; - if let Some(vma) = mmap_read.vma_lookup(vma_addr) { - let user_page_addr = vma_addr + (page_index << PAGE_SHIFT); - vma.zap_page_range_single(user_page_addr, PAGE_SIZE); + if let Some(unchecked_vma) = mmap_read.vma_lookup(vma_addr) { + if let Some(vma) = check_vma(unchecked_vma, range_ptr) { + let user_page_addr = vma_addr + (page_index << PAGE_SHIFT); + vma.zap_page_range_single(user_page_addr, PAGE_SIZE); + } } drop(mmap_read); -- cgit v1.2.3 From 4cb9e13fec0de7c942f5f927469beb8e48ddd20f Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Wed, 18 Feb 2026 11:53:27 +0000 Subject: rust_binder: avoid reading the written value in offsets array When sending a transaction, its offsets array is first copied into the target proc's vma, and then the values are read back from there. This is normally fine because the vma is a read-only mapping, so the target process cannot change the value under us. However, if the target process somehow gains the ability to write to its own vma, it could change the offset before it's read back, causing the kernel to misinterpret what the sender meant. If the sender happens to send a payload with a specific shape, this could in the worst case lead to the receiver being able to privilege escalate into the sender. The intent is that gaining the ability to change the read-only vma of your own process should not be exploitable, so remove this TOCTOU read even though it's unexploitable without another Binder bug. Cc: stable Fixes: eafedbc7c050 ("rust_binder: add Rust Binder driver") Reported-by: Jann Horn Reviewed-by: Jann Horn Signed-off-by: Alice Ryhl Acked-by: Liam R. Howlett Link: https://patch.msgid.link/20260218-binder-vma-check-v2-2-60f9d695a990@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/thread.rs | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/thread.rs b/drivers/android/binder/thread.rs index 0b62d24b2118..c004214b1662 100644 --- a/drivers/android/binder/thread.rs +++ b/drivers/android/binder/thread.rs @@ -1015,12 +1015,9 @@ impl Thread { // Copy offsets if there are any. if offsets_size > 0 { - { - let mut reader = - UserSlice::new(UserPtr::from_addr(trd_data_ptr.offsets as _), offsets_size) - .reader(); - alloc.copy_into(&mut reader, aligned_data_size, offsets_size)?; - } + let mut offsets_reader = + UserSlice::new(UserPtr::from_addr(trd_data_ptr.offsets as _), offsets_size) + .reader(); let offsets_start = aligned_data_size; let offsets_end = aligned_data_size + offsets_size; @@ -1041,11 +1038,9 @@ impl Thread { .step_by(size_of::()) .enumerate() { - let offset: usize = view - .alloc - .read::(index_offset)? - .try_into() - .map_err(|_| EINVAL)?; + let offset = offsets_reader.read::()?; + view.alloc.write(index_offset, &offset)?; + let offset: usize = offset.try_into().map_err(|_| EINVAL)?; if offset < end_of_previous_object || !is_aligned(offset, size_of::()) { pr_warn!("Got transaction with invalid offset."); -- cgit v1.2.3 From 2e303f0febb65a434040774b793ba8356698802b Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Tue, 24 Feb 2026 18:16:39 +0000 Subject: rust_binder: call set_notification_done() without proc lock Consider the following sequence of events on a death listener: 1. The remote process dies and sends a BR_DEAD_BINDER message. 2. The local process invokes the BC_CLEAR_DEATH_NOTIFICATION command. 3. The local process then invokes the BC_DEAD_BINDER_DONE. Then, the kernel will reply to the BC_DEAD_BINDER_DONE command with a BR_CLEAR_DEATH_NOTIFICATION_DONE reply using push_work_if_looper(). However, this can result in a deadlock if the current thread is not a looper. This is because dead_binder_done() still holds the proc lock during set_notification_done(), which called push_work_if_looper(). Normally, push_work_if_looper() takes the thread lock, which is fine to take under the proc lock. But if the current thread is not a looper, then it falls back to delivering the reply to the process work queue, which involves taking the proc lock. Since the proc lock is already held, this is a deadlock. Fix this by releasing the proc lock during set_notification_done(). It was not intentional that it was held during that function to begin with. I don't think this ever happens in Android because BC_DEAD_BINDER_DONE is only invoked in response to BR_DEAD_BINDER messages, and the kernel always delivers BR_DEAD_BINDER to a looper. So there's no scenario where Android userspace will call BC_DEAD_BINDER_DONE on a non-looper thread. Cc: stable Fixes: eafedbc7c050 ("rust_binder: add Rust Binder driver") Reported-by: syzbot+c8287e65a57a89e7fb72@syzkaller.appspotmail.com Tested-by: syzbot+c8287e65a57a89e7fb72@syzkaller.appspotmail.com Signed-off-by: Alice Ryhl Reviewed-by: Gary Guo Reviewed-by: Andreas Hindborg Link: https://patch.msgid.link/20260224-binder-dead-binder-done-proc-lock-v1-1-bbe1b8a6e74a@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/process.rs | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/process.rs b/drivers/android/binder/process.rs index 41de5593197c..f06498129aa9 100644 --- a/drivers/android/binder/process.rs +++ b/drivers/android/binder/process.rs @@ -1295,7 +1295,8 @@ impl Process { } pub(crate) fn dead_binder_done(&self, cookie: u64, thread: &Thread) { - if let Some(death) = self.inner.lock().pull_delivered_death(cookie) { + let death = self.inner.lock().pull_delivered_death(cookie); + if let Some(death) = death { death.set_notification_done(thread); } } -- cgit v1.2.3 From a0b9b0f1433c845bda708753db32befef78e0f1f Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Wed, 18 Feb 2026 15:13:23 +0000 Subject: rust_binder: use lock_vma_under_rcu() in use_page_slow() There's no reason to lock the whole mm when we are doing operations on the vma if we can help it, so to reduce contention, use the lock_vma_under_rcu() abstraction. Signed-off-by: Alice Ryhl Reviewed-by: Jann Horn Reviewed-by: Liam R. Howlett Link: https://patch.msgid.link/20260218-binder-vma-rcu-v1-1-8bd45b2b1183@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/page_range.rs | 37 ++++++++++++++++++------------------ 1 file changed, 19 insertions(+), 18 deletions(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/page_range.rs b/drivers/android/binder/page_range.rs index 67aae783e8b8..9dfc154e5dd4 100644 --- a/drivers/android/binder/page_range.rs +++ b/drivers/android/binder/page_range.rs @@ -435,24 +435,25 @@ impl ShrinkablePageRange { // // Using `mmput_async` avoids this, because then the `mm` cleanup is instead queued to a // workqueue. - check_vma( - MmWithUser::into_mmput_async(self.mm.mmget_not_zero().ok_or(ESRCH)?) - .mmap_read_lock() - .vma_lookup(vma_addr) - .ok_or(ESRCH)?, - self, - ) - .ok_or(ESRCH)? - .vm_insert_page(user_page_addr, &new_page) - .inspect_err(|err| { - pr_warn!( - "Failed to vm_insert_page({}): vma_addr:{} i:{} err:{:?}", - user_page_addr, - vma_addr, - i, - err - ) - })?; + let mm = MmWithUser::into_mmput_async(self.mm.mmget_not_zero().ok_or(ESRCH)?); + { + let vma_read; + let mmap_read; + let vma = if let Some(ret) = mm.lock_vma_under_rcu(vma_addr) { + vma_read = ret; + check_vma(&vma_read, self) + } else { + mmap_read = mm.mmap_read_lock(); + mmap_read + .vma_lookup(vma_addr) + .and_then(|vma| check_vma(vma, self)) + }; + + match vma { + Some(vma) => vma.vm_insert_page(user_page_addr, &new_page)?, + None => return Err(ESRCH), + } + } let inner = self.lock.lock(); -- cgit v1.2.3 From ec327abae5edd1d5b60ea9f920212970133171d2 Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Sat, 14 Mar 2026 11:19:51 +0000 Subject: rust_binder: use AssertSync for BINDER_VM_OPS When declaring an immutable global variable in Rust, the compiler checks that it looks thread safe, because it is generally safe to access said global variable. When using C bindings types for these globals, we don't really want this check, because it is conservative and assumes pointers are not thread safe. In the case of BINDER_VM_OPS, this is a challenge when combined with the patch 'userfaultfd: introduce vm_uffd_ops' [1], which introduces a pointer field to vm_operations_struct. It previously only held function pointers, which are considered thread safe. Rust Binder should not be assuming that vm_operations_struct contains no pointer fields, so to fix this, use AssertSync (which Rust Binder has already declared for another similar global of type struct file_operations with the same problem). This ensures that even if another commit adds a pointer field to vm_operations_struct, this does not cause problems. Fixes: 8ef2c15aeae0 ("rust_binder: check ownership before using vma") Cc: stable Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202603121235.tpnRxFKO-lkp@intel.com/ Link: https://lore.kernel.org/r/20260306171815.3160826-8-rppt@kernel.org [1] Signed-off-by: Alice Ryhl Reviewed-by: Gary Guo Link: https://patch.msgid.link/20260314111951.4139029-1-aliceryhl@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/android/binder/page_range.rs | 8 +++++--- drivers/android/binder/rust_binder_main.rs | 2 +- 2 files changed, 6 insertions(+), 4 deletions(-) (limited to 'drivers/android') diff --git a/drivers/android/binder/page_range.rs b/drivers/android/binder/page_range.rs index 9dfc154e5dd4..b57e0c7ba3f1 100644 --- a/drivers/android/binder/page_range.rs +++ b/drivers/android/binder/page_range.rs @@ -13,6 +13,8 @@ // // The shrinker will use trylock methods because it locks them in a different order. +use crate::AssertSync; + use core::{ marker::PhantomPinned, mem::{size_of, size_of_val, MaybeUninit}, @@ -143,14 +145,14 @@ pub(crate) struct ShrinkablePageRange { } // We do not define any ops. For now, used only to check identity of vmas. -static BINDER_VM_OPS: bindings::vm_operations_struct = pin_init::zeroed(); +static BINDER_VM_OPS: AssertSync = AssertSync(pin_init::zeroed()); // To ensure that we do not accidentally install pages into or zap pages from the wrong vma, we // check its vm_ops and private data before using it. fn check_vma(vma: &virt::VmaRef, owner: *const ShrinkablePageRange) -> Option<&virt::VmaMixedMap> { // SAFETY: Just reading the vm_ops pointer of any active vma is safe. let vm_ops = unsafe { (*vma.as_ptr()).vm_ops }; - if !ptr::eq(vm_ops, &BINDER_VM_OPS) { + if !ptr::eq(vm_ops, &BINDER_VM_OPS.0) { return None; } @@ -342,7 +344,7 @@ impl ShrinkablePageRange { // SAFETY: We own the vma, and we don't use any methods on VmaNew that rely on // `vm_ops`. - unsafe { (*vma.as_ptr()).vm_ops = &BINDER_VM_OPS }; + unsafe { (*vma.as_ptr()).vm_ops = &BINDER_VM_OPS.0 }; Ok(num_pages) } diff --git a/drivers/android/binder/rust_binder_main.rs b/drivers/android/binder/rust_binder_main.rs index aa5f2a75adb4..014010662df8 100644 --- a/drivers/android/binder/rust_binder_main.rs +++ b/drivers/android/binder/rust_binder_main.rs @@ -306,7 +306,7 @@ impl kernel::Module for BinderModule { /// Makes the inner type Sync. #[repr(transparent)] pub struct AssertSync(T); -// SAFETY: Used only to insert `file_operations` into a global, which is safe. +// SAFETY: Used only to insert C bindings types into globals, which is safe. unsafe impl Sync for AssertSync {} /// File operations that rust_binderfs.c can use. -- cgit v1.2.3