- Linux-stable-mirror - lists.linaro.org

[PATCH 092/103] binder: check for binder_thread allocation failure in binder_poll()

by micky387

From: Eric Biggers <ebiggers(a)google.com> If the kzalloc() in binder_get_thread() fails, binder_poll() dereferences the resulting NULL pointer. Fix it by returning POLLERR if the memory allocation failed. This bug was found by syzkaller using fault injection. Reported-by: syzbot <syzkaller(a)googlegroups.com> Fixes: 457b9a6f09f0 ("Staging: android: add binder driver") Cc: stable(a)vger.kernel.org Signed-off-by: Eric Biggers <ebiggers(a)google.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index abf05aa..ed23507 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -4289,6 +4289,8 @@ static unsigned int binder_poll(struct file *filp, bool wait_for_proc_work; thread = binder_get_thread(proc); + if (!thread) + return POLLERR; binder_inner_proc_lock(thread->proc); thread->looper |= BINDER_LOOPER_STATE_POLL; -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 091/103] ANDROID: binder: remove WARN() for redundant txn error

by micky387

From: Todd Kjos <tkjos(a)android.com> binder_send_failed_reply() is called when a synchronous transaction fails. It reports an error to the thread that is waiting for the completion. Given that the transaction is synchronous, there should never be more than 1 error response to that thread -- this was being asserted with a WARN(). However, when exercising the driver with syzbot tests, cases were observed where multiple "synchronous" requests were sent without waiting for responses, so it is possible that multiple errors would be reported to the thread. This testing was conducted with panic_on_warn set which forced the crash. This is easily reproduced by sending back-to-back "synchronous" transactions without checking for any response (eg, set read_size to 0): bwr.write_buffer = (uintptr_t)&bc1; bwr.write_size = sizeof(bc1); bwr.read_buffer = (uintptr_t)&br; bwr.read_size = 0; ioctl(fd, BINDER_WRITE_READ, &bwr); sleep(1); bwr2.write_buffer = (uintptr_t)&bc2; bwr2.write_size = sizeof(bc2); bwr2.read_buffer = (uintptr_t)&br; bwr2.read_size = 0; ioctl(fd, BINDER_WRITE_READ, &bwr2); sleep(1); The first transaction is sent to the servicemanager and the reply fails because no VMA is set up by this client. After binder_send_failed_reply() is called, the BINDER_WORK_RETURN_ERROR is sitting on the thread's todo list since the read_size was 0 and the client is not waiting for a response. The 2nd transaction is sent and the BINDER_WORK_RETURN_ERROR has not been consumed, so the thread's reply_error.cmd is still set (normally cleared when the BINDER_WORK_RETURN_ERROR is handled). Therefore when the servicemanager attempts to reply to the 2nd failed transaction, the error is already set and it triggers this warning. This is a user error since it is not waiting for the synchronous transaction to complete. If it ever does check, it will see an error. Changed the WARN() to a pr_warn(). Signed-off-by: Todd Kjos <tkjos(a)android.com> Reported-by: syzbot <syzkaller(a)googlegroups.com> Cc: stable <stable(a)vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index c88582a..abf05aa 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -1919,8 +1919,14 @@ static void binder_send_failed_reply(struct binder_transaction *t, &target_thread->todo); wake_up_interruptible(&target_thread->wait); } else { - WARN(1, "Unexpected reply error: %u\n", - target_thread->reply_error.cmd); + /* + * Cannot get here for normal operation, but + * we can if multiple synchronous transactions + * are sent without blocking for responses. + * Just ignore the 2nd error in this case. + */ + pr_warn("Unexpected reply error: %u\n", + target_thread->reply_error.cmd); } binder_inner_proc_unlock(target_thread->proc); binder_thread_dec_tmpref(target_thread); -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 090/103] binder: replace "%p" with "%pK"

by micky387

From: Todd Kjos <tkjos(a)android.com> The format specifier "%p" can leak kernel addresses. Use "%pK" instead. There were 4 remaining cases in binder.c. Signed-off-by: Todd Kjos <tkjos(a)google.com> Cc: stable <stable(a)vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index 7ab7726..c88582a 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -2101,7 +2101,7 @@ static void binder_transaction_buffer_release(struct binder_proc *proc, int debug_id = buffer->debug_id; binder_debug(BINDER_DEBUG_TRANSACTION, - "%d buffer release %d, size %zd-%zd, failed at %p\n", + "%d buffer release %d, size %zd-%zd, failed at %pK\n", proc->pid, buffer->debug_id, buffer->data_size, buffer->offsets_size, failed_at); @@ -3613,7 +3613,7 @@ static int binder_thread_write(struct binder_proc *proc, } } binder_debug(BINDER_DEBUG_DEAD_BINDER, - "%d:%d BC_DEAD_BINDER_DONE %016llx found %p\n", + "%d:%d BC_DEAD_BINDER_DONE %016llx found %pK\n", proc->pid, thread->pid, (u64)cookie, death); if (death == NULL) { @@ -4921,7 +4921,7 @@ static void print_binder_transaction_ilocked(struct seq_file *m, spin_lock(&t->lock); to_proc = t->to_proc; seq_printf(m, - "%s %d: %p from %d:%d to %d:%d code %x flags %x pri %ld r%d", + "%s %d: %pK from %d:%d to %d:%d code %x flags %x pri %ld r%d", prefix, t->debug_id, t, t->from ? t->from->proc->pid : 0, t->from ? t->from->pid : 0, @@ -4945,7 +4945,7 @@ static void print_binder_transaction_ilocked(struct seq_file *m, } if (buffer->target_node) seq_printf(m, " node %d", buffer->target_node->debug_id); - seq_printf(m, " size %zd:%zd data %p\n", + seq_printf(m, " size %zd:%zd data %pK\n", buffer->data_size, buffer->offsets_size, buffer->data); } -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 086/103] binder: fix use-after-free in binder_transaction()

by micky387

From: Todd Kjos <tkjos(a)android.com> User-space normally keeps the node alive when creating a transaction since it has a reference to the target. The local strong ref keeps it alive if the sending process dies before the target process processes the transaction. If the source process is malicious or has a reference counting bug, this can fail. In this case, when we attempt to decrement the node in the failure path, the node has already been freed. This is fixed by taking a tmpref on the node while constructing the transaction. To avoid re-acquiring the node lock and inner proc lock to increment the proc's tmpref, a helper is used that does the ref increments on both the node and proc. Signed-off-by: Todd Kjos <tkjos(a)google.com> Cc: stable <stable(a)vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 93 ++++++++++++++++++++++++++++++++++-------------- 1 file changed, 66 insertions(+), 27 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index 3f9d702..fa23e45 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -2582,6 +2582,48 @@ static bool binder_proc_transaction(struct binder_transaction *t, return true; } +/** + * binder_get_node_refs_for_txn() - Get required refs on node for txn + * @node: struct binder_node for which to get refs + * @proc: returns @node->proc if valid + * @error: if no @proc then returns BR_DEAD_REPLY + * + * User-space normally keeps the node alive when creating a transaction + * since it has a reference to the target. The local strong ref keeps it + * alive if the sending process dies before the target process processes + * the transaction. If the source process is malicious or has a reference + * counting bug, relying on the local strong ref can fail. + * + * Since user-space can cause the local strong ref to go away, we also take + * a tmpref on the node to ensure it survives while we are constructing + * the transaction. We also need a tmpref on the proc while we are + * constructing the transaction, so we take that here as well. + * + * Return: The target_node with refs taken or NULL if no @node->proc is NULL. + * Also sets @proc if valid. If the @node->proc is NULL indicating that the + * target proc has died, @error is set to BR_DEAD_REPLY + */ +static struct binder_node *binder_get_node_refs_for_txn( + struct binder_node *node, + struct binder_proc **procp, + uint32_t *error) +{ + struct binder_node *target_node = NULL; + + binder_node_inner_lock(node); + if (node->proc) { + target_node = node; + binder_inc_node_nilocked(node, 1, 0, NULL); + binder_inc_node_tmpref_ilocked(node); + node->proc->tmp_ref++; + *procp = node->proc; + } else + *error = BR_DEAD_REPLY; + binder_node_inner_unlock(node); + + return target_node; +} + static void binder_transaction(struct binder_proc *proc, struct binder_thread *thread, struct binder_transaction_data *tr, int reply, @@ -2685,43 +2727,35 @@ static void binder_transaction(struct binder_proc *proc, ref = binder_get_ref_olocked(proc, tr->target.handle, true); if (ref) { - binder_inc_node(ref->node, 1, 0, NULL); - target_node = ref->node; - } - binder_proc_unlock(proc); - if (target_node == NULL) { + target_node = binder_get_node_refs_for_txn( + ref->node, &target_proc, + &return_error); + } else { binder_user_error("%d:%d got transaction to invalid handle\n", - proc->pid, thread->pid); + proc->pid, thread->pid); return_error = BR_FAILED_REPLY; - return_error_param = -EINVAL; - return_error_line = __LINE__; - goto err_invalid_target_handle; } + binder_proc_unlock(proc); } else { mutex_lock(&context->context_mgr_node_lock); target_node = context->binder_context_mgr_node; - if (target_node == NULL) { + if (target_node) + target_node = binder_get_node_refs_for_txn( + target_node, &target_proc, + &return_error); + else return_error = BR_DEAD_REPLY; - mutex_unlock(&context->context_mgr_node_lock); - return_error_line = __LINE__; - goto err_no_context_mgr_node; - } - binder_inc_node(target_node, 1, 0, NULL); mutex_unlock(&context->context_mgr_node_lock); } - e->to_node = target_node->debug_id; - binder_node_lock(target_node); - target_proc = target_node->proc; - if (target_proc == NULL) { - binder_node_unlock(target_node); - return_error = BR_DEAD_REPLY; + if (!target_node) { + /* + * return_error is set above + */ + return_error_param = -EINVAL; return_error_line = __LINE__; goto err_dead_binder; } - binder_inner_proc_lock(target_proc); - target_proc->tmp_ref++; - binder_inner_proc_unlock(target_proc); - binder_node_unlock(target_node); + e->to_node = target_node->debug_id; if (security_binder_transaction(proc->tsk, target_proc->tsk) < 0) { return_error = BR_FAILED_REPLY; @@ -3071,6 +3105,8 @@ static void binder_transaction(struct binder_proc *proc, if (target_thread) binder_thread_dec_tmpref(target_thread); binder_proc_dec_tmpref(target_proc); + if (target_node) + binder_dec_node_tmpref(target_node); /* * write barrier to synchronize with initialization * of log entry @@ -3089,6 +3125,8 @@ err_bad_parent: err_copy_data_failed: trace_binder_transaction_failed_buffer_release(t->buffer); binder_transaction_buffer_release(target_proc, t->buffer, offp); + if (target_node) + binder_dec_node_tmpref(target_node); target_node = NULL; t->buffer->transaction = NULL; binder_alloc_free_buf(&target_proc->alloc, t->buffer); @@ -3103,13 +3141,14 @@ err_bad_call_stack: err_empty_call_stack: err_dead_binder: err_invalid_target_handle: -err_no_context_mgr_node: if (target_thread) binder_thread_dec_tmpref(target_thread); if (target_proc) binder_proc_dec_tmpref(target_proc); - if (target_node) + if (target_node) { binder_dec_node(target_node, 1, 0); + binder_dec_node_tmpref(target_node); + } binder_debug(BINDER_DEBUG_FAILED_TRANSACTION, "%d:%d transaction failed %d/%d, size %lld-%lld line %d\n", -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 073/103] binder: free memory on error

by micky387

From: Christian Brauner <christian.brauner(a)ubuntu.com> On binder_init() the devices string is duplicated and smashed into individual device names which are passed along. However, the original duplicated string wasn't freed in case binder_init() failed. Let's free it on error. Signed-off-by: Christian Brauner <christian.brauner(a)ubuntu.com> Cc: stable <stable(a)vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index c40cd0c..ddd4235 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -5236,7 +5236,7 @@ static int __init init_binder_device(const char *name) static int __init binder_init(void) { int ret; - char *device_name, *device_names; + char *device_name, *device_names, *device_tmp; struct binder_device *device; struct hlist_node *tmp; @@ -5287,7 +5287,8 @@ static int __init binder_init(void) } strcpy(device_names, binder_devices_param); - while ((device_name = strsep(&device_names, ","))) { + device_tmp = device_names; + while ((device_name = strsep(&device_tmp, ","))) { ret = init_binder_device(device_name); if (ret) goto err_init_binder_device_failed; @@ -5301,6 +5302,9 @@ err_init_binder_device_failed: hlist_del(&device->hlist); kfree(device); } + + kfree(device_names); + err_alloc_device_names_failed: debugfs_remove_recursive(binder_debugfs_dir_entry_root); -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 072/103] ANDROID: binder: add hwbinder, vndbinder to BINDER_DEVICES.

by micky387

From: Martijn Coenen <maco(a)android.com> These will be required going forward. Change-Id: I960e205e7974079313da208959794c58577bdaa9 Signed-off-by: Martijn Coenen <maco(a)android.com> Cc: stable <stable(a)vger.kernel.org> # 4.11+ Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/android/Kconfig b/drivers/android/Kconfig index 832e885..4d4cdc1a 100644 --- a/drivers/android/Kconfig +++ b/drivers/android/Kconfig @@ -22,7 +22,7 @@ config ANDROID_BINDER_IPC config ANDROID_BINDER_DEVICES string "Android Binder devices" depends on ANDROID_BINDER_IPC - default "binder,hwbinder" + default "binder,hwbinder,vndbinder" ---help--- Default value for the binder.devices parameter. -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 044/103] binder: use group leader instead of open thread

by micky387

From: Todd Kjos <tkjos(a)android.com> The binder allocator assumes that the thread that called binder_open will never die for the lifetime of that proc. That thread is normally the group_leader, however it may not be. Use the group_leader instead of current. Signed-off-by: Todd Kjos <tkjos(a)google.com> Cc: stable <stable(a)vger.kernel.org> # 4.4+ Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index fbe0eb9..a24ad74 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -3039,8 +3039,8 @@ static int binder_open(struct inode *nodp, struct file *filp) proc = kzalloc(sizeof(*proc), GFP_KERNEL); if (proc == NULL) return -ENOMEM; - get_task_struct(current); - proc->tsk = current; + get_task_struct(current->group_leader); + proc->tsk = current->group_leader; INIT_LIST_HEAD(&proc->todo); init_waitqueue_head(&proc->wait); proc->default_priority = task_nice(current); -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 034/103] binder: Use wake up hint for synchronous transactions.

by micky387

From: Riley Andrews <riandrews(a)google.com> Use wake_up_interruptible_sync() to hint to the scheduler binder transactions are synchronous wakeups. Disable preemption while waking to avoid ping-ponging on the binder lock. Signed-off-by: Todd Kjos <tkjos(a)google.com> Signed-off-by: Omprakash Dhyade <odhyade(a)codeaurora.org> Cc: stable <stable(a)vger.kernel.org> # 4.4+ Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index b51fedc..6adf0bf 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -2199,8 +2199,12 @@ static void binder_transaction(struct binder_proc *proc, list_add_tail(&t->work.entry, target_list); tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE; list_add_tail(&tcomplete->entry, &thread->todo); - if (target_wait) - wake_up_interruptible(target_wait); + if (target_wait) { + if (reply || !(t->flags & TF_ONE_WAY)) + wake_up_interruptible_sync(target_wait); + else + wake_up_interruptible(target_wait); + } return; err_translate_failed: -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 033/103] Revert "android: binder: Sanity check at binder ioctl"

by micky387

From: Todd Kjos <tkjos(a)android.com> This reverts commit a906d6931f3ccaf7de805643190765ddd7378e27. The patch introduced a race in the binder driver. An attempt to fix the race was submitted in "[PATCH v2] android: binder: fix dangling pointer comparison", however the conclusion in the discussion for that patch was that the original patch should be reverted. The reversion is being done as part of the fine-grained locking patchset since the patch would need to be refactored when proc->vmm_vm_mm is removed from struct binder_proc and added in the binder allocator. Signed-off-by: Todd Kjos <tkjos(a)google.com> Cc: stable <stable(a)vger.kernel.org> # 4.6+ Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/android/binder.c | 5 ----- 1 file changed, 5 deletions(-) diff --git a/drivers/android/binder.c b/drivers/android/binder.c index 15b263a..b51fedc 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -3246,10 +3246,6 @@ static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) /*pr_info("binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/ - if (unlikely(current->mm != proc->vma_vm_mm)) { - pr_err("current mm mismatch proc mm\n"); - return -EINVAL; - } trace_binder_ioctl(cmd, arg); ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2); @@ -3465,7 +3461,6 @@ static int binder_open(struct inode *nodp, struct file *filp) return -ENOMEM; get_task_struct(current); proc->tsk = current; - proc->vma_vm_mm = current->mm; INIT_LIST_HEAD(&proc->todo); init_waitqueue_head(&proc->wait); proc->default_priority = task_nice(current); -- 2.7.4

7 years, 4 months

1
0
0 0

[PATCH 030/103] ANDROID: binder: add padding to binder_fd_array_object.

by micky387

From: Martijn Coenen <maco(a)android.com> binder_fd_array_object starts with a 4-byte header, followed by a few fields that are 8 bytes when ANDROID_BINDER_IPC_32BIT=N. This can cause alignment issues in a 64-bit kernel with a 32-bit userspace, as on x86_32 an 8-byte primitive may be aligned to a 4-byte address. Pad with a __u32 to fix this. Signed-off-by: Martijn Coenen <maco(a)android.com> Cc: stable <stable(a)vger.kernel.org> # 4.11+ Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- include/uapi/linux/android/binder.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/uapi/linux/android/binder.h b/include/uapi/linux/android/binder.h index 51f891f..7668b57 100644 --- a/include/uapi/linux/android/binder.h +++ b/include/uapi/linux/android/binder.h @@ -132,6 +132,7 @@ enum { /* struct binder_fd_array_object - object describing an array of fds in a buffer * @hdr: common header structure + * @pad: padding to ensure correct alignment * @num_fds: number of file descriptors in the buffer * @parent: index in offset array to buffer holding the fd array * @parent_offset: start offset of fd array in the buffer @@ -152,6 +153,7 @@ enum { */ struct binder_fd_array_object { struct binder_object_header hdr; + __u32 pad; binder_size_t num_fds; binder_size_t parent; binder_size_t parent_offset; -- 2.7.4

7 years, 4 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror