[PATCH] Bluetooth: hci_debugfs: serialize force_bredr_smp writes

[PATCH] Bluetooth: hci_debugfs: serialize force_bredr_smp writes

[Cen Zhang]

force_bredr_smp_write() calls smp_force_bredr() without hci_dev_lock().
That helper checks HCI_FORCE_BREDR_SMP, updates hdev->smp_bredr_data, and
then toggles the force flag.  Two same-value writers can therefore both
pass the -EALREADY gate and run the same enable or disable transition on
one hdev.

The buggy scenario involves two paths, with each column showing the order
within that path:

debugfs writer A:                         debugfs writer B:
  1. parse enable = false                   1. parse enable = false
  2. test HCI_FORCE_BREDR_SMP as set        2. test HCI_FORCE_BREDR_SMP as set
  3. read hdev->smp_bredr_data              3. read hdev->smp_bredr_data
  4. clear hdev->smp_bredr_data             4. clear hdev->smp_bredr_data
  5. call smp_del_chan(chan)                5. call smp_del_chan(chan)
  6. toggle HCI_FORCE_BREDR_SMP             6. toggle HCI_FORCE_BREDR_SMP

If writer B reaches step 3 after writer A has cleared
hdev->smp_bredr_data but before writer A toggles HCI_FORCE_BREDR_SMP,
writer B still passed the old flag check and can pass NULL into
smp_del_chan().  The double toggle can also leave the force flag
mismatched with the requested state.

Take hci_dev_lock() around smp_force_bredr() in the debugfs write path so
each request observes and applies one stable BR/EDR SMP transition.

Validation reproduced this kernel report:
KASAN null-ptr-deref in smp_del_chan+0x31/0x90
RIP: 0033:0x7faae680d340
RIP: 0010:smp_del_chan+0x31/0x90 [bluetooth]
Read of size 8
Call trace:
  dump_stack_lvl+0x66/0xa0
  kasan_report+0xe0/0x110
  smp_del_chan+0x31/0x90
  smp_force_bredr+0x69/0xc0
  trace_clock_x86_tsc+0x20/0x20
  srso_alias_return_thunk+0x5/0xfbef5
  lock_acquire+0xd0/0x300
  ksys_write+0xd2/0x170
  full_proxy_write+0x9e/0xd0
  vfs_write+0x1b0/0x810
  find_held_lock+0x2b/0x80
  do_user_addr_fault+0x65a/0x890
  rcu_is_watching+0x20/0x50
  do_syscall_64+0x115/0x6a0 (arch/x86/entry/syscall_64.c:87)
  entry_SYSCALL_64_after_hwframe+0x77/0x7f

Fixes: 300acfdec916 ("Bluetooth: Introduce force_bredr_smp debugfs option for testing")
Assisted-by: Codex:gpt-5.5
Signed-off-by: Cen Zhang <zzzccc427@gmail.com>
---
 net/bluetooth/hci_debugfs.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/bluetooth/hci_debugfs.c b/net/bluetooth/hci_debugfs.c
index 99e2e9fc70e8..89cc93f5984e 100644
--- a/net/bluetooth/hci_debugfs.c
+++ b/net/bluetooth/hci_debugfs.c
@@ -524,7 +524,9 @@ static ssize_t force_bredr_smp_write(struct file *file,
 	if (err)
 		return err;
 
+	hci_dev_lock(hdev);
 	err = smp_force_bredr(hdev, enable);
+	hci_dev_unlock(hdev);
 	if (err)
 		return err;
 
-- 
2.43.0

RE: Bluetooth: hci_debugfs: serialize force_bredr_smp writes

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 3165 bytes --]

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=1113387

---Test result---

Test Summary:
CheckPatch                    FAIL      0.53 seconds
VerifyFixes                   PASS      0.06 seconds
VerifySignedoff               PASS      0.06 seconds
GitLint                       PASS      0.18 seconds
SubjectPrefix                 PASS      0.05 seconds
BuildKernel                   PASS      26.74 seconds
CheckAllWarning               PASS      29.66 seconds
CheckSparse                   PASS      27.99 seconds
BuildKernel32                 PASS      25.94 seconds
CheckKernelLLVM               SKIP      0.00 seconds
TestRunnerSetup               PASS      581.10 seconds
TestRunner_l2cap-tester       PASS      60.19 seconds
TestRunner_iso-tester         PASS      79.55 seconds
TestRunner_bnep-tester        PASS      18.66 seconds
TestRunner_mgmt-tester        FAIL      212.01 seconds
TestRunner_rfcomm-tester      PASS      24.97 seconds
TestRunner_sco-tester         PASS      32.66 seconds
TestRunner_ioctl-tester       PASS      28.46 seconds
TestRunner_mesh-tester        FAIL      25.84 seconds
TestRunner_smp-tester         PASS      23.43 seconds
TestRunner_userchan-tester    PASS      19.82 seconds
TestRunner_6lowpan-tester     PASS      22.93 seconds
IncrementalBuild              PASS      25.13 seconds

Details
##############################
Test: CheckPatch - FAIL
Desc: Run checkpatch.pl script
Output:
Bluetooth: hci_debugfs: serialize force_bredr_smp writes
WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)
#108: 
  2. test HCI_FORCE_BREDR_SMP as set        2. test HCI_FORCE_BREDR_SMP as set

total: 0 errors, 1 warnings, 0 checks, 9 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
      mechanically convert to the typical style using --fix or --fix-inplace.

/github/workspace/src/patch/14635033.patch has style problems, please review.

NOTE: Ignored message types: UNKNOWN_COMMIT_ID

NOTE: If any of the errors are false positives, please report
      them to the maintainer, see CHECKPATCH in MAINTAINERS.


##############################
Test: CheckKernelLLVM - SKIP
Desc: Build kernel with LLVM + context analysis
Output:
Clang not found
##############################
Test: TestRunner_mgmt-tester - FAIL
Desc: Run mgmt-tester with test-runner
Output:
Total: 494, Passed: 489 (99.0%), Failed: 1, Not Run: 4

Failed Test Cases
Read Exp Feature - Success                           Failed       0.245 seconds
##############################
Test: TestRunner_mesh-tester - FAIL
Desc: Run mesh-tester with test-runner
Output:
Total: 10, Passed: 8 (80.0%), Failed: 2, Not Run: 0

Failed Test Cases
Mesh - Send cancel - 1                               Timed out    2.549 seconds
Mesh - Send cancel - 2                               Timed out    1.982 seconds


https://github.com/bluez/bluetooth-next/pull/332

---
Regards,
Linux Bluetooth