Skip to content
Snippets Groups Projects
  1. Apr 29, 2023
  2. Apr 25, 2023
  3. Apr 18, 2023
    • Hui Peng's avatar
      Fix an OOB bug in set_data · 58802164
      Hui Peng authored
      Plus move macros used in struct bt_oob_data_s
      to bluetooth.h
      
      Bug: 274722185
      Test: manual
      Ignore-AOSP-First: security
      Tag: #security
      Change-Id: Ie12feb4090a1eb88f5c9e097546f55a076839fb0
      58802164
    • Hui Peng's avatar
      Fix an OOB write bug in gatt_process_notification · fdaaa82d
      Hui Peng authored
      Bug: 276975913
      Test: manual
      Ignore-AOSP-First: security
      Tag: security
      Change-Id: I38353a573168e18f06b2b311e532a937044fd92f
      fdaaa82d
    • Hui Peng's avatar
      Fix an OOB bug in set_data · 0592ed17
      Hui Peng authored
      Bug: 274722163
      Test: manual
      Ignore-AOSP-First: security
      Tag: #security
      Change-Id: Ie4b30bbc19ba0bd191839af35880a4831d8005b1
      0592ed17
  4. Apr 14, 2023
  5. Apr 13, 2023
  6. Apr 04, 2023
  7. Apr 03, 2023
  8. Mar 31, 2023
  9. Mar 29, 2023
    • Brian Delwiche's avatar
      Fix OOB read in btm_ble_periodic_av_sync_lost · c077ffbe
      Brian Delwiche authored
      btm_ble_periodic_av_sync_lost internally calls the function
      btm_ble_get_psync_index_from_handle, which polls the internal periodic
      sync buffer and returns a matching index if one exists.  If no matching
      handle is found, it returns MAX_SYNC_TRANSACTION.
      
      However, here the calling function lacks the check for this case present
      in similar functions.  If no handle is matched, it will attempt to index
      the buffer with MAX_SYNC_TRANSACTION, which will overrun it by a single
      width and lead to OOB access.
      
      Add handling for this case.
      
      Bug: 273502002
      Test: atest bluetooth_test_gd_unit, atest net_test_stack_btm, validated
      against researcher POC
      Tag: #security
      Ignore-AOSP-First: Security
      
      Change-Id: I2e1e95b277f81b2668f721a7693df50841968ec5
      c077ffbe
  10. Mar 28, 2023
    • tyiu's avatar
      Fix gatt_end_operation buffer overflow · 7236e449
      tyiu authored
      Added boundary check for gatt_end_operation to prevent writing out of
      boundary.
      
      Since response of the GATT server is handled in
      gatt_client_handle_server_rsp() and gatt_process_read_rsp(), the maximum
      lenth that can be passed into the handlers is bounded by
      GATT_MAX_MTU_SIZE, which is set to 517, which is greater than
      GATT_MAX_ATTR_LEN which is set to 512. The fact that there is no spec
      that gaurentees MTU response to be less than or equal to 512 bytes can
      cause a buffer overflow when performing memcpy without length check.
      
      Bug: 261068592
      Test: No test since not affecting behavior
      Tag: #security
      Ignore-AOSP-First: security
      
      Change-Id: I49e2797cd9300ee4cd69f2c7fa5f0073db78b873
      7236e449
  11. Mar 24, 2023
    • Hui Peng's avatar
      Fix a OOB bug in bta_hh_co_get_rpt_rsp · f173fcb4
      Hui Peng authored
      Fix to the regression reported in b/264708304 and b/266585826 added:
      The root cause of the regression, the sensor HAL layer expects the HID feature
      reports to contain 40 bytes, even less bytes are contained in the data
      field.
      
      This updated fix restores the length of data fields with the len arg.
      
      Bug: 259675705
      Test: manual verification with a Pixel 6 and LinkBuds
      Ignore-AOSP-First: security
      Tag: security
      Change-Id: I02f16c360965b049fc6c8fdfa0132b7aa54bc1d3
      f173fcb4
  12. Mar 21, 2023
  13. Mar 13, 2023
  14. Mar 11, 2023
  15. Mar 06, 2023
  16. Feb 16, 2023
    • Brian Delwiche's avatar
      Fix potential use after free in pan_api.cc · f4bd0731
      Brian Delwiche authored
      Structure length is checked in pan_api.cc after the structure may
      be freed, leading to a potential use after free.
      
      Save the buffer length to a local instead.  Note that BNEP_WriteBuf
      may alter the length being written internally; this does not appear
      to be an issue in this use case because the octet count being tracked
      is used only for logging purposes within PAN.
      
      Bug: 259939435
      Test: atest bluetooth_test_gd_unit, validate against researcher POC
      Tag: #security
      Ignore-AOSP-First: Security
      Change-Id: I613b3dd3684182bdc725f9e1512061484448d367
      f4bd0731
  17. Feb 15, 2023
    • Brian Delwiche's avatar
      Validate buffer length in sdpu_build_uuid_seq · 367ed057
      Brian Delwiche authored
      sdpu_build_uuid_seq accepts a UUID sequence of arbitrary length
      but does not validate against the boundaries of the buffer it's
      filling.  This can lead to an OOB write.
      
      Add validation.
      
      Bug: 239414876
      Test: atest: bluetooth, validated against POC
      Tag: #security
      Ignore-AOSP-First: Security
      
      Change-Id: I6c0b91428bd37d73ae707b8a1843338998fb9562
      367ed057
  18. Feb 09, 2023
  19. Feb 07, 2023
  20. Feb 06, 2023
Loading