---
- branch: netbsd-9
  date: Sun Apr 12 08:48:57 UTC 2020
  files:
  - new: 1.75.20.2
    old: 1.75.20.1
    path: src/sys/dev/acpi/acpi_ec.c
    pathrev: src/sys/dev/acpi/acpi_ec.c@1.75.20.2
    type: modified
  id: 20200412T084857Z.093e36dc92c95a0774149d0efe948d4bd9e60257
  log: "Pull up following revision(s) (requested by riastradh in ticket #829):\n\n\tsys/dev/acpi/acpi_ec.c:
    revision 1.78\n\tsys/dev/acpi/acpi_ec.c: revision 1.79\n\tsys/dev/acpi/acpi_ec.c:
    revision 1.80\n\tsys/dev/acpi/acpi_ec.c: revision 1.81\n\nRevert acpi_ec.c 1.77.\nWe
    will do this another way.\nok msaitoh\n\nRevert acpi_ec.c 1.76.\nWe will do this
    another way, and separate KNF fixes from the critical\nfunctional change.\nok
    msaitoh\n\nKNF\n\nReject overly large widths, from mlelstv.\nWe are returning
    an ACPI_INTEGER (= uint64_t), so it doesn't make\nsense to handle more than 64
    bits.\n\nApparently there are some ACPIs out there that ask for unreasonably\nlarge
    widths here.  Just reject those requests, rather than writing\npast the caller's
    stack buffer.\n\nPreviously we attempted to fix this by copying byte by byte as
    large\nas the caller asked, in order to avoid the undefined behaviour of\nshifting
    past the size of ACPI_INTEGER, but that just turned a shift\n(which might have
    been harmless on real machines) into a stack buffer\noverflow (!).\n\nok msaitoh\n"
  module: src
  subject: 'CVS commit: [netbsd-9] src/sys/dev/acpi'
  unixtime: '1586681337'
  user: martin