linux-stable.git/drivers/edac, branch v3.16.78

EDAC: Fix global-out-of-bounds write when setting edac_mc_poll_msec

2019-10-31T22:14:58+00:00

commit d8655e7630dafa88bc37f101640e39c736399771 upstream.

Commit 9da21b1509d8 ("EDAC: Poll timeout cannot be zero, p2") assumes
edac_mc_poll_msec to be unsigned long, but the type of the variable still
remained as int. Setting edac_mc_poll_msec can trigger out-of-bounds
write.

Reproducer:

  # echo 1001 > /sys/module/edac_core/parameters/edac_mc_poll_msec

KASAN report:

  BUG: KASAN: global-out-of-bounds in edac_set_poll_msec+0x140/0x150
  Write of size 8 at addr ffffffffb91b2d00 by task bash/1996

  CPU: 1 PID: 1996 Comm: bash Not tainted 5.2.0-rc6+ #23
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-2.fc30 04/01/2014
  Call Trace:
   dump_stack+0xca/0x13e
   print_address_description.cold+0x5/0x246
   __kasan_report.cold+0x75/0x9a
   ? edac_set_poll_msec+0x140/0x150
   kasan_report+0xe/0x20
   edac_set_poll_msec+0x140/0x150
   ? dimmdev_location_show+0x30/0x30
   ? vfs_lock_file+0xe0/0xe0
   ? _raw_spin_lock+0x87/0xe0
   param_attr_store+0x1b5/0x310
   ? param_array_set+0x4f0/0x4f0
   module_attr_store+0x58/0x80
   ? module_attr_show+0x80/0x80
   sysfs_kf_write+0x13d/0x1a0
   kernfs_fop_write+0x2bc/0x460
   ? sysfs_kf_bin_read+0x270/0x270
   ? kernfs_notify+0x1f0/0x1f0
   __vfs_write+0x81/0x100
   vfs_write+0x1e1/0x560
   ksys_write+0x126/0x250
   ? __ia32_sys_read+0xb0/0xb0
   ? do_syscall_64+0x1f/0x390
   do_syscall_64+0xc1/0x390
   entry_SYSCALL_64_after_hwframe+0x49/0xbe
  RIP: 0033:0x7fa7caa5e970
  Code: 73 01 c3 48 8b 0d 28 d5 2b 00 f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00 83 3d 99 2d 2c 00 00 75 10 b8 01 00 00 00 04
  RSP: 002b:00007fff6acfdfe8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
  RAX: ffffffffffffffda RBX: 0000000000000005 RCX: 00007fa7caa5e970
  RDX: 0000000000000005 RSI: 0000000000e95c08 RDI: 0000000000000001
  RBP: 0000000000e95c08 R08: 00007fa7cad1e760 R09: 00007fa7cb36a700
  R10: 0000000000000073 R11: 0000000000000246 R12: 0000000000000005
  R13: 0000000000000001 R14: 00007fa7cad1d600 R15: 0000000000000005

  The buggy address belongs to the variable:
   edac_mc_poll_msec+0x0/0x40

  Memory state around the buggy address:
   ffffffffb91b2c00: 00 00 00 00 fa fa fa fa 00 00 00 00 fa fa fa fa
   ffffffffb91b2c80: 00 00 00 00 fa fa fa fa 00 00 00 00 fa fa fa fa
  >ffffffffb91b2d00: 04 fa fa fa fa fa fa fa 04 fa fa fa fa fa fa fa
                     ^
   ffffffffb91b2d80: 04 fa fa fa fa fa fa fa 00 00 00 00 00 00 00 00
   ffffffffb91b2e00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Fix it by changing the type of edac_mc_poll_msec to unsigned int.
The reason why this patch adopts unsigned int rather than unsigned long
is msecs_to_jiffies() assumes arg to be unsigned int. We can avoid
integer conversion bugs and unsigned int will be large enough for
edac_mc_poll_msec.

Reviewed-by: James Morse 
Fixes: 9da21b1509d8 ("EDAC: Poll timeout cannot be zero, p2")
Signed-off-by: Eiichi Tsukata 
Signed-off-by: Tony Luck 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

EDAC, {i7core,sb,skx}_edac: Fix uncorrected error counting

2019-02-11T17:53:23+00:00

commit 432de7fd7630c84ad24f1c2acd1e3bb4ce3741ca upstream.

The count of errors is picked up from bits 52:38 of the machine check
bank status register. But this is the count of *corrected* errors. If an
uncorrected error is being logged, the h/w sets this field to 0. Which
means that when edac_mc_handle_error() is called, the EDAC core will
carefully add zero to the appropriate uncorrected error counts.

Signed-off-by: Tony Luck 
[ Massage commit message. ]
Signed-off-by: Borislav Petkov 
Cc: Aristeu Rozanski 
Cc: Mauro Carvalho Chehab 
Cc: Qiuxu Zhuo 
Cc: linux-edac 
Link: http://lkml.kernel.org/r/20180928213934.19890-1-tony.luck@intel.com
[bwh: Backported to 3.16: Drop change in skx_edac.c]
Signed-off-by: Ben Hutchings

EDAC, i7core: Fix memleaks and use-after-free on probe and remove

2018-12-16T22:08:04+00:00

commit 6c974d4dfafe5e9ee754f2a6fba0eb1864f1649e upstream.

Make sure to free and deregister the addrmatch and chancounts devices
allocated during probe in all error paths. Also fix use-after-free in a
probe error path and in the remove success path where the devices were
being put before before deregistration.

Signed-off-by: Johan Hovold 
Cc: Mauro Carvalho Chehab 
Cc: linux-edac 
Fixes: 356f0a30860d ("i7core_edac: change the mem allocation scheme to make Documentation/kobject.txt happy")
Link: http://lkml.kernel.org/r/20180612124335.6420-2-johan@kernel.org
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC: i7core: Return proper error codes for kzalloc() errors

2018-12-16T22:08:03+00:00

commit e97d7e38162dc305b4734a316ca758a2bbd1fa6e upstream.

... instead of possibly uninitialized return value.

Signed-off-by: Takashi Iwai 
Link: http://lkml.kernel.org/r/1423046938-18111-5-git-send-email-tiwai@suse.de
[ Add a commit message, albeit a small one. ]
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC: Fix memleak in module init error path

2018-12-16T22:08:03+00:00

commit 4708aa85d50cc6e962dfa8acf5ad4e0d290a21db upstream.

Make sure to use put_device() to free the initialised struct device so
that resources managed by driver core also gets released in the event of
a registration failure.

Signed-off-by: Johan Hovold 
Cc: Denis Kirjanov 
Cc: Mauro Carvalho Chehab 
Cc: linux-edac 
Fixes: 2d56b109e3a5 ("EDAC: Handle error path in edac_mc_sysfs_init() properly")
Link: http://lkml.kernel.org/r/20180612124335.6420-1-johan@kernel.org
Signed-off-by: Borislav Petkov 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

EDAC, octeon: Fix an uninitialized variable warning

2018-06-16T21:21:40+00:00

commit 544e92581a2ac44607d7cc602c6b54d18656f56d upstream.

Fix an uninitialized variable warning in the Octeon EDAC driver, as seen
in MIPS cavium_octeon_defconfig builds since v4.14 with Codescape GNU
Tools 2016.05-03:

  drivers/edac/octeon_edac-lmc.c In function ‘octeon_lmc_edac_poll_o2’:
  drivers/edac/octeon_edac-lmc.c:87:24: warning: ‘((long unsigned int*)&int_reg)[1]’ may \
    be used uninitialized in this function [-Wmaybe-uninitialized]
    if (int_reg.s.sec_err || int_reg.s.ded_err) {
                        ^
Iinitialise the whole int_reg variable to zero before the conditional
assignments in the error injection case.

Signed-off-by: James Hogan 
Acked-by: David Daney 
Cc: linux-edac 
Cc: linux-mips@linux-mips.org
Fixes: 1bc021e81565 ("EDAC: Octeon: Add error injection support")
Link: http://lkml.kernel.org/r/20171113161206.20990-1-james.hogan@mips.com
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC: Increment correct counter in edac_inc_ue_error()

2016-08-22T21:37:51+00:00

commit 993f88f1cc7f0879047ff353e824e5cc8f10adfc upstream.

Fix typo in edac_inc_ue_error() to increment ue_noinfo_count instead of
ce_noinfo_count.

Signed-off-by: Emmanouil Maroudas 
Cc: Mauro Carvalho Chehab 
Cc: linux-edac 
Fixes: 4275be635597 ("edac: Change internal representation to work with layers")
Link: http://lkml.kernel.org/r/1461425580-5898-1-git-send-email-emmanouil.maroudas@gmail.com
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC: i7core, sb_edac: Don't return NOTIFY_BAD from mce_decoder callback

2016-06-15T20:29:29+00:00

commit c4fc1956fa31003bfbe4f597e359d751568e2954 upstream.

Both of these drivers can return NOTIFY_BAD, but this terminates
processing other callbacks that were registered later on the chain.
Since the driver did nothing to log the error it seems wrong to prevent
other interested parties from seeing it. E.g. neither of them had even
bothered to check the type of the error to see if it was a memory error
before the return NOTIFY_BAD.

Signed-off-by: Tony Luck 
Acked-by: Aristeu Rozanski 
Acked-by: Mauro Carvalho Chehab 
Cc: linux-edac 
Link: http://lkml.kernel.org/r/72937355dd92318d2630979666063f8a2853495b.1461864507.git.tony.luck@intel.com
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC, amd64_edac: Shift wrapping issue in f1x_get_norm_dct_addr()

2016-04-30T22:05:43+00:00

commit 6f3508f61c814ee852c199988a62bd954c50dfc1 upstream.

dct_sel_base_off is declared as a u64 but we're only using the lower 32
bits because of a shift wrapping bug. This can possibly truncate the
upper 16 bits of DctSelBaseOffset[47:26], causing us to misdecode the CS
row.

Fixes: c8e518d5673d ('amd64_edac: Sanitize f10_get_base_addr_offset')
Signed-off-by: Dan Carpenter 
Cc: Aravind Gopalakrishnan 
Cc: linux-edac 
Link: http://lkml.kernel.org/r/20160120095451.GB19898@mwanda
Signed-off-by: Borislav Petkov 
Signed-off-by: Ben Hutchings

EDAC: Robustify workqueues destruction

2016-02-02T17:38:36+00:00

commit fcd5c4dd8201595d4c598c9cca5e54760277d687 upstream.

EDAC workqueue destruction is really fragile. We cancel delayed work
but if it is still running and requeues itself, we still go ahead and
destroy the workqueue and the queued work explodes when workqueue core
attempts to run it.

Make the destruction more robust by switching op_state to offline so
that requeuing stops. Cancel any pending work *synchronously* too.

  EDAC i7core: Driver loaded.
  general protection fault: 0000 [#1] SMP
  CPU 12
  Modules linked in:
  Supported: Yes
  Pid: 0, comm: kworker/0:1 Tainted: G          IE   3.0.101-0-default #1 HP ProLiant DL380 G7
  RIP: 0010:[]  [] __queue_work+0x17/0x3f0
  < ... regs ...>
  Process kworker/0:1 (pid: 0, threadinfo ffff88019def6000, task ffff88019def4600)
  Stack:
   ...
  Call Trace:
   call_timer_fn
   run_timer_softirq
   __do_softirq
   call_softirq
   do_softirq
   irq_exit
   smp_apic_timer_interrupt
   apic_timer_interrupt
   intel_idle
   cpuidle_idle_call
   cpu_idle
  Code: ...
  RIP  __queue_work
   RSP <...>

Signed-off-by: Borislav Petkov 
Signed-off-by: Luis Henriques