linux.git/drivers/base, branch v5.4.51

regmap: Fix memory leak from regmap_register_patch

2020-06-30T19:36:56+00:00

[ Upstream commit 95b2c3ec4cb1689db2389c251d39f64490ba641c ]

When a register patch is registered the reg_sequence is copied but the
memory allocated is never freed. Add a kfree in regmap_exit to clean it
up.

Fixes: 22f0d90a3482 ("regmap: Support register patch sets")
Signed-off-by: Charles Keepax 
Link: https://lore.kernel.org/r/20200617152129.19655-1-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown 
Signed-off-by: Sasha Levin

drivers: base: Fix NULL pointer exception in __platform_driver_probe() if a driver developer is foolish

2020-06-24T15:50:24+00:00

[ Upstream commit 388bcc6ecc609fca1b4920de7dc3806c98ec535e ]

If platform bus driver registration is failed then, accessing
platform bus spin lock (&drv->driver.bus->p->klist_drivers.k_lock)
in __platform_driver_probe() without verifying the return value
__platform_driver_register() can lead to NULL pointer exception.

So check the return value before attempting the spin lock.

One such example is below:

For a custom usecase, I have intentionally failed the platform bus
registration and I expected all the platform device/driver
registrations to fail gracefully. But I came across this panic
issue.

[    1.331067] BUG: kernel NULL pointer dereference, address: 00000000000000c8
[    1.331118] #PF: supervisor write access in kernel mode
[    1.331163] #PF: error_code(0x0002) - not-present page
[    1.331208] PGD 0 P4D 0
[    1.331233] Oops: 0002 [#1] PREEMPT SMP
[    1.331268] CPU: 3 PID: 1 Comm: swapper/0 Tainted: G        W         5.6.0-00049-g670d35fb0144 #165
[    1.331341] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015
[    1.331406] RIP: 0010:_raw_spin_lock+0x15/0x30
[    1.331588] RSP: 0000:ffffc9000001be70 EFLAGS: 00010246
[    1.331632] RAX: 0000000000000000 RBX: 00000000000000c8 RCX: 0000000000000001
[    1.331696] RDX: 0000000000000001 RSI: 0000000000000092 RDI: 0000000000000000
[    1.331754] RBP: 00000000ffffffed R08: 0000000000000501 R09: 0000000000000001
[    1.331817] R10: ffff88817abcc520 R11: 0000000000000670 R12: 00000000ffffffed
[    1.331881] R13: ffffffff82dbc268 R14: ffffffff832f070a R15: 0000000000000000
[    1.331945] FS:  0000000000000000(0000) GS:ffff88817bd80000(0000) knlGS:0000000000000000
[    1.332008] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    1.332062] CR2: 00000000000000c8 CR3: 000000000681e001 CR4: 00000000003606e0
[    1.332126] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    1.332189] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    1.332252] Call Trace:
[    1.332281]  __platform_driver_probe+0x92/0xee
[    1.332323]  ? rtc_dev_init+0x2b/0x2b
[    1.332358]  cmos_init+0x37/0x67
[    1.332396]  do_one_initcall+0x7d/0x168
[    1.332428]  kernel_init_freeable+0x16c/0x1c9
[    1.332473]  ? rest_init+0xc0/0xc0
[    1.332508]  kernel_init+0x5/0x100
[    1.332543]  ret_from_fork+0x1f/0x30
[    1.332579] CR2: 00000000000000c8
[    1.332616] ---[ end trace 3bd87f12e9010b87 ]---
[    1.333549] note: swapper/0[1] exited with preempt_count 1
[    1.333592] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000009
[    1.333736] Kernel Offset: disabled

Note, this can only be triggered if a driver errors out from this call,
which should never happen.  If it does, the driver needs to be fixed.

Signed-off-by: Kuppuswamy Sathyanarayanan 
Link: https://lore.kernel.org/r/20200408214003.3356-1-sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

x86/speculation: Add Special Register Buffer Data Sampling (SRBDS) mitigation

2020-06-10T18:24:57+00:00

commit 7e5b3c267d256822407a22fdce6afdf9cd13f9fb upstream

SRBDS is an MDS-like speculative side channel that can leak bits from the
random number generator (RNG) across cores and threads. New microcode
serializes the processor access during the execution of RDRAND and
RDSEED. This ensures that the shared buffer is overwritten before it is
released for reuse.

While it is present on all affected CPU models, the microcode mitigation
is not needed on models that enumerate ARCH_CAPABILITIES[MDS_NO] in the
cases where TSX is not supported or has been disabled with TSX_CTRL.

The mitigation is activated by default on affected processors and it
increases latency for RDRAND and RDSEED instructions. Among other
effects this will reduce throughput from /dev/urandom.

* Enable administrator to configure the mitigation off when desired using
  either mitigations=off or srbds=off.

* Export vulnerability status via sysfs

* Rename file-scoped macros to apply for non-whitelist table initializations.

 [ bp: Massage,
   - s/VULNBL_INTEL_STEPPING/VULNBL_INTEL_STEPPINGS/g,
   - do not read arch cap MSR a second time in tsx_fused_off() - just pass it in,
   - flip check in cpu_set_bug_bits() to save an indentation level,
   - reflow comments.
   jpoimboe: s/Mitigated/Mitigation/ in user-visible strings
   tglx: Dropped the fused off magic for now
 ]

Signed-off-by: Mark Gross 
Signed-off-by: Borislav Petkov 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Tony Luck 
Reviewed-by: Pawan Gupta 
Reviewed-by: Josh Poimboeuf 
Tested-by: Neelima Krishnan 
Signed-off-by: Greg Kroah-Hartman

component: Silence bind error on -EPROBE_DEFER

2020-05-27T15:46:31+00:00

[ Upstream commit 7706b0a76a9697021e2bf395f3f065c18f51043d ]

If a component fails to bind due to -EPROBE_DEFER we should not log an
error as this is not a real failure.

Fixes messages like:
vc4-drm soc:gpu: failed to bind 3f902000.hdmi (ops vc4_hdmi_ops): -517
vc4-drm soc:gpu: master bind failed: -517

Signed-off-by: James Hilliard 
Link: https://lore.kernel.org/r/20200411190241.89404-1-james.hilliard1@gmail.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

Revert "software node: Simplify software_node_release() function"

2020-05-10T08:31:30+00:00

commit 7589238a8cf37331607c3222a64ac3140b29532d upstream.

This reverts commit 3df85a1ae51f6b256982fe9d17c2dc5bfb4cc402.

The reverted commit says "It's possible to release the node ID
immediately when fwnode_remove_software_node() is called, no need to
wait for software_node_release() with that." However, releasing the node
ID before waiting for software_node_release() to be called causes the
node ID to be released before the kobject and the underlying sysfs
entry; this means there is a period of time where a sysfs entry exists
that is associated with an unallocated node ID.

Once consequence of this is that there is a race condition where it is
possible to call fwnode_create_software_node() with no parent node
specified (NULL) and have it fail with -EEXIST because the node ID that
was assigned is still associated with a stale sysfs entry that hasn't
been cleaned up yet.

Although it is difficult to reproduce this race condition under normal
conditions, it can be deterministically reproduced with the following
minconfig on UML:

CONFIG_KUNIT_DRIVER_PE_TEST=y
CONFIG_DEBUG_KERNEL=y
CONFIG_DEBUG_OBJECTS=y
CONFIG_DEBUG_OBJECTS_TIMERS=y
CONFIG_DEBUG_KOBJECT_RELEASE=y
CONFIG_KUNIT=y

Running the tests with this configuration causes the following failure:


kobject: 'node0' ((____ptrval____)): kobject_release, parent (____ptrval____) (delayed 400)
	ok 1 - pe_test_uints
sysfs: cannot create duplicate filename '/kernel/software_nodes/node0'
CPU: 0 PID: 28 Comm: kunit_try_catch Not tainted 5.6.0-rc3-next-20200227 #14

kobject_add_internal failed for node0 with -EEXIST, don't try to register things with the same name in the same directory.
kobject: 'node0' ((____ptrval____)): kobject_release, parent (____ptrval____) (delayed 100)
	# pe_test_uint_arrays: ASSERTION FAILED at drivers/base/test/property-entry-test.c:123
	Expected node is not error, but is: -17
	not ok 2 - pe_test_uint_arrays


Reported-by: Heidi Fahim 
Signed-off-by: Brendan Higgins 
Reviewed-by: Heikki Krogerus 
Cc: 5.3+  # 5.3+
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM: sleep: core: Switch back to async_schedule_dev()

2020-05-02T06:48:53+00:00

commit 09beebd8f93b3c8bf894e342f0a203a5c612478c upstream.

Commit 8b9ec6b73277 ("PM core: Use new async_schedule_dev command")
introduced a new function for better performance.

However commit f2a424f6c613 ("PM / core: Introduce dpm_async_fn()
helper") went back to the non-optimized version, async_schedule().

So switch back to the sync_schedule_dev() to improve performance

Fixes: f2a424f6c613 ("PM / core: Introduce dpm_async_fn() helper")
Signed-off-by: Kai-Heng Feng 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM: sleep: wakeup: Skip wakeup_source_sysfs_remove() if device is not there

2020-04-17T08:50:12+00:00

commit 87de6594dc45dbf6819f3e0ef92f9331c5a9444c upstream.

Skip wakeup_source_sysfs_remove() to fix a NULL pinter dereference via
ws->dev, if the wakeup source is unregistered before registering the
wakeup class from device_add().

Fixes: 2ca3d1ecb8c4 ("PM / wakeup: Register wakeup class kobj after device is added")
Signed-off-by: Neeraj Upadhyay 
Cc: 5.4+  # 5.4+
[ rjw: Subject & changelog, white space ]
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM / Domains: Allow no domain-idle-states DT property in genpd when parsing

2020-04-17T08:50:12+00:00

commit 56cb26891ea4180121265dc6b596015772c4a4b8 upstream.

Commit 2c361684803e ("PM / Domains: Don't treat zero found compatible idle
states as an error"), moved of_genpd_parse_idle_states() towards allowing
none compatible idle state to be found for the device node, rather than
returning an error code.

However, it didn't consider that the "domain-idle-states" DT property may
be missing as it's optional, which makes of_count_phandle_with_args() to
return -ENOENT. Let's fix this to make the behaviour consistent.

Fixes: 2c361684803e ("PM / Domains: Don't treat zero found compatible idle states as an error")
Reported-by: Benjamin Gaignard 
Cc: 4.20+  # 4.20+
Reviewed-by: Sudeep Holla 
Signed-off-by: Ulf Hansson 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

firmware: fix a double abort case with fw_load_sysfs_fallback

2020-04-17T08:50:05+00:00

[ Upstream commit bcfbd3523f3c6eea51a74d217a8ebc5463bcb7f4 ]

fw_sysfs_wait_timeout may return err with -ENOENT
at fw_load_sysfs_fallback and firmware is already
in abort status, no need to abort again, so skip it.

This issue is caused by concurrent situation like below:
when thread 1# wait firmware loading, thread 2# may write
-1 to abort loading and wakeup thread 1# before it timeout.
so wait_for_completion_killable_timeout of thread 1# would
return remaining time which is != 0 with fw_st->status
FW_STATUS_ABORTED.And the results would be converted into
err -ENOENT in __fw_state_wait_common and transfered to
fw_load_sysfs_fallback in thread 1#.
The -ENOENT means firmware status is already at ABORTED,
so fw_load_sysfs_fallback no need to get mutex to abort again.
-----------------------------
thread 1#,wait for loading
fw_load_sysfs_fallback
 ->fw_sysfs_wait_timeout
    ->__fw_state_wait_common
       ->wait_for_completion_killable_timeout

in __fw_state_wait_common,
...
93    ret = wait_for_completion_killable_timeout(&fw_st->completion, timeout);
94    if (ret != 0 && fw_st->status == FW_STATUS_ABORTED)
95       return -ENOENT;
96    if (!ret)
97	 return -ETIMEDOUT;
98
99    return ret < 0 ? ret : 0;
-----------------------------
thread 2#, write -1 to abort loading
firmware_loading_store
 ->fw_load_abort
   ->__fw_load_abort
     ->fw_state_aborted
       ->__fw_state_set
         ->complete_all

in __fw_state_set,
...
111    if (status == FW_STATUS_DONE || status == FW_STATUS_ABORTED)
112       complete_all(&fw_st->completion);
-------------------------------------------
BTW,the double abort issue would not cause kernel panic or create an issue,
but slow down it sometimes.The change is just a minor optimization.

Signed-off-by: Junyong Sun 
Acked-by: Luis Chamberlain 
Link: https://lore.kernel.org/r/1583202968-28792-1-git-send-email-sunjunyong@xiaomi.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

drivers/base/memory.c: indicate all memory blocks as removable

2020-04-01T09:02:02+00:00

commit 53cdc1cb29e87ce5a61de5bb393eb08925d14ede upstream.

We see multiple issues with the implementation/interface to compute
whether a memory block can be offlined (exposed via
/sys/devices/system/memory/memoryX/removable) and would like to simplify
it (remove the implementation).

1. It runs basically lockless. While this might be good for performance,
   we see possible races with memory offlining that will require at
   least some sort of locking to fix.

2. Nowadays, more false positives are possible. No arch-specific checks
   are performed that validate if memory offlining will not be denied
   right away (and such check will require locking). For example, arm64
   won't allow to offline any memory block that was added during boot -
   which will imply a very high error rate. Other archs have other
   constraints.

3. The interface is inherently racy. E.g., if a memory block is detected
   to be removable (and was not a false positive at that time), there is
   still no guarantee that offlining will actually succeed. So any
   caller already has to deal with false positives.

4. It is unclear which performance benefit this interface actually
   provides. The introducing commit 5c755e9fd813 ("memory-hotplug: add
   sysfs removable attribute for hotplug memory remove") mentioned

	"A user-level agent must be able to identify which sections
	 of memory are likely to be removable before attempting the
	 potentially expensive operation."

   However, no actual performance comparison was included.

Known users:

 - lsmem: Will group memory blocks based on the "removable" property. [1]

 - chmem: Indirect user. It has a RANGE mode where one can specify
          removable ranges identified via lsmem to be offlined. However,
          it also has a "SIZE" mode, which allows a sysadmin to skip the
          manual "identify removable blocks" step. [2]

 - powerpc-utils: Uses the "removable" attribute to skip some memory
          blocks right away when trying to find some to offline+remove.
          However, with ballooning enabled, it already skips this
          information completely (because it once resulted in many false
          negatives). Therefore, the implementation can deal with false
          positives properly already. [3]

According to Nathan Fontenot, DLPAR on powerpc is nowadays no longer
driven from userspace via the drmgr command (powerpc-utils).  Nowadays
it's managed in the kernel - including onlining/offlining of memory
blocks - triggered by drmgr writing to /sys/kernel/dlpar.  So the
affected legacy userspace handling is only active on old kernels.  Only
very old versions of drmgr on a new kernel (unlikely) might execute
slower - totally acceptable.

With CONFIG_MEMORY_HOTREMOVE, always indicating "removable" should not
break any user space tool.  We implement a very bad heuristic now.
Without CONFIG_MEMORY_HOTREMOVE we cannot offline anything, so report
"not removable" as before.

Original discussion can be found in [4] ("[PATCH RFC v1] mm:
is_mem_section_removable() overhaul").

Other users of is_mem_section_removable() will be removed next, so that
we can remove is_mem_section_removable() completely.

[1] http://man7.org/linux/man-pages/man1/lsmem.1.html
[2] http://man7.org/linux/man-pages/man8/chmem.8.html
[3] https://github.com/ibm-power-utilities/powerpc-utils
[4] https://lkml.kernel.org/r/20200117105759.27905-1-david@redhat.com

Also, this patch probably fixes a crash reported by Steve.
http://lkml.kernel.org/r/CAPcyv4jpdaNvJ67SkjyUJLBnBnXXQv686BiVW042g03FUmWLXw@mail.gmail.com

Reported-by: "Scargall, Steve" 
Suggested-by: Michal Hocko 
Signed-off-by: David Hildenbrand 
Signed-off-by: Andrew Morton 
Reviewed-by: Nathan Fontenot 
Acked-by: Michal Hocko 
Cc: Dan Williams 
Cc: Greg Kroah-Hartman 
Cc: "Rafael J. Wysocki" 
Cc: Badari Pulavarty 
Cc: Robert Jennings 
Cc: Heiko Carstens 
Cc: Karel Zak 
Cc: 
Link: http://lkml.kernel.org/r/20200128093542.6908-1-david@redhat.com
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman