linux-stable.git/drivers/base, branch v4.9.231

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

2020-06-30T19:38:24+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

2020-06-11T07:22:23+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

2020-05-27T14:41:46+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

driver core: platform: fix u32 greater or equal to zero comparison

2020-02-28T14:42:37+00:00

[ Upstream commit 0707cfa5c3ef58effb143db9db6d6e20503f9dec ]

Currently the check that a u32 variable i is >= 0 is always true because
the unsigned variable will never be negative, causing the loop to run
forever.  Fix this by changing the pre-decrement check to a zero check on
i followed by a decrement of i.

Addresses-Coverity: ("Unsigned compared against 0")
Fixes: 39cc539f90d0 ("driver core: platform: Prevent resouce overflow from causing infinite loops")
Signed-off-by: Colin Ian King 
Reviewed-by: Rafael J. Wysocki 
Link: https://lore.kernel.org/r/20200116175758.88396-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

driver core: Print device when resources present in really_probe()

2020-02-28T14:42:32+00:00

[ Upstream commit 7c35e699c88bd60734277b26962783c60e04b494 ]

If a device already has devres items attached before probing, a warning
backtrace is printed.  However, this backtrace does not reveal the
offending device, leaving the user uninformed.  Furthermore, using
WARN_ON() causes systems with panic-on-warn to reboot.

Fix this by replacing the WARN_ON() by a dev_crit() message.
Abort probing the device, to prevent doing more damage to the device's
resources.

Signed-off-by: Geert Uytterhoeven 
Link: https://lore.kernel.org/r/20191206132219.28908-1-geert+renesas@glider.be
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

driver core: platform: Prevent resouce overflow from causing infinite loops

2020-02-28T14:42:32+00:00

[ Upstream commit 39cc539f90d035a293240c9443af50be55ee81b8 ]

num_resources in the platform_device struct is declared as a u32.  The
for loops that iterate over num_resources use an int as the counter,
which can cause infinite loops on architectures with smaller ints.
Change the loop counters to u32.

Signed-off-by: Simon Schwartz 
Link: https://lore.kernel.org/r/2201ce63a2a171ffd2ed14e867875316efcf71db.camel@theschwartz.xyz
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

drivers/base/platform.c: kmemleak ignore a known leak

2019-12-05T14:35:04+00:00

[ Upstream commit 967d3010df8b6f6f9aa95c198edc5fe3646ebf36 ]

unreferenced object 0xffff808ec6dc5a80 (size 128):
  comm "swapper/0", pid 1, jiffies 4294938063 (age 2560.530s)
  hex dump (first 32 bytes):
    ff ff ff ff 00 00 00 00 6b 6b 6b 6b 6b 6b 6b 6b  ........kkkkkkkk
    6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b  kkkkkkkkkkkkkkkk
  backtrace:
    [<00000000476dcf8c>] kmem_cache_alloc_trace+0x430/0x500
    [<000000004f708d37>] platform_device_register_full+0xbc/0x1e8
    [<000000006c2a7ec7>] acpi_create_platform_device+0x370/0x450
    [<00000000ef135642>] acpi_default_enumeration+0x34/0x78
    [<000000003bd9a052>] acpi_bus_attach+0x2dc/0x3e0
    [<000000003cf4f7f2>] acpi_bus_attach+0x108/0x3e0
    [<000000003cf4f7f2>] acpi_bus_attach+0x108/0x3e0
    [<000000002968643e>] acpi_bus_scan+0xb0/0x110
    [<0000000010dd0bd7>] acpi_scan_init+0x1a8/0x410
    [<00000000965b3c5a>] acpi_init+0x408/0x49c
    [<00000000ed4b9fe2>] do_one_initcall+0x178/0x7f4
    [<00000000a5ac5a74>] kernel_init_freeable+0x9d4/0xa9c
    [<0000000070ea6c15>] kernel_init+0x18/0x138
    [<00000000fb8fff06>] ret_from_fork+0x10/0x1c
    [<0000000041273a0d>] 0xffffffffffffffff

Then, faddr2line pointed out this line,

/*
 * This memory isn't freed when the device is put,
 * I don't have a nice idea for that though.  Conceptually
 * dma_mask in struct device should not be a pointer.
 * See http://thread.gmane.org/gmane.linux.kernel.pci/9081
 */
pdev->dev.dma_mask =
	kmalloc(sizeof(*pdev->dev.dma_mask), GFP_KERNEL);

Since this leak has existed for more than 8 years and it does not
reference other parts of the memory, let kmemleak ignore it, so users
don't need to waste time reporting this in the future.

Link: http://lkml.kernel.org/r/20181206160751.36211-1-cai@gmx.us
Signed-off-by: Qian Cai 
Reviewed-by: Andrew Morton 
Cc: Greg Kroah-Hartman 
Cc: "Rafael J . Wysocki" 
Cc: Catalin Marinas 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Sasha Levin

mm/memory_hotplug: Do not unlock when fails to take the device_hotplug_lock

2019-11-28T17:29:02+00:00

[ Upstream commit d2ab99403ee00d8014e651728a4702ea1ae5e52c ]

When adding the memory by probing memory block in sysfs interface, there is an
obvious issue that we will unlock the device_hotplug_lock when fails to takes it.

That issue was introduced in Commit 8df1d0e4a265
("mm/memory_hotplug: make add_memory() take the device_hotplug_lock")

We should drop out in time when fails to take the device_hotplug_lock.

Fixes: 8df1d0e4a265 ("mm/memory_hotplug: make add_memory() take the device_hotplug_lock")
Reported-by: Yang yingliang 
Signed-off-by: zhong jiang 
Reviewed-by: Oscar Salvador 
Reviewed-by: David Hildenbrand 
Acked-by: Michal Hocko 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

mm/memory_hotplug: make add_memory() take the device_hotplug_lock

2019-11-28T17:28:51+00:00

[ Upstream commit 8df1d0e4a265f25dc1e7e7624ccdbcb4a6630c89 ]

add_memory() currently does not take the device_hotplug_lock, however
is aleady called under the lock from
	arch/powerpc/platforms/pseries/hotplug-memory.c
	drivers/acpi/acpi_memhotplug.c
to synchronize against CPU hot-remove and similar.

In general, we should hold the device_hotplug_lock when adding memory to
synchronize against online/offline request (e.g.  from user space) - which
already resulted in lock inversions due to device_lock() and
mem_hotplug_lock - see 30467e0b3be ("mm, hotplug: fix concurrent memory
hot-add deadlock").  add_memory()/add_memory_resource() will create memory
block devices, so this really feels like the right thing to do.

Holding the device_hotplug_lock makes sure that a memory block device
can really only be accessed (e.g. via .online/.state) from user space,
once the memory has been fully added to the system.

The lock is not held yet in
	drivers/xen/balloon.c
	arch/powerpc/platforms/powernv/memtrace.c
	drivers/s390/char/sclp_cmd.c
	drivers/hv/hv_balloon.c
So, let's either use the locked variants or take the lock.

Don't export add_memory_resource(), as it once was exported to be used by
XEN, which is never built as a module.  If somebody requires it, we also
have to export a locked variant (as device_hotplug_lock is never
exported).

Link: http://lkml.kernel.org/r/20180925091457.28651-3-david@redhat.com
Signed-off-by: David Hildenbrand 
Reviewed-by: Pavel Tatashin 
Reviewed-by: Rafael J. Wysocki 
Reviewed-by: Rashmica Gupta 
Reviewed-by: Oscar Salvador 
Cc: Benjamin Herrenschmidt 
Cc: Paul Mackerras 
Cc: Michael Ellerman 
Cc: "Rafael J. Wysocki" 
Cc: Len Brown 
Cc: Greg Kroah-Hartman 
Cc: Boris Ostrovsky 
Cc: Juergen Gross 
Cc: Nathan Fontenot 
Cc: John Allen 
Cc: Michal Hocko 
Cc: Dan Williams 
Cc: Joonsoo Kim 
Cc: Vlastimil Babka 
Cc: Mathieu Malaterre 
Cc: Pavel Tatashin 
Cc: YASUAKI ISHIMATSU 
Cc: Balbir Singh 
Cc: Haiyang Zhang 
Cc: Heiko Carstens 
Cc: Jonathan Corbet 
Cc: Kate Stewart 
Cc: "K. Y. Srinivasan" 
Cc: Martin Schwidefsky 
Cc: Michael Neuling 
Cc: Philippe Ombredanne 
Cc: Stephen Hemminger 
Cc: Thomas Gleixner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Sasha Levin

component: fix loop condition to call unbind() if bind() fails

2019-11-25T08:52:23+00:00

[ Upstream commit bdae566d5d9733b6e32b378668b84eadf28a94d4 ]

During component_bind_all(), if bind() fails for any
particular component associated with a master, unbind()
should be called for all previous components in that
master's match array, whose bind() might have completed
successfully. As per the current logic, if bind() fails
for the component at position 'n' in the master's match
array, it would start calling unbind() from component in
'n'th position itself and work backwards, and will always
skip calling unbind() for component in 0th position in the
master's match array.
Fix this by updating the loop condition, and the logic to
refer to the components in master's match array, so that
unbind() is called for all components starting from 'n-1'st
position in the array, until (and including) component in
0th position.

Signed-off-by: Banajit Goswami 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin