linux.git/drivers/base, branch v6.6.16

arm64: irq: set the correct node for VMAP stack

2024-02-05T20:14:14+00:00

[ Upstream commit 75b5e0bf90bffaca4b1f19114065dc59f5cc161f ]

In current code, init_irq_stacks() will call cpu_to_node().
The cpu_to_node() depends on percpu "numa_node" which is initialized in:
     arch_call_rest_init() --> rest_init() -- kernel_init()
	--> kernel_init_freeable() --> smp_prepare_cpus()

But init_irq_stacks() is called in init_IRQ() which is before
arch_call_rest_init().

So in init_irq_stacks(), the cpu_to_node() does not work, it
always return 0. In NUMA, it makes the node 1 cpu accesses the IRQ stack which
is in the node 0.

This patch fixes it by:
  1.) export the early_cpu_to_node(), and use it in the init_irq_stacks().
  2.) change init_irq_stacks() to __init function.

Reviewed-by: Catalin Marinas 
Signed-off-by: Huang Shijie 
Link: https://lore.kernel.org/r/20231124031513.81548-1-shijie@os.amperecomputing.com
Signed-off-by: Will Deacon 
Signed-off-by: Sasha Levin

rtc: Extend timeout for waiting for UIP to clear to 1s

2024-02-01T00:18:56+00:00

commit cef9ecc8e938dd48a560f7dd9be1246359248d20 upstream.

Specs don't say anything about UIP being cleared within 10ms. They
only say that UIP won't occur for another 244uS. If a long NMI occurs
while UIP is still updating it might not be possible to get valid
data in 10ms.

This has been observed in the wild that around s2idle some calls can
take up to 480ms before UIP is clear.

Adjust callers from outside an interrupt context to wait for up to a
1s instead of 10ms.

Cc:   # 6.1.y
Fixes: ec5895c0f2d8 ("rtc: mc146818-lib: extract mc146818_avoid_UIP")
Reported-by: Carsten Hatger 
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217626
Tested-by: Mateusz Jończyk 
Reviewed-by: Mateusz Jończyk 
Acked-by: Mateusz Jończyk 
Signed-off-by: Mario Limonciello 
Link: https://lore.kernel.org/r/20231128053653.101798-5-mario.limonciello@amd.com
Signed-off-by: Alexandre Belloni 
Signed-off-by: Greg Kroah-Hartman

rtc: Add support for configuring the UIP timeout for RTC reads

2024-02-01T00:18:56+00:00

commit 120931db07b49252aba2073096b595482d71857c upstream.

The UIP timeout is hardcoded to 10ms for all RTC reads, but in some
contexts this might not be enough time. Add a timeout parameter to
mc146818_get_time() and mc146818_get_time_callback().

If UIP timeout is configured by caller to be >=100 ms and a call
takes this long, log a warning.

Make all callers use 10ms to ensure no functional changes.

Cc:   # 6.1.y
Fixes: ec5895c0f2d8 ("rtc: mc146818-lib: extract mc146818_avoid_UIP")
Signed-off-by: Mario Limonciello 
Tested-by: Mateusz Jończyk 
Reviewed-by: Mateusz Jończyk 
Acked-by: Mateusz Jończyk 
Link: https://lore.kernel.org/r/20231128053653.101798-4-mario.limonciello@amd.com
Signed-off-by: Alexandre Belloni 
Signed-off-by: Greg Kroah-Hartman

PM: sleep: Fix possible deadlocks in core system-wide PM code

2024-02-01T00:18:49+00:00

commit 7839d0078e0d5e6cc2fa0b0dfbee71de74f1e557 upstream.

It is reported that in low-memory situations the system-wide resume core
code deadlocks, because async_schedule_dev() executes its argument
function synchronously if it cannot allocate memory (and not only in
that case) and that function attempts to acquire a mutex that is already
held.  Executing the argument function synchronously from within
dpm_async_fn() may also be problematic for ordering reasons (it may
cause a consumer device's resume callback to be invoked before a
requisite supplier device's one, for example).

Address this by changing the code in question to use
async_schedule_dev_nocall() for scheduling the asynchronous
execution of device suspend and resume functions and to directly
run them synchronously if async_schedule_dev_nocall() returns false.

Link: https://lore.kernel.org/linux-pm/ZYvjiqX6EsL15moe@perf/
Reported-by: Youngmin Nam 
Signed-off-by: Rafael J. Wysocki 
Reviewed-by: Stanislaw Gruszka 
Tested-by: Youngmin Nam 
Reviewed-by: Ulf Hansson 
Cc: 5.7+  # 5.7+: 6aa09a5bccd8 async: Split async_schedule_node_domain()
Cc: 5.7+  # 5.7+: 7d4b5d7a37bd async: Introduce async_schedule_dev_nocall()
Cc: 5.7+  # 5.7+
Signed-off-by: Greg Kroah-Hartman

software node: Let args be NULL in software_node_get_reference_args

2024-01-25T23:35:51+00:00

[ Upstream commit 1eaea4b3604eb9ca7d9a1e73d88fc121bb4061f5 ]

fwnode_get_property_reference_args() may not be called with args argument
NULL and while OF already supports this. Add the missing NULL check.

The purpose is to be able to count the references.

Fixes: b06184acf751 ("software node: Add software_node_get_reference_args()")
Signed-off-by: Sakari Ailus 
Reviewed-by: Andy Shevchenko 
Reviewed-by: Heikki Krogerus 
Link: https://lore.kernel.org/r/20231109101010.1329587-3-sakari.ailus@linux.intel.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

base/node.c: initialize the accessor list before registering

2024-01-25T23:35:51+00:00

[ Upstream commit 48b5928e18dc27e05cab3dc4c78cd8a15baaf1e5 ]

The current code registers the node as available in the node array
before initializing the accessor list.  This makes it so that
anything which might access the accessor list as a result of
allocations will cause an undefined memory access.

In one example, an extension to access hmat data during interleave
caused this undefined access as a result of a bulk allocation
that occurs during node initialization but before the accessor
list is initialized.

Initialize the accessor list before making the node generally
available to the global system.

Fixes: 08d9dbe72b1f ("node: Link memory nodes to their compute nodes")
Signed-off-by: Gregory Price 
Link: https://lore.kernel.org/r/20231030044239.971756-1-gregory.price@memverge.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

class: fix use-after-free in class_register()

2024-01-25T23:35:41+00:00

commit 93ec4a3b76404bce01bd5c9032bef5df6feb1d62 upstream.

The lock_class_key is still registered and can be found in
lock_keys_hash hlist after subsys_private is freed in error
handler path.A task who iterate over the lock_keys_hash
later may cause use-after-free.So fix that up and unregister
the lock_class_key before kfree(cp).

On our platform, a driver fails to kset_register because of
creating duplicate filename '/class/xxx'.With Kasan enabled,
it prints a invalid-access bug report.

KASAN bug report:

BUG: KASAN: invalid-access in lockdep_register_key+0x19c/0x1bc
Write of size 8 at addr 15ffff808b8c0368 by task modprobe/252
Pointer tag: [15], memory tag: [fe]

CPU: 7 PID: 252 Comm: modprobe Tainted: G        W
 6.6.0-mainline-maybe-dirty #1

Call trace:
dump_backtrace+0x1b0/0x1e4
show_stack+0x2c/0x40
dump_stack_lvl+0xac/0xe0
print_report+0x18c/0x4d8
kasan_report+0xe8/0x148
__hwasan_store8_noabort+0x88/0x98
lockdep_register_key+0x19c/0x1bc
class_register+0x94/0x1ec
init_module+0xbc/0xf48 [rfkill]
do_one_initcall+0x17c/0x72c
do_init_module+0x19c/0x3f8
...
Memory state around the buggy address:
ffffff808b8c0100: 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a 8a
ffffff808b8c0200: 8a 8a 8a 8a 8a 8a 8a 8a fe fe fe fe fe fe fe fe
>ffffff808b8c0300: fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe
                                     ^
ffffff808b8c0400: 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03

As CONFIG_KASAN_GENERIC is not set, Kasan reports invalid-access
not use-after-free here.In this case, modprobe is manipulating
the corrupted lock_keys_hash hlish where lock_class_key is already
freed before.

It's worth noting that this only can happen if lockdep is enabled,
which is not true for normal system.

Fixes: dcfbb67e48a2 ("driver core: class: use lock_class_key already present in struct subsys_private")
Cc: stable 
Signed-off-by: Jing Xia 
Signed-off-by: Xuewen Yan 
Link: https://lore.kernel.org/r/20231220024603.186078-1-jing.xia@unisoc.com
Signed-off-by: Greg Kroah-Hartman

devcoredump: Send uevent once devcd is ready

2023-12-13T17:45:33+00:00

commit af54d778a03853801d681c98c0c2a6c316ef9ca7 upstream.

dev_coredumpm() creates a devcoredump device and adds it
to the core kernel framework which eventually end up
sending uevent to the user space and later creates a
symbolic link to the failed device. An application
running in userspace may be interested in this symbolic
link to get the name of the failed device.

In a issue scenario, once uevent sent to the user space
it start reading '/sys/class/devcoredump/devcdX/failing_device'
to get the actual name of the device which might not been
created and it is in its path of creation.

To fix this, suppress sending uevent till the failing device
symbolic link gets created and send uevent once symbolic
link is created successfully.

Fixes: 833c95456a70 ("device coredump: add new device coredump class")
Signed-off-by: Mukesh Ojha 
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1700232572-25823-1-git-send-email-quic_mojha@quicinc.com
Signed-off-by: Greg Kroah-Hartman

mm/memory_hotplug: add missing mem_hotplug_lock

2023-12-13T17:45:24+00:00

commit 001002e73712cdf6b8d9a103648cda3040ad7647 upstream.

From Documentation/core-api/memory-hotplug.rst:
When adding/removing/onlining/offlining memory or adding/removing
heterogeneous/device memory, we should always hold the mem_hotplug_lock
in write mode to serialise memory hotplug (e.g. access to global/zone
variables).

mhp_(de)init_memmap_on_memory() functions can change zone stats and
struct page content, but they are currently called w/o the
mem_hotplug_lock.

When memory block is being offlined and when kmemleak goes through each
populated zone, the following theoretical race conditions could occur:
CPU 0:					     | CPU 1:
memory_offline()			     |
-> offline_pages()			     |
	-> mem_hotplug_begin()		     |
	   ...				     |
	-> mem_hotplug_done()		     |
					     | kmemleak_scan()
					     | -> get_online_mems()
					     |    ...
-> mhp_deinit_memmap_on_memory()	     |
  [not protected by mem_hotplug_begin/done()]|
  Marks memory section as offline,	     |   Retrieves zone_start_pfn
  poisons vmemmap struct pages and updates   |   and struct page members.
  the zone related data			     |
   					     |    ...
   					     | -> put_online_mems()

Fix this by ensuring mem_hotplug_lock is taken before performing
mhp_init_memmap_on_memory().  Also ensure that
mhp_deinit_memmap_on_memory() holds the lock.

online/offline_pages() are currently only called from
memory_block_online/offline(), so it is safe to move the locking there.

Link: https://lkml.kernel.org/r/20231120145354.308999-2-sumanthk@linux.ibm.com
Fixes: a08a2ae34613 ("mm,memory_hotplug: allocate memmap from the added memory range")
Signed-off-by: Sumanth Korikkar 
Reviewed-by: Gerald Schaefer 
Acked-by: David Hildenbrand 
Cc: Alexander Gordeev 
Cc: Aneesh Kumar K.V 
Cc: Anshuman Khandual 
Cc: Heiko Carstens 
Cc: Michal Hocko 
Cc: Oscar Salvador 
Cc: Vasily Gorbik 
Cc: kernel test robot 
Cc: 	[5.15+]
Signed-off-by: Andrew Morton 
Signed-off-by: Greg Kroah-Hartman

drivers/base/cpu: crash data showing should depends on KEXEC_CORE

2023-12-13T17:45:24+00:00

commit 4e9e2e4c65136dfd32dd0afe555961433d1cf906 upstream.

After commit 88a6f8994421 ("crash: memory and CPU hotplug sysfs
attributes"), on x86_64, if only below kernel configs related to kdump are
set, compiling error are triggered.

----
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_CRASH_DUMP=y
CONFIG_CRASH_HOTPLUG=y
------

------------------------------------------------------
drivers/base/cpu.c: In function `crash_hotplug_show':
drivers/base/cpu.c:309:40: error: implicit declaration of function `crash_hotplug_cpu_support'; did you mean `crash_hotplug_show'? [-Werror=implicit-function-declaration]
  309 |         return sysfs_emit(buf, "%d\n", crash_hotplug_cpu_support());
      |                                        ^~~~~~~~~~~~~~~~~~~~~~~~~
      |                                        crash_hotplug_show
cc1: some warnings being treated as errors
------------------------------------------------------

CONFIG_KEXEC is used to enable kexec_load interface, the
crash_notes/crash_notes_size/crash_hotplug showing depends on
CONFIG_KEXEC is incorrect. It should depend on KEXEC_CORE instead.

Fix it now.

Link: https://lkml.kernel.org/r/20231128055248.659808-1-bhe@redhat.com
Fixes: 88a6f8994421 ("crash: memory and CPU hotplug sysfs attributes")
Signed-off-by: Baoquan He 
Tested-by: Ignat Korchagin 	[compile-time only]
Tested-by: Alexander Gordeev 
Reviewed-by: Eric DeVolder 
Cc: 
Signed-off-by: Andrew Morton 
Signed-off-by: Greg Kroah-Hartman