At the moment, UBSAN report will be serialized using a spin_lock().  On
RT-systems, spinlocks are turned to rt_spin_lock and may sleep.  This will
result to the following splat if the undefined behavior is in a context
that can sleep:

[ 6951.484876] BUG: sleeping function called from invalid context at /src/linux/kernel/locking/rtmutex.c:968
[ 6951.484882] in_atomic(): 1, irqs_disabled(): 128, pid: 3447, name: make
[ 6951.484884] 1 lock held by make/3447:
[ 6951.484885]  #0: 000000009a966332 (&mm->mmap_sem){++++}, at: do_page_fault+0x140/0x4f8
[ 6951.484895] irq event stamp: 6284
[ 6951.484896] hardirqs last  enabled at (6283): [<ffff000011326520>] _raw_spin_unlock_irqrestore+0x90/0xa0
[ 6951.484901] hardirqs last disabled at (6284): [<ffff0000113262b0>] _raw_spin_lock_irqsave+0x30/0x78
[ 6951.484902] softirqs last  enabled at (2430): [<ffff000010088ef8>] fpsimd_restore_current_state+0x60/0xe8
[ 6951.484905] softirqs last disabled at (2427): [<ffff000010088ec0>] fpsimd_restore_current_state+0x28/0xe8
[ 6951.484907] Preemption disabled at:
[ 6951.484907] [<ffff000011324a4c>] rt_mutex_futex_unlock+0x4c/0xb0
[ 6951.484911] CPU: 3 PID: 3447 Comm: make Tainted: G        W         5.2.14-rt7-01890-ge6e057589653 #911
[ 6951.484913] Call trace:
[ 6951.484913]  dump_backtrace+0x0/0x148
[ 6951.484915]  show_stack+0x14/0x20
[ 6951.484917]  dump_stack+0xbc/0x104
[ 6951.484919]  ___might_sleep+0x154/0x210
[ 6951.484921]  rt_spin_lock+0x68/0xa0
[ 6951.484922]  ubsan_prologue+0x30/0x68
[ 6951.484924]  handle_overflow+0x64/0xe0
[ 6951.484926]  __ubsan_handle_add_overflow+0x10/0x18
[ 6951.484927]  __lock_acquire+0x1c28/0x2a28
[ 6951.484929]  lock_acquire+0xf0/0x370
[ 6951.484931]  _raw_spin_lock_irqsave+0x58/0x78
[ 6951.484932]  rt_mutex_futex_unlock+0x4c/0xb0
[ 6951.484933]  rt_spin_unlock+0x28/0x70
[ 6951.484934]  get_page_from_freelist+0x428/0x2b60
[ 6951.484936]  __alloc_pages_nodemask+0x174/0x1708
[ 6951.484938]  alloc_pages_vma+0x1ac/0x238
[ 6951.484940]  __handle_mm_fault+0x4ac/0x10b0
[ 6951.484941]  handle_mm_fault+0x1d8/0x3b0
[ 6951.484942]  do_page_fault+0x1c8/0x4f8
[ 6951.484943]  do_translation_fault+0xb8/0xe0
[ 6951.484945]  do_mem_abort+0x3c/0x98
[ 6951.484946]  el0_da+0x20/0x24

The spin_lock() will protect against multiple CPUs to output a report
together, I guess to prevent them from being interleaved.  However, they
can still interleave with other messages (and even splat from
__might_sleep).

So the lock usefulness seems pretty limited.  Rather than trying to
accomodate RT-system by switching to a raw_spin_lock(), the lock is now
completely dropped.

Link: http://lkml.kernel.org/r/20190920100835.14999-1-julien.grall@arm.com


Signed-off-by: Julien Grall <julien.grall@arm.com>
Reported-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>