include/linux/percpu_counter.h: race in uniprocessor percpu_counter_add()

The percpu interface is supposed to be preempt and irq safe.

But:
The uniprocessor implementation of percpu_counter_add() is not irq safe:
if an interrupt happens during the +=, then the result is undefined.

Therefore: switch from preempt_disable() to local_irq_save().
This prevents interrupts from interrupting the +=, and as a side effect
prevents preemption.

Link: https://lkml.kernel.org/r/20221216150441.200533-2-manfred@colorfullife.com
Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
Cc: "Sun, Jiebin" <jiebin.sun@intel.com>
Cc: <1vier1@web.de>
Cc: Alexander Sverdlin <alexander.sverdlin@siemens.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h
index a3aae8d57a421f454e26a97b6babdcd8f65827cd..521a733e21a920cc04a401788048143ecb64f236 100644 (file)
--- a/include/linux/percpu_counter.h
+++ b/include/linux/percpu_counter.h
@@ -152,9 +152,11 @@ __percpu_counter_compare(struct percpu_counter *fbc, s64 rhs, s32 batch)
 static inline void
 percpu_counter_add(struct percpu_counter *fbc, s64 amount)
 {
-       preempt_disable();
+       unsigned long flags;
+
+       local_irq_save(flags);
        fbc->count += amount;
-       preempt_enable();
+       local_irq_restore(flags);
 }
 
 /* non-SMP percpu_counter_add_local is the same with percpu_counter_add */