x86/mce/dev-mcelog: Fix potential memory access error
Subsystems (3)
PATCH 1/1 x86/mce/dev-mcelog: Fix potential memory access error
Using set_bit() to set a bit in an integer is not a good idea, since the function expects an unsigned long as argument, which can be 64bit wide.
Coverity reports this problem as
High:Out-of-bounds access(INCOMPATIBLE_CAST)
CWE119: Out-of-bounds access to a scalar
Pointer "&mcelog->flags" points to an object whose effective type is "unsigned int" (32 bits, unsigned) but is dereferenced as a wider "unsigned long" (64 bits, unsigned). This may lead to memory corruption.
/home/heyuan.shy/git-repo/linux/arch/x86/kernel/cpu/mce/dev-mcelog.c:dev_mce_log
Just use BIT instead.
--- arch/x86/kernel/cpu/mce/dev-mcelog.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kernel/cpu/mce/dev-mcelog.c +1 -1
@@ -51,7 +51,7 @@ static int dev_mce_log(struct notifier_block *nb, unsigned long val,* earlier errors are the more interesting ones: */ if (entry >= mcelog->len) { -set_bit (MCE_OVERFLOW , (unsigned long *)&mcelog->flags); + mcelog->flags |=BIT (MCE_OVERFLOW ); goto unlock; }
Signature · 1 line
1.8.3.1
Luck, Tony
maintainer- set_bit(MCE_OVERFLOW, (unsigned long *)&mcelog->flags);
+ mcelog->flags |= BIT(MCE_OVERFLOW);
set_bit() is an atomic operation because it might race with the code to get and clear this bit:
do {
flags = mcelog->flags;
} while (cmpxchg(&mcelog->flags, flags, 0) != flags);Originally this was needed because mcelog could be called from #MC context.
That's all changed now and the machine check notifier chain routines are called in a kernel thread. So some other locking (mutex?) could be used to protect access to mcelog->flags.
-Tony