x86/cpu_entry_area: Prevent wraparound in setup_cpu_entry_area_ptes() on 32bit

commit f6c4fd50 upstream.

The loop which populates the CPU entry area PMDs can wrap around on 32bit
machines when the number of CPUs is small.

It worked wonderful for NR_CPUS=64 for whatever reason and the moron who
wrote that code did not bother to test it with !SMP.

Check for the wraparound to fix it.

Fixes: 92a0f81d ("x86/cpu_entry_area: Move it out of the fixmap")
Reported-by: kernel test robot <fengguang.wu@intel.com>
Signed-off-by: Thomas "Feels stupid" Gleixner <tglx@linutronix.de>
Tested-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arch/x86/mm/cpu_entry_area.c

@@ -122,7 +122,8 @@ static __init void setup_cpu_entry_area_ptes(void)
 	start = CPU_ENTRY_AREA_BASE;
 	end = start + CPU_ENTRY_AREA_MAP_SIZE;
 
-	for (; start < end; start += PMD_SIZE)
+	/* Careful here: start + PMD_SIZE might wrap around */
+	for (; start < end && start >= CPU_ENTRY_AREA_BASE; start += PMD_SIZE)
 		populate_extra_pte(start);
 #endif
 }