x86/topology: Use total_cpus not nr_cpu_ids for logical packages

nr_cpu_ids can be limited on the command line via nr_cpus=. That can break the
logical package management because it results in a smaller number of packages,
but the cpus to online are occupying the full package space as the hyper
threads are enumerated after the physical cores typically.

total_cpus is the real possible cpu space not limited by nr_cpus command line
and gives us the proper number of packages.

Reported-by: Mike Galbraith <[email protected]>
Fixes: 1f12e32 ("x86/topology: Create logical package id")
Signed-off-by: Thomas Gleixner <[email protected]>
Cc: Xiong Zhou <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Borislav Petkov <[email protected]>
Cc: Andreas Herrmann <[email protected]>
Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1603181254330.3978@nanos

arch/x86/kernel/smpboot.c

@@ -325,9 +325,14 @@ static void __init smp_init_package_map(void)
 	 * By not including this we'll sometimes over-estimate the number of
 	 * logical packages by the amount of !present siblings, but this is
 	 * still better than MAX_LOCAL_APIC.
+	 *
+	 * We use total_cpus not nr_cpu_ids because nr_cpu_ids can be limited
+	 * on the command line leading to a similar issue as the HT disable
+	 * problem because the hyperthreads are usually enumerated after the
+	 * primary cores.
 	 */
 	ncpus = boot_cpu_data.x86_max_cores;
-	__max_logical_packages = DIV_ROUND_UP(nr_cpu_ids, ncpus);
+	__max_logical_packages = DIV_ROUND_UP(total_cpus, ncpus);
 
 	/*
 	 * Possibly larger than what we need as the number of apic ids per