提交 dd988528 编写于 作者: F Fernando Luis Vazquez Cao 提交者: Linus Torvalds

Use the APIC to determine the hardware processor id - x86_64

hard_smp_processor_id used to be just a macro that hard-coded
hard_smp_processor_id to 0 in the non SMP case.  When booting non SMP kernels
on hardware where the boot ioapic id is not 0 this turns out to be a problem.
This is happens frequently in the case of kdump and once in a great while in
the case of real hardware.

Use the APIC to determine the hardware processor id in both UP and SMP kernels
to fix this issue.

Notice that hard_smp_processor_id is only used by SMP code or by code that
works with apics so we do not need to handle the case when apics are not
present and hard_smp_processor_id should never be called there.
Signed-off-by: NFernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Cc: "Luck, Tony" <tony.luck@intel.com>
Acked-by: NAndi Kleen <ak@suse.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
上级 a36166c6
......@@ -57,12 +57,6 @@ static inline int num_booting_cpus(void)
#define raw_smp_processor_id() read_pda(cpunumber)
static inline int hard_smp_processor_id(void)
{
/* we don't want to mark this access volatile - bad code generation */
return GET_APIC_ID(*(unsigned int *)(APIC_BASE+APIC_ID));
}
extern int __cpu_disable(void);
extern void __cpu_die(unsigned int cpu);
extern void prefill_possible_map(void);
......@@ -71,10 +65,14 @@ extern unsigned __cpuinitdata disabled_cpus;
#define NO_PROC_ID 0xFF /* No processor magic marker */
#else /* CONFIG_SMP */
#define hard_smp_processor_id() 0
#endif /* CONFIG_SMP */
static inline int hard_smp_processor_id(void)
{
/* we don't want to mark this access volatile - bad code generation */
return GET_APIC_ID(*(unsigned int *)(APIC_BASE+APIC_ID));
}
/*
* Some lowlevel functions might want to know about
* the real APIC ID <-> CPU # mapping.
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册