[Xen-devel] Re: [RFC, PATCH 9/24] i386 Vmi smp support

[Chris Wright <chrisw@xxxxxxxxxxxx>]

* Zachary Amsden (zach@xxxxxxxxxx) wrote:
> SMP bootstrapping support.  Just as in the physical platform model,
> the BSP is responsible for initializing the AP state prior to execution.
> The dependence on lots of processor state information is a design choice
> of our implementation.  Conceivably, this could be a hypercall that
> awakens the same start of day state on APs as on the BSP.
> 
> It is likely the AP startup and the start-of-day model will eventually
> merge into a more common interface.
> 
> --- linux-2.6.16-rc5.orig/arch/i386/mach-vmi/smpboot_hooks.c  2006-03-08 
> 11:02:12.000000000 -0800
> +++ linux-2.6.16-rc5/arch/i386/mach-vmi/smpboot_hooks.c       2006-03-08 
> 11:02:16.000000000 -0800
> +#ifdef CONFIG_SMP

couldn't this be:
obj-$(CONFIG_X86_SMP)   += smpboot_hooks.o

> +
> +#ifdef CONFIG_HOTPLUG_CPU
> +#define DEFAULT_SEND_IPI     (1)
> +#else
> +#define DEFAULT_SEND_IPI     (0)
> +#endif
> +
> +int no_broadcast=DEFAULT_SEND_IPI;
> +
> +APState ap;

struct...

> +
> +void __init
> +smpboot_startup_ipi_hook(int phys_apicid, unsigned long start_eip,
> +                         unsigned long start_esp)
> +{
> +        /* We require phys_acpicid to be the cpu number. */
> +        if (hypervisor_found) {

Another nit, inconsistent use of hypervisor_found vs.
vmi_hypervisor_found().

> +                /* Default everything to zero.  This is fine for most GPRs. 
> */
> +                memset(&ap, 0, sizeof(APState));
> +
> +                /* Set up AP's per-cpu GDT. */
> +                memcpy(get_cpu_gdt_table(phys_apicid), cpu_gdt_table,
> +                       GDT_SIZE);
> +                ap.gdtr_limit = GDT_SIZE - 1;
> +                ap.gdtr_base = (unsigned long) 
> get_cpu_gdt_table(phys_apicid);
> +                
> +                ap.idtr_limit = IDT_ENTRIES * 8 - 1;
> +                ap.idtr_base = (unsigned long) idt_table;
> +
> +                ap.ldtr = 0;
> +                
> +                ap.cs = __KERNEL_CS;
> +                ap.eip = (unsigned long) start_eip;
> +                ap.ss = __KERNEL_DS;
> +                ap.esp = (unsigned long) start_esp;
> +                
> +                ap.ds = __USER_DS;
> +                ap.es = __USER_DS;
> +                ap.fs = 0;
> +                ap.gs = 0;
> +                
> +                ap.eflags = 0;
> +                
> +#ifdef CONFIG_X86_PAE
> +                /* efer should match BSP efer. */
> +                if (cpu_has_nx) {
> +                        unsigned l, h;
> +                        rdmsr(MSR_EFER, l, h);
> +                        ap.efer = (unsigned long long) h << 32 | l;
> +                }
> +#endif
> +
> +                ap.cr3 = __pa(swapper_pg_dir);
> +                /* Protected mode, paging, AM, WP, NE, MP. */
> +                ap.cr0 = 0x80050023;
> +                ap.cr4 = mmu_cr4_features;
> +                
> +                vmi_set_initial_ap_state(__pa(&ap), phys_apicid);

I think the big difference here is the use of apicid for processor id.
The whole of apic usage model is quite different from Xen approach.
Need to look more at that.

> +        }
> +}
> +
> +void __init smpboot_pre_start_secondary_hook(void)
> +{
> +        if (vmi_hypervisor_found()) {
> +                *(unsigned long *) trampoline_base = 0xa5a5a5a5;
> +        }
> +}
> +
> +static __init int no_ipi_broadcast(char *str)
> +{
> +     get_option(&str, &no_broadcast);
> +     printk ("Using %s mode\n", no_broadcast ? "No IPI Broadcast" :
> +                                                                             
>         "IPI Broadcast");
> +     return 1;
> +}
> +
> +__setup("no_ipi_broadcast", no_ipi_broadcast);
> +
> +static int __init print_ipi_mode(void)
> +{
> +     printk ("Using IPI %s mode\n", no_broadcast ? "No-Shortcut" :
> +                                                                             
>         "Shortcut");
> +     return 0;
> +}

Be nice to avoid the pure duplication.

> +
> +late_initcall(print_ipi_mode);
> +#endif
> +
> Index: linux-2.6.16-rc5/arch/i386/kernel/smpboot.c
> ===================================================================
> --- linux-2.6.16-rc5.orig/arch/i386/kernel/smpboot.c  2006-03-08 
> 10:53:46.000000000 -0800
> +++ linux-2.6.16-rc5/arch/i386/kernel/smpboot.c       2006-03-08 
> 11:02:16.000000000 -0800
> @@ -111,7 +111,7 @@ EXPORT_SYMBOL(x86_cpu_to_apicid);
>  
>  extern unsigned char trampoline_data [];
>  extern unsigned char trampoline_end  [];
> -static unsigned char *trampoline_base;
> +unsigned char *trampoline_base;
>  static int trampoline_exec;
>  
>  static void map_cpu_to_logical_apicid(void);
> @@ -507,6 +507,7 @@ static void __devinit start_secondary(vo
>        * booting is too fragile that we want to limit the
>        * things done here to the most necessary things.
>        */
> +     smpboot_pre_start_secondary_hook();
>       cpu_init();
>       preempt_disable();
>       smp_callin();
> @@ -782,6 +783,10 @@ wakeup_secondary_cpu(int phys_apicid, un
>       else
>               num_starts = 0;
>  
> +     smpboot_startup_ipi_hook(phys_apicid, (unsigned long) start_secondary,
> +             (unsigned long) stack_start.esp);
> +
> +
>       /*
>        * Run STARTUP IPI loop.
>        */
> Index: linux-2.6.16-rc5/include/asm-i386/mach-default/smpboot_hooks.h
> ===================================================================
> --- linux-2.6.16-rc5.orig/include/asm-i386/mach-default/smpboot_hooks.h       
> 2006-03-08 10:53:46.000000000 -0800
> +++ linux-2.6.16-rc5/include/asm-i386/mach-default/smpboot_hooks.h    
> 2006-03-08 11:02:16.000000000 -0800
> @@ -42,3 +42,15 @@ static inline void smpboot_setup_io_apic
>       if (!skip_ioapic_setup && nr_ioapics)
>               setup_IO_APIC();
>  }
> +
> +static inline void smpboot_startup_ipi_hook(int phys_apicid, 
> +                                            unsigned long start_eip,
> +                                            unsigned long start_esp)
> +{
> +
> +}
> +
> +static inline void smpboot_pre_start_secondary_hook(void)
> +{
> +
> +}
> Index: linux-2.6.16-rc5/include/asm-i386/mach-vmi/smpboot_hooks.h
> ===================================================================
> --- linux-2.6.16-rc5.orig/include/asm-i386/mach-vmi/smpboot_hooks.h   
> 2006-03-08 11:02:12.000000000 -0800
> +++ linux-2.6.16-rc5/include/asm-i386/mach-vmi/smpboot_hooks.h        
> 2006-03-08 11:02:16.000000000 -0800
> @@ -0,0 +1,51 @@
> +/* 
> + * include/asm-i386/mach-default/smpboot_hooks.h

wrong subarch in comment ;-)

> + *
> + * Portions Copyright 2005 VMware, Inc.
> + */
> +
> +static inline void smpboot_clear_io_apic_irqs(void)
> +{
> +     io_apic_irqs = 0;
> +}
> +
> +static inline void smpboot_setup_warm_reset_vector(unsigned long start_eip)
> +{
> +     CMOS_WRITE(0xa, 0xf);
> +     local_flush_tlb();
> +     Dprintk("1.\n");
> +     *((volatile unsigned short *) TRAMPOLINE_HIGH) = start_eip >> 4;
> +     Dprintk("2.\n");
> +     *((volatile unsigned short *) TRAMPOLINE_LOW) = start_eip & 0xf;
> +     Dprintk("3.\n");
> +}
> +
> +static inline void smpboot_restore_warm_reset_vector(void)
> +{
> +     /*
> +      * Install writable page 0 entry to set BIOS data area.
> +      */
> +     local_flush_tlb();
> +
> +     /*
> +      * Paranoid:  Set warm reset code and vector here back
> +      * to default values.
> +      */
> +     CMOS_WRITE(0, 0xf);
> +
> +     *((volatile long *) phys_to_virt(0x467)) = 0;
> +}
> +
> +static inline void smpboot_setup_io_apic(void)
> +{
> +     /*
> +      * Here we can be sure that there is an IO-APIC in the system. Let's
> +      * go and set it up:
> +      */
> +     if (!skip_ioapic_setup && nr_ioapics)
> +             setup_IO_APIC();
> +}
> +
> +extern void smpboot_startup_ipi_hook(int phys_apicid, unsigned long 
> start_eip,
> +                             unsigned long start_esp);
> +extern void smpboot_pre_start_secondary_hook(void);
> Index: linux-2.6.16-rc5/include/asm-i386/mach-vmi/entry_arch.h
> ===================================================================
> --- linux-2.6.16-rc5.orig/include/asm-i386/mach-vmi/entry_arch.h      
> 2006-03-08 11:02:12.000000000 -0800
> +++ linux-2.6.16-rc5/include/asm-i386/mach-vmi/entry_arch.h   2006-03-08 
> 11:02:16.000000000 -0800
> @@ -0,0 +1,34 @@
> +/*
> + * This file is designed to contain the BUILD_INTERRUPT specifications for
> + * all of the extra named interrupt vectors used by the architecture.
> + * Usually this is the Inter Process Interrupts (IPIs)
> + */
> +
> +/*
> + * The following vectors are part of the Linux architecture, there
> + * is no hardware IRQ pin equivalent for them, they are triggered
> + * through the ICC by us (IPIs)
> + */
> +#ifdef CONFIG_X86_SMP
> +BUILD_INTERRUPT(reschedule_interrupt,RESCHEDULE_VECTOR)
> +BUILD_INTERRUPT(invalidate_interrupt,INVALIDATE_TLB_VECTOR)
> +BUILD_INTERRUPT(call_function_interrupt,CALL_FUNCTION_VECTOR)
> +#endif
> +
> +/*
> + * every pentium local APIC has two 'local interrupts', with a
> + * soft-definable vector attached to both interrupts, one of
> + * which is a timer interrupt, the other one is error counter
> + * overflow. Linux uses the local APIC timer interrupt to get
> + * a much simpler SMP time architecture:
> + */
> +#ifdef CONFIG_X86_LOCAL_APIC
> +BUILD_INTERRUPT(apic_timer_interrupt,LOCAL_TIMER_VECTOR)
> +BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR)
> +BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR)
> +
> +#ifdef CONFIG_X86_MCE_P4THERMAL
> +BUILD_INTERRUPT(thermal_interrupt,THERMAL_APIC_VECTOR)
> +#endif
> +
> +#endif

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel