Re: [Xen-devel] [PATCH v5 08/13] x86/traps: factor out pv_trap_init

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Tue, Jun 27, 2017 at 12:26:52PM -0600, Jan Beulich wrote:
> >>> Wei Liu <wei.liu2@xxxxxxxxxx> 06/26/17 6:29 PM >>>
> >--- a/xen/arch/x86/traps.c
> >+++ b/xen/arch/x86/traps.c
> >@@ -1934,21 +1934,29 @@ void __init init_idt_traps(void)
>      >this_cpu(compat_gdt_table) = boot_cpu_compat_gdt_table;
>  >}
>  >
> >+void __init pv_trap_init(void)
> >+{
> >+    /* The 32-on-64 hypercall vector is only accessible from ring 1. */
> >+    _set_gate(idt_table + HYPERCALL_VECTOR,
> >+              SYS_DESC_trap_gate, 1, entry_int82);
> >+
> >+    /* Fast trap for int80 (faster than taking the #GP-fixup path). */
> >+    _set_gate(idt_table + LEGACY_SYSCALL_VECTOR, SYS_DESC_trap_gate, 3,
> >+              &int80_direct_trap);
> >+
> >+    open_softirq(NMI_MCE_SOFTIRQ, nmi_mce_softirq);
> >+}
> 
> Any reason you don't move this into pv/traps.c right away?

Because nmi_mce_softirq needs to be moved at the same time. That can't
be undone until I refactor the code further in later patch(es).

> 
> >void __init trap_init(void)
>  >{
>      >unsigned int vector;
>  >
> >+    pv_trap_init();
> 
> I think this would better be called after ...
> 
> >/* Replace early pagefault with real pagefault handler. */
>      >set_intr_gate(TRAP_page_fault, &page_fault);
>  
> ... this, i.e. in line with where most of the code was before.
> 

NP.

> Jan
> 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel