[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-devel] [PATCH] Improved pending interrupt handling during VMXexits


  • To: "Leendert van Doorn" <leendert@xxxxxxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Ling, Xiaofeng" <xiaofeng.ling@xxxxxxxxx>
  • Date: Wed, 29 Jun 2005 09:07:28 +0800
  • Delivery-date: Wed, 29 Jun 2005 01:06:24 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcV7hHb5cQpXekLtTQKc0X5f3s3MjAAwZxjQ
  • Thread-topic: [Xen-devel] [PATCH] Improved pending interrupt handling during VMXexits

Does this patch fix the problem of double fault/triple fault?

Leendert van Doorn <> wrote:
> This patch improves VMX exit handling while software interrupts are
> pending: These should not stored as a VMX entry exception but should
> instead just be restarted. This was the cause of the VMRESUME
> failure.  
> 
>       Leendert
> 
> Signed-Off-By: Leendert van Doorn <leendert@xxxxxxxxxxxxxx>
> 
> 
> --- xeno-unstable.orig/xen/arch/x86/vmx.c     2005-06-23
> 09:58:11.000000000 -0400 +++
> xeno-unstable.mine/xen/arch/x86/vmx.c 2005-06-27 21:46:12.000000000
> -0400 @@ -1159,15 +1159,19 @@ 
> 
>      __vmread(IDT_VECTORING_INFO_FIELD, &idtv_info_field);
>      if (idtv_info_field & INTR_INFO_VALID_MASK) {
> -        __vmwrite(VM_ENTRY_INTR_INFO_FIELD, idtv_info_field);
> -        if ((idtv_info_field & 0xff) == 14) {
> -            unsigned long error_code;
> +     if ((idtv_info_field & 0x0700) != 0x400) { /* exclude soft ints */
> +            __vmwrite(VM_ENTRY_INTR_INFO_FIELD, idtv_info_field);
> 
> -            __vmread(VM_EXIT_INTR_ERROR_CODE, &error_code);
> -            printk("#PG error code: %lx\n", error_code);
> -        }
> -        VMX_DBG_LOG(DBG_LEVEL_1, "idtv_info_field=%x",
> -                idtv_info_field);
> +         if (idtv_info_field & 0x800) { /* valid error code */
> +             unsigned long error_code;
> +             printk("VMX exit %x: %x/%lx\n",
> +                     exit_reason, idtv_info_field, error_code);
> +             __vmread(VM_EXIT_INTR_ERROR_CODE, &error_code);
> +             __vmwrite(VM_ENTRY_EXCEPTION_ERROR_CODE, error_code);
> +         } else
> +             printk("VMX exit %x: %x\n", exit_reason, idtv_info_field);
> +     }
> +        VMX_DBG_LOG(DBG_LEVEL_1, "idtv_info_field=%x",
> +idtv_info_field);
>      }
> 
>      /* don't bother H/W interrutps */
> @@ -1399,7 +1403,6 @@
>  #else
>      asm volatile("movq %0,%%cr2": :"r" (d->arch.arch_vmx.cpu_cr2)); 
> #endif -
>  }
> 
>  #endif /* CONFIG_VMX */
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.