[Qemu-devel] [PATCH] SVM: acknowledge interrupt only after it is taken

[Qemu-devel] [PATCH] SVM: acknowledge interrupt only after it is taken

[Alexander Graf]

SVM specifies that the V_IRQ mask is only to be removed, if the
interrupt that is to be delivered actually is delivered.

As of the SVM rewrite, this mask is always unmasked when the main cpu
loop is processed, leaving a corner case where calling the interrupt
handler causes a #PF. In that case (booting Linux / starting gfxboot)
the current implementation tells the VMM the interrupt is taken, even
though it is not.

This patch modifies the VIRQ unmasking to occur after do_interrupt,
making gfxboot work again.
---
 cpu-exec.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/cpu-exec.c b/cpu-exec.c
index 9a35a59..981ba86 100644
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -401,11 +401,11 @@ int cpu_exec(CPUState *env1)
                             int intno;
                             /* FIXME: this should respect TPR */
                             svm_check_intercept(SVM_EXIT_VINTR);
-                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
                             intno = ldl_phys(env->vm_vmcb + offsetof(struct vmcb, control.int_vector));
                             if (loglevel & CPU_LOG_TB_IN_ASM)
                                 fprintf(logfile, "Servicing virtual hardware INT=0x%02x\n", intno);
                             do_interrupt(intno, 0, 0, 0, 1);
+                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
                             next_tb = 0;
 #endif
                         }
-- 
1.5.6

Re: [Qemu-devel] [PATCH] SVM: acknowledge interrupt only after it is taken

[Alexander Graf]

Alexander Graf wrote:
> SVM specifies that the V_IRQ mask is only to be removed, if the
> interrupt that is to be delivered actually is delivered.
>
> As of the SVM rewrite, this mask is always unmasked when the main cpu
> loop is processed, leaving a corner case where calling the interrupt
> handler causes a #PF. In that case (booting Linux / starting gfxboot)
> the current implementation tells the VMM the interrupt is taken, even
> though it is not.
>
> This patch modifies the VIRQ unmasking to occur after do_interrupt,
> making gfxboot work again.
>   

(wow, I actually forgot that line even though working with git - phew)

Signed-off-by: Alexander Graf <agraf@suse.de>

> ---
>  cpu-exec.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/cpu-exec.c b/cpu-exec.c
> index 9a35a59..981ba86 100644
> --- a/cpu-exec.c
> +++ b/cpu-exec.c
> @@ -401,11 +401,11 @@ int cpu_exec(CPUState *env1)
>                              int intno;
>                              /* FIXME: this should respect TPR */
>                              svm_check_intercept(SVM_EXIT_VINTR);
> -                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
>                              intno = ldl_phys(env->vm_vmcb + offsetof(struct vmcb, control.int_vector));
>                              if (loglevel & CPU_LOG_TB_IN_ASM)
>                                  fprintf(logfile, "Servicing virtual hardware INT=0x%02x\n", intno);
>                              do_interrupt(intno, 0, 0, 0, 1);
> +                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
>                              next_tb = 0;
>  #endif
>                          }
>   

Re: [Qemu-devel] [PATCH] SVM: acknowledge interrupt only after it is taken

[Aurelien Jarno]

On Fri, Dec 12, 2008 at 05:48:41PM +0100, Alexander Graf wrote:
> SVM specifies that the V_IRQ mask is only to be removed, if the
> interrupt that is to be delivered actually is delivered.
> 
> As of the SVM rewrite, this mask is always unmasked when the main cpu
> loop is processed, leaving a corner case where calling the interrupt
> handler causes a #PF. In that case (booting Linux / starting gfxboot)
> the current implementation tells the VMM the interrupt is taken, even
> though it is not.
> 
> This patch modifies the VIRQ unmasking to occur after do_interrupt,
> making gfxboot work again.

Thanks, applied.

> ---
>  cpu-exec.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/cpu-exec.c b/cpu-exec.c
> index 9a35a59..981ba86 100644
> --- a/cpu-exec.c
> +++ b/cpu-exec.c
> @@ -401,11 +401,11 @@ int cpu_exec(CPUState *env1)
>                              int intno;
>                              /* FIXME: this should respect TPR */
>                              svm_check_intercept(SVM_EXIT_VINTR);
> -                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
>                              intno = ldl_phys(env->vm_vmcb + offsetof(struct vmcb, control.int_vector));
>                              if (loglevel & CPU_LOG_TB_IN_ASM)
>                                  fprintf(logfile, "Servicing virtual hardware INT=0x%02x\n", intno);
>                              do_interrupt(intno, 0, 0, 0, 1);
> +                            env->interrupt_request &= ~CPU_INTERRUPT_VIRQ;
>                              next_tb = 0;
>  #endif
>                          }
> -- 
> 1.5.6
> 
> 
> 
> 

-- 
  .''`.  Aurelien Jarno	            | GPG: 1024D/F1BCDB73
 : :' :  Debian developer           | Electrical Engineer
 `. `'   aurel32@debian.org         | aurelien@aurel32.net
   `-    people.debian.org/~aurel32 | www.aurel32.net