From: Jan Kiszka <jan.kiszka@web.de>
To: "Serge E. Hallyn" <serge@hallyn.com>
Cc: Avi Kivity <avi@redhat.com>, KVM mailing list <kvm@vger.kernel.org>
Subject: Re: buggy emulate_int_real
Date: Wed, 13 Apr 2011 01:31:42 +0200 [thread overview]
Message-ID: <4DA4E0DE.7070903@web.de> (raw)
In-Reply-To: <20110412223940.GA3485@hallyn.com>
[-- Attachment #1: Type: text/plain, Size: 4312 bytes --]
On 2011-04-13 00:39, Serge E. Hallyn wrote:
> Quoting Serge E. Hallyn (serge@hallyn.com):
>> Quoting Jan Kiszka (jan.kiszka@web.de):
>>> From a brief refresh-reading, I would say
>>>
>>> if (interrupt.soft || kvm_exception_is_soft(nr))
>>> increment_eip_by_inst_len
>>>
>>> but only for those interrupts/exceptions which were raised by the
>>> triggering instructions, _not_ for exceptions raise while processing
>>> them (e.g. a page fault while accessing the IDT).
>>>
>>> Jan
>>
>> Ok, thanks guys. I'm currently trying a compile of the following.
>
> With the obvious fixes, that turns into the below, which seems to be
> working for me:
>
> From 555696f355ec2db9f4919f5e363355a2671b15a5 Mon Sep 17 00:00:00 2001
> From: Serge E. Hallyn <serge.hallyn@canonical.com>
> Date: Tue, 12 Apr 2011 22:18:40 +0100
> Subject: [PATCH 1/1] kvm: fix push of wrong eip when doing softint
>
> When doing a soft int, we need to bump eip before pushing it to
> the stack. Otherwise we'll do the int a second time.
>
> Signed-off-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
> ---
> arch/x86/kvm/vmx.c | 12 +++++++++---
> arch/x86/kvm/x86.c | 3 ++-
> arch/x86/kvm/x86.h | 2 +-
> 3 files changed, 12 insertions(+), 5 deletions(-)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index bf89ec2..3aad96c 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -1053,7 +1053,10 @@ static void vmx_queue_exception(struct kvm_vcpu *vcpu, unsigned nr,
> }
>
> if (vmx->rmode.vm86_active) {
> - if (kvm_inject_realmode_interrupt(vcpu, nr) != EMULATE_DONE)
> + int inc_eip = 0;
> + if (kvm_exception_is_soft(nr))
> + inc_eip = vcpu->arch.event_exit_inst_len;
> + if (kvm_inject_realmode_interrupt(vcpu, nr, inc_eip) != EMULATE_DONE)
> kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
> return;
> }
> @@ -2871,7 +2874,10 @@ static void vmx_inject_irq(struct kvm_vcpu *vcpu)
>
> ++vcpu->stat.irq_injections;
> if (vmx->rmode.vm86_active) {
> - if (kvm_inject_realmode_interrupt(vcpu, irq) != EMULATE_DONE)
> + int inc_eip = 0;
> + if (vcpu->arch.interrupt.soft)
> + inc_eip = vcpu->arch.event_exit_inst_len;
> + if (kvm_inject_realmode_interrupt(vcpu, irq, inc_eip) != EMULATE_DONE)
> kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
> return;
> }
> @@ -2905,7 +2911,7 @@ static void vmx_inject_nmi(struct kvm_vcpu *vcpu)
>
> ++vcpu->stat.nmi_injections;
> if (vmx->rmode.vm86_active) {
> - if (kvm_inject_realmode_interrupt(vcpu, NMI_VECTOR) != EMULATE_DONE)
> + if (kvm_inject_realmode_interrupt(vcpu, NMI_VECTOR, 0) != EMULATE_DONE)
> kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
> return;
> }
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index bcc0efc..8911622 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -4293,7 +4293,7 @@ static void init_emulate_ctxt(struct kvm_vcpu *vcpu)
> memcpy(c->regs, vcpu->arch.regs, sizeof c->regs);
> }
>
> -int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq)
> +int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq, int inc_eip)
> {
> struct decode_cache *c = &vcpu->arch.emulate_ctxt.decode;
> int ret;
> @@ -4303,6 +4303,7 @@ int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq)
> vcpu->arch.emulate_ctxt.decode.op_bytes = 2;
> vcpu->arch.emulate_ctxt.decode.ad_bytes = 2;
> vcpu->arch.emulate_ctxt.decode.eip = vcpu->arch.emulate_ctxt.eip;
> + vcpu->arch.emulate_ctxt.decode.eip += inc_eip;
I would fold this into the previous line.
> ret = emulate_int_real(&vcpu->arch.emulate_ctxt, &emulate_ops, irq);
>
> if (ret != X86EMUL_CONTINUE)
> diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
> index c600da8..e407ed3 100644
> --- a/arch/x86/kvm/x86.h
> +++ b/arch/x86/kvm/x86.h
> @@ -77,7 +77,7 @@ static inline u32 bit(int bitno)
>
> void kvm_before_handle_nmi(struct kvm_vcpu *vcpu);
> void kvm_after_handle_nmi(struct kvm_vcpu *vcpu);
> -int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq);
> +int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq, int inc_eip);
>
> void kvm_write_tsc(struct kvm_vcpu *vcpu, u64 data);
>
Looks consistent to me.
Jan
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 259 bytes --]
next prev parent reply other threads:[~2011-04-12 23:31 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-08 21:09 buggy emulate_int_real Serge E. Hallyn
2011-04-10 8:30 ` Avi Kivity
2011-04-12 7:53 ` Serge E. Hallyn
2011-04-12 8:02 ` Avi Kivity
2011-04-12 13:57 ` Serge E. Hallyn
2011-04-12 14:14 ` Avi Kivity
2011-04-12 14:12 ` Serge E. Hallyn
2011-04-12 14:16 ` Avi Kivity
2011-04-12 15:42 ` Jan Kiszka
2011-04-12 18:31 ` Serge E. Hallyn
2011-04-12 20:51 ` Jan Kiszka
2011-04-12 21:25 ` Serge E. Hallyn
2011-04-12 22:39 ` Serge E. Hallyn
2011-04-12 23:31 ` Jan Kiszka [this message]
2011-04-13 13:24 ` Serge E. Hallyn
2011-04-13 13:29 ` Avi Kivity
2011-04-13 13:52 ` Serge E. Hallyn
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4DA4E0DE.7070903@web.de \
--to=jan.kiszka@web.de \
--cc=avi@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=serge@hallyn.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox