From: Paolo Bonzini <pbonzini@redhat.com>
To: Bandan Das <bsd@redhat.com>
Cc: kvm@vger.kernel.org, guangrong.xiao@linux.intel.com,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/5] mmu: don't set the present bit unconditionally
Date: Tue, 28 Jun 2016 22:21:56 +0200 [thread overview]
Message-ID: <e35deee6-146a-9525-2c91-cd0560db12b6@redhat.com> (raw)
In-Reply-To: <jpgh9cdgqzx.fsf@linux.bootlegged.copy>
On 28/06/2016 19:30, Bandan Das wrote:
> Paolo Bonzini <pbonzini@redhat.com> writes:
>> On 28/06/2016 06:32, Bandan Das wrote:
>>> + bool execonly = !(context->guest_rsvd_check.bad_mt_xwr &
>>> + (1ull << VMX_EPT_EXECUTABLE_MASK));
>>>
>>> if (set_mmio_spte(vcpu, sptep, gfn, pfn, pte_access))
>>> return 0;
>>>
>>> - spte = PT_PRESENT_MASK;
>>> + if (!execonly)
>>> + spte |= PT_PRESENT_MASK;
>>
>> This needs a comment:
>>
>> /*
>> * There are two cases in which execonly is false: 1) for
>> * non-EPT page tables, in which case we need to set the
>> * P bit; 2) for EPT page tables where an X-- page table
>> * entry is invalid, in which case we need to force the R
>> * bit of the page table entry to 1.
>> */
>
> I think this should be: 2) for EPT page tables where an X-- page
> table entry is invalid and a EPT misconfig is injected to the guest
> before we reach here.
Right, I messed this one up. shadow_user_mask and ACC_USER_MASK work
the same way on processors that do not support execonly.
Paolo
>> BUILD_BUG_ON(PT_PRESENT_MASK != VMX_EPT_READABLE_MASK);
>> if (!execonly)
>> spte |= PT_PRESENT_MASK;
>>
>>
>>> if (!speculative)
>>> spte |= shadow_accessed_mask;
>>>
>>> if (enable_ept) {
>>> - kvm_mmu_set_mask_ptes(0ull,
>>> + kvm_mmu_set_mask_ptes(PT_PRESENT_MASK,
>>
>> This should be VMX_EPT_READABLE_MASK.
>>
>>> @@ -2531,6 +2535,7 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
>>> else
>>> spte |= shadow_nx_mask;
>>>
>>> + /* In the EPT case, shadow_user_mask is PT_PRESENT_MASK */
>>
>> I don't think this comment is necessary, but it's better to add one in
>> FNAME(gpte_access).
>>
>> /*
>> * In the EPT case, a page table can be executable but not
>> * readable (on some processors). Therefore, set_spte does not
>> * automatically set bit 0 if execute-only is supported.
>> * Instead, since EPT page tables do not have a U bit, we
>> * repurpose ACC_USER_MASK to signify readability. Likewise,
>> * when EPT is in use shadow_user_mask is set to
>> * VMX_EPT_READABLE_MASK.
>> */
>>
>>
>> Thanks,
>>
>> Paolo
>>
>>> if (pte_access & ACC_USER_MASK)
>>> spte |= shadow_user_mask;
>>
>>
>> Paolo
>>
>>> (enable_ept_ad_bits) ? VMX_EPT_ACCESS_BIT : 0ull,
>>> (enable_ept_ad_bits) ? VMX_EPT_DIRTY_BIT : 0ull,
>>> 0ull, VMX_EPT_EXECUTABLE_MASK);
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe kvm" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
next prev parent reply other threads:[~2016-06-28 20:22 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-06-28 4:32 [PATCH 0/5] Add support for EPT execute only for nested hypervisors Bandan Das
2016-06-28 4:32 ` [PATCH 1/5] mmu: mark spte present if the x bit is set Bandan Das
2016-06-28 8:44 ` Paolo Bonzini
2016-06-28 17:33 ` Bandan Das
2016-06-28 20:17 ` Paolo Bonzini
2016-06-28 20:37 ` Bandan Das
2016-06-28 20:49 ` Paolo Bonzini
2016-06-28 21:04 ` Bandan Das
2016-06-29 3:01 ` Xiao Guangrong
2016-07-05 3:06 ` Wanpeng Li
2016-07-05 10:50 ` Paolo Bonzini
2016-07-05 11:29 ` Wanpeng Li
2016-06-28 4:32 ` [PATCH 2/5] mmu: pass execonly value when initializing rsvd bits Bandan Das
2016-06-29 3:07 ` Xiao Guangrong
2016-06-28 4:32 ` [PATCH 3/5] mmu: don't set the present bit unconditionally Bandan Das
2016-06-28 8:57 ` Paolo Bonzini
2016-06-28 17:30 ` Bandan Das
2016-06-28 20:21 ` Paolo Bonzini [this message]
2016-07-05 5:50 ` Wanpeng Li
2016-07-05 10:50 ` Paolo Bonzini
2016-06-29 3:17 ` Xiao Guangrong
2016-06-29 8:18 ` Paolo Bonzini
2016-06-30 7:18 ` Xiao Guangrong
2016-06-28 4:32 ` [PATCH 4/5] mmu: remove is_present_gpte() Bandan Das
2016-06-28 4:32 ` [PATCH 5/5] nvmx: advertise support for ept execute only Bandan Das
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=e35deee6-146a-9525-2c91-cd0560db12b6@redhat.com \
--to=pbonzini@redhat.com \
--cc=bsd@redhat.com \
--cc=guangrong.xiao@linux.intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
/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
all inboxes | Powered by JetHome®