From: Krish Sadhukhan <krish.sadhukhan@oracle.com>
To: Nadav Amit <nadav.amit@gmail.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org
Subject: Re: [kvm-unit-tests PATCH] x86: vmx: Mask undefined bits in exit qualifications
Date: Mon, 17 Jun 2019 15:22:58 -0700 [thread overview]
Message-ID: <720b1ba2-11aa-6baf-9f21-aa3e1e324afa@oracle.com> (raw)
In-Reply-To: <A9500030-816E-49F7-84C7-6176C722C2B0@gmail.com>
On 06/17/2019 12:52 PM, Nadav Amit wrote:
>> On May 3, 2019, at 10:49 AM, nadav.amit@gmail.com wrote:
>>
>> From: Nadav Amit <nadav.amit@gmail.com>
>>
>> On EPT violation, the exit qualifications may have some undefined bits.
>>
>> Bit 6 is undefined if "mode-based execute control" is 0.
>>
>> Bits 9-11 are undefined unless the processor supports advanced VM-exit
>> information for EPT violations.
>>
>> Right now on KVM these bits are always undefined inside the VM (i.e., in
>> an emulated VM-exit). Mask these bits to avoid potential false
>> indication of failures.
>>
>> Signed-off-by: Nadav Amit <nadav.amit@gmail.com>
>> ---
>> x86/vmx.h | 20 ++++++++++++--------
>> x86/vmx_tests.c | 4 ++++
>> 2 files changed, 16 insertions(+), 8 deletions(-)
>>
>> diff --git a/x86/vmx.h b/x86/vmx.h
>> index cc377ef..5053d6f 100644
>> --- a/x86/vmx.h
>> +++ b/x86/vmx.h
>> @@ -603,16 +603,20 @@ enum vm_instruction_error_number {
>> #define EPT_ADDR_MASK GENMASK_ULL(51, 12)
>> #define PAGE_MASK_2M (~(PAGE_SIZE_2M-1))
>>
>> -#define EPT_VLT_RD 1
>> -#define EPT_VLT_WR (1 << 1)
>> -#define EPT_VLT_FETCH (1 << 2)
>> -#define EPT_VLT_PERM_RD (1 << 3)
>> -#define EPT_VLT_PERM_WR (1 << 4)
>> -#define EPT_VLT_PERM_EX (1 << 5)
>> +#define EPT_VLT_RD (1ull << 0)
>> +#define EPT_VLT_WR (1ull << 1)
>> +#define EPT_VLT_FETCH (1ull << 2)
>> +#define EPT_VLT_PERM_RD (1ull << 3)
>> +#define EPT_VLT_PERM_WR (1ull << 4)
>> +#define EPT_VLT_PERM_EX (1ull << 5)
>> +#define EPT_VLT_PERM_USER_EX (1ull << 6)
>> #define EPT_VLT_PERMS (EPT_VLT_PERM_RD | EPT_VLT_PERM_WR | \
>> EPT_VLT_PERM_EX)
>> -#define EPT_VLT_LADDR_VLD (1 << 7)
>> -#define EPT_VLT_PADDR (1 << 8)
>> +#define EPT_VLT_LADDR_VLD (1ull << 7)
>> +#define EPT_VLT_PADDR (1ull << 8)
>> +#define EPT_VLT_GUEST_USER (1ull << 9)
>> +#define EPT_VLT_GUEST_WR (1ull << 10)
This one should be named EPT_VLT_GUEST_RW, assuming you are naming them
according to the 1-setting of the bits.
>> +#define EPT_VLT_GUEST_EX (1ull << 11)
>>
>> #define MAGIC_VAL_1 0x12345678ul
>> #define MAGIC_VAL_2 0x87654321ul
>> diff --git a/x86/vmx_tests.c b/x86/vmx_tests.c
>> index c52ebc6..b4129e1 100644
>> --- a/x86/vmx_tests.c
>> +++ b/x86/vmx_tests.c
>> @@ -2365,6 +2365,10 @@ static void do_ept_violation(bool leaf, enum ept_access_op op,
>>
>> qual = vmcs_read(EXI_QUALIFICATION);
>>
>> + /* Mask undefined bits (which may later be defined in certain cases). */
>> + qual &= ~(EPT_VLT_GUEST_USER | EPT_VLT_GUEST_WR | EPT_VLT_GUEST_EX |
>> + EPT_VLT_PERM_USER_EX);
>> +
The "DIAGNOSE" macro doesn't check any of these bits, so this masking
seems redundant.
Also, don't we need to check for the relevant conditions before masking
the bits ? For example, EPT_VLT_PERM_USER_EX is dependent on "mode-based
execute control" VM-execution control" and the other ones depend on bit
7 and 8 of the Exit Qualification field.
>> diagnose_ept_violation_qual(expected_qual, qual);
>> TEST_EXPECT_EQ(expected_qual, qual);
>>
>> --
>> 2.17.1
> Ping.
>
next prev parent reply other threads:[~2019-06-17 22:23 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-03 17:49 [kvm-unit-tests PATCH] x86: vmx: Mask undefined bits in exit qualifications nadav.amit
2019-06-17 19:52 ` Nadav Amit
2019-06-17 22:22 ` Krish Sadhukhan [this message]
2019-06-17 22:46 ` Nadav Amit
2019-06-17 23:52 ` Krish Sadhukhan
2019-06-18 8:50 ` Paolo Bonzini
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=720b1ba2-11aa-6baf-9f21-aa3e1e324afa@oracle.com \
--to=krish.sadhukhan@oracle.com \
--cc=kvm@vger.kernel.org \
--cc=nadav.amit@gmail.com \
--cc=pbonzini@redhat.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;
as well as URLs for NNTP newsgroup(s).