linux-kernel - Re: [PATCH v2 07/12] KVM: SVM: Adding support for configuring x2APIC MSRs interception

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 18 Apr 2022 14:17:15 +0300
From:   Maxim Levitsky <mlevitsk@...hat.com>
To:     Suravee Suthikulpanit <suravee.suthikulpanit@....com>,
        linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     pbonzini@...hat.com, seanjc@...gle.com, joro@...tes.org,
        jon.grimm@....com, wei.huang2@....com, terry.bowman@....com
Subject: Re: [PATCH v2 07/12] KVM: SVM: Adding support for configuring
 x2APIC MSRs interception

On Tue, 2022-04-12 at 06:58 -0500, Suravee Suthikulpanit wrote:
> When enabling x2APIC virtualization (x2AVIC), the interception of
> x2APIC MSRs must be disabled to let the hardware virtualize guest
> MSR accesses.
> 
> Current implementation keeps track of list of MSR interception state
> in the svm_direct_access_msrs array. Therefore, extends the array to
> include x2APIC MSRs.
> 
> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@....com>
> ---
>  arch/x86/kvm/svm/svm.c | 29 ++++++++++++++++++++++++++++-
>  arch/x86/kvm/svm/svm.h |  5 +++--
>  2 files changed, 31 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> index 5ec770a1b4e8..c85663b62d4e 100644
> --- a/arch/x86/kvm/svm/svm.c
> +++ b/arch/x86/kvm/svm/svm.c
> @@ -76,7 +76,7 @@ static uint64_t osvw_len = 4, osvw_status;
>  
>  static DEFINE_PER_CPU(u64, current_tsc_ratio);
>  
> -static const struct svm_direct_access_msrs {
> +static struct svm_direct_access_msrs {
>  	u32 index;   /* Index of the MSR */
>  	bool always; /* True if intercept is initially cleared */
>  } direct_access_msrs[MAX_DIRECT_ACCESS_MSRS] = {
> @@ -774,6 +774,32 @@ static void add_msr_offset(u32 offset)
>  	BUG();
>  }
>  
> +static void init_direct_access_msrs(void)
> +{
> +	int i, j;
> +
> +	/* Find first MSR_INVALID */
> +	for (i = 0; i < MAX_DIRECT_ACCESS_MSRS; i++) {
> +		if (direct_access_msrs[i].index == MSR_INVALID)
> +			break;
> +	}
> +	BUG_ON(i >= MAX_DIRECT_ACCESS_MSRS);
> +
> +	/*
> +	 * Initialize direct_access_msrs entries to intercept X2APIC MSRs
> +	 * (range 0x800 to 0x8ff)
> +	 */
> +	for (j = 0; j < 0x100; j++) {
> +		direct_access_msrs[i + j].index = APIC_BASE_MSR + j;
> +		direct_access_msrs[i + j].always = false;
> +	}

That looks *much cleaner* code wise even though it is slower 
because now we have 256 more msrs in this list.

So the best of the two worlds I think would be to add only APIC msrs that AVIC
actually handles to the list.

SDM has a table of these registers in 

"15.29.3.1 Virtual APIC Register Accesses"

I would add the registers that are either read/write allowed 
or at least cause traps.

Besides this, the patch looks fine.

Best regards,
	Maxim Levitsky


> +	BUG_ON(i + j >= MAX_DIRECT_ACCESS_MSRS);
> +
> +	/* Initialize last entry */
> +	direct_access_msrs[i + j].index = MSR_INVALID;
> +	direct_access_msrs[i + j].always = true;
> +}
> +
>  static void init_msrpm_offsets(void)
>  {
>  	int i;
> @@ -4739,6 +4765,7 @@ static __init int svm_hardware_setup(void)
>  	memset(iopm_va, 0xff, PAGE_SIZE * (1 << order));
>  	iopm_base = page_to_pfn(iopm_pages) << PAGE_SHIFT;
>  
> +	init_direct_access_msrs();
>  	init_msrpm_offsets();
>  
>  	supported_xcr0 &= ~(XFEATURE_MASK_BNDREGS | XFEATURE_MASK_BNDCSR);
> diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h
> index c44326eeb3f2..e340c86941be 100644
> --- a/arch/x86/kvm/svm/svm.h
> +++ b/arch/x86/kvm/svm/svm.h
> @@ -29,8 +29,9 @@
>  #define	IOPM_SIZE PAGE_SIZE * 3
>  #define	MSRPM_SIZE PAGE_SIZE * 2
>  
> -#define MAX_DIRECT_ACCESS_MSRS	20
> -#define MSRPM_OFFSETS	16
> +#define MAX_DIRECT_ACCESS_MSRS	(20 + 0x100)
> +#define MSRPM_OFFSETS	30
> +
>  extern u32 msrpm_offsets[MSRPM_OFFSETS] __read_mostly;
>  extern bool npt_enabled;
>  extern bool intercept_smi;