Re: [PATCH v3 01/10] KVM: MMU: fix decoding cache type from MTRR

[Bandan Das]

Alex Williamson <alex.williamson@xxxxxxxxxx> writes:

> On Wed, 2015-05-13 at 14:42 +0800, Xiao Guangrong wrote:
>> There are some bugs in current get_mtrr_type();
>> 1: bit 1 of mtrr_state->enabled is corresponding bit 11 of
>>    IA32_MTRR_DEF_TYPE MSR which completely control MTRR's enablement
>>    that means other bits are ignored if it is cleared
>> 
>> 2: the fixed MTRR ranges are controlled by bit 0 of
>>    mtrr_state->enabled (bit 10 of IA32_MTRR_DEF_TYPE)
>> 
>> 3: if MTRR is disabled, UC is applied to all of physical memory rather
>>    than mtrr_state->def_type
>> 
>> Signed-off-by: Xiao Guangrong <guangrong.xiao@xxxxxxxxxxxxxxx>
>> ---
>>  arch/x86/kvm/mmu.c | 14 ++++++--------
>>  1 file changed, 6 insertions(+), 8 deletions(-)
>
>
> I'm seeing a significant regression in boot performance on Intel
> hardware with assigned devices that bisects back to this patch.  There's
> a long delay with Seabios between the version splash and execution of
> option ROMs, and a _very_ long delay with OVMF before the display is
> initialized.  The delay is long enough that users are reporting their
> previously working VM is hung with 100% CPU usage on v4.2-rc1.  Thanks,
>
> Alex
>
>> 
>> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
>> index b78e83f..d00cebd 100644
>> --- a/arch/x86/kvm/mmu.c
>> +++ b/arch/x86/kvm/mmu.c
>> @@ -2393,19 +2393,20 @@ EXPORT_SYMBOL_GPL(kvm_mmu_unprotect_page);
>>  static int get_mtrr_type(struct mtrr_state_type *mtrr_state,
>>  			 u64 start, u64 end)
>>  {
>> -	int i;
>>  	u64 base, mask;
>>  	u8 prev_match, curr_match;
>> -	int num_var_ranges = KVM_NR_VAR_MTRR;
>> +	int i, num_var_ranges = KVM_NR_VAR_MTRR;
>>  
>> -	if (!mtrr_state->enabled)
>> -		return 0xFF;
>> +	/* MTRR is completely disabled, use UC for all of physical memory. */
>> +	if (!(mtrr_state->enabled & 0x2))
>> +		return MTRR_TYPE_UNCACHABLE;

Setting this to UC if MTRR is disabled is too restrictive maybe..

Bandan

>>  	/* Make end inclusive end, instead of exclusive */
>>  	end--;
>>  
>>  	/* Look in fixed ranges. Just return the type as per start */
>> -	if (mtrr_state->have_fixed && (start < 0x100000)) {
>> +	if (mtrr_state->have_fixed && (mtrr_state->enabled & 0x1) &&
>> +	      (start < 0x100000)) {
>>  		int idx;
>>  
>>  		if (start < 0x80000) {
>> @@ -2428,9 +2429,6 @@ static int get_mtrr_type(struct mtrr_state_type *mtrr_state,
>>  	 * Look of multiple ranges matching this address and pick type
>>  	 * as per MTRR precedence
>>  	 */
>> -	if (!(mtrr_state->enabled & 2))
>> -		return mtrr_state->def_type;
>> -
>>  	prev_match = 0xFF;
>>  	for (i = 0; i < num_var_ranges; ++i) {
>>  		unsigned short start_state, end_state;
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/