>>> On 15.08.18 at 13:39, <andrew.coop...@citrix.com> wrote:
> On 15/08/18 12:34, Jan Beulich wrote:
>> It took till the 4.5 backports of the L1TF prereqs that gcc 8.2 finally
>> noticed that the vPMU callers, not checking the function's return value,
>> may consume uninitialized data. Guard against this by storing zero on
>> the error path.
>>
>> Signed-off-by: Jan Beulich <jbeul...@suse.com>
>
> vPMU really is the gift which keeps on giving...
>
>>
>> --- a/xen/include/asm-x86/hvm/vmx/vmcs.h
>> +++ b/xen/include/asm-x86/hvm/vmx/vmcs.h
>> @@ -586,7 +586,10 @@ static inline int vmx_read_guest_msr(con
>> const struct vmx_msr_entry *ent = vmx_find_msr(v, msr, VMX_MSR_GUEST);
>>
>> if ( !ent )
>> + {
>> + *val = 0;
>> return -ESRCH;
>> + }
>>
>> *val = ent->data;
>>
>
> You want a similar adjustment in vmx_read_guest_loadonly_msr()
> immediately below, which I believe is only applicable to staging atm.
Oh, indeed - coming from 4.5 I didn't pay enough attention here.
> With that, Reviewed-by: Andrew Cooper <andrew.coop...@citrix.com>
Thanks, Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
