[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2 4/7] x86/vmx: add do_vmtrace_op



----- 19 cze 2020 o 17:50, Jan Beulich jbeulich@xxxxxxxx napisał(a):

> On 19.06.2020 17:30, Roger Pau Monné wrote:
>> On Fri, Jun 19, 2020 at 01:41:03AM +0200, Michał Leszczyński wrote:
>>> --- a/xen/arch/x86/hvm/hvm.c
>>> +++ b/xen/arch/x86/hvm/hvm.c
>>> @@ -1612,6 +1612,24 @@ int hvm_vcpu_initialise(struct vcpu *v)
>>>      return rc;
>>>  }
>>>  
>>> +void hvm_vmtrace_destroy(struct vcpu *v)
>>> +{
>>> +    unsigned int i;
>>> +    struct page_info *pg;
>>> +    struct ipt_state *ipt = v->arch.hvm.vmx.ipt_state;
>>> +    mfn_t buf_mfn = ipt->output_base >> PAGE_SHIFT;
>> 
>> Does this build? I think you are missing a _mfn(...) here?
> 
> This as well as ...
> 
>>> +    size_t buf_size = ipt->output_mask.size + 1;
>>> +
>>> +    xfree(ipt);
>>> +    v->arch.hvm.vmx.ipt_state = NULL;
>>> +
>>> +    for ( i = 0; i < (buf_size >> PAGE_SHIFT); i++ )
>>> +    {
>>> +        pg = mfn_to_page(_mfn(mfn_add(buf_mfn, i)));
> 
> ... the extra _mfn() here suggest the code was only ever built in
> release mode so far.
> 
> Jan


Ah, I forgot to enable developer checks. This will be corrected in v3.

ml



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.