Re: [PATCH] xen/x86: Identify which vcpu's CR4 is being badly modified

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 26/03/14 13:17, Jan Beulich wrote:
>>>> On 25.03.14 at 20:40, <andrew.cooper3@citrix.com> wrote:
>> When the toolstack is setting vcpu state on behalf of a migrating guest, the
>> domain/vcpu reference from gdprintk() identifies the toolstack, not the
>> affected domain.
>>
>> After this change, the error looks more like:
>>
>>   domain.c:632:d0v3 Attempt to change d1v0's CR4 flags 00002660 -> 01875000
> And is the leading d0v3 really useful here in any way? I'd really like to
> see unmotivated uses of gdprintk() replaced by printk(XENLOG_G_...).
>
> Jan

Personally, I find it useful to distinguish between "the toolstack tried
something" and "a guest tried something", although the file/line
reference is quite useless.

Perhaps printk(XENLOG_G_WARNING "%pv attempted to change %pv's CR4 flags
...", current, v ....) ?

~Andrew

>
>> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
>> CC: Keir Fraser <keir@xen.org>
>> CC: Jan Beulich <JBeulich@suse.com>
>> ---
>>  xen/arch/x86/domain.c |    4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/xen/arch/x86/domain.c b/xen/arch/x86/domain.c
>> index b48f2dc..59e2729 100644
>> --- a/xen/arch/x86/domain.c
>> +++ b/xen/arch/x86/domain.c
>> @@ -652,8 +652,8 @@ unsigned long pv_guest_cr4_fixup(const struct vcpu *v, 
>> unsigned long guest_cr4)
>>  
>>      if ( (guest_cr4 & hv_cr4_mask) != (hv_cr4 & hv_cr4_mask) )
>>          gdprintk(XENLOG_WARNING,
>> -                 "Attempt to change CR4 flags %08lx -> %08lx\n",
>> -                 hv_cr4, guest_cr4);
>> +                 "Attempt to change %pv's CR4 flags %08lx -> %08lx\n",
>> +                 v, hv_cr4, guest_cr4);
>>  
>>      return (hv_cr4 & hv_cr4_mask) | (guest_cr4 & ~hv_cr4_mask);
>>  }
>> -- 
>> 1.7.10.4
>
>