Re: [Xen-devel] [PATCH] pvh: add a call to pit_init to init spinlock

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 10.03.14 at 23:09, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx> wrote:
> On Mon, 10 Mar 2014 09:10:23 +0000
> "Jan Beulich" <JBeulich@xxxxxxxx> wrote:
> 
>> >>> On 08.03.14 at 03:28, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>
>> >>> wrote:
>> > --- a/xen/arch/x86/hvm/hvm.c
>> > +++ b/xen/arch/x86/hvm/hvm.c
>> > @@ -1299,6 +1299,9 @@ int hvm_vcpu_initialise(struct vcpu *v)
>> >          v->arch.hvm_vcpu.hcall_64bit = 1;    /* PVH 32bitfixme. */
>> >          /* This is for hvm_long_mode_enabled(v). */
>> >          v->arch.hvm_vcpu.guest_efer = EFER_SCE | EFER_LMA |
>> > EFER_LME; +
>> > +        if ( v->vcpu_id == 0 )
>> > +            pit_init(v, cpu_khz);
>> 
>> I know I'm asking a lot, but did you consider using the opportunity
>> to do what the HVM call site says - move this to
>> hvm_domain_initialise()? Or, considering that PV wants this too,
>> even move it to arch_domain_create()?
> 
> I started that route.. but thought it might be too intrusive for 4.4.
> I had it moved to hvm_domain_initialise, but arch_domain_create is 
> even better. It's hard to anticipate what any maintainer would prefer, 
> the impression so far is to not touch anything un-needed or un-necessary... 

Properly done cleanup, even more so in a case like this where a
comment even says this should be done, is always welcome.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel