Re: [PATCH] kvm: remove invalid check for debugfs_create_dir()

[Sean Christopherson <sean.j.christopherson@intel.com>]

On Wed, Jun 12, 2019 at 05:46:22PM +0200, Greg Kroah-Hartman wrote:
> On Wed, Jun 12, 2019 at 08:40:21AM -0700, Sean Christopherson wrote:
> > On Wed, Jun 12, 2019 at 04:50:33PM +0200, Greg Kroah-Hartman wrote:
> > > debugfs_create_dir() can never return NULL, so no need to check for an
> > > impossible thing.
> > > 
> > > It's also not needed to ever check the return value of this function, so
> > > just remove the check entirely, and indent the previous line to a sane
> > > formatting :)
> > > 
> > > Cc: Paolo Bonzini <pbonzini@redhat.com>
> > > Cc: "Radim Krčmář" <rkrcmar@redhat.com>
> > > Cc: kvm@vger.kernel.org
> > > Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> > > ---
> > >  virt/kvm/kvm_main.c | 4 +---
> > >  1 file changed, 1 insertion(+), 3 deletions(-)
> > > 
> > > diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> > > index ca54b09adf5b..4b4ef642d8fa 100644
> > > --- a/virt/kvm/kvm_main.c
> > > +++ b/virt/kvm/kvm_main.c
> > > @@ -2605,9 +2605,7 @@ static int kvm_create_vcpu_debugfs(struct kvm_vcpu *vcpu)
> > >  
> > >  	snprintf(dir_name, sizeof(dir_name), "vcpu%d", vcpu->vcpu_id);
> > >  	vcpu->debugfs_dentry = debugfs_create_dir(dir_name,
> > > -								vcpu->kvm->debugfs_dentry);
> > > -	if (!vcpu->debugfs_dentry)
> > > -		return -ENOMEM;
> > > +						  vcpu->kvm->debugfs_dentry);
> > >  
> > >  	ret = kvm_arch_create_vcpu_debugfs(vcpu);
> > >  	if (ret < 0) {
> > > -- 
> > > 2.22.0
> > 
> > Any objection to me pulling this into a series to clean up similar issues
> > in arch/x86/kvm/debugfs.c -> kvm_arch_create_vcpu_debugfs(), and to
> > change kvm_create_vcpu_debugfs() to not return success/failure?  It'd be
> > nice to fix everything in a single shot.
> 
> It was on my todo list to do the cleanup of the x86 kvm stuff.  I
> figured this arch-neutral fix was independent, but if you want me to do
> it all at once, I'll be glad to do so later this week.

My preference would be to get it done all at once, having a discrepancy
between kvm_main and x86 leads to a bit of head scratching.

The x86 change would be independent, the piece that's not is converting
kvm_create_vcpu_debugfs() to handle failure internally and not return
errors, which would depend on this patch's removal of "return -ENOMEM".
At a minimum I think that change should be paired with this fix.