[Skiboot] [PATCH-RESEND] phb4: Reset FIR/NFIR registers before PHB4 probe

[Michael Neuling]

On Tue, 2018-04-03 at 06:21 +0530, Vaibhav Jain wrote:
> Thanks looking into this patch Mikey,
> 
> Michael Neuling <mikey at neuling.org> writes:
> > > >  hw/phb4.c | 9 +++++++++
> > > >  1 file changed, 9 insertions(+)
> > > > 
> > > > diff --git a/hw/phb4.c b/hw/phb4.c
> > > > index e45be01f..c5a33a0f 100644
> > > > --- a/hw/phb4.c
> > > > +++ b/hw/phb4.c
> > > > @@ -5315,6 +5315,15 @@ static void phb4_probe_stack(struct dt_nod
> > 
> > Can we put this in phb4_init_hw() instead or is that too late?
> 
> Its too late to put this in phb4_init_hw() as phb4_probe_stack() will do
> the first test mmio to read the PHB version register. This is done just
> after resetting the ETU. Without this change the PHB version was being
> reported as 0xffffffffffffffff.

OK, I think that's resonable. Thanks for fixing.

Mikey

> 
> > 
> > > > *stk_node, uint32_t pec_index,
> > > >  		return;
> > > >  	}
> > > >  
> > > > +	/* Clear errors in PFIR and NFIR */
> > > > +	xscom_read(gcid, pci_stack + XPEC_PCI_STK_PCI_FIR, &val);
> > > > +	prlog_once(PR_DEBUG, "PFIR: %llx\n", val);
> > 
> > Can we removed these reads and prints.  I'm not sure they add any value.
> 
> Agreed they dont add much value. Was trying to follow the same
> convention to dump ETU reset register before its being overwritten. Will
> send a v2 with these debug prints removed.
>