Re: [Qemu-devel] [Qemu-ppc] [PULL 41/44] spapr: change default interrupt mode to 'dual'

[David Gibson <david@gibson.dropbear.id.au>]

[-- Attachment #1: Type: text/plain, Size: 4061 bytes --]

On Mon, Jul 15, 2019 at 12:19:03PM +0200, Cédric Le Goater wrote:
> On 11/07/2019 03:26, David Gibson wrote:
> > On Wed, Jul 10, 2019 at 06:26:09PM +0200, Laurent Vivier wrote:
> >> On 29/05/2019 08:50, David Gibson wrote:
> >>> From: Cédric Le Goater <clg@kaod.org>
> >>>
> >>> Now that XIVE support is complete (QEMU emulated and KVM devices),
> >>> change the pseries machine to advertise both interrupt modes: XICS
> >>> (P7/P8) and XIVE (P9).
> >>>
> >>> The machine default interrupt modes depends on the version. Current
> >>> settings are:
> >>>
> >>>     pseries   default interrupt mode
> >>>
> >>>     4.1       dual
> >>>     4.0       xics
> >>>     3.1       xics
> >>>     3.0       legacy xics (different IRQ number space layout)
> >>>
> >>> Signed-off-by: Cédric Le Goater <clg@kaod.org>
> >>> Message-Id: <20190522074016.10521-3-clg@kaod.org>
> >>> Reviewed-by: Greg Kurz <groug@kaod.org>
> >>> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> >>> ---
> >>>  hw/ppc/spapr.c | 3 ++-
> >>>  1 file changed, 2 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> >>> index 39e698e9b0..4fd16b43f0 100644
> >>> --- a/hw/ppc/spapr.c
> >>> +++ b/hw/ppc/spapr.c
> >>> @@ -4352,7 +4352,7 @@ static void spapr_machine_class_init(ObjectClass *oc, void *data)
> >>>      smc->default_caps.caps[SPAPR_CAP_LARGE_DECREMENTER] = SPAPR_CAP_ON;
> >>>      smc->default_caps.caps[SPAPR_CAP_CCF_ASSIST] = SPAPR_CAP_OFF;
> >>>      spapr_caps_add_properties(smc, &error_abort);
> >>> -    smc->irq = &spapr_irq_xics;
> >>> +    smc->irq = &spapr_irq_dual;
> >>>      smc->dr_phb_enabled = true;
> >>>  }
> >>>  
> >>> @@ -4430,6 +4430,7 @@ static void spapr_machine_4_0_class_options(MachineClass *mc)
> >>>      spapr_machine_4_1_class_options(mc);
> >>>      compat_props_add(mc->compat_props, hw_compat_4_0, hw_compat_4_0_len);
> >>>      smc->phb_placement = phb_placement_4_0;
> >>> +    smc->irq = &spapr_irq_xics;
> >>>  }
> >>>  
> >>>  DEFINE_SPAPR_MACHINE(4_0, "4.0", false);
> >>>
> >>
> >> This patch breaks the '-no-reboot' parameter (I think the "dual" mode
> >> breaks the -no-reboot parameter)
> >>
> >> After grub loads the kernel and starts it, the kernel aborts:
> >>
> >> OF stdout device is: /vdevice/vty@71000000
> >> Preparing to boot Linux version 4.18.0-112.el8.ppc64le
> >> (mockbuild@ppc-061.build.eng.bos.redhat.com) (gcc version 8.3.1 20190507
> >> (Red Hat 8.3.1-4) (GCC)) #1 SMP Fri Jul 5 11:21:28 UTC 2019
> >> Detected machine type: 0000000000000101
> >> command line: BOOT_IMAGE=/vmlinuz-4.18.0-112.el8.ppc64le
> >> root=/dev/mapper/rhel_ibm--p8--kvm--03--guest--02-root ro
> >> crashkernel=auto rd.lvm.lv=rhel_ibm-p8-kvm-03-guest-02/root
> >> rd.lvm.lv=rhel_ibm-p8-kvm-03-guest-02/swap
> >> Max number of cores passed to firmware: 256 (NR_CPUS = 2048)
> >> Calling ibm,client-architecture-support...[lvivier@localhost ~]$
> >>
> >> I bisected to this patch, and then after I understood the problem is
> >> with the -no-reboot parameter as the machine is reset by the CAS
> >> negotiation... and the -no-reboot prevents this reset.
> >>
> >> I don't know if it's a real problem or not.
> > 
> > Ah, bother.  I didn't think of the interaction between the CAS reboot
> > and -no-reboot.  I guess that's more reason to work out a way to do
> > the xics/xive switch without a full reset.  People were already not
> > thrilled with the extra reboots here.
> 
> QEMU builds a device tree depending on the interrupt mode negotiated 
> at CAS time. Can we dynamically add/remove nodes ?

I believe we can make essentially arbitrary change to the device tree
as part of the CAS process without a reboot.

Re-doing this is kind of awkward with the current structure, but it's
probably what we're going to want.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]