qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kurz <groug@kaod.org>
To: "Cédric Le Goater" <clg@kaod.org>
Cc: qemu-ppc@nongnu.org, qemu-devel@nongnu.org,
	David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [Qemu-devel] [Qemu-ppc] [PATCH v4 1/3] spapr: introduce a fixed IRQ number space
Date: Fri, 6 Jul 2018 16:09:26 +0200	[thread overview]
Message-ID: <20180706160926.471a1930@bahia.lan> (raw)
In-Reply-To: <20180706153624.77494d28@bahia.lan>

On Fri, 6 Jul 2018 15:36:24 +0200
Greg Kurz <groug@kaod.org> wrote:

[...]

> Hmm... if you put a VIO net and two VIO vty in the domain XML, libvirt
> will generate reg == 0x1000 for the VIO net and reg == 0x30001000 for the
> second VIO vty... this will necessarily collide, won't it ?
> 
> With a 256 VIO devices limit, libvirt can only add 255 devices since
> the nvram is created by QEMU by default (libvirt can only change its
> reg using -global).
> 
> As David mentioned in another mail:
> 
> 	VIO net devices start at reg=0x1000
> 	VIO scsi devices start at reg=0x2000
> 	VIO nvram devices start at reg=0x3000
> 	VIO vty devices start at reg=0x30000000
> 	    and increment by 0x1000 each type
> 
> 
> The values for net, scsi and nvram overlap... which makes me wonder why
> do we even care to have per-type base value !?! Anyway, I don't think
> it's important for what you're trying to achieve.
> 
> Basically libvirt can generate regs in two distinct ranges:
> 
> - one for scsi/net/nvram:
> 
> smallest possible reg: 0x1000
> largest possible reg: 0x2000 + 254 * 0x1000 = 0x100000
> 
> ie, 254 scsi devices starting at 0x2000 and 1 nvram
> 

Oops, it is 255 scsi devices plus 1 nvram so the largest
reg in this range is 0x101000

> - one for vty 
> 
> smallest possible reg: 0x30000000
> largest possible reg: 0x30000000 + 253 * 0x1000 = 0x300fd000
> 
> ie, 254 vty devices
> 

and 255 vty devices, ie, largest reg is 0x300fe000.

> Thinking about the bit shifting magic that is needed to convert
> reg into a usable index makes my brain hurt, but I'll happily
> review anything you propose :)
> 
> > +    }
> > +}
> > +
> >  static void spapr_vio_busdev_realize(DeviceState *qdev, Error **errp)
> >  {
> >      sPAPRMachineState *spapr = SPAPR_MACHINE(qdev_get_machine());
> > @@ -485,10 +512,18 @@ static void spapr_vio_busdev_realize(DeviceState *qdev, Error **errp)
> >      }
> >  
> >      if (!dev->irq) {
> > -        dev->irq = spapr_irq_findone(spapr, &local_err);
> > -        if (local_err) {
> > -            error_propagate(errp, local_err);
> > -            return;
> > +        if (SPAPR_MACHINE_GET_CLASS(spapr)->legacy_irq_allocation) {
> > +            dev->irq = spapr_irq_findone(spapr, &local_err);
> > +            if (local_err) {
> > +                error_propagate(errp, local_err);
> > +                return;
> > +            }
> > +        } else {
> > +            dev->irq = spapr_vio_reg_to_irq(dev->reg);
> > +            if (dev->irq == SPAPR_IRQ_PCI_LSI) {
> > +                error_setg(errp, "Too many VIO devices");
> > +                return;
> > +            }
> >          }
> >      }
> >  
> > @@ -557,7 +592,7 @@ VIOsPAPRBus *spapr_vio_bus_init(void)
> >      /* Create bus on bridge device */
> >      qbus = qbus_create(TYPE_SPAPR_VIO_BUS, dev, "spapr-vio");
> >      bus = SPAPR_VIO_BUS(qbus);
> > -    bus->next_reg = 0x71000000;
> > +    bus->next_reg = SPAPR_VIO_REG_BASE;
> >  
> >      /* hcall-vio */
> >      spapr_register_hypercall(H_VIO_SIGNAL, h_vio_signal);
> > diff --git a/hw/ppc/Makefile.objs b/hw/ppc/Makefile.objs
> > index bcab6323b7ed..4ab556467289 100644
> > --- a/hw/ppc/Makefile.objs
> > +++ b/hw/ppc/Makefile.objs
> > @@ -4,7 +4,7 @@ obj-y += ppc.o ppc_booke.o fdt.o
> >  obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o
> >  obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o
> >  obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o spapr_rng.o
> > -obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o
> > +obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o
> >  # IBM PowerNV
> >  obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.o
> >  ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)  
> 
> 

  reply	other threads:[~2018-07-06 14:09 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-06  9:07 [Qemu-devel] [PATCH v4 0/3] spapr: introduce a fixed IRQ number space and an IRQ controller backend Cédric Le Goater
2018-07-06  9:07 ` [Qemu-devel] [PATCH v4 1/3] spapr: introduce a fixed IRQ number space Cédric Le Goater
2018-07-06 13:27   ` Cédric Le Goater
2018-07-06 14:16     ` Greg Kurz
2018-07-06 16:43       ` Cédric Le Goater
2018-07-06 13:36   ` Greg Kurz
2018-07-06 14:09     ` Greg Kurz [this message]
2018-07-20  2:38     ` David Gibson
2018-07-23 15:20       ` Cédric Le Goater
2018-07-20  2:18   ` David Gibson
2018-07-23 16:04     ` Cédric Le Goater
2018-07-06  9:07 ` [Qemu-devel] [PATCH v4 2/3] spapr: introduce a IRQ controller backend to the machine Cédric Le Goater
2018-07-06  9:07 ` [Qemu-devel] [PATCH v4 3/3] spapr: increase the size of the IRQ number space Cédric Le Goater

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180706160926.471a1930@bahia.lan \
    --to=groug@kaod.org \
    --cc=clg@kaod.org \
    --cc=david@gibson.dropbear.id.au \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).