hw/display/Kconfig | 2 +- hw/hppa/Kconfig | 1 + hw/i386/Kconfig | 1 + hw/ppc/Kconfig | 1 + 4 files changed, 4 insertions(+), 1 deletion(-)
When virtio-vga was added, the intention was to only support it for
those machines where the firmware does not know about virtio-gpu,
and supported VGA legacy hardware before virtio-{gpu,vga} were
introduced.
The Kconfig switch however enabled virtio-vga for all machines with
a PCI bus, and libvirt then prefers it even on hardware where
virtio-gpu would be preferrable. At least for now, only enable
virtio-vga for PC, hppa and pSeries machines, as was the case
before Kconfig dependencies were introduced.
Reported-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/display/Kconfig | 2 +-
hw/hppa/Kconfig | 1 +
hw/i386/Kconfig | 1 +
hw/ppc/Kconfig | 1 +
4 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/hw/display/Kconfig b/hw/display/Kconfig
index 86c1d544c5..72be57a403 100644
--- a/hw/display/Kconfig
+++ b/hw/display/Kconfig
@@ -100,7 +100,7 @@ config VIRTIO_GPU
config VIRTIO_VGA
bool
- default y if PCI_DEVICES
+ # defaults to "N", enabled by specific boards
depends on VIRTIO_PCI
select VGA
diff --git a/hw/hppa/Kconfig b/hw/hppa/Kconfig
index 7334f57081..6e5d74a825 100644
--- a/hw/hppa/Kconfig
+++ b/hw/hppa/Kconfig
@@ -2,6 +2,7 @@ config DINO
bool
imply PCI_DEVICES
imply E1000_PCI
+ imply VIRTIO_VGA
select PCI
select SERIAL
select ISA_BUS
diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig
index 8e8444dbf9..a6aed7c131 100644
--- a/hw/i386/Kconfig
+++ b/hw/i386/Kconfig
@@ -19,6 +19,7 @@ config PC
imply TPM_CRB
imply TPM_TIS
imply VGA_PCI
+ imply VIRTIO_VGA
select FDC
select I8259
select I8254
diff --git a/hw/ppc/Kconfig b/hw/ppc/Kconfig
index ae07b4da63..a3465155f0 100644
--- a/hw/ppc/Kconfig
+++ b/hw/ppc/Kconfig
@@ -2,6 +2,7 @@ config PSERIES
bool
imply PCI_DEVICES
imply TEST_DEVICES
+ imply VIRTIO_VGA
select DIMM
select PCI
select SPAPR_VSCSI
--
2.20.1
On 21/03/2019 15.29, Paolo Bonzini wrote: > When virtio-vga was added, the intention was to only support it for > those machines where the firmware does not know about virtio-gpu, > and supported VGA legacy hardware before virtio-{gpu,vga} were > introduced. > > The Kconfig switch however enabled virtio-vga for all machines with > a PCI bus, and libvirt then prefers it even on hardware where > virtio-gpu would be preferrable. At least for now, only enable > virtio-vga for PC, hppa and pSeries machines, as was the case > before Kconfig dependencies were introduced. > > Reported-by: Laszlo Ersek <lersek@redhat.com> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > --- > hw/display/Kconfig | 2 +- > hw/hppa/Kconfig | 1 + > hw/i386/Kconfig | 1 + > hw/ppc/Kconfig | 1 + > 4 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/hw/display/Kconfig b/hw/display/Kconfig > index 86c1d544c5..72be57a403 100644 > --- a/hw/display/Kconfig > +++ b/hw/display/Kconfig > @@ -100,7 +100,7 @@ config VIRTIO_GPU > > config VIRTIO_VGA > bool > - default y if PCI_DEVICES > + # defaults to "N", enabled by specific boards > depends on VIRTIO_PCI > select VGA > > diff --git a/hw/hppa/Kconfig b/hw/hppa/Kconfig > index 7334f57081..6e5d74a825 100644 > --- a/hw/hppa/Kconfig > +++ b/hw/hppa/Kconfig > @@ -2,6 +2,7 @@ config DINO > bool > imply PCI_DEVICES > imply E1000_PCI > + imply VIRTIO_VGA > select PCI > select SERIAL > select ISA_BUS > diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig > index 8e8444dbf9..a6aed7c131 100644 > --- a/hw/i386/Kconfig > +++ b/hw/i386/Kconfig > @@ -19,6 +19,7 @@ config PC > imply TPM_CRB > imply TPM_TIS > imply VGA_PCI > + imply VIRTIO_VGA > select FDC > select I8259 > select I8254 > diff --git a/hw/ppc/Kconfig b/hw/ppc/Kconfig > index ae07b4da63..a3465155f0 100644 > --- a/hw/ppc/Kconfig > +++ b/hw/ppc/Kconfig > @@ -2,6 +2,7 @@ config PSERIES > bool > imply PCI_DEVICES > imply TEST_DEVICES > + imply VIRTIO_VGA > select DIMM > select PCI > select SPAPR_VSCSI Reviewed-by: Thomas Huth <thuth@redhat.com>
On Thu, 21 Mar 2019 at 14:41, Paolo Bonzini <pbonzini@redhat.com> wrote: > > When virtio-vga was added, the intention was to only support it for > those machines where the firmware does not know about virtio-gpu, > and supported VGA legacy hardware before virtio-{gpu,vga} were > introduced. > > The Kconfig switch however enabled virtio-vga for all machines with > a PCI bus, and libvirt then prefers it even on hardware where > virtio-gpu would be preferrable. At least for now, only enable > virtio-vga for PC, hppa and pSeries machines, as was the case > before Kconfig dependencies were introduced. Do we have a reference to a bug report against libvirt for this? > Reported-by: Laszlo Ersek <lersek@redhat.com> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > --- > hw/display/Kconfig | 2 +- > hw/hppa/Kconfig | 1 + > hw/i386/Kconfig | 1 + > hw/ppc/Kconfig | 1 + > 4 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/hw/display/Kconfig b/hw/display/Kconfig > index 86c1d544c5..72be57a403 100644 > --- a/hw/display/Kconfig > +++ b/hw/display/Kconfig > @@ -100,7 +100,7 @@ config VIRTIO_GPU > > config VIRTIO_VGA > bool > - default y if PCI_DEVICES > + # defaults to "N", enabled by specific boards > depends on VIRTIO_PCI > select VGA Could we have a comment that says what the criterion for a board needing to enable it is, so that if we add new boards in future we know whether they need to enable it or not? (If the answer is "never for any new board", that's useful to say.) thanks -- PMM
On 03/21/19 15:29, Paolo Bonzini wrote: > When virtio-vga was added, the intention was to only support it for > those machines where the firmware does not know about virtio-gpu, > and supported VGA legacy hardware before virtio-{gpu,vga} were > introduced. > > The Kconfig switch however enabled virtio-vga for all machines with > a PCI bus, and libvirt then prefers it even on hardware where > virtio-gpu would be preferrable. At least for now, only enable > virtio-vga for PC, hppa and pSeries machines, as was the case > before Kconfig dependencies were introduced. > > Reported-by: Laszlo Ersek <lersek@redhat.com> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > --- > hw/display/Kconfig | 2 +- > hw/hppa/Kconfig | 1 + > hw/i386/Kconfig | 1 + > hw/ppc/Kconfig | 1 + > 4 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/hw/display/Kconfig b/hw/display/Kconfig > index 86c1d544c5..72be57a403 100644 > --- a/hw/display/Kconfig > +++ b/hw/display/Kconfig > @@ -100,7 +100,7 @@ config VIRTIO_GPU > > config VIRTIO_VGA > bool > - default y if PCI_DEVICES > + # defaults to "N", enabled by specific boards > depends on VIRTIO_PCI > select VGA > > diff --git a/hw/hppa/Kconfig b/hw/hppa/Kconfig > index 7334f57081..6e5d74a825 100644 > --- a/hw/hppa/Kconfig > +++ b/hw/hppa/Kconfig > @@ -2,6 +2,7 @@ config DINO > bool > imply PCI_DEVICES > imply E1000_PCI > + imply VIRTIO_VGA > select PCI > select SERIAL > select ISA_BUS > diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig > index 8e8444dbf9..a6aed7c131 100644 > --- a/hw/i386/Kconfig > +++ b/hw/i386/Kconfig > @@ -19,6 +19,7 @@ config PC > imply TPM_CRB > imply TPM_TIS > imply VGA_PCI > + imply VIRTIO_VGA > select FDC > select I8259 > select I8254 > diff --git a/hw/ppc/Kconfig b/hw/ppc/Kconfig > index ae07b4da63..a3465155f0 100644 > --- a/hw/ppc/Kconfig > +++ b/hw/ppc/Kconfig > @@ -2,6 +2,7 @@ config PSERIES > bool > imply PCI_DEVICES > imply TEST_DEVICES > + imply VIRTIO_VGA > select DIMM > select PCI > select SPAPR_VSCSI > I tried to apply this for testing, on top of c692931cda9dc6cbc16b89d5094a725a10dbb641, but git-am failed. The conflicts are in "hw/hppa/Kconfig" and "hw/i386/Kconfig". In the former, the conflict is due to E1000_PCI, in the latter, the conflict occurs due to VGA_PCI -- upstream doesn't seem to have either of those. Thanks Laszlo
----- Original Message ----- > From: "Laszlo Ersek" <lersek@redhat.com> > To: "Paolo Bonzini" <pbonzini@redhat.com>, qemu-devel@nongnu.org > Cc: thuth@redhat.com > Sent: Thursday, March 21, 2019 8:34:46 PM > Subject: Re: [PATCH] virtio-vga: only enable for specific boards > > On 03/21/19 15:29, Paolo Bonzini wrote: > > When virtio-vga was added, the intention was to only support it for > > those machines where the firmware does not know about virtio-gpu, > > and supported VGA legacy hardware before virtio-{gpu,vga} were > > introduced. > > > > The Kconfig switch however enabled virtio-vga for all machines with > > a PCI bus, and libvirt then prefers it even on hardware where > > virtio-gpu would be preferrable. At least for now, only enable > > virtio-vga for PC, hppa and pSeries machines, as was the case > > before Kconfig dependencies were introduced. > > > > Reported-by: Laszlo Ersek <lersek@redhat.com> > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > > --- > > hw/display/Kconfig | 2 +- > > hw/hppa/Kconfig | 1 + > > hw/i386/Kconfig | 1 + > > hw/ppc/Kconfig | 1 + > > 4 files changed, 4 insertions(+), 1 deletion(-) > > > > diff --git a/hw/display/Kconfig b/hw/display/Kconfig > > index 86c1d544c5..72be57a403 100644 > > --- a/hw/display/Kconfig > > +++ b/hw/display/Kconfig > > @@ -100,7 +100,7 @@ config VIRTIO_GPU > > > > config VIRTIO_VGA > > bool > > - default y if PCI_DEVICES > > + # defaults to "N", enabled by specific boards > > depends on VIRTIO_PCI > > select VGA > > > > diff --git a/hw/hppa/Kconfig b/hw/hppa/Kconfig > > index 7334f57081..6e5d74a825 100644 > > --- a/hw/hppa/Kconfig > > +++ b/hw/hppa/Kconfig > > @@ -2,6 +2,7 @@ config DINO > > bool > > imply PCI_DEVICES > > imply E1000_PCI > > + imply VIRTIO_VGA > > select PCI > > select SERIAL > > select ISA_BUS > > diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig > > index 8e8444dbf9..a6aed7c131 100644 > > --- a/hw/i386/Kconfig > > +++ b/hw/i386/Kconfig > > @@ -19,6 +19,7 @@ config PC > > imply TPM_CRB > > imply TPM_TIS > > imply VGA_PCI > > + imply VIRTIO_VGA > > select FDC > > select I8259 > > select I8254 > > diff --git a/hw/ppc/Kconfig b/hw/ppc/Kconfig > > index ae07b4da63..a3465155f0 100644 > > --- a/hw/ppc/Kconfig > > +++ b/hw/ppc/Kconfig > > @@ -2,6 +2,7 @@ config PSERIES > > bool > > imply PCI_DEVICES > > imply TEST_DEVICES > > + imply VIRTIO_VGA > > select DIMM > > select PCI > > select SPAPR_VSCSI > > > > I tried to apply this for testing, on top of > c692931cda9dc6cbc16b89d5094a725a10dbb641, but git-am failed. The > conflicts are in "hw/hppa/Kconfig" and "hw/i386/Kconfig". In the former, > the conflict is due to E1000_PCI, in the latter, the conflict occurs due > to VGA_PCI -- upstream doesn't seem to have either of those. Yeah, it is on top of my tree. You can find it as tag for-upstream at https://github.com/bonzini/qemu, but it is not tested yet. Paolo
On 03/22/19 23:00, Paolo Bonzini wrote: > > > ----- Original Message ----- >> From: "Laszlo Ersek" <lersek@redhat.com> >> To: "Paolo Bonzini" <pbonzini@redhat.com>, qemu-devel@nongnu.org >> Cc: thuth@redhat.com >> Sent: Thursday, March 21, 2019 8:34:46 PM >> Subject: Re: [PATCH] virtio-vga: only enable for specific boards >> >> On 03/21/19 15:29, Paolo Bonzini wrote: >>> When virtio-vga was added, the intention was to only support it for >>> those machines where the firmware does not know about virtio-gpu, >>> and supported VGA legacy hardware before virtio-{gpu,vga} were >>> introduced. >>> >>> The Kconfig switch however enabled virtio-vga for all machines with >>> a PCI bus, and libvirt then prefers it even on hardware where >>> virtio-gpu would be preferrable. At least for now, only enable >>> virtio-vga for PC, hppa and pSeries machines, as was the case >>> before Kconfig dependencies were introduced. >>> >>> Reported-by: Laszlo Ersek <lersek@redhat.com> >>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> >>> --- >>> hw/display/Kconfig | 2 +- >>> hw/hppa/Kconfig | 1 + >>> hw/i386/Kconfig | 1 + >>> hw/ppc/Kconfig | 1 + >>> 4 files changed, 4 insertions(+), 1 deletion(-) >>> >>> diff --git a/hw/display/Kconfig b/hw/display/Kconfig >>> index 86c1d544c5..72be57a403 100644 >>> --- a/hw/display/Kconfig >>> +++ b/hw/display/Kconfig >>> @@ -100,7 +100,7 @@ config VIRTIO_GPU >>> >>> config VIRTIO_VGA >>> bool >>> - default y if PCI_DEVICES >>> + # defaults to "N", enabled by specific boards >>> depends on VIRTIO_PCI >>> select VGA >>> >>> diff --git a/hw/hppa/Kconfig b/hw/hppa/Kconfig >>> index 7334f57081..6e5d74a825 100644 >>> --- a/hw/hppa/Kconfig >>> +++ b/hw/hppa/Kconfig >>> @@ -2,6 +2,7 @@ config DINO >>> bool >>> imply PCI_DEVICES >>> imply E1000_PCI >>> + imply VIRTIO_VGA >>> select PCI >>> select SERIAL >>> select ISA_BUS >>> diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig >>> index 8e8444dbf9..a6aed7c131 100644 >>> --- a/hw/i386/Kconfig >>> +++ b/hw/i386/Kconfig >>> @@ -19,6 +19,7 @@ config PC >>> imply TPM_CRB >>> imply TPM_TIS >>> imply VGA_PCI >>> + imply VIRTIO_VGA >>> select FDC >>> select I8259 >>> select I8254 >>> diff --git a/hw/ppc/Kconfig b/hw/ppc/Kconfig >>> index ae07b4da63..a3465155f0 100644 >>> --- a/hw/ppc/Kconfig >>> +++ b/hw/ppc/Kconfig >>> @@ -2,6 +2,7 @@ config PSERIES >>> bool >>> imply PCI_DEVICES >>> imply TEST_DEVICES >>> + imply VIRTIO_VGA >>> select DIMM >>> select PCI >>> select SPAPR_VSCSI >>> >> >> I tried to apply this for testing, on top of >> c692931cda9dc6cbc16b89d5094a725a10dbb641, but git-am failed. The >> conflicts are in "hw/hppa/Kconfig" and "hw/i386/Kconfig". In the former, >> the conflict is due to E1000_PCI, in the latter, the conflict occurs due >> to VGA_PCI -- upstream doesn't seem to have either of those. > > Yeah, it is on top of my tree. You can find it as tag for-upstream at > https://github.com/bonzini/qemu, but it is not tested yet. I've now tested this at your commit 938912a86611 ("virtio-vga: only enable for specific boards", 2019-03-21). Tested-by: Laszlo Ersek <lersek@redhat.com> Reviewed-by: Laszlo Ersek <lersek@redhat.com> Thank you! Laszlo
© 2016 - 2024 Red Hat, Inc.