[Qemu-devel] [PATCH] floppy: fix I/O ranges in both portio and acpi dsdt

[Qemu-devel] [PATCH] floppy: fix I/O ranges in both portio and acpi dsdt

[Gerd Hoffmann]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 hw/block/fdc.c            | 2 +-
 hw/i386/acpi-dsdt-isa.dsl | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/hw/block/fdc.c b/hw/block/fdc.c
index d8a8edd..c761291 100644
--- a/hw/block/fdc.c
+++ b/hw/block/fdc.c
@@ -2186,7 +2186,7 @@ static void fdctrl_realize_common(FDCtrl *fdctrl, Error **errp)
 }
 
 static const MemoryRegionPortio fdc_portio_list[] = {
-    { 1, 5, 1, .read = fdctrl_read, .write = fdctrl_write },
+    { 0, 6, 1, .read = fdctrl_read, .write = fdctrl_write },
     { 7, 1, 1, .read = fdctrl_read, .write = fdctrl_write },
     PORTIO_END_OF_LIST(),
 };
diff --git a/hw/i386/acpi-dsdt-isa.dsl b/hw/i386/acpi-dsdt-isa.dsl
index 89caa16..f71d7df 100644
--- a/hw/i386/acpi-dsdt-isa.dsl
+++ b/hw/i386/acpi-dsdt-isa.dsl
@@ -58,7 +58,7 @@ Scope(\_SB.PCI0.ISA) {
             }
         }
         Name(_CRS, ResourceTemplate() {
-            IO(Decode16, 0x03F2, 0x03F2, 0x00, 0x04)
+            IO(Decode16, 0x03F0, 0x03F0, 0x00, 0x06)
             IO(Decode16, 0x03F7, 0x03F7, 0x00, 0x01)
             IRQNoFlags() { 6 }
             DMA(Compatibility, NotBusMaster, Transfer8) { 2 }
-- 
1.8.3.1

Re: [Qemu-devel] [PATCH] floppy: fix I/O ranges in both portio and acpi dsdt

[Kevin Wolf]

Am 28.05.2015 um 17:37 hat Gerd Hoffmann geschrieben:
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---
>  hw/block/fdc.c            | 2 +-
>  hw/i386/acpi-dsdt-isa.dsl | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)

The commit message could be a bit longer, I don't think this change is
obvious.

My FDC spec says that this version of the controller doesn't have the
SRA register any more, and the emulation was added in the context of a
Sun4m patch (commit 8c6a4d774). Therefore my default assumption would be
that it's intentionally not accessible and not needed on isa-fdc.

I'm happy to get that assumption corrected, but it needs a non-empty
commit message for that.

Kevin


> diff --git a/hw/block/fdc.c b/hw/block/fdc.c
> index d8a8edd..c761291 100644
> --- a/hw/block/fdc.c
> +++ b/hw/block/fdc.c
> @@ -2186,7 +2186,7 @@ static void fdctrl_realize_common(FDCtrl *fdctrl, Error **errp)
>  }
>  
>  static const MemoryRegionPortio fdc_portio_list[] = {
> -    { 1, 5, 1, .read = fdctrl_read, .write = fdctrl_write },
> +    { 0, 6, 1, .read = fdctrl_read, .write = fdctrl_write },
>      { 7, 1, 1, .read = fdctrl_read, .write = fdctrl_write },
>      PORTIO_END_OF_LIST(),
>  };
> diff --git a/hw/i386/acpi-dsdt-isa.dsl b/hw/i386/acpi-dsdt-isa.dsl
> index 89caa16..f71d7df 100644
> --- a/hw/i386/acpi-dsdt-isa.dsl
> +++ b/hw/i386/acpi-dsdt-isa.dsl
> @@ -58,7 +58,7 @@ Scope(\_SB.PCI0.ISA) {
>              }
>          }
>          Name(_CRS, ResourceTemplate() {
> -            IO(Decode16, 0x03F2, 0x03F2, 0x00, 0x04)
> +            IO(Decode16, 0x03F0, 0x03F0, 0x00, 0x06)
>              IO(Decode16, 0x03F7, 0x03F7, 0x00, 0x01)
>              IRQNoFlags() { 6 }
>              DMA(Compatibility, NotBusMaster, Transfer8) { 2 }
> -- 
> 1.8.3.1
> 

Re: [Qemu-devel] [PATCH] floppy: fix I/O ranges in both portio and acpi dsdt

[Gerd Hoffmann]

On Do, 2015-05-28 at 18:36 +0200, Kevin Wolf wrote:
> Am 28.05.2015 um 17:37 hat Gerd Hoffmann geschrieben:
> > Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> > ---
> >  hw/block/fdc.c            | 2 +-
> >  hw/i386/acpi-dsdt-isa.dsl | 2 +-
> >  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> The commit message could be a bit longer, I don't think this change is
> obvious.
> 
> My FDC spec says that this version of the controller doesn't have the
> SRA register any more, and the emulation was added in the context of a
> Sun4m patch (commit 8c6a4d774). Therefore my default assumption would be
> that it's intentionally not accessible and not needed on isa-fdc.
> 
> I'm happy to get that assumption corrected, but it needs a non-empty
> commit message for that.

/me throws the towel.

I know next to nothing about floppy.  It's just that all parties ...

  (1) hw/block/fdc.c
  (2) hw/i386/acpi-dsdt-isa.dsl
  (3) hw/isa/lpc_ich9.c

... should agree on what the floppy io ranges are.  Right now they
don't, for whatever reasons.  And I simply don't know what are the
correct ranges.  So I leave fixing this up to whomever knows better this
than me.

cheers,
  Gerd