piix: Set I440FXState member pci_info.w32 in one place
commit97a83ec3a9d83f2e86b8b93178d8e8b64ccc7486
authorMarkus Armbruster <armbru@redhat.com>
Wed, 15 Jun 2016 17:56:30 +0000 (15 19:56 +0200)
committerMichael S. Tsirkin <mst@redhat.com>
Mon, 4 Jul 2016 11:50:59 +0000 (4 14:50 +0300)
tree8f7d00bfabb8144921601e2a32223fa259c17945
parent10d01f73e39100701028c7badd6ece52990cf758
piix: Set I440FXState member pci_info.w32 in one place

Range pci_info.w32 records the location of the PCI hole.

It's initialized to empty when QOM zeroes I440FXState.  That's a fine
value for a still unknown PCI hole.

i440fx_init() sets pci_info.w32.begin = below_4g_mem_size.  Changes
the PCI hole from empty to [below_4g_mem_size, UINT64_MAX].  That's a
bogus value.

i440fx_pcihost_initfn() sets pci_info.end = IO_APIC_DEFAULT_ADDRESS.
Since i440fx_init() ran already, this changes the PCI hole to
[below_4g_mem_size, IO_APIC_DEFAULT_ADDRESS-1].  That's the correct
value.

Setting the bounds of the PCI hole in two separate places is
confusing, and begs the question whether the bogus intermediate value
could be used by something, or what would happen if we somehow managed
to realize an i440FX device without having run the board init function
i440fx_init() first.

Avoid the confusion by setting the (constant) upper bound along with
the lower bound in i440fx_init().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
hw/pci-host/piix.c