QEMU commit b7ee1603c16c1feb0d439d2ddf6cf824119d0aab improved the PCI bridge
device handling. PCI bridges use different header format from ordinary PCI
devices, especially BAR registers do not exist. OpenBIOS tried to configure
the bridges based on the non-BAR data which was assumed to be BARs.
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@505 f158a5a8-5612-0410-a976-696ce0be7e32
This changes pocket index calculation to support more than two pockets.
As an example default number of pockets is set to 4.
This should ease nesting more than two levels using pockets as
temporary storage.
Signed-off-by: Igor Kovalenko <igor.v.kovalenko@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@494 f158a5a8-5612-0410-a976-696ce0be7e32
Wrong encoding of "#address-cells" property can lead to a failure
fetching correct value in my-#acells method. According to docs
properties must be encoded with "encode-int". I spent some time
looking at the docs and it is clear that encode-int produces
quad-sized result (of size /l bytes).
This patch fixes pci helpers to encode to 32bit instead of 64bit values,
and correctes my-address and my-unit methods to read 32bit data
to match encode-int rules
modules/bindings.c: also fixed set_int_property and get_int_property
to match encode-int rules
Signed-off-by: Igor Kovalenko <igor.v.kovalenko@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@493 f158a5a8-5612-0410-a976-696ce0be7e32
struct ata_sector was meant to facilitate taking upper and lower byte of
the sector. However, the implementation is incorrect, as "struct" and
"union" are swapped in the definition. What's worse, it's an overkill
for that simple task. The code should be transparent without knowing
how struct ata_sector is defined.
[laurent: Original patch modified to remove useless "& 0xff"]
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@489 f158a5a8-5612-0410-a976-696ce0be7e32
This is needed on Fedora 11 (GNU ld 2.19.51.0.2). Otherwise, qemu
reports on statup:
invalid/unsupported opcode: 00 - 18 - 01 (00004070) 00000004 1
invalid/unsupported opcode: 00 - 04 - 17 (000095c8) 000095ec 0
I checked Linux sources (arch/powerpc/kernel/vmlinux.lds.S) and I see
that .sbss is put to the .bss section in front of .bss itself.
The problem is discussed at
https://bugzilla.redhat.com/show_bug.cgi?id=494075
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@488 f158a5a8-5612-0410-a976-696ce0be7e32
Use firmware configuration device for boot device, kernel, initrd and
kernel command line parameters on PPC, Sparc32 and Sparc64.
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@479 f158a5a8-5612-0410-a976-696ce0be7e32
Expose PCI controller information
Linux tries to detect which PCI controller it's supposed to work with.
On PPC32 there is fallback code that doesn't get compiled with
CONFIG_PPC64, so we really have to expose a controller or we don't have
pci config space accessor functions.
Let's expose a un3-agp if we have a mac99 model, as that's what is
supposed to be in there anyways.
Signed-off-by: Alexander Graf <alex@csgraf.de>
Signed-off-by: Laurent Vivier <Laurent@vivier.eu>
git-svn-id: svn://coreboot.org/openbios/openbios-devel@466 f158a5a8-5612-0410-a976-696ce0be7e32
Expose NVRAM compatibility
Linux requires the NVRAM to expose a 'nvram,flash' compatibility
in the device tree. Let's expose it so it's happy.
v2 adds a NEWWORLD macro to protect nvram,flash getting exposed in
oldworld macs.
Signed-off-by: Alexander Graf <alex@csgraf.de>
Signed-off-by: Laurent Vivier <Laurent@vivier.eu>
git-svn-id: svn://coreboot.org/openbios/openbios-devel@465 f158a5a8-5612-0410-a976-696ce0be7e32
PPC64: Switch to 32-bit mode in interrupts
All the code in OpenBIOS runs in 32 bits, which is fine. Now, when
an interrupt arrives, the 970 CPU automatically sets the MSR_SF bit,
making us run in 64 bit mode.
That breaks address calculation, because 'lis' and 'ba' get sign
extended.
In order to circumvent that mess, let's just go back to 32 bit mode
whenever we get an interrupt.
Signed-off-by: Alexander Graf <alex@csgraf.de>
Signed-off-by: Laurent Vivier <Laurent@vivier.eu>
git-svn-id: svn://coreboot.org/openbios/openbios-devel@464 f158a5a8-5612-0410-a976-696ce0be7e32
While booting a 64bit kernel, there is a small timeframe where OF and the kernel communicate with each other. Within that timeframe, DSI/ISI interrupts may occur, because some memory is not mapped yet.
Right now in case that happens, we jump into the DSI/ISI interrupt handler which clobbers the high 32 bits of the kernel's registers. In order to circumvent that, let's save/restore all 64 bits of all kernel registers when we get a DSI/ISI interrupt.
This patch enables a PPC64 Linux kernel to boot up to the point where it tries to set up the SLB entries (slbmte), which is not yet implemented in qemu.
v2 implements handling for the HIOR register, bringing interrupt handlers to RAM.
Signed-off-by: Alexander Graf <alex@csgraf.de>
git-svn-id: svn://coreboot.org/openbios/openbios-devel@462 f158a5a8-5612-0410-a976-696ce0be7e32
It gets up to the boot prompt and works rather good so far, though I haven't been able to run a kernel yet.
For more recent PowerPC CPUs the PTE layout has changed, so we need to take that into account and create PTEs according to the new layout and at the new physical positions.
Signed-off-by: Alexander Graf <alex@csgraf.de>
Signed-off-by: Laurent Vivier <Laurent@vivier.eu>
git-svn-id: svn://coreboot.org/openbios/openbios-devel@461 f158a5a8-5612-0410-a976-696ce0be7e32
It is needed to boot Fedora-10 as we must specify the path to the yaboot.conf.
To boot Fedora:
boot cd: conf=cd:,\ppc\ppc32\yaboot.conf
so "bootpath" is "cd:", and "bootargs" is "conf=cd:,\ppc\ppc32\yaboot.conf"
git-svn-id: svn://coreboot.org/openbios/openbios-devel@456 f158a5a8-5612-0410-a976-696ce0be7e32
