Support for AMD's latest G-Series system-on-chip (SoC) family of x86
processors has been fully released into the coreboot community. I very
much appreciate everyone who contributed comments on the early code posted
to Gerrit, specifically Kyösti, Paul, Ron, Stepan, and the folks at Sage.
I think I have addressed all of the comments that I can to-date. Shoot me
an e-mail if theres something I missed. Please review any of the code
that interests you and post comments.
The set of patches for this release include the following:
Platform Initialization (AGESA):
The AGESA/PI code is a dump from AMDs internal BIOS team and therefore
doesnt necessarily meet coreboot.org standards. The hope is that this
one will see more regular updates than previous releases.
24f441d AMD Kabini: Add AGESA/PI code for new processor family
(http://review.coreboot.org/#/c/3836/)
7e1b725 AMD Kabini: Add "const" modifier to AGESA function parameters
(http://review.coreboot.org/#/c/3837/)
9a24f45 AMD AGESA: Add missing breaks to switch statement in one file
(http://review.coreboot.org/#/c/3839/)
265edf4 AMD AGESA: Fix comment for `PCIE_DDI_DATA_INITIALIZER`
(http://review.coreboot.org/#/c/3840/)
6c4c278 AMD SATA: Correct "them implement" to "then implement" in
comments
(http://review.coreboot.org/#/c/3841/)
AGESA Wrapper code for the Family 16 (Kabini) CPU
334df15 AMD Kabini: Add CPU AGESA wrapper for new AMD processor family
(http://review.coreboot.org/#/c/3781/)
AGESA Wrapper code for the Family 16 (Kabini) northbridge / graphics
e67b533 AMD Kabini: Add northbridge AGESA wrapper (new AMD processor)
(http://review.coreboot.org/#/c/3782/)
6c7cffc AMD Kabini: Add map_oprom() function for Vendor/Device IDs
(http://review.coreboot.org/#/c/3806/)
AGESA Wrapper code modifications to Hudson code for the integrated
southbridge (Yangtze)
480b69a AMD Kabini: Modify Hudson southbridge to support new AMD
processor
(http://review.coreboot.org/#/c/3783)
d7cc030 AMD Hudson/Yangtze: Enable support for SATA port multipliers
(http://review.coreboot.org/#/c/3808/)
d71234f AMD Hudson: Add wrapper functions to enable LPC Super I/O ports
(http://review.coreboot.org/#/c/3842/)
Note: The original code in early_setup.c was used during
Kabini/Yangtze development. It is not used on Olive
Hill.
By splitting it out as a separate changelist, the hope is
that it will be useful on future AMD-based boards that
have Super I/O chips.
AMD Olive Hill development board (DB-FT3)
44ee758 AMD Olive Hill: Add new AMD mainboard using Kabini processor
(http://review.coreboot.org/#/c/3784/)
fcee4b2 AMD Olive Hill: Eliminate unnecessary memory copy
(http://review.coreboot.org/#/c/3818/)
6ec03a2 AMD Olive Hill: Enable HDMI audio setting in build options
(http://review.coreboot.org/#/c/3814/)
c6424d6 AMD Olive Hill: Change SB800 references to Yangtze
(http://review.coreboot.org/#/c/3815/)
594cfc1 AMD Olive Hill: Remove default VBIOS vendor/device ID
(http://review.coreboot.org/#/c/3816/)
79028f8 AMD Olive Hill: Enable WARNINGS_ARE_ERRORS (remove override)
(http://review.coreboot.org/#/c/3819/)
ACPI changes for consistency with community ACPI generalization work on
Persimmon, Parmer, ASUS F2A85-M:
fda169a AMD Fam16kb: Split DSDT into common sections
(http://review.coreboot.org/#/c/3821/)
19b3483 AMD Fam16: Add secondary bus number to CRES method
(http://review.coreboot.org/#/c/3822/)
0578a98 AMD Fam16: Add OSC method to PCI0
(http://review.coreboot.org/#/c/3823/)
After some toing and froing with Aaron Durbin over the last 6 hours
this is where we are:
mrc.cache is now working but cold boot still takes roughly 5 seconds to
get to SeaBIOS display. Warm boot/reset is instantaneous. Aaron says the
mrc.bin blob is not happy (and we can't have an unhappy blob now, can
we).
In order to get mrc.cache working this is what's required:
1. Remove 'select CHROMEOS' from src/mainboard/samsung/lumpy/Kconfig.
2. Add this patch:
diff --git a/src/vendorcode/google/chromeos/gnvs.h
b/src/vendorcode/google/chrom
index 00fe443..5384423 100644
--- a/src/vendorcode/google/chromeos/gnvs.h
+++ b/src/vendorcode/google/chromeos/gnvs.h
@@ -61,9 +61,13 @@ typedef struct {
// dbe
} __attribute__((packed)) chromeos_acpi_t;
+#if CONFIG_CHROMEOS
extern chromeos_acpi_t *vboot_data;
void chromeos_init_vboot(chromeos_acpi_t *chromeos);
void chromeos_set_me_hash(u32*, int);
void acpi_get_vdat_info(uint64_t *vdat_addr, uint32_t *vdat_size);
+#else
+static inline void chromeos_init_vboot(chromeos_acpi_t *chromeos) {}
+#endif
#endif
3. make oldconfig
4. make clean
5. make (obviously)
Additionally, to get timing information in cbmem -c and cbmem -t the
following was necessary:
1. Add 'select LAPIC_MONOTONIC_TIMER' to
src/mainboard/samsung/lumpy/Kconfig
2. make oldconfig
3. make (again obviously)
The cbmem output is attached. Where to go from here? Many thanks to
Aaron for all his help.
On Fri, 26 Jul 2013 19:17:50 -0400
David Giard <swe3tdave(a)gmail.com> wrote:
> Will remember to put a temporary jumper on the reset switch
> header next time.
Which will not help on new Intel-based boards AFAIK because they have
edge-triggered reset inputs. I have documented the issues David
mentioned and more here: http://flashrom.org/ISP
--
Kind regards/Mit freundlichen Grüßen, Stefan Tauner
Hello! Someone recently contacted me directly about my MA785GM-US2H coreboot port, and they weren't able to get the microcode to generate:
HOSTCC cbfstool/cbfstool (link)
LD cpu_microcode_blob.o
coreboot/util/crossgcc/xgcc/bin/i386-elf-ld: no input files
make: *** [build/cpu_microcode_blob.o] Error 1
I've tested the compilation process here too and I get the same problem, not sure how to fix it. It doesn't work for me no matter what board I select in Kconfig.
Please note I am not subscribed to the coreboot mailing list so if you have a response, I kindly suggest you CC the email to me directly.
Alec Ari <neotheuser(a)ymail.com>
