The legacy DMA is not used by linux. This change frees up those IO
ports.
When FSP-S runs, it re-enables the legacy DMA IO region, so we need to
disable it again.
BOOTBLOCK: PMx00: 0xe3060bf3
ROMSTAGE - Before FSP: PMx00: 0xe3060bf3
ROMSTAGE - After FSP: PMx00: 0xe3060bf7
BUG=b:180949454
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I7792d1f8ea40eb1c7f6cca67e9907208884ac694
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51076
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
With this change NMI works in the kernel:
----------------
| NMI testsuite:
--------------------
remote IPI: ok |
local IPI: ok |
--------------------
Good, all 2 testcases passed! |
---------------------------------
See setup_lapic() for where this gets configured.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ia391ec5a015d909462ff8aaf3cb047c6fd45fe0a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50562
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Mathew King <mathewk@chromium.org>
This causes the linux kernel to complain:
32/64X address mismatch in FADT/Pm1aEventBlock: 0x00000400/0x00000000FED80800
32/64X address mismatch in FADT/Pm1aControlBlock: 0x00000404/0x00000000FED80804
32/64X address mismatch in FADT/PmTimerBlock: 0x00000408/0x00000000FED80808
32/64X address mismatch in FADT/Gpe0Block: 0x00000420/0x00000000FED80814
The linux kernel also verifies that the PM Timer block only uses IO
ports.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I612b6bfb67d8559127ab2ee8a2fb828493820e31
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51074
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Printing this in both bootblock and romstage is redundant, since the CPU
family and model aren't expected to change between bootblock and
romstage entry.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
Change-Id: Id7c6aea0d8a6dac39114593584e534661faea89d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51062
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
The PSP soft fuse bit 6 doesn't do what the comment above it says. See
NDA document #55758 for details.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ic84cf6e1eee30af92cd700dc4bf78290143bf88b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51040
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
This differs slightly from picasso. The PCI BAR region is between TOM1
and CONFIG_MMCONF_BASE_ADDRESS. This matches what the Intel platforms
are doing. It also matches what linux derives from the e820 tables:
> [mem 0xd0000000-0xf7ffffff] available for PCI devices
Picasso currently declares the region between TOM and IO_APIC_ADDR.
This region includes MMCONF. We don't want to map any PCI BARs in this
region.
TEST=Boot majolica and check logs
pci_bus 0000:00: root bus resource [io 0x0000-0x0cf7 window]
pci_bus 0000:00: root bus resource [io 0x0d00-0xffff window]
pci_bus 0000:00: root bus resource [mem 0x000a0000-0x000bffff]
pci_bus 0000:00: root bus resource [mem 0x000c0000-0x000dffff]
pci_bus 0000:00: root bus resource [mem 0xd0000000-0xf7ffffff]
pci_bus 0000:00: root bus resource [bus 00-3f]
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I4ff02012795e2166e3a4197071b1136727089318
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50893
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The devices were copied from picasso with the following modifications:
* UART{2,3} were deleted
* I2C{0,1} were added
* eMMC was removed since it hasn't been validated
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Iddfb975e9292785d0951dd7bb31c1997d2185abd
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50573
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
We were missing this, so we ran into the scope assert in
acpi_device_write_pci_dev for the data fabric PCI devices.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I566791527ba839ba52ec5fa28f0f6c25f547d1da
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50815
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
TEST=Boot majolica to linux and see IO-APIC logs
ACPI: Local APIC address 0xfee00000
ACPI: LAPIC_NMI (acpi_id[0xff] high edge lint[0x1])
IOAPIC[0]: apic_id 16, version 33, address 0xfec00000, GSI 0-23
ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
ACPI: IRQ0 used by override.
ACPI: IRQ9 used by override.
Using ACPI (MADT) for SMP configuration information
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ib8094c3edf401659d9d740e2cc6266ddd5f91da9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50803
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Since all bridges to the internal buses have the same PCI ID, we can
just add this one ID to the pci_driver struct and don't need to use a
list of PCI IDs.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ice024b91f49f03995acbd8dfc8b33d3ae3559dde
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50804
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This fixes the undefined reference for NVB0, NVB1, and NVB2.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ib4ba24b66b9ae7899ccd40f91cdd23074f6afc4b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50614
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Additionally to the PCI IDs of Cezanne it also handles the Renoir ones.
The main difference between those two is that Renoir has two core
complexes while Cezanne only has one core complex. I haven't seen
incompatible changes between those two though, so for example the fabric
IDs are the same and the one that's only present in Renoir is just not
used in Cezanne. Also adding the ACPI parts for those don't have
anything to do with those differences.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I3b2517bc15d872f41183a33857333f1972ff2cb9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50706
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
I'm not 100% sure yet if this code will be common for all AMD SoCs, so
I'll add a copy for Cezanne for now. This part of the code should
probably be reworked after the initial bringup of Cezanne anyway.
DF MMIO register configuration at the beginning of
data_fabric_set_mmio_np:
=== Data Fabric MMIO configuration registers ===
Addresses are shifted to the right by 16 bits.
idx control base limit
0 a3 fc00 febf
1 a3 1000000 fffcffff
2 a3 d000 f7ff
3 a0 0 0
4 a3 fed0 fed0
5 a0 0 0
6 a0 0 0
7 a0 0 0
DF MMIO register configuration at the end of data_fabric_set_mmio_np:
=== Data Fabric MMIO configuration registers ===
Addresses are shifted to the right by 16 bits.
idx control base limit
0 a3 fc00 febf
1 a3 1000000 fffcffff
2 a3 d000 f7ff
3 10a3 fed0 fedf
4 a0 0 0
5 a0 0 0
6 a0 0 0
7 a0 0 0
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia243a0cad311eb210d14d6242c52f599db22515c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50624
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This will be common for all boards, so move it to the chipset device
tree.
TEST=CPU cluster and LAPIC still show up in console logs.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia49e7b4cfc09c60b6152b8ccc47f37b6adc1e319
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50613
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The MADT doesn't populate the IO-APICs yet since we need FSP to
configure those.
The FADT differs from picasso in the following ways:
* The duty_offset is supposed to be 0
* Don't clear x_firmware_ctl_l
* Make the extended addresses use MMIO
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ib6c3a01084a0de33894885b47c637a292d252ed4
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50510
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
TEST=Boot majolica and see microcode update
CBFS: Found 'cpu_microcode_blob.bin' @0x6900 size 0x15c0 in mcache @0xcf7fe9d8
microcode: patch id to apply = 0x0a50000b
microcode: being updated to patch id = 0x0a50000b succeeded
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: If50b1d8b3ebf4b3e6f8a9dd3ab96073e0cb92424
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50616
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
When USE_PSPSECUREOS isn't selected, we don't even get post codes on
Majolica, so remove this option.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I0ec976f40d962a2d9f2bd36dc97d86526bd661ae
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50601
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Without this part being present in amdfw the PSP won't enter its normal
operation mode, but goes into recovery mode instead.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: If13b5495e9f246afd0317daaa53c3d2cefbaa4dc
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50600
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Matt Papageorge <matthewpapa07@gmail.com>
A copy of Picasso's include/nvs.h was added to Cezanne right before the
commit d6ccbb9d48 that removed it for the
other mainboards and SoCs, so apply the equivalent change here as well
to keep everything in sync.
Change-Id: I76b551c05b3c3028a3afb3bc3b77df2401aed7a8
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50563
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
These definitions were identical to picasso. The only thing I changed
was that I renamed Misc1 and Misc2 to HPET_L and HPET_H.
This change still doesn't write the PCI_IRQ register for all the PCI
devices. We need to refactor the picasso pci_gpp code first.
TEST=Boot majolica and see FCH IRQs being programmed.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ic7e637f234d3af426959a9bbd82a0dcf25bb3c8e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50451
Reviewed-by: Mathew King <mathewk@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
TEST=Majolica still gets to SeaBIOS. Like before this patch the PSP
still has the recovery flag set in its return value, but we likely still
have some problem in the amdfw part or miss some PSP initialization in
FSP.
Change-Id: I9f343452ef2ea6b01f9b2fd0cf6371218d046046
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50537
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The RTC functionality will be used by elog.
Change-Id: I3a8d0a353620f64207d5ba8e17c145090f0c7506
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50543
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Since SMM is in TSEG on the platforms which is the default, drop the
SMM_TSEG condition for the default of SMM_TSEG_SIZE.
Change-Id: I7bd965c0794efa12ea4886a55522cc5193a1d3ac
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50498
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Only the ACPI enable/disable functionality is implemented and sleep is
also not implemented yet. This will be added in future patches.
Change-Id: I7701944023ce2e86586679c32c4138d4488768a1
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50488
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This doesn't select HAVE_ACPI_TABLES, so no ACPI tables will be
generated for now. There's also no globalnvs.asl that corresponds to
nvs.h yet. The added nvs.h has some currently unused fields, but still
having them in the struct aligns it with Picasso and also might reduce
the noise in future ACPI patches a bit. When most of the ACPI code for
Cezanne has landed, we need to do a cleanup though.
Change-Id: I3d658d284fa67e4da43a89d74686445fd5e93b1f
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50487
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The MCA MSRs aren't getting cleared and no microcode update gets applied
for now. Both will be added later.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I38ce5d11787ffefdd0183c5540ae2683158cbee8
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50482
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The common code gets moved to soc/amd/common/block/cpu/smm, since it is
related to the CPU cores and soc/amd/common/block/smi is about the SMI/
SCI functionality in the FCH part. Also relocation_handler gets renamed
to smm_relocation_handler to keep it clear what it does, since it got
moved to another compilation unit.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I45224131dfd52247018c5ca19cb37c44062b03eb
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50462
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
