Age | Commit message (Collapse) | Author |
|
The overhead of pushing variables to stack exceeded the number
of instructions the actual MMIO operation took and the build of
google/aleena with inlined accessors turned out to be just
slightly (<2 KiB) smaller for the entire romstage or ramstage.
Simple read-modify-write MMIO cycles should optimise better now.
IO cycles with index/data register are borderline, at
first sight assembly looked better by not inlining them.
Change-Id: If2c37c9886a0151183aa6dd80eb068d6c67b3848
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37211
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
|
|
If one wishes to use the functions guarded here, he
has to have datasheet open anyways. It should be clear
from there which regions are supported and which are not.
TEST=Reproducible build of google/aleena.
Change-Id: I0c1f0c9c9a6711532c5078c08cdf9e6612f3bc9c
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37210
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
|
|
The standard PCI register space for D14F0 is accessible at 0xfed80000.
Add functions for use as helpers.
Change-Id: Icbf5bdc449322c3f5e59e6126d709cb2808591d5
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34914
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
|
|
Include the file containing the typedefs for uint_*.
Change-Id: I3eae80a677e9d6932dc115523da2c0819a371fa7
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35268
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
|
|
Add a method for the soc/amd/<product> to indicate what AcpiMmio
ranges are supported. Induce a build error if soc or mainboard
code is added which attempts to use an unsupported block.
This patch attempts to dissuade accessing unsupported blocks without
requiring the complexity of structures or reinitializing at the
beginning of a new stage.
TEST=boot grunt, force build errors by removing blocks in iomap.h
BUG=b:131682806
Change-Id: I2121df108fd3caf07e5588bc3201bcdd8dcaaa00
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32934
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
|
|
Document the AcpiMmio individual blocks better. This is in response
to a request in gerrit for
69486cac7: Create AcpiMmio functionality from stoneyridge
Correct comments that were inadvertently left in place from older
patches.
Change-Id: I4c16a866de5622e8cfbd3a08816b9d3182950d0e
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32931
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
|
|
Change all arguments named "offset" to "reg" to match the others.
These should have gone into change
69486cac7: Create AcpiMmio functionality from stoneyridge
Change-Id: Ifdd00d0a5d1e03bfa68a13eeece2d2cfd56aa39d
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32930
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
|
|
Reorder the biosram prototypes to match the rest of the file. Remove
prototypes for asf 32-bit functions that have not been implemented.
Change-Id: Ic2663158d8a71952c26eb37f34342a6ea5e58a42
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32929
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
|
|
Move the stoneyridge AcpiMmio code into soc/amd/common.
The SB800 southbridge introduced the MMIO hardware blocks at 0xfed80000
commonly known as AcpiMmio. Implementations beginning with Mullins
enable decode in PMx04. Older designs use PMx24 and allow for
configuring the base address. Future work may support the older version.
Comparing the documentation for AMD's RRGs and BKDGs, it is evident that
the block locations have not been reassigned across products. In some
cases, address locations are deprecated and new ones consumed, e.g. the
early GPIO blocks were simpler at offset 0x100 and the newer GPIO banks
are now at 0x1500, 0x1600, and 0x1700.
Note: Do not infer the definitions within the hardware blocks are
consistent across family/model products.
BUG=b:131682806
Change-Id: I083b6339cd29e72289e63c9331a815c46d71600d
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32649
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
|