| Age | Commit message (Collapse) | Author |
|---|
|
Introduce a new cbmem id to indicate romstage information. Proper
coordination with ramstage and romstage can use this cbmem entity
to communicate between one another.
Change-Id: Id785f429eeff5b015188c36eb932e6a6ce122da8
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2790
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
The current ramstage code contains uses of symbols that cause issues
when the ramstage is relocatable. There are 2 scenarios resolved by this
patch:
1. Absolute symbols that are actually sizes/limits. The symbols are
problematic when relocating a program because there is no way to
distinguish a symbol that shouldn't be relocated and one that can.
The only way to handle these symbols is to write a program to post
process the relocations and keep a whitelist of ones that shouldn't
be relocated. I don't believe that is a route that should be taken
so fix the users of these sizes/limits encoded as absolute symbols
to calculate the size at runtime or dereference a variable in memory
containing the size/limit.
2. Absoulte symbols that were relocated to a fixed address. These
absolute symbols are generated by assembly files to be placed at a
fixed location. Again, these symbols are problematic because one
can't distinguish a symbol that can't be relocated. The symbols
are again resolved at runtime to allow for proper relocation.
For the symbols defining a size either use 2 symbols and calculate the
difference or provide a variable in memory containing the size.
Change-Id: I1ef2bfe6fd531308218bcaac5dcccabf8edf932c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2789
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
There is a need to calculate the proper placement for an rmodule
in memory. e.g. loading a compressed rmodule from flash into ram
can be an issue. Determining the placement is hard since the header
is not readable until it is decompressed so choosing the wrong location
may require a memmove() after decompression. This patch provides
a function to perform this calculation by finding region below a given
address while making an assumption on the size of the rmodule header..
Change-Id: I2703438f58ae847ed6e80b58063ff820fbcfcbc0
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2788
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds a function for using the DCCMVAC instruction (dcache clean
by MVA at point of coherency (main memory)). We already have the
inline defined, it's just not used by anything.
Change-Id: Ia0641566a8881335bed8da2963e1db8321d74267
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2871
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
This adds a helper function for dcache ops by MVA which will perform
the specified operation on a given memory range. This will make it
more trivial to add other data cache maintenance routines.
Change-Id: I01d746d5fd2f4138257ca9cab9e9d738e73f8633
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2870
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
This clarifies and/or fixes formatting of some comments and
alphabetizes some function prototypes and inlines. It also
corrects references to "modified virtual address" (MVA).
Change-Id: Ibcdda4febf915cc4a1996a5bbb4ffecbcb50a324
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2869
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
This removes some old macros that we no longer use.
Change-Id: I9d87beb5c2deca228cdf89a98e54b2779be0f0ea
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2868
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
This just moves cache maintenance stuff from the armv7 bootblock
code to cache.c
Change-Id: I0b3ab58a1d8a3fe3d9568e02e156a36b6f33ca0b
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2867
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
The cbfs stage loading routine already zeros out the full
memory region that a stage will be loaded. Therefore, it is
unnecessary to to clear the bss again after once ramstage starts.
Change-Id: Icc7021329dbf59bef948a41606f56746f21b507f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2865
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Change-Id: I4aba6cc490ab24c6db345c0c5a64a6a9985ed0ab
Signed-off-by: Siyuan Wang <SiYuan.Wang@amd.com>
Signed-off-by: Siyuan Wang <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/2864
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
For diagnostic purposes we had been dumping the assembly
code when stages.o was built. We've past the need to do this
and it's confusing to watch.
Change-Id: Ib84cb73ed9dad3454efcb2be90d990ce88575229
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2555
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
|
|
Remove the spurious creation of a start symbol, and use the
stage_entry symbol directly.
Change-Id: Ia62d5c056ac8b20c8ffdb78bff3d306065b6c45f
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2560
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
|
|
Changes:
- Get rid of the h8scm mainboard specific code and use the
platform generic function wrapper that was added in change
http://review.coreboot.org/#/c/2777/
AMD Fam15: Add SPD read functions to wrapper code
- Move DIMM addresses into devicetree.cb
Notes:
- The DIMM reads only happen in romstage, so the function is not
available in ramstage. Point the read-SPD callback to a generic
function in ramstage.
Change-Id: I575221039ad65a59ae0f93397ef1038b669e81c7
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2829
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Changes:
- Get rid of the dinar mainboard specific code and use the
platform generic function wrapper that was added in change
http://review.coreboot.org/#/c/2777/
AMD Fam15: Add SPD read functions to wrapper code
- Move DIMM addresses into devicetree.cb
Notes:
- The DIMM reads only happen in romstage, so the function is not
available in ramstage. Point the read-SPD callback to a generic
function in ramstage.
- select_socket() and restore_socket() were created from code that
was removed from AmdMemoryReadSPD() in dimmSpd.c. The functionality
is specific to the dinar mainboard configuration and was therefore
split from the generic read SPD functionality.
Change-Id: I1e4b9a20dc497c15dbde6d89865bd5ee7501cdc0
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2830
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Changes:
- Get rid of the s8226 mainboard specific code and use the
platform generic function wrapper that was added in change
http://review.coreboot.org/#/c/2777/
AMD Fam15: Add SPD read functions to wrapper code
- Move DIMM addresses into devicetree.cb
Notes:
- The DIMM reads only happen in romstage, so the function is not
available in ramstage. Point the read-SPD callback to a generic
function in ramstage.
- select_socket() and restore_socket() started by duplicating
sp5100_set_gpio() and sp5100_restore_gpio(), which were in
dimmSpd.c. In addition to renaming the functions to more
specifically state their purpose, some cleanup and magic number
reduction was done.
Change-Id: I1eaf64986ef4fa3f89aed2b69d3f9c8c913f726f
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2827
Tested-by: build bot (Jenkins)
Reviewed-by: Siyuan Wang <wangsiyuanbuaa@gmail.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Changes:
- Get rid of the h8qgi mainboard specific code and use the
platform generic function wrapper that was added in change
http://review.coreboot.org/#/c/2777/
AMD Fam15: Add SPD read functions to wrapper code
- Move DIMM addresses into devicetree.cb
Notes:
- The DIMM reads only happen in romstage, so the function is not
available in ramstage. Point the read-SPD callback to a generic
function in ramstage.
- select_socket() and restore_socket() started by duplicating
sp5100_set_gpio() and sp5100_restore_gpio(), which were in
dimmSpd.c. In addition to renaming the functions to more
specifically state their purpose, some cleanup and magic number
reduction was done.
Change-Id: I346ebd8399d4ba3e280576e667fdc62fa75a63b8
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2828
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
This code is taken from an EDID reader written at Red Hat.
The key function is
int decode_edid(unsigned char *edid, int size, struct edid *out)
Which takes a pointer to an EDID blob, and a size, and decodes it into
a machine-independent format in out, which may be used for driving
chipsets. The EDID blob might come for IO, or a compiled-in EDID
BLOB, or CBFS.
Also included are the changes needed to use the EDID code on Link.
Change-Id: I66b275b8ed28fd77cfa5978bdec1eeef9e9425f1
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2837
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
This is a new state machine. It is more programmatic, in the
case of auxio, and has much more symbolic naming, and very few
"magic" numbers, except in the case of undocumented settings.
As before, the 'pre-computed' IO ops are encoded in the iodefs
table. A function, run, is passed and index into the table and
runs the ops.
A new operator, I, has been added. When the I operator is hit,
run() returns the index of the next operator in the table.
The i915lightup function runs the table. All the AUX channel ops
have been removed from the table, however, and are now called as
functions, using the previously committed auxio function.
The iodefs table has been grouped into blocks of ops, which end in
an I operator. As the lightup function progresses through startup,
and the run() returns, the lightup function performs aux channel
operations.
This code is symbolic enough, I hope, that it will make haswell
graphics bringup simpler.
i915io.c, and the core of the code in i915lightup.c, were
programatically generated, starting with IO logs from the DRM
startup code in the kernel. It is possible to apply the tools that
do this generation to newer IO logs from the kernel.
Change-Id: I8a8e121dc0d9674f0c6a866343b28e179a1e3d8a
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2836
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Add a new operator, P, for the state machine, meaning
implement a palette fill.
Implement a function (palette) that fills the palette when the
P operator is hit.
This replaces 256 lines in the state machine table with 1.
Change-Id: I67d9219fe7de0ecf1fb9faf92130c00c9f5f8e88
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/2835
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
There are some external libraries that are built within
coreboot's environment that expect a more common C standard
environment. That includes things like inttypes.h and UINTx_MAX
macros. This provides the minimal amount of #defines and files
to build vboot_reference.
Change-Id: I95b1f38368747af7b63eaca3650239bb8119bb13
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2859
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This is not needed in haswell.
Change-Id: I23817c2e01be33855f9d5a5e389e8ccb7954c0e2
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/2847
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: I853e381240b539b204c653404ca3d46369109219
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2846
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: I4585288905cf7374e671894ab37f125220ae535e
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2843
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
For full integration of FUI into coreboot, we need aux channel
communcations. The intel_dp.c is a file taken from Linux and is
used for aux channel comms. This file has been cut down to work
with coreboot. For now it is associated with the link mainboard
until we get a better handle on how this all fits together. This
code is almost certainly usable on other platforms in the long term.
But one step at a time.
Change-Id: I7be4c56e0a7903f3901ac86e12b28f3bdc0f7947
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/2834
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds a new API for cache maintenance operations. The idea is
to be more explicit about operations that are going on so it's easier
to manage branch predictor, cache, and TLB cleans and invalidations.
Also, this adds some operations that were missing but required early
on, such as branch predictor invalidation. Instruction and sync
barriers were wrong earlier as well since the imported API assumed
we compield with -march=armv5 (which we don't) and was missing
wrappers for the native ARMv7 ISB/DSB/DMB instructions.
For now, this is a start and it gives us something we can easily use
in libpayload for doing things like cleaning and invalidating dcache
when doing DMA transfers.
TODO:
- Set cache policy explicitly before re-enabling. Right now it's left
at default.
- Finish deprecating old cache maintenance API.
- We do an extra icache/dcache flush when going from bootblock to
romstage.
Change-Id: I7390981190e3213f4e1431f8e56746545c5cc7c9
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2729
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Coreboot's ramstage defines certain sections/symbols in its fixed
static linker script. It uses these sections/symbols for locating the
drivers as well as its own program information. Add these sections
and symbols to the rmodule linker script so that ramstage can be
linked as an rmodule. These sections and symbols are a noop for other
rmodule-linked programs, but they are vital to the ramstage.
Also add a comment in coreboot_ram.ld to mirror any changes made there
to the rmodule linker script.
Change-Id: Ib9885a00e987aef0ee1ae34f1d73066e15bca9b1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2786
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
This fixes a trivial error with the recovery mode GPIO index.
Change-Id: I7290c1e23cdddaf91c9021d4e4252c0c772b6eab
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2825
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
SPI accesses can be slow depending on the setup and the access pattern.
The current SPI hardware setup to cache and prefetch. The alternative
cbfs_load_payload() function takes advantage of the caching in the CPU
because the ROM is cached as write protected as well as the SPI's
hardware's caching/prefetching implementation. The CPU will fetch
consecutive aligned cachelines which will hit the ROM as
cacheline-aligned addresses. Once the payload is mirrored into RAM the
segment loading can take place by reading RAM instead of ROM.
With the alternative cbfs_load_payload() the boot time on a baskingridge
board saves ~100ms. This savings is observed using cbmem.py after
performing warm reboots and looking at TS_SELFBOOT_JUMP (99) entries.
This is booting with a depthcharge payload whose payload file fits
within the SMM_DEFAULT_SIZE (0x10000 bytes).
Datapoints with TS_LOAD_PAYLOAD (90) & TS_SELFBOOT_JUMP (99) cbmem entries:
Baseline Alt
-------- --------
90:3,859,310 (473) 90:3,863,647 (454)
99:3,989,578 (130,268) 99:3,888,709 (25,062)
90:3,899,450 (477) 90:3,860,926 (463)
99:4,029,459 (130,008) 99:3,890,583 (29,657)
90:3,834,600 (466) 90:3,890,564 (465)
99:3,964,535 (129,934) 99:3,920,213 (29,649)
Booted baskingridge many times and observed 100ms reduction in
TS_SELFBOOT_JUMP times (time to load payload).
Change-Id: I27b2dec59ecd469a4906b4179b39928e9201db81
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2783
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
In order to prepare the ramstage to be linked by the rmodule linker the
stack needs to be self-contained within the ramstage objects. The
reasoning is that the rmodule linker provides a way to define a heap,
but it doesn't currently have a region for the stack.
The downside to this is that memory footprint of the ramstage can change
when compared before this change. The size difference stems from the
link ordering of the objects as the stack is now defined within
c_start.S. The size fluctuation ranges from 0 to CONFIG_STACK_SIZE - 1
because of the previous behavior or aligning to CONFIG_STACK_SIZE. It
should be noted that such an alignment is unnecessary for 32-bit x86 as
the alignment requirement for the stacks are 4 byte alignment. Also the
memory footprint is still dominated by CONFIG_RAMTOP and CONFIG_RAMBASE.
Change-Id: I63a4ddd249104bc27aff2ab6b39fc6db12b54028
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2785
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
In certain situations boot speed can be increased by providing an
alternative implementation to cbfs_load_payload(). The
ALT_CBFS_LOAD_PAYLOAD option allows for the mainboard or chipset to
provide its own implementation.
Booted baskingridge board with alternative and regular
cbfs_load_payload().
Change-Id: I547ac9881a82bacbdb3bbdf38088dfcc22fd0c2c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2782
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
At some point, compiles with USB Debug port stopped working. This change makes
a trivial reordering in the code and adds two makefile entries to make it build
without errors. It also works on stout.
Build and boot as normal. Works. Enable CONFIG_USB, connect USB debug hardware
to the correct port (on stout, that's the one on the left nearest the back) and
watch for output.
Change-Id: I7fbb7983a19b0872e2d9e4248db8949e72beaaa0
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/2784
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
Change:
This is the initial step for moving the AMD F15 & HUDSON1,2,3
SPD-read callout out of the mainboard directories and into
the wrapper. The next step is to update the platforms to use
this routine in BiosCallouts.c and to delete the code from the
mainboard directories. The DIMM addresses should be moved into
devicetree.cb.
If there are significant differences or reasons that the mainboard
needs to override this code, it's perfectly reasonable to keep using
the version in the mainboard, but this allows us to remove duplicated
code and simplify the mainboard directories.
Notes:
This started by duplicating what was in Dinar, and was changed to
use the devicetree.cb structures. Significant cleanup and magic
number reduction was done as well.
It is intended that this file will not be included in ramstage as
the DIMM init is all done in romstage.
This is similar to what was done for Parmer/Thatcher in commit
7fb692bd - http://review.coreboot.org/#/c/2190/
Fam15tn: Move SPD read from mainboards into wrapper
Yes, it would make sense to split this into two separate files
and move the SMBus initialization and access into the southbridge
wrapper. Maybe that can come next.
Change-Id: I4e00ada288e1486cf30684403505e475f9093ec2
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2777
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
cbfstool usage change:
The "-a" parameter for "cbfstool locate" is switched to "-P/--page-size".
The "locate" command was used to find a place to store ELF stage image in one
memory page. Its argument "-a (alignment)" was actually specifying the page size
instead of doing memory address alignment. This can be confusing when people are
trying to put a blob in aligned location (ex, microcode needs to be aligned in
0x10), and see this:
cbfstool coreboot.rom locate -f test.bin -n test -a 0x40000
# output: 0x44, which does not look like aligned to 0x40000.
To prevent confusion, it's now switched to "-P/--page-size".
Verified by building i386/axus/tc320 (with page limitation 0x40000):
cbfstool coreboot.rom locate -f romstage_null.bin -n romstage -P 0x40000
# output: 0x44
Signed-off-by: Hung-Te Lin <hungte@chromium.org>
Change-Id: I0893adde51ebf46da1c34913f9c35507ed8ff731
Reviewed-on: http://review.coreboot.org/2730
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
|
|
A fix to eliminate warnings when building romstage files with ChromeOS
compilers
Change-Id: Ia5d7bbdde3aa3439fd493f5795f2cc2bf4c4c187
Signed-off-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-on: http://review.coreboot.org/2781
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
There should be a fixed 10ms wait after sending an INIT IPI. The
previous implementation was just waiting up to 10ms for the IPI to
complete the send. That is not correct. The 10ms is unconditional
according to the documentation. No ill effects were observed with the
previous behavior, but it's important to follow the documentation.
Change-Id: Ib31d49ac74808f6eb512310e9f54a8f4abc3bfd7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2780
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This patch parallelizes the AP startup for Haswell-based devices. It
does not touch the generic secondary startup code. Instead it provides
its own MP support matching up with the Haswell BWG. It seemed to be too
much trouble to support the old startup way and this new way. Because of
that parallel loading is the only thing supported.
A couple of things to note:
1. Micrcode needs to be loaded twice. Once before MTRR and caching is
enabled. And a second time after SMM relocation.
2. The sipi_vector is entirely self-contained. Once it is loaded and
written back to RAM the APs do not access memory outside of the
sipi_vector load location until a sync up in ramstage.
3. SMM relocation is kicked off by an IPI to self w/ SMI set as the
destination mode.
The following are timings from cbmem with dev mode disabled and recovery mode
enabled to boot directly into the kernel. This was done on the
baskingridge CRB with a 4-core 8-thread CPU and 2 DIMMs 1GiB each. The
kernel has console enabled on the serial port. Entry 70 is the device
initialization, and that is where the APs are brought up. With these two
examples it looks to shave off ~200 ms of boot time.
Before:
1:55,382
2:57,606 (2,223)
3:3,108,983 (3,051,377)
4:3,110,084 (1,101)
8:3,113,109 (3,024)
9:3,156,694 (43,585)
10:3,156,815 (120)
30:3,157,110 (295)
40:3,158,180 (1,069)
50:3,160,157 (1,977)
60:3,160,366 (208)
70:4,221,044 (1,060,677)
75:4,221,062 (18)
80:4,227,185 (6,122)
90:4,227,669 (484)
99:4,265,596 (37,927)
1000:4,267,822 (2,225)
1001:4,268,507 (685)
1002:4,268,780 (272)
1003:4,398,676 (129,896)
1004:4,398,979 (303)
1100:7,477,601 (3,078,621)
1101:7,480,210 (2,608)
After:
1:49,518
2:51,778 (2,259)
3:3,081,186 (3,029,407)
4:3,082,252 (1,066)
8:3,085,137 (2,884)
9:3,130,339 (45,202)
10:3,130,518 (178)
30:3,130,544 (26)
40:3,131,125 (580)
50:3,133,023 (1,897)
60:3,133,278 (255)
70:4,009,259 (875,980)
75:4,009,273 (13)
80:4,015,947 (6,674)
90:4,016,430 (482)
99:4,056,265 (39,835)
1000:4,058,492 (2,226)
1001:4,059,176 (684)
1002:4,059,450 (273)
1003:4,189,333 (129,883)
1004:4,189,770 (436)
1100:7,262,358 (3,072,588)
1101:7,263,926 (1,567)
Booted the baskingridge board as noted above. Also analyzed serial
messages with pcserial enabled.
Change-Id: Ifedc7f787953647c228b11afdb725686e38c4098
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2779
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This patch only applies to CONFIG_MICROCODE_IN_CBFS. The intel microcode
update routine would always walk the CBFS for the microcode file. Then
it would loop through the whole file looking for a match then load the
microcode. This process was maintained for intel_update_microcode_from_cbfs(),
however 2 new functions were exported:
1. const void *intel_microcode_find(void)
2. void intel_microcode_load_unlocked(const void *microcode_patch)
The first locates a matching microcode while the second loads that
mircocode. These new functions can then be used to cache the found
microcode blob w/o having to re-walk the CBFS.
Booted baskingridge board to Linux and noted that all microcode
revisions match on all the CPUs.
Change-Id: Ifde3f3e5c100911c4f984dd56d36664a8acdf7d5
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2778
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Having this header file in the mainboard directory breaks
the dinar build on cygwin because the header file in the
dinar mainboard is used instead of the correct header file
src/vendorcode/amd/cimx/sb700/OEM.h. The build probably works
fine on Linux systems because, due to case-sensitivity, Oem.h
will not match the #include "OEM.h" statement in
src/southbridge/amd/cimx/sb700/Platform.h.
The Oem.h file in the dinar mainboard is not used by any other
source files, and the defines in the dinar mainboard are duplicated
by defines in the correct OEM.h file. Therefore, the file can be
safely removed.
Change-Id: I81b97eca8116d63644d335edc3bb51f90c7094d9
Signed-off-by: Kimarie Hoot <kimarie.hoot@se-eng.com>
Reviewed-on: http://review.coreboot.org/2776
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
|
|
The non-relocatable SMM code was changed to link against libgcc a while back
so that printk could use built-in division instead of a hand crafted div()
function. However, the relocatable SMM code was not adapted by mistake.
This patch links the relocatable SMM against libgcc, too, so we can enable it
for Haswell.
Change-Id: Ia64a78e2e62348d115ae4ded52d1a02c74c5cea4
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2727
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
There are changes coming to perform more complex tasks after cache-as-ram
has been torn down but before ramstage is loaded. Therefore, add the
romstage_after_car() function to call after cache-as-ram is torn down.
Its responsibility is for loading the ramstage and any other complex
tasks. For example, the saving of OS-controlled memory in the resume
path has now been moved into C instead of assembly.
Change-Id: Ie0c229cf83a9271c8995b31c534c8e5a696b164e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2757
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The save_mrc_data() was previously called conditionally
in the raminit code. The save_mrc_data() function was called
in the non-S3 wake paths. However, the common romstage_common()
code was checking cbmem initialization things on s3 wake. Between
the two callers cbmem_initialize() was being called twice in the
non-s3 wake paths. Moreover, saving of the mrc data was not allowed
when CONFIG_EARLY_CBMEM_INIT wasn't enabled.
Therefore, move the save_mrc_data() to romstage_common. It already has
the knowledge of the wake path. Also remove the CONFIG_EARLY_CBMEM_INIT
requirement from save_mrc_data() as well as the call to cbmem_initialize().
Change-Id: I7f0e4d752c92d9d5eedb8fa56133ec190caf77da
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2756
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Instead of hard coding the policy for the stack and MTRR values after
the cache-as-ram is torn down, allow for the C code to pass those
policies back to the cache-as-ram assembly file. That way, ramstage
relocation can use a different stack as well as different MTRR policies.
Change-Id: Ied024d933f96a12ed0703c51c506586f4b50bd14
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2755
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This commit pulls in all the common logic for romstage into
the Haswell cpu directory. The bits specific to the mainboard
still reside under their respective directories. The calling
sequence bounces from the cpu directory to mainboard then back
to the cpu directory. The reasoning is that Haswell systems use
cache-as-ram for backing memory in romstage. The stack is used to
allocate structures. However, now changes can be made to the
romstage for Haswell and apply to all boards.
Change-Id: I2bf08013c46a99235ffe4bde88a935c3378eb341
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2754
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This structure is not used nor the variable being instantiated on the
stack. Remove them.
Change-Id: If3abe2dd77104eff49665dd33570b07179bf34f5
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2753
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
It was found that the Haswell reference code was smashing through the
stack into the reference code's heap implementation. The reason for this
is because our current CAR allocation is too small. Moreover there are
quite a few things to coordinate between 2 code bases to get correct.
This commit separates the CAR into 2 parts:
1. MRC CAR usage.
2. Coreboot CAR usage.
Pointers from one region can be passed between the 2 modules, but one
should not be able to affect the others as checking has been put into
place in both modules.
The CAR size has effectively been doubled from 0x20000 (128 KiB) to
0x40000 (256KiB). Not all of that increase was needed, but enforcing
a power of 2 size only utilizes 1 MTRR.
Old CAR layout with a single contiguous stack with the region starting
at CONFIG_DCACHE_RAM_BASE:
+---------------------------------------+ Offset CONFIG_DCACHE_RAM_SIZE
| MRC global variables |
| CONFIG_DCACHE_RAM_MRC_VAR_SIZE bytes |
+---------------------------------------+
| ROM stage stack |
| |
| |
+---------------------------------------+
| MRC Heap 30000 bytes |
+---------------------------------------+
| ROM stage console |
| CONFIG_CONSOLE_CAR_BUFFER_SIZE bytes |
+---------------------------------------+
| ROM stage CAR_GLOBAL variables |
+---------------------------------------+ Offset 0
There was some hard coded offsets in the reference code wrapper to start
the heap past the console buffer. Even with this commit the console
can smash into the following region depending on what size
CONFIG_CONSOLE_CAR_BUFFER_SIZE is.
As noted above This change splits the CAR region into 2 parts starting
at CONFIG_DCACHE_RAM_BASE:
+---------------------------------------+
| MRC Region |
| CONFIG_DCACHE_RAM_MRC_VAR_SIZE bytes |
+---------------------------------------+ Offset CONFIG_DCACHE_RAM_SIZE
| ROM stage stack |
| |
| |
+---------------------------------------+
| ROM stage console |
| CONFIG_CONSOLE_CAR_BUFFER_SIZE bytes |
+---------------------------------------+
| ROM stage CAR_GLOBAL variables |
+---------------------------------------+ Offset 0
Another variable was add, CONFIG_DCACHE_RAM_ROMSTAGE_STACK_SIZE,
which represents the expected stack usage for the romstage. A marker
is checked at the base of the stack to determine if either the stack
was smashed or the console encroached on the stack.
Change-Id: Id76f2fe4a5cf1c776c8f0019f406593f68e443a7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2752
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Add an rmodules class so that there are default rules for compiling
files that will be linked by the rmodule linker. Also, add a new type
for SIPI vectors.
Change-Id: Ided9e15577b34aff34dc23e5e16791c607caf399
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2751
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This is causing a hang in depthcharge. For now just disable
this port.
Change-Id: I87a6db2d8361588e82eee640c74cea690115bed5
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/2764
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
By including the heap in the bss output section the size is accounted
for in a elf PT_LOAD segment. Without this change the heap wasn't being
put into a PT_LOAD segment. The result is a nop w.r.t. functionality,
but readelf and company will have proper MemSiz fields.
Change-Id: Ibfe9bb87603dcd4c5ff1c57c6af910bbba96b02b
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2750
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
There is a plan to utlize rmodules for loading ramstage as a
relocatable module. However, the rmodule header may change.
In order to provide some wiggle room for changing the contents
of the rmodule header add some padding. This won't stop the need
for coordinating properly between the romstage loader that may be
in readonly flash and rmodule header fields. But it will provide
for a way to make certain assumptions about alignment of the
rmodule's program when the rmodule is compressed in the flash.
Change-Id: I9ac5cf495c0bce494e7eaa3bd2f2bd39889b4c52
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2749
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Haswell's BCLK is fised at 100MHz like Sandy/Ivy. Add Haswell's model
to the switch statement.
Change-Id: Ib9e2afc04eba940bfcee92a6ee5402759b21cc45
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2747
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
