The caching policy for romstage was previously using a 32KiB
of cache-as-ram for both the MRC wrapper and the romstage stack/data.
It also used a 32KiB code cache region. The BWG's limitations for
the code and data region before memory is up was wrong. It consists
of a 16-way set associative 1MiB cache. As long as enough addresses
are not read there isn't a risk of evicting the data/stack.
Now create a 64KiB cache-as-ram region split evenly between romstage
and the MRC wrapper. Additionally cache the memory just below
4GiB in CBFS size. This will cover any code and read-only data needed.
BUG=chrome-os-partner:22858
BRANCH=None
TEST=Built and booted quickly with corresponding changes to MRC warpper.
CQ-DEPEND=CL:*146175
Change-Id: I021cecb886a9c0622005edc389136d22905d4520
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/172150
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4868
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Like the bunit and dunit, add the punit accessor functions.
BUG=chrome-os-partner:23085
BRANCH=None
TEST=Built.
Change-Id: Ifd7184dfca8c0491c107bc1c562ea1ded444e372
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171931
Reviewed-on: http://review.coreboot.org/4867
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
- Set config0 defaults for hysteresis disable, pad bypass, etc.
- Set config1 power-on defaults.
- Set pad_val for input as default.
BUG=chrome-os-partner:22863
TEST=Manual. Enable GPIO_DEBUG and verify pad registers are set
according to expectation. Also verify bayleybay still boots to payload
loading.
Change-Id: I0f1c9e4d4f39c5c56d7e14a82eb4825612e19420
Reviewed-on: https://chromium-review.googlesource.com/171903
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Commit-Queue: Shawn Nematbakhsh <shawnn@chromium.org>
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Tested-by: Shawn Nematbakhsh <shawnn@chromium.org>
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4866
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
We should not have x86 specific includes in lib/.
Change-Id: I18fa9c8017d65c166ffd465038d71f35b30d6f3d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5156
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@google.com>
Needs printk and is not a console core function.
Change-Id: Id90a363eca133af4469663c1e8b504baa70471e0
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5155
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
These options seem to control the behavior of the encoder/decoder,
with comments citing a trade-off between memory usage and performance.
I removed these in a separate patch to make reverting in the future
easier, if we find these options are useful.
Change-Id: I24cb7101b89e60f4fb96777e3681c03d2a62e3d5
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/5084
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Rather than using [hn]to[nh] whenever accessing a member of the CBFS
header, deserialize the header when opening the CBFS image. The header
is no longer a pointer inside the CBFS buffer, but a separate struct,
a copy of the original header in a host-friendly format. This kills
more of the ntohl usage.
Change-Id: I5f8a5818b9d5a2d1152b1906249c4a5847d02bac
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/5121
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Basic ACPI support for this old platform. Created by copying and
tweaking similar motherboard ACPI implementations in coreboot.
Works reasonably well under Linux, providing HPET-timers
and more under linux (tested under OpenSUSE 12.2 kernel 3.4.63-2.44).
Not tested under Windows.
Change-Id: I69431be962a0d272db398ecf4ac9f0249de8ebab
Signed-off-by: Oskar Enoksson <enok@lysator.liu.se>
Reviewed-on: http://review.coreboot.org/5185
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
There are EHCI compatible host controllers on ARM without PCI bus
architecture. Currently we have not come across one with the debug
capability though.
Change-Id: I8775c9814f6fdf8754f97265118a7186369d721d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5175
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
USB device end toggles data PID when we ACK'd the zero-length data
packet. As USB host we need to toggle data PID too or the next data
received would get discarded.
Change-Id: I3203bc874c7ded9244c7548a666d7041a0fbb379
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4775
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
This claim is useless when done before EHCI controller reset. Code in
usbdebug_init_() already sets this properly after reset, see use of
DBGP_OWNER.
Change-Id: Ic17493fe4edbbbed6ebcbef35a264fbf188f1fba
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4709
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Read from USB endpoint_in 8 bytes at a time, the maximum what
EHCI debug port capability has to offer.
Change-Id: I3d012d758a24b24f894e587b301f620933331407
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4700
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
This was designed as a micro-optimization for x86, but it is only used
once. Let the compiler decide if optimizing this is worth the effort.
Change-Id: I5939efa34f0e9d16643893ca04675247842e7db5
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/5085
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
LzmaEnc.c was full of shadow definitions. Luckily, shadow definitions
were not used after the scope in which they were redefined, so it is
possible to just remove them.
Tested by successfully booting qemu i440fx to grub2 payload.
Change-Id: I01d44db59882114ffe64434b655b931f3beec8e2
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/5082
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Now that we can set CC to an arbitrary compiler, fix issues that clang
finds. Luckily, there were only two trivial errors.
Change-Id: I0fd1f0f263a8ab7004f39cd36ed42d1a1cba5c04
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/5081
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
LDN is 8-bit but coreboot squeezes unrelated info: VLDN in this field.
Increase to 16-bit to handle this.
Change-Id: I97af1b32dcfaed84980fa3aa4c317dfab6fad6d8
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/5165
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Be more conservative and only add VGA devices' prefetchable
resources as write-combining in the address space. Previously
all prefetchable memory was added as a write-combining memory
type. Some hardware incorrectly advertises its BAR as
prefetchable when it shouldn't be.
A new memranges_add_resources_filter() function is added
to provide additional filtering on device and resource.
Change-Id: I3fc55b90d8c5b694c5aa9e2f34db1b4ef845ce10
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/5169
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Tested-by: build bot (Jenkins)
It was a typo.
Change-Id: I82964b5ed7e7749ba141aeb3ee8dc4c107bcd7a9
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/5127
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
The calculations for static allocation are no longer valid.
Change-Id: I6740cdcec789abddf78485a0edaf24882ef8c2a5
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4569
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Unused and hard-coded to use uart8250 on IO.
Change-Id: I3f84c50039a450a2ae97a5fd2af89992f8567e6c
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5137
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Also prepare this console for use in romstage.
Change-Id: I26a4d4b5db1e44a261396a21bb0f0574d72aa86d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5136
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Also relocate and split header files, there is some interest
for EHCI debug support without PCI.
Change-Id: Ibe91730eb72dfe0634fb38bdd184043495e2fb08
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/5129
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
If the MTRR usage exceeds the BIOS allocation for MTRR usage
re-try without the WRCOMB type.
Change-Id: Ie70ce84994428ff6700c36310264c3c44d9ed128
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/5151
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Tested-by: build bot (Jenkins)
The memranges_update_tag() function replaces all instances
that are tagged with old_tag and update to new_tag. This
can be helpful in the MTRR code by adjusting the address
space if certain memory types cause the MTRR usage to
become too large.
Change-Id: Ie5c405204de2fdd9fd1dd5d6190b223925d6d318
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/5150
Tested-by: build bot (Jenkins)
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Flash prefers 32-bit sequential access. On some platforms ROM is
not cached due to i.a. MTRR shortage. Moreover ROM caching is not
currently enabled by default. With this patch payload decompression
is sped up by theoretical factor of 4.
Test on X201, with caching disabled:
Before:
90:load payload 4,470,841 (24,505)
99:selfboot jump 6,073,812 (1,602,971)
After:
90:load payload 4,530,979 (17,728)
99:selfboot jump 5,103,408 (572,429)
Change-Id: Id17e61316dbbf73f4a837bf173f88bf26c01c62b
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/5144
Reviewed-by: Aaron Durbin <adurbin@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
And use it in fit.c and remove one more use of htonl.
Change-Id: Ibf18dcc0a7f08d75c2374115de0db7a4bf64ec1e
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/5120
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The punit is responsible for a number of things. Without
performing the sequence included it won't change processor
frequency when requested and apparently there are some bizarre
hangs introduced if this sequence isn't included either. Lastly,
this needs to come after microcode has been loaded. As that is
done in bootblock the ordering is correct.
One other side effect is that this fixes the graphics devices'
device id. Before it was showing up as the same device id of the
SoC transaction router.
BUG=chrome-os-partner:22880
BUG=chrome-os-partner:23085
BUG=chrome-os-partner:22876
BRANCH=None
TEST=Built and booted.
Change-Id: Ib7be1d4b365e9a45647c778ee5f91de497c55bf1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171862
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4864
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Start loading microcode in the bootblock. This way
no caching has been set up and cache-as-ram mode
will be running in a validated configruation (with ucode
patch).
BUG=chrome-os-partner:22858
BRANCH=None
TEST=Built and booted. Confirmed microcode is loaded.
Change-Id: I6fd1d8e55bcc9d799b11d9faed771ac50dc120a2
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171861
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4863
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The TCO timer always starts ticking out of reset.
However, depending on microcode loading and punit
initialization the TCO timing out has a different
impact on the sytem. Without loading microcode
or initializing the punit the tco times out and
nothing happens. However, when microcode is loaded
a timeout will reset the system. Lastly, if the
punit is initialized but the microcode isn't loaded
the TCO timeout will shut down the system.
To fix all the weird symptoms disable the TCO.
BUG=chrome-os-partner:22858
BRANCH=None
TEST=Built and booted with microcode loading. Reset doesn't
occur.
Change-Id: I49cd62f510726a96bf734ae728a352c671d1561e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171860
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4862
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Apparently there was another BAR living at 0x5c in the LPC
bridge that mapped the PUNIT registers. EDS 2.0 released
and this register is now documented.
BUG=chrome-os-partner:23085
BRANCH=None
TEST=Built and booted.
Change-Id: I5892c2a14923b57826060e92b4335cb1952ea057
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171612
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4861
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The 316 microcode is the newest version. Include that in the build.
BUG=chrome-os-partner:22858
BRANCH=None
TEST=Built and partially booted with microcode loading. Noted 316
loaded.
Change-Id: Iba01dd58688737ae38bc58a84014ee9526540db1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171611
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4860
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Allow for one to write an individual byte of a 32-bit register
when sending a read/write through the IOSF messaging system.
Add PUNIT registers and fields for early sequencing.
BUG=chrome-os-partner:23085
BRANCH=None
TEST=Built and partially booted with changes that use PUNIT
registers and individual byte en fields.
Change-Id: I929fb5c51d805c55c478cab884e3572254987fc7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171710
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4859
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The mrc_wrapper.h was changed to protect against ABI differences
between the two sets of compilers and flags used. This requires
a prope shim for the console output funciton.
BUG=chrome-os-partner:23048
BRANCH=None
TEST=Built and booted successfully.
Change-Id: I976e692e66dcfc0eacadae6173abfd9b81e31137
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/171580
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4858
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This commit always selects COLLECT_TIMESTAMPS and starts
tracking TSC values from the early stages of bootblock.
The initial timestamp value is saved in mm0 and mm1 while
in bootlbock. This approach works because romcc is not configured
to use mmx registers for its compilation.
Additionally, the romstage api with the mainboard was changed to
always pass around a pointer to a romstage_params structure as the
timestamps are saved in there until ram is up.
BUG=chrome-os-partner:22873
BRANCH=None
TEST=Built and booted with added code to print out timestamps at
end of ramstage. Everything looks legit.
Change-Id: Iba8d5fff1654afa6471088c46a357474ba533236
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/170950
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4856
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
When I changed mkpayload, I did not realize we had a duplicate
block of code in the linux payload code. Have it use the same
header generator as the standard payload code does.
Change-Id: Ie39540089ce89b704290c89127da4c7b051ecb0e
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/5115
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Aaron Durbin