Some new fields were added to the edid data structure, and the edid code was
changed to put estimated values into those fields which were ultimately passed
into depthcharge or other payloads. On snow we do things different and just
declare an edid structure statically which didn't have those members. The rows
and columns of the graphics console were 0, and that confused the framebuffer
driver and made it loop forever.
Change-Id: I6ca3bd948482b347a6a981e83b82b10dca995e5e
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/61057
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Reviewed-on: http://review.coreboot.org/4341
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
- Update RAM_ID table.
- Add DEVSLP0 signal to NGFF SATA port.
Note: After this change, old Micron 2GB boards will no longer boot.
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Change-Id: Id68a1d6ace2702cca9c37305726cd55a0bde5005
Reviewed-on: https://gerrit.chromium.org/gerrit/60167
Tested-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-by: Dave Parker <dparker@chromium.org>
Commit-Queue: Dave Parker <dparker@chromium.org>
Reviewed-on: http://review.coreboot.org/4340
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Change-Id: I9791beff44535a0a130292414fcd9875b497b1ca
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4492
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
No ROMCC involved, no need to include .c files in romstage.c.
Change-Id: I8a2aaf84276f2931d0a0557ba29e359fa06e2fba
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4501
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
walkcbfs() is used only with ROMCC. Besides finding stages during the
bootblock, it's also used when applying microcode updates during the
bootblock phase. The function used to return only a pointer to the data of
the CBFS file, while making the header completely inaccessible. Since the
header contains the length of the CBFS file, the caller did not have a way
to know how long the data was. Then, other conventions had to be used to
determine the EOF, which might present problems if the user replaces the
CBFS file. This is not an issue when jumping to a stage (romstage), but can
present problems when accessing a microcode file which has not been
NULL-terminated.
Refactor walkcbfs_asm to return a pointer to the CBFS file header rather
than the data. Rename walkcbfs() to walkcbfs_head(), and reimplement a new
walkcbfs() based on walkcbfs_head(). Thus current usage of walkcbfs()
remains unaffected.
The code has been verified to run successfully under qemu.
Subsequent patches will change usage of walkcbfs() to walkcbfs_head where
knowing the length of the data is needed.
Change-Id: I21cbf19e130e1480e2749754e5d5130d36036f8e
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/4504
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
During the update_fit step, 'file_length' is used to determine how many
bytes are left in the CBFS file. It was decremented in a loop from an
array 'mcus[num_mcus].size', but 'num_mcus' was incremented right before.
Since 'mcus' is memset(0) externally, 'file_length' was never decremented.
The loop exited when it reached a dummy terminator, usually 48 bytes of 0
which are internationally added to microcode blobs in coreboot. However,
if that terminator is removed, the loop doesn't stop and continues until
it segfaults.
Change-Id: I840727add69379ffef75b694d90402ed89769e3b
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/4508
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
If the unpacking was interrupt by Ctrl-C, probably part of
an archive is unpacked. If we run buildgcc again, the
incomplete folder would be and skipped.
We can create a file to tell the script the unpacking is done.
Change-Id: Id9eb74d119e22b62c70dca9b38a92c3dbdf0f64c
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4512
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The architecture information has been added to the folders to
build archives for crossgcc. We need to change the .gitignore
to keep ignoring them.
Change-Id: Ic18685e507f1b09088120eee6047d49141e29906
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4513
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
I am planning to write a document for new man, helping them to build a
working image from knowing nothing about Coreboot. The previous
LinuxBIOS-AMD64.tex was a good script.
LinuxBIOS was a out-of-date name. It needs to be changed it to Coreboot.
And the new document is not specific to AMD, so we get rid of the AMD64 from
the file name.
This is a separate patch for the renaming.
LinuxBIOS-AMD64.tex was still about SVN and buildtarget. Later patch
will come up for updating those old thing. And this new document,
CorebootBuildingGuide.tex, is planning to be only about building.
Navigating the source code is moved to later advanced document.
Change-Id: Ia8a2fd9db51e9870c1d645067bcfdc91ae1bf90a
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4487
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Make boards take less vertical space, and link to board pages
Change-Id: Ifdd062a15191809b75422416c874161d9114363d
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4493
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Instead of having global variables put them on the stack.
Change-Id: I462e3b245612ff2dfb077da1cbcc5ac88f8b8e48
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/4288
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
It was suggested to eliminate the lock for sprintf. One way to do it is
to make the fake tx_byte into a closure. This patch allows it.
It's a bit tricky since we need to preserve compatibility with romcc.
Change-Id: I877ef0cef54dcbb0589fe858c485f76f3dd27ece
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/4287
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
These were terribly under-documented
Change-Id: I285ea083110d87076281e81065f5f38d0c688358
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4491
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Newer mainboards that use haswell -- and, presumably, chipsets to come -- need
some support functions. Add them in the drivers/intel/gma directory.
Currently, this is one file: intel_ddi.c, but more may come.
Compilation of this file is controlled by INTEL_DDI, defined
in the Kconfig as default n and used in the Makefile.inc
Change-Id: I501ee291c0d4589925ed3e478f67106337fcad31
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/60612
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-on: http://review.coreboot.org/4337
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
We use junit style output these days.
Change-Id: I4110ec10bf0e9f4354ee08e7e1c5a81ae605fee0
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4484
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Add ACPI Methods to enable and disable power limiting with PL1.
This can be used in ACPI Thermal Zone or in EC ACPI _QXX events.
This commit adds new unused methods and is fully tested with the
subsequent commit that makes use of these methods.
Change-Id: I9d8d23bfe9cf7c756ff8ab0412e5a010826b12db
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/60546
Reviewed-on: http://review.coreboot.org/4334
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
1) fix enable of power aware interrupt routing
2) set BIOS_RESET_CPL to 3 instead of 1
3) mirror PKG power limit values from MSR to MMIO on all SKUs
4) mirror DDR power limit values from MMIO to MSR
5) remove DMI settings that were from snb/ivb as they do
not apply to haswell
1) verify power aware interrupt routing is working by looking
in /proc/interrupts to see interrupts routed to both cores
instead of always to core0
BEFORE: 58: 4943 0 PCI-MSI-edge ahci
AFTER: 58: 4766 334 PCI-MSI-edge ahci
2) read back BIOS_RESET_CPL to verify it is == 3
localhost ~ # iotools mmio_read32 0xfed15da8
0x00000003
3) read PKG power limit from MMIO and verify it is the same
as the MSR value
localhost ~ # rdmsr 0 0x610
0x0000809600dc8078
localhost ~ # iotools mmio_read32 0xfed159a0
0x00dc8078
localhost ~ # iotools mmio_read32 0xfed159a4
0x00008096
4) read DDR power limit from MSR and verify it is the same
as the MMIO value (note this is zero based on current MRC input)
localhost ~ # rdmsr 0 0x618
0x0000000000000000
localhost ~ # iotools mmio_read32 0xfed158e0
0x00000000
localhost ~ # iotools mmio_read32 0xfed158e4
0x00000000
Change-Id: I6cc4c5b2a81304e9deaad8cffcaf604ebad60b29
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/60544
Reviewed-on: http://review.coreboot.org/4333
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Limit power to 12W at 73C and remove limit at 68C.
To have the CPU consume maximum power it is necessary to stress
both the CPU and the GPU. Bastion (chrome.supergiantgames.com)
and/or webglsamples.googlecode.com can be useful for this.
Testing this properly requires a script to report the running
average power readings. The watch_power.sh script is attached
to this issue in the partner tracker.
1) Run watch_power.sh continuously:
localhost ~ # watch -n 0 bash -e /tmp/watch_power.sh
2) Start Bastion (or other stress apps). The power draw should
be close to 15W if under enough load.
3) Watch until temperature climbs above 73C and is caught by
the thermal zone 10 second poll, this can be sped up by blocking
or removing the fan.
4) The ACPI thermal zone states should change to reflect that
active[2] is now enabled and power consumption should drop to 12W.
5) Stop the stress apps and wait until the CPU cools off again,
enable the fan again if it was removed.
6) The ACPI thermal zone state should switch back to active[3].
Change-Id: Ie6714a8543d4f06edf8513086fc9c968273bdb23
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/60545
Reviewed-on: http://review.coreboot.org/4335
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The elog code calculates flash offsets and their equivalent
addresses in the memory address space. However, it assumes
the detected flash size is entirely mapped into the address
space. This can lead to incorrect calculations. Add code
to allow ROM_SIZE to be less than detected flash size. The
underlying assumption is that the first ROM_SIZE bytes are
programmed into the larger device.
Change-Id: Id848f136515289b40594b7d3762e26e3e55da62f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/60501
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4332
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The original intention was to only run UPDATE_FIT when a microcode file was
included in CBFS. This happens when either CPU_MICROCODE_CBFS_GENERATE or
CPU_MICROCODE_CBFS_EXTERNAL is selected, however, the makefile checked that
CPU_MICROCODE_IN_CBFS was selected instead. The end result was that on
hasswell, the UPDATE-FIT step was always run, even when no microcode was
included, generating a build error.
Instead, introduce a new variable which tells if a microcode update is
added in CBFS during the build.
Change-Id: I28638912ed6f77761ef8a584f7636dc907b7a9b7
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/4480
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
No need to show the choice of USB port or controller in case of older
hardware where location for usbdebug was hardwired.
Change-Id: Ia186bf2c6ed60be2834cf6fd0a1965c8bf81ed4d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4290
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Use a file in CBFS for keyboard layout and ethernet MAC instead
of scanning FMAP.
Change-Id: I7658c7c4e389deb20d7d8f57cce8b568efdc575d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/4307
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Variables in coreboot and not in line with GRUB ones. E.g. HOSTCC is both
HOST_CC and BUILD_CC for GRUB (consult INSTALL for more details) and
what coreboot calls CC is TARGET_CC for GRUB.
Current code plugs this by defining variables explicitly but it has a nasty
effect that make stops caring about flags added in makefile itself. Undef
as many variables as possible but still pass them to configure for them to
have correct effect and keep CC assignment as my make version doesn't undefine
it even when instructed to do so.
Tested with qemu.
Change-Id: I9d18f557138a20ae3918d698dee8f5b5c5738f75
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/4310
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The Intel GMA driver is in, this CL splices in the Makefile bits.
Change-Id: Icf42a537575b8cc90a679ec1fc15b09294630611
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/60346
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-on: http://review.coreboot.org/4331
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
USE_XARGS mode builds n boards in parallel (with 1 CPU each) instead of
building 1 board with n CPUs.
This requires the main build system to work under such circumstances.
Change-Id: Ib4571a78dfe78fd61ae5b26c18be9745bd8b3d52
Reviewed-on: http://review.coreboot.org/4485
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This makes USE_XARGS-abuild unhappy due to races
Change-Id: I1237468366c7f8af7eacd572c2bd32df9a3d58ca
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4486
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
These functions are not all used yet, but do compile and are partially used
in the FUI testing.
They were extracted from the 3.4 kernel using coccinnelle filters. The .c files
are only compiled in if CONFIG_INTEL_DP is set.
Change-Id: Id95622a75aa02b496c9ea4717cb143394a8332e3
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/60245
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4329
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Removed two unnecessary register sets, and did the power well a bit
more correctly. Also, added a register definition include file so we can
used constants instead of magic numbers.
We also set registers to common initialized values that are
needed for FUI, VBIOS, and kernel. This set of registers
appears to be an absolute bare minimum. Since we're hoping to use
FUI for all chipsets from this one forward, we unconditionally do the
setting here.
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Change-Id: Ife3f661ba010214d92b646b336f2b06645119f17
Reviewed-on: https://gerrit.chromium.org/gerrit/59988
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-on: http://review.coreboot.org/4328
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The new edid functions support converting the edid to an lb_framebuffer.
Use them. Also, since panels seem to set bits per color instead of bits
per pixel, just force the right value in the edid struct.
Add helpful comment because people don't always believe we need to set
the pallette.
While we're at it, fix a problem that caused it to not compile.
Change-Id: I645edc4e442d9b96303d9e17f175458dc7ef28b6
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/57619
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-on: http://review.coreboot.org/4327
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
- updates from 1.6.0 ref code
- remove the step comments as they are no longer even close
- add constants for LPT revisions
build and boot on Falco
Check that RCBA+2300[1] is set:
> mmio_read32 0xfed1e300
0x00000002
Change-Id: I8b3c5fda3f3170455699a7834239cb991603e7a8
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/59821
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4326
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
There's a need to determine if a specific gpio pin is
is set up to be a native function or not. Implement this.
Change-Id: I91d57a549e0f4fddc0b1849e5f74320fc839642c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/59589
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4324
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The BIOS spec for LynxPoint calls out additional
programming steps for the PCIe Root Ports. Implement those
steps from the BIOS spec. These steps are completed before
deeper PCIe probing. The "late" programming was removed as
that was applicable to Cougar/Panther point where this
code was originally copied, though there was some overlap.
Change-Id: I64f25e4451e035d98ca6b66b0335bd280b70b074
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/59558
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4323
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
PCIe Root Ports should be disabled based on pin ownership
and the strapping configuration. Implement this logic
for LynxPoint. The chip_ops->enable_dev() path is no
longer used. Instead the PCIe driver handles the enabling
and disabling of devices. This allows for having an empty
or incomplete device tree since those "allocated" devices
do not travel through the chip_ops->enable_dev() path.
The coalescing was tested to be working properly, however
not all configurations were tested.
Change-Id: I1e8bfe5e447b72ff8a4b04b650982d8c1ae0823c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/59424
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4322
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Don't force dev mode. Allow users to enter / exit dev mode as normal.
Change-Id: I168eb04a8ac102a8c4a1ca8936f78f62b001e0eb
Reviewed-on: https://gerrit.chromium.org/gerrit/59492
Commit-Queue: Shawn Nematbakhsh <shawnn@chromium.org>
Tested-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-by: Dave Parker <dparker@chromium.org>
Reviewed-on: http://review.coreboot.org/4321
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
So far this is used by the USB driver, and instead of
having ifdefs all throughout that code, implement the same
API on x86 and ARM.
Change-Id: I8093ad818ad2e38a0901787aa8674faf591d580c
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/56105
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-on: http://review.coreboot.org/4320
Tested-by: build bot (Jenkins)
On some systems there may be 2GB SKU that is the same as the
4GB SKU but just one channel of memory. In that case we need
to ensure that both copies of the same SPD source end up
populated by ensuring that repeated entries are included by
using $+ instead of $^.
Alternatively we could do the check inside romstage, but it
is already set to behave this way if the SPD gets populated
correctly.
I changed spd_index to 3 in falco romstage to force it to
pretend it was a 2GB config of the same memory, then booted
to ensure it was indeed limited to 2GB.
memcfg channel[0] config (00780008):
ECC inactive
enhanced interleave mode on
rank interleave on
DIMMA 2048 MB width x16 single rank, selected
DIMMB 0 MB width x16 single rank
memcfg channel[1] config (00600000):
ECC inactive
enhanced interleave mode on
rank interleave on
DIMMA 0 MB width x8 single rank, selected
DIMMB 0 MB width x8 single rank
Change-Id: Ibfe5051ccda2fe69e8caff3f3c264116e3411c65
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/59483
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: Jay Kim <yongjaek@chromium.org>
Reviewed-on: http://review.coreboot.org/4319
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The numbers alone are hard to parse, so add
some timestamp names to make it easier to read.
Change-Id: Ie32d3e7ca759bd15e7c160bdd829dec19943e6cb
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/65333
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Reviewed-on: http://review.coreboot.org/4314
Tested-by: build bot (Jenkins)
On ARM the timestamps are already in micro seconds, so
no need to convert them.
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: If7363b0703e144bde62d9dab4ba845e1ace5bd18
Reviewed-on: https://gerrit.chromium.org/gerrit/63991
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-on: http://review.coreboot.org/4313
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
It's a start...
Change-Id: Ibdb0b64ab0349df58bcad5ce553bf0dbec636925
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/4483
Tested-by: build bot (Jenkins)
- Add -ffreestanding and -fomit-frame-pointer for all
platforms.
- Add ARMv7 specific flags to the armv7 Makefile
Change-Id: I71ab1b096e505940cc20c266bccd43917bcfad3a
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/56104
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-on: http://review.coreboot.org/4317
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Propagated from
http://review.coreboot.org/3347http://review.coreboot.org/3374
The cause of this issue is:
USB devices use bit 11(0x0b) of GP0_STS represents S3 wake up event,
but this bit is not clear after wake up. So OS thinks there is a
wake up signal and wake up immediately.
Both amd/olivehill and asrock/imb-a180 have been validated.
Change-Id: I7c26cb07bcd2e62bb792809b67314e5155c6adf6
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4261
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The AML code of PTS and WAK for southbridge are in
UINT8 AlibSsdtKB[], Proc/GNB/Modules/GnbInitKB/AlibSsdtKB.h.
It was integrated into SSDT even it was called by nobody.
The source ASL was provided by AGESA for reference, but it
has been scrubbed when it was ported to Coreboot.
Without the calls, Olive Hill can not wake up if it boots Windows.
Both amd/olivehill and asrock/imb-a180 have been validated.
Change-Id: Ia7bba29904dbd6f33fdb08bf88bb499005ef561b
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4260
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The bug is hard to find. We were adding the feature of fan control. We
met some strange things which could not be explained. Like, sometimes
adding printk let the error disappear. Then we traced the code by hardware
debug tool (HDT). It turned out the data in stack was overwritten.
The values of AccessWidthxx are
{ AccessWidth8 = 1,
AccessWidth16,
AccessWidth32,}
For the case of AccessWidth8, we only need to access the index/data
once. But ReadECmsg and WriteECmsg did the loop twice, 1 more time
than they are supposed to do. The data in stack next to "Value" would
be overwritten.
For all the cases, the code should be
OpFlag = OpFlag & 0x7f;
switch (OpFlag) {
case 1: /* AccessWidth8 */
OpFlag = 0;break;
case 2: /* AccessWidth16 */
OpFlag = 1;break;
case 3: /* AccessWidth32 */
OpFlag = 3;break;
case 4: /* AccessWidth64 */
OpFlag = 7;break;
default:
error;
}
Actually, the caller only takes AccessWidth8 as the parameter. We can ignore other
cases for now.
That is an AGESA bug. AMD's AGESA team own this code. They have given the
response that they are going to update this in next release. I presume let them
decide the proper way to fix that. Before that, I change the code as little
as possible to make it run without crash.
Change-Id: I566f74c242ce93f4569eedf69ca07d2fb7fb368d
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/4297
Reviewed-by: Bruce Griffith <Bruce.Griffith@se-eng.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Gabe Black