The ssdt2 generation code was calling acpigen_patch_len().
However, none of the entries had AML object lengths that
needed patching. That resulted in the following message:
ASSERTION FAILED: file 'src/arch/x86/boot/acpigen.c', line 52
Additionally, this caused an errant write to a memory address
whose value was in the variable ltop. This was the 0 address.
Change-Id: I44abf5a4e4225220575aee6b5c9bb6b0be093a28
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/56299
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4182
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The ACPI code was defining two EHCI controllers and ignoring
the XHCI controller. This changes the second EHCI controller
to be XHCI instead and changes the wake resource to indicate
S3 and not S4.
cat /proc/acpi/wakeup
Device S-state Status Sysfs node
HDEF S4 *disabled pci:0000:00:1b.0
EHCI S3 *enabled pci:0000:00:1d.0
XHCI S3 *enabled pci:0000:00:14.0
Change-Id: If28775e6ef8608c22c85ca91d91d1f598ec7755d
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/56263
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4181
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The architecture name for our ARM port is armv7, not arm.
Hence, none of those flags were ever actually used.
Fix the architecture name and remove the flags, they should
not be set in xcompile, but in the Makefile, like in coreboot.
Change-Id: Id9c5db7ebceafddb58a1ce1988417f09c074ba6c
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/56084
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4179
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Enable GPIO SMI for GPIO34 and set it as inverted so it
is only generated when it is raised by the EC.
1) ec console command: lidopen
2) wait until booted to developer screen
3) ec console command: lidclose
4) ensure system turns off
Change-Id: I7d50f171f3f4539c7c264103d1ffc7c5d0f1c7ba
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/56052
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4177
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Two structures in the USB EHCI stack were pointing
to hardware but not marked attribute((packed)) hence
leaving it to GCC to correctly align the data structures.
Next, the number of reserved bytes in hc_op_t was wrong
(but implicitly aligned to the correct values on x86)
It seems this worked fine on x86, but on ARM it was doing
the wrong thing.
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: I94bed4850ded7d3f7bbc7ff3079c103c6054c22d
Reviewed-on: https://gerrit.chromium.org/gerrit/55555
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4174
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
In order to report whether coreboot enabled a SerialIO device
in ACPI mode we had been relying on reading NVS in the _STA
method for the SerialIO device.
The ACPI _STA method has restrictions on what it can access
and is unable to access OperationRegions outside its scope
which means it should not be trying to read NVS.
This change adds a new SSDT to the ACPI tables and fills it
with constants that indicate whether or not a device is enabled
in ACPI mode.
The ACPI code is changed to read these variables from the
SSDT and use that instead of trying to query a variable in NVS.
Attempt to use lpt-clk driver to probe the
device clocks for SerialIO devices and see that the kernel
does not complain about accessing the GNVS region.
Change-Id: I8538bee4390daed4ecca679496ab0cb313f174ce
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/51369
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4170
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
- Disable EC software sync for now
- Report correct EC active firmware mode
- Force enable developer mode by default
- Set up PCH generic decode regions in romstage
- Pass the oprom_is_loaded flag into vboot handoff data
Change-Id: Ib7ab35e6897c19455cbeecba88160ae830ea7984
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/51155
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4169
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The update-fit command takes in a parameter for number of slots
in the FIT table. It then processes the microcobe blob in cbfs
adding those entries to the FIT table. However, the tracking of
the number of mircocode updates was incremented before validating
the update. Therefore, move the sanity checking before an increment
of the number of updates.
Change-Id: Ie8290f53316b251e500b88829fdcf9b5735c1b0e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/50319
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4161
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
In order for the FIT entries to be populated in the table the
update-fit command needs to be done on the coreboot image. That
way the microcode entries are added to the table properly.
Change-Id: I44595aee1ca710f4f04d482d8900cf95fbc1797f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/50317
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4159
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
In order to probe the gpio-lynxpoint kernel driver the
LP GPIO controller needs to be exposed as a specific
ACPI device.
This also allows the resources to be exposed to the OS via
this device instead of the catch-all LPC device.
Ensure the driver loads at boot:
gpiochip_find_base: found new base at 162
gpiochip_add: registered GPIOs 162 to 255 on device: INT33C7:00
Also ensure the driver is visible in sysfs:
$ cat /sys/devices/platform/INT33C7:00/gpio/gpiochip162/label
INT33C7:00
Change-Id: I9f79c008f88da9b67ed1cdfdb9d3a581ce8f05ff
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/50215
Reviewed-on: http://review.coreboot.org/4158
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
In the process of getting rid of compiler includes during in coreboot
and libpayload, we defined size_t and ssize_t ourselves, using a GCC
macro for size_t: __SIZE_TYPE__. Unfortunately, there is no
__SSIZE_TYPE__, so we temporarily redefine unsigned to signed to make
__SIZE_TYPE__ __SSIZE_TYPE__.
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: I4cf4eb0fdaa4db64277c2585fe2c1bdc0acdf02b
Reviewed-on: https://gerrit.chromium.org/gerrit/49947
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4156
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Now that we have RW ramstage we don't need to have the
management engine lock down step done in a final SMM.
ME: mkhi_end_of_post
ME: END OF POST message successful (0)
PCI: 00:16.0: Disabling device
Change-Id: I9db4e72e38be58cc875c1622a966d8fcacc83280
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/49757
Reviewed-on: http://review.coreboot.org/4153
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
There were two undefined MBP types that are now defined.
These include NFC status and some interesting timing data.
ME: Wake Event to ME Reset: 6 ms
ME: ME Reset to Platform Reset: 7 ms
ME: Platform Reset to CPU Reset: 51 ms
Change-Id: I67bf1f303f3c32497041e64c40eb9ccb6a63d88a
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/49756
Reviewed-on: http://review.coreboot.org/4152
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Instead of having an OS re-parse cbmem book-keeping records
for the cbmem allocator just to get the console buffer export
the pointer to the memory console directly in a field named 'CBMC'.
This field lives in the GNVS table.
Change-Id: Ief0c4da7b18df66feb9c816c9f4abdf5a72bd3a4
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/49764
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4149
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
If elog_clear is called before other elog functions, for instance if it's
called through an SMI immediately after the system boots, then the elog data
structures won't have been set up and the system will go off the deep end.
This change adds a call to elog_init to elog_clear to make sure things things
are always initialized before we start using them.
Before this change, this command would cause
the system to lock up if run immediately after boot:
echo 1 > /sys/firmware/gsmi/clear_eventlog
After this change, that results in the log being cleared correctly.
Change-Id: I45027f0dbfa40ca8c581954a93b14b4fedce91ed
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/49303
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Reviewed-on: http://review.coreboot.org/4144
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Slight tweaks found when looking at latest ref code when
investigating package C-state issues.
A few bits in the clock gating register don't match the
documentation and are also cleaned up.
Change-Id: I36ced7280c160b114c70b2eeafc8b24813ff2f6a
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/49330
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4142
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The mainboard_interrupt_handlers() argument for the function
pointer was using void * as the type. This does not allow the compiler
to catch type differences for the arguments. Thus, some code has been
committed which violates the new interrupt callbacks not taking any
arguments. Make sure the compiler provides a type checking benefit.
Change-Id: Ie20699a368e70c33a9a9912e0fcd63f1e6bb4f18
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/48970
Reviewed-on: http://review.coreboot.org/4141
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Commit b8ad224 changed the memory address in lb_cbmem_ref coreboot
table entries from a pointer to a uint64_t. This change was introduced
to make the cbmem utility work on both 32bit and 64bit userland.
Unfortunately, this broke the cbmem utility running on older versions
of coreboot because they were still providing a 32bit only field for
the address while the cbmem utility would now take the following 4
bytes as upper 32bits of a pointer that can obviously not be
mmapped. This change checks if the size of the lb_cbmem_ref structure
provided by coreboot is smaller than expected, and if so, ignore the
upper 32bit of the address read.
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: If4c8e9b72b2a38c961c11d7071b728e61e5f1d18
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4139
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The format of this function changed but was not updated in
all mainboards. This fixes all Sandybridge/Ivybridge boards.
The int15 handler no longer takes a regs structure as an
argument and instead uses global variables. The yabel interface
is now similar enough that we can drop the duplicate handler.
Change-Id: Icdaae4d6d50884f6d7bce7a167d48cb1d4807010
Reviewed-on: https://gerrit.chromium.org/gerrit/48969
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4135
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Recent commit proposal by Ron Minnich proposes to move to native gfx init for
qemu. Unfortunately we didn't have native init for default qemu video (cirrus)
Here is one extracted from GRUB one which I wrote couple of years ago.
Change-Id: Icb89cf918ef5d276bcc703c48c568e7b9c1be756
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/4270
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Per our discussions with Gerd, qemu will now always do native graphics
on coreboot. The VGA BIOS capability is not needed and will no longer
be supported. Attempts to build without native graphics will result in
an error.
This code builds for both x86 emulation targets. I'm hitting an issue
testing that is unrelated to coreboot; if someone can test, that
would be helpful. Be sure to start qemu with -vga std.
We also add a test for the PCI BAR being zero and return silently if it
is.
Change-Id: I66188f61e1bac7ad93c989cc10f3e0b55140e148
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: http://review.coreboot.org/4258
Tested-by: build bot (Jenkins)
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
SandyBridge raminit uses this CMOS option. If it is not declared, the build
fails when USE_OPTION_TABLE is selected.
Change-Id: I1ba1f994d4ea3824dc66e8f35d0b5b24b88d4dd6
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/4269
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
the part !CAR && PRE_RAM is obviously meant as dummies. Unfortunately
cbmemc_tx_byte has wrong number of arguments and hence causes compilation
failure.
Found out when compiling for vexpress-a9.
Change-Id: Ic84d142bac5c455c2371fbc9439c898de04a974e
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: http://review.coreboot.org/4267
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
The chromeos_acpi driver sysfs naming is not what
crossystem expects if there is just one entry in the package
because it does not add a ".#" suffix in that case.
Specify all the expected GPIOs on wtm2 as undefined, which
should be 0xFF and not 0x00 becuase 0 is a valid GPIO.
Change-Id: I9b17e9bab94219695e65b17914c84acf02a0983b
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/50337
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4162
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
On haswell ULT systems there is a 24MHz clock that continuously runs
when deep package c-states are entered. The 100MHz BCLK is shut down
in the lower c-states. When the package wakes back up a conversion
formula needs to be applied. The 24MHz calibration is done using the
internal PCODE unit.
Change-Id: I6be7702fb1de1429273724536f5af9125b98da64
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/48292
Tested-by: Stefan Reinauer <reinauer@google.com>
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4136
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The c-states are configured according to the BWG, however the
package c-states are disabled as they currently cause platform
instability. The exposed ACPI c-state to processor c-state mapping
are as follows for ULT boards:
ACPI(C1) = MWAIT(C1E)
ACPI(C2) = MWAIT(C7S long latency)
ACPI(C3) = MWAIT(C10)
The non-ULT boards have an expoed c-state mapping:
ACPI(C1) = MWAIT(C1E)
ACPI(C2) = MWAIT(C3)
ACPI(C3) = MWAIT(C7S)
Included in this patch is removing the updating of current limit
registers as some of the MSRs are different and the proper values
are currently unknown. Lastly, some of the MSRs were renamed to
match the BWG.
Booted 3.8 kernel and used powertop to note package, core, and acpi
c-state residency.
Change-Id: Ia428d4a4979ba3cba44eb9faa96f74b7d3f22dfe
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/48291
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4133
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
