Skylake/Kabylake SOC has two possible ways to make serial
console functional.
1. Legacy IO based access using Port 0x3F8.
2. LPSS UART PCI based access.
This patch to provide option to select index for LPSS
UART port: 0 = LPSS UART0, 1 = LPSS UART1, 2 = LPSS UART2
PCI based LPSS UART2 is by default enabled for Chrome Design.
Change-Id: I9647820fe59b5d1a1001a611b9ae3580946da0ae
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20996
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Instead of enabling power button SMI unconditionally, add a boot state
handler to enable power button SMI just before jumping to
payload. This ensures that:
1. We do not respond to power button SMI until we know that coreboot
is done.
2. On resume, there is no need to enable power button SMI. This avoids
any power button presses during resume path from triggering a
shutdown.
BUG=b:64811381
Change-Id: Icc52dc0103555602c23e09660bc38bb4bfddbc11
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/21082
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Split `i2c.h` into three pieces to ease reuse of the generic defi-
nitions. No code is changed.
* `i2c.h` - keeps the generic definitions
* `i2c_simple.h` - holds the current, limited to one controller driver
per board, devicetree independent I2C interface
* `i2c_bus.h` - will become the devicetree compatible interface for
native I2C (e.g. non-SMBus) controllers
Change-Id: I382d45c70f9314588663e1284f264f877469c74d
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20845
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
We do not need or use the Management Engine MBP HOB so that
step can be skipped when FSP initializes the ME.
BUG=b:64479422
TEST=boot with FSP debug enabled binary and ensure that the
output indicates this step is being skipped:
Skipping MBP data due to SkipMbpHob set!
Change-Id: I5ea22ec4b8b47fa17b1cf2bf562337bfaad5ec0d
Signed-off-by: Duncan Laurie <dlaurie@google.com>
Reviewed-on: https://review.coreboot.org/20951
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Need to perform a dummy read in order to activate LPSS UART's
16550 8-bit compatibility mode.
TEST=Able to get serial log in both 32 bit and 8 bit mode through
LPSS UART2 based on CONFIG_DRIVERS_UART_8250MEM_32 and
CONFIG_DRIVERS_UART_8250MEM selection.
Change-Id: I5f23fef4522743efd49167afb04d56032e16e417
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20939
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
1. Add a new variable to GNVS to store information during S3 suspend
whether UART debug port controller is enabled.
2. On resume, read stored GNVS variable to decide if UART debug port
controller needs to be initialized.
3. Provide helpers functions required by intel/common UART driver for
enabling controller on S3 resume.
BUG=b:64030366
TEST=Verified behavior with different combinations:
1. Serial console enabled in coreboot: No change in behavior.
2. Serial console enabled only in kernel: coreboot initializes debug
controller on S3 resume.
3. Serial console not enabled in coreboot and kernel: coreboot skips
initialization of debug controller on S3 resume.
Change-Id: Iad1cc974bc396ecd55b05ebb6591eec6cedfa16c
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/20886
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
1. Create a new function uart_lpss_init which takes the UART LPSS
controller out of reset and initializes and enables clock.
2. Instead of passing in m/n clock divider values as parameters to
uart_common_init, introduce Kconfig variables so that uart_lpss_init
can use the values directly without having to query the SoC.
BUG=b:64030366
TEST=Verified that UART still works on APL and KBL boards.
Change-Id: I74d01b0037d8c38fe6480c38ff2283d76097282a
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/20884
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
In _FSL ACPI notification 0x83 was incorrectly being sent to DPTF.
When there should be no notification on fan speed change.
Change-Id: I66efa7a7feb911a458829a54dbd0afefabd42394
Signed-off-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Reviewed-on: https://review.coreboot.org/20875
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Internal PME is detected when bit PME_B0_STS is set. Following devices
causes internal PME.
- Integrated LAN
- HD Audio/Audio DSP
- SATA
- XHCI ('USB3')
- ME Maskable Host Wake
In SPT, PCIEXPWAK_STS bit isn't getting set due to known bug.
So scan all PCIe RP for PME status bit & update event log accordingly.
BUG=b:36992859
TEST=Build for Soraka, Verify resume due to PME on root port is logged
in elog.
Change-Id: I879a7c332e62ab598942b29d31bad84619b35ea7
Signed-off-by: Naresh G Solanki <naresh.solanki@intel.com>
Reviewed-on: https://review.coreboot.org/20532
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
If IGD PCI device is disabled:
1. BAR for the device will be 0.
2. There is no need to allocate framebuffer for this device.
Some early SOCs don't have GFX model fuse by default hence
we need to add a check to ensure PCI device is enable. This
code to avoid die inside coreboot for missing resources.
Change-Id: Ic31d3e57ba730f6b569bf2cc3bdc54cb369b8caf
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20829
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
As per discussion with CSME team, ME is NOT using PCI Config
Space register HFSTS2 Bit 10 to update ME power-gated status.
ME goes to CM0-PG state after ME device becomes idle after
Bit 2 of MMIO register offset 0x800 (D0i3 Control - HECI1_D0I3C)
is being set.
And to retrieve the PG status of ME, one should read from the
PWRMBASE+offset 0x590 (which should give the value 0xF9) and
PWRMBASE+offset 0x594 (which should give the value 0xFF).
But, also it needs some time for the ME FW to go to idle state
and reflect these values in PWRMBASE registers after D0i3 bit
is being set. This does not happen instantly.
So, in coreboot, if we read the ME PG state in finalize.c, which
happens just after FSP Notify phase, where actually ME D0i3 bit
is set, we do not read the correct PG state values (i.e, 0xF9
and 0xFF).
But, once it boots to Kernel, if we read those same registers
through iotool mmio_read32 command, we get correct values.
So, removing the ME PG state prints from coreboot, since it is
actually showing wrong information, although ME Power Gating is
successful.
Change-Id: Idd31a9803b4c9db7d4bb8bbec5374583a8df0c41
Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com>
Reviewed-on: https://review.coreboot.org/20172
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
init_igd_opregion itself is supposed to return cb_err so this adds
error handling, just like other implentations of write_acpi_tables do it.
this had been found by coverity:
*** CID 1378270: Error handling issues (CHECKED_RETURN)
/src/soc/intel/skylake/igd.c: 147 in write_acpi_igd_opregion()
141 /* If IGD is disabled, exit here */
142 if (pci_read_config16(device, PCI_VENDOR_ID) == 0xFFFF)
143 return current;
144
145 printk(BIOS_DEBUG, "ACPI: * IGD OpRegion\n");
146 opregion = (igd_opregion_t *)current;
CID 1378270: Error handling issues (CHECKED_RETURN)
Calling "init_igd_opregion" without checking return value
(as is done elsewhere 5 out of 6 times).
147 init_igd_opregion(opregion);
148 update_igd_opregion(opregion);
149 current += sizeof(igd_opregion_t);
150 current = acpi_align_current(current);
TEST=Built
Change-Id: If6f5d53037f093607d89cfe8faf193d55de7f6c4
Found-by: Coverity (CID 1378270: Error handling issues (CHECKED_RETURN))
Signed-off-by: Martin Kepplinger <martink@posteo.de>
Reviewed-on: https://review.coreboot.org/20766
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
The values were verified to be correct with the KabylakeFsp0001 from
github. Skylake FSP documentation disagrees so YMMV.
Change-Id: I1ee04dbbed48d5376dbc24ae70753b059f2646eb
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20765
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
coreboot was setting SPI FPR register to protect the
mrc_cache data range stored in flash. This programming was being done
after FSP Notify 1.
But, FSP was locking the SPI by setting FLOCKDN Bit during Notify
phase 1, due to which coreboot was unable to protect that range.
As solution, FSP introduced a new UPD SpiFlashCfgLockDown to skip
the lockdown of flash on interest of bootloader. Set that UPD to 0
to skip the lockdown of FAST_SPI flash from FSP.
The same is being done from coreboot after end of Post at finalize.c
file.
BUG=b:63049493
BRANCH=none
TEST=FPR can be set properly to protect the mrc_cache range. The
issue reported in the bug doesn't come when both software and
hardware WP is enabled with this patch.
Change-Id: I3ffca2f1b05ab2e4ef631275ef7c3a6e23e393aa
Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com>
Reviewed-on: https://review.coreboot.org/20645
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Enable SOC_INTEL_COMMON_GFX_OPREGION for all FSP versions.
Allows to get rid of opregion.c, as it's no longer needed.
Change-Id: I39190488e12917a09dbf7ee3947a33940ebc290b
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/20222
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
This patch is to provide an additional read LPC pci offset register
BIOS_CONTROL (BC) - offset 0xDC to ensure that the last write is
successful.
Change-Id: I308c0622d348fc96c410a04ab4081bb6af98e874
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20678
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
`CONFIG_PRE_GRAPHICS_DELAY` was only applied on a dead code path in
`igd.c` that is guarded by always selected `CONFIG_ADD_VBT_DATA_FILE`.
Nobody missed it for nearly a year, plus, it's not applied on the GOP
path, let's drop it.
Change-Id: I0b70cce3a3f2b50cb4e72c4d927b35510ff362a2
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20111
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-by: Sumeet R Pawnikar <sumeet.r.pawnikar@intel.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This quirk was superseded a view lines above. Also the whole path is
guarded by `CONFIG_ADD_VBT_DATA_FILE` which is always selected for
nearly a year now.
Change-Id: I7fc5184d6e81e4588616e0302dee410e74bdab5a
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20110
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Sumeet R Pawnikar <sumeet.r.pawnikar@intel.com>
It looks like this code was written with completely different semantics
in mind. Controllers, channels and DIMMs are all presented in their phy-
sical order (i.e. gaps are not closed). So we have to look at the whole
structure and not only the first n respective entries.
Change-Id: I8a9039f73f1befdd09c1fc8e17cd3f6e08e0cd47
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20650
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
We've just decided to remove the only known use of the VBSD_SW_WP flag
in vboot (https://chromium-review.googlesource.com/c/575389), since it
was unused and never reliable on all platforms anyway. Therefore, we can
now also remove the coreboot infrastructure that supported it. It
doesn't really hurt anyone, but removing it saves a small bit of effort
for future platforms.
Change-Id: I6706eba2761a73482e03f3bf46343cf1d84f154b
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/20628
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
HECI2 and HECI3 devices are “function disable” during FSP
Silicon Init phase. Device will not be visible over PCI bus
hence removing these devices from wake source list.
Change-Id: I0de665e039d74e49e5a22db9714bc9fee734e681
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20613
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
It should never be globally exposed. Remove the global symbol
and make it static.
Change-Id: I3b85f3bbf6a73d480cdefdcdec26e137e3a3f75f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/20584
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
The FSP 2.0 path uses postcar to decompress ramstage. Since postcar
is entirely RAM based there's no need to have an excessively large
stack for the lzma decompression buffer. Therefore, reduce the stack
required to 1 KiB like apollolake.
Change-Id: I45e5c283f8ae87e701c94d6a123463dddde3f221
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/20536
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This reverts commit 5535cead (intel/skylake: Disable SaGv in
recovery mode).
Commit 5535cead disables SaGv in recovery mode to save few seconds
booting time as we were doing memory training on every recovery flow.
Now we don't need to perform MRC training on every recovery boot
due to RECOVERY_MRC_CACHE implementation in place. Hence we don't
need to define different SaGv policy between Normal (developer) mode
and recovery mode to save few seconds.
Using different SaGv parameters between recovery and all other mode
has some significent drawbacks over warm reboot cycle. We are seeing
a MRC traning hang in eve/soraka/poppy devices with below use case.
Step 1: Boot system in developer mode (first time RW_MRC training)
Step 2: Set recovery_request=1 (using crossystem) and issue “reboot”
from OS
Step 3: System will perform recovery mode MRC training and boot to
OS (first time RECOVERY_MRC training)
Step 4: Issue “reboot” from OS console.
Step 5: System wil boot in developer mode (using RW_MRC cache)
Step 6: Set recovery_request=1 (using crossystem) and issue “reboot”
from OS
Step 7: System will pick RECOVERY_MRC_CACHE and will hang during
MRC training.
This patch fixes issue mentioned above and ensures system boot to
OS without any hang if we change mode (dev<->recovery) over warm
reset.
BUG=b:63515071
BRANCH=none
TEST=manual stress testing of dev<->recovery mode over warm boot.
No MRC hang with this fix on eve/soraka/poppy devices.
Change-Id: I8d094a8b6d78ea3bf8f929870a4a179495c29c78
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20516
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Don't need this additional 2ms delay as PCR read after sideband write
help to fix original hard hang issue.
This reverts commit d4b6ac19b0.
Change-Id: I4232cba5b92e17f728795f7c282af6161e385e9b
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20462
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
BIOS must ensure to read same PCR offset after PCR write operation
is done.
BUG=b:35587084
BRANCH=eve
TEST=manual stress testing of D0<->D3 transition on eve failing
unit. No hard hang with this fix.
Change-Id: Id3d567aab517b16ff99a526fc29c2d71bf4042d0
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20461
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
The PMC of PCH-H requires a different destination id.
TEST=Run on kontron/bsl6 and observed that PM registers are correctly
dumped at start of romstage.
Change-Id: I862e4df986f1cdea34f8fa45d016fb6b51f29122
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20479
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Move the generic I/O decode range setup before the console init.
TEST=Run on kontron/bsl6 which requires 0xa80/0xa81 decoded to
initialize serial ports. Serial console works from boot-
block on.
Change-Id: I9829f188c80eb73f6cd91b0c22e1c07da5745ad6
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20478
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
CONFIG_SOC_INTEL_COMMON_BLOCK_SGX controls building. The SGX feature
is still enabled from devicetree.cb. As of now this SGX init supports
only KBL (SKL not tested). Support of SGX for new SOCs would be added
incrementally in this common code base.
Change-Id: I0fbba364b7342e686a2287ea1a910ef9a4eed595
Signed-off-by: Pratik Prajapati <pratikkumar.v.prajapati@intel.com>
Reviewed-on: https://review.coreboot.org/20173
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch basically does two things -
1) Remove unnecessary setting of flex_ratio to TDP nominal:
Factory configured (default) Max Non-TURBO ratio(P1) is already cofigured
in MSR_PLATFORM_INFO(0xCE).
If this Maximum Non-TURBO Ratio(P1) needs to be modified, it should be done
using MSR_FLEX_RATIO (0x194).
Here, in this code, the FLEX_RATIO is being modified by the TDP Nominal
Ratio, reading the MSR_CONFIG_TDP_NOMINAL(0x648). But this value is
actually less than the factory configured Maximum Non TURBO Ratio (P1).
So, this code is actually not required.
Also, the Bit 12 in PCH Soft Strap Register is already set in descriptor.
This Bit implies Processor Boot Max Frequency -
0 = Disable Boot Max Frequency
1 = Enable Boot Max Frequency (Default)
This setting determines if the processor will operate at maximum frequency
at power-on and boot.
Thus this patch will avoid one extra platform warm reset now onwards.
2) Remove wrongly setting Max Frequency in Bootblock phase:
In the function set_max_frequency(), the P-State max ratio was set to
TDP Nominal ratio if C-TDP was enabled, else it was set to Max Non
Trbo ratio.
But, when the cpu gets reset, it will operate with the Max-Non Turbo
ratio only, which is greater than the TDP Nominal ratio.
So, no need to set back the ratio to TDP Nominal which is lower than
the currently operating frequency.
BUG=none
BRANCH=none
TEST=Build and boot poppy
Change-Id: I24bfc86ddf0f038d85da938e41e950382fe2a6c3
Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com>
Reviewed-on: https://review.coreboot.org/20050
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
There is no choice, if not leaving it with the default the build
will fail.
Change-Id: Id91e3ce87f8ced3001fcd2125f8f6781b270f5bc
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/20402
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Return CB_SUCCESS and CB_ERR instead of some integer.
Preparation to merge intel/soc and intel/nb opregion implementations.
Change-Id: Ib99fcfe347b98736979fc82ab3de48bfc6fc7dcd
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/20220
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Martin Roth <martinroth@google.com>
In the D0 and D3 ACPI methods use word access to the PME status and
control register. This brings the code inline with the Intel reference
code and matches how the kernel handles access to this register.
BUG=b:35587084
BRANCH=eve
TEST=manual stress testing of D0<>D3 transition across multiple devices
Change-Id: I53f7465d6ad5da1780a5641ff52056445ebaca8b
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/20364
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
For the skylake/kabylake generation of PCH there is an ACPI workaround
for emmc/sd power state that involves disabling and re-enabling dynamic
clock gating after enabling power to the controller, before setting the
power state to D0.
Under certain conditions we have observed that the controller is not
powered and ready by the time the kernel attempts to read the PME
control and status register and so the system will hang while attempting
to read PCI config register 0x84.
To ensure that the controller is ready add a 2ms delay after re-enabling
dynamic clock gating and before setting the power state to D0.
This issue has been observed on eMMC, but the same workaround exists for
the SD card interface so the same delay is added there.
BUG=b:35587084
BRANCH=eve
TEST=manual stress testing of D0<>D3 transition across many devices
shows no hard hang after 2 days.
Change-Id: If0f0323cf5437c54c907c332937b5de9dda2d8f6
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/20363
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
As per latest BWG, ucode reloading should be done at the end
of Mp Init, i.e., after PRMRR and other features are enabled.
No reloading specifically after SMM Relocation is required.
As, in the Common CPU MP Init code, we are already doing a
uCode load at the end of MP Init Feature Programming, hence,
the uCode loading after SMM relocation can be removed.
Change-Id: Ib1957c5fe5a8c83bb20b978a9841670b0c3e8846
Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com>
Reviewed-on: https://review.coreboot.org/20306
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
USB port status register can be used to decide if a particular port
was responsible for generating PME# resulting in device wake:
1. CSC bit is set and port is capable of waking on connect/disconnect
2. PLC bit is set and port is in resume state
BUG=b:37088992
TEST=Verified with wake on USB2.0 port 3, mosys shows:
19 | 2017-06-08 15:43:30 | Wake Source | PME - XHCI (USB 2.0 port) | 3
Change-Id: Ie4fa87393d8f096c4b3dca5f7a97f194cb065468
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/20122
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
There are many good reasons why we may want to run some sort of generic
callback before we're executing a reset. Unfortunateley, that is really
hard right now: code that wants to reset simply calls the hard_reset()
function (or one of its ill-differentiated cousins) which is directly
implemented by a myriad of different mainboards, northbridges, SoCs,
etc. More recent x86 SoCs have tried to solve the problem in their own
little corner of soc/intel/common, but it's really something that would
benefit all of coreboot.
This patch expands the concept onto all boards: hard_reset() and friends
get implemented in a generic location where they can run hooks before
calling the platform-specific implementation that is now called
do_hard_reset(). The existing Intel reset_prepare() gets generalized as
soc_reset_prepare() (and other hooks for arch, mainboard, etc. can now
easily be added later if necessary). We will also use this central point
to ensure all platforms flush their cache before reset, which is
generally useful for all cases where we're trying to persist information
in RAM across reboots (like the new persistent CBMEM console does).
Also remove cpu_reset() completely since it's not used anywhere and
doesn't seem very useful compared to the others.
Change-Id: I41b89ce4a923102f0748922496e1dd9bce8a610f
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/19789
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
