Many aggregate WMI drivers do not use -EPROBE_DEFER when they
cannot find a WMI device during probe, instead they require
all WMI devices associated with an platform device to become
available at once. This is currently achieved by adding those
WMI devices to the wmi_block_list before they are registered,
which is then used by the deprecated GUID-based functions to
search for WMI devices.
Replace this approach with a device link which defers probing
of the WMI device until the associated platform device has finished
probing (and has registered all WMI devices). New aggregate WMI
drivers should not rely on this behaviour.
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Link: https://lore.kernel.org/r/20231020211005.38216-2-W_Armin@gmx.de
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
For a long time now the acpi_video driver reports evdev brightness up/down
key events for the brightness hotkeys on most (non ancient) laptops.
asus-wmi also reports evdev brightness up/down key events for these
keys leading to each press being reported twice and e.g. GNOME increasing
the brightness by 2 steps instead of 1 step.
Use the acpi_video_handles_brightness_key_presses() helper to detect if
acpi_video is reporting brightness key-presses and if it is then don't
report the same events also from the asus-wmi driver.
Note there is a chance that this may lead to regressions where
the brightness hotkeys stop working because they are not actually
reported by the acpi_video driver. Unfortunately the only way to
find out if this is a problem is to try.
To at least avoid regressions on old hw using the eeepc-wmi driver,
implement this as a key filter in asus-nb-wmi so that the eeepc-wmi
driver is not affected.
Reported-by: James John <me@donjajo.com>
Closes: https://lore.kernel.org/platform-driver-x86/a2c441fe-457e-44cf-a146-0ecd86b037cf@donjajo.com/
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20231021094841.7419-1-hdegoede@redhat.com
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
strncpy() is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous
interfaces.
We expect ec_fw_string to be NUL-terminated based on its use with format
strings in thinkpad_acpi.c:
11241 | pr_notice("ThinkPad firmware release %s doesn't match the known patterns\n",
11242 | ec_fw_string);
Moreover, NUL-padding is not required since ec_fw_string is explicitly
zero-initialized:
11185 | char ec_fw_string[18] = {0};
When carefully copying bytes from one buffer to another in
pre-determined blocks (like what's happening here with dmi_data):
| static void find_new_ec_fwstr(const struct dmi_header *dm, void *private)
| {
| char *ec_fw_string = (char *) private;
| const char *dmi_data = (const char *)dm;
| /*
| * ThinkPad Embedded Controller Program Table on newer models
| *
| * Offset | Name | Width | Description
| * ----------------------------------------------------
| * 0x00 | Type | BYTE | 0x8C
| * 0x01 | Length | BYTE |
| * 0x02 | Handle | WORD | Varies
| * 0x04 | Signature | BYTEx6 | ASCII for "LENOVO"
| * 0x0A | OEM struct offset | BYTE | 0x0B
| * 0x0B | OEM struct number | BYTE | 0x07, for this structure
| * 0x0C | OEM struct revision | BYTE | 0x01, for this format
| * 0x0D | ECP version ID | STR ID |
| * 0x0E | ECP release date | STR ID |
| */
|
| /* Return if data structure not match */
| if (dm->type != 140 || dm->length < 0x0F ||
| memcmp(dmi_data + 4, "LENOVO", 6) != 0 ||
| dmi_data[0x0A] != 0x0B || dmi_data[0x0B] != 0x07 ||
| dmi_data[0x0C] != 0x01)
| return;
|
| /* fwstr is the first 8byte string */
| strncpy(ec_fw_string, dmi_data + 0x0F, 8);
... we shouldn't be using a C string api. Let's instead use memcpy() as
this more properly relays the intended behavior.
Do note that ec_fw_string will still end up being NUL-terminated since
we are memcpy'ing only 8 bytes into a buffer full of 18 zeroes. There's
still some trailing NUL-bytes there. To ensure this behavior, let's add
a BUILD_BUG_ON checking the length leaves space for at least one
trailing NUL-byte.
Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
Link: https://github.com/KSPP/linux/issues/90
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Mark Pearson <mpearson-lenovo@squebb.ca>
Reviewed-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20231020-strncpy-drivers-platform-x86-thinkpad_acpi-c-v1-1-312f2e33034f@google.com
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Summary of changes:
- CPU 0 hotplug support is deprecated in the upstream kernel. This
causes failures during level change for several customers. So, add a
silent workaround to use Cgroup v2, without user option.
- Increase number of CPUs in a single request
- Fix turbo mode enable/disable issue
- Error handling for invalid input options
This version addresses issues with:
- When CPU 0 hotplug is not possible, try cgroup v2 isolation
without any user input
- Fix turbo mode enable/disable swapped
- Sanitize command line integer and hex arguments
- Add more error messages
- Increase CPU count in one request
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
From kernel version 6.5, CPU 0 hotplug capability is deprecated.
If some SST profile doesn't have CPU 0, then it is no longer possible to
offline CPU 0. This means that user space threads will still run on
CPU 0.
To workaround this issue, use cgroup v2 isolation feature. Whenever there
/sys/devices/system/cpu/cpu0/online file is absent or open fails, isolate
CPU 0 via CPU cgroup v2 isolation. Also add a command line option to
force even if the /sys/devices/system/cpu/cpu0/online is present.
The previous commit "01bcb56f059e ("tools/power/x86/intel-speed-select:
Prevent CPU 0 offline") was just warning about this issue based on the
kernel version 6.5 and above. With this new approach, instead of warning
take action to mitigate the issue.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
With the increase in the CPU count, this count needs to be updated.
Increase max CPU count to 512.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When core-power is getting enabled, if the feaure is not supported,
display error.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
The command for turbo-mode enable and disable is swapped. Fix that.
Previously turbo-mode enable was actually disabling and disable was
enabling.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
TRL (turbo ratio limit) argument is passed in hex string. Clarify that
in the help.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
If the command takes some integer arguments, make sure the command
contains only digits. Same for Hex arguments. Otherwise return error.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
There have been instances when the default size (1M) of the STB is not
sufficient to get the complete traces of the failure. In such scenarios
we can use a module_param to enable full trace that shall contain more
debugging data. This is not a regular case and hence not enabling this
capability by default.
With this change, there will be two cases on how the driver fetches the
stb data:
1) A special case (proposed now) - which is required only for certain
platforms. Here, a new module param will be supplied to the driver that
will have a special PMFW supporting enhanced dram sizes for getting
the stb data. Without the special PMFW support, just setting the module
param will not help to get the enhanced stb data.
To adapt to this change, we will have a new amd_pmc_stb_handle_efr() to
handle enhanced firmware reporting mechanism. Note that, since num_samples
based r/w pointer offset calculation is not required for enhanced firmware
reporting we will have this mailbox command sent only in case of regular
STB cases.
2) Current code branch which fetches the stb data based on the parameters
like the num_samples, fsize and the r/w pointer.
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Co-developed-by: Harsh Jain <Harsh.Jain@amd.com>
Signed-off-by: Harsh Jain <Harsh.Jain@amd.com>
Signed-off-by: Sanket Goswami <Sanket.Goswami@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Link: https://lore.kernel.org/r/20231010145003.139932-3-Shyam-sundar.S-k@amd.com
[ij: Renamed flex_arr -> stb_data_arr]
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
In amd_pmc_stb_debugfs_open_v2(), the stb buffer is created based on the
num_samples and the read/write pointer offset. This holds good when the
num_samples reported by PMFW is less than S2D_TELEMETRY_BYTES_MAX; where
the stb buffer gets filled from 0th position until
S2D_TELEMETRY_BYTES_MAX - 1 based on the read/write pointer offset.
But when the num_samples exceeds the S2D_TELEMETRY_BYTES_MAX, the current
code does not handle it well as it does not account for the cases where
the stb buffer has to filled up as a circular buffer.
Handle this scenario into two cases, where first memcpy will have the
samples from location:
(num_samples % S2D_TELEMETRY_BYTES_MAX) - (S2D_TELEMETRY_BYTES_MAX - 1)
and next memcpy will have the newest ones i.e.
0 - (num_samples % S2D_TELEMETRY_BYTES_MAX - 1)
Suggested-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Sanket Goswami <Sanket.Goswami@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Link: https://lore.kernel.org/r/20231010145003.139932-2-Shyam-sundar.S-k@amd.com
[ij: renamed flex_arr -> stb_data_arr]
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
AMD MI300 MCM provides GET_METRICS_TABLE message to retrieve
all the system management information from SMU.
The metrics table is made available as hexadecimal sysfs binary file
under per socket sysfs directory created at
/sys/devices/platform/amd_hsmp/socket%d/metrics_bin
Metrics table definitions will be documented as part of Public PPR.
The same is defined in the amd_hsmp.h header.
Signed-off-by: Suma Hegde <suma.hegde@amd.com>
Reviewed-by: Naveen Krishna Chatradhi <nchatrad@amd.com>
Link: https://lore.kernel.org/r/20231010120310.3464066-2-suma.hegde@amd.com
[ij: lseek -> lseek(), dram -> DRAM in dev_err()]
[ij: added period to terminate a documentation sentence]
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Scan image loading flow for newer IFS generations are slightly different
from that of current generation. In newer schemes, loading need not be
done once for each socket as was done in gen0.
Also the width of NUM_CHUNKS bitfield in SCAN_HASHES_STATUS MSR has
increased from 8 -> 16 bits. Similarly there are width differences for
CHUNK_AUTHENTICATION_STATUS too.
Further the parameter to AUTHENTICATE_AND_COPY_CHUNK is passed
differently in newer generations.
Signed-off-by: Jithu Joseph <jithu.joseph@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Tested-by: Pengfei Xu <pengfei.xu@intel.com>
Link: https://lore.kernel.org/r/20231005195137.3117166-4-jithu.joseph@intel.com
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
The hardware definition of every TPMI feature contains a major and minor
version. When there is a change in the MMIO offset or change in the
definition of a field, hardware will change major version. For addition
of new fields without modifying existing MMIO offsets or fields, only the
minor version is changed.
Driver is developed to support uncore frequency control (UFS) for a major
and minor version. If the hardware changes major version, since offsets
and definitions are changed, driver cannot continue to provide UFS
interface to users. Driver can still function with minor version change
as it will just miss the new functionality added by the hardware.
The current implementation logs information message and skips adding
uncore sysfs entry for a resource for any version mismatch. Check major
and minor version mismatch for every valid resource and fail on any major
version mismatch by logging an error message. A valid resource has a
version which is not 0xFF.
If there is mismatch with the minor version, continue with a log message.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://lore.kernel.org/r/20231003184916.1860084-4-srinivas.pandruvada@linux.intel.com
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
The hardware definition of every TPMI feature contains a major and minor
version. When there is a change in the MMIO offset or change in the
definition of a field, hardware will change major version. For addition
of new fields without modifying existing MMIO offsets or fields, only the
minor version is changed.
Driver is developed to support SST functionality for a major and minor
version. If the hardware changes major version, since offsets and
definitions are changed, driver cannot continue to provide SST interface
to users. Driver can still function with a minor version change as it will
just miss the new functionality added by the hardware. The current
implementation doesn't ignore any version change.
If there is mismatch with the minor version, continue with an information
log message. If there is mismatch with the major version, log error and
exit.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://lore.kernel.org/r/20231003184916.1860084-3-srinivas.pandruvada@linux.intel.com
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
