This version addresses issues with:
- Support of SST BF/TF support per level
- Increase number of CPUs displayed
- Present all TRL levels for turbo-freq
- Fix display for unsupported levels
- Support multiple dies
- Increase die count
- Change CPU display for non compute domain
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
SST BF and TF can be enabled/disabled per level. So check the current
level support from the mask of supported levels.
This change from a single level to mask for info.sst_tf_support and
info.sst_tf_support is indicated by API version change. Use as mask for
API version above 2. In this way there is no change in behavior when
running on older kernel with API version 2.
Since the tool can support now API version 3, update the supported API
version.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
During call to "intel-speed-select turbo-freq info" some junk values are
reported for unsupported levels. Initialize the structure fact_info with
0s, so that isst_fact_display_information() will skip "0" values in the
frequency.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
When the die id is same as punit compute die ID, treat them same. In this
case, when for_each_online_power_domain_in_set() is called, then don't
loop for each punit in a die. Just loop for all punits in a package.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
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>
This version addresses issues with:
- CPU count display for power domain != 0
- Support more than 8 sockets
- Error on max CPU count exceeds in one request
- Prevent trying CPU 0 hotplug for kernel version 6.5 or later
- Change mem-frequency display to max-mem-frequency
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Kernel 6.5 version deprecated CPU 0 hotplug. This will cause all
requests to fail to offline CPU 0. Check version number of kernel
and ignore CPU 0 hotplug request with debug aid to use cgroup
isolation feature for CPU 0.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
There is a limit on number of CPUs in one request. This is set to 256.
Currently tool silently ignores request for count over 256. Give an
error message to indicate this.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Fix CPU count display for power domain != 0. In the function
punit_id is always 0, so it never incremented cpu count for power
domain id != 0.
Update punit_id after call to update_punit_cpu_info() to what is
actually received from the kernel.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
This version addresses issues with core power configuration for
non CPU dies. Also address issue with JSON formatting of output.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When core-power configuration or enabled is modified, this is only done
for compute dies. But the config must also be set to cores with no CPUs.
Without this the configuration is not affective.
On displaying config information, allow display for non compute dies
also.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Update tool and supported API version. This is the first version which
supports newer Xeon platforms with TPMI support.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: subject and changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
There are some differences compared to Sapphire Rapids. So, add a separate
API.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
On supported systems, it is possiible to isolate CPUs instead of
CPU online/offline. This is optional and can be specified using
-g option when running as daemon or in combination with -o option
for SST-PP level change.
CPU isolation doesn't isolate IRQs. So IRQs needs to be moved away
from isoolated CPUs. This can be done via IRQ sysfs or irqbalance daemon.
The IRQ balance daemon is also capable to parse thermal HFI messages to
move IRQs away from CPUS, which are supposed be isolated. But this
requires version released after July 2022.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
During perf level change cpuset is allocated but not freed.
Add free_cpu_set() in success and failure path.
Although this is not an issue, as the program will exit after
processing of online/offline, but for completeness add the
free_cpu_set().
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Some operations applies to cpu-power-domain only. Add check for cpu id
for these functions.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
If the new TDP level requetsted is same as the current TDP level, don't
call into driver to change level.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Non-cpu power domain does not support cpu clos config.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Get punit core mapping information using format of MSR 0x54. Based
on the API version, decode is done using new format. The new format
also include a power domain ID. TPMI SST information is for each
power domain.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
In some cases, the output format may be different with different
api_version because of different capabilities or for backward
capabilities reason.
Introduce api_version() to get the api_version of the platform running.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Platform specific code also needs to give debug output.
Introduce is_debug_enabled() for this purpose.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Different api_version suggests different kernel driver used and
different interface is used to communication with the hardware.
Allow setting platform specific callbacks based on api_version.
Currently, all platforms with api_version 1 uses Mbox/MMIO interfaces.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Allow platform specific implementation to adjust the uncore frequency.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
isst_get_pbf_info_complete does nothing but just free the core_mask.
Remove the function and do free core_mask directly and free core mask in
the caller.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
isst-core.c should contain generic core APIs only.
Platform specific implementations/configurations should be removed from
this file.
Introduce isst-core-mbox.c and move all mbox/mmio specific functions to
this file.
Introduce struct isst_platform_ops which contains a series of callbacks
that used by the core APIs but need platform specific implementation.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
isst_fill_platform_info fills platform specific information.
And it is the proper place to set platform specific callbacks, as done in
next patch.
As the platform specific callbacks are needed in all cases, including
isst_print_platform_information.
The best way to achieve both is to invoke isst_fill_platform_info
unconditionally, and make isst_print_platform_information leverage the
data already filled.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Remove hardcoded DISP_FREQ_MULTIPLIER in the code and use
isst_get_disp_freq_multiplier() instead.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
isst-config.c should only contain generic code.
Move mbox functions which are platform specific code to isst-core.c.
As there are some platform specific parameters set via generic
application options, introduce isst_update_platform_param to pass these
parameters to platform specific code.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
The main thing done in isst_print_extended_platform_info is to get the
isst feature status by checking one of the power domains of the
platform.
This can be done using the for_each_online_power_domain_in_set()
function, which makes the code clean and easier to read.
No functional changes are expected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
for_each_online_package_in_set is actually used to invoke callback for
each power domain.
This is not a problem when there is a single power domain within a
package/die, but it does not reflect the truth in multi-punit case.
Rename for_each_online_package_in_set to
for_each_online_power_domain_in_set.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
New platforms may have more than 1 punit in a Package/Die, thus it can
have multiple power domains in a Package/Die. Package id and die id is not
sufficient to refer to a specific Power domain.
Introduce support for multi-punit per package/die.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Punit id can also be retrieved from ISST_IF_GET_PHY_ID.
punit id is unique within a Package/Die, and together with Package id and
Die id, they can be used to refer to a specific SST power domain.
For current platforms, Punit id is always Zero. So no functional changes
are expected for the current platforms.
While here, prevent issuing IOCTL if the file /dev/isst_interface can't be
opened.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
This release adds following change:
- Minor fixes for coverity static analysis
- Don't read cpufreq on offline CPUs
- SST turbo-freq enable on auto mode when user disables SMT from
kernel command line
- Fix uncore frequency display
- Set uncore frequency max/min limits on perf level change
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
When perf level is changed, uncore limits can change. Set the uncore
limits via Linux uncore sysfs, when user changes perf level with
-o option.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Need memory frequency quirk as Sapphire Rapids in Emerald Rapids.
So add Emerald Rapids CPU model check in is_spr_platform().
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: Subject, changelog and code edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
When SMT is disabled from kernel command line, sibling CPUs still
appears in the sysfs as offline CPUs. This is a problem when turbo-freq
is enabled in auto mode. They are still assigned to CLOS value
of 3 as they are still in the present CPU list. But they are not in the
sibling list of a CPU. When the CPU is a high priority CPU, because of
sibling it will be still set to CLOS to 3 as CLOS is assigned at core
level not at CPU level.
So, avoid setting CLOS 3 to offline CPU.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Due to some recent kernel changes, reading cpufreq attributes like
scaling_max_freq on offline CPUs returns error. So avoid reading
cpufreq attributes on offline CPUs.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
strlen() and strtok() takes null-termimated strings as input.
Make sure these strings are null-terminated before using them.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
struct isst_id *id is a pointer, comparing it with less than zero is wrong.
The check is there to make sure the id->pkg and id->die is set to -1, when
it is illegal or unavailable. Here comparing with MAX_PACKAGE_COUNT and
MAX_DIE_PER_PACKAGE is sufficient.
Hence remove the wrong check.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[srinivas.pandruvada@linux.intel.com: Subject and changelog edits]
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Update version number.
This version includes fixes for:
- fix build failure when using gcc options -Wl,--as-needed
- Fix warning for perf_cap.cpu may be uninitialized
- Fix off by one check for MAX_DIE_PER_PACKAGE
- Fix issue with use of get_physical_die_id instead of
get_physical_die_id
Optimizations:
- Removed unused interfaces and functions
- Better handle package, die, cpu combination by
defining a struct and set at one place instead
at each user level.
New functional change:
- Warn if turbo is disabled and SST turbo-freq feature is requested
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Optimize CPU initialization.
Do cpu related initialization in one function, including setting the cpu
present_cpumask, target_cpumask, and cpu_map and core_count arrays.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
cpu_map already has the cpu package id, die id information.
Thus there is no need to re-evaluating sysfs attributes or stored data
file to get the package id and die id of a given CPU each time.
In order to unitlize this, cpu_map needs to be created unconditionally.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
