[WIP][DNM] Add s2idle docs #562
Conversation
github-actions
bot
assigned
devarsht,
DhruvaG2000,
praneethbajjuri,
r-vignesh,
uditkumarti and
vishalmti
github-actions
bot
requested review from
devarsht,
praneethbajjuri,
r-vignesh,
uditkumarti and
vishalmti
|
New warnings found with rstcheck: |
| * **System State:** (Optional) Power Down or Retention | ||
| 4. **PSCI Enforcement:** The PSCI implementation verifies that all other cores in the requested node are indeed idle. If they are not, the request is denied (to prevent race conditions). | ||
|
|
||
| ************************************* |
There was a problem hiding this comment.
The section identifier line should be of equal length to the section name.
| Suspend-to-Idle (s2idle), also known as "freeze," is a generic, pure software, light-weight variant of system suspend. | ||
| In this state, the Linux kernel freezes user space tasks, suspends devices, and then puts all CPUs into their deepest available idle state. | ||
|
|
||
| ********************** |
There was a problem hiding this comment.
The section identifier line should be of equal length to the section name.
|
|
||
| ************************** | ||
| OS Initiated (OSI) Mode | ||
| ************************** |
There was a problem hiding this comment.
The section identifier line should be of equal length to the section name.
DhruvaG2000
force-pushed
the
s2idle_doc
branch
from
b935f77 to
ef4ad74
Compare
|
@cshilwant , and rest I'll mark the PR ready for review once it has undergone LPM team internal review. Till then requesting you to hold off reviews. Thanks! |
source/linux/Foundational_Components/Power_Management/pm_psci_s2idle.rst
Outdated
Show resolved
Hide resolved
| On ARM64 platforms, both ``s2idle`` and ``deep`` states can achieve similar power savings (e.g., suspending to RAM / DDR Self-Refresh). | ||
| The primary differences lie in the software execution flow, specifically how CPUs are managed and which PSCI APIs are invoked. | ||
|
|
||
| .. list-table:: S2Idle vs Deep Sleep (ARM64) |
There was a problem hiding this comment.
Is it necessary to specify ARM64 platforms since you already specified ARM64 previously and this page will (presumably) only be used in ARM64 platforms?
| - **Offlined**: Non-boot CPUs are hot-unplugged (removed) from the system via ``CPU_OFF``. | ||
|
|
||
| * - **Entry Path** | ||
| - **cpuidle**: Uses the standard CPU idle framework and governance. It runtime suspends each driver to make sure it's idle. |
There was a problem hiding this comment.
| - **cpuidle**: Uses the standard CPU idle framework and governance. It runtime suspends each driver to make sure it's idle. | |
| - **cpuidle**: Uses the standard CPU idle framework and governance. Each driver is put into an idle state by runtime suspend. |
source/linux/Foundational_Components/Power_Management/pm_psci_s2idle.rst
Outdated
Show resolved
Hide resolved
source/linux/Foundational_Components/Power_Management/pm_psci_s2idle.rst
Outdated
Show resolved
Hide resolved
source/linux/Foundational_Components/Power_Management/pm_psci_s2idle.rst
Outdated
Show resolved
Hide resolved
| In this state, the Linux kernel freezes user space tasks, suspends devices, and then puts all CPUs into their deepest available idle state. | ||
|
|
||
| ********************************** | ||
| S2Idle vs Deep Sleep (mem) |
There was a problem hiding this comment.
Wondering if this whole section should be moved to be the last / one of the last sections? I think explaining PSCI would be good to do before going into comparisons on how it operates differently than Deep Sleep
| 5. Thaw tasks | ||
|
|
||
| The `cpuidle` driver calls the PSCI `CPU_SUSPEND` API to transition the CPU (and potentially | ||
| higher-level topology nodes like clusters) into a low-power state. The effectiveness of s2idle |
There was a problem hiding this comment.
Assume the reader doesn't know what the topology nodes / clusters are. This should either be explained or the information in the parenthesis should be removed.
There was a problem hiding this comment.
ok since we will need more docs on how we're grouping things and what is cluster/ system definition I will send it in next PR. removing for now
| 2. Suspend devices | ||
| | | ||
| v | ||
| 3. cpuidle driver -----------> CPU_SUSPEND (SMC/HVC) |
There was a problem hiding this comment.
ok will add explanation for secure monitor call.
| | Enter low-power | ||
| | hardware state |
There was a problem hiding this comment.
| | Enter low-power | |
| | hardware state | |
| | CPU enters low-power | |
| | hardware state |
Only suggesting this because I think the cpuidle to PSCI firmware can be a little confusing.
DhruvaG2000
force-pushed
the
s2idle_doc
branch
from
ef4ad74 to
fcd21dd
Compare
|
New warnings found with rstcheck: |
8 participants
Document all about s2idle and PSCI and how the whole stack helps us in selecting between low power modes on the TI AM62L