Re: [Patch v9] crypto: qce - Add runtime PM and interconnect bandwidth scaling support

On Wed, Jul 29, 2026 at 04:34:55PM +0530, Kuldeep Singh wrote:
> From: Udit Tiwari <[email protected]>
> 
> The Qualcomm Crypto Engine (QCE) driver currently lacks support for
> runtime power management (PM) and interconnect bandwidth control.
> As a result, the hardware remains fully powered and clocks stay
> enabled even when the device is idle. Additionally, static
> interconnect bandwidth votes are held indefinitely, preventing the
> system from reclaiming unused bandwidth.
> 
> Address this by enabling runtime PM and dynamic interconnect
> bandwidth scaling to allow the system to suspend the device when idle
> and scale interconnect usage based on actual demand. Improve overall
> system efficiency by reducing power usage and optimizing interconnect
> resource allocation.
> 
> Signed-off-by: Udit Tiwari <[email protected]>
> Tested-by: Pankaj Patil <[email protected]>
> Signed-off-by: Kuldeep Singh <[email protected]>
> ---
> I am taking care of follow-up of this patch from the original author,
> with their consent, while preserving the original authorship.
> 
> Changes in v9:
> - Rebase to latest linux-next/master(tag: next-20260728).
> - Link to v8:
>   https://lore.kernel.org/linux-arm-msm/[email protected]/
> 
> Changes in v8:
> - Drop pm_clk framework (devm_pm_clk_create/pm_clk_add/pm_clk_suspend/
>   pm_clk_resume); use devm_clk_get_optional() and direct
>   clk_prepare_enable()/clk_disable_unprepare() in runtime PM callbacks.
>   This removes the CONFIG_PM_CLK dependency and the build error reported
>   by the kernel test robot.
> - Replace icc_disable() with icc_set_bw(path, 0, 0) in runtime suspend
>   to avoid corrupting the internal 'enabled' flag, which would cause
>   subsequent icc_set_bw() calls in resume to be silently skipped during
>   aggregation.
> - Fix ICC vote ordering: cast bandwidth vote before enabling clocks in
>   resume; disable clocks before dropping ICC vote in suspend.
> - Use PM_RUNTIME_ACQUIRE_AUTOSUSPEND()/PM_RUNTIME_ACQUIRE_ERR() wrapper
>   macros instead of raw ACQUIRE() in both qce_handle_queue() and probe.
> - Drop __maybe_unused from runtime PM callbacks; use RUNTIME_PM_OPS /
>   SYSTEM_SLEEP_PM_OPS (non-SET_ prefix) and pm_ptr(&qce_crypto_pm_ops).
> - Drop unnecessary ret = 0 initializations in qce_handle_queue() and
>   qce_runtime_resume().
> - Extend probe comment to explain ICC + clock ordering rationale.
> - Link to v7:
>   https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v7:
> - Use ACQUIRE guard in probe to simplify runtime PM management and error
>   paths.
> - Drop redundant icc_enable() call in runtime resume path.
> - Explicitly call pm_clk_suspend(dev) and pm_clk_resume(dev) within the
>   custom runtime PM callbacks. Since custom callbacks are provided to
>   handle interconnect scaling, the standard PM clock helpers must be
>   invoked manually to ensure clocks are gated/ungated.
> - Link to v6:
>  https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v6:
> - Adopt ACQUIRE(pm_runtime_active_try, ...) for scoped runtime PM
>   management in qce_handle_queue(). This removes the need for manual
>   put calls and goto labels in the error paths, as suggested by Konrad.
> - Link to v5:
>   https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v5:
> - Drop Reported-by and Closes tags for kernel test robot W=1 warnings,
>   as the issue was fixed within the same patch series.
> - Fix a minor comment indentation/style issue.
> - Link to v4:
>   https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v4:
> - Annotate runtime PM callbacks with __maybe_unused to silence W=1
>   warnings.
> - Add Reported-by and Closes tags for kernel test robot warning.
> - Link to v3:
>   https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v3:
> - Switch from manual clock management to PM clock helpers
>   (devm_pm_clk_create() + pm_clk_add()); no direct clk_* enable/disable
>   in runtime callbacks.
> - Replace pm_runtime_get_sync() with pm_runtime_resume_and_get(); remove
>   pm_runtime_put_noidle() on error.
> - Define PM ops using helper macros and reuse runtime callbacks for
>   system sleep via pm_runtime_force_suspend()/pm_runtime_force_resume().
> - Link to v2:
>   https://lore.kernel.org/lkml/[email protected]/
> 
> Changes in v2:
> - Extend suspend/resume support to include runtime PM and ICC scaling.
> - Register dev_pm_ops and implement runtime_suspend/resume callbacks.
> - Link to v1:
>   https://lore.kernel.org/lkml/[email protected]/
> ---
>  drivers/crypto/qce/core.c | 99 ++++++++++++++++++++++++++++++++++++---
>  1 file changed, 92 insertions(+), 7 deletions(-)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt