[PATCH] spi: tegra210-quad: Add runtime autosuspend support

[PATCH] spi: tegra210-quad: Add runtime autosuspend support

[Vishwaroop A]

Using Tegra234, it was found that it takes about 10us to disable clocks
and 20us to enable clocks, adding about 30us overhead per operation.
For a 4MB firmware update with 16,384 page programs, this results in
~491ms total overhead (12% impact).

With Tegra234, flash operations were observed to occur in bursts with
50-200μs gaps between page programs. Testing on Tegra234 with various
operation patterns shows 500ms delay provides optimal balance, and for
longer operations (>500ms), the overhead is negligible. Therefore,
update the driver to use pm-runtime autosuspend with the default timeout
of 500ms to reduce the clock gating overhead during consecutive QSPI
transfers.

Signed-off-by: Vishwaroop A <va@nvidia.com>
---
 drivers/spi/spi-tegra210-quad.c | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/drivers/spi/spi-tegra210-quad.c b/drivers/spi/spi-tegra210-quad.c
index 7ea5aa993596..7fac695b5142 100644
--- a/drivers/spi/spi-tegra210-quad.c
+++ b/drivers/spi/spi-tegra210-quad.c
@@ -1000,7 +1000,8 @@ static int tegra_qspi_setup(struct spi_device *spi)
 
 	spin_unlock_irqrestore(&tqspi->lock, flags);
 
-	pm_runtime_put(tqspi->dev);
+	pm_runtime_mark_last_busy(tqspi->dev);
+	pm_runtime_put_autosuspend(tqspi->dev);
 
 	return 0;
 }
@@ -1765,6 +1766,14 @@ static int tegra_qspi_probe(struct platform_device *pdev)
 	init_completion(&tqspi->rx_dma_complete);
 	init_completion(&tqspi->xfer_completion);
 
+	/*
+	 * Set autosuspend delay to 500ms. Testing shows this value eliminates
+	 * suspend/resume overhead during burst operations while allowing quick
+	 * suspension during idle. For longer operations, the overhead is negligible.
+	 */
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 500);
+	pm_runtime_use_autosuspend(&pdev->dev);
+
 	pm_runtime_enable(&pdev->dev);
 	ret = pm_runtime_resume_and_get(&pdev->dev);
 	if (ret < 0) {
@@ -1781,7 +1790,8 @@ static int tegra_qspi_probe(struct platform_device *pdev)
 	tqspi->spi_cs_timing2 = tegra_qspi_readl(tqspi, QSPI_CS_TIMING2);
 	tqspi->def_command2_reg = tegra_qspi_readl(tqspi, QSPI_COMMAND2);
 
-	pm_runtime_put(&pdev->dev);
+	pm_runtime_mark_last_busy(&pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
 
 	ret = request_threaded_irq(tqspi->irq, NULL,
 				   tegra_qspi_isr_thread, IRQF_ONESHOT,
@@ -1802,6 +1812,7 @@ static int tegra_qspi_probe(struct platform_device *pdev)
 exit_free_irq:
 	free_irq(qspi_irq, tqspi);
 exit_pm_disable:
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
 	pm_runtime_force_suspend(&pdev->dev);
 	tegra_qspi_deinit_dma(tqspi);
 	return ret;
@@ -1814,6 +1825,7 @@ static void tegra_qspi_remove(struct platform_device *pdev)
 
 	spi_unregister_controller(host);
 	free_irq(tqspi->irq, tqspi);
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
 	pm_runtime_force_suspend(&pdev->dev);
 	tegra_qspi_deinit_dma(tqspi);
 }
@@ -1839,7 +1851,8 @@ static int __maybe_unused tegra_qspi_resume(struct device *dev)
 
 	tegra_qspi_writel(tqspi, tqspi->command1_reg, QSPI_COMMAND1);
 	tegra_qspi_writel(tqspi, tqspi->def_command2_reg, QSPI_COMMAND2);
-	pm_runtime_put(dev);
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
 
 	return spi_controller_resume(host);
 }
-- 
2.17.1

Re: [PATCH] spi: tegra210-quad: Add runtime autosuspend support

[Mark Brown]

On Tue, 24 Feb 2026 09:24:52 +0000, Vishwaroop A wrote:
> Using Tegra234, it was found that it takes about 10us to disable clocks
> and 20us to enable clocks, adding about 30us overhead per operation.
> For a 4MB firmware update with 16,384 page programs, this results in
> ~491ms total overhead (12% impact).
> 
> With Tegra234, flash operations were observed to occur in bursts with
> 50-200μs gaps between page programs. Testing on Tegra234 with various
> operation patterns shows 500ms delay provides optimal balance, and for
> longer operations (>500ms), the overhead is negligible. Therefore,
> update the driver to use pm-runtime autosuspend with the default timeout
> of 500ms to reduce the clock gating overhead during consecutive QSPI
> transfers.
> 
> [...]

Applied to

   https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next

Thanks!

[1/1] spi: tegra210-quad: Add runtime autosuspend support
      commit: 7c12f6ead4672cb08b74e6f6115eb04dca8ccfa4

All being well this means that it will be integrated into the linux-next
tree (usually sometime in the next 24 hours) and sent to Linus during
the next merge window (or sooner if it is a bug fix), however if
problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing
and review of the tree, please engage with people reporting problems and
send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they
should be sent as incremental updates against current git, existing
patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying
to this mail.

Thanks,
Mark