From: Thomas Zimmermann <tzimmermann@suse.de>
To: Devarsh Thakkar <devarsht@ti.com>,
David Airlie <airlied@gmail.com>, Simona Vetter <simona@ffwll.ch>,
Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
Maxime Ripard <mripard@kernel.org>, Rob Herring <robh@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
Conor Dooley <conor+dt@kernel.org>,
Neil Armstrong <neil.armstrong@linaro.org>,
Bjorn Andersson <bjorn.andersson@oss.qualcomm.com>,
dri-devel@lists.freedesktop.org, devicetree@vger.kernel.org,
linux-kernel@vger.kernel.org
Cc: praneeth@ti.com, vigneshr@ti.com, s-jain1@ti.com,
s-wang12@ti.com, r-donadkar@ti.com, r-sharma3@ti.com, afd@ti.com
Subject: Re: [PATCH 4/6] drm/tiny: panel-ssd16xx: Add power management support
Date: Tue, 5 May 2026 09:08:14 +0200 [thread overview]
Message-ID: <af9a1cde-00be-49fd-b236-1cb2bfb4837e@suse.de> (raw)
In-Reply-To: <20260430183311.2978142-5-devarsht@ti.com>
Hi
Am 30.04.26 um 20:33 schrieb Devarsh Thakkar:
> Add system suspend/resume and runtime PM with idle timeout support to
> the SSD16xx driver.
I'll go over this patch after the core driver from patch 3 is ready.
Best regards
Thomas
>
> E-paper panels are bistable: they continue to display the last rendered
> image indefinitely even in deep sleep or when power is cut to the
> controller IC. This makes it practical to suspend the controller
> aggressively — including while a display application is still open and the
> user is simply reading — without any visible disruption. The driver
> exploits this by firing the autosuspend timer after an idle timeout,
> putting the controller into deep sleep regardless of CRTC state, and waking
> it transparently on the next frame update.
>
> The SSD16xx family supports two deep sleep modes:
>
> Mode 1 (RAM retained): used for runtime idle. Display RAM survives the
> sleep, so resume only requires a hardware reset and re-initialisation; no
> full redraw is needed, keeping wake latency minimal.
>
> Mode 2 (RAM lost): used for system suspend. Maximises power savings at
> the cost of a complete controller re-init and repaint on resume.
>
> Runtime PM: Each hardware-touching callback wraps its SPI access with
> pm_runtime_resume_and_get() / pm_runtime_put_autosuspend() pairs so the PM
> reference count drops to zero after every update. This allows the
> autosuspend timer to fire while the CRTC is still enabled, putting the
> controller into Mode 1 deep sleep between updates without requiring the
> application to close or the display pipeline to be torn down.
>
> An autosuspend delay of 35 seconds keeps the panel active across typical
> user-interaction gaps while still capturing long idle periods.
>
> System suspend/resume: On system suspend the driver switches to Mode 2 and
> quiesces the atomic pipeline via drm_mode_config_helper_suspend(). If the
> autosuspend timer already fired (device is RPM_SUSPENDED), the driver wakes
> the panel via HWRESET (the only exit from deep sleep since SPI is
> unresponsive), sends Mode 2 directly, and tracks the state with a
> pm_force_suspended flag so the matching force_resume is called only when
> needed. On resume, Mode 1 is restored for subsequent runtime PM cycles and
> the atomic pipeline is rebuilt via drm_mode_config_helper_resume().
>
> Per-client initialisation refresh: A drm_driver.master_set callback arms
> init_refresh_pending when a new client opens the DRM device as a DRM
> master, ensuring its first frame uses the configured init-refresh waveform.
> A paired master_drop callback clears the flag on fd close by a DRM master.
> This is kept separate from the runtime resume path since Mode 1 retains RAM
> and a full refresh on every runtime wake would be wasteful.
>
> Signed-off-by: Devarsh Thakkar <devarsht@ti.com>
> ---
> drivers/gpu/drm/tiny/panel-ssd16xx.c | 203 +++++++++++++++++++++++++--
> 1 file changed, 193 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/tiny/panel-ssd16xx.c b/drivers/gpu/drm/tiny/panel-ssd16xx.c
> index b232837c54ff..6bf763667d82 100644
> --- a/drivers/gpu/drm/tiny/panel-ssd16xx.c
> +++ b/drivers/gpu/drm/tiny/panel-ssd16xx.c
> @@ -12,6 +12,7 @@
> #include <linux/delay.h>
> #include <linux/module.h>
> #include <linux/of.h>
> +#include <linux/pm_runtime.h>
> #include <linux/property.h>
> #include <linux/spi/spi.h>
>
> @@ -88,6 +89,7 @@ MODULE_PARM_DESC(color_mode,
>
> /* SPI command codes (common) */
> #define SSD16XX_CMD_DRIVER_OUTPUT_CONTROL 0x01
> +#define SSD16XX_CMD_DEEP_SLEEP_MODE 0x10
> #define SSD16XX_CMD_DATA_ENTRY_MODE 0x11
> #define SSD16XX_CMD_SW_RESET 0x12
> #define SSD16XX_CMD_MASTER_ACTIVATION 0x20
> @@ -100,6 +102,9 @@ MODULE_PARM_DESC(color_mode,
> #define SSD16XX_CMD_SET_RAM_X_ADDRESS_COUNTER 0x4E
> #define SSD16XX_CMD_SET_RAM_Y_ADDRESS_COUNTER 0x4F
>
> +/* Runtime PM autosuspend delay (ms): keep display active across typical gaps */
> +#define SSD16XX_PM_AUTOSUSPEND_DELAY_MS 35000
> +
> /*
> * Data Entry Mode (command 0x11) AM/IDY/IDX bit encoding (common).
> *
> @@ -405,8 +410,10 @@ struct ssd16xx_panel {
> bool init_refresh_pending; /* First frame after refresh_mode_init enable */
> bool first_clear_done; /* clear_on_init has already fired once */
> bool display_cleared_on_deinit; /* Avoid redundant clear in atomic_disable/master_drop */
> + bool pm_force_suspended; /* pm_runtime_force_suspend was called in pm_suspend */
>
> - int orientation; /* Display orientation in degrees: 0/90/180/270 */
> + int orientation; /* Display orientation in degrees: 0/90/180/270 */
> + u8 deep_sleep_mode; /* Deep sleep mode to use on next disable */
> enum ssd16xx_refresh_mode refresh_mode; /* Active refresh mode */
> enum ssd16xx_color_mode color_mode; /* Active color mode (BW or 3-color) */
> bool fast_lut_pending; /* LUT pre-load needed before next fast refresh */
> @@ -1526,10 +1533,9 @@ static void ssd16xx_plane_atomic_update(struct drm_plane *plane,
> struct drm_rect rect;
> int ret;
>
> - drm_dbg(&panel->drm, "plane_atomic_update: fb=%p, initialized=%d\n",
> - fb, panel->initialized);
> + drm_dbg(&panel->drm, "plane_atomic_update: fb=%p\n", fb);
>
> - if (!fb || !panel->initialized)
> + if (!fb)
> return;
>
> /*
> @@ -1541,6 +1547,12 @@ static void ssd16xx_plane_atomic_update(struct drm_plane *plane,
> return;
> }
>
> + ret = pm_runtime_resume_and_get(panel->drm.dev);
> + if (ret < 0) {
> + drm_err(&panel->drm, "plane_atomic_update: failed to resume: %d\n", ret);
> + return;
> + }
> +
> if (!drm_atomic_helper_damage_merged(old_state, new_state, &rect)) {
> rect.x1 = 0;
> rect.y1 = 0;
> @@ -1598,6 +1610,9 @@ static void ssd16xx_plane_atomic_update(struct drm_plane *plane,
> panel->init_refresh_pending = false;
> panel->border_waveform_pending = true;
> }
> +
> + pm_runtime_mark_last_busy(panel->drm.dev);
> + pm_runtime_put_autosuspend(panel->drm.dev);
> }
>
> static const struct drm_plane_helper_funcs ssd16xx_plane_helper_funcs = {
> @@ -1656,6 +1671,10 @@ static void ssd16xx_crtc_atomic_disable(struct drm_crtc *crtc,
> if (panel->clear_on_disable < 0 || panel->display_cleared_on_deinit)
> goto out;
>
> + ret = pm_runtime_resume_and_get(panel->drm.dev);
> + if (ret < 0)
> + goto out;
> +
> drm_dbg(&panel->drm, "clear_on_disable: running, mode=%d\n",
> panel->clear_on_disable);
> ret = ssd16xx_clear_display(panel,
> @@ -1663,10 +1682,12 @@ static void ssd16xx_crtc_atomic_disable(struct drm_crtc *crtc,
> panel->clear_on_disable));
> if (ret) {
> drm_err(&panel->drm, "atomic_disable: clear failed: %d\n", ret);
> + pm_runtime_put_autosuspend(panel->drm.dev);
> goto out;
> }
>
> panel->display_cleared_on_deinit = true;
> + pm_runtime_put_sync_suspend(panel->drm.dev);
> out:
> drm_dev_exit(idx);
> }
> @@ -1682,6 +1703,12 @@ static void ssd16xx_crtc_atomic_enable(struct drm_crtc *crtc,
> return;
>
> panel->display_cleared_on_deinit = false;
> + ret = pm_runtime_resume_and_get(panel->drm.dev);
> + if (ret < 0) {
> + drm_err(&panel->drm, "crtc_atomic_enable: failed to resume: %d\n", ret);
> + drm_dev_exit(idx);
> + return;
> + }
>
> drm_dbg(&panel->drm, "atomic_enable: %dx%d\n",
> crtc_state->mode.hdisplay, crtc_state->mode.vdisplay);
> @@ -1689,12 +1716,19 @@ static void ssd16xx_crtc_atomic_enable(struct drm_crtc *crtc,
> panel->width = crtc_state->mode.hdisplay;
> panel->height = crtc_state->mode.vdisplay;
>
> - ret = ssd16xx_hw_init(panel);
> - if (ret) {
> - drm_err(&panel->drm, "crtc_atomic_enable: HW init failed: %d\n", ret);
> - goto out;
> + /*
> + * pm_runtime_resume_and_get() triggers pm_runtime_resume which runs
> + * hw_init and sets initialized. If runtime PM is disabled the callback
> + * never fires, so fall back to running hw_init directly here.
> + */
> + if (!panel->initialized) {
> + ret = ssd16xx_hw_init(panel);
> + if (ret) {
> + drm_err(&panel->drm, "crtc_atomic_enable: HW init failed: %d\n", ret);
> + goto out;
> + }
> + panel->initialized = true;
> }
> - panel->initialized = true;
>
> /* Clear display on first app launch if configured */
> ret = ssd16xx_clear_display_on_init(panel);
> @@ -1714,6 +1748,8 @@ static void ssd16xx_crtc_atomic_enable(struct drm_crtc *crtc,
> }
>
> out:
> + pm_runtime_mark_last_busy(panel->drm.dev);
> + pm_runtime_put_autosuspend(panel->drm.dev);
> drm_dev_exit(idx);
> }
>
> @@ -1730,7 +1766,7 @@ static void ssd16xx_crtc_atomic_flush(struct drm_crtc *crtc,
> struct drm_rect full;
> int ret, idx;
>
> - if (!panel->reinit_pending || !panel->initialized)
> + if (!panel->reinit_pending)
> return;
>
> if (!drm_dev_enter(&panel->drm, &idx))
> @@ -1738,6 +1774,13 @@ static void ssd16xx_crtc_atomic_flush(struct drm_crtc *crtc,
>
> panel->reinit_pending = false;
>
> + ret = pm_runtime_resume_and_get(panel->drm.dev);
> + if (ret < 0) {
> + drm_err(&panel->drm, "atomic_flush: failed to resume: %d\n", ret);
> + drm_dev_exit(idx);
> + return;
> + }
> +
> drm_dbg(&panel->drm, "atomic_flush: reinit, orientation=%u°\n",
> panel->orientation);
>
> @@ -1762,6 +1805,8 @@ static void ssd16xx_crtc_atomic_flush(struct drm_crtc *crtc,
> }
>
> out:
> + pm_runtime_mark_last_busy(panel->drm.dev);
> + pm_runtime_put_autosuspend(panel->drm.dev);
> drm_dev_exit(idx);
> }
>
> @@ -2189,11 +2234,17 @@ static void ssd16xx_drm_master_drop(struct drm_device *drm,
> if (panel->clear_on_close < 0 || panel->display_cleared_on_deinit)
> return;
>
> + ret = pm_runtime_resume_and_get(drm->dev);
> + if (ret < 0)
> + return;
> +
> ret = ssd16xx_clear_display_on_exit(panel);
> if (ret)
> drm_err(drm, "master_drop: clear on close failed: %d\n", ret);
>
> panel->display_cleared_on_deinit = true;
> + /* sync suspend — bypass autosuspend, sleep immediately after clear */
> + pm_runtime_put_sync_suspend(drm->dev);
> }
>
> static struct drm_driver ssd16xx_drm_driver = {
> @@ -2496,6 +2547,20 @@ static int ssd16xx_probe(struct spi_device *spi)
> if (ret)
> return ret;
>
> + /* Default to runtime sleep mode (RAM retained if supported) */
> + panel->deep_sleep_mode = panel->controller_cfg->deep_sleep_mode_level1;
> + /*
> + * Mark the device active before enabling runtime PM. The SPI device
> + * persists across module reload cycles; pm_runtime_enable() alone
> + * does not clear a stale dev->power.runtime_error left by a previous
> + * failed hw_init, which would cause pm_runtime_resume_and_get() to
> + * return -EINVAL on every subsequent call.
> + */
> + pm_runtime_set_active(dev);
> + pm_runtime_enable(dev);
> + pm_runtime_set_autosuspend_delay(dev, SSD16XX_PM_AUTOSUSPEND_DELAY_MS);
> + pm_runtime_use_autosuspend(dev);
> +
> drm_dbg(drm, "SSD16xx e-paper display initialized (%dx%d, %d° rotation)\n",
> panel->width, panel->height, panel->orientation);
>
> @@ -2508,6 +2573,8 @@ static void ssd16xx_remove(struct spi_device *spi)
> {
> struct ssd16xx_panel *panel = spi_get_drvdata(spi);
>
> + pm_runtime_dont_use_autosuspend(&spi->dev);
> + pm_runtime_disable(&spi->dev);
> drm_dev_unplug(&panel->drm);
> drm_atomic_helper_shutdown(&panel->drm);
> }
> @@ -2516,9 +2583,124 @@ static void ssd16xx_shutdown(struct spi_device *spi)
> {
> struct ssd16xx_panel *panel = spi_get_drvdata(spi);
>
> + pm_runtime_dont_use_autosuspend(&spi->dev);
> + pm_runtime_disable(&spi->dev);
> drm_atomic_helper_shutdown(&panel->drm);
> }
>
> +static int ssd16xx_pm_suspend(struct device *dev)
> +{
> + struct drm_device *drm = dev_get_drvdata(dev);
> + struct ssd16xx_panel *panel = to_ssd16xx_panel(drm);
> + int ret;
> +
> + /* System suspend: Mode 2 (max savings, RAM lost, full re-init on resume). */
> + panel->deep_sleep_mode = panel->controller_cfg->deep_sleep_mode_level2;
> + dev_dbg(dev, "system suspend: entering deep sleep mode 0x%02x\n",
> + panel->deep_sleep_mode);
> +
> + ret = drm_mode_config_helper_suspend(drm);
> + if (ret)
> + return ret;
> +
> + /*
> + * With per-update PM gating the panel may already be in MODE_1 deep
> + * sleep (autosuspend fired while the user was reading).
> + * pm_runtime_force_suspend() skips its callback when the device is
> + * already RPM_SUSPENDED, so MODE_2 would never reach the hardware.
> + *
> + * Per the datasheet, HWRESET is the only way to exit deep sleep (SPI
> + * is unresponsive in both modes). After the reset, MODE_2 can be
> + * sent immediately — no full hw_init is required. The RPM state
> + * remains RPM_SUSPENDED; pm_runtime_force_suspend() is skipped since
> + * the device is already in the right state.
> + */
> + if (pm_runtime_status_suspended(dev)) {
> + int err = 0;
> +
> + dev_dbg(dev, "system suspend: upgrading MODE_1 -> MODE_2 via HWRESET\n");
> + ssd16xx_hw_reset(panel);
> + ssd16xx_send_cmd(panel, SSD16XX_CMD_DEEP_SLEEP_MODE, &err);
> + ssd16xx_send_data(panel, panel->controller_cfg->deep_sleep_mode_level2, &err);
> + panel->pm_force_suspended = false; /* Skip force_resume on resume */
> + return err;
> + }
> +
> + /* Force runtime-suspended state for clean resume. */
> + pm_runtime_force_suspend(dev);
> + panel->pm_force_suspended = true; /* Resume must call force_resume */
> +
> + return 0;
> +}
> +
> +static int ssd16xx_pm_resume(struct device *dev)
> +{
> + struct drm_device *drm = dev_get_drvdata(dev);
> + struct ssd16xx_panel *panel = to_ssd16xx_panel(drm);
> +
> + dev_dbg(dev, "system resume: restoring state\n");
> +
> + /*
> + * Only call force_resume if we actually called force_suspend during
> + * system suspend. The MODE_1->MODE_2 upgrade path skips force_suspend
> + * since the device is already RPM_SUSPENDED.
> + */
> + if (panel->pm_force_suspended) {
> + pm_runtime_force_resume(dev);
> + panel->pm_force_suspended = false;
> + }
> +
> + /* Restore Mode 1 (RAM retained) for subsequent runtime PM cycles. */
> + panel->deep_sleep_mode = panel->controller_cfg->deep_sleep_mode_level1;
> +
> + return drm_mode_config_helper_resume(drm);
> +}
> +
> +static int ssd16xx_pm_runtime_suspend(struct device *dev)
> +{
> + struct ssd16xx_panel *panel = to_ssd16xx_panel(dev_get_drvdata(dev));
> + int err = 0;
> +
> + dev_dbg(dev, "runtime suspend: entering deep sleep mode 0x%02x\n",
> + panel->deep_sleep_mode);
> +
> + ssd16xx_send_cmd(panel, SSD16XX_CMD_DEEP_SLEEP_MODE, &err);
> + ssd16xx_send_data(panel, panel->deep_sleep_mode, &err);
> + panel->initialized = false;
> +
> + return err;
> +}
> +
> +static int ssd16xx_pm_runtime_resume(struct device *dev)
> +{
> + struct ssd16xx_panel *panel = to_ssd16xx_panel(dev_get_drvdata(dev));
> + int ret;
> +
> + dev_dbg(dev, "runtime resume: initialized=%d%s\n",
> + panel->initialized,
> + !panel->initialized ? " (running hw_init)" : "");
> +
> + /*
> + * pm_runtime_suspend clears initialized after sending the deep sleep
> + * command. If a new app opens before the suspend fires (the put was
> + * cancelled), initialized is still true and hw_init is unnecessary.
> + */
> + if (!panel->initialized) {
> + ret = ssd16xx_hw_init(panel);
> + if (ret)
> + return ret;
> +
> + panel->initialized = true;
> + }
> +
> + return 0;
> +}
> +
> +static const struct dev_pm_ops ssd16xx_pm_ops = {
> + SYSTEM_SLEEP_PM_OPS(ssd16xx_pm_suspend, ssd16xx_pm_resume)
> + RUNTIME_PM_OPS(ssd16xx_pm_runtime_suspend, ssd16xx_pm_runtime_resume, NULL)
> +};
> +
> static const struct of_device_id ssd16xx_of_match[] = {
> { .compatible = "gooddisplay,gdey042t81", .data = (void *)GDEY042T81 },
> { }
> @@ -2535,6 +2717,7 @@ static struct spi_driver ssd16xx_spi_driver = {
> .driver = {
> .name = "ssd16xx",
> .of_match_table = ssd16xx_of_match,
> + .pm = pm_ptr(&ssd16xx_pm_ops),
> },
> .probe = ssd16xx_probe,
> .remove = ssd16xx_remove,
--
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
next prev parent reply other threads:[~2026-05-05 7:08 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-30 18:33 [PATCH 0/6] Add DRM driver for Solomon SSD16xx e-paper display controllers Devarsh Thakkar
2026-04-30 18:33 ` [PATCH 1/6] dt-bindings: vendor-prefixes: Add Dalian Good Display Co., Ltd Devarsh Thakkar
2026-05-04 10:18 ` Krzysztof Kozlowski
2026-04-30 18:33 ` [PATCH 2/6] dt-bindings/display: Add Solomon SSD16xx e-paper controller binding Devarsh Thakkar
2026-04-30 18:33 ` [PATCH 3/6] drm/tiny: Add DRM driver for Solomon SSD16xx e-paper display controllers Devarsh Thakkar
2026-05-05 7:05 ` Thomas Zimmermann
2026-04-30 18:33 ` [PATCH 4/6] drm/tiny: panel-ssd16xx: Add power management support Devarsh Thakkar
2026-05-05 7:08 ` Thomas Zimmermann [this message]
2026-04-30 18:33 ` [PATCH 5/6] MAINTAINERS: Add entry for Solomon SSD16xx DRM driver Devarsh Thakkar
2026-04-30 18:33 ` [PATCH 6/6] arm64: defconfig: Enable DRM_PANEL_SSD16XX Devarsh Thakkar
2026-05-01 9:30 ` Krzysztof Kozlowski
2026-05-05 6:25 ` Devarsh Thakkar
2026-05-05 12:42 ` Andrew Davis
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=af9a1cde-00be-49fd-b236-1cb2bfb4837e@suse.de \
--to=tzimmermann@suse.de \
--cc=afd@ti.com \
--cc=airlied@gmail.com \
--cc=bjorn.andersson@oss.qualcomm.com \
--cc=conor+dt@kernel.org \
--cc=devarsht@ti.com \
--cc=devicetree@vger.kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=krzk+dt@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=maarten.lankhorst@linux.intel.com \
--cc=mripard@kernel.org \
--cc=neil.armstrong@linaro.org \
--cc=praneeth@ti.com \
--cc=r-donadkar@ti.com \
--cc=r-sharma3@ti.com \
--cc=robh@kernel.org \
--cc=s-jain1@ti.com \
--cc=s-wang12@ti.com \
--cc=simona@ffwll.ch \
--cc=vigneshr@ti.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox