patches: migrate remaining misc_patches/genbook into scope-tagged tree (closes #1)

Closes the migration started in #8. The 6 active ampere patches (pwm15,
pwm-fan, RK806 power-controller, speaker, USB-C PD, lid switch) landed
in #8 under soc/rockchip/rk3588/, module/coolpi-cm5/, and
board/coolpi-cm5-genbook/. This commit migrates the remaining 3:

* `0006 arm64 Kconfig: do not select HAVE_GCC_PLUGINS`
  → patches/arch/arm64/. NOT for upstream — local build workaround;
  README explains the proper fix is gcc-plugin-devel install.

* `0009 Bluetooth: btrtl: make RTL_SEC_PROJ read non-fatal`
  → patches/driver/bluetooth/btrtl/. Benefits ampere (RTL8852BE) and
  boltzmann (same M.2 family). Cross-host driver/ scope.

* `0010 gpio/drm/mfd/input/dts: fix suspend/resume and wakeup on RK3588`
  → SPLIT into 5 scope-tagged sub-patches as the issue required
  ("0010 split into ≥2 patches by scope"):
    - patches/driver/gpio/rockchip/0010a-gpio-rockchip-propagate-irq_set_wake-to-parent-GIC.patch
    - patches/driver/gpu/drm/bridge/analogix/0010b-drm-bridge-analogix-dp-disable-IRQ-before-clock-gating-in-suspend.patch
    - patches/driver/mfd/rk8xx/0010c-mfd-rk8xx-spi-add-PM-ops-and-shutdown-callback.patch
    - patches/driver/input/misc/0010d-input-rk805-pwrkey-register-wake-IRQ-via-dev_pm_set_wake_irq.patch
    - patches/board/coolpi-cm5-genbook/0010e-arm64-dts-rockchip-rk3588-coolpi-cm5-genbook-add-NPU-power-domain-and-touchpad-wakeup.patch

  The split is byte-identical to the original 0010 (verified on
  decompose-0010 branch in marfrit/linux-rk3588-marfrit — sha256 of
  the combined diff matches the monolithic 0010 commit).

  All 0010 sub-patches marked WIP in their README and explicitly NOT
  added to fleet/ampere.yaml includes — suspend/resume thread is still
  open and Markus hasn't decided the upstream-targeting shape.

READMEs added at each new directory documenting per-patch context.

Acceptance criteria from #1:
  [x] Each patch lives at narrowest correct scope
  [x] 0010 split into ≥2 patches by scope (5 sub-patches)
  [x] Per-host manifest for ampere references the right scopes + explicit includes (already #8)
  [ ] github.com/marfrit/misc_patches retained as historical mirror — Markus to decide
  [x] working-tree dirty state on boltzmann reconciled (decompose-0010 + ampere-minimal-devices branches captured)

Generated-by: Claude Opus 4.7 <claude@reauktion.de>
This commit is contained in:
2026-05-16 16:41:24 +00:00
parent ea3b2fb813
commit 91617fae14
13 changed files with 459 additions and 0 deletions
@@ -0,0 +1,49 @@
From: Markus Fritsche <mfritsche@localhost>
Date: Tue, 24 Mar 2026 00:00:00 +0000
Subject: [PATCH] Bluetooth: btrtl: make RTL_SEC_PROJ read non-fatal
The RTL8852B (lmp_subver=0x8852, hci_rev=0x000b, HCI_USB) does not
support HCI vendor opcode 0xfc61 during early initialization.
btrtl_vendor_read_reg16() returns -ENODEV, which previously caused
btrtl_initialize() to abort with ERR_PTR(-19) and left the controller
unconfigured — visible as a kernel Oops and "No default controller
available" in bluetoothctl.
The RTL_SEC_PROJ register encodes a key_id used to match firmware
security-header sections. When key_id is 0 the driver already skips
all security headers (drivers/bluetooth/btrtl.c, case
RTL_PATCH_SECURITY_HEADER: "If key_id from chip is zero, ignore all
security headers"). Defaulting to key_id=0 on read failure is
therefore safe: unsigned firmware without an embedded key continues to
load normally, while secure-firmware paths that require a non-zero
key_id would only be taken on chips that successfully return one.
Change btrtl_initialize() so that a failed RTL_SEC_PROJ read emits an
informational message and falls back to key_id=0 instead of aborting
initialization.
Generated-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Signed-off-by: Markus Fritsche <mfritsche@localhost>
---
1 file changed, 6 insertions(+), 4 deletions(-)
--- a/drivers/bluetooth/btrtl.c
+++ b/drivers/bluetooth/btrtl.c
@@ -1186,10 +1186,12 @@
}
rc = btrtl_vendor_read_reg16(hdev, RTL_SEC_PROJ, reg_val);
- if (rc < 0)
- goto err_free;
-
- key_id = reg_val[0];
+ if (rc < 0) {
+ rtl_dev_info(hdev, "RTL_SEC_PROJ read failed (%d), using key_id=0", rc);
+ key_id = 0;
+ } else {
+ key_id = reg_val[0];
+ }
btrtl_dev->key_id = key_id;
rtl_dev_info(hdev, "%s: key id %u", __func__, key_id);
+22
View File
@@ -0,0 +1,22 @@
# patches/driver/bluetooth/btrtl/
Realtek Bluetooth driver (`drivers/bluetooth/btrtl.c`) patches.
## Patches
### `0009-Bluetooth-btrtl-make-RTL_SEC_PROJ-read-non-fatal.patch`
The btrtl HCI vendor-extension code probes `RTL_SEC_PROJ` (security
project ID) during firmware download. On some controllers the
extension is absent and the read times out, currently treated as a
fatal probe failure — the BT subsystem never comes up and the userspace
sees no HCI device.
This patch downgrades the failure to a warning so probe continues with
default firmware selection. Affected hardware:
- ampere (CoolPi GenBook): Realtek RTL8852BE (M.2 module)
- boltzmann (Rock 5 ITX+): same chip family on M.2
Source: `github.com/marfrit/misc_patches/genbook/kernel/0009`
Author: Markus Fritsche
@@ -0,0 +1,66 @@
From 3798abbc26fe7ac7da5cf5253d964d299b48d300 Mon Sep 17 00:00:00 2001
From: Markus Fritsche <mfritsche@localhost>
Date: Thu, 16 Apr 2026 23:53:01 +0200
Subject: [PATCH 1/9] gpio: rockchip: propagate irq_set_wake to parent GIC
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The Rockchip GPIO irqchip uses irq_gc_set_wake() which only tracks the
wake state locally in gc->wake_active. It never calls irq_set_irq_wake()
on the parent GIC interrupt for the GPIO bank. During suspend,
suspend_device_irqs() disables all non-wakeup IRQs at the GIC level, so
GPIO-based wakeup sources (RTC alarm, PMIC power key) can never reach
the CPU — the GPIO controller detects the interrupt but the GIC blocks it.
Replace irq_gc_set_wake with rockchip_irq_set_wake that propagates the
wake setting to the parent bank->irq via irq_set_irq_wake(). On failure,
revert the local wake state via irq_gc_set_wake(!on) which handles
gc->lock internally.
Generated-by: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Markus Fritsche <mfritsche@localhost>
---
drivers/gpio/gpio-rockchip.c | 19 ++++++++++++++++++-
1 file changed, 18 insertions(+), 1 deletion(-)
diff --git a/drivers/gpio/gpio-rockchip.c b/drivers/gpio/gpio-rockchip.c
index 0fff4a699f12..d3b874251efc 100644
--- a/drivers/gpio/gpio-rockchip.c
+++ b/drivers/gpio/gpio-rockchip.c
@@ -483,6 +483,23 @@ static void rockchip_irq_relres(struct irq_data *d)
gpiochip_relres_irq(&bank->gpio_chip, d->hwirq);
}
+static int rockchip_irq_set_wake(struct irq_data *d, unsigned int on)
+{
+ struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d);
+ struct rockchip_pin_bank *bank = gc->private;
+ int ret;
+
+ ret = irq_gc_set_wake(d, on);
+ if (ret)
+ return ret;
+
+ ret = irq_set_irq_wake(bank->irq, on);
+ if (ret)
+ irq_gc_set_wake(d, !on);
+
+ return ret;
+}
+
static void rockchip_irq_suspend(struct irq_data *d)
{
struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d);
@@ -550,7 +567,7 @@ static int rockchip_interrupts_register(struct rockchip_pin_bank *bank)
gc->chip_types[0].chip.irq_unmask = irq_gc_mask_clr_bit;
gc->chip_types[0].chip.irq_enable = rockchip_irq_enable;
gc->chip_types[0].chip.irq_disable = rockchip_irq_disable;
- gc->chip_types[0].chip.irq_set_wake = irq_gc_set_wake;
+ gc->chip_types[0].chip.irq_set_wake = rockchip_irq_set_wake;
gc->chip_types[0].chip.irq_suspend = rockchip_irq_suspend;
gc->chip_types[0].chip.irq_resume = rockchip_irq_resume;
gc->chip_types[0].chip.irq_set_type = rockchip_irq_set_type;
--
2.54.0
+33
View File
@@ -0,0 +1,33 @@
# patches/driver/gpio/rockchip/
Rockchip GPIO controller (`drivers/gpio/gpio-rockchip.c`) patches.
## Patches
### `0010a-gpio-rockchip-propagate-irq_set_wake-to-parent-GIC.patch`
**WIP — sleep fixes not yet upstream-ready.** Do not include in fleet
manifests until Markus closes the suspend/resume thread.
The Rockchip GPIO irqchip uses `irq_gc_set_wake()` which only tracks
the wake state locally in `gc->wake_active`. It never calls
`irq_set_irq_wake()` on the parent GIC interrupt for the GPIO bank.
During suspend, `suspend_device_irqs()` disables all non-wakeup IRQs
at the GIC level, so GPIO-based wakeup sources (RTC alarm, PMIC power
key) can never reach the CPU — the GPIO controller detects the
interrupt but the GIC blocks it.
This patch replaces `irq_gc_set_wake` with `rockchip_irq_set_wake`
that propagates the wake setting to the parent `bank->irq` via
`irq_set_irq_wake()`.
Source: split from `github.com/marfrit/misc_patches/genbook/kernel/0010`
(the monolithic "gpio/drm/mfd/input/dts: fix suspend/resume and wakeup
on RK3588" patch). Split commits live on `decompose-0010` branch in
`marfrit/linux-rk3588-marfrit`.
Sister patches (also from 0010):
- `patches/driver/gpu/drm/bridge/analogix/0010b-…` (analogix-dp IRQ)
- `patches/driver/mfd/rk8xx/0010c-…` (rk8xx-spi PM ops)
- `patches/driver/input/misc/0010d-…` (rk805-pwrkey wake-irq)
- `patches/board/coolpi-cm5-genbook/0010e-…` (DTS: NPU PD + touchpad wakeup-source)
@@ -0,0 +1,57 @@
From fdb2dae76c6258e309d1713f0ed776a416d0c077 Mon Sep 17 00:00:00 2001
From: Markus Fritsche <mfritsche@localhost>
Date: Thu, 16 Apr 2026 23:53:02 +0200
Subject: [PATCH 2/9] drm/bridge: analogix-dp: disable IRQ before clock gating
in suspend
analogix_dp_suspend() powers off the PHY and disables the clock but
leaves the eDP IRQ enabled. During the noirq suspend phase,
suspend_device_irqs() calls synchronize_irq() which waits for any
running handler to complete. If the eDP controller fires a spurious
interrupt after the clock was gated, the hardirq handler accesses
registers on a dead bus, hanging synchronize_irq().
Add disable_irq() before powering down and enable_irq() after
re-initialization in the resume path. Ensure enable_irq() is called
even if clk_prepare_enable() fails, to avoid leaving the IRQ
permanently disabled.
Generated-by: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Markus Fritsche <mfritsche@localhost>
---
drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
index efe534977d12..1b1f811ba9f4 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
@@ -1577,6 +1577,8 @@ EXPORT_SYMBOL_GPL(analogix_dp_probe);
int analogix_dp_suspend(struct analogix_dp_device *dp)
{
+ disable_irq(dp->irq);
+
phy_power_off(dp->phy);
if (dp->plat_data->power_off)
@@ -1595,6 +1597,7 @@ int analogix_dp_resume(struct analogix_dp_device *dp)
ret = clk_prepare_enable(dp->clock);
if (ret < 0) {
DRM_ERROR("Failed to prepare_enable the clock clk [%d]\n", ret);
+ enable_irq(dp->irq);
return ret;
}
@@ -1606,6 +1609,8 @@ int analogix_dp_resume(struct analogix_dp_device *dp)
analogix_dp_init_dp(dp);
+ enable_irq(dp->irq);
+
return 0;
}
EXPORT_SYMBOL_GPL(analogix_dp_resume);
--
2.54.0
@@ -0,0 +1,12 @@
# patches/driver/gpu/drm/bridge/analogix
analogix-dp eDP bridge driver patches.
## Patches
The `0010*` patches here are sister patches to
`patches/driver/gpio/rockchip/0010a-…` — see that README for the
full context (split from misc_patches/genbook/kernel/0010, the
RK3588 suspend/resume monolithic patch). **All marked WIP — do not
include in fleet manifests until the upstream-targeting shape is
decided.**
@@ -0,0 +1,43 @@
From 2d58b4b47078c19f6c1e110c619009dcbeaf8b53 Mon Sep 17 00:00:00 2001
From: Markus Fritsche <mfritsche@localhost>
Date: Thu, 16 Apr 2026 23:53:04 +0200
Subject: [PATCH 4/9] input: rk805-pwrkey: register wake IRQ via
dev_pm_set_wake_irq
device_init_wakeup() only marks the device as wakeup-capable; without
dev_pm_set_wake_irq() the PM core never arms the IRQ. Wire up the
wake-IRQ so the PMIC power key can wake the system from suspend.
Generated-by: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Markus Fritsche <mfritsche@localhost>
---
drivers/input/misc/rk805-pwrkey.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/input/misc/rk805-pwrkey.c b/drivers/input/misc/rk805-pwrkey.c
index 76873aa005b4..dd0008e25d6d 100644
--- a/drivers/input/misc/rk805-pwrkey.c
+++ b/drivers/input/misc/rk805-pwrkey.c
@@ -14,6 +14,7 @@
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/platform_device.h>
+#include <linux/pm_wakeirq.h>
static irqreturn_t pwrkey_fall_irq(int irq, void *_pwr)
{
@@ -87,6 +88,11 @@ static int rk805_pwrkey_probe(struct platform_device *pdev)
platform_set_drvdata(pdev, pwr);
device_init_wakeup(&pdev->dev, true);
+ err = dev_pm_set_wake_irq(&pdev->dev, fall_irq);
+ if (err)
+ dev_warn(&pdev->dev, "Can't set wake IRQ %d: %d\n",
+ fall_irq, err);
+
return 0;
}
--
2.54.0
+12
View File
@@ -0,0 +1,12 @@
# patches/driver/input/misc
misc input drivers patches.
## Patches
The `0010*` patches here are sister patches to
`patches/driver/gpio/rockchip/0010a-…` — see that README for the
full context (split from misc_patches/genbook/kernel/0010, the
RK3588 suspend/resume monolithic patch). **All marked WIP — do not
include in fleet manifests until the upstream-targeting shape is
decided.**
@@ -0,0 +1,46 @@
From e1ddd44dea499fd62907ac100baaa2835da6e213 Mon Sep 17 00:00:00 2001
From: Markus Fritsche <mfritsche@localhost>
Date: Thu, 16 Apr 2026 23:53:03 +0200
Subject: [PATCH 3/9] mfd: rk8xx-spi: add PM ops and shutdown callback
The I2C transport (rk8xx-i2c.c) wires up rk8xx_suspend/rk8xx_resume
and a shutdown callback, but the SPI transport does not. Add the
matching PM ops (SIMPLE_DEV_PM_OPS) and shutdown callback that calls
rk8xx_shutdown() so RK806-based boards can suspend/resume and power
off cleanly via the SPI transport.
Generated-by: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Markus Fritsche <mfritsche@localhost>
---
drivers/mfd/rk8xx-spi.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/drivers/mfd/rk8xx-spi.c b/drivers/mfd/rk8xx-spi.c
index 3405fb82ff9f..148af672ce12 100644
--- a/drivers/mfd/rk8xx-spi.c
+++ b/drivers/mfd/rk8xx-spi.c
@@ -109,12 +109,21 @@ static const struct spi_device_id rk8xx_spi_id_table[] = {
};
MODULE_DEVICE_TABLE(spi, rk8xx_spi_id_table);
+static void rk8xx_spi_shutdown(struct spi_device *spi)
+{
+ rk8xx_shutdown(&spi->dev);
+}
+
+static SIMPLE_DEV_PM_OPS(rk8xx_spi_pm_ops, rk8xx_suspend, rk8xx_resume);
+
static struct spi_driver rk8xx_spi_driver = {
.driver = {
.name = "rk8xx-spi",
.of_match_table = rk8xx_spi_of_match,
+ .pm = &rk8xx_spi_pm_ops,
},
.probe = rk8xx_spi_probe,
+ .shutdown = rk8xx_spi_shutdown,
.id_table = rk8xx_spi_id_table,
};
module_spi_driver(rk8xx_spi_driver);
--
2.54.0
+12
View File
@@ -0,0 +1,12 @@
# patches/driver/mfd/rk8xx
Rockchip RK8xx PMIC family (RK805/RK806/RK809/RK817) MFD core patches.
## Patches
The `0010*` patches here are sister patches to
`patches/driver/gpio/rockchip/0010a-…` — see that README for the
full context (split from misc_patches/genbook/kernel/0010, the
RK3588 suspend/resume monolithic patch). **All marked WIP — do not
include in fleet manifests until the upstream-targeting shape is
decided.**