On Mon, Aug 17, 2015 at 3:52 PM, Yingjoe Chen [off-list ref] wrote:

On Fri, 2015-08-14 at 16:38 +0800, maoguang.meng at mediatek.com wrote:

quoted

From: Maoguang Meng <redacted>

This patch implement irq_set_wake to get who is wakeup source and
setup on suspend resume.

Signed-off-by: Maoguang Meng <redacted>

---
changes since v3:
-add a comment in mtk_eint_chip_read_mask.
-delete ALIGN when allocate eint_offsets.ports.
-fix unrelated change.

changes since v2:
-modify irq_wake to handle irq wakeup source.
-allocate two buffers separately.
-fix some codestyle.

Changes since v1:
-implement irq_wake handler.
---

 drivers/pinctrl/mediatek/pinctrl-mt8173.c     |  1 +
 drivers/pinctrl/mediatek/pinctrl-mtk-common.c | 91 ++++++++++++++++++++++++++-
 drivers/pinctrl/mediatek/pinctrl-mtk-common.h |  4 ++
 3 files changed, 95 insertions(+), 1 deletion(-)

diff --git a/drivers/pinctrl/mediatek/pinctrl-mt8173.c b/drivers/pinctrl/mediatek/pinctrl-mt8173.c
index d0c811d..ad27184 100644
--- a/drivers/pinctrl/mediatek/pinctrl-mt8173.c
+++ b/drivers/pinctrl/mediatek/pinctrl-mt8173.c

@@ -385,6 +385,7 @@ static struct platform_driver mtk_pinctrl_driver = {
      .driver = {
              .name = "mediatek-mt8173-pinctrl",
              .of_match_table = mt8173_pctrl_match,
+             .pm = &mtk_eint_pm_ops,
      },
 };

diff --git a/drivers/pinctrl/mediatek/pinctrl-mtk-common.c b/drivers/pinctrl/mediatek/pinctrl-mtk-common.c
index ad1ea16..fe34ce9 100644
--- a/drivers/pinctrl/mediatek/pinctrl-mtk-common.c
+++ b/drivers/pinctrl/mediatek/pinctrl-mtk-common.c

@@ -33,6 +33,7 @@
 #include <linux/mfd/syscon.h>
 #include <linux/delay.h>
 #include <linux/interrupt.h>
+#include <linux/pm.h>
 #include <dt-bindings/pinctrl/mt65xx.h>

 #include "../core.h"

@@ -1062,6 +1063,77 @@ static int mtk_eint_set_type(struct irq_data *d,
      return 0;
 }

+static int mtk_eint_irq_set_wake(struct irq_data *d, unsigned int on)
+{
+     struct mtk_pinctrl *pctl = irq_data_get_irq_chip_data(d);
+     int shift = d->hwirq & 0x1f;
+     int reg = d->hwirq >> 5;
+
+     if (on)
+             pctl->wake_mask[reg] |= BIT(shift);
+     else
+             pctl->wake_mask[reg] &= ~BIT(shift);
+
+     return 0;
+}

Hi Maoguang,

You changed from set_bit/clear_bit to this, but didn't add any locking.
Since this is basic read/modify/write, is it OK to do it without
locking?

I believe calling .irq_set_wake() concurrently with itself is
protected by irq_get_desc_buslock():

 int irq_set_irq_wake(unsigned int irq, unsigned int on)
 {
         unsigned long flags;
         struct irq_desc *desc = irq_get_desc_buslock(irq, &flags,
IRQ_GET_DESC_CHECK_GLOBAL);
         int ret = 0;
...
                         ret = set_irq_wake_real(irq, on);
...
         irq_put_desc_busunlock(desc, flags);
         return ret;
 }


I'm not 100% sure about the .suspend/.resume paths, but I don't think
they can occur during .irq_set_wake(), either.
Nor were they protected by the set_bit/clear_bit implementation.

-Dan

Joe.C


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/