[PATCH 3/6] mailbox: Add support for ST's Mailbox IP

On Thu, 23 Jul 2015, Alexey Klimov wrote:

On Fri, Jul 17, 2015 at 3:04 PM, Lee Jones [off-list ref] wrote:

quoted

ST's platforms currently support a maximum of 5 Mailboxes, one for
each of the supported co-processors situated on the platform.  Each
Mailbox is divided up into 4 instances which consist of 32 channels.
Messages are passed between the application and co-processors using
shared memory areas.  It is the Client's responsibility to manage
these areas.

Signed-off-by: Lee Jones <redacted>
---
 drivers/mailbox/Kconfig       |   7 +
 drivers/mailbox/Makefile      |   2 +
 drivers/mailbox/mailbox-sti.c | 562 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 571 insertions(+)
 create mode 100644 drivers/mailbox/mailbox-sti.c

[..]

quoted

+static irqreturn_t sti_mbox_thread_handler(int irq, void *data)
+{
+       struct sti_mbox_device *mdev = data;
+       struct sti_mbox_pdata *pdata = dev_get_platdata(mdev->dev);
+       struct mbox_chan *chan;
+       unsigned int instance;
+
+       for (instance = 0; instance < pdata->num_inst; instance++) {
+keep_looking:
+               chan = sti_mbox_irq_to_channel(mdev, instance);
+               if (!chan)
+                       continue;
+
+               mbox_chan_received_data(chan, NULL);
+               sti_mbox_clear_irq(chan);
+               sti_mbox_enable_channel(chan);
+               goto keep_looking;
+       }
+
+       return IRQ_HANDLED;
+}
+
+static irqreturn_t sti_mbox_irq_handler(int irq, void *data)
+{
+       struct sti_mbox_device *mdev = data;
+       struct sti_mbox_pdata *pdata = dev_get_platdata(mdev->dev);
+       struct sti_channel *chan_info;
+       struct mbox_chan *chan;
+       unsigned int instance;
+       int ret = IRQ_NONE;
+
+       for (instance = 0; instance < pdata->num_inst; instance++) {
+               chan = sti_mbox_irq_to_channel(mdev, instance);
+               if (!chan)
+                       continue;
+               chan_info = chan->con_priv;
+
+               if (!sti_mbox_channel_is_enabled(chan)) {
+                       dev_warn(mdev->dev,
+                                "Unexpected IRQ: %s\n"
+                                "  instance: %d: channel: %d [enabled: %x]\n",
+                                mdev->name, chan_info->instance,
+                                chan_info->channel, mdev->enabled[instance]);
+                       ret = IRQ_HANDLED;
+                       continue;
+               }
+
+               sti_mbox_disable_channel(chan);
+               ret = IRQ_WAKE_THREAD;
+       }
+
+       if (ret == IRQ_NONE)
+               dev_err(mdev->dev, "Spurious IRQ - was a channel requested?\n");
+
+       return ret;
+}

With such usage of ret variable can it happen that handling of last
but one channel/instance will set ret to IRQ_WAKE_THREAD and at the
same time handling of last channel/instance will set ret to
IRQ_HANDLED during iteration loop and finally generic subsystem will
not wake up thread handler because it will receive IRQ_HANDLED?
Just checking.

Yes, I guess that it theoretically possible.  Now fixed.

Thanks for the spot.

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog