On Saturday 08 July 2006 16:24, [EMAIL PROTECTED] wrote:
> I recently thought about using coldplug (/etc/hotplug/*.rc) scripts to 
> automagically load modules responsible for ACPI-controlled devices such 
> as button, battery, laptop extras. To accomplish this I wrote a trivial 
> kernel patch which exports the ACPI device _HID and _CID attributes in 
> the corresponding sysfs node. I also wrote a script for /etc/hotplug 
> which scans the /sys tree and loads modules according to HIDs exported.

Funny, I was just thinking about this, too :-)  I think it's a good
idea to be able to automatically load the correct drivers for ACPI
devices.

But I think it'd be better for the kernel to expose ACPI devices to
user-space as PNP devices, and to use uevents for PNP devices just
like we do for PCI devices.  I don't like the idea of having separate
user-space stuff for PNPBIOS vs PNPACPI devices.

I don't know very much about udev.  Would the following patch, in
combination with CONFIG_PNPACPI, be enough to do what you want?

Index: work-mm7/drivers/pnp/driver.c
===================================================================
--- work-mm7.orig/drivers/pnp/driver.c  2006-07-09 20:51:31.000000000 -0600
+++ work-mm7/drivers/pnp/driver.c       2006-07-09 20:55:09.000000000 -0600
@@ -191,9 +191,56 @@
        return 0;
 }
 
+static int pnp_uevent(struct device *dev, char **envp, int num_envp,
+                     char *buffer, int buffer_size)
+{
+       struct pnp_dev *pdev;
+       struct pnp_id *pos;
+       int i = 0;
+       int length = 0;
+       int ids = 0;
+       char *id_buffer, *p;
+
+       if (!dev)
+               return -ENODEV;
+
+       pdev = to_pnp_dev(dev);
+       if (!pdev)
+               return -ENODEV;
+
+       for (pos = pdev->id; pos; pos = pos->next)
+               ids++;
+
+       id_buffer = kmalloc(ids * 9, GFP_KERNEL);
+       if (!id_buffer)
+               return -ENOMEM;
+
+       p = id_buffer;
+       for (pos = pdev->id; pos; pos = pos->next)
+               p += scnprintf(p, 9, "%s%c", pos->id, pos->next ? ':' : '\0');
+
+       if (add_uevent_var(envp, num_envp, &i,
+                          buffer, buffer_size, &length,
+                          "PNP_ID=%s", id_buffer)) {
+               kfree(id_buffer);
+               return -ENOMEM;
+       }
+
+       kfree(id_buffer);
+
+       if (add_uevent_var(envp, num_envp, &i,
+                          buffer, buffer_size, &length,
+                          "PNP_DEVICE_NODE=%s", dev->bus_id))
+               return -ENOMEM;
+
+       envp[i] = NULL;
+       return 0;
+}
+
 struct bus_type pnp_bus_type = {
        .name   = "pnp",
        .match  = pnp_bus_match,
+       .uevent = pnp_uevent,
        .probe  = pnp_device_probe,
        .remove = pnp_device_remove,
        .suspend = pnp_bus_suspend,
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to