[PATCH][0/2] [RESEND] Hotplug variable patches

[PATCH][0/2] [RESEND] Hotplug variable patches

[Roland Dreier]

(Resending because this seems to have gotten dropped)

Hi Greg,

When I wanted to implement some environment variables in my hotplug
method, I looked for an example of how to do it.  I noticed two
things: adding values ends up being kind of messy, and the hotplug
method in drivers/usb/core/usb.c is subtly wrong!  These two patches
attempt to fix both of those problems.

Let me know if you don't like the HOTPLUG_ENV_VAR name and want
something different.

If you apply these, I'll send patches to use HOTPLUG_ENV_VAR in net/,
drivers/ieee1394/, etc.

Thanks,
  Roland

[PATCH][1/2] [RESEND] kobject: add HOTPLUG_ENV_VAR

[Roland Dreier]

Add a HOTPLUG_ENV_VAR macro to <linux/kobject.h>.  There's a lot of
boilerplate code involved in setting environment variables in a
hotplug method, so we should have a convenience macro to consolidate
it (and avoid subtle bugs).


Signed-off-by: Roland Dreier <roland@topspin.com>

Index: linux-bk/include/linux/kobject.h
===================================================================
--- linux-bk.orig/include/linux/kobject.h	2004-09-12 15:17:02.000000000 -0700
+++ linux-bk/include/linux/kobject.h	2004-09-12 19:44:31.000000000 -0700
@@ -95,6 +95,20 @@
 			int num_envp, char *buffer, int buffer_size);
 };
 
+#define HOTPLUG_ENV_VAR(_buf, _size, _envp, _nenvp, _index, _fmt, _arg...) \
+	({								\
+		int len, ret = 0;					\
+		_envp[_index++] = _buf;					\
+		len = snprintf(_buf, _size, _fmt, ## _arg);		\
+		if (_size - len <= 0 || _index >= _nenvp)		\
+			ret = -ENOMEM;					\
+		else {							\
+			_buf  += len + 1;				\
+			_size -= len + 1;				\
+		}							\
+		ret;							\
+	})
+
 struct kset {
 	struct subsystem	* subsys;
 	struct kobj_type	* ktype;

[PATCH][2/2] [RESEND] USB: use HOTPLUG_ENV_VAR in core/usb.c

[Roland Dreier]

Use the new HOTPLUG_ENV_VAR macro is drivers/usb/core/usb.c.  In
addition to cleaning up the code, this fixes a (probably harmless) bug
here: for each value added to the environment, the code did

	length += sprintf(...);

and then

	scratch += length;

which means that we skip the sum of the lengths of all the values
we've put so far, rather than just the length of the value we just
put.  This is probably harmless since we're unlikely to run out of
space but if nothing else it's setting a bad example....

I've tested this on a system with USB floppy and CD-ROM; hotplug gets
the same environment with the patch as without.


Signed-off-by: Roland Dreier <roland@topspin.com>

Index: linux-bk/drivers/usb/core/usb.c
===================================================================
--- linux-bk.orig/drivers/usb/core/usb.c	2004-09-12 15:21:27.000000000 -0700
+++ linux-bk/drivers/usb/core/usb.c	2004-09-12 19:28:23.000000000 -0700
@@ -564,9 +564,7 @@
 {
 	struct usb_interface *intf;
 	struct usb_device *usb_dev;
-	char *scratch;
 	int i = 0;
-	int length = 0;
 
 	if (!dev)
 		return -ENODEV;
@@ -591,8 +589,6 @@
 		return -ENODEV;
 	}
 
-	scratch = buffer;
-
 #ifdef	CONFIG_USB_DEVICEFS
 	/* If this is available, userspace programs can directly read
 	 * all the device descriptors we don't tell them about.  Or
@@ -600,37 +596,27 @@
 	 *
 	 * FIXME reduce hardwired intelligence here
 	 */
-	envp [i++] = scratch;
-	length += snprintf (scratch, buffer_size - length,
+	if (HOTPLUG_ENV_VAR(buffer, buffer_size, envp, num_envp, i,
 			    "DEVICE=/proc/bus/usb/%03d/%03d",
-			    usb_dev->bus->busnum, usb_dev->devnum);
-	if ((buffer_size - length <= 0) || (i >= num_envp))
+			    usb_dev->bus->busnum, usb_dev->devnum))
 		return -ENOMEM;
-	++length;
-	scratch += length;
 #endif
 
 	/* per-device configurations are common */
-	envp [i++] = scratch;
-	length += snprintf (scratch, buffer_size - length, "PRODUCT=%x/%x/%x",
+	if (HOTPLUG_ENV_VAR(buffer, buffer_size, envp, num_envp, i,
+			    "PRODUCT=%x/%x/%x",
 			    usb_dev->descriptor.idVendor,
 			    usb_dev->descriptor.idProduct,
-			    usb_dev->descriptor.bcdDevice);
-	if ((buffer_size - length <= 0) || (i >= num_envp))
+			    usb_dev->descriptor.bcdDevice))
 		return -ENOMEM;
-	++length;
-	scratch += length;
 
 	/* class-based driver binding models */
-	envp [i++] = scratch;
-	length += snprintf (scratch, buffer_size - length, "TYPE=%d/%d/%d",
+	if (HOTPLUG_ENV_VAR(buffer, buffer_size, envp, num_envp, i,
+			    "TYPE=%d/%d/%d",
 			    usb_dev->descriptor.bDeviceClass,
 			    usb_dev->descriptor.bDeviceSubClass,
-			    usb_dev->descriptor.bDeviceProtocol);
-	if ((buffer_size - length <= 0) || (i >= num_envp))
+			    usb_dev->descriptor.bDeviceProtocol))
 		return -ENOMEM;
-	++length;
-	scratch += length;
 
 	if (usb_dev->descriptor.bDeviceClass == 0) {
 		struct usb_host_interface *alt = intf->cur_altsetting;
@@ -639,18 +625,14 @@
 		 * agents are called for all interfaces, and can use
 		 * $DEVPATH/bInterfaceNumber if necessary.
 		 */
-		envp [i++] = scratch;
-		length += snprintf (scratch, buffer_size - length,
-			    "INTERFACE=%d/%d/%d",
-			    alt->desc.bInterfaceClass,
-			    alt->desc.bInterfaceSubClass,
-			    alt->desc.bInterfaceProtocol);
-		if ((buffer_size - length <= 0) || (i >= num_envp))
+		if (HOTPLUG_ENV_VAR(buffer, buffer_size, envp, num_envp, i,
+				    "INTERFACE=%d/%d/%d",
+				    alt->desc.bInterfaceClass,
+				    alt->desc.bInterfaceSubClass,
+				    alt->desc.bInterfaceProtocol))
 			return -ENOMEM;
-		++length;
-		scratch += length;
-
 	}
+
 	envp[i++] = NULL;
 
 	return 0;

[PATCH][0/2] [RESEND] Hotplug variable patches

[Roland Dreier]

(Resending because this seems to have gotten dropped)

Hi Greg,

When I wanted to implement some environment variables in my hotplug
method, I looked for an example of how to do it.  I noticed two
things: adding values ends up being kind of messy, and the hotplug
method in drivers/usb/core/usb.c is subtly wrong!  These two patches
attempt to fix both of those problems.

Let me know if you don't like the HOTPLUG_ENV_VAR name and want
something different.

If you apply these, I'll send patches to use HOTPLUG_ENV_VAR in net/,
drivers/ieee1394/, etc.

Thanks,
  Roland