Re: [PATCH] CDRW packet writing support for 2.6.7-bk13

[Christoph Hellwig <hch@infradead.org>]

On Sun, Jul 04, 2004 at 01:58:42PM -0700, Andrew Morton wrote:
> Peter Osterlund <petero2@telia.com> wrote:
> >
> > But anyway, if __bdevname() leaks a module reference it should get
> >  fixed, right?
> 
> Yes.  The questions is, where's the bug?

in __bdevname.  It shouldn't try the get_gendisk.  I introduced that long
ago, but it's wrong.  Use bdevname where you can, else we can't do much
else than printing the namjor/minor sanely.

The packet writing code should always use bdevname() (see my previous
mail on how to get there), and here's a patch to fix __bdevname:


--- 1.124/fs/partitions/check.c	2004-06-18 08:43:53 +02:00
+++ edited/fs/partitions/check.c	2004-07-05 01:04:11 +02:00
@@ -137,25 +137,14 @@
 EXPORT_SYMBOL(bdevname);
 
 /*
- * NOTE: this cannot be called from interrupt context.
- *
- * But in interrupt context you should really have a struct
- * block_device anyway and use bdevname() above.
+ * There's very little reason to use this, you should really
+ * have a struct block_device just about everywhere and use
+ * bdevname() instead.
  */
 const char *__bdevname(dev_t dev, char *buffer)
 {
-	struct gendisk *disk;
-	int part;
-
-	disk = get_gendisk(dev, &part);
-	if (disk) {
-		buffer = disk_name(disk, part, buffer);
-		put_disk(disk);
-	} else {
-		snprintf(buffer, BDEVNAME_SIZE, "unknown-block(%u,%u)",
+	snprintf(buffer, BDEVNAME_SIZE, "unknown-block(%u,%u)",
 				MAJOR(dev), MINOR(dev));
-	}
-
 	return buffer;
 }