On Sat, Jan 21, 2012 at 12:41:56AM +0900, Akinobu Mita wrote:

2012/1/21 Konrad Rzeszutek Wilk [off-list ref]:

quoted

On Sat, Jan 21, 2012 at 12:15:26AM +0900, Akinobu Mita wrote:

quoted

Use bitmap_set and bitmap_clear rather than modifying individual bits
in a memory region.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Jeremy Fitzhardinge <redacted>
Cc: Konrad Rzeszutek Wilk <redacted>
Cc: xen-devel@lists.xensource.com
Cc: virtualization@lists.linux-foundation.org
---
 drivers/block/xen-blkfront.c |    7 +++----
 1 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 2f22874..619868d 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c

@@ -43,6 +43,7 @@

 #include <linux/slab.h>
 #include <linux/mutex.h>
 #include <linux/scatterlist.h>
+#include <linux/bitmap.h>

 #include <xen/xen.h>
 #include <xen/xenbus.h>

@@ -177,8 +178,7 @@ static int xlbd_reserve_minors(unsigned int minor, unsigned int nr)

      spin_lock(&minor_lock);
      if (find_next_bit(minors, end, minor) >= end) {
-             for (; minor < end; ++minor)
-                     __set_bit(minor, minors);
+             bitmap_set(minors, minor, nr);

Hm, I would have thought the last argument should have been 'end'?

'end' is the index of the last bit to clear.  But the last argument of
bitmap_clear() is the number of bits to clear.  So I think 'nr' is correct.

Ah, I see it.

quoted

Did you test this patch with a large amount of minors?

Sorry I didn't do runtime test.

Please do.