On 08/23/2011 11:15 AM, Ali Bahar wrote:
Hi Larry,
On Tue, Aug 23, 2011 at 10:38:57AM -0500, Larry Finger wrote:
On 08/23/2011 12:53 AM, Ali Bahar wrote:
Realtek's latest is v2.6.6. A few types and definitions have been updated in
these files.
But rtl871x_recv.h's addr_t change has not been accepted: it will break on
64-bit platforms.
Signed-off-by: Ali Bahar<ali@xxxxxxxxxxxxxxx>
index c48757f..3c766c0 100644
--- a/drivers/staging/rtl8712/rtl8712_recv.h
+++ b/drivers/staging/rtl8712/rtl8712_recv.h
-- snip --
+ ******************************************************************************/
#ifndef _RTL8712_RECV_H_
#define _RTL8712_RECV_H_
#include "osdep_service.h"
#include "drv_types.h"
-#define NR_RECVBUFF (8)
+#define NR_RECVBUFF (4)
Realtek may have made this change, but I'm not sure it should be
done here. Under heavy network and CPU loads, I'm not sure that 8
receive buffers would be enough, but cutting it to 4 seems unwise.
Perhaps an embedded system with limited RAM would find that change
useful. I propose leaving it at 8, but add a comment.
OK.
#define NR_PREALLOC_RECV_SKB (8)
#define RXDESC_SIZE 24
#define RXDESC_OFFSET RXDESC_SIZE
@@ -115,7 +140,7 @@ union recv_frame {
union {
struct list_head list;
struct recv_frame_hdr hdr;
- addr_t mem[RECVFRAME_HDR_ALIGN>>2];
+ uint mem[RECVFRAME_HDR_ALIGN>>2];
Does this work with 64-bit systems? I have not tested, but changing
from addr_t to uint looks wrong.
Yeah, I spent a fair bit of time looking at this, back then.
The key point is that 'mem' is not used anymore! The function which
used it, mp_ioctl_xmit_packet_hdl, has been gutted.
That aside, I checked it for size; but .hdr is still its biggest
member. So the size has not been affected.
And I did not think of any unorthodox manipulations of the other
members which may choke on this change. So I did not go thru all the
OK, but if mem is not used anymore. should it be deleted?
Larry
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
