Hi Bin,
On 04.02.2017 06:57, Bin Meng wrote:
On Mon, Jan 23, 2017 at 10:55 PM, Stefan Roese <s...@denx.de> wrote:
Without configuring these registers in the SPI controller, the Linux
MTD device driver is not able to correctly read/write to the SPI
NOR chip at all. In fact, the chip is not detected at all.
Signed-off-by: Stefan Roese <s...@denx.de>
Cc: Bin Meng <bmeng...@gmail.com>
Cc: Simon Glass <s...@chromium.org>
Cc: Jagan Teki <jt...@openedev.com>
---
v2:
- Moved code into the ICH SPI driver as suggested by Simon and Bin
drivers/spi/ich.c | 51 +++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 51 insertions(+)
diff --git a/drivers/spi/ich.c b/drivers/spi/ich.c
index caf0103dc3..586b4e9024 100644
--- a/drivers/spi/ich.c
+++ b/drivers/spi/ich.c
@@ -18,6 +18,39 @@
#include "ich.h"
+#define SPI_OPMENU_0 0x01 /* WRSR: Write Status Register */
+#define SPI_OPTYPE_0 0x01 /* Write, no address */
+
+#define SPI_OPMENU_1 0x02 /* BYPR: Byte Program */
+#define SPI_OPTYPE_1 0x03 /* Write, address required */
+
+#define SPI_OPMENU_2 0x03 /* READ: Read Data */
+#define SPI_OPTYPE_2 0x02 /* Read, address required */
+
+#define SPI_OPMENU_3 0x05 /* RDSR: Read Status Register */
+#define SPI_OPTYPE_3 0x00 /* Read, no address */
+
+#define SPI_OPMENU_4 0x20 /* SE20: Sector Erase 0x20 */
+#define SPI_OPTYPE_4 0x03 /* Write, address required */
+
+#define SPI_OPMENU_5 0x9f /* RDID: Read ID */
+#define SPI_OPTYPE_5 0x00 /* Read, no address */
+
+#define SPI_OPMENU_6 0xd8 /* BED8: Block Erase 0xd8 */
+#define SPI_OPTYPE_6 0x03 /* Write, address required */
+
+#define SPI_OPMENU_7 0x0b /* FAST: Fast Read */
+#define SPI_OPTYPE_7 0x02 /* Read, address required */
+
nits: can we move these defines to ich.h?
Sure, no problem.
In ich.h, I see such macros are defined using enum,
You are referring to these here?
enum {
SPI_OPCODE_TYPE_READ_NO_ADDRESS = 0,
SPI_OPCODE_TYPE_WRITE_NO_ADDRESS = 1,
SPI_OPCODE_TYPE_READ_WITH_ADDRESS = 2,
SPI_OPCODE_TYPE_WRITE_WITH_ADDRESS = 3
};
not sure the
coding standard here, but I guess we may either follow the enum
defines, or use a new patch to convert enum to #defines...
I personally prefer #defines in such cases. So if nobody objects,
I can send a patch to move those enums to defines.
Thanks,
Stefan
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
