On 15 May 2015 at 07:49, Wang Haikun Haikun.Wang@freescale.com wrote:
On 5/14/2015 6:50 PM, Jagan Teki wrote:
On 14 May 2015 at 14:57, Haikun Wang haikun.wang@freescale.com wrote:
Atmel DataFlash chips have commands different from common spi flash commands. Atmel DataFlash also have special page-size. This driver add support for accessing Atmel DataFlash. It is based on the Driver Model. Example: => sf probe 1:0 SPI DataFlash: Detected AT45DB021B with page size 264 Bytes, erase size 264 Bytes, total 264 KiB, revision d => sf erase 0 42000 SF: 270336 bytes @ 0x0 Erased: OK => mw.l 82000000 45444342 20000 => sf write 82000000 0 42000 SF: 270336 bytes @ 0x0 Written: OK => sf read 83000000 0 42000 SF: 270336 bytes @ 0x0 Read: OK => cmp.b 82000000 83000000 42000 Total of 270336 byte(s) were the same
Signed-off-by: Haikun Wang haikun.wang@freescale.com
Verified with AT45DB021B on LS1021AQDS.
Changes in v4:
- Use dev_get_priv and dev_get_uclass_priv
- Add test log to commit message
Changes in v3:
- Rename file spi_dataflash.c to sf_dataflash.c
- Add comment for array "dataflash_data"
Changes in v2:
- Correct comment style
- Use get_timer in dataflash_waitready to check whether timeout
- Remove struct spi_flash * in struct dataflash, and get it from udevice->uclass_priv
- Replace spi_flash_write_common with spi_flash_cmd_write
- Replace spi_flash_read with spi_flash_cmd_read
- Change type of varible "status" form char to u8 in dataflash_status
- Change add_dataflash's argument type due to <change 3>
- Add claim_bus and release_bus in erase/write/read due to <change 4 & 5>
Changes in v1: None drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/sf_dataflash.c | 711 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 712 insertions(+) create mode 100644 drivers/mtd/spi/sf_dataflash.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index c61b784..8adec3d 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -15,6 +15,7 @@ endif #ifndef CONFIG_DM_SPI obj-$(CONFIG_SPI_FLASH) += sf_probe.o #endif +obj-$(CONFIG_DM_SF_DATAFLASH) += sf_dataflash.o
This looks confuse to me, CONFIG_SF_DATAFLASH looks simpler.
Fine.
For long run all the driver stuff could move to DM, no need to specify explicitly.
BTW: Did you define this any board config?
I will add it to LS1021AQDS board config and submit a separate patch.
Can you please send the patch as well, I will apply both at once.
Applying new feature driver with proper usage in particular board config with well tested - make me more confident to apply [for new feature-list]
thanks!