On 2014-08-05 09:20, Hannes Petermaier wrote:
hi folks,

i am actually trying to enable the "enhanced data area" on an 4GB emmc device, this i've done successfully with the linux mmc-utils and the card has after this "partitioning" 2GB user-partition.

Linux kernel deals correctly with the card.

U-Boot doesn't:

"
MMC: block number 0x301 exceeds max(0x0)
spl: mmc blk read err - 0
### ERROR ### Please RESET the board ###
"

So i debugged around and found the capacity-calculation, which i don't fully understand.
"
    if (mmc->high_capacity) {
        csize = (mmc->csd[1] & 0x3f) << 16
            | (mmc->csd[2] & 0xffff0000) >> 16;
        cmult = 8;
    } else {
        csize = (mmc->csd[1] & 0x3ff) << 2
            | (mmc->csd[2] & 0xc0000000) >> 30;
        cmult = (mmc->csd[2] & 0x00038000) >> 15;
    }
"
The else path does for my opinion the right calculation according the mmc-specification, the "high_capacity" path i do not understand at all, neither i found some specification for this.

In fact my 4GB card (with 2GB enhanced area) works only with "the else" path.
my csd Registers:

MMC-card dection:
ocr    : 0xc0ff8080
csd[0] : 0xd04f0132
csd[1] : 0x0f5a13ff
csd[2] : 0xffffffe7
csd[3] : 0x8a4000f1
c_size : 0x00000000003fffff
c_mult : 0x0000000000000008

U-Boot (BuR V2.0)# mmc info
Device: OMAP SD/MMC
Manufacturer ID: fe
OEM: 14e
Name: MMC04
Tran Speed: 52000000
Rd Block Len: 512
MMC version 4.41
High Capacity: Yes
Capacity: 4 TiB
Bus Width: 1-bit

any ideas ?

best regards,
Hannes

I've digged a bit deeper and found an (for me, as newbie in this) interesting webpage:

http://www.hjreggel.net/cardspeed/special-sd.html

there i've found the key to understand the "high_capacity" path:
"...In CSD Version 2.0, the C_SIZE field was extended to 22 bit. The C_SIZE_MULT was dropped, assuming a preset multiplier of 2^10 . The READ_BL_LEN was kept, but the only value allowed is 9, indicating a block length of 512 bytes."

So the code should be as following:

-       if (mmc->high_capacity) {
+       if (mmc->high_capacity && mmc->read_bl_len == MMC_MAX_BLOCK_LEN) {

i'am correct in my thinking ? please some comment.
if yes - i will format a patch for this

best regards,
Hannes




_______________________________________________
U-Boot mailing list
[email protected]
http://lists.denx.de/mailman/listinfo/u-boot

Reply via email to