This has been tested on the RB800, but should work on the RB333, RB600, and
RB1100 as well.
It's currently missing ECC support, but read and write are complete.
Details
Details
Diff Detail
Diff Detail
- Repository
- rS FreeBSD src repository - subversion
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| sys/dev/nand/nfc_rb.c | ||
|---|---|---|
| 3 ↗ | (On Diff #4660) | Why this? |
| 103–110 ↗ | (On Diff #4660) | This looks like fun. It also looks like code copied from somewhere else... |
| 232 ↗ | (On Diff #4660) | read_region? Is that really right? |
| 257 ↗ | (On Diff #4660) | I didn't see where you asserted the CS for this chip. Are you relying on the boot loader to do that, or is it hard wired? |
| sys/dev/nand/nfc_rb.c | ||
|---|---|---|
| 3 ↗ | (On Diff #4660) | This started life as a copy of nfc_mv.c. The only remaining original method, though, is write_buf, which is almost identical to the original. |
| 103–110 ↗ | (On Diff #4660) | Copied in spirit from the Linux driver (had to get the eccpositions from somewhere). |
| 232 ↗ | (On Diff #4660) | The Linux driver does a memcpy(), so this seemed appropriate. Is there a better option? |
| 257 ↗ | (On Diff #4660) | CS is the nce GPIO. I don't see a deselect_cs, so just enabled it in the send_command() and send_address() methods. I guess it can be moved here. |
Comment Actions
This looks good to me. Feel free to do or not do the comments (except the last one: the code under review is fine and preferable).
| sys/dev/nand/nfc_rb.c | ||
|---|---|---|
| 3 ↗ | (On Diff #4660) | That's de-minimus, also the write_buf method looks to be one of the few ways to express this... |
| 232 ↗ | (On Diff #4660) | No, this is good. You are reading a buffer, not a FIFO like I though originally... |
| 257 ↗ | (On Diff #4660) | I wouldn't. I'd just enable it like you are now until the NAND system can cope with multiple CS for real. Better to have everybody do the same thing... |