Skip to contentSkip to conversation

Different erase block sizes in a JFFS2 partion.

DISCUSSION Joakim Tjernlund 2 participants 3 messages Sep 28, 2006 – Sep 29, 2006 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

Got multiple NOR chips with top and bottom boot blocks with diffrent EB size then the rest of the blocks.
I want to put a JFFS2 FS spanning over several chips which makes the FS cross top/bottom boot blocks.
Since(last I looked) JFFS2 can't deal with diffrent EB sizes on the same partition I need do something about it.
I want to use point()/unpoint() and a cached memory mapping for burst access and I don't know how to do that yet and I would like some ideas.

One idea I have is to reserve the boot blocks with a new type of JFFS2 node(similar to JFFS2_FEATURE_ROCOMPAT) that spans the whole boot block. JFFS2 should not try to move or delete nodes of this type. Not sure if thats a good idea or if its hard to impl. yet.

Better ideas anyone?

#1 lore raw
Quoted 2 lines · Joakim Tjernlund

Got multiple NOR chips with top and bottom boot blocks with diffrent EB size then the rest of the blocks.

Intel P30? IIRC, the MTD driver maps those into a set of larger virtual eraseblocks. You might not be able to even distinguish that they exist.

josh

#2 lore raw
Quoted 7 lines · Josh Boyer
Quoted 2 lines · Joakim Tjernlund

Got multiple NOR chips with top and bottom boot blocks with diffrent EB size then the rest of the blocks.

Intel P30? IIRC, the MTD driver maps those into a set of larger virtual eraseblocks. You might not be able to even distinguish that they exist.

yes, Intel P30.
The cfi_cmdset_0001 driver maps them like this:
0: offset=0x0,size=0x8000,blocks=4
1: offset=0x20000,size=0x20000,blocks=255
2: offset=0x2000000,size=0x8000,blocks=4
3: offset=0x2020000,size=0x20000,blocks=255

First, this map is wrong. 2: and 3: should be swapped, that is, the small blocks should be at the end of the device.

the small blocks fits into a larger one so I should have no problem reading/writing to such virtual block, but I don't think erase will work as JFFS2 will only erase one of the small blocks instead of all 4. Maybe I am missing something?