Skip to contentSkip to conversation

JFFS2 internals question

DISCUSSION David Müller (ELSOFT AG) 3 participants 7 messages Aug 14, 2003 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

Hello

As i'm not a JFFS2 guru i hope someone familar with the JFFS2 internals could answer this question:

Situation:

JFFS2 filesystem on a NAND flash device:

Is it true that the read calls emitted from the JFFS2 code to the MTD HW driver do not cross NAND block boundries (nand_read_ecc() has its start and end point of reading in the same NAND block)?

TIA

Dave

#1 lore raw
Quoted 3 lines · David Müller (ELSOFT AG)

Is it true that the read calls emitted from the JFFS2 code to the MTD HW driver do not cross NAND block boundries (nand_read_ecc() has its start and end point of reading in the same NAND block)?

I'd first like to know why you think you need the answer to this... :)

Signature · 1 line
dwmw2
#2 lore raw
Quoted 9 lines · David Woodhouse
Quoted 3 lines · David Müller (ELSOFT AG)

Is it true that the read calls emitted from the JFFS2 code to the MTD HW driver do not cross NAND block boundries (nand_read_ecc() has its start and end point of reading in the same NAND block)?

I'd first like to know why you think you need the answer to this... :)

I need it to test a theory.

Dave

#3 lore raw
Quoted 8 lines · David Müller (ELSOFT AG)
Quoted 7 lines · David Woodhouse
Quoted 3 lines · David Müller (ELSOFT AG)

Is it true that the read calls emitted from the JFFS2 code to the MTD HW driver do not cross NAND block boundries (nand_read_ecc() has its start and end point of reading in the same NAND block)?

Quoted 4 lines
Quoted 2 lines

I'd first like to know why you think you need the answer to this... :)

I need it to test a theory.

I think we do make read calls which cross a page boundary, yes.

Signature · 1 line
dwmw2
#4 lore raw
Quoted 1 line · David

I think we do make read calls which cross a page boundary, yes.

And the inability to handle such calls was the reason that the old DiskOnChip driver didn't work with jffs2, as I recall.

-Jasmine.

#5 lore raw
In reply to David Woodhouse
Quoted 12 lines · David Woodhouse
Quoted 10 lines · David Müller (ELSOFT AG)
Quoted 8 lines · David Woodhouse
Quoted 3 lines · David Müller (ELSOFT AG)

Is it true that the read calls emitted from the JFFS2 code to the MTD HW driver do not cross NAND block boundries (nand_read_ecc() has its start and end point of reading in the same NAND block)?

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

I'd first like to know why you think you need the answer to this... :)

I need it to test a theory.

I think we do make read calls which cross a page boundary, yes.

Sorry, i was talking about block boundary, not page boundary. Or is it just a typo?

Dave

#6 lore raw
Quoted 2 lines · David Müller (ELSOFT AG)

Sorry, i was talking about block boundary, not page boundary. Or is it just a typo?

Oh, sorry. We'll not read over _logical_ eraseblock boundaries but on large chips we may artificially increase the eraseblock size to be 2* or 4* the physical eraseblock size, to reduce the number of 'struct jffs2_eraseblock' we have to allocate. In that case, yes, we'll be reading across eraseblock boundaries.

Signature · 1 line
dwmw2