[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: New jffs2 files stay uncompressed?

joakim.tjernlund@xxxxxxx.se said:
> Sorry to bug you with all these questions, but .. 

Np. Questions which have a simple answer are fine. Questions I have to 
think about, like the last one, you might have to wait for answers to :)

> In jffs2_scan_inode_node() a raw node is created even for nodes that does
> not have the ACCURATE bit set, but if a CRC error is found in the data
> part no raw inode is created. Should raw nodes be created for nodes
> that doesn't have the ACCURATE bit set or can we just mark the space
> DIRTY?  

We should be able to just mark the space dirty.

If we're doing that, it could also make sense to free the raw_node_ref
whenever we obsolete a node later, too. We'd want to start using
doubly-linked lists then though - removing nodes is a PITA at the moment.
We'd have to see how that tradeoff worked - fewer, larger, raw_node_refs.
Wanna instrument it so it keeps a count of how many are clean/dirty during
normal operations?

Note that we can't do this for NAND flash when nodes are just marked
obsolete in memory, not actually marked as such on the flash - at least we
can't do it for dirent nodes. Unless we come up with a different
implementation of jffs2_garbage_collect_dirent() that is.


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo@xxxxxxx.com