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

Re: [patch?] Re: Do ramdisk exec's map direct to buffer cache?

bjorn@xxxxxxx.se said:
> I would not say it's "lots of RAM".. ok it's obviously not scaleable,
> but in absolute numbers for a moderately sized R/W filesystem part on
> an embedded device, it's no problem. What's needed to make it really
> scaleable is a checkpointing node (just like normal log-structured
> filesystems).

Actually, I've just checked the numbers and it's not anywhere near as 
RAM-hungry as I'd thought -  I appear to have been talking crap.

The filesystem on my toy at the moment is about 4.5MB in 424 files:

 no_jffs_file = 424,         sizeof(struct jffs_file) == 88 
 no_jffs_node = 515,         sizeof(struct jffs_node) == 48
 no_jffs_node_ref = 516,     sizeof(struct jffs_node_ref) == 8
 no_jffs_fm = 578,           sizeof(struct jffs_fm) == 20

Rounding the sizes up to the next power of two and adding them up, it only 
comes to about 110KB. If were were to stop using kmalloc() and use our own 
slab pools, we could cut that down to ~75KB.