jffs2/nand flash

[David Woodhouse]

On Wed, 2003-12-10 at 10:56 +0000, John Stirling wrote:
> On Tue, 2003-12-09 at 18:45, David Woodhouse wrote:
> > On Tue, 2003-12-09 at 17:55 +0000, John Stirling wrote:
> > > jffs2_flash_writev is called with vecs[0]=68, vecs[1]=489. This only
> > > results in one page write even though there is more than 512 bytes of
> > > data. It is called from jffs2_full_dnode (in write.c).
> > > alloc_mode=ALLOC_GC. Is this expected operation ? It seems to be causing
> > > the reported errors.
> > 
> > Yes, it's expected. The remainder should remain in the wbuf, and later
> > writes will be appended, before it's flushed. It'll also be flushed if
> > appropriate on sync(), fsync(), umount() etc.
> > 
> ok thanks. Am I right in thinking that fsync should follow fairly soon
> after the page write assuming no other page writes follow (2 secs) ?
> Doesn't look like it is happening on my system which is probably my
> fault..
Not fsync() but a flush should happen then, yes.

Compile with CONFIG_JFFS2_FS_DEBUG=1.  Run 'sync' after writing your
file. Show me the output.


-- 
dwmw2