[ECOS] JFFS2 Garbage Collection Error? (NO! rbtree.c bug)
Bob Koninckx
bob.koninckx@mech.kuleuven.ac.be
Mon Jul 14 19:24:00 GMT 2003
It has to do with the eCos C-library which only buffers 255 bytes of
data. PAGE_CACHE_SIZE is defined as 4 kilobytes. I use the same solution
as yours (which should be probably ok according to the response I got
from David then), and have had no more problems since.
I would be great if it could be confirmed that this is indeed a/the
correct solution, 'cause, just like you, I am not really sure about what
I have been doing there.
Bob
On Mon, 2003-07-14 at 20:26, Scott Wilkinson wrote:
> >
> > That's great! Are there any other [similar] issues like this in that
> > code?
>
> Actually, there is one other thing that I had to do to get my tests to
> run. Perhaps David can comment...
>
> If on a JFFS2 file you were to:
>
> open();
> write( fd, buf, 100 );
> close();
>
> an then
>
> open();
> read( fd, buf, 50 );
> read( fd, buf, 50 ); /* This would fail */
> close();
>
> The second read was failing for me. I had to comment out a check in the
> JFFS2 read.c file. I am not sure what it was doing, but I can't run
> with it in.
>
> Take a look at the code below. Hopefully my mail client will not munge
> it too hard. Look for my "#if 0"
>
> Scott
>
> from: packages/fs/jffs2/current/src/read.c
>
> int jffs2_read_inode_range(struct jffs2_sb_info *c, struct
> jffs2_inode_info *f,
> unsigned char *buf, uint32_t offset, uint32_t len)
> {
> uint32_t end = offset + len;
> struct jffs2_node_frag *frag;
> int ret;
>
> D1(printk(KERN_DEBUG "jffs2_read_inode_range: ino #%u, range 0x%08x-0x%08x\n",
> f->inocache->ino, offset, offset+len));
>
> frag = jffs2_lookup_node_frag(&f->fragtree, offset);
>
> /* XXX FIXME: Where a single physical node actually shows up in two
> frags, we read it twice. Don't do that. */
> /* Now we're pointing at the first frag which overlaps our page */
> while(offset < end) {
> D2(printk(KERN_DEBUG "jffs2_read_inode_range: offset %d, end %d\n", offset, end));
> if (!frag || frag->ofs > offset) {
> uint32_t holesize = end - offset;
> if (frag) {
> D1(printk(KERN_NOTICE "Eep. Hole in ino #%u fraglist. frag->ofs = 0x%08x, offset = 0x%08x\n", f->inocache->ino, frag->ofs, offset));
> holesize = min(holesize, frag->ofs - offset);
> D1(jffs2_print_frag_list(f));
> }
> D1(printk(KERN_DEBUG "Filling non-frag hole from %d-%d\n", offset, offset+holesize));
> memset(buf, 0, holesize);
> buf += holesize;
> offset += holesize;
> continue;
> #if 0
> /* SCOTT - This had to be taken out to get tests to run.
> Seems like it is checking for overlap, but perhaps there
> is a logic problem somewhere because when I print the frag_list
> I don't see any overlap. */
> } else if (frag->ofs < offset && (offset & (PAGE_CACHE_SIZE-1)) != 0) {
> D1(printk(KERN_NOTICE "Eep. Overlap in ino #%u fraglist. frag->ofs = 0x%08x, offset = 0x%08x\n",
> f->inocache->ino, frag->ofs, offset));
> D1(jffs2_print_frag_list(f));
> memset(buf, 0, end - offset);
> return -EIO;
> #endif
> } else if (!frag->node) {
> uint32_t holeend = min(end, frag->ofs + frag->size);
> D1(printk(KERN_DEBUG "Filling frag hole from %d-%d (frag 0x%x 0x%x)\n", offset, holeend, frag->ofs, frag->ofs + frag->size));
> memset(buf, 0, holeend - offset);
> buf += holeend - offset;
> offset = holeend;
> frag = frag_next(frag);
> continue;
> } else {
> uint32_t readlen;
> uint32_t fragofs; /* offset within the frag to start reading */
>
> fragofs = offset - frag->ofs;
> readlen = min(frag->size - fragofs, end - offset);
> D1(printk(KERN_DEBUG "Reading %d-%d from node at 0x%08x (%d)\n",
> frag->ofs+fragofs, frag->ofs+fragofs+readlen,
> ref_offset(frag->node->raw), ref_flags(frag->node->raw)));
> ret = jffs2_read_dnode(c, frag->node, buf, fragofs + frag->ofs - frag->node->ofs, readlen);
> D2(printk(KERN_DEBUG "node read done\n"));
> if (ret) {
> D1(printk(KERN_DEBUG"jffs2_read_inode_range error %d\n",ret));
> memset(buf, 0, readlen);
> return ret;
> }
> buf += readlen;
> offset += readlen;
> frag = frag_next(frag);
> D2(printk(KERN_DEBUG "node read was OK. Looping\n"));
> }
> }
> return 0;
> }
--
----------------------------------------------------------------------
ir. Bob Koninckx
Katholieke Universiteit Leuven
Division Production Engineering, tel. +32 16 322535
Machine Design and Automation fax. +32 16 322987
Celestijnenlaan 300B bob.koninckx@mech.kuleuven.ac.be
B-3001 Leuven Belgium http://www.mech.kuleuven.ac.be/pma
----------------------------------------------------------------------
--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss
More information about the Ecos-discuss
mailing list