Skip to contentSkip to conversation

Updated 2.4 htree patches available for 2.4.21rc5

DISCUSSION Theodore Ts'o 2 participants 2 messages Mar 8, 2003 – Mar 9, 2003

lore
Share & tools
lore
#0 lore raw

I've backported all of the bugfixes to the 2.5 dxdir/htree patches to 2.4, and have created a new set of patches for Linux 2.4.21rc5. At this point it *looks* like we've fixed all of the htree bugs that people have reported, including the brelse bug, the memory leak bugs, and the NFS compatibility problems.

I've done *very* light testing, and things seem to work, but I'm now looking for some brave souls/guinea pigs to give this some more strenous testing. Please don't use this on production systems just yet, until some people report success. In particular, I'm looking for people who had trouble with NFS to confirm whether or not this patch fixes their problems or not. If you do try out the patch, please let me know how well (or how poorly) it works.

The patches can be found at:

	http://thunk.org/tytso/linux/extfs-2.4-update/extfs-update-2.4.21rc5

Also available are the patches broken out into separate diffs, here:

	http://thunk.org/tytso/linux/extfs-2.4-update/broken-out-2.4.21rc5/

In the broken out directory, only the first patch is the htree patch. The rest are some other 2.5 features which I've backported to 2.4, including forward compatibility for upcoming future extfs upgrades including dynamic filesystem resizing and expanded inodes for things like subsecond timestamps. The patches also add support for setting default mount options via tune2fs, and the improved Orlov block allocator.

So if you're feeling adventurour, please backup your systems (you should *always* backup your systems, even when you're not trying out new kernel patches :-), and give these patches a whirl, and let me know how it goes. Thanks!

- Ted

#1 lore raw

Still getting some NFS htree strangeness

Quoted 5 lines · Theodore Ts'o

I've backported all of the bugfixes to the 2.5 dxdir/htree patches to 2.4, and have created a new set of patches for Linux 2.4.21rc5. At this point it *looks* like we've fixed all of the htree bugs that people have reported, including the brelse bug, the memory leak bugs, and the NFS compatibility problems.

I'm still getting odd results from programs doing readdir over NFS. I haven't really nailed down what's happening yet, but I'm getting readdir failing with EOVERFLOW. Strace doesn't show any syscalls failing, and an inspection of the glibc source of readdir() shows that it can return EOVERFLOW, but I don't really understand the code yet. It may be that it assumes that the getdents "offset" is really a file offset rather than a random number.

I'm inclined to suspect that it is a glibc bug, but I'm not willing to say so until I've worked out what's really going on. Also, even if it were, it would be worth working around if possible (otherwise htree+nfs is still effectively useless).

I'll see if I can get something more concrete showing what's going on.

J