loop semi-broken in 1.3.80

Ray_Van_Tassle-CRV004@email.mot.com
1 Apr 96 08:34:58 -0600


I thought the problem in 1.3.72 was with loop to a file on an msdos
partition. But with 1.3.80, I can get the same failure with an ext2 file.
Here's what I do:
>losetup /dev/loop0 /tmp/looptst0 (or maybe the other way around)
>mkfs -text2 /dev/loop0
>losetup -d /dev/loop0
All works ok, then---
>mount /tmp/looptst0 /mnt -oloop
It mounts ok. then---
>cp *.zip /mnt ; rm -f /mnt/*.zip

Repeating this--eventually the system will lock up.
I have tracked it down this far:
loop calls getblk (which never returns),
getblk does while(!free_list[isize]) refill_freelist(size).
It seems to hang in the whileloop--no buffers ever get put on the free_list.

I'm trying to track this down, but I'm having a hard time figuring out how
it's *supposed* to work---the documentation/comments are very sparse!
Any comments, suggestions, verification would be appreciated.

Ray Van Tassle
rayvt@comm.mot.com