[uml-user] ReiserFS anybody?

[Peter J. B. <br...@cl...>]

Hi Chris, 

I am having bad luck getting Reiserfs to run under user mode Linux.  I
am running it on a block device with kernel 2.4.5-ac9, with a patch to
do nested transactions for InterMezzo (I'll get you that, when I'm
sure it works :). 

When I do 

mount -t reiserfs /dev/ubd/2 /mnt 

it takes forever (45 secs) but returns fine. 

cp -r /etc /mnt 

panics after a while: below is the stack trace.  Any ideas?  It may have
to do with some funkyness of the ubd devices since things
are fine on real devices. 

It's easily reproducable and I can track back with gdb as you can see
below, let me know if you need more information. 

- Peter -




(gdb) bt
#0  stop () at user_util.c:203
#1  0x10063d30 in reiserfs_panic (sb=0x50073e98, 
    fmt=0x1012dcc0 "journal-1577: handle trans id %d != current trans
id %d\n")
    at prints.c:332
#2  0x1006f96e in journal_mark_dirty (th=0x50073e78,
p_s_sb=0x50073e98, 
    bh=0x50930b00) at journal.c:2118
#3  0x1005c743 in reiserfs_update_sd (th=0x50073e78, inode=0x50ebf080)
    at inode.c:1123
#4  0x1005c916 in reiserfs_write_inode (inode=0x50ebf080, do_sync=1)
    at inode.c:1236
#5  0x1003cfb5 in try_to_sync_unused_inodes () at inode.c:185
#6  0x1003d6fd in prune_icache (goal=18) at inode.c:682
#7  0x1003d737 in shrink_icache_memory (priority=6, gfp_mask=4) at
inode.c:703
#8  0x100232d2 in do_try_to_free_pages (gfp_mask=4, user=0) at
vmscan.c:877
#9  0x10023362 in kswapd (unused=0x0) at vmscan.c:941
#10 0x100c1f61 in new_thread_proc (t=0x50070000) at process_kern.c:115