I added to block_write_full_page() the debugging code suggested by
Alexander Viro:
if (inode->i_dev == 0x305 && inode->i_ino == 48991) // Md
printk("block_write_full_page: writing page %d, size %Ld\n",
page->index, inode->i_size);
and I have just been able to trigger the bug (two times):
root@wonderland:vc/2:/var/lib/news$cp active active.ok
root@wonderland:vc/2:/var/lib/news$diff -u active active.ok
root@wonderland:vc/2:/var/lib/news$/etc/init.d/inn stop
Stopping news server: innd
block_write_full_page: writing page 0, size 3075
root@wonderland:vc/2:/var/lib/news$diff -u active active.ok
--- active Sun Sep 17 16:09:01 2000
+++ active.ok Sun Sep 17 21:03:16 2000
@@ -1,11 +1,11 @@
control 0000004774 0000004775 y
control.cancel 0000021917 0000021889 n
junk 0000001777 0000001768 y
-fido.ita.ridere 0000014632 0000014600 y
+fido.ita.ridere 0000014634 0000014600 y
[...]
-rw-r--r-- 1 news news 3075 Sep 17 16:09 active
-rw-r--r-- 1 root root 3075 Sep 17 21:03 active.ok
I'm running -test8 with the block_truncate_page() patch.
BTW, I see devfs triggers modules loading on readlink(2). Should this
be considered a feature?
Every time I run "ls --color /dev" the modules corresponding to devices
which have a symlink /dev/audio are loaded, so there is no much point
in having on demand loading...
-- ciao, Marco- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 21:00:15 EST