Date: Fri, 27 Aug 2004 21:52:45 +0400 From: Pavel Merdine <fbsdlist@merdin.com> To: freebsd-stable@freebsd.org Subject: ffs_alloc panic patch Message-ID: <1076237332.20040827215245@kaluga.ru>
next in thread | raw e-mail | index | archive | help
Hello,
I'd like to propose the following patch:
------------------------------------------------------------------
--- ffs_alloc.c.orig Fri May 14 19:03:50 2004
+++ ffs_alloc.c Sat Aug 7 03:36:32 2004
@@ -712,7 +712,7 @@
minbfree = 1;
cgsize = fs->fs_fsize * fs->fs_fpg;
dirsize = fs->fs_avgfilesize * fs->fs_avgfpdir;
- curdirsize = avgndir ? (cgsize - avgbfree * fs->fs_bsize) / avgndir : 0;
+ curdirsize = avgndir ? (cgsize - avgbfree * fs->fs_bsize) / avgndir : 512;
if (dirsize < curdirsize)
dirsize = curdirsize;
maxcontigdirs = min((avgbfree * fs->fs_bsize) / dirsize, 255);
------------------------------------------------------------------
It should solve the problem of "panic: integer divide fault" on the
last line of the code above. The panic happens on:
1. directory creation
2. when disk is empty
3. when avg_dir_size avg_file_size are set to values, so
avg_dir_size * avg_file_size is negative in integer.
example: avg_dir_size = 8000, avf_file_size = 375000.
(curdirsize == 0 and dirsize < 0, so dirsize = 0)
I'm sure it's very old bug.
I know that maybe those params are wrong. And maybe newfs or tunefs
should check them.
But I'm sure that there should be some check in the code above to
eliminate ANY possibility of panic.
Panic is VERY undesirable situation. And I'm in doubt why those people
who wrote ffs like panics so devotedly:
# grep -c "panic" ffs_alloc.c ffs_softdep.c
ffs_alloc.c:37
ffs_softdep.c:108
I think such things are not acceptable in production environment. Why
those functions cannot just return a failure state and leave system
working?
--
/ Pavel Merdine
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1076237332.20040827215245>