[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Scary bugs



On Fri, 28 Jan 2000, Richard Braakman wrote:
>I vote these "most likely to delay the release".  They are difficult
>packages that we can't really do without, and seem to have been
>abandoned by their maintainers.  Please help.
>
>Package: quota (debian/main).
>Maintainer: Heiko Schlittermann <heiko@lotte.sax.de>
>[HELP] Mail to maintainer bounced.
>  41753 quota: repquota and quotacheck segfault
>  47222 quota: edquota - confusing
>  47718 edquota will not allow soft or hard limits above 4194303 blocks

Is 47718 critical?  4194303 blocks is one K less than 4G.  As on 32bit
platforms we can currently only support 2G files (actually 1 byte less than
2G) does it matter that we can only set a quota of 4G?
For 64bit platforms it is an issue, but I think that there are more
significant 64bit issues with the quota system than that.

For 47222, if software that is confusing is a release critical bug then how
can we ever release anything?!?!
I've just read that bug report and will correspond with the user off this
list.

For 41753, it was reported against 2.2.9.  The quota code in the kernel has
changed quite a bit since then.  I think that if it can't be reproduced on
the current kernels then we should consider it squashed.
I have CC'd this email to the reporter of the original bug for comments.  A
section of the 2.2.10 patch is below.

diff -u --recursive --new-file v2.2.10/linux/fs/dquot.c linux/fs/dquot.c
--- v2.2.10/linux/fs/dquot.c    Fri May  7 18:04:12 1999
+++ linux/fs/dquot.c    Mon Aug  9 12:04:41 1999
@@ -21,6 +21,17 @@
  *             Revised list management to avoid races
  *             -- Bill Hawes, <whawes@star.net>, 9/98
  *
+ *             Fixed races in dquot_transfer(), dqget() and dquot_alloc_...().
+ *             As the consequence the locking was moved from dquot_decr_...(),
+ *             dquot_incr_...() to calling functions.
+ *             invalidate_dquots() now writes modified dquots.
+ *             Serialized quota_off() and quota_on() for mount point.
+ *             Fixed a few bugs in grow_dquots. Fixed deadlock in write_dquot().
+ *             reset_dquot_ptrs() now traverse through inodes not filps.
+ *             add_dquot_ref() restarts after blocking
+ *             Added check for bogus uid and fixed check for group in quotactl.+ *             Jan Kara, <jack@atrey.karlin.mff.cuni.cz>, 4-6/99
+ *
  * (C) Copyright 1994 - 1997 Marco van Wieringen
  */ 

So what's the chances of me taking over the quota package with Brian May
sponsoring it?
I will not make any committment to maintaining it for a long time, but I can
do it at the moment.


-- 
Confucius wrote: To be fond of knowledge is better than simply acquiring
it, and to take delight in it is far better than simply being fond of it.


Reply to: