| To: | John M Trostel <jtrostel@xxxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: Repeat: Difficulties with large ACL/DACL counts |
| From: | Tim Shimmin <tes@xxxxxxx> |
| Date: | Wed, 22 May 2002 10:08:19 +1000 |
| Cc: | XFS-List <linux-xfs@xxxxxxxxxxx>, "acl-devel@xxxxxxxxxxx" <acl-devel@xxxxxxxxxxx>, Andreas Gruenbacher <ag@xxxxxxxxxxx>, Nathan Scott <nathans@xxxxxxxxxxxxxxxxxxxxxxx> |
| In-reply-to: | <1022008566.1989.12.camel@jtsdell>; from jtrostel@xxxxxxxxxxxxxx on Tue, May 21, 2002 at 03:16:05PM -0400 |
| References: | <1022008566.1989.12.camel@jtsdell> |
| Sender: | owner-linux-xfs@xxxxxxxxxxx |
Hi John, We haven't forgotten. cmd/xfstests/067 was written to reproduce your problem. Been a bit busy with other stuff. --Tim On Tue, May 21, 2002 at 03:16:05PM -0400, John M Trostel wrote: > There are really 2 problems described below. The first problem in which > having _exactly_ 21 Access ACEs and _exactly_ 21 Default ACEs on a > directory leads to system panics REMAINS. > > I suspect this is related to whether or not the ACEs and DACEs are small > enough to be represented as 'short-form' EAs. Note that it occurs > between 20 entries (4 + 20*12 = 244 bytes) and 21 entries ( 4 + 21*12 = > 256 bytes). > > The second problem, in which the 17th created ACE resulted in an > 'arguement list too long' message, has indeed been corrected. > > |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | TAKE - Trade in panic() for BUG() in debug kernel, Eric Sandeen |
|---|---|
| Next by Date: | TAKE - xfs_attr_leaf.c, Timothy Shimmin |
| Previous by Thread: | Repeat: Difficulties with large ACL/DACL counts, John M Trostel |
| Next by Thread: | TAKE - More I/O path work, Steve Lord |
| Indexes: | [Date] [Thread] [Top] [All Lists] |