[Mono-bugs] [Bug 74649][Min] Changed - NUnit test in corlib fails
Thu, 21 Apr 2005 10:00:11 -0400 (EDT)
Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.
Changed by firstname.lastname@example.org.
--- shadow/74649 2005-04-21 08:34:32.000000000 -0400
+++ shadow/74649.tmp.20000 2005-04-21 10:00:11.000000000 -0400
@@ -202,6 +202,24 @@
opened by another process via the technique you employ there. I'm
investigating, but it seems like the only way to do this reliably on
FreeBSD is to read /dev/kmem.
And again I say, Feh.
+------- Additional Comments From email@example.com 2005-04-21 10:00 -------
+Here's the scenario:
+1) there are items in the share table with refs > 1
+2) someone tries to open a file that is already listed in the table
+3) the share check denies permission
+4) an additional check is made (this /proc/xx/fd/yy lookup) to see if
+in fact the share check is out of date and can therefore be ignored
+If a platform can't do step 4) then it would just return a share
+permission failure from CreateFile().
+Your bug is stating that there are items in the share table with
+refs==0, and someone is trying to release them again. The bug
+requires figuring out how the refs managed to get to 0 without the
+entry being deleted.