Perhaps the old app is calling sys_old_getrlimit() from
linux/kernel/sys.c. It truncates rlimits to 0x7FFFFFFF
if it's bigger than that. 0x7FFFFFFF used to be the old
RLIM_INFINITY in 2.2 [actually, ((long)(~0UL>>1))]. In
2.4, RLIM_INFINITY is (~0UL).
So if you call sys_setrlimit() with the old RLIM_INFINITY from 2.2
OR with the result from sys_old_getrlimit(), then the new limit
will be 0x7FFFFFFF instead of unlimited.
Looks like someone forgot to implement sys_old_setrlimit(),
which would have been the right thing to do.
Now all we can do is to hack sys_setrlimit and let it translate
0x7FFFFFFF to RLIM_INFINITY.
The following untested and uncompiled patch might do it, or not...
--- linux-2.4.17-pre2/kernel/sys.c.orig Tue Sep 18 23:10:43 2001
+++ linux-2.4.17-pre2/kernel/sys.c Wed Dec 5 23:30:50 2001
@@ -1120,6 +1120,16 @@
return -EINVAL;
if(copy_from_user(&new_rlim, rlim, sizeof(*rlim)))
return -EFAULT;
+#if !defined(__ia64__)
+ /*
+ * In 2.2, RLIMIT_INFINITY was defined as ((long)(~0UL>>1)).
+ * Reckognize it and translate it to the new RLIMIT_INFINITY.
+ */
+ if ((long)new_rlim.rlim_cur == ((long)(~0UL>>1)))
+ new_rlim.rlim_cur = RLIMIT_INFINITY;
+ if ((long)new_rlim.rlim_max == ((long)(~0UL>>1)))
+ new_rlim.rlim_max = RLIMIT_INFINITY;
+#endif
old_rlim = current->rlim + resource;
if (((new_rlim.rlim_cur > old_rlim->rlim_max) ||
(new_rlim.rlim_max > old_rlim->rlim_max)) &&
Mike.
-- "Only two things are infinite, the universe and human stupidity, and I'm not sure about the former" -- Albert Einstein.- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/