Re: audit 0.9.17 released
On Wed, 2005-07-13 at 17:19 -0400, Linda Knippers wrote:
Since that last problem was so easy, I'll mention another that I
have less info on as I've only seen it twice, both times on my ia64
box, and I couldn't capture much data.
The problem I've seen is a panic while shutting down when there's
a heavy audit load. The message is this:
Bah. I should know better than to set precedents like that. audit.76 is
uploading; audit.77 is building with this...
--- linux-2.6.9/kernel/auditsc.c~ 2005-07-13 21:34:12.000000000 +0100
+++ linux-2.6.9/kernel/auditsc.c 2005-07-13 22:29:33.000000000 +0100
@@ -802,7 +802,7 @@ static void audit_log_task_info(struct a
up_read(&mm->mmap_sem);
}
-static void audit_log_exit(struct audit_context *context)
+static void audit_log_exit(struct audit_context *context, unsigned int gfp_mask)
{
int i;
struct audit_buffer *ab;
@@ -810,7 +810,7 @@ static void audit_log_exit(struct audit_
struct audit_watch_info *winfo;
struct hlist_node *pos;
- ab = audit_log_start(context, GFP_KERNEL, AUDIT_SYSCALL);
+ ab = audit_log_start(context, gfp_mask, AUDIT_SYSCALL);
if (!ab)
return; /* audit_panic has been called */
audit_log_format(ab, "arch=%x syscall=%d",
@@ -957,9 +947,11 @@ void audit_free(struct task_struct *tsk)
return;
/* Check for system calls that do not go through the exit
- * function (e.g., exit_group), then free context block. */
+ * function (e.g., exit_group), then free context block.
+ * We use GFP_ATOMIC here because we might be doing this
+ * in the context of the idle thread */
if (context->in_syscall && context->auditable)
- audit_log_exit(context);
+ audit_log_exit(context, GFP_ATOMIC);
audit_free_context(context);
}
@@ -1064,7 +1056,7 @@ void audit_syscall_exit(struct task_stru
return;
if (context->in_syscall && context->auditable)
- audit_log_exit(context);
+ audit_log_exit(context, GFP_KERNEL);
context->in_syscall = 0;
context->auditable = 0;
--
dwmw2