I forgot about "make xcheck", it crashes the kernel. Fortunately the
kernel dumps the stack trace. Trust me, it wasn't easy to notice the
missing return ;) I am wondering why the compiler doesn't complain.
Roland, this all needs more fixes. Look at the fixed code,
utrace = target->utrace;
if (!utrace)
return ERR;
spin_lock(utrace->lock);
Now, if we race with another task doing utrace_task_alloc() and see
->utrace != NULL, why should we see the correctly initialized *utrace?
utrace_task_alloc() needs wmb(), and the code like above read_barrier_depends().
UPD: tested the kernel with this patch, now late-ptrace-may-attach-check
crashes the kernel silently (no output under kvm).
Signed-off-by: Oleg Nesterov <o...@redhat.com>
---
kernel/utrace.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- UTRACE-PTRACE/kernel/utrace.c~UTRACE_ATTACH_FIX_UTRACE_CK 2009-11-16
00:02:08.000000000 +0100
+++ UTRACE-PTRACE/kernel/utrace.c 2009-11-16 00:06:26.000000000 +0100
@@ -281,7 +281,7 @@ struct utrace_engine *utrace_attach_task
if (!(flags & UTRACE_ATTACH_CREATE)) {
if (unlikely(!utrace))
- ERR_PTR(-ENOENT);
+ return ERR_PTR(-ENOENT);
spin_lock(&utrace->lock);
engine = matching_engine(utrace, flags, ops, data);
if (engine)
