[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210922110836.304335737@infradead.org>
Date: Wed, 22 Sep 2021 13:05:13 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: gor@...ux.ibm.com, jpoimboe@...hat.com, jikos@...nel.org,
mbenes@...e.cz, pmladek@...e.com, mingo@...nel.org
Cc: linux-kernel@...r.kernel.org, peterz@...radead.org,
joe.lawrence@...hat.com, fweisbec@...il.com, tglx@...utronix.de,
hca@...ux.ibm.com, svens@...ux.ibm.com, sumanthk@...ux.ibm.com,
live-patching@...r.kernel.org, paulmck@...nel.org
Subject: [RFC][PATCH 7/7] livepatch,context_tracking: Avoid disturbing NOHZ_FULL tasks
When a task is stuck in NOHZ_FULL usermode, we can simply mark the
livepatch state complete.
Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
---
kernel/livepatch/transition.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
--- a/kernel/livepatch/transition.c
+++ b/kernel/livepatch/transition.c
@@ -270,13 +270,24 @@ static int klp_check_task(struct task_st
{
int ret;
- if (task_curr(task))
+ if (task_curr(task)) {
+ if (context_tracking_state_cpu(task_cpu(task)) == CONTEXT_USER) {
+ /*
+ * If we observe the CPU being in USER context, they
+ * must issue an smp_mb() before doing much kernel
+ * space and as such will observe the patched state,
+ * mark it clean.
+ */
+ goto complete;
+ }
return -EBUSY;
+ }
ret = klp_check_stack(task, arg);
if (ret)
return ret;
+complete:
clear_tsk_thread_flag(task, TIF_PATCH_PENDING);
task->patch_state = klp_target_state;
return 0;
Powered by blists - more mailing lists