From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756827AbaFYLJr (ORCPT ); Wed, 25 Jun 2014 07:09:47 -0400 Received: from ip4-83-240-18-248.cust.nbox.cz ([83.240.18.248]:41099 "EHLO ip4-83-240-18-248.cust.nbox.cz" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756181AbaFYLHR (ORCPT ); Wed, 25 Jun 2014 07:07:17 -0400 From: Jiri Slaby To: linux-kernel@vger.kernel.org Cc: tj@kernel.org, rostedt@goodmis.org, mingo@redhat.com, akpm@linux-foundation.org, andi@firstfloor.org, paulmck@linux.vnet.ibm.com, pavel@ucw.cz, jirislaby@gmail.com, Vojtech Pavlik , Michael Matz , Jiri Kosina , Jiri Slaby , Frederic Weisbecker Subject: [PATCH -repost 10/21] kgr: kthreads support Date: Wed, 25 Jun 2014 13:07:04 +0200 Message-Id: <1403694435-3180-10-git-send-email-jslaby@suse.cz> X-Mailer: git-send-email 2.0.0 In-Reply-To: <1403694435-3180-1-git-send-email-jslaby@suse.cz> References: <1403694435-3180-1-git-send-email-jslaby@suse.cz> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Wake up kthreads so that they cycle through kgr_task_safe either by an explicit call to it or implicitly via try_to_freeze. This ensures nobody should use the old version of the code and kgraft core can push everybody to use the new version by switching to the fast path. Signed-off-by: Jiri Slaby Cc: Tejun Heo Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Ingo Molnar --- kernel/kgraft.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) diff --git a/kernel/kgraft.c b/kernel/kgraft.c index b0c712e8297c..d21eaad8d48d 100644 --- a/kernel/kgraft.c +++ b/kernel/kgraft.c @@ -53,7 +53,7 @@ static void kgr_stub_slow(unsigned long ip, unsigned long parent_ip, { struct kgr_loc_caches *c = ops->private; - if (kgr_task_in_progress(current) && current->mm) { + if (kgr_task_in_progress(current)) { pr_info("kgr: slow stub: calling old code at %lx\n", c->old); kgr_set_regs_ip(regs, c->old + MCOUNT_INSN_SIZE); @@ -71,11 +71,7 @@ static bool kgr_still_patching(void) read_lock(&tasklist_lock); for_each_process(p) { - /* - * TODO - * kernel thread codepaths not supported and silently ignored - */ - if (kgr_task_in_progress(p) && p->mm) { + if (kgr_task_in_progress(p)) { pr_info("pid %d (%s) still in kernel after timeout\n", p->pid, p->comm); failed = true; @@ -123,13 +119,23 @@ static void kgr_work_fn(struct work_struct *work) mutex_unlock(&kgr_in_progress_lock); } -static void kgr_mark_processes(void) +static void kgr_handle_processes(void) { struct task_struct *p; read_lock(&tasklist_lock); - for_each_process(p) + for_each_process(p) { kgr_mark_task_in_progress(p); + + /* wake up kthreads, they will clean the progress flag */ + if (!p->mm) { + /* + * this is incorrect for kthreads waiting still for + * their first wake_up. + */ + wake_up_process(p); + } + } read_unlock(&tasklist_lock); } @@ -274,8 +280,7 @@ static int kgr_patch_code(const struct kgr_patch_fun *patch_fun, bool final) * kgr_start_patching -- the entry for a kgraft patch * @patch: patch to be applied * - * Start patching of code that is neither running in IRQ context nor - * kernel thread. + * Start patching of code that is not running in IRQ context. */ int kgr_start_patching(const struct kgr_patch *patch) { @@ -314,7 +319,7 @@ int kgr_start_patching(const struct kgr_patch *patch) kgr_patch = patch; mutex_unlock(&kgr_in_progress_lock); - kgr_mark_processes(); + kgr_handle_processes(); /* * give everyone time to exit kernel, and check after a while -- 2.0.0