Attachment 304428 Details for Bug 432979 – Patch to reduce ftrace printk output

[patch] Patch to reduce ftrace printk output

mrg-ftrace-remove-prink.patch (text/plain), 2.24 KB, created by Clark Williams on 2008-05-02 22:09:49 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Clark Williams

Created: 2008-05-02 22:09:49 UTC

Size: 2.24 KB

patch

obsolete

>Subject: [PATCH] remove printks from ftrace max latencies
>
>From: Steven Rostedt <srostedt@redhat.com>
>
>When ftrace wakeup latency or one of the irqs off /preempt off variants hit a max latency they perform a printk. Unfortunately, these latencies happen very often in the scheduler, or other areas that are dangerous to perform a printk.
>
>Since the latencies are recorded in userspace debugfs files, there's no reason to spam the logs with new latencies, and cause areas of concern. This patch simply removes the printks from those critical sections.
>---
>
> kernel/trace/trace_irqsoff.c      |   14 --------------
> kernel/trace/trace_sched_wakeup.c |   13 -------------
> 2 files changed, 0 insertions(+), 27 deletions(-)
>
>
>diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c
>index 0547950..0bd7e16 100644
>--- a/kernel/trace/trace_irqsoff.c
>+++ b/kernel/trace/trace_irqsoff.c
>@@ -166,20 +166,6 @@ check_critical_timing(struct trace_array *tr,
> 
> 	update_max_tr_single(tr, current, cpu);
> 
>-	if (tracing_thresh) {
>-		printk(KERN_INFO "(%16s-%-5d|#%d):"
>-			" %lu us critical section violates %lu us threshold.\n",
>-				current->comm, current->pid,
>-				raw_smp_processor_id(),
>-				latency, nsecs_to_usecs(tracing_thresh));
>-	} else {
>-		printk(KERN_INFO "(%16s-%-5d|#%d):"
>-		       " new %lu us maximum-latency critical section.\n",
>-				current->comm, current->pid,
>-				raw_smp_processor_id(),
>-				latency);
>-	}
>-
> 	max_sequence++;
> 
> out_unlock:
>diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c
>index 93e037a..ffd0871 100644
>--- a/kernel/trace/trace_sched_wakeup.c
>+++ b/kernel/trace/trace_sched_wakeup.c
>@@ -109,19 +109,6 @@ wakeup_sched_switch(struct task_struct *prev, struct task_struct *next)
> 
> 	update_max_tr(tr, wakeup_task, wakeup_cpu);
> 
>-	if (tracing_thresh) {
>-		printk(KERN_INFO "(%16s-%-5d|#%d):"
>-			" %lu us wakeup latency violates %lu us threshold.\n",
>-				wakeup_task->comm, wakeup_task->pid,
>-				raw_smp_processor_id(),
>-				latency, nsecs_to_usecs(tracing_thresh));
>-	} else {
>-		printk(KERN_INFO "(%16s-%-5d|#%d):"
>-			" new %lu us maximum wakeup latency.\n",
>-				wakeup_task->comm, wakeup_task->pid,
>-				cpu, latency);
>-	}
>-
> out_unlock:
> 	__wakeup_reset(tr);
> 	spin_unlock_irqrestore(&wakeup_lock, flags);

Actions: View | Diff

Attachments on bug 432979: 296255 | 304428