From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754668AbYJUQnV (ORCPT ); Tue, 21 Oct 2008 12:43:21 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751997AbYJUQnH (ORCPT ); Tue, 21 Oct 2008 12:43:07 -0400 Received: from hrndva-omtalb.mail.rr.com ([71.74.56.124]:43258 "EHLO hrndva-omtalb.mail.rr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751822AbYJUQnF (ORCPT ); Tue, 21 Oct 2008 12:43:05 -0400 Message-Id: <20081021164302.575069241@goodmis.org> References: <20081021164018.889518687@goodmis.org> User-Agent: quilt/0.46-1 Date: Tue, 21 Oct 2008 12:40:20 -0400 From: Steven Rostedt To: linux-kernel@vger.kernel.org Cc: Ingo Molnar , Thomas Gleixner , Peter Zijlstra , Andrew Morton , David Miller , Linus Torvalds , Steven Rostedt Subject: [PATCH 2/2] ftrace: release functions from hash Content-Disposition: inline; filename=ftrace-release-funcs-from-hash.patch Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The x86 architecture uses a static recording of mcount caller locations and is not affected by this patch. For architectures still using the dynamic ftrace daemon, this patch is critical. It removes the race between the recording of a function that calls mcount, the unloading of a module, and the ftrace daemon updating the call sites. This patch adds the releasing of the hash functions that the daemon uses to update the mcount call sites. When a module is unloaded, not only are the replaced call site table update, but now so is the hash recorded functions that the ftrace daemon will use. Again, architectures that implement MCOUNT_RECORD are not affected by this (which currently only x86 has). Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+) Index: linux-compile.git/kernel/trace/ftrace.c =================================================================== --- linux-compile.git.orig/kernel/trace/ftrace.c 2008-10-21 09:34:51.000000000 -0400 +++ linux-compile.git/kernel/trace/ftrace.c 2008-10-21 10:31:24.000000000 -0400 @@ -164,10 +164,14 @@ static DEFINE_SPINLOCK(ftrace_hash_lock) #define ftrace_hash_lock(flags) spin_lock_irqsave(&ftrace_hash_lock, flags) #define ftrace_hash_unlock(flags) \ spin_unlock_irqrestore(&ftrace_hash_lock, flags) +static void ftrace_release_hash(unsigned long start, unsigned long end); #else /* This is protected via the ftrace_lock with MCOUNT_RECORD. */ #define ftrace_hash_lock(flags) do { (void)(flags); } while (0) #define ftrace_hash_unlock(flags) do { } while(0) +static inline void ftrace_release_hash(unsigned long start, unsigned long end) +{ +} #endif /* @@ -354,6 +358,8 @@ void ftrace_release(void *start, unsigne } } spin_unlock(&ftrace_lock); + + ftrace_release_hash(s, e); } static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip) @@ -1686,6 +1692,44 @@ void __init ftrace_init(void) ftrace_disabled = 1; } #else /* CONFIG_FTRACE_MCOUNT_RECORD */ + +static void ftrace_release_hash(unsigned long start, unsigned long end) +{ + struct dyn_ftrace *rec; + struct hlist_node *t, *n; + struct hlist_head *head, temp_list; + unsigned long flags; + int i, cpu; + + preempt_disable_notrace(); + + /* disable incase we call something that calls mcount */ + cpu = raw_smp_processor_id(); + per_cpu(ftrace_shutdown_disable_cpu, cpu)++; + + ftrace_hash_lock(flags); + + for (i = 0; i < FTRACE_HASHSIZE; i++) { + INIT_HLIST_HEAD(&temp_list); + head = &ftrace_hash[i]; + + /* all CPUS are stopped, we are safe to modify code */ + hlist_for_each_entry_safe(rec, t, n, head, node) { + if (rec->flags & FTRACE_FL_FREE) + continue; + + if ((rec->ip >= start) && (rec->ip < end)) + ftrace_free_rec(rec); + } + } + + ftrace_hash_unlock(flags); + + per_cpu(ftrace_shutdown_disable_cpu, cpu)--; + preempt_enable_notrace(); + +} + static int ftraced(void *ignore) { unsigned long usecs; --