Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org, Li Huafei <lihuafei1@huawei.com>
Subject: [PATCH 4.19 06/34] tracing/kprobe: Fix kprobe_on_func_entry() modification
Date:   Fri, 24 Sep 2021 14:44:00 +0200
Message-Id: <20210924124330.173774576@linuxfoundation.org>
In-Reply-To: <20210924124329.965218583@linuxfoundation.org>
References: <20210924124329.965218583@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Li Huafei <lihuafei1@huawei.com>

The commit 960434acef37 ("tracing/kprobe: Fix to support kretprobe
events on unloaded modules") backport from v5.11, which modifies the
return value of kprobe_on_func_entry(). However, there is no adaptation
modification in create_trace_kprobe(), resulting in the exact opposite
behavior. Now we need to return an error immediately only if
kprobe_on_func_entry() returns -EINVAL.

Fixes: 960434acef37 ("tracing/kprobe: Fix to support kretprobe events on unloaded modules")
Signed-off-by: Li Huafei <lihuafei1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 kernel/trace/trace_kprobe.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -836,8 +836,9 @@ static int create_trace_kprobe(int argc,
 			pr_info("Failed to parse either an address or a symbol.\n");
 			return ret;
 		}
+		/* Defer the ENOENT case until register kprobe */
 		if (offset && is_return &&
-		    !kprobe_on_func_entry(NULL, symbol, offset)) {
+		    kprobe_on_func_entry(NULL, symbol, offset) == -EINVAL) {
 			pr_info("Given offset is not valid for return probe.\n");
 			return -EINVAL;
 		}