summaryrefslogtreecommitdiff
blob: d5602616880959bfb1df87ee71e11fd35a5c32f5 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
From 2e5b17d01347d3c3118be2b8ad63d20415dbb1f0 Mon Sep 17 00:00:00 2001
From: Kay Sievers <kay@vrfy.org>
Date: Thu, 24 Jul 2014 23:37:35 +0200
Subject: [PATCH] udev: exclude MD from block device ownership event locking

MD instantiates devices at open(). This is incomptible with the
locking logic, as the "change" event emitted when stopping a
device will bring it back.
---
 src/udev/udevd.c | 23 +++++++----------------
 1 file changed, 7 insertions(+), 16 deletions(-)

diff --git a/src/udev/udevd.c b/src/udev/udevd.c
index a45d324..db935d6 100644
--- a/src/udev/udevd.c
+++ b/src/udev/udevd.c
@@ -285,26 +285,17 @@ static void worker_new(struct event *event)
                                 udev_event->exec_delay = exec_delay;
 
                         /*
-                         * Take a "read lock" on the device node; this establishes
+                         * Take a shared lock on the device node; this establishes
                          * a concept of device "ownership" to serialize device
-                         * access. External processes holding a "write lock" will
+                         * access. External processes holding an exclusive lock will
                          * cause udev to skip the event handling; in the case udev
-                         * acquired the lock, the external process will block until
+                         * acquired the lock, the external process can block until
                          * udev has finished its event handling.
                          */
-
-                        /*
-                         * <kabi_> since we make check - device seems unused - we try
-                         *         ioctl to deactivate - and device is found to be opened
-                         * <kay> sure, you try to take a write lock
-                         * <kay> if you get it udev is out
-                         * <kay> if you can't get it, udev is busy
-                         * <kabi_> we cannot deactivate openned device  (as it is in-use)
-                         * <kay> maybe we should just exclude dm from that thing entirely
-                         * <kabi_> IMHO this sounds like a good plan for this moment
-                         */
-                        if (streq_ptr("block", udev_device_get_subsystem(dev)) &&
-                            !startswith(udev_device_get_sysname(dev), "dm-")) {
+                        if (!streq_ptr(udev_device_get_action(dev), "remove") &&
+                            streq_ptr("block", udev_device_get_subsystem(dev)) &&
+                            !startswith(udev_device_get_sysname(dev), "dm-") &&
+                            !startswith(udev_device_get_sysname(dev), "md")) {
                                 struct udev_device *d = dev;
 
                                 if (streq_ptr("partition", udev_device_get_devtype(d)))
-- 
1.8.5.5