From: Kay Sievers <kay.sievers@vrfy.org>
To: Michael Guntsche <mike@it-loops.com>
Cc: Alan Jenkins <sourcejedi.lkml@googlemail.com>,
linux-kernel <linux-kernel@vger.kernel.org>,
linux-hotplug <linux-hotplug@vger.kernel.org>
Subject: Re: [BUG] No ttyS0 - Strange udev 8250_pnp interaction
Date: Wed, 07 Oct 2009 21:23:22 +0000 [thread overview]
Message-ID: <1254950602.2307.1.camel@yio.site> (raw)
In-Reply-To: <ac3eb2510910071353v3bbe5bc7lc41e184fbc1ff7ed@mail.gmail.com>
On Wed, 2009-10-07 at 22:53 +0200, Kay Sievers wrote:
> On Wed, Oct 7, 2009 at 22:31, Michael Guntsche <mike@it-loops.com> wrote:
> > add /module/8250_pnp (module)
> > remove /devices/platform/serial8250/tty/ttyS0 (tty)
> > add /devices/pnp0/00:05/tty/ttyS0 (tty)
>
> That's pretty odd, that the loading of a module removes pre-existing
> devices created by other code, and creates new ones with the same name
> and major/minor.
>
> As Alan mentioned, we could serialize events which carry the the same
> major/minor number, that should work around such behavior.
Michael, you mentioned trying the udev git version. Any chance to check
if that fixes it?
Thanks,
Kay
diff --git a/udev/udevd.c b/udev/udevd.c
index 62c6436..dfdbb4c 100644
--- a/udev/udevd.c
+++ b/udev/udevd.c
@@ -118,6 +118,8 @@ struct event {
const char *devpath;
size_t devpath_len;
const char *devpath_old;
+ dev_t devnum;
+ bool is_block;
};
static struct event *node_to_event(struct udev_list_node *node)
@@ -410,6 +412,8 @@ static void event_queue_insert(struct udev_device *dev)
event->devpath = udev_device_get_devpath(dev);
event->devpath_len = strlen(event->devpath);
event->devpath_old = udev_device_get_devpath_old(dev);
+ event->devnum = udev_device_get_devnum(dev);
+ event->is_block = (strcmp("block", udev_device_get_subsystem(dev)) = 0);
udev_queue_export_device_queued(udev_queue_export, dev);
info(event->udev, "seq %llu queued, '%s' '%s'\n", udev_device_get_seqnum(dev),
@@ -473,7 +477,7 @@ static int mem_size_mb(void)
}
/* lookup event for identical, parent, child device */
-static int devpath_busy(struct event *event)
+static bool is_devpath_busy(struct event *event)
{
struct udev_list_node *loop;
size_t common;
@@ -488,18 +492,21 @@ static int devpath_busy(struct event *event)
/* event we checked earlier still exists, no need to check again */
if (loop_event->seqnum = event->delaying_seqnum)
- return 2;
+ return true;
/* found ourself, no later event can block us */
if (loop_event->seqnum >= event->seqnum)
break;
+ /* check major/minor */
+ if (major(event->devnum) != 0 && event->devnum = loop_event->devnum && event->is_block = loop_event->is_block)
+ return true;
+
/* check our old name */
- if (event->devpath_old != NULL)
- if (strcmp(loop_event->devpath, event->devpath_old) = 0) {
- event->delaying_seqnum = loop_event->seqnum;
- return 3;
- }
+ if (event->devpath_old != NULL && strcmp(loop_event->devpath, event->devpath_old) = 0) {
+ event->delaying_seqnum = loop_event->seqnum;
+ return true;
+ }
/* compare devpath */
common = MIN(loop_event->devpath_len, event->devpath_len);
@@ -511,26 +518,26 @@ static int devpath_busy(struct event *event)
/* identical device event found */
if (loop_event->devpath_len = event->devpath_len) {
event->delaying_seqnum = loop_event->seqnum;
- return 4;
+ return true;
}
/* parent device event found */
if (event->devpath[common] = '/') {
event->delaying_seqnum = loop_event->seqnum;
- return 5;
+ return true;
}
/* child device event found */
if (loop_event->devpath[common] = '/') {
event->delaying_seqnum = loop_event->seqnum;
- return 6;
+ return true;
}
/* no matching device */
continue;
}
- return 0;
+ return false;
}
static void events_start(struct udev *udev)
@@ -544,7 +551,7 @@ static void events_start(struct udev *udev)
continue;
/* do not start event if parent or child event is still running */
- if (devpath_busy(event) != 0) {
+ if (is_devpath_busy(event)) {
dbg(udev, "delay seq %llu (%s)\n", event->seqnum, event->devpath);
continue;
}
next prev parent reply other threads:[~2009-10-07 21:23 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20091007175403.GA335@trillian.comsick.at>
2009-10-07 19:46 ` [BUG] No ttyS0 - Strange udev 8250_pnp interaction Alan Jenkins
2009-10-07 20:06 ` Kay Sievers
2009-10-07 20:12 ` Kay Sievers
2009-10-07 20:31 ` Michael Guntsche
2009-10-07 20:53 ` Kay Sievers
2009-10-07 21:10 ` Alan Cox
2009-10-07 21:15 ` Kay Sievers
2009-10-07 21:23 ` Kay Sievers [this message]
2009-10-07 21:52 ` Michael Guntsche
2009-10-07 22:11 ` Kay Sievers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1254950602.2307.1.camel@yio.site \
--to=kay.sievers@vrfy.org \
--cc=linux-hotplug@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mike@it-loops.com \
--cc=sourcejedi.lkml@googlemail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).