Integrate udevsettle with udevtrigger?

Integrate udevsettle with udevtrigger?

[Bryan Kadzban]

[-- Attachment #1: Type: text/plain, Size: 2048 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

It seems to me that udevtrigger and udevsettle are strongly related
(generate all the uevents, then wait for them to finish -- this seems
like it's is all part of the same "task").  Therefore I don't really see
any reason to run udevtrigger without udevsettle.  I doubt any distro's
bootscripts are going to run the one without the other, for instance,
but even running udevtrigger manually should (IMO) wait for events to
settle out.

Would it make sense to anyone else to integrate the two programs, by
adding an option to udevtrigger that would tell it to wait?

If so, I've attached one possible patch to do that, against udev-090.
It adds support in udevtrigger (and documentation) for a --wait option,
which when specified will cause udevtrigger to call a new do_uevent_wait
function.  do_uevent_wait is basically the guts out of udevsettle.
- --wait will also take an argument: the number of seconds to wait (this
is the same option that udevsettle took).

(I basically cut-and-pasted udevsettle's code into the function; it's
possible that something got screwed up there.  The code compiles OK for
me, though, and it hasn't leaked any more uevents than "udevtrigger &&
udevsettle" over ~6-8 bootups.  To check for leaks, I'm doing what
Alexander did[1]: putting an ACTION=="add" RUN+="bug" rule last,
clearing out the /dev/bug file after the udevtrigger --wait call,
sleeping 6 seconds, and looking for anything in /dev/bug.  Both setups
leak events in the presence of usb-storage devices, but I think that's a
known issue.  If I have only USB-HID devices plugged in, that works OK
whether I use "udevtrigger --wait" or "udevtrigger && udevsettle".)

Comments?

[1] http://sourceforge.net/mailarchive/message.php?msg_id=15413263

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFERCHJi5mq6LtIi44RA6erAJ9DcSoVaSgSIhvWrpXdLXNu9j+MnwCfX6iE
ZGwFpzNryvZsou+xW5W1jj4=
=K4JV
-----END PGP SIGNATURE-----

[-- Attachment #2: udev-090-integrate-settle-with-trigger.patch --]
[-- Type: text/plain, Size: 13364 bytes --]

diff -Naur udev-090/udevsettle.8 udev-090-patched/udevsettle.8
--- udev-090/udevsettle.8	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevsettle.8	1969-12-31 19:00:00.000000000 -0500
@@ -1,32 +0,0 @@
-.\" ** You probably do not want to edit this file directly **
-.\" It was generated using the DocBook XSL Stylesheets (version 1.69.1).
-.\" Instead of manually editing it, you probably should edit the DocBook XML
-.\" source for it and then use the DocBook XSL Stylesheets to regenerate it.
-.TH "UDEVSETTLE" "8" "March 2006" "udev" "udevsettle"
-.\" disable hyphenation
-.nh
-.\" disable justification (adjust text to left margin only)
-.ad l
-.SH "NAME"
-udevsettle \- wait until queued kernel/udev events are handled
-.SH "SYNOPSIS"
-.HP 11
-\fBudevsettle\fR [\fB\-\-timeout=\fR\fB\fIseconds\fR\fR]
-.SH "DESCRIPTION"
-.PP
-Waits watching the udev event queue and exits if all current events are handled.
-.SH "OPTIONS"
-.TP
-\fB\-\-timeout=\fR\fB\fIseconds\fR\fR
-maximum seconds to wait for the queue to become empty.
-.SH "ENVIRONMENT"
-.TP
-\fBUDEV_LOG\fR
-Overrides the syslog priority specified in the config file.
-.SH "AUTHOR"
-.PP
-Written by Kay Sievers
-<kay.sievers@vrfy.org>.
-.SH "SEE ALSO"
-.PP
-\fBudev\fR(7)
diff -Naur udev-090/udevsettle.c udev-090-patched/udevsettle.c
--- udev-090/udevsettle.c	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevsettle.c	1969-12-31 19:00:00.000000000 -0500
@@ -1,156 +0,0 @@
-/*
- * udevsettle.c
- *
- * Copyright (C) 2006 Kay Sievers <kay@vrfy.org>
- *
- *	This program is free software; you can redistribute it and/or modify it
- *	under the terms of the GNU General Public License as published by the
- *	Free Software Foundation version 2 of the License.
- * 
- *	This program is distributed in the hope that it will be useful, but
- *	WITHOUT ANY WARRANTY; without even the implied warranty of
- *	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- *	General Public License for more details.
- * 
- *	You should have received a copy of the GNU General Public License along
- *	with this program; if not, write to the Free Software Foundation, Inc.,
- *	675 Mass Ave, Cambridge, MA 02139, USA.
- *
- */
-
-#include <stdlib.h>
-#include <stddef.h>
-#include <string.h>
-#include <stdio.h>
-#include <unistd.h>
-#include <errno.h>
-#include <dirent.h>
-#include <fcntl.h>
-#include <syslog.h>
-#include <sys/stat.h>
-#include <sys/types.h>
-
-#include "udev.h"
-#include "udevd.h"
-
-#define DEFAULT_TIMEOUT			180
-#define LOOP_PER_SECOND			20
-
-static const char *udev_log_str;
-
-#ifdef USE_LOG
-void log_message(int priority, const char *format, ...)
-{
-	va_list args;
-
-	if (priority > udev_log_priority)
-		return;
-
-	va_start(args, format);
-	vsyslog(priority, format, args);
-	va_end(args);
-}
-#endif
-
-int main(int argc, char *argv[], char *envp[])
-{
-	char queuename[PATH_SIZE];
-	char filename[PATH_SIZE];
-	unsigned long long seq_kernel;
-	unsigned long long seq_udev;
-	char seqnum[32];
-	int fd;
-	ssize_t len;
-	int timeout = DEFAULT_TIMEOUT;
-	int loop;
-	int i;
-	int rc = 1;
-
-	logging_init("udevsettle");
-	udev_config_init();
-	dbg("version %s", UDEV_VERSION);
-
-	udev_log_str = getenv("UDEV_LOG");
-
-	for (i = 1 ; i < argc; i++) {
-		char *arg = argv[i];
-
-		if (strncmp(arg, "--timeout=", 10) == 0) {
-			char *str = &arg[10];
-
-			timeout = atoi(str);
-			dbg("timeout=%i", timeout);
-			if (timeout <= 0) {
-				fprintf(stderr, "Invalid timeout value.\n");
-				goto exit;
-			}
-		} else {
-			fprintf(stderr, "Usage: udevsettle [--timeout=<seconds>]\n");
-			goto exit;
-		}
-	}
-
-	sysfs_init();
-	strlcpy(queuename, udev_root, sizeof(queuename));
-	strlcat(queuename, "/" EVENT_QUEUE_DIR, sizeof(queuename));
-
-	loop = timeout * LOOP_PER_SECOND;
-	while (loop--) {
-		/* wait for events in queue to finish */
-		while (loop--) {
-			struct stat statbuf;
-
-			if (stat(queuename, &statbuf) < 0) {
-				info("queue is empty");
-				break;
-			}
-			usleep(1000 * 1000 / LOOP_PER_SECOND);
-		}
-		if (loop <= 0) {
-			info("timeout waiting for queue");
-			goto exit;
-		}
-
-		/* read current kernel seqnum */
-		strlcpy(filename, sysfs_path, sizeof(filename));
-		strlcat(filename, "/kernel/uevent_seqnum", sizeof(filename));
-		fd = open(filename, O_RDONLY);
-		if (fd < 0)
-			goto exit;
-		len = read(fd, seqnum, sizeof(seqnum)-1);
-		close(fd);
-		if (len <= 0)
-			goto exit;
-		seqnum[len] = '\0';
-		seq_kernel = strtoull(seqnum, NULL, 10);
-		info("kernel seqnum = %llu", seq_kernel);
-
-		/* read current udev seqnum */
-		strlcpy(filename, udev_root, sizeof(filename));
-		strlcat(filename, "/" EVENT_SEQNUM, sizeof(filename));
-		fd = open(filename, O_RDONLY);
-		if (fd < 0)
-			goto exit;
-		len = read(fd, seqnum, sizeof(seqnum)-1);
-		close(fd);
-		if (len <= 0)
-			goto exit;
-		seqnum[len] = '\0';
-		seq_udev = strtoull(seqnum, NULL, 10);
-		info("udev seqnum = %llu", seq_udev);
-
-		/* make sure all kernel events have arrived in the queue */
-		if (seq_udev >= seq_kernel) {
-			info("queue is empty and no pending events left");
-			rc = 0;
-			goto exit;
-		}
-		usleep(1000 * 1000 / LOOP_PER_SECOND);
-		info("queue is empty, but events still pending");
-	}
-
-exit:
-	sysfs_cleanup();
-	logging_close();
-	return rc;
-}
diff -Naur udev-090/udevsettle.xml udev-090-patched/udevsettle.xml
--- udev-090/udevsettle.xml	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevsettle.xml	1969-12-31 19:00:00.000000000 -0500
@@ -1,82 +0,0 @@
-<?xml version='1.0'?>
-<!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
-  "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
-
-<article>
-  <articleinfo>
-    <title>xmlto</title>
-    <author>
-      <firstname>Kay</firstname>
-      <surname>Sievers</surname>
-      <email>kay.sievers@vrfy.org</email>
-    </author>
-    <copyright>
-      <year>2006</year>
-      <holder>Kay Sievers</holder>
-    </copyright>
-  </articleinfo>
-
-  <section>
-    <title>udevsettle</title>
-    <refentry>
-      <refentryinfo>
-        <title>udevsettle</title>
-        <date>March 2006</date>
-        <productname>udev</productname>
-      </refentryinfo>
-
-      <refmeta>
-        <refentrytitle>udevsettle</refentrytitle>
-        <manvolnum>8</manvolnum>
-      </refmeta>
-
-      <refnamediv>
-        <refname>udevsettle</refname><refpurpose>wait until queued kernel/udev events are handled</refpurpose>
-      </refnamediv>
-
-      <refsynopsisdiv>
-        <cmdsynopsis>
-          <command>udevsettle</command>
-          <arg><option>--timeout=<replaceable>seconds</replaceable></option></arg>
-        </cmdsynopsis>
-      </refsynopsisdiv>
-
-      <refsect1><title>DESCRIPTION</title>
-        <para>Waits watching the udev event queue and exits if all current events are handled.</para>
-      </refsect1>
-
-      <refsect1><title>OPTIONS</title>
-        <variablelist>
-          <varlistentry>
-            <term><option>--timeout=<replaceable>seconds</replaceable></option></term>
-            <listitem>
-              <para>maximum seconds to wait for the queue to become empty.</para>
-            </listitem>
-          </varlistentry>
-        </variablelist>
-      </refsect1>
-
-      <refsect1><title>ENVIRONMENT</title>
-        <variablelist>
-          <varlistentry>
-            <term><option>UDEV_LOG</option></term>
-            <listitem>
-              <para>Overrides the syslog priority specified in the config file.</para>
-            </listitem>
-          </varlistentry>
-        </variablelist>
-     </refsect1>
-
-      <refsect1><title>AUTHOR</title>
-        <para>Written by Kay Sievers <email>kay.sievers@vrfy.org</email>.</para>
-      </refsect1>
-
-      <refsect1>
-        <title>SEE ALSO</title>
-        <para><citerefentry>
-            <refentrytitle>udev</refentrytitle><manvolnum>7</manvolnum>
-        </citerefentry></para>
-      </refsect1>
-    </refentry>
-  </section>
-</article>
diff -Naur udev-090/udevtrigger.8 udev-090-patched/udevtrigger.8
--- udev-090/udevtrigger.8	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevtrigger.8	2006-04-17 18:43:17.000000000 -0400
@@ -11,7 +11,7 @@
 udevtrigger \- request kernel devices events for coldplug
 .SH "SYNOPSIS"
 .HP 12
-\fBudevtrigger\fR [\fB\-\-verbose\fR] [\fB\-\-dry\-run\fR]
+\fBudevtrigger\fR [\fB\-\-verbose\fR] [\fB\-\-dry\-run\fR] [\fB\-\-wait\fR\fB[=\fIseconds\fR]\fR]
 .SH "DESCRIPTION"
 .PP
 Trigger kernel device uevents to replay missing events at system coldplug.
@@ -22,6 +22,11 @@
 .TP
 \fB\-\-dry\-run\fR
 don't actually trigger the event.
+.TP
+\fB\-\-wait\fR\fB[=\fIseconds\fR]\fR
+Whether to wait for the uevent queue to empty. If specified,
+\fIseconds\fR
+determines how long to wait. Default is 180.
 .SH "ENVIRONMENT"
 .TP
 \fBUDEV_LOG\fR
diff -Naur udev-090/udevtrigger.c udev-090-patched/udevtrigger.c
--- udev-090/udevtrigger.c	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevtrigger.c	2006-04-17 18:40:08.000000000 -0400
@@ -32,6 +32,10 @@
 #include <sys/types.h>
 
 #include "udev.h"
+#include "udevd.h"
+
+#define DEFAULT_TIMEOUT			180
+#define LOOP_PER_SECOND			20
 
 static const char *udev_log_str;
 static int verbose;
@@ -318,9 +322,85 @@
 	}
 }
 
+static int do_uevent_wait(int timeout)
+{
+	char queuename[PATH_SIZE];
+	char filename[PATH_SIZE];
+	unsigned long long seq_kernel;
+	unsigned long long seq_udev;
+	char seqnum[32];
+	int fd;
+	ssize_t len;
+	int loop;
+	int rc = 1;
+
+	strlcpy(queuename, udev_root, sizeof(queuename));
+	strlcat(queuename, "/" EVENT_QUEUE_DIR, sizeof(queuename));
+
+	loop = timeout * LOOP_PER_SECOND;
+	while (loop--) {
+		/* wait for events in queue to finish */
+		while (loop--) {
+			struct stat statbuf;
+
+			if (stat(queuename, &statbuf) < 0) {
+				info("queue is empty");
+				break;
+			}
+			usleep(1000 * 1000 / LOOP_PER_SECOND);
+		}
+		if (loop <= 0) {
+			info("timeout waiting for queue");
+			goto exit;
+		}
+
+		/* read current kernel seqnum */
+		strlcpy(filename, sysfs_path, sizeof(filename));
+		strlcat(filename, "/kernel/uevent_seqnum", sizeof(filename));
+		fd = open(filename, O_RDONLY);
+		if (fd < 0)
+			goto exit;
+		len = read(fd, seqnum, sizeof(seqnum)-1);
+		close(fd);
+		if (len <= 0)
+			goto exit;
+		seqnum[len] = '\0';
+		seq_kernel = strtoull(seqnum, NULL, 10);
+		info("kernel seqnum = %llu", seq_kernel);
+
+		/* read current udev seqnum */
+		strlcpy(filename, udev_root, sizeof(filename));
+		strlcat(filename, "/" EVENT_SEQNUM, sizeof(filename));
+		fd = open(filename, O_RDONLY);
+		if (fd < 0)
+			goto exit;
+		len = read(fd, seqnum, sizeof(seqnum)-1);
+		close(fd);
+		if (len <= 0)
+			goto exit;
+		seqnum[len] = '\0';
+		seq_udev = strtoull(seqnum, NULL, 10);
+		info("udev seqnum = %llu", seq_udev);
+
+		/* make sure all kernel events have arrived in the queue */
+		if (seq_udev >= seq_kernel) {
+			info("queue is empty and no pending events left");
+			rc = 0;
+			goto exit;
+		}
+		usleep(1000 * 1000 / LOOP_PER_SECOND);
+		info("queue is empty, but events still pending");
+	}
+
+exit:
+	return rc;
+}
+
 int main(int argc, char *argv[], char *envp[])
 {
 	int i;
+	int wait_for_uevents = 0, uevent_wait_timeout = DEFAULT_TIMEOUT;
+	int rc = 0;
 
 	logging_init("udevtrigger");
 	udev_config_init();
@@ -335,8 +415,23 @@
 			verbose = 1;
 		else if (strcmp(arg, "--dry-run") == 0 || strcmp(arg, "-n") == 0)
 			dry_run = 1;
+		else if (strncmp(arg, "--wait", 6) == 0) {
+			wait_for_uevents = 1;
+
+			if (arg[6] == '=') {
+				char *str = &arg[7];
+
+				uevent_wait_timeout = atoi(str);
+				dbg("timeout=%i", uevent_wait_timeout);
+				if(uevent_wait_timeout <= 0) {
+					fprintf(stderr, "Invalid timeout value.\n");
+					goto exit;
+				}
+			}
+		}
 		else {
-			fprintf(stderr, "Usage: udevtrigger [--verbose] [--dry-run]\n");
+			fprintf(stderr, "Usage: udevtrigger [--verbose] [--dry-run] [--wait[=<seconds>]]\n");
+			fprintf(stderr, "Default is to not wait.  If no <seconds> timeout is provided, default is to wait %d seconds.\n", DEFAULT_TIMEOUT);
 			goto exit;
 		}
 	}
@@ -348,8 +443,11 @@
 	udev_scan_block();
 	exec_lists();
 
+	if(wait_for_uevents)
+		rc = do_uevent_wait(uevent_wait_timeout);
+
 	sysfs_cleanup();
 exit:
 	logging_close();
-	return 0;
+	return rc;
 }
diff -Naur udev-090/udevtrigger.xml udev-090-patched/udevtrigger.xml
--- udev-090/udevtrigger.xml	2006-04-15 13:32:38.000000000 -0400
+++ udev-090-patched/udevtrigger.xml	2006-04-17 18:43:11.000000000 -0400
@@ -39,6 +39,7 @@
           <command>udevtrigger</command>
           <arg><option>--verbose</option></arg>
           <arg><option>--dry-run</option></arg>
+          <arg><option>--wait<optional>=<replaceable>seconds</replaceable></optional></option></arg>
         </cmdsynopsis>
       </refsynopsisdiv>
 
@@ -60,6 +61,12 @@
               <para>don't actually trigger the event.</para>
             </listitem>
           </varlistentry>
+          <varlistentry>
+            <term><option>--wait<optional>=<replaceable>seconds</replaceable></optional></option></term>
+            <listitem>
+              <para>Whether to wait for the uevent queue to empty. If specified, <replaceable>seconds</replaceable> determines how long to wait. Default is 180.</para>
+            </listitem>
+          </varlistentry>
         </variablelist>
       </refsect1>
 

Re: Integrate udevsettle with udevtrigger?

[Alexander E. Patrakov]

Bryan Kadzban wrote:
> (I basically cut-and-pasted udevsettle's code into the function; it's
> possible that something got screwed up there.  The code compiles OK for
> me, though, and it hasn't leaked any more uevents than "udevtrigger &&
> udevsettle" over ~6-8 bootups.  To check for leaks, I'm doing what
> Alexander did[1]: putting an ACTION="add" RUN+="bug" rule last,
> clearing out the /dev/bug file after the udevtrigger --wait call,
> sleeping 6 seconds, and looking for anything in /dev/bug.  Both setups
> leak events in the presence of usb-storage devices, but I think that's a
> known issue.  If I have only USB-HID devices plugged in, that works OK
> whether I use "udevtrigger --wait" or "udevtrigger && udevsettle".)

Does the leak go away if you use the "ub" driver instead of "usb-storage"?

-- 
Alexander E. Patrakov


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Alexander E. Patrakov]

Bryan Kadzban wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
> 
> It seems to me that udevtrigger and udevsettle are strongly related
> (generate all the uevents, then wait for them to finish -- this seems
> like it's is all part of the same "task").  Therefore I don't really see
> any reason to run udevtrigger without udevsettle.  I doubt any distro's
> bootscripts are going to run the one without the other, for instance,
> but even running udevtrigger manually should (IMO) wait for events to
> settle out.
> 
> Would it make sense to anyone else to integrate the two programs, by
> adding an option to udevtrigger that would tell it to wait?

It doesn't make sense to remove udevsettle as a separate program. Reason: it is
needed on its own in the udev_retry script. See
http://wiki.linuxfromscratch.org/lfs/ticket/1769.

-- 
Alexander E. Patrakov



-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Bryan Kadzban]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Alexander E. Patrakov wrote:
> Bryan Kadzban wrote:
> 
>> Would it make sense to anyone else to integrate the two programs,
>> by adding an option to udevtrigger that would tell it to wait?
> 
> It doesn't make sense to remove udevsettle as a separate program. 
> Reason: it is needed on its own in the udev_retry script.
> 

Ah.  OK, I didn't look into anyone's udev_retry script (my /usr isn't a
separate partition, so anything that fails the first time through, like
PCI bus modaliases, etc., is going to fail any other time through).

I'd ask whether it would make sense to automatically exec udevsettle if
- --wait was given to udevtrigger, but it's likely not worth it.

So never mind then.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFERFmpi5mq6LtIi44RA+sSAJ9QtaA8zM/8vBLfxhM8ms4oCKwCfwCggLH7
CFIPd0d2l8xpiA17CDdA9G0=Yx9W
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Marco d'Itri]

On Apr 18, Bryan Kadzban <bryan@kadzban.is-a-geek.net> wrote:

> It seems to me that udevtrigger and udevsettle are strongly related
I think it would be a good idea to have a single executable to save
space in the initramfs and install disks, but it should be possible to
independently use each function.

-- 
ciao,
Marco


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Kay Sievers]

On Mon, Apr 17, 2006 at 07:16:30PM -0400, Bryan Kadzban wrote:
> It seems to me that udevtrigger and udevsettle are strongly related
> (generate all the uevents, then wait for them to finish -- this seems
> like it's is all part of the same "task").  Therefore I don't really see
> any reason to run udevtrigger without udevsettle.  I doubt any distro's
> bootscripts are going to run the one without the other, for instance,
> but even running udevtrigger manually should (IMO) wait for events to
> settle out.
> 
> Would it make sense to anyone else to integrate the two programs, by
> adding an option to udevtrigger that would tell it to wait?

udevsettle is needed in some other cases too, like repartitioning
a disk and wait for the partition devices to be handled. Or triggering
a "uevent" for a device to update its state like a changed volume label
and wait for udev to finish updating the /dev/disk/by-* links.

> If so, I've attached one possible patch to do that, against udev-090.
> It adds support in udevtrigger (and documentation) for a --wait option,
> which when specified will cause udevtrigger to call a new do_uevent_wait
> function.  do_uevent_wait is basically the guts out of udevsettle.
> - --wait will also take an argument: the number of seconds to wait (this
> is the same option that udevsettle took).
> 
> (I basically cut-and-pasted udevsettle's code into the function; it's
> possible that something got screwed up there.  The code compiles OK for
> me, though, and it hasn't leaked any more uevents than "udevtrigger &&
> udevsettle" over ~6-8 bootups.  To check for leaks, I'm doing what
> Alexander did[1]: putting an ACTION="add" RUN+="bug" rule last,
> clearing out the /dev/bug file after the udevtrigger --wait call,
> sleeping 6 seconds, and looking for anything in /dev/bug.  Both setups
> leak events in the presence of usb-storage devices, but I think that's a
> known issue.

There are some cases where the kernel schedules work and there is
no event pending, not in the kernel and not for udev, but not all devices
are created. You need to add that logic to the stuff that depends on the
devices, like waiting for all devices mentioned in fstab or try to detect
the kernel thread that signifies the pending work, that's nothing
we can really solve in udev.

Kay


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Andrey Borzenkov]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 18 April 2006 07:14, Bryan Kadzban wrote:
> Ah.  OK, I didn't look into anyone's udev_retry script (my /usr isn't a
> separate partition, so anything that fails the first time through, like
> PCI bus modaliases, etc., is going to fail any other time through).
>

At least on Mandriva udev dynamically creates (updates) rules for persistent 
cdrom and interfaces names. This fails on coldplug when root is still mounted 
read-only, so retrying after root has been remounted rw makes sense.

- -andrey
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFERQ0PR6LMutpd94wRAiciAJ4oHAXV/sJBJoQS0EbJgWjgGOJn8QCfTfip
nDXBafHUJtmqq1dPo8E72iA=OpOa
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid\x110944&bid$1720&dat\x121642
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Integrate udevsettle with udevtrigger?

[Bryan Kadzban]

[-- Attachment #1: Type: text/plain, Size: 2566 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Alexander E. Patrakov wrote:
> Bryan Kadzban wrote:
> 
>> Both setups leak events in the presence of usb-storage devices, but
>> I think that's a known issue.
> 
> Does the leak go away if you use the "ub" driver instead of 
> "usb-storage"?

Once I figured out *how* to use it, I've been able to boot once with no
leaks, and then I booted once with a leak.

(I had some problems because I enabled CONFIG_USB_LIBUSUAL.  When that
was on, hotplugging a USB device wouldn't load ub, and if I manually
modprobed ub, I didn't get any device files from udev.  With LIBUSUAL, I
also leaked USB-HID events on *every* boot; I assume this was from extra
overhead from the LIBUSUAL code.  Finally I disabled LIBUSUAL but kept
ub turned on, and that worked much better.)

I've attached the /dev/bugreport created by the udev script on the one
boot that leaked, if it matters.  It looks like both input and storage
class devices were leaking in this case.

And I realize how much of a pain it is to look into the future and say
"there are no more uevents coming", but it sure would be nice if there
was a way for the kernel to know for sure.  :-P

What might be possible is some kind of new kobject from each module,
which means "ok, I'm done with all my uevents now".  udevd could just
ignore the uevent; the new kobject could sit under the /sys/module/*
directory for the module in question.  Then it might be possible for the
wait loop in udevsettle to check whether everything in /sys/module has
the proper subdirectory, in addition to its current checks.

The catch with that is that even if /sys/module/* is a correct place to
store information like this, there's still a race between modprobe and
udevsettle.  (Say udevd has just modprobed a driver which hasn't
registered in /sys/module at all yet.  Assume also that the modprobe
process has exited, so udevd's "current sequence number" has updated.
If udevsettle runs at that point, it'll think everything is OK.  Unless
the registration in /sys/module is guaranteed to happen before modprobe
exits?  That would be a fix for the race, assuming the sequence number
exported by udevd doesn't update until modprobe exits.)  That may not be
a good place to store that information either, I don't know.

Comments?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFERXi7i5mq6LtIi44RA2WLAJwMro+YuDQ+gqkZntuoSJ6qxacIQACfZ0se
rEauY2HnpxWkoNbqZ271DcQ=
=nCnh
-----END PGP SIGNATURE-----

[-- Attachment #2: bugreport --]
[-- Type: text/plain, Size: 4111 bytes --]

UDEV_LOG=3
ACTION=add
DEVPATH=/devices/pci0000:00/0000:00:10.3/usb4/4-3
SUBSYSTEM=usb
SEQNUM=536
PHYSDEVBUS=usb
PHYSDEVDRIVER=usb
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/module/ub
SUBSYSTEM=module
SEQNUM=539
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/bus/usb/drivers/ub
SUBSYSTEM=drivers
SEQNUM=540
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2
SUBSYSTEM=usb
SEQNUM=541
PHYSDEVBUS=usb
PHYSDEVDRIVER=usb
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/devices/pci0000:00/0000:00:10.3/usb4/4-3/4-3:1.0
SUBSYSTEM=usb
SEQNUM=537
PHYSDEVBUS=usb
DEVICE=/proc/bus/usb/004/003
PRODUCT=8ec/8/100
TYPE=0/0/0
INTERFACE=8/6/80
MODALIAS=usb:v08ECp0008d0100dc00dsc00dp00ic08isc06ip50
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/class/usb_device/usbdev4.3
SUBSYSTEM=usb_device
SEQNUM=538
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.3/usb4/4-3
PHYSDEVBUS=usb
PHYSDEVDRIVER=usb
MAJOR=189
MINOR=386
UDEVD_EVENT=1
DEVNAME=/dev/bus/usb/4/3
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/block/uba
SUBSYSTEM=block
SEQNUM=544
MINOR=0
MAJOR=180
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.3/usb4/4-3/4-3:1.0
PHYSDEVBUS=usb
PHYSDEVDRIVER=ub
UDEVD_EVENT=1
ID_PATH=pci-0000:00:10.3-usb-0:3:1.0
DEVNAME=/dev/uba
DEVLINKS=/dev/disk/by-path/pci-0000:00:10.3-usb-0:3:1.0
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/module/usbhid
SUBSYSTEM=module
SEQNUM=546
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/bus/usb/drivers/usbhid
SUBSYSTEM=drivers
SEQNUM=547
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/block/uba/uba1
SUBSYSTEM=block
SEQNUM=545
MINOR=1
MAJOR=180
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.3/usb4/4-3/4-3:1.0
PHYSDEVBUS=usb
PHYSDEVDRIVER=ub
UDEVD_EVENT=1
ID_PATH=pci-0000:00:10.3-usb-0:3:1.0
ID_FS_USAGE=filesystem
ID_FS_TYPE=vfat
ID_FS_VERSION=FAT16
ID_FS_UUID=5F3D-53A4
ID_FS_LABEL=STORE'N'GO
ID_FS_LABEL_SAFE=STORE_N_GO
DEVNAME=/dev/uba1
DEVLINKS=/dev/disk/by-path/pci-0000:00:10.3-usb-0:3:1.0-part1 /dev/disk/by-uuid/5F3D-53A4 /dev/disk/by-label/STORE_N_GO
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2/1-2:1.0
SUBSYSTEM=usb
SEQNUM=542
PHYSDEVBUS=usb
DEVICE=/proc/bus/usb/001/002
PRODUCT=45e/39/121
TYPE=0/0/0
INTERFACE=3/1/2
MODALIAS=usb:v045Ep0039d0121dc00dsc00dp00ic03isc01ip02
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/class/usb_device/usbdev1.2
SUBSYSTEM=usb_device
SEQNUM=543
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2
PHYSDEVBUS=usb
PHYSDEVDRIVER=usb
MAJOR=189
MINOR=1
UDEVD_EVENT=1
DEVNAME=/dev/bus/usb/1/2
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/class/input/input2
SUBSYSTEM=input
SEQNUM=548
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2/1-2:1.0
PHYSDEVBUS=usb
PHYSDEVDRIVER=usbhid
PRODUCT=3/45e/39/121
NAME="Microsoft Microsoft IntelliMouse��� Optical"
PHYS="usb-0000:00:10.0-2/input0"
UNIQ=""
EV=7
KEY=1f0000 0 0 0 0 0 0 0 0
REL=103
MODALIAS=input:b0003v045Ep0039e0121-e0,1,2,k110,111,112,113,114,r0,1,8,amlsfw
UDEVD_EVENT=1
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/class/input/input2/mouse0
SUBSYSTEM=input
SEQNUM=549
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2/1-2:1.0
PHYSDEVBUS=usb
PHYSDEVDRIVER=usbhid
MAJOR=13
MINOR=32
UDEVD_EVENT=1
DEVNAME=/dev/input/mouse0
_SEPARATOR=--------------------------------------
UDEV_LOG=3
ACTION=add
DEVPATH=/class/input/input2/event2
SUBSYSTEM=input
SEQNUM=550
PHYSDEVPATH=/devices/pci0000:00/0000:00:10.0/usb1/1-2/1-2:1.0
PHYSDEVBUS=usb
PHYSDEVDRIVER=usbhid
MAJOR=13
MINOR=66
UDEVD_EVENT=1
DEVNAME=/dev/input/event2
_SEPARATOR=--------------------------------------