* [Patch] udev: new dasd_id tool
@ 2006-11-27 13:15 Volker Sameske
2006-11-27 15:52 ` Kay Sievers
` (11 more replies)
0 siblings, 12 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-27 13:15 UTC (permalink / raw)
To: linux-hotplug
The currently used uid (volume label) for s390 DASD devices is not really
unique. For that reason a unique uid sysfs-attribute for DASD devices has
been introduced. This new dasd_id tool replaces the current one and
evaluates this new sysfs attribute. In case it is not available, dasd_id
returns the old volume label instead.
Signed-of-by: Volker Sameske <sameske@de.ibm.com>
---
etc/udev/debian/persistent.rules | 2
etc/udev/redhat/50-udev.rules | 2
etc/udev/rules.d/60-persistent-storage.rules | 4
etc/udev/slackware/udev.rules | 2
etc/udev/suse/60-persistent-storage.rules | 2
extras/dasd_id/dasd_id.8 | 68 +-
extras/dasd_id/dasd_id.c | 731 +++++++++++++++++++++------
7 files changed, 636 insertions(+), 175 deletions(-)
diff -Nurp udev-103-orig/etc/udev/debian/persistent.rules udev-103-work/etc/udev/debian/persistent.rules
--- udev-103-orig/etc/udev/debian/persistent.rules 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/etc/udev/debian/persistent.rules 2006-11-27 13:10:17.000000000 +0100
@@ -30,7 +30,7 @@ KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SER
IMPORT{program}="scsi_id -g -x -a -s $devpath -d $tempnode"
KERNEL="dasd*[!0-9]", \
- IMPORT{program}="dasd_id --export $tempnode"
+ IMPORT{program}="dasd_id -u -e -d $tempnode"
KERNEL="*[!0-9]", \
IMPORT{program}="edd_id --export $tempnode"
diff -Nurp udev-103-orig/etc/udev/redhat/50-udev.rules udev-103-work/etc/udev/redhat/50-udev.rules
--- udev-103-orig/etc/udev/redhat/50-udev.rules 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/etc/udev/redhat/50-udev.rules 2006-11-27 13:11:06.000000000 +0100
@@ -277,7 +277,7 @@ KERNEL="sd*[!0-9]|sr*", SYSFS{ieee1394_
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/usb_id -x"
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/scsi_id -g -x -s %p -d $tempnode"
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/scsi_id -g -x -a -s %p -d $tempnode"
-KERNEL="dasd*[!0-9]", IMPORT{program}="/lib/udev/dasd_id --export $tempnode"
+KERNEL="dasd*[!0-9]", IMPORT{program}="/lib/udev/dasd_id -u -e -d $tempnode"
KERNEL="sd*[!0-9]|sr*|dasd*[!0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
# for partitions import parent information
diff -Nurp udev-103-orig/etc/udev/rules.d/60-persistent-storage.rules udev-103-work/etc/udev/rules.d/60-persistent-storage.rules
--- udev-103-orig/etc/udev/rules.d/60-persistent-storage.rules 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/etc/udev/rules.d/60-persistent-storage.rules 2006-11-27 13:08:08.000000000 +0100
@@ -23,7 +23,9 @@ KERNEL="sd*[!0-9]|sr*|st*", ATTRS{ieee1
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="usb_id -x"
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="scsi_id -g -x -s %p -d $tempnode"
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="scsi_id -g -x -a -s %p -d $tempnode"
-KERNEL="dasd*[!0-9]", IMPORT{program}="dasd_id --export $tempnode"
+KERNEL="dasd*[!0-9]", IMPORT{program}="dasd_id -u -e -d $tempnode"
+KERNEL="dasd*[!0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="dasd/$env{ID_SERIAL}/device"
+KERNEL="dasd*[0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="dasd/$env{ID_SERIAL}/part%n"
KERNEL="sd*[!0-9]|sr*|dasd*[!0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
KERNEL="sd*[0-9]|dasd*[0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}-part%n"
KERNEL="st*", ENV{ID_SERIAL}="?*", SYMLINK+="tape/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
diff -Nurp udev-103-orig/etc/udev/slackware/udev.rules udev-103-work/etc/udev/slackware/udev.rules
--- udev-103-orig/etc/udev/slackware/udev.rules 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/etc/udev/slackware/udev.rules 2006-11-27 13:12:15.000000000 +0100
@@ -60,7 +60,7 @@ KERNEL="sd*[!0-9]|sr*", SYSFS{ieee1394_
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/usb_id -x"
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/scsi_id -g -x -s %p -d $tempnode"
KERNEL="sd*[!0-9]|sr*", ENV{ID_SERIAL}="", IMPORT{program}="/lib/udev/scsi_id -g -x -a -s %p -d $tempnode"
-KERNEL="dasd*[!0-9]", IMPORT{program}="/lib/udev/dasd_id --export $tempnode"
+KERNEL="dasd*[!0-9]", IMPORT{program}="/lib/udev/dasd_id -u -e -d $tempnode"
KERNEL="sd*[!0-9]|sr*|dasd*[!0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
# for partitions import parent information
diff -Nurp udev-103-orig/etc/udev/suse/60-persistent-storage.rules udev-103-work/etc/udev/suse/60-persistent-storage.rules
--- udev-103-orig/etc/udev/suse/60-persistent-storage.rules 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/etc/udev/suse/60-persistent-storage.rules 2006-11-27 13:12:43.000000000 +0100
@@ -23,7 +23,7 @@ KERNEL="sd*[!0-9]|sr*|st*", ATTRS{ieee1
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="usb_id -x"
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="scsi_id -g -x -s %p -d $tempnode"
KERNEL="sd*[!0-9]|sr*|st*", ENV{ID_SERIAL}="", IMPORT{program}="scsi_id -g -x -a -s %p -d $tempnode"
-KERNEL="dasd*[!0-9]", IMPORT{program}="dasd_id --export $tempnode"
+KERNEL="dasd*[!0-9]", IMPORT{program}="dasd_id -u -e -d $tempnode"
KERNEL="sd*[!0-9]|sr*|dasd*[!0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
KERNEL="sd*[0-9]|dasd*[0-9]", ENV{ID_SERIAL}="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}-part%n"
KERNEL="sd*[!0-9]|sr*|dasd*[!0-9]", ENV{ID_SERIAL}="?*", ENV{ID_VENDOR}="ATA", SYMLINK+="disk/by-id/ata-$env{ID_MODEL}_$env{ID_SERIAL_SHORT}"
diff -Nurp udev-103-orig/extras/dasd_id/dasd_id.8 udev-103-work/extras/dasd_id/dasd_id.8
--- udev-103-orig/extras/dasd_id/dasd_id.8 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/extras/dasd_id/dasd_id.8 2006-11-27 10:59:13.000000000 +0100
@@ -1,28 +1,62 @@
-.TH DASD_ID 8 "November 2005" "" "Linux Administrator's Manual"
+.TH DASD_ID 8 "November 2006" "" "Linux Administrator's Manual"
.SH NAME
-dasd_id \- udev callout to read label from s390 block device
+.B "dasd_id "
+\- udev callout to read unique id from s390 DASD device
.SH SYNOPSIS
-.BI dasd_id
-[\fI--export\fP] \fIblockdevice\fP
-.SH "DESCRIPTION"
-.B dasd_id
-is normally called from a udev rule, to provide udev with a unique string and
-additional information (type, serial) for an S390 drive. Udev can use this
+.BI "dasd_id [-u] [-e] {-l " <busid>
+.BI "| -b " <blockdev>
+.BI " | -d " <devnode>
+.BI "}"
+.sp
+.BI "dasd_id [-h] [-?] [-v]"
+
+.SH DESCRIPTION
+.B dasd_id
+displays specific information about a specified DASD device.
+It is normally called from a udev rule, to provide udev with a unique id string and
+additional information (type, serial) for an S390 DASD drive. Udev can use this
information to create symlinks in /dev/disk/by-id and /dev/disk/by-label
to the real device node.
-.SH USAGE
-.B dasd_id
-opens the blockdevice node specified at the commandline and prints the
-information chosen by the options.
+
.SH OPTIONS
-The following commandline switches are supported to specify what dasd_id
-should print:
+
+.TP
+.BI "-u|--uid"
+Print DASD unique ID.
+.TP
+.BI "-b|--busid " <busid>
+Use the bus ID as input parameter, e.g. 0.0.e910.
+.TP
+.BI "-l|--block " <blockdev>
+Use the block device name as input parameter, e.g. dasdb.
+.TP
+.BI "-d|--devnode " <devnode>
+Use a device node as input parameter, e.g. /dev/dasdb.
.TP
-.BI --export
-print all values (instead of only the serial)
+.BI "-e|--export"
+Print all values (ID_BUS, ID_TYPE, ID_SERIAL).
+.TP
+.BI "-h|-?|--help"
+Print usage text.
+.TP
+.BI "-v|--version"
+Print version number.
+
+.SH EXAMPLES
+dasd_id -u -b 0.0.e910
+
+dasd_id -u -l dasdb
+
+dasd_id -u -d /dev/dasdb
+
+All three examples should return the same unique ID for
+the same DASD device, e.g. IBM.75000000092461.e900.10.
+
+In case this uid is not available, dasd_id will return
+the volume label instead, e.g. 0XE910.
.RE
.SH SEE ALSO
.BR udev (7)
.SH AUTHORS
-Developed by Arnd Bergmann <arnd@arndb.de> and Kay Sievers <kay.sievers@vrfy.org>.
+Volker Sameske <sameske@de.ibm.com>
diff -Nurp udev-103-orig/extras/dasd_id/dasd_id.c udev-103-work/extras/dasd_id/dasd_id.c
--- udev-103-orig/extras/dasd_id/dasd_id.c 2006-11-27 09:05:30.000000000 +0100
+++ udev-103-work/extras/dasd_id/dasd_id.c 2006-11-27 12:49:40.000000000 +0100
@@ -1,92 +1,164 @@
/*
- * dasdlabel - read label from s390 block device
+ * This file contains the dasdinfo tool which can be used
+ * to get a unique DASD ID
*
- * Copyright (C) 2004 Arnd Bergmann <arnd@arndb.de>
+ * Copyright IBM Corp. 2006
+ *
+ * Author(s): Volker Sameske (sameske@de.ibm.com)
*
- * 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.
*/
-#ifndef _GNU_SOURCE
-#define _GNU_SOURCE 1
-#endif
-
-#ifdef HAVE_CONFIG_H
-# include <config.h>
-#endif
-
#include <stdlib.h>
#include <stdio.h>
+#include <getopt.h>
#include <string.h>
-#include <unistd.h>
#include <fcntl.h>
+#include <unistd.h>
#include <errno.h>
+#include <dirent.h>
+#include <sys/utsname.h>
+#include <sys/types.h>
+#include <sys/stat.h>
#include <sys/ioctl.h>
+#include <sys/sysmacros.h>
-#include "../../udev.h"
-
-#ifdef USE_LOG
-void log_message(int priority, const char *format, ...)
-{
- va_list args;
- static int udev_log = -1;
-
- if (udev_log = -1) {
- const char *value;
+#define RELEASE_STRING "1.0.0"
+#define READCHUNK 80
+#define BLKSSZGET _IO(0x12,104)
+#define DASD_IOCTL_LETTER 'D'
+#define BIODASDINFO _IOR(DASD_IOCTL_LETTER,1,struct dasd_information)
+#define BIODASDINFO2 _IOR(DASD_IOCTL_LETTER,3,struct dasd_information2)
+#define TEMP_DEV_MAX_RETRIES 1000
+
+static const char tool_name[] = "zSeries DASD information program";
+static const char copyright_notice[] = "Copyright IBM Corp. 2006";
+
+struct volume_label {
+ char volkey[4];
+ char vollbl[4];
+ char volid[6];
+} __attribute__ ((packed));
+
+struct dasd_information2 {
+ unsigned int devno;
+ unsigned int real_devno;
+ unsigned int schid;
+ unsigned int cu_type : 16;
+ unsigned int cu_model : 8;
+ unsigned int dev_type : 16;
+ unsigned int dev_model : 8;
+ unsigned int open_count;
+ unsigned int req_queue_len;
+ unsigned int chanq_len;
+ char type[4];
+ unsigned int status;
+ unsigned int label_block;
+ unsigned int FBA_layout;
+ unsigned int characteristics_size;
+ unsigned int confdata_size;
+ char characteristics[64];
+ char configuration_data[256];
+ unsigned int format;
+ unsigned int features;
+ unsigned int reserved0;
+ unsigned int reserved1;
+ unsigned int reserved2;
+ unsigned int reserved3;
+ unsigned int reserved4;
+ unsigned int reserved5;
+ unsigned int reserved6;
+ unsigned int reserved7;
+};
- value = getenv("UDEV_LOG");
- if (value)
- udev_log = log_priority(value);
- else
- udev_log = LOG_ERR;
- }
+struct dasd_information {
+ unsigned int devno;
+ unsigned int real_devno;
+ unsigned int schid;
+ unsigned int cu_type : 16;
+ unsigned int cu_model : 8;
+ unsigned int dev_type : 16;
+ unsigned int dev_model : 8;
+ unsigned int open_count;
+ unsigned int req_queue_len;
+ unsigned int chanq_len;
+ char type[4];
+ unsigned int status;
+ unsigned int label_block;
+ unsigned int FBA_layout;
+ unsigned int characteristics_size;
+ unsigned int confdata_size;
+ char characteristics[64];
+ char configuration_data[256];
+};
- if (priority > udev_log)
- return;
+struct dasd_data
+{
+ struct dasd_information2 dasd_info;
+ int dasd_info_version;
+ int blksize;
+};
- va_start(args, format);
- vsyslog(priority, format, args);
- va_end(args);
+static void dinfo_print_version (void)
+{
+ printf ("%s version %s.\n", tool_name, RELEASE_STRING);
+ printf ("%s.\n", copyright_notice);
}
-#endif
-/*
- * Only compile this on S/390. Doesn't make any sense
- * for other architectures.
- */
+static void dinfo_print_usage(char *cmd)
+{
+ printf(""
+ "Display specific information about a specified DASD device.\n"
+ "\n"
+ "Usage: %s [-u] [-e] {-l <busid> | -b <blockdev> | -d <devnode>}\n"
+ " %s [-h] [-?] [-v]\n"
+ "\n"
+ "where:\n"
+ " -u|--uid\n"
+ " print DASD unique ID\n"
+ " -b|--busid <busid>\n"
+ " bus ID, e.g. 0.0.e910\n"
+ " -l|--block <blockdev>\n"
+ " block device name, e.g. dasdb\n"
+ " -d|--devnode <devnode>\n"
+ " device node, e.g. /dev/dasda\n"
+ " -e|--export\n"
+ " print all values (ID_BUS, ID_TYPE, ID_SERIAL)\n"
+ " -h|-?|--help\n"
+ " prints this usage text\n"
+ " -v|--version\n"
+ " prints the version number\n"
+ "\n"
+ "Example: %s -u -l dasda\n"
+ "\n",cmd,cmd,cmd);
+}
-static unsigned char EBCtoASC[256] +static char EBCtoASC[256] {
/* 0x00 NUL SOH STX ETX *SEL HT *RNL DEL */
0x00, 0x01, 0x02, 0x03, 0x07, 0x09, 0x07, 0x7F,
/* 0x08 -GE -SPS -RPT VT FF CR SO SI */
0x07, 0x07, 0x07, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
-/* 0x10 DLE DC1 DC2 DC3 -RES -NL BS -POC
- -ENP ->LF */
+/* 0x10 DLE DC1 DC2 DC3 -RES -NL BS -POC */
0x10, 0x11, 0x12, 0x13, 0x07, 0x0A, 0x08, 0x07,
-/* 0x18 CAN EM -UBS -CU1 -IFS -IGS -IRS -ITB
- -IUS */
+/* 0x18 CAN EM -UBS -CU1 -IFS -IGS -IRS -ITB */
0x18, 0x19, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
-/* 0x20 -DS -SOS FS -WUS -BYP LF ETB ESC
- -INP */
+/* 0x20 -DS -SOS FS -WUS -BYP LF ETB ESC */
0x07, 0x07, 0x1C, 0x07, 0x07, 0x0A, 0x17, 0x1B,
-/* 0x28 -SA -SFE -SM -CSP -MFA ENQ ACK BEL
- -SW */
+/* 0x28 -SA -SFE -SM -CSP -MFA ENQ ACK BEL */
0x07, 0x07, 0x07, 0x07, 0x07, 0x05, 0x06, 0x07,
/* 0x30 ---- ---- SYN -IR -PP -TRN -NBS EOT */
0x07, 0x07, 0x16, 0x07, 0x07, 0x07, 0x07, 0x04,
/* 0x38 -SBS -IT -RFF -CU3 DC4 NAK ---- SUB */
0x07, 0x07, 0x07, 0x07, 0x14, 0x15, 0x07, 0x1A,
-/* 0x40 SP RSP ä ---- */
+/* 0x40 SP RSP ? ---- */
0x20, 0xFF, 0x83, 0x84, 0x85, 0xA0, 0x07, 0x86,
/* 0x48 . < ( + | */
0x87, 0xA4, 0x9B, 0x2E, 0x3C, 0x28, 0x2B, 0x7C,
/* 0x50 & ---- */
0x26, 0x82, 0x88, 0x89, 0x8A, 0xA1, 0x8C, 0x07,
-/* 0x58 ß ! $ * ) ; */
+/* 0x58 ? ! $ * ) ; */
0x8D, 0xE1, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0xAA,
-/* 0x60 - / ---- Ä ---- ---- ---- */
+/* 0x60 - / ---- ? ---- ---- ---- */
0x2D, 0x2F, 0x07, 0x8E, 0x07, 0x07, 0x07, 0x8F,
/* 0x68 ---- , % _ > ? */
0x80, 0xA5, 0x07, 0x2C, 0x25, 0x5F, 0x3E, 0x3F,
@@ -98,7 +170,7 @@ static unsigned char EBCtoASC[256] 0x07, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
/* 0x88 h i ---- ---- ---- */
0x68, 0x69, 0xAE, 0xAF, 0x07, 0x07, 0x07, 0xF1,
-/* 0x90 ° j k l m n o p */
+/* 0x90 ? j k l m n o p */
0xF8, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70,
/* 0x98 q r ---- ---- */
0x71, 0x72, 0xA6, 0xA7, 0x91, 0x07, 0x92, 0x07,
@@ -106,157 +178,510 @@ static unsigned char EBCtoASC[256] 0xE6, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
/* 0xA8 y z ---- ---- ---- ---- */
0x79, 0x7A, 0xAD, 0xAB, 0x07, 0x07, 0x07, 0x07,
-/* 0xB0 ^ ---- § ---- */
+/* 0xB0 ^ ---- ? ---- */
0x5E, 0x9C, 0x9D, 0xFA, 0x07, 0x07, 0x07, 0xAC,
/* 0xB8 ---- [ ] ---- ---- ---- ---- */
0xAB, 0x07, 0x5B, 0x5D, 0x07, 0x07, 0x07, 0x07,
/* 0xC0 { A B C D E F G */
0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
-/* 0xC8 H I ---- ö ---- */
+/* 0xC8 H I ---- ? ---- */
0x48, 0x49, 0x07, 0x93, 0x94, 0x95, 0xA2, 0x07,
/* 0xD0 } J K L M N O P */
0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50,
-/* 0xD8 Q R ---- ü */
+/* 0xD8 Q R ---- ? */
0x51, 0x52, 0x07, 0x96, 0x81, 0x97, 0xA3, 0x98,
/* 0xE0 \ S T U V W X */
0x5C, 0xF6, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58,
-/* 0xE8 Y Z ---- Ö ---- ---- ---- */
+/* 0xE8 Y Z ---- ? ---- ---- ---- */
0x59, 0x5A, 0xFD, 0x07, 0x99, 0x07, 0x07, 0x07,
/* 0xF0 0 1 2 3 4 5 6 7 */
0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
-/* 0xF8 8 9 ---- ---- Ü ---- ---- ---- */
+/* 0xF8 8 9 ---- ---- ? ---- ---- ---- */
0x38, 0x39, 0x07, 0x07, 0x9A, 0x07, 0x07, 0x07
};
-static void vtoc_ebcdic_dec (const unsigned char *source, unsigned char *target, int l)
+
+static char * dinfo_ebcdic_dec (char *source, char *target, int l)
{
int i;
- for (i = 0; i < l; i++)
- target[i]=(char)EBCtoASC[(unsigned char)(source[i])];
+ for (i = 0; i < l; i++)
+ target[i]ëCtoASC[(unsigned char)(source[i])];
+
+ return target;
}
-/*
- * struct dasd_information_t
- * represents any data about the data, which is visible to userspace
- */
-typedef struct dasd_information_t {
- unsigned int devno; /* S/390 devno */
- unsigned int real_devno; /* for aliases */
- unsigned int schid; /* S/390 subchannel identifier */
- unsigned int cu_type : 16; /* from SenseID */
- unsigned int cu_model : 8; /* from SenseID */
- unsigned int dev_type : 16; /* from SenseID */
- unsigned int dev_model : 8; /* from SenseID */
- unsigned int open_count;
- unsigned int req_queue_len;
- unsigned int chanq_len; /* length of chanq */
- char type[4]; /* from discipline.name, 'none' for unknown */
- unsigned int status; /* current device level */
- unsigned int label_block; /* where to find the VOLSER */
- unsigned int FBA_layout; /* fixed block size (like AIXVOL) */
- unsigned int characteristics_size;
- unsigned int confdata_size;
- char characteristics[64]; /* from read_device_characteristics */
- char configuration_data[256]; /* from read_configuration_data */
-} dasd_information_t;
+static int dinfo_read_dasd_uid (char *uidfile, char *readbuf, int readbuflen)
+{
+ FILE *dasduid;
+ int offset = 0;
-#define DASD_IOCTL_LETTER 'D'
-#define BIODASDINFO _IOR(DASD_IOCTL_LETTER,1,dasd_information_t)
-#define BLKSSZGET _IO(0x12,104)
+ if ((dasduid = fopen(uidfile,"r")) = NULL) {
+ printf("Could not open uid-file %s.\n", uidfile);
+ return -1;
+ }
-unsigned char serial[8];
+ while (fgets(readbuf + offset, READCHUNK, dasduid) &&
+ readbuf[strlen(readbuf)-1] != '\n' ) {
+ offset += READCHUNK-1;
+ if ( offset+READCHUNK-1 >= readbuflen ) {
+ readbuf = realloc(readbuf,
+ readbuflen + READCHUNK-1);
+ readbuflen += READCHUNK-1;
+ }
+ }
+ fclose(dasduid);
+ return 0;
+}
-static int dasd_id(int fd)
+static int dinfo_read_dasd_vlabel (char *device, struct volume_label *vlabel,
+ char *readbuf)
{
- int blocksize;
- dasd_information_t info;
- unsigned char *data;
- unsigned char *label_raw;
+ struct dasd_data data;
+ struct volume_label tmp;
+ unsigned int vlsize = sizeof(struct volume_label);
+ unsigned long vlabel_start;
+ char vollbl[5];
+ int f;
+
+ if ((f = open(device, O_RDONLY)) < 0) {
+ printf("Could not open device node.\n");
+ goto error;
+ }
- if (ioctl(fd, BIODASDINFO, &info) != 0) {
- dbg("not a dasd");
- return -1;
+ if (ioctl(f, BLKSSZGET, &data.blksize) != 0) {
+ printf("Unable to figure out block size.\n");
+ goto error;
}
- if (ioctl(fd, BLKSSZGET, &blocksize) != 0) {
- err("failed to get blocksize");
- return -1;
+ if (ioctl(f, BIODASDINFO2, &data.dasd_info) = 0)
+ data.dasd_info_version = 2;
+ else {
+ if (ioctl(f, BIODASDINFO, &data.dasd_info) != 0) {
+ printf("Unable to figure out DASD informations.\n");
+ goto error;
}
+ }
- if (lseek(fd,info.label_block * blocksize, SEEK_SET) = -1) {
- err("seek failed on dasd");
- return -1;
- }
+ vlabel_start = data.dasd_info.label_block * data.blksize;
+ if (lseek(f, vlabel_start, SEEK_SET) < 0)
+ goto error;
+
+ bzero(vlabel, vlsize);
+
+ if (read(f, vlabel, vlsize) != vlsize) {
+ printf("Could not read volume label.\n");
+ goto error;
+ }
+
+ if (data.dasd_info.FBA_layout) {
+ bzero(&tmp, vlsize);
+ memcpy(&tmp, vlabel, vlsize);
+ memcpy(vlabel->vollbl, &tmp, vlsize-4);
+ }
+
+ close(f);
+
+ bzero(readbuf, 7);
+ bzero(vollbl, 5);
+ strncpy(vollbl, vlabel->vollbl, 4);
+ dinfo_ebcdic_dec(vollbl, vollbl, 4);
+
+ if ((strncmp(vollbl, "VOL1", 4) = 0) || (strncmp(vollbl, "LNX1", 4) = 0)) {
+ strncpy(readbuf, vlabel->volid, 6);
+ dinfo_ebcdic_dec(readbuf, readbuf, 6);
+ } else
+ strncpy(readbuf, " ", 6);
+
+
+ return 0;
+error:
+ close(f);
+ return -1;
+}
+
+static void* dinfo_malloc(size_t size)
+{
+ void* result;
+
+ result = malloc(size);
+ if (result = NULL) {
+ printf("Could not allocate %lld bytes of memory",
+ (unsigned long long) size);
+ }
+ return result;
+}
+
+static char* dinfo_make_path(char* dirname, char* filename)
+{
+ char* result;
+ size_t len;
+
+ len = strlen(dirname) + strlen(filename) + 2;
+ result = (char *) dinfo_malloc(len);
+ if (result = NULL)
+ return NULL;
+ sprintf(result, "%s/%s", dirname, filename);
+ return result;
+}
+
+static int dinfo_create_devnode(dev_t dev, char** devno)
+{
+ char* result;
+ char* pathname[] = { "/dev", getenv("TMPDIR"), "/tmp",
+ getenv("HOME"), "." , "/"};
+ char filename[] = "dasdinfo0000";
+ mode_t mode;
+ unsigned int path;
+ int retry;
+ int rc;
+ int fd;
+
+ mode = S_IFBLK | S_IRWXU;
+
+ /* Try several locations for the temporary device node. */
+ for (path=0; path < sizeof(pathname) / sizeof(pathname[0]); path++) {
+ if (pathname[path] = NULL)
+ continue;
+ for (retry=0; retry < TEMP_DEV_MAX_RETRIES; retry++) {
+ sprintf(filename, "dasdinfo%04d", retry);
+ result = dinfo_make_path(pathname[path], filename);
+ if (result = NULL)
+ return -1;
+ rc = mknod(result, mode, dev);
+ if (rc = 0) {
+ /* Need this test to cover
+ * 'nodev'-mounted
+ * filesystems. */
+ fd = open(result, O_RDWR);
+ if (fd != -1) {
+ close(fd);
+ *devno = result;
+ return 0;
+ }
+ remove(result);
+ retry = TEMP_DEV_MAX_RETRIES;
+ } else if (errno != EEXIST)
+ retry = TEMP_DEV_MAX_RETRIES;
+ free(result);
+ }
+ }
+ printf("Error: Unable to create temporary device node");
+ return -1;
+}
+
+static void dinfo_free_devnode(char* device)
+{
+ if (remove(device)) {
+ printf("Warning: Could not remove "
+ "temporary file %s", device);
+ }
+}
- data = malloc(blocksize);
- if (data = NULL)
- return -1;
-
- if (read(fd, data, blocksize) = -1) {
- err("read disklabel failed");
- free(data);
- return -1;
+static int dinfo_extract_dev(dev_t *dev, char *str, int readbuflen)
+{
+ char tmp[readbuflen];
+ char *p = NULL;
+ int ma, mi;
+
+ bzero(tmp, readbuflen);
+ strncpy(tmp, str, readbuflen);
+ if (( p = strchr(tmp, ':')) = NULL) {
+ printf("Error: unable to extract major/minor\n");
+ return -1;
+ }
+
+ *p = '\0';
+ ma = atoi(tmp);
+ mi = atoi(p + sizeof(char));
+
+ *dev = makedev(ma, mi);
+
+ return 0;
+}
+
+static int dinfo_get_dev_from_blockdev(char *blockdev, dev_t *dev)
+{
+ FILE *dasddev;
+ int offset = 0;
+ char *devfile = NULL;
+ char *readbuf = NULL;
+ int readbuflen = READCHUNK;
+
+ if ((devfile = dinfo_malloc(readbuflen)) = NULL)
+ return -1;
+
+ sprintf(devfile,"/sys/block/%s/dev", blockdev);
+
+ if ((readbuf = dinfo_malloc(readbuflen)) = NULL) {
+ printf("Error: Not enough memory to allocate readbuffer\n");
+ return -1;
+ }
+
+ if ((dasddev = fopen(devfile,"r")) = NULL) {
+ printf("Could not open %s.\n", devfile);
+ return -1;
+ }
+
+ while (fgets(readbuf + offset, READCHUNK, dasddev) &&
+ readbuf[strlen(readbuf)-1] != '\n' ) {
+ offset += READCHUNK-1;
+ if ( offset+READCHUNK-1 >= readbuflen ) {
+ readbuf = realloc(readbuf,
+ readbuflen + READCHUNK-1);
+ readbuflen += READCHUNK-1;
}
+ }
+ fclose(dasddev);
- if ((!info.FBA_layout) && (!strcmp(info.type, "ECKD")))
- label_raw = &data[8];
- else
- label_raw = &data[4];
- serial[6] = '\0';
+ if (dinfo_extract_dev(dev, readbuf, readbuflen) != 0)
+ return -1;
- vtoc_ebcdic_dec(label_raw, serial, 6);
- free(data);
+ return 0;
+}
- return 0;
+static char *dinfo_get_blockdev_from_busid(char *busid, char **blkdev)
+{
+ char tmp[READCHUNK];
+ DIR *directory = NULL;
+ struct dirent *dir_entry = NULL;
+
+ sprintf(tmp,"/sys/bus/ccw/devices/%s/", busid);
+ directory = opendir(tmp);
+
+ while ((dir_entry = readdir(directory)) != NULL) {
+ if (strncmp(dir_entry->d_name, "block:", 6) = 0) {
+ if ((*blkdev = dinfo_malloc(_D_ALLOC_NAMLEN(dir_entry))) = NULL)
+ return NULL;
+ strcpy(*blkdev, strchr(dir_entry->d_name, ':') + sizeof(char));
+ break;
+ }
}
-int main(int argc, char *argv[])
+ closedir(directory);
+
+ return *blkdev;
+}
+
+static int dinfo_get_uid_from_devnode(char **uidfile, char *devnode)
{
- const char *node = NULL;
- int i;
- int export = 0;
- int fd;
- int rc = 0;
+ struct stat stat_buffer;
+ char stat_dev[READCHUNK];
+ char sys_dev_path[READCHUNK];
+ char *readbuf;
+ DIR *directory = NULL;
+ struct dirent *dir_entry = NULL;
+ FILE *block_dev;
+ int readbuflen = READCHUNK;
+ int offset;
+
+ if (stat(devnode, &stat_buffer) != 0) {
+ printf("Error: could not stat %s\n", devnode);
+ return -1;
+ }
+
+ sprintf(stat_dev, "%d:%d", major(stat_buffer.st_rdev),
+ minor(stat_buffer.st_rdev));
+
+ if ((directory = opendir("/sys/block/")) = NULL) {
+ printf("Error: could not open directory /sys/block\n");
+ return -1;
+ }
+
+ if ((readbuf = dinfo_malloc(readbuflen)) = NULL) {
+ printf("Error: Not enough memory to allocate readbuffer\n");
+ return -1;
+ }
+
+ while ((dir_entry = readdir(directory)) != NULL) {
+ sprintf(sys_dev_path, "/sys/block/%s/dev", dir_entry->d_name);
- logging_init("dasd_id");
+ if ((block_dev = fopen(sys_dev_path,"r")) = NULL)
+ continue;
- for (i = 1; i < argc; i++) {
- char *arg = argv[i];
+ offset = 0;
+ while (fgets(readbuf + offset, READCHUNK, block_dev) &&
+ readbuf[strlen(readbuf)-1] != '\n' ) {
+ offset += READCHUNK-1;
+ if ( offset+READCHUNK-1 >= readbuflen ) {
+ readbuf = realloc(readbuf,
+ readbuflen + READCHUNK-1);
+ readbuflen += READCHUNK-1;
+ }
+ }
+ fclose(block_dev);
- if (strcmp(arg, "--export") = 0) {
+ if (strncmp(stat_dev, readbuf, strlen(stat_dev)) = 0) {
+ sprintf(*uidfile,"/sys/block/%s/device/uid", dir_entry->d_name);
+ break;
+ }
+ }
+
+ closedir(directory);
+ return 0;
+}
+
+int main(int argc, char * argv[])
+{
+ struct utsname uname_buf;
+ int version, release;
+ char *uidfile = NULL;
+ char *device = NULL;
+ char *readbuf = NULL;
+ int readbuflen = READCHUNK;
+ dev_t dev;
+ int export = 0;
+ int c;
+ int print_uid = 0;
+ char *blockdev = NULL;
+ char *busid = NULL;
+ char *devnode = NULL;
+ struct volume_label vlabel;
+
+ while (1) {
+ int option_index = 0;
+ static struct option long_options[] = {
+ {"uid", 0, 0, 'u'},
+ {"busid", 1, 0, 'b'},
+ {"block", 1, 0, 'l'},
+ {"devnode", 1, 0, 'd'},
+ {"export", 0, 0, 'e'},
+ {"help", 0, 0, 'h'},
+ {"version", 0, 0, 'v'},
+ {0, 0, 0, 0}
+ };
+
+ c = getopt_long (argc, argv, "vh?eub:l:d:",
+ long_options, &option_index);
+ if (c = -1)
+ break;
+
+ switch (c) {
+ case 'u':
+ print_uid = 1;
+ break;
+ case 'b':
+ busid=strdup(optarg);
+ break;
+ case 'l':
+ blockdev=strdup(optarg);
+ break;
+ case 'd':
+ devnode=strdup(optarg);
+ break;
+ case 'e':
export = 1;
- } else
- node = arg;
+ break;
+ case 'h':
+ case '?':
+ dinfo_print_usage(argv[0]);
+ exit(0);
+ case 'v':
+ dinfo_print_version();
+ exit(0);
+ default:
+ dinfo_print_usage(argv[0]);
+ exit(0);
+ }
}
- if (!node) {
- err("no node specified");
- rc = 1;
- goto exit;
+
+ uname(&uname_buf);
+ sscanf(uname_buf.release, "%d.%d", &version,&release);
+ if (strcmp(uname_buf.sysname,"Linux") ||
+ version < 2 || release < 6) {
+ printf("%s %d.%d is not supported\n", uname_buf.sysname,
+ version,release);
+ return -1;
}
- fd = open(node, O_RDONLY);
- if (fd < 0) {
- err("unable to open '%s'", node);
- rc = 1;
- goto exit;
+ if (!busid && !blockdev && !devnode) {
+ printf("Error: please specify a device using either -b, -l or -d\n");
+ return -1;
}
- if (dasd_id(fd) < 0) {
- err("dasd_id failed: %s", strerror(errno));
- rc = 1;
+ if ((busid && blockdev) || (busid && devnode) || (blockdev && devnode)) {
+ printf("Error: please specify device only once, either -b, -l or -d\n");
+ return -1;
}
+ if (!print_uid) {
+ printf("Error: no action specified (e.g. -u)\n");
+ return -1;
+ }
+
+ if (((readbuf = dinfo_malloc(readbuflen)) = NULL) ||
+ ((uidfile = dinfo_malloc(readbuflen)) = NULL) ||
+ ((device = dinfo_malloc(readbuflen)) = NULL))
+ return -1;
+
if (export) {
- printf("ID_TYPE=disk\n");
- printf("ID_SERIAL=%s\n",serial);
printf("ID_BUSÌw\n");
- } else
- printf("%s\n", serial);
+ printf("ID_TYPE=disk\n");
+ }
- close(fd);
-exit:
- logging_close();
- return rc;
-}
+ /* try to read the uid attribute */
+ if (busid) {
+ sprintf(uidfile,"/sys/bus/ccw/devices/%s/uid", busid);
+ } else if (blockdev) {
+ sprintf(uidfile,"/sys/block/%s/device/uid", blockdev);
+ } else if (devnode) {
+ if (dinfo_get_uid_from_devnode(&uidfile, devnode) != 0)
+ goto error;
+ }
+ if (dinfo_read_dasd_uid(uidfile, readbuf, readbuflen) = 0) {
+ if (export) {
+ printf("ID_SERIAL=%s",readbuf);
+ } else
+ printf("%s",readbuf);
+
+ goto out;
+ }
+
+ /* there is no uid, try to read the volume serial */
+ if (busid) {
+ char *blockdev_name = NULL;
+
+ if (dinfo_get_blockdev_from_busid(busid, &blockdev_name) = NULL)
+ goto error;
+
+ if (dinfo_get_dev_from_blockdev(blockdev_name, &dev) != 0)
+ goto error;
+
+ if(dinfo_create_devnode(dev, &device) != 0)
+ goto error;
+
+ free(blockdev_name);
+
+ } else if (blockdev) {
+ if (dinfo_get_dev_from_blockdev(blockdev, &dev) != 0)
+ goto error;
+
+ if(dinfo_create_devnode(dev, &device) != 0)
+ goto error;
+
+ } else if (devnode) {
+ strcpy(device, devnode);
+ }
+
+ if (dinfo_read_dasd_vlabel(device, &vlabel, readbuf) = 0) {
+ if (export) {
+ printf("ID_SERIAL=%6.6s\n",readbuf);
+ } else
+ printf("%6.6s\n", readbuf);
+ if (busid || blockdev)
+ dinfo_free_devnode(device);
+
+ goto out;
+ }
+
+error:
+ printf("Error: could not read unique DASD ID\n");
+
+out:
+ free(uidfile);
+ free(device);
+ free(readbuf);
+
+ return 0;
+}
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
@ 2006-11-27 15:52 ` Kay Sievers
2006-11-27 17:05 ` Volker Sameske
` (10 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-27 15:52 UTC (permalink / raw)
To: linux-hotplug
On 11/27/06, Volker Sameske <sameske@de.ibm.com> wrote:
>
> The currently used uid (volume label) for s390 DASD devices is not really
> unique. For that reason a unique uid sysfs-attribute for DASD devices has
> been introduced. This new dasd_id tool replaces the current one and
> evaluates this new sysfs attribute. In case it is not available, dasd_id
> returns the old volume label instead.
> etc/udev/debian/persistent.rules | 2
> etc/udev/redhat/50-udev.rules | 2
> etc/udev/slackware/udev.rules | 2
> etc/udev/suse/60-persistent-storage.rules | 2
You can't patch any of these rules, they are usually copies _from_ the
distro as examples, not _to_ the distro.
> - * 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.
Why do you want to remove the license?
> - value = getenv("UDEV_LOG");
> - if (value)
> - udev_log = log_priority(value);
> - else
> - udev_log = LOG_ERR;
Why don't we want that anymore?
> +static int dinfo_create_devnode(dev_t dev, char** devno)
> +{
> + char* result;
> + char* pathname[] = { "/dev", getenv("TMPDIR"), "/tmp",
> + getenv("HOME"), "." , "/"};
What's are you trying to fight here? That's absolutely not needed,
just let udev provide you a temporary node, or use /dev, which is
obviously always writable in a udev setup.
> + sprintf(tmp,"/sys/bus/ccw/devices/%s/", busid);
> + directory = opendir(tmp);
You can't reference /sys/bus/ directly, look at udev_sysfs.c, how to do it.
> + while ((dir_entry = readdir(directory)) != NULL) {
> + if (strncmp(dir_entry->d_name, "block:", 6) = 0) {
You can't use these links, they will go away some day.
> + sscanf(uname_buf.release, "%d.%d", &version,&release);
> + if (strcmp(uname_buf.sysname,"Linux") ||
> + version < 2 || release < 6) {
No need for that, there is no udev for kernels < 2.6.
Instead of all that, can't you just read the values from sysfs with
udev using $attr{} in a rule, and possibly fall back to the old tool
only if you don't find the values there?
Thanks,
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
2006-11-27 15:52 ` Kay Sievers
@ 2006-11-27 17:05 ` Volker Sameske
2006-11-27 17:41 ` Kay Sievers
` (9 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-27 17:05 UTC (permalink / raw)
To: linux-hotplug
> > etc/udev/debian/persistent.rules | 2
> > etc/udev/redhat/50-udev.rules | 2
> > etc/udev/slackware/udev.rules | 2
> > etc/udev/suse/60-persistent-storage.rules | 2
>
> You can't patch any of these rules, they are usually copies _from_ the
> distro as examples, not _to_ the distro.
Good point, I'll remove these 4 chunks.
> > - * 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.
>
> Why do you want to remove the license?
Right, but I did not really remove it. I wrote a new tool
with a new header. And I have to follow IBM directives, how such
a header has to look like. But it is still GPL.
> > - value = getenv("UDEV_LOG");
> > - if (value)
> > - udev_log = log_priority(value);
> > - else
> > - udev_log = LOG_ERR;
>
> Why don't we want that anymore?
Well, initially I wrote a new tool for the multipath-tools
package. It was called dasdinfo. To a later point in time the
idea came to replace the existing udev dasd_id with it. So
even if the patch looks like I've changed dasd_id, I wrote a
new tool.
But good point. I'll think about adding these lines.
> > +static int dinfo_create_devnode(dev_t dev, char** devno)
> > +{
> > + char* result;
> > + char* pathname[] = { "/dev", getenv("TMPDIR"), "/tmp",
> > + getenv("HOME"), "." , "/"};
>
> What's are you trying to fight here? That's absolutely not needed,
> just let udev provide you a temporary node, or use /dev, which is
> obviously always writable in a udev setup.
Agree, if this would be only a udev-tool. But this tool will be
also used within multipath-tools and there I have to create my own
temporary device nodes. So I can't remove that.
> > + sprintf(tmp,"/sys/bus/ccw/devices/%s/", busid);
> > + directory = opendir(tmp);
>
> You can't reference /sys/bus/ directly, look at udev_sysfs.c, how to do
it.
Initially I had a sysfs-mountpoint detection. But it has been
removed during a long discussion about that topic and after some
review cycles. There are many many other tools with hard-coded
sysfs mount points.
And this sysfs-mount-point-detection takes some time since it will be
done for every device. And there could be several hundred DASD devices
on a single IBM System z.
> > + while ((dir_entry = readdir(directory)) != NULL) {
> > + if (strncmp(dir_entry->d_name, "block:", 6) = 0) {
>
> You can't use these links, they will go away some day.
>
> > + sscanf(uname_buf.release, "%d.%d", &version,&release);
> > + if (strcmp(uname_buf.sysname,"Linux") ||
> > + version < 2 || release < 6) {
>
> No need for that, there is no udev for kernels < 2.6.
Yes, but since this tools is not only used for udev I'll keep this check.
> Instead of all that, can't you just read the values from sysfs with
> udev using $attr{} in a rule, and possibly fall back to the old tool
> only if you don't find the values there?
Yes, if udev would be the only exploiter of this tool. Udev is one of
the main exploiter, but not the single one.
I'll wait some more time in case there are some more comments. Then
I'll post the changed patch.
thank you and regards,
Volker
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
2006-11-27 15:52 ` Kay Sievers
2006-11-27 17:05 ` Volker Sameske
@ 2006-11-27 17:41 ` Kay Sievers
2006-11-27 17:53 ` Patrick Mansfield
` (8 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-27 17:41 UTC (permalink / raw)
To: linux-hotplug
On Mon, 2006-11-27 at 18:05 +0100, Volker Sameske wrote:
> > > etc/udev/debian/persistent.rules | 2
> > > etc/udev/redhat/50-udev.rules | 2
> > > etc/udev/slackware/udev.rules | 2
> > > etc/udev/suse/60-persistent-storage.rules | 2
> >
> > You can't patch any of these rules, they are usually copies _from_ the
> > distro as examples, not _to_ the distro.
>
> Good point, I'll remove these 4 chunks.
>
> > > - * 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.
> >
> > Why do you want to remove the license?
>
> Right, but I did not really remove it. I wrote a new tool
> with a new header. And I have to follow IBM directives, how such
> a header has to look like. But it is still GPL.
>
> > > - value = getenv("UDEV_LOG");
> > > - if (value)
> > > - udev_log = log_priority(value);
> > > - else
> > > - udev_log = LOG_ERR;
> >
> > Why don't we want that anymore?
>
> Well, initially I wrote a new tool for the multipath-tools
> package. It was called dasdinfo. To a later point in time the
> idea came to replace the existing udev dasd_id with it. So
> even if the patch looks like I've changed dasd_id, I wrote a
> new tool.
> But good point. I'll think about adding these lines.
>
> > > +static int dinfo_create_devnode(dev_t dev, char** devno)
> > > +{
> > > + char* result;
> > > + char* pathname[] = { "/dev", getenv("TMPDIR"), "/tmp",
> > > + getenv("HOME"), "." , "/"};
> >
> > What's are you trying to fight here? That's absolutely not needed,
> > just let udev provide you a temporary node, or use /dev, which is
> > obviously always writable in a udev setup.
>
> Agree, if this would be only a udev-tool. But this tool will be
> also used within multipath-tools and there I have to create my own
> temporary device nodes. So I can't remove that.
>
> > > + sprintf(tmp,"/sys/bus/ccw/devices/%s/", busid);
> > > + directory = opendir(tmp);
> >
> > You can't reference /sys/bus/ directly, look at udev_sysfs.c, how to do
> it.
>
> Initially I had a sysfs-mountpoint detection. But it has been
> removed during a long discussion about that topic and after some
> review cycles. There are many many other tools with hard-coded
> sysfs mount points.
> And this sysfs-mount-point-detection takes some time since it will be
> done for every device. And there could be several hundred DASD devices
> on a single IBM System z.
No that's not what I mean, /sys is even hard-coded in udev, only
overwritable for the test-suite. But /sys/block may live
in /sys/subsystem/block and all that stuff that is queued at the moment
for sysfs changes. There will be backwards compat-links in sysfs for
existing apps to work, but in the udev tree, nothing new is accepted,
that relies on the compat-links.
> > > + while ((dir_entry = readdir(directory)) != NULL) {
> > > + if (strncmp(dir_entry->d_name, "block:", 6) = 0) {
> >
> > You can't use these links, they will go away some day.
> >
> > > + sscanf(uname_buf.release, "%d.%d", &version,&release);
> > > + if (strcmp(uname_buf.sysname,"Linux") ||
> > > + version < 2 || release < 6) {
> >
> > No need for that, there is no udev for kernels < 2.6.
>
> Yes, but since this tools is not only used for udev I'll keep this check.
>
> > Instead of all that, can't you just read the values from sysfs with
> > udev using $attr{} in a rule, and possibly fall back to the old tool
> > only if you don't find the values there?
>
> Yes, if udev would be the only exploiter of this tool. Udev is one of
> the main exploiter, but not the single one.
Oh, then your tool probably doesn't want to hide in /lib/udev/ too,
right? The udev tarball does not want to carry around generic or
generally useful tools, only things which are private to udev and can be
changed to our needs without really caring about possible other users.
It may make sense, to remove the current tool from the udev tree, so
this udev version can just rely on the sysfs-value to be present.
Your new tool would go into its own package, from a distribution point
of view. Packagers can still use and integrate the new tool into
udev-rules, if they really feel the need for that kind of compatibility,
which I don't expect for udev's need.
Does that make sense?
Thanks,
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (2 preceding siblings ...)
2006-11-27 17:41 ` Kay Sievers
@ 2006-11-27 17:53 ` Patrick Mansfield
2006-11-28 8:54 ` Kay Sievers
` (7 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Patrick Mansfield @ 2006-11-27 17:53 UTC (permalink / raw)
To: linux-hotplug
On Mon, Nov 27, 2006 at 06:41:28PM +0100, Kay Sievers wrote:
> On Mon, 2006-11-27 at 18:05 +0100, Volker Sameske wrote:
> > Yes, if udev would be the only exploiter of this tool. Udev is one of
> > the main exploiter, but not the single one.
>
> Oh, then your tool probably doesn't want to hide in /lib/udev/ too,
> right? The udev tarball does not want to carry around generic or
> generally useful tools, only things which are private to udev and can be
> changed to our needs without really caring about possible other users.
>
> It may make sense, to remove the current tool from the udev tree, so
> this udev version can just rely on the sysfs-value to be present.
Kay -
The same is true for scsi_id, it's used by both udev and multipath.
-- Patrick Mansfield
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (3 preceding siblings ...)
2006-11-27 17:53 ` Patrick Mansfield
@ 2006-11-28 8:54 ` Kay Sievers
2006-11-28 15:29 ` Volker Sameske
` (6 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-28 8:54 UTC (permalink / raw)
To: linux-hotplug
On Mon, 2006-11-27 at 09:53 -0800, Patrick Mansfield wrote:
> On Mon, Nov 27, 2006 at 06:41:28PM +0100, Kay Sievers wrote:
> > On Mon, 2006-11-27 at 18:05 +0100, Volker Sameske wrote:
>
> > > Yes, if udev would be the only exploiter of this tool. Udev is one of
> > > the main exploiter, but not the single one.
> >
> > Oh, then your tool probably doesn't want to hide in /lib/udev/ too,
> > right? The udev tarball does not want to carry around generic or
> > generally useful tools, only things which are private to udev and can be
> > changed to our needs without really caring about possible other users.
> >
> > It may make sense, to remove the current tool from the udev tree, so
> > this udev version can just rely on the sysfs-value to be present.
>
> The same is true for scsi_id, it's used by both udev and multipath.
It's true, but not the same. Scsi_id is there for ages, and we install a
symlink in /sbin for that reason. That will not happen for new tools,
especially not when the information is available in sysfs today.
Thanks,
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (4 preceding siblings ...)
2006-11-28 8:54 ` Kay Sievers
@ 2006-11-28 15:29 ` Volker Sameske
2006-11-28 15:53 ` Kay Sievers
` (5 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-28 15:29 UTC (permalink / raw)
To: linux-hotplug
> No that's not what I mean, /sys is even hard-coded in udev, only
> overwritable for the test-suite. But /sys/block may live
> in /sys/subsystem/block and all that stuff that is queued at the moment
> for sysfs changes. There will be backwards compat-links in sysfs for
> existing apps to work, but in the udev tree, nothing new is accepted,
> that relies on the compat-links.
Okay, now I see your point. I can change that.
> Oh, then your tool probably doesn't want to hide in /lib/udev/ too,
> right?
Right
> The udev tarball does not want to carry around generic or
> generally useful tools, only things which are private to udev and can be
> changed to our needs without really caring about possible other users.
> It may make sense, to remove the current tool from the udev tree, so
> this udev version can just rely on the sysfs-value to be present.
No, since the uid sysfs attribute is new and not part of older
distributions. Without that uid attribute we have to open the
device to read the not really unique volume label. So dasd_id
still needs to be part of udev.
> Your new tool would go into its own package, from a distribution point
> of view. Packagers can still use and integrate the new tool into
> udev-rules, if they really feel the need for that kind of compatibility,
> which I don't expect for udev's need.
Would be one approach. But if this tool goes into the multipath package
we had an udev dependency on the mp-tools package. Since every Linux user
uses udev and only a subset uses the multipath package we would force
some users to install mp-tool, even if they don't need it. The other
way around would be better.
> Does that make sense?
I understand your point. But what you suggest is to have two different
tools for udev and mp-tools wich implement the same functionality.
That's something I would like to avoid.
regards,
Volker
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (5 preceding siblings ...)
2006-11-28 15:29 ` Volker Sameske
@ 2006-11-28 15:53 ` Kay Sievers
2006-11-28 16:22 ` Volker Sameske
` (4 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-28 15:53 UTC (permalink / raw)
To: linux-hotplug
On Tue, 2006-11-28 at 16:29 +0100, Volker Sameske wrote:
> > No that's not what I mean, /sys is even hard-coded in udev, only
> > overwritable for the test-suite. But /sys/block may live
> > in /sys/subsystem/block and all that stuff that is queued at the moment
> > for sysfs changes. There will be backwards compat-links in sysfs for
> > existing apps to work, but in the udev tree, nothing new is accepted,
> > that relies on the compat-links.
>
> Okay, now I see your point. I can change that.
Fine.
> > Oh, then your tool probably doesn't want to hide in /lib/udev/ too,
> > right?
>
> Right
The current multipath already uses the current dasd_id in /lib/udev?
> > The udev tarball does not want to carry around generic or
> > generally useful tools, only things which are private to udev and can be
> > changed to our needs without really caring about possible other users.
>
> > It may make sense, to remove the current tool from the udev tree, so
> > this udev version can just rely on the sysfs-value to be present.
>
> No, since the uid sysfs attribute is new and not part of older
> distributions. Without that uid attribute we have to open the
> device to read the not really unique volume label. So dasd_id
> still needs to be part of udev.
Well, we usually don't care about backwards compat in that direction. A
udev version can simply require a minimal kernel version, and this
should be fine.
But how do we handle people already using the current persistent links?
Your new tool creates different names for the same devices on recent
kernels, right?
> > Your new tool would go into its own package, from a distribution point
> > of view. Packagers can still use and integrate the new tool into
> > udev-rules, if they really feel the need for that kind of compatibility,
> > which I don't expect for udev's need.
>
> Would be one approach. But if this tool goes into the multipath package
> we had an udev dependency on the mp-tools package. Since every Linux user
> uses udev and only a subset uses the multipath package we would force
> some users to install mp-tool, even if they don't need it. The other
> way around would be better.
>
> > Does that make sense?
>
> I understand your point. But what you suggest is to have two different
> tools for udev and mp-tools wich implement the same functionality.
> That's something I would like to avoid.
Right, that's always good to avoid.
Thanks,
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (6 preceding siblings ...)
2006-11-28 15:53 ` Kay Sievers
@ 2006-11-28 16:22 ` Volker Sameske
2006-11-28 16:33 ` Kay Sievers
` (3 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-28 16:22 UTC (permalink / raw)
To: linux-hotplug
> The current multipath already uses the current dasd_id in /lib/udev?
Yes, the current mp-tools package calls the symlink /sbin/dasd_id
> Well, we usually don't care about backwards compat in that direction. A
> udev version can simply require a minimal kernel version, and this
> should be fine.
>
> But how do we handle people already using the current persistent links?
> Your new tool creates different names for the same devices on recent
> kernels, right?
Yes. If the new uid is available in a distribution dasd_id returns:
ID_BUSÌw
ID_TYPE=disk
ID_SERIAL=IBM.75000000092461.4d00.34
if not, you'll get:
ID_BUSÌw
ID_TYPE=disk
ID_SERIAL=0X4D34
> Right, that's always good to avoid.
So you can see my problem :)
regards,
Volker
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (7 preceding siblings ...)
2006-11-28 16:22 ` Volker Sameske
@ 2006-11-28 16:33 ` Kay Sievers
2006-11-29 6:55 ` Volker Sameske
` (2 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-28 16:33 UTC (permalink / raw)
To: linux-hotplug
On Tue, 2006-11-28 at 17:22 +0100, Volker Sameske wrote:
> > The current multipath already uses the current dasd_id in /lib/udev?
>
> Yes, the current mp-tools package calls the symlink /sbin/dasd_id
>
> > Well, we usually don't care about backwards compat in that direction. A
> > udev version can simply require a minimal kernel version, and this
> > should be fine.
> >
> > But how do we handle people already using the current persistent links?
> > Your new tool creates different names for the same devices on recent
> > kernels, right?
>
> Yes. If the new uid is available in a distribution dasd_id returns:
>
> ID_BUSÌw
> ID_TYPE=disk
> ID_SERIAL=IBM.75000000092461.4d00.34
>
> if not, you'll get:
>
> ID_BUSÌw
> ID_TYPE=disk
> ID_SERIAL=0X4D34
So, how do we handle possible users who use the old name? Always create
both links? Warn on update?
> > Right, that's always good to avoid.
> So you can see my problem :)
Yes and no, it would be fine in its own package too, I think. But let's
solve the real problems first ...
Thanks,
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (8 preceding siblings ...)
2006-11-28 16:33 ` Kay Sievers
@ 2006-11-29 6:55 ` Volker Sameske
2006-11-29 9:11 ` Kay Sievers
2006-11-29 12:03 ` Volker Sameske
11 siblings, 0 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-29 6:55 UTC (permalink / raw)
To: linux-hotplug
> > > Well, we usually don't care about backwards compat in that
direction. A
> > > udev version can simply require a minimal kernel version, and this
> > > should be fine.
> > >
> > > But how do we handle people already using the current persistent
links?
> > > Your new tool creates different names for the same devices on recent
> > > kernels, right?
> >
> > Yes. If the new uid is available in a distribution dasd_id returns:
> >
> > ID_BUSÌw
> > ID_TYPE=disk
> > ID_SERIAL=IBM.75000000092461.4d00.34
> >
> > if not, you'll get:
> >
> > ID_BUSÌw
> > ID_TYPE=disk
> > ID_SERIAL=0X4D34
>
> So, how do we handle possible users who use the old name? Always create
> both links? Warn on update?
I see your point, but from my experiences changes like new sysfs
attributes only come with new versions of a distribution or with
bigger updates, e.g. SUSE service packs or RedHat Updates.
I do not see a problem with new versions of distributions,
e.g. SLES9 -> SLES10. New distribution, new device nodes, new game.
If the change comes with a service pack within the lifetime of a
distribution it is a distributors decision what to do.
They could just change the device nodes and affected system tools
(e.g. yast) and document that possibly some user tools needs to
be adapted too.
Or the distributor decides do create both device nodes using two
different udev rules. I could add a --old option to dasd_id which
always returns the old version. The current implementation decides
itself which version to return.
But I would prefer the first approach. The old uid is not unique.
So using these device nodes could lead in worst case to data loss.
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (9 preceding siblings ...)
2006-11-29 6:55 ` Volker Sameske
@ 2006-11-29 9:11 ` Kay Sievers
2006-11-29 12:03 ` Volker Sameske
11 siblings, 0 replies; 13+ messages in thread
From: Kay Sievers @ 2006-11-29 9:11 UTC (permalink / raw)
To: linux-hotplug
On Wed, 2006-11-29 at 07:55 +0100, Volker Sameske wrote:
> > > > Well, we usually don't care about backwards compat in that
> direction. A
> > > > udev version can simply require a minimal kernel version, and this
> > > > should be fine.
> > > >
> > > > But how do we handle people already using the current persistent
> links?
> > > > Your new tool creates different names for the same devices on recent
> > > > kernels, right?
> > >
> > > Yes. If the new uid is available in a distribution dasd_id returns:
> > >
> > > ID_BUSÌw
> > > ID_TYPE=disk
> > > ID_SERIAL=IBM.75000000092461.4d00.34
> > >
> > > if not, you'll get:
> > >
> > > ID_BUSÌw
> > > ID_TYPE=disk
> > > ID_SERIAL=0X4D34
> >
> > So, how do we handle possible users who use the old name? Always create
> > both links? Warn on update?
>
> I see your point, but from my experiences changes like new sysfs
> attributes only come with new versions of a distribution or with
> bigger updates, e.g. SUSE service packs or RedHat Updates.
>
> I do not see a problem with new versions of distributions,
> e.g. SLES9 -> SLES10. New distribution, new device nodes, new game.
>
> If the change comes with a service pack within the lifetime of a
> distribution it is a distributors decision what to do.
> They could just change the device nodes and affected system tools
> (e.g. yast) and document that possibly some user tools needs to
> be adapted too.
> Or the distributor decides do create both device nodes using two
> different udev rules. I could add a --old option to dasd_id which
> always returns the old version. The current implementation decides
> itself which version to return.
>
> But I would prefer the first approach. The old uid is not unique.
> So using these device nodes could lead in worst case to data loss.
Well, for a distributor, which doesn't support the old names, there is
no need for that tool at all then, because he can just rely on the sysfs
attribute with the same result. :)
You should probably add the old name to the --export switch, so udev can
create both links if needed. Can the old version be computed from the
sysfs value, or do we need to read it from the device?
Kay
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [Patch] udev: new dasd_id tool
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
` (10 preceding siblings ...)
2006-11-29 9:11 ` Kay Sievers
@ 2006-11-29 12:03 ` Volker Sameske
11 siblings, 0 replies; 13+ messages in thread
From: Volker Sameske @ 2006-11-29 12:03 UTC (permalink / raw)
To: linux-hotplug
> Well, for a distributor, which doesn't support the old names, there is
> no need for that tool at all then, because he can just rely on the sysfs
> attribute with the same result. :)
Well, SLES9 and RHEL4 support the old names and I am not sure if the
new uid attribute is already part of them.
> You should probably add the old name to the --export switch, so udev can
> create both links if needed.
Yes, I can do that. --export could return:
ID_BUSÌw
ID_TYPE=disk
ID_SERIAL=IBM.75000000092461.4d00.34
ID_SERIAL_OLD=0X4D34
> Can the old version be computed from the
> sysfs value, or do we need to read it from the device?
It has to be read from the device, usually from the third block.
The root user can change it to nearly any 6-char string. That's
the reason it is not unique.
So, let me summarize the options I see:
(1) Replace the old dasd_id with the new dasd_id which implements the
new uid and some additional stuff (e.g. temp. device nodes) which is
required to use the uid also for the multipath tools.
(2) Enhance the existing dasd_id to be able to handle the new uids
and include my new tool with a different name into the mp-tools
package. We had two tools wich implement the same functionality twice
and using different ways.
(3) Leave the old dasd_id as it is (only supporting volume labels) and
implement the new uid additionally only with udev instruments, without
calling an external tool. Not sure if this is possible. So we had
two different device nodes (volume label und uid). I would try to
get my new tool into the mp-tools package using a different name.
(4) Move dasd_id from udev to mp-tools which would create a
dependency on mp-tools.
Did I miss one? In my opinion (2) and (4) are options we definitely
do not want.
regards,
Volker
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
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
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2006-11-29 12:03 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-11-27 13:15 [Patch] udev: new dasd_id tool Volker Sameske
2006-11-27 15:52 ` Kay Sievers
2006-11-27 17:05 ` Volker Sameske
2006-11-27 17:41 ` Kay Sievers
2006-11-27 17:53 ` Patrick Mansfield
2006-11-28 8:54 ` Kay Sievers
2006-11-28 15:29 ` Volker Sameske
2006-11-28 15:53 ` Kay Sievers
2006-11-28 16:22 ` Volker Sameske
2006-11-28 16:33 ` Kay Sievers
2006-11-29 6:55 ` Volker Sameske
2006-11-29 9:11 ` Kay Sievers
2006-11-29 12:03 ` Volker Sameske
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).