busybox/util-linux/mdev.c

1272 lines
34 KiB
C
Raw Normal View History

/* vi: set sw=4 ts=4: */
/*
* mdev - Mini udev for busybox
2006-01-25 05:38:53 +05:30
*
* Copyright 2005 Rob Landley <rob@landley.net>
* Copyright 2005 Frank Sorenson <frank@tuxrocks.com>
*
* Licensed under GPLv2, see file LICENSE in this source tree.
*/
//config:config MDEV
//config: bool "mdev (17 kb)"
//config: default y
//config: help
//config: mdev is a mini-udev implementation for dynamically creating device
//config: nodes in the /dev directory.
//config:
//config: For more information, please see docs/mdev.txt
//config:
//config:config FEATURE_MDEV_CONF
//config: bool "Support /etc/mdev.conf"
//config: default y
//config: depends on MDEV
//config: help
//config: Add support for the mdev config file to control ownership and
//config: permissions of the device nodes.
//config:
//config: For more information, please see docs/mdev.txt
//config:
//config:config FEATURE_MDEV_RENAME
//config: bool "Support subdirs/symlinks"
//config: default y
//config: depends on FEATURE_MDEV_CONF
//config: help
//config: Add support for renaming devices and creating symlinks.
//config:
//config: For more information, please see docs/mdev.txt
//config:
//config:config FEATURE_MDEV_RENAME_REGEXP
//config: bool "Support regular expressions substitutions when renaming device"
//config: default y
//config: depends on FEATURE_MDEV_RENAME
//config: help
//config: Add support for regular expressions substitutions when renaming
//config: device.
//config:
//config:config FEATURE_MDEV_EXEC
//config: bool "Support command execution at device addition/removal"
//config: default y
//config: depends on FEATURE_MDEV_CONF
//config: help
//config: This adds support for an optional field to /etc/mdev.conf for
//config: executing commands when devices are created/removed.
//config:
//config: For more information, please see docs/mdev.txt
//config:
//config:config FEATURE_MDEV_LOAD_FIRMWARE
//config: bool "Support loading of firmware"
//config: default y
//config: depends on MDEV
//config: help
//config: Some devices need to load firmware before they can be usable.
//config:
//config: These devices will request userspace look up the files in
//config: /lib/firmware/ and if it exists, send it to the kernel for
//config: loading into the hardware.
//config:
//config:config FEATURE_MDEV_DAEMON
//config: bool "Support daemon mode"
//config: default y
//config: depends on MDEV
//config: help
//config: Adds the -d option to run mdev in daemon mode handling hotplug
//config: events from the kernel like udev. If the system generates many
//config: hotplug events this mode of operation will consume less
//config: resources than registering mdev as hotplug helper or using the
//config: uevent applet.
//applet:IF_MDEV(APPLET(mdev, BB_DIR_SBIN, BB_SUID_DROP))
//kbuild:lib-$(CONFIG_MDEV) += mdev.o
//usage:#define mdev_trivial_usage
//usage: "[-s]" IF_FEATURE_MDEV_DAEMON(" | [-df]")
//usage:#define mdev_full_usage "\n\n"
//usage: "mdev -s is to be run during boot to scan /sys and populate /dev.\n"
//usage: IF_FEATURE_MDEV_DAEMON(
//usage: "mdev -d[f]: daemon, listen on netlink.\n"
//usage: " -f: stay in foreground.\n"
//usage: )
//usage: "\n"
//usage: "Bare mdev is a kernel hotplug helper. To activate it:\n"
//usage: " echo /sbin/mdev >/proc/sys/kernel/hotplug\n"
//usage: IF_FEATURE_MDEV_CONF(
//usage: "\n"
//usage: "It uses /etc/mdev.conf with lines\n"
//usage: " [-][ENV=regex;]...DEVNAME UID:GID PERM"
//usage: IF_FEATURE_MDEV_RENAME(" [>|=PATH]|[!]")
//usage: IF_FEATURE_MDEV_EXEC(" [@|$|*PROG]")
//usage: "\n"
//usage: "where DEVNAME is device name regex, @major,minor[-minor2], or\n"
//usage: "environment variable regex. A common use of the latter is\n"
//usage: "to load modules for hotplugged devices:\n"
//usage: " $MODALIAS=.* 0:0 660 @modprobe \"$MODALIAS\"\n"
//usage: )
//usage: "\n"
//usage: "If /dev/mdev.seq file exists, mdev will wait for its value\n"
//usage: "to match $SEQNUM variable. This prevents plug/unplug races.\n"
//usage: "To activate this feature, create empty /dev/mdev.seq at boot.\n"
//usage: "\n"
//usage: "If /dev/mdev.log file exists, debug log will be appended to it."
#include "libbb.h"
#include "common_bufsiz.h"
#include "xregex.h"
#include <linux/netlink.h>
/* "mdev -s" scans /sys/class/xxx, looking for directories which have dev
* file (it is of the form "M:m\n"). Example: /sys/class/tty/tty0/dev
* contains "4:0\n". Directory name is taken as device name, path component
* directly after /sys/class/ as subsystem. In this example, "tty0" and "tty".
* Then mdev creates the /dev/device_name node.
2009-04-20 05:08:08 +05:30
* If /sys/class/.../dev file does not exist, mdev still may act
* on this device: see "@|$|*command args..." parameter in config file.
*
* mdev w/o parameters is called as hotplug helper. It takes device
* and subsystem names from $DEVPATH and $SUBSYSTEM, extracts
* maj,min from "/sys/$DEVPATH/dev" and also examines
* $ACTION ("add"/"delete") and $FIRMWARE.
*
* If action is "add", mdev creates /dev/device_name similarly to mdev -s.
* (todo: explain "delete" and $FIRMWARE)
*
* If /etc/mdev.conf exists, it may modify /dev/device_name's properties.
*
* Leading minus in 1st field means "don't stop on this line", otherwise
* search is stopped after the matching line is encountered.
*
* $envvar=regex format is useful for loading modules for hot-plugged devices
2009-04-20 05:08:08 +05:30
* which do not have driver loaded yet. In this case /sys/class/.../dev
* does not exist, but $MODALIAS is set to needed module's name
* (actually, an alias to it) by kernel. This rule instructs mdev
* to load the module and exit:
* $MODALIAS=.* 0:0 660 @modprobe "$MODALIAS"
* The kernel will generate another hotplug event when /sys/class/.../dev
* file appears.
*
* When line matches, the device node is created, chmod'ed and chown'ed,
* moved to path, and if >path, a symlink to moved node is created,
* all this if /sys/class/.../dev exists.
* Examples:
* =loop/ - moves to /dev/loop
* >disk/sda%1 - moves to /dev/disk/sdaN, makes /dev/sdaN a symlink
2009-04-20 05:08:08 +05:30
*
* Then "command args..." is executed (via sh -c 'command args...').
* @:execute on creation, $:on deletion, *:on both.
2009-04-20 05:08:08 +05:30
* This happens regardless of /sys/class/.../dev existence.
*/
/* Kernel's hotplug environment constantly changes.
* Here are new cases I observed on 3.1.0:
*
* Case with $DEVNAME and $DEVICE, not just $DEVPATH:
* ACTION=add
* BUSNUM=001
* DEVICE=/proc/bus/usb/001/003
* DEVNAME=bus/usb/001/003
* DEVNUM=003
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5
* DEVTYPE=usb_device
* MAJOR=189
* MINOR=2
* PRODUCT=18d1/4e12/227
* SUBSYSTEM=usb
* TYPE=0/0/0
*
* Case with $DEVICE, but no $DEVNAME - apparenty, usb iface notification?
* "Please load me a module" thing?
* ACTION=add
* DEVICE=/proc/bus/usb/001/003
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0
* DEVTYPE=usb_interface
* INTERFACE=8/6/80
* MODALIAS=usb:v18D1p4E12d0227dc00dsc00dp00ic08isc06ip50
* PRODUCT=18d1/4e12/227
* SUBSYSTEM=usb
* TYPE=0/0/0
*
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5
* DEVTYPE=scsi_host
* SUBSYSTEM=scsi
*
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/scsi_host/host5
* SUBSYSTEM=scsi_host
*
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/target5:0:0
* DEVTYPE=scsi_target
* SUBSYSTEM=scsi
*
* Case with strange $MODALIAS:
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/target5:0:0/5:0:0:0
* DEVTYPE=scsi_device
* MODALIAS=scsi:t-0x00
* SUBSYSTEM=scsi
*
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/target5:0:0/5:0:0:0/scsi_disk/5:0:0:0
* SUBSYSTEM=scsi_disk
*
* ACTION=add
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/target5:0:0/5:0:0:0/scsi_device/5:0:0:0
* SUBSYSTEM=scsi_device
*
* Case with explicit $MAJOR/$MINOR (no need to read /sys/$DEVPATH/dev?):
* ACTION=add
* DEVNAME=bsg/5:0:0:0
* DEVPATH=/devices/pci0000:00/0000:00:02.1/usb1/1-5/1-5:1.0/host5/target5:0:0/5:0:0:0/bsg/5:0:0:0
* MAJOR=253
* MINOR=1
* SUBSYSTEM=bsg
*
* ACTION=add
* DEVPATH=/devices/virtual/bdi/8:16
* SUBSYSTEM=bdi
*
* ACTION=add
* DEVNAME=sdb
* DEVPATH=/block/sdb
* DEVTYPE=disk
* MAJOR=8
* MINOR=16
* SUBSYSTEM=block
*
* Case with ACTION=change:
* ACTION=change
* DEVNAME=sdb
* DEVPATH=/block/sdb
* DEVTYPE=disk
* DISK_MEDIA_CHANGE=1
* MAJOR=8
* MINOR=16
* SUBSYSTEM=block
*/
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#define DEBUG_LVL 2
#if DEBUG_LVL >= 1
# define dbg1(...) do { if (G.verbose) bb_error_msg(__VA_ARGS__); } while(0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg1s(msg) do { if (G.verbose) bb_simple_error_msg(msg); } while(0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#else
# define dbg1(...) ((void)0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg1s(msg) ((void)0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#endif
#if DEBUG_LVL >= 2
# define dbg2(...) do { if (G.verbose >= 2) bb_error_msg(__VA_ARGS__); } while(0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg2s(msg) do { if (G.verbose >= 2) bb_simple_error_msg(msg); } while(0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#else
# define dbg2(...) ((void)0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg2s(msg) ((void)0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#endif
#if DEBUG_LVL >= 3
# define dbg3(...) do { if (G.verbose >= 3) bb_error_msg(__VA_ARGS__); } while(0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg3s(msg) do { if (G.verbose >= 3) bb_simple_error_msg(msg); } while(0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#else
# define dbg3(...) ((void)0)
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
# define dbg3s(msg) ((void)0)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
#endif
#ifndef SO_RCVBUFFORCE
#define SO_RCVBUFFORCE 33
#endif
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
static const char keywords[] ALIGN1 = "add\0remove\0"; // "change\0"
enum { OP_add, OP_remove };
struct envmatch {
struct envmatch *next;
char *envname;
regex_t match;
};
struct rule {
bool keep_matching;
bool regex_compiled;
mode_t mode;
int maj, min0, min1;
struct bb_uidgid_t ugid;
char *envvar;
char *ren_mov;
IF_FEATURE_MDEV_EXEC(char *r_cmd;)
regex_t match;
struct envmatch *envmatch;
};
struct globals {
int root_major, root_minor;
smallint verbose;
char *subsystem;
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
char *subsys_env; /* for putenv("SUBSYSTEM=subsystem") */
#if ENABLE_FEATURE_MDEV_CONF
const char *filename;
parser_t *parser;
struct rule **rule_vec;
unsigned rule_idx;
#endif
struct rule cur_rule;
char timestr[sizeof("HH:MM:SS.123456")];
} FIX_ALIASING;
#define G (*(struct globals*)bb_common_bufsiz1)
#define INIT_G() do { \
setup_common_bufsiz(); \
IF_NOT_FEATURE_MDEV_CONF(G.cur_rule.maj = -1;) \
IF_NOT_FEATURE_MDEV_CONF(G.cur_rule.mode = 0660;) \
} while (0)
/* Prevent infinite loops in /sys symlinks */
#define MAX_SYSFS_DEPTH 3
/* We use additional bytes in make_device() */
#define SCRATCH_SIZE 128
#if ENABLE_FEATURE_MDEV_CONF
static void make_default_cur_rule(void)
{
memset(&G.cur_rule, 0, sizeof(G.cur_rule));
G.cur_rule.maj = -1; /* "not a @major,minor rule" */
G.cur_rule.mode = 0660;
}
static void clean_up_cur_rule(void)
{
struct envmatch *e;
free(G.cur_rule.envvar);
free(G.cur_rule.ren_mov);
if (G.cur_rule.regex_compiled)
regfree(&G.cur_rule.match);
IF_FEATURE_MDEV_EXEC(free(G.cur_rule.r_cmd);)
e = G.cur_rule.envmatch;
while (e) {
free(e->envname);
regfree(&e->match);
e = e->next;
}
make_default_cur_rule();
}
static char *parse_envmatch_pfx(char *val)
{
struct envmatch **nextp = &G.cur_rule.envmatch;
for (;;) {
struct envmatch *e;
char *semicolon;
char *eq = strchr(val, '=');
if (!eq /* || eq == val? */)
return val;
if (endofname(val) != eq)
return val;
semicolon = strchr(eq, ';');
if (!semicolon)
return val;
/* ENVVAR=regex;... */
*nextp = e = xzalloc(sizeof(*e));
nextp = &e->next;
e->envname = xstrndup(val, eq - val);
*semicolon = '\0';
xregcomp(&e->match, eq + 1, REG_EXTENDED);
*semicolon = ';';
val = semicolon + 1;
}
}
static void parse_next_rule(void)
{
/* Note: on entry, G.cur_rule is set to default */
while (1) {
char *tokens[4];
char *val;
/* No PARSE_EOL_COMMENTS, because command may contain '#' chars */
if (!config_read(G.parser, tokens, 4, 3, "# \t", PARSE_NORMAL & ~PARSE_EOL_COMMENTS))
break;
/* Fields: [-]regex uid:gid mode [alias] [cmd] */
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("token1:'%s'", tokens[1]);
/* 1st field */
val = tokens[0];
G.cur_rule.keep_matching = ('-' == val[0]);
val += G.cur_rule.keep_matching; /* swallow leading dash */
val = parse_envmatch_pfx(val);
if (val[0] == '@') {
/* @major,minor[-minor2] */
/* (useful when name is ambiguous:
* "/sys/class/usb/lp0" and
* "/sys/class/printer/lp0")
*/
int sc = sscanf(val, "@%u,%u-%u", &G.cur_rule.maj, &G.cur_rule.min0, &G.cur_rule.min1);
if (sc < 2 || G.cur_rule.maj < 0) {
bb_error_msg("bad @maj,min on line %d", G.parser->lineno);
goto next_rule;
}
if (sc == 2)
G.cur_rule.min1 = G.cur_rule.min0;
} else {
char *eq = strchr(val, '=');
if (val[0] == '$') {
/* $ENVVAR=regex ... */
val++;
if (!eq) {
bb_error_msg("bad $envvar=regex on line %d", G.parser->lineno);
goto next_rule;
}
G.cur_rule.envvar = xstrndup(val, eq - val);
val = eq + 1;
}
xregcomp(&G.cur_rule.match, val, REG_EXTENDED);
G.cur_rule.regex_compiled = 1;
}
/* 2nd field: uid:gid - device ownership */
if (get_uidgid(&G.cur_rule.ugid, tokens[1]) == 0) {
bb_error_msg("unknown user/group '%s' on line %d", tokens[1], G.parser->lineno);
goto next_rule;
}
/* 3rd field: mode - device permissions */
G.cur_rule.mode = bb_parse_mode(tokens[2], G.cur_rule.mode);
/* 4th field (opt): ">|=alias" or "!" to not create the node */
val = tokens[3];
if (ENABLE_FEATURE_MDEV_RENAME && val && strchr(">=!", val[0])) {
char *s = skip_non_whitespace(val);
G.cur_rule.ren_mov = xstrndup(val, s - val);
val = skip_whitespace(s);
}
if (ENABLE_FEATURE_MDEV_EXEC && val && val[0]) {
const char *s = "$@*";
const char *s2 = strchr(s, val[0]);
if (!s2) {
bb_error_msg("bad line %u", G.parser->lineno);
goto next_rule;
}
IF_FEATURE_MDEV_EXEC(G.cur_rule.r_cmd = xstrdup(val);)
}
return;
next_rule:
clean_up_cur_rule();
} /* while (config_read) */
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("config_close(G.parser)");
config_close(G.parser);
G.parser = NULL;
return;
}
/* If mdev -s, we remember rules in G.rule_vec[].
* Otherwise, there is no point in doing it, and we just
* save only one parsed rule in G.cur_rule.
*/
static const struct rule *next_rule(void)
{
struct rule *rule;
/* Open conf file if we didn't do it yet */
if (!G.parser && G.filename) {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("config_open('%s')", G.filename);
G.parser = config_open2(G.filename, fopen_for_read);
G.filename = NULL;
}
if (G.rule_vec) {
/* mdev -s */
/* Do we have rule parsed already? */
if (G.rule_vec[G.rule_idx]) {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("< G.rule_vec[G.rule_idx:%d]=%p", G.rule_idx, G.rule_vec[G.rule_idx]);
return G.rule_vec[G.rule_idx++];
}
make_default_cur_rule();
} else {
/* not mdev -s */
clean_up_cur_rule();
}
/* Parse one more rule if file isn't fully read */
rule = &G.cur_rule;
if (G.parser) {
parse_next_rule();
if (G.rule_vec) { /* mdev -s */
rule = xmemdup(&G.cur_rule, sizeof(G.cur_rule));
G.rule_vec = xrealloc_vector(G.rule_vec, 4, G.rule_idx);
G.rule_vec[G.rule_idx++] = rule;
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("> G.rule_vec[G.rule_idx:%d]=%p", G.rule_idx, G.rule_vec[G.rule_idx]);
}
}
return rule;
}
static int env_matches(struct envmatch *e)
{
while (e) {
int r;
char *val = getenv(e->envname);
if (!val)
return 0;
r = regexec(&e->match, val, /*size*/ 0, /*range[]*/ NULL, /*eflags*/ 0);
if (r != 0) /* no match */
return 0;
e = e->next;
}
return 1;
}
#else
# define next_rule() (&G.cur_rule)
#endif
static void mkdir_recursive(char *name)
{
/* if name has many levels ("dir1/dir2"),
* bb_make_directory() will create dir1 according to umask,
* not according to its "mode" parameter.
* Since we run with umask=0, need to temporarily switch it.
*/
umask(022); /* "dir1" (if any) will be 0755 too */
bb_make_directory(name, 0755, FILEUTILS_RECUR);
umask(0);
}
/* Builds an alias path.
* This function potentionally reallocates the alias parameter.
* Only used for ENABLE_FEATURE_MDEV_RENAME
*/
static char *build_alias(char *alias, const char *device_name)
{
char *dest;
/* ">bar/": rename to bar/device_name */
/* ">bar[/]baz": rename to bar[/]baz */
dest = strrchr(alias, '/');
if (dest) { /* ">bar/[baz]" ? */
*dest = '\0'; /* mkdir bar */
mkdir_recursive(alias);
*dest = '/';
if (dest[1] == '\0') { /* ">bar/" => ">bar/device_name" */
dest = alias;
alias = concat_path_file(alias, device_name);
free(dest);
}
}
return alias;
}
/* mknod in /dev based on a path like "/sys/block/hda/hda1"
* NB1: path parameter needs to have SCRATCH_SIZE scratch bytes
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
* after NUL, but we promise to not mangle it (IOW: to restore NUL if needed).
* NB2: "mdev -s" may call us many times, do not leak memory/fds!
*
* device_name = $DEVNAME (may be NULL)
* path = /sys/$DEVPATH
*/
static void make_device(char *device_name, char *path, int operation)
{
int major, minor, type, len;
char *path_end = path + strlen(path);
/* Try to read major/minor string. Note that the kernel puts \n after
* the data, so we don't need to worry about null terminating the string
* because sscanf() will stop at the first nondigit, which \n is.
* We also depend on path having writeable space after it.
*/
major = -1;
if (operation == OP_add) {
strcpy(path_end, "/dev");
len = open_read_close(path, path_end + 1, SCRATCH_SIZE - 1);
*path_end = '\0';
if (len < 1) {
if (!ENABLE_FEATURE_MDEV_EXEC)
return;
2009-04-20 03:07:07 +05:30
/* no "dev" file, but we can still run scripts
* based on device name */
} else if (sscanf(path_end + 1, "%u:%u", &major, &minor) == 2) {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("dev %u,%u", major, minor);
} else {
major = -1;
}
}
/* else: for delete, -1 still deletes the node, but < -1 suppresses that */
2006-01-25 05:38:53 +05:30
/* Determine device name */
if (!device_name) {
/*
* There was no $DEVNAME envvar (for example, mdev -s never has).
* But it is very useful: it contains the *path*, not only basename,
* Thankfully, uevent file has it.
* Example of .../sound/card0/controlC0/uevent file on Linux-3.7.7:
* MAJOR=116
* MINOR=7
* DEVNAME=snd/controlC0
*/
strcpy(path_end, "/uevent");
len = open_read_close(path, path_end + 1, SCRATCH_SIZE - 1);
if (len < 0)
len = 0;
*path_end = '\0';
path_end[1 + len] = '\0';
device_name = strstr(path_end + 1, "\nDEVNAME=");
if (device_name) {
device_name += sizeof("\nDEVNAME=")-1;
strchrnul(device_name, '\n')[0] = '\0';
} else {
/* Fall back to just basename */
device_name = (char*) bb_basename(path);
}
}
/* Determine device type */
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
/*
* http://kernel.org/doc/pending/hotplug.txt says that only
* "/sys/block/..." is for block devices. "/sys/bus" etc is not.
* But since 2.6.25 block devices are also in /sys/class/block.
* We use strstr("/block/") to forestall future surprises.
*/
2008-10-29 06:49:15 +05:30
type = S_IFCHR;
libbb: introduce and use is_prefixed_with() function old new delta is_prefixed_with - 18 +18 complete_username 78 77 -1 man_main 737 735 -2 fsck_device 429 427 -2 unpack_ar_archive 80 76 -4 strip_unsafe_prefix 105 101 -4 singlemount 1054 1050 -4 rtc_adjtime_is_utc 90 86 -4 resolve_mount_spec 88 84 -4 parse_one_line 1029 1025 -4 parse_conf 1460 1456 -4 may_wakeup 83 79 -4 loadkmap_main 219 215 -4 get_irqs_from_stat 103 99 -4 get_header_cpio 913 909 -4 findfs_main 79 75 -4 fbsplash_main 1230 1226 -4 load_crontab 776 771 -5 expand_vars_to_list 1151 1146 -5 date_main 881 876 -5 skip_dev_pfx 30 24 -6 make_device 2199 2193 -6 complete_cmd_dir_file 773 767 -6 run_applet_and_exit 715 708 -7 uudecode_main 321 313 -8 pwdx_main 197 189 -8 execute 568 560 -8 i2cdetect_main 1186 1176 -10 procps_scan 1242 1230 -12 procps_read_smaps 1017 1005 -12 process_module 746 734 -12 patch_main 1903 1891 -12 nfsmount 3572 3560 -12 stack_machine 126 112 -14 process_timer_stats 449 435 -14 match_fstype 111 97 -14 do_ipaddr 1344 1330 -14 open_list_and_close 359 343 -16 get_header_tar 1795 1779 -16 prepend_new_eth_table 340 323 -17 fsck_main 1811 1794 -17 find_iface_state 56 38 -18 dnsd_main 1321 1303 -18 base_device 179 158 -21 find_keyword 104 82 -22 handle_incoming_and_exit 2785 2762 -23 parse_and_put_prompt 774 746 -28 modinfo 347 317 -30 find_action 204 171 -33 update_passwd 1470 1436 -34 ------------------------------------------------------------------------------ (add/remove: 1/0 grow/shrink: 0/49 up/down: 18/-540) Total: -522 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-03-12 22:18:34 +05:30
if (strstr(path, "/block/") || (G.subsystem && is_prefixed_with(G.subsystem, "block")))
2008-10-29 06:49:15 +05:30
type = S_IFBLK;
2006-01-25 05:38:53 +05:30
#if ENABLE_FEATURE_MDEV_CONF
G.rule_idx = 0; /* restart from the beginning (think mdev -s) */
#endif
for (;;) {
const char *str_to_match;
regmatch_t off[1 + 9 * ENABLE_FEATURE_MDEV_RENAME_REGEXP];
char *command;
char *alias;
char aliaslink = aliaslink; /* for compiler */
char *node_name;
const struct rule *rule;
str_to_match = device_name;
rule = next_rule();
#if ENABLE_FEATURE_MDEV_CONF
if (!env_matches(rule->envmatch))
continue;
if (rule->maj >= 0) { /* @maj,min rule */
if (major != rule->maj)
continue;
if (minor < rule->min0 || minor > rule->min1)
continue;
memset(off, 0, sizeof(off));
goto rule_matches;
}
if (rule->envvar) { /* $envvar=regex rule */
str_to_match = getenv(rule->envvar);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("getenv('%s'):'%s'", rule->envvar, str_to_match);
if (!str_to_match)
continue;
}
/* else: str_to_match = device_name */
if (rule->regex_compiled) {
int regex_match = regexec(&rule->match, str_to_match, ARRAY_SIZE(off), off, 0);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("regex_match for '%s':%d", str_to_match, regex_match);
//bb_error_msg("matches:");
//for (int i = 0; i < ARRAY_SIZE(off); i++) {
// if (off[i].rm_so < 0) continue;
// bb_error_msg("match %d: '%.*s'\n", i,
// (int)(off[i].rm_eo - off[i].rm_so),
// device_name + off[i].rm_so);
//}
if (regex_match != 0
/* regexec returns whole pattern as "range" 0 */
|| off[0].rm_so != 0
|| (int)off[0].rm_eo != (int)strlen(str_to_match)
) {
continue; /* this rule doesn't match */
2009-04-20 03:07:07 +05:30
}
}
/* else: it's final implicit "match-all" rule */
rule_matches:
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg2("rule matched, line %d", G.parser ? G.parser->lineno : -1);
#endif
/* Build alias name */
alias = NULL;
if (ENABLE_FEATURE_MDEV_RENAME && rule->ren_mov) {
aliaslink = rule->ren_mov[0];
if (aliaslink == '!') {
/* "!": suppress node creation/deletion */
major = -2;
}
else if (aliaslink == '>' || aliaslink == '=') {
if (ENABLE_FEATURE_MDEV_RENAME_REGEXP) {
char *s;
char *p;
unsigned n;
/* substitute %1..9 with off[1..9], if any */
n = 0;
s = rule->ren_mov;
while (*s)
if (*s++ == '%')
n++;
p = alias = xzalloc(strlen(rule->ren_mov) + n * strlen(str_to_match));
s = rule->ren_mov + 1;
while (*s) {
*p = *s;
if ('%' == *s) {
unsigned i = (s[1] - '0');
if (i <= 9 && off[i].rm_so >= 0) {
n = off[i].rm_eo - off[i].rm_so;
strncpy(p, str_to_match + off[i].rm_so, n);
p += n - 1;
s++;
}
}
p++;
s++;
}
} else {
alias = xstrdup(rule->ren_mov + 1);
}
}
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("alias:'%s'", alias);
command = NULL;
IF_FEATURE_MDEV_EXEC(command = rule->r_cmd;)
if (command) {
/* Are we running this command now?
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
* Run @cmd on create, $cmd on delete, *cmd on any
*/
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
if ((command[0] == '@' && operation == OP_add)
|| (command[0] == '$' && operation == OP_remove)
|| (command[0] == '*')
) {
command++;
} else {
command = NULL;
}
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("command:'%s'", command);
/* "Execute" the line we found */
node_name = device_name;
if (ENABLE_FEATURE_MDEV_RENAME && alias) {
node_name = alias = build_alias(alias, device_name);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg3("alias2:'%s'", alias);
}
if (operation == OP_add && major >= 0) {
char *slash = strrchr(node_name, '/');
if (slash) {
*slash = '\0';
mkdir_recursive(node_name);
*slash = '/';
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
if (ENABLE_FEATURE_MDEV_CONF) {
dbg1("mknod %s (%d,%d) %o"
" %u:%u",
node_name, major, minor, rule->mode | type,
rule->ugid.uid, rule->ugid.gid
);
} else {
dbg1("mknod %s (%d,%d) %o",
node_name, major, minor, rule->mode | type
);
}
if (mknod(node_name, rule->mode | type, makedev(major, minor)) && errno != EEXIST)
bb_perror_msg("can't create '%s'", node_name);
if (ENABLE_FEATURE_MDEV_CONF) {
chmod(node_name, rule->mode);
chown(node_name, rule->ugid.uid, rule->ugid.gid);
}
if (major == G.root_major && minor == G.root_minor)
symlink(node_name, "root");
if (ENABLE_FEATURE_MDEV_RENAME && alias) {
if (aliaslink == '>') {
//TODO: on devtmpfs, device_name already exists and symlink() fails.
//End result is that instead of symlink, we have two nodes.
//What should be done?
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("symlink: %s", device_name);
symlink(node_name, device_name);
}
}
}
if (ENABLE_FEATURE_MDEV_EXEC && command) {
/* setenv will leak memory, use putenv/unsetenv/free */
char *s = xasprintf("%s=%s", "MDEV", node_name);
putenv(s);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("running: %s", command);
if (system(command) == -1)
bb_perror_msg("can't run '%s'", command);
bb_unsetenv_and_free(s);
}
if (operation == OP_remove && major >= -1) {
if (ENABLE_FEATURE_MDEV_RENAME && alias) {
if (aliaslink == '>') {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("unlink: %s", device_name);
unlink(device_name);
}
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("unlink: %s", node_name);
unlink(node_name);
}
if (ENABLE_FEATURE_MDEV_RENAME)
free(alias);
/* We found matching line.
* Stop unless it was prefixed with '-'
*/
if (!ENABLE_FEATURE_MDEV_CONF || !rule->keep_matching)
break;
} /* for (;;) */
}
static ssize_t readlink2(char *buf, size_t bufsize)
{
// Grr... gcc 8.1.1:
// "passing argument 2 to restrict-qualified parameter aliases with argument 1"
// dance around that...
char *obuf FIX_ALIASING;
obuf = buf;
return readlink(buf, obuf, bufsize);
}
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
/* File callback for /sys/ traversal.
* We act only on "/sys/.../dev" (pseudo)file
*/
static int FAST_FUNC fileAction(struct recursive_state *state,
const char *fileName,
struct stat *statbuf UNUSED_PARAM)
{
size_t len = strlen(fileName) - 4; /* can't underflow */
char *path = state->userData; /* char array[PATH_MAX + SCRATCH_SIZE] */
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
char subsys[PATH_MAX];
int res;
/* Is it a ".../dev" file? (len check is for paranoid reasons) */
if (strcmp(fileName + len, "/dev") != 0 || len >= PATH_MAX - 32)
return FALSE; /* not .../dev */
strcpy(path, fileName);
path[len] = '\0';
/* Read ".../subsystem" symlink in the same directory where ".../dev" is */
strcpy(subsys, path);
strcpy(subsys + len, "/subsystem");
res = readlink2(subsys, sizeof(subsys)-1);
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
if (res > 0) {
subsys[res] = '\0';
free(G.subsystem);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
if (G.subsys_env) {
bb_unsetenv_and_free(G.subsys_env);
G.subsys_env = NULL;
}
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
/* Set G.subsystem and $SUBSYSTEM from symlink's last component */
G.subsystem = strrchr(subsys, '/');
if (G.subsystem) {
G.subsystem = xstrdup(G.subsystem + 1);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
G.subsys_env = xasprintf("%s=%s", "SUBSYSTEM", G.subsystem);
putenv(G.subsys_env);
}
}
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
make_device(/*DEVNAME:*/ NULL, path, OP_add);
return TRUE;
}
/* Directory callback for /sys/ traversal */
static int FAST_FUNC dirAction(struct recursive_state *state,
const char *fileName UNUSED_PARAM,
struct stat *statbuf UNUSED_PARAM)
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
{
return (state->depth >= MAX_SYSFS_DEPTH ? SKIP : TRUE);
}
2007-06-13 13:04:15 +05:30
/* For the full gory details, see linux/Documentation/firmware_class/README
*
* Firmware loading works like this:
* - kernel sets FIRMWARE env var
* - userspace checks /lib/firmware/$FIRMWARE
* - userspace waits for /sys/$DEVPATH/loading to appear
* - userspace writes "1" to /sys/$DEVPATH/loading
* - userspace copies /lib/firmware/$FIRMWARE into /sys/$DEVPATH/data
* - userspace writes "0" (worked) or "-1" (failed) to /sys/$DEVPATH/loading
* - kernel loads firmware into device
*/
static void load_firmware(const char *firmware, const char *sysfs_path)
2007-06-13 13:04:15 +05:30
{
int cnt;
int firmware_fd, loading_fd;
2007-06-13 13:04:15 +05:30
/* check for /lib/firmware/$FIRMWARE */
mdev: create devices from /sys/dev Currently some new devices that have a bus but no class will be missed by mdev coldplug device creation after boot. This happens because mdev recursively searches /sys/class which will by definition only find class devices. Some important devices such as iio and gpiochip does not have a class. But users will need them. This switches from using /sys/class as the place to look for devices to create to using /sys/dev where all char and block devices are listed. The subsystem lookup code that provide the G.subsystem environment variable is changed from using the directory name of the class device to instead dereference the "subsystem" symlink for the device, and look at the last element of the path of the symlink for the subsystem, which will work with class devices and bus devices alike. (The new bus-only devices only symlink to the /sys/bus/* hierarchy.) We delete the legacy kernel v2.6.2x /sys/block device path code as part of this change. It's too old to be kept alive. Tested on kernel v4.6-rc2 with a bunch of devices, including some IIO and gpiochip devices. With a print inserted before make_device() the log looks like so: Create device from "/sys/dev/char/1:1", subsystem "mem" Create device from "/sys/dev/char/1:2", subsystem "mem" Create device from "/sys/dev/char/1:3", subsystem "mem" Create device from "/sys/dev/char/1:5", subsystem "mem" (...) Create device from "/sys/dev/block/179:56", subsystem "block" Create device from "/sys/dev/block/179:64", subsystem "block" function old new delta mdev_main 1388 1346 -42 dirAction 134 14 -120 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes Cc: Isaac Dunham <ibid.ag@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jonathan Cameron <jic23@cam.ac.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-07 17:39:01 +05:30
firmware_fd = -1;
if (chdir("/lib/firmware") == 0)
firmware_fd = open(firmware, O_RDONLY); /* can fail */
2007-06-13 13:04:15 +05:30
/* check for /sys/$DEVPATH/loading ... give 30 seconds to appear */
xchdir(sysfs_path);
for (cnt = 0; cnt < 30; ++cnt) {
loading_fd = open("loading", O_WRONLY);
if (loading_fd >= 0)
goto loading;
sleep(1);
2007-06-13 13:04:15 +05:30
}
goto out;
2007-06-13 13:04:15 +05:30
loading:
cnt = 0;
if (firmware_fd >= 0) {
int data_fd;
/* tell kernel we're loading by "echo 1 > /sys/$DEVPATH/loading" */
if (full_write(loading_fd, "1", 1) != 1)
goto out;
/* load firmware into /sys/$DEVPATH/data */
data_fd = open("data", O_WRONLY);
if (data_fd < 0)
goto out;
cnt = bb_copyfd_eof(firmware_fd, data_fd);
if (ENABLE_FEATURE_CLEAN_UP)
close(data_fd);
}
2007-06-13 13:04:15 +05:30
/* Tell kernel result by "echo [0|-1] > /sys/$DEVPATH/loading"
* Note: we emit -1 also if firmware file wasn't found.
* There are cases when otherwise kernel would wait for minutes
* before timing out.
*/
2007-06-13 13:04:15 +05:30
if (cnt > 0)
full_write(loading_fd, "0", 1);
2007-06-13 13:04:15 +05:30
else
full_write(loading_fd, "-1", 2);
2007-06-13 13:04:15 +05:30
out:
xchdir("/dev");
2007-06-13 13:04:15 +05:30
if (ENABLE_FEATURE_CLEAN_UP) {
close(firmware_fd);
close(loading_fd);
}
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
static char *curtime(void)
{
struct timeval tv;
gettimeofday(&tv, NULL);
sprintf(
strftime_HHMMSS(G.timestr, sizeof(G.timestr), &tv.tv_sec),
".%06u",
(unsigned)tv.tv_usec
);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
return G.timestr;
}
static void open_mdev_log(const char *seq, unsigned my_pid)
{
int logfd = open("mdev.log", O_WRONLY | O_APPEND);
if (logfd >= 0) {
xmove_fd(logfd, STDERR_FILENO);
G.verbose = 2;
applet_name = xasprintf("%s[%s]", applet_name, seq ? seq : utoa(my_pid));
}
}
/* If it exists, does /dev/mdev.seq match $SEQNUM?
* If it does not match, earlier mdev is running
* in parallel, and we need to wait.
* Active mdev pokes us with SIGCHLD to check the new file.
*/
static int
wait_for_seqfile(unsigned expected_seq)
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
{
/* We time out after 2 sec */
static const struct timespec ts = { 0, 32*1000*1000 };
int timeout = 2000 / 32;
int seq_fd = -1;
int do_once = 1;
sigset_t set_CHLD;
sigemptyset(&set_CHLD);
sigaddset(&set_CHLD, SIGCHLD);
sigprocmask(SIG_BLOCK, &set_CHLD, NULL);
for (;;) {
int seqlen;
char seqbuf[sizeof(long)*3 + 2];
unsigned seqbufnum;
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
if (seq_fd < 0) {
seq_fd = open("mdev.seq", O_RDWR);
if (seq_fd < 0)
break;
close_on_exec_on(seq_fd);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
}
seqlen = pread(seq_fd, seqbuf, sizeof(seqbuf) - 1, 0);
if (seqlen < 0) {
close(seq_fd);
seq_fd = -1;
break;
}
seqbuf[seqlen] = '\0';
if (seqbuf[0] == '\n' || seqbuf[0] == '\0') {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
/* seed file: write out seq ASAP */
xwrite_str(seq_fd, utoa(expected_seq));
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
xlseek(seq_fd, 0, SEEK_SET);
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
dbg2s("first seq written");
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
break;
}
seqbufnum = atoll(seqbuf);
if (seqbufnum == expected_seq) {
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
/* correct idx */
break;
}
if (seqbufnum > expected_seq) {
/* a later mdev runs already (this was seen by users to happen) */
/* do not overwrite seqfile on exit */
close(seq_fd);
seq_fd = -1;
break;
}
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
if (do_once) {
dbg2("%s mdev.seq='%s', need '%u'", curtime(), seqbuf, expected_seq);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
do_once = 0;
}
if (sigtimedwait(&set_CHLD, NULL, &ts) >= 0) {
dbg3("woken up");
continue; /* don't decrement timeout! */
}
if (--timeout == 0) {
dbg1("%s mdev.seq='%s'", "timed out", seqbuf);
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
break;
}
}
sigprocmask(SIG_UNBLOCK, &set_CHLD, NULL);
return seq_fd;
}
static void signal_mdevs(unsigned my_pid)
{
procps_status_t* p = NULL;
while ((p = procps_scan(p, PSSCAN_ARGV0)) != NULL) {
if (p->pid != my_pid
&& p->argv0
&& strcmp(bb_basename(p->argv0), "mdev") == 0
) {
kill(p->pid, SIGCHLD);
}
}
}
static void process_action(char *temp, unsigned my_pid)
{
char *fw;
char *seq;
char *action;
char *env_devname;
char *env_devpath;
unsigned seqnum = seqnum; /* for compiler */
int seq_fd;
smalluint op;
/* Hotplug:
* env ACTION=... DEVPATH=... SUBSYSTEM=... [SEQNUM=...] mdev
* ACTION can be "add", "remove", "change"
* DEVPATH is like "/block/sda" or "/class/input/mice"
*/
env_devname = getenv("DEVNAME"); /* can be NULL */
G.subsystem = getenv("SUBSYSTEM");
action = getenv("ACTION");
env_devpath = getenv("DEVPATH");
if (!action || !env_devpath /*|| !G.subsystem*/)
bb_show_usage();
fw = getenv("FIRMWARE");
seq = getenv("SEQNUM");
op = index_in_strings(keywords, action);
if (my_pid)
open_mdev_log(seq, my_pid);
seq_fd = -1;
if (my_pid && seq) {
seqnum = atoll(seq);
seq_fd = wait_for_seqfile(seqnum);
}
dbg1("%s "
"ACTION:%s SEQNUM:%s SUBSYSTEM:%s DEVNAME:%s DEVPATH:%s"
"%s%s",
curtime(),
action, seq, G.subsystem, env_devname, env_devpath,
fw ? " FW:" : "", fw ? fw : ""
);
snprintf(temp, PATH_MAX, "/sys%s", env_devpath);
if (op == OP_remove) {
/* Ignoring "remove firmware". It was reported
* to happen and to cause erroneous deletion
* of device nodes. */
if (!fw)
make_device(env_devname, temp, op);
}
else {
make_device(env_devname, temp, op);
if (ENABLE_FEATURE_MDEV_LOAD_FIRMWARE) {
if (op == OP_add && fw)
load_firmware(fw, temp);
}
}
if (seq_fd >= 0) {
xwrite_str(seq_fd, utoa(seqnum + 1));
signal_mdevs(my_pid);
}
}
static void initial_scan(char *temp)
{
struct stat st;
xstat("/", &st);
G.root_major = major(st.st_dev);
G.root_minor = minor(st.st_dev);
putenv((char*)"ACTION=add");
/* Create all devices from /sys/dev hierarchy */
recursive_action("/sys/dev",
ACTION_RECURSE | ACTION_FOLLOWLINKS,
fileAction, dirAction, temp);
}
#if ENABLE_FEATURE_MDEV_DAEMON
/* uevent applet uses 16k buffer, and mmaps it before every read */
# define BUFFER_SIZE (2 * 1024)
# define RCVBUF (2 * 1024 * 1024)
# define MAX_ENV 32
static void daemon_loop(char *temp, int fd)
{
for (;;) {
char netbuf[BUFFER_SIZE];
char *env[MAX_ENV];
char *s, *end;
ssize_t len;
int idx;
len = safe_read(fd, netbuf, sizeof(netbuf) - 1);
if (len < 0) {
libbb: reduce the overhead of single parameter bb_error_msg() calls Back in 2007, commit 0c97c9d43707 ("'simple' error message functions by Loic Grenie") introduced bb_simple_perror_msg() to allow for a lower overhead call to bb_perror_msg() when only a string was being printed with no parameters. This saves space for some CPU architectures because it avoids the overhead of a call to a variadic function. However there has never been a simple version of bb_error_msg(), and since 2007 many new calls to bb_perror_msg() have been added that only take a single parameter and so could have been using bb_simple_perror_message(). This changeset introduces 'simple' versions of bb_info_msg(), bb_error_msg(), bb_error_msg_and_die(), bb_herror_msg() and bb_herror_msg_and_die(), and replaces all calls that only take a single parameter, or use something like ("%s", arg), with calls to the corresponding 'simple' version. Since it is likely that single parameter calls to the variadic functions may be accidentally reintroduced in the future a new debugging config option WARN_SIMPLE_MSG has been introduced. This uses some macro magic which will cause any such calls to generate a warning, but this is turned off by default to avoid use of the unpleasant macros in normal circumstances. This is a large changeset due to the number of calls that have been replaced. The only files that contain changes other than simple substitution of function calls are libbb.h, libbb/herror_msg.c, libbb/verror_msg.c and libbb/xfuncs_printf.c. In miscutils/devfsd.c, networking/udhcp/common.h and util-linux/mdev.c additonal macros have been added for logging so that single parameter and multiple parameter logging variants exist. The amount of space saved varies considerably by architecture, and was found to be as follows (for 'defconfig' using GCC 7.4): Arm: -92 bytes MIPS: -52 bytes PPC: -1836 bytes x86_64: -938 bytes Note that for the MIPS architecture only an exception had to be made disabling the 'simple' calls for 'udhcp' (in networking/udhcp/common.h) because it made these files larger on MIPS. Signed-off-by: James Byrne <james.byrne@origamienergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2019-07-02 15:05:03 +05:30
bb_simple_perror_msg_and_die("read");
}
end = netbuf + len;
*end = '\0';
idx = 0;
s = netbuf;
while (s < end && idx < MAX_ENV) {
if (endofname(s)[0] == '=') {
env[idx++] = s;
putenv(s);
}
s += strlen(s) + 1;
}
process_action(temp, 0);
while (idx)
bb_unsetenv(env[--idx]);
}
}
#endif
int mdev_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
int mdev_main(int argc UNUSED_PARAM, char **argv)
{
enum {
MDEV_OPT_SCAN = 1 << 0,
MDEV_OPT_DAEMON = 1 << 1,
MDEV_OPT_FOREGROUND = 1 << 2,
};
int opt;
RESERVE_CONFIG_BUFFER(temp, PATH_MAX + SCRATCH_SIZE);
INIT_G();
/* We can be called as hotplug helper */
/* Kernel cannot provide suitable stdio fds for us, do it ourself */
bb_sanitize_stdio();
2009-04-20 03:07:07 +05:30
/* Force the configuration file settings exactly */
umask(0);
xchdir("/dev");
opt = getopt32(argv, "s" IF_FEATURE_MDEV_DAEMON("df"));
#if ENABLE_FEATURE_MDEV_CONF
G.filename = "/etc/mdev.conf";
if (opt & (MDEV_OPT_SCAN|MDEV_OPT_DAEMON)) {
/* Same as xrealloc_vector(NULL, 4, 0): */
G.rule_vec = xzalloc((1 << 4) * sizeof(*G.rule_vec));
}
#endif
#if ENABLE_FEATURE_MDEV_DAEMON
if (opt & MDEV_OPT_DAEMON) {
/*
* Daemon mode listening on uevent netlink socket.
*/
int fd;
/* Subscribe for UEVENT kernel messages */
/* Without a sufficiently big RCVBUF, a ton of simultaneous events
* can trigger ENOBUFS on read, which is unrecoverable.
* Reproducer:
* mdev -d
* find /sys -name uevent -exec sh -c 'echo add >"{}"' ';'
*/
fd = create_and_bind_to_netlink(NETLINK_KOBJECT_UEVENT, 1 << 0, RCVBUF);
/*
* Make inital scan after the uevent socket is alive and
* _before_ we fork away.
*/
initial_scan(temp);
if (!(opt & MDEV_OPT_FOREGROUND))
bb_daemonize_or_rexec(0, argv);
open_mdev_log(NULL, getpid());
daemon_loop(temp, fd);
}
#endif
if (opt & MDEV_OPT_SCAN) {
/*
* Scan: mdev -s
*/
initial_scan(temp);
} else {
process_action(temp, getpid());
mdev: improve $SEQ handling; improve debug logging Sequential run of concurrent mdev's was too simplistic: they waited for /dev/mdev.seq to match. This could sometimes cause cumulative loss of time on the order of a second. Added SIGCHLD signaling from exiting mdev to all other mdev's. Added debugging required to see that code actually works as intended. Example of /dev/mdev.log (with "woken up" elevated from dbg lvl 3 to 2): mdev[1023]: first seq written ^^^^ seq, not pid mdev[1023]: 35.022395 ACTION:add SUBSYSTEM:module DEVNAME:(null) DEVPATH:/module/lib80211 mdev[1023]: rule matched, line -1 ^^^^^^^ means "default rule" mdev[1023]: 35.022676 exiting ^^^^^^^^^ second,usec timestamp mdev[1024]: 35.069691 ACTION:add SUBSYSTEM:vc DEVNAME:vcs9 DEVPATH:/devices/virtual/vc/vcs9 mdev[1024]: dev 7,9 mdev[1025]: 35.069889 waiting for '1024' mdev[1026]: 35.069946 waiting for '1024' mdev[1027]: 35.070151 waiting for '1024' mdev[1024]: rule matched, line -1 mdev[1024]: mknod vcs9 (7,9) 20660 0:0 mdev[1024]: 35.070346 exiting mdev[1025]: woken up mdev[1026]: woken up mdev[1025]: 35.071213 ACTION:add SUBSYSTEM:vc DEVNAME:vcsa9 DEVPATH:/devices/virtual/vc/vcsa9 ^^^^^^^^^ took only a millisecond to start running after prev mdev exited mdev[1025]: dev 7,137 mdev[1027]: woken up mdev[1025]: rule matched, line -1 mdev[1025]: mknod vcsa9 (7,137) 20660 0:0 mdev[1025]: 35.072109 exiting function old new delta mdev_main 849 1372 +523 curtime - 59 +59 dirAction 87 134 +47 static.ts - 8 +8 keywords 19 12 -7 make_device 2189 2119 -70 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2013-02-27 15:21:41 +05:30
dbg1("%s exiting", curtime());
}
if (ENABLE_FEATURE_CLEAN_UP)
RELEASE_CONFIG_BUFFER(temp);
return EXIT_SUCCESS;
}