e2fsck: /dev/hda2 is mounted.

e2fsck: /dev/hda2 is mounted.

[George Spelvin]

After a recent upgrade of a Debian system, I discovered that
systems weren't rebooting; they would always halt in the automatic
fsck of other file systems.

Upon investigation, it turned out that fsck didn't understand the
kernel "root=302" parameter and was trying to run e2fsck on the root
file system.  This produced an error, causing the boot process
to drop into a recovery shell.

System is i386, current Debian sid (unstable).  But I compiled
util-linux from the git sources to eliminate Debianisms from
the problem.

Running the v2.28-rc1 release (8c97556ecfc0af358d35dc7a842aaadde5eeafd3)
I got some compile warnings (appended for reference), but a
working execurable which could be run as follows:

LIBMOUNT_DEBUG=all LIBBLKID_DEBUG=all ./fsck -C -M -A -V

32305: libmount:     INIT: library debug mask: 0xffff
32305: libmount:     INIT: library version: 2.28.
32305: libmount:     INIT:     feature: btrfs
32305: libmount:     INIT:     feature: assert
32305: libmount:     INIT:     feature: debug
Available "LIBMOUNT_DEBUG=<name>[,...]|<mask>" debug masks:
   all      [0xffff] : info about all subsystems
   cache    [0x0004] : paths and tags cache
   cxt      [0x0200] : library context (handler)
   diff     [0x0400] : mountinfo changes tracking
   fs       [0x0040] : FS abstraction
   help     [0x0001] : this help
   locks    [0x0010] : mtab and utab locking
   loop     [0x2000] : loop devices routines
   options  [0x0008] : mount options parsing
   tab      [0x0020] : fstab, mtab, mounninfo routines
   update   [0x0080] : mtab, utab updates
   utils    [0x0100] : misc library utils
   monitor  [0x0800] : mount tables monitor
   btrfs    [0x1000] : btrfs specific routines
32305: libmount:    CACHE: [0x983a048]: alloc
lt-fsck from util-linux 2.28-rc1
32305: libmount:      TAB: [0x983a068]: alloc
32305: libmount:      TAB: [0x983a068]: /etc/fstab: start parsing [entries=0, filter=not]
32305: libmount:      TAB: [0x983a068]: add entry: /dev/hda1 none
32305: libmount:      TAB: [0x983a068]: add entry: /dev/hda2 /
32305: libmount:      TAB: [0x983a068]: add entry: proc /proc
32305: libmount:      TAB: [0x983a068]: add entry: tmpfs /tmp
32305: libmount:      TAB: [0x983a068]: add entry: /dev/fd0 /a
32305: libmount:      TAB: [0x983a068]: add entry: /dev/hdc /cd
32305: libmount:      TAB: [0x983a068]: add entry: /dev/hdd /dvd
32305: libmount:      TAB: [0x983a068]: add entry: /dev/hda3 /home
32305: libmount:       FS: [0x983a9c8]: free [refcount=0]
32305: libmount:      TAB: [0x983a068]: /etc/fstab: stop parsing (8 entries)
32305: libmount:      TAB: [0x983a068]: parsing done [filename=/etc/fstab, rc=0]
32305: libmount:    CACHE: [0x983a048]: canonicalize path /dev/hda2
32305: libmount:    CACHE: [0x983a048]: add entry [ 1] (path): /dev/hda2: /dev/hda2
32305: libmount:    CACHE: [0x983a048]: canonicalize path /dev/hda3
32305: libmount:    CACHE: [0x983a048]: add entry [ 2] (path): /dev/hda3: /dev/hda3
Checking all file systems.
32305: libmount:      TAB: [0x983a068]: lookup TARGET: '/'
32305: libmount:      TAB: [0x983a160]: alloc
32305: libmount:    UTILS: mtab: /etc/mtab
32305: libmount:    UTILS: /etc/mtab: irregular/non-writable
32305: libmount:      TAB: [0x983a160]: mtab parse: #1 read mountinfo
32305: libmount:      TAB: [0x983a160]: /proc/self/mountinfo: start parsing [entries=0, filter=not]
32305: libmount:      TAB: [0x983a160]: add entry: /dev/root /
32305: libmount:    CACHE: canonicalize path /proc/self/mountinfo
32305: libmount:      TAB: TID for /proc/self/mountinfo is 32305
32305: libmount:      TAB: [0x983a160]: root FS: 302
32305: libmount:    CACHE: [0x983a048]: canonicalize path 302
32305: libmount:    CACHE: [0x983a048]: add entry [ 3] (path): 302: 302
32305: libmount:      TAB: [0x983a160]: canonical root FS: 302
32305: libmount:      TAB: [0x983a160]: add entry: sysfs /sys
32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run
32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run/lock
32305: libmount:      TAB: [0x983a160]: add entry: proc /proc
32305: libmount:      TAB: [0x983a160]: add entry: devtmpfs /dev
32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run/shm
32305: libmount:      TAB: [0x983a160]: add entry: devpts /dev/pts
32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /tmp
32305: libmount:      TAB: [0x983a160]: add entry: /dev/hda3 /home
32305: libmount:       FS: [0x983ddc8]: free [refcount=0]
32305: libmount:      TAB: [0x983a160]: /proc/self/mountinfo: stop parsing (10 entries)
32305: libmount:      TAB: [0x983a160]: parsing done [filename=/proc/self/mountinfo, rc=0]
32305: libmount:      TAB: [0x983a160]: mtab parse: #2 read utab
32305: libmount:      TAB: [0x983a160]: lookup SOURCE: '/dev/hda2'
32305: libmount:      TAB: [0x983a160]: lookup SRCPATH: '/dev/hda2'
32305: libmount:      TAB: [0x983a160]: lookup canonical SRCPATH: '/dev/hda2'
/dev/hda2 is not mounted
[/sbin/fsck.ext4 (1) -- /] fsck.ext4 -C0 /dev/hda2 
32318: libblkid:     INIT: library debug mask: 0xffff
32318: libblkid:     INIT: library version: 2.27.0 [02-Nov-2015]
Available "LIBBLKID_DEBUG=<name>[,...]|<mask>" debug masks:
   all      [0xffff] : info about all subsystems
   cache    [0x0004] : blkid tags cache
   config   [0x0008] : config file utils
   dev      [0x0010] : device utils
   devname  [0x0020] : /proc/partitions evaluation
   devno    [0x0040] : convertions to device name
   evaluate [0x0080] : tags resolving
   help     [0x0001] : this help
   lowprobe [0x0100] : superblock/raids/partitions probing
   probe    [0x0200] : devices verification
   read     [0x0400] : cache parsing
   save     [0x0800] : cache writing
   tag      [0x1000] : tags utils
32318: libblkid:    CACHE: creating blkid cache (using default cache)
32318: libblkid:   CONFIG: reading config file: /etc/blkid.conf.
32318: libblkid:   CONFIG: /etc/blkid.conf: does not exist, using built-in default
32318: libblkid:      TAG: looking for /dev/hda2 in cache
e2fsck 1.43-WIP (18-May-2015)
/dev/hda2 is mounted.
e2fsck: Cannot continue, aborting.


I'm not quite sure where in the code this is *supposed* to be parsed
so I haven't found the bug yet.  It *used* to work, but I fear it's
been months since my last reboot and I don't have a list of everything
that changed.



Other problems encountered during compilation:

checking linux/gsmmux.h usability... no
checking linux/gsmmux.h presence... yes
configure: WARNING: linux/gsmmux.h: present but cannot be compiled
configure: WARNING: linux/gsmmux.h:     check for missing prerequisite headers?
configure: WARNING: linux/gsmmux.h: see the Autoconf documentation
configure: WARNING: linux/gsmmux.h:     section "Present But Cannot Be Compiled"
configure: WARNING: linux/gsmmux.h: proceeding with the compiler's result
configure: WARNING:     ## ------------------------------ ##
configure: WARNING:     ## Report this to kzak@redhat.com ##
configure: WARNING:     ## ------------------------------ ##
checking for linux/gsmmux.h... no


The details from config.log look more like a problem with the headers
rather than util-linux, but it says report it, so I am:

configure:18012: checking linux/gsmmux.h usability
configure:18012: gcc -c -g -O2  conftest.c >&5
In file included from /usr/include/linux/gsmmux.h:4:0,
                 from conftest.c:79:
/usr/include/linux/if.h:211:19: error: field 'ifru_addr' has incomplete type
   struct sockaddr ifru_addr;
                   ^
/usr/include/linux/if.h:212:19: error: field 'ifru_dstaddr' has incomplete type
   struct sockaddr ifru_dstaddr;
                   ^
/usr/include/linux/if.h:213:19: error: field 'ifru_broadaddr' has incomplete type
   struct sockaddr ifru_broadaddr;
                   ^
/usr/include/linux/if.h:214:19: error: field 'ifru_netmask' has incomplete type
   struct sockaddr ifru_netmask;
                   ^
/usr/include/linux/if.h:215:20: error: field 'ifru_hwaddr' has incomplete type
   struct  sockaddr ifru_hwaddr;
                    ^
configure:18012: $? = 1



There were also two batches of complaints about mismatched types in printf:

  CC       libfdisk/src/libfdisk_la-alignment.lo
In file included from libfdisk/src/fdiskP.h:25:0,
                 from libfdisk/src/alignment.c:7:
libfdisk/src/alignment.c: In function 'fdisk_align_lba':
libfdisk/src/alignment.c:115:29: warning: format '%ju' expects argument of type 'uintmax_t', but argument 6 has type 'long unsigned int' [-Wformat=]
   DBG(CXT, ul_debugobj(cxt, "LBA %ju -aligned-%s-> %ju [grain=%jus]",
                             ^
./include/debug.h:63:4: note: in definition of macro '__UL_DBG'
    x; \
    ^
libfdisk/src/alignment.c:115:3: note: in expansion of macro 'DBG'
   DBG(CXT, ul_debugobj(cxt, "LBA %ju -aligned-%s-> %ju [grain=%jus]",
   ^


  CC       sys-utils/lsns-lsns.o
In file included from sys-utils/lsns.c:45:0:
sys-utils/lsns.c: In function 'add_namespace':
sys-utils/lsns.c:346:26: warning: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'ino_t {aka long long unsigned int}' [-Wformat=]
  DBG(NS, ul_debugobj(ns, "new %s[%lu]", ns_names[type], ino));
                          ^
./include/debug.h:63:4: note: in definition of macro '__UL_DBG'
    x; \
    ^
sys-utils/lsns.c:346:2: note: in expansion of macro 'DBG'
  DBG(NS, ul_debugobj(ns, "new %s[%lu]", ns_names[type], ino));
  ^
sys-utils/lsns.c: In function 'add_process_to_namespace':
sys-utils/lsns.c:362:26: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'ino_t {aka long long unsigned int}' [-Wformat=]
  DBG(NS, ul_debugobj(ns, "add process [%p] pid=%d to %s[%lu]", proc, proc->pid,
                          ^
./include/debug.h:63:4: note: in definition of macro '__UL_DBG'
    x; \
    ^
sys-utils/lsns.c:362:2: note: in expansion of macro 'DBG'
  DBG(NS, ul_debugobj(ns, "add process [%p] pid=%d to %s[%lu]", proc, proc->pid,
  ^
sys-utils/lsns.c: In function 'add_scols_line':
sys-utils/lsns.c:440:20: warning: format '%lu' expects argument of type 'long unsigned int', but argument 3 has type 'ino_t {aka long long unsigned int}' [-Wformat=]
    xasprintf(&str, "%lu", ns->id);
                    ^

Re: e2fsck: /dev/hda2 is mounted.

[George Spelvin]

I dug back through the commit log, and discovered that reverting
7a4f98854283a5b98e76902b12685ad1999b2802 
fixes the regression.

I assume reverting this isn't the preferred solution, but it
at least narrows down the issue.

Re: e2fsck: /dev/hda2 is mounted.

[Karel Zak]

On Sun, Mar 13, 2016 at 06:14:14AM -0400, George Spelvin wrote:
> 32305: libmount:      TAB: [0x983a160]: /proc/self/mountinfo: start parsing [entries=0, filter=not]
> 32305: libmount:      TAB: [0x983a160]: add entry: /dev/root /
> 32305: libmount:    CACHE: canonicalize path /proc/self/mountinfo
> 32305: libmount:      TAB: TID for /proc/self/mountinfo is 32305
> 32305: libmount:      TAB: [0x983a160]: root FS: 302
> 32305: libmount:    CACHE: [0x983a048]: canonicalize path 302
> 32305: libmount:    CACHE: [0x983a048]: add entry [ 3] (path): 302: 302
> 32305: libmount:      TAB: [0x983a160]: canonical root FS: 302
> 32305: libmount:      TAB: [0x983a160]: add entry: sysfs /sys
> 32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run
> 32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run/lock
> 32305: libmount:      TAB: [0x983a160]: add entry: proc /proc
> 32305: libmount:      TAB: [0x983a160]: add entry: devtmpfs /dev
> 32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /run/shm
> 32305: libmount:      TAB: [0x983a160]: add entry: devpts /dev/pts
> 32305: libmount:      TAB: [0x983a160]: add entry: tmpfs /tmp
> 32305: libmount:      TAB: [0x983a160]: add entry: /dev/hda3 /home
> 32305: libmount:       FS: [0x983ddc8]: free [refcount=0]
> 32305: libmount:      TAB: [0x983a160]: /proc/self/mountinfo: stop parsing (10 entries)
> 32305: libmount:      TAB: [0x983a160]: parsing done [filename=/proc/self/mountinfo, rc=0]
> 32305: libmount:      TAB: [0x983a160]: mtab parse: #2 read utab
> 32305: libmount:      TAB: [0x983a160]: lookup SOURCE: '/dev/hda2'
> 32305: libmount:      TAB: [0x983a160]: lookup SRCPATH: '/dev/hda2'
> 32305: libmount:      TAB: [0x983a160]: lookup canonical SRCPATH: '/dev/hda2'
> /dev/hda2 is not mounted

 Can you send me your /proc/self/mountinfo? (or at least root fs entry
 from the file)

 Thanks,
    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

Re: e2fsck: /dev/hda2 is mounted.

[George Spelvin]

Thanks for the help.  I don't quite understand the structure of libmount
well enough to know where a fix *should* go.

> Can you send me your /proc/self/mountinfo? (or at least root fs entry
> from the file)

12 0 3:2 / / rw,relatime - ext4 /dev/root rw,errors=remount-ro,data=ordered
13 12 0:12 / /sys rw,nosuid,nodev,noexec,relatime - sysfs sysfs rw
14 12 0:13 / /run rw,nosuid,noexec,relatime - tmpfs tmpfs rw,size=207448k,mode=755
15 14 0:14 / /run/lock rw,nosuid,nodev,noexec,relatime - tmpfs tmpfs rw,size=5120k
16 12 0:4 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
17 12 0:6 / /dev rw,nosuid,relatime - devtmpfs devtmpfs rw,size=10240k,nr_inodes=221580,mode=755
18 14 0:15 / /run/shm rw,nosuid,nodev,noexec,relatime - tmpfs tmpfs rw,size=1391640k
19 17 0:11 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=5,mode=620
20 12 0:16 / /tmp rw,noexec,relatime - tmpfs tmpfs rw,size=524288k
21 12 3:3 / /home rw,relatime - ext4 /dev/hda3 rw,data=ordered

And here's /etc/mtab -> /proc/mounts:

/dev/root / ext4 rw,relatime,errors=remount-ro,data=ordered 0 0                 
sysfs /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0                            
tmpfs /run tmpfs rw,nosuid,noexec,relatime,size=207448k,mode=755 0 0            tmpfs /run/lock tmpfs rw,nosuid,nodev,noexec,relatime,size=5120k 0 0            
proc /proc proc rw,nosuid,nodev,noexec,relatime 0 0                             
devtmpfs /dev devtmpfs rw,nosuid,relatime,size=10240k,nr_inodes=221580,mode=755 0 0                                                                             
tmpfs /run/shm tmpfs rw,nosuid,nodev,noexec,relatime,size=1391640k 0 0          
devpts /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620 0 0             
tmpfs /tmp tmpfs rw,noexec,relatime,size=524288k 0 0                            
/dev/hda3 /home ext4 rw,relatime,data=ordered 0 0             16,0-1        All

Re: e2fsck: /dev/hda2 is mounted.

[Ruediger Meier]

On Sunday 13 March 2016, George Spelvin wrote:

> There were also two batches of complaints about mismatched types in
> printf:
> [...]

These warnings are fixed already in e1a15f8f. They are indicating that 
you are using a 32bit system. Maybe that's important to reproduce the 
bug.

cu,
Rudi

Re: e2fsck: /dev/hda2 is mounted.

[Ruediger Meier]

On Monday 14 March 2016, George Spelvin wrote:
> Thanks for the help.  I don't quite understand the structure of
> libmount well enough to know where a fix *should* go.
>
> > Can you send me your /proc/self/mountinfo? (or at least root fs
> > entry from the file)

> 12 0 3:2 / / rw,relatime - ext4 /dev/root
> [...]

Maybe it's not related but I've noticed a similar issue where 
getmntent() was not able to parse long lines in /proc/mounts. It was 
with musl and I've asked there
http://permalink.gmane.org/gmane.linux.lib.musl.general/9548

I've got no reply, so maybe it was a stupid question and we are doing 
something wrong ;)

cu,
Rudi

Re: e2fsck: /dev/hda2 is mounted.

[Karel Zak]

On Tue, Mar 15, 2016 at 12:38:10AM +0100, Ruediger Meier wrote:
> On Monday 14 March 2016, George Spelvin wrote:
> > Thanks for the help.  I don't quite understand the structure of
> > libmount well enough to know where a fix *should* go.
> >
> > > Can you send me your /proc/self/mountinfo? (or at least root fs
> > > entry from the file)
> 
> > 12 0 3:2 / / rw,relatime - ext4 /dev/root
> > [...]
> 
> Maybe it's not related but I've noticed a similar issue where 
> getmntent() was not able to parse long lines in /proc/mounts. It was 
> with musl and I've asked there
> http://permalink.gmane.org/gmane.linux.lib.musl.general/9548
> 
> I've got no reply, so maybe it was a stupid question and we are doing 
> something wrong ;)

The problem in this case is /dev/root, libmount tries to convert it to
the real device name by root= command line option, but the conversion
is pretty poor and it does not support all relevant use-cases. 

I'll try to improve it. It also seems that in many cases (except
btrfs) we can follow maj:min from mountinfo entry rather than parse
the complicated root= option.

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

Re: e2fsck: /dev/hda2 is mounted.

[Karel Zak]

On Sun, Mar 13, 2016 at 06:14:14AM -0400, George Spelvin wrote:
> Upon investigation, it turned out that fsck didn't understand the
> kernel "root=302" parameter and was trying to run e2fsck on the root
> file system.  This produced an error, causing the boot process
> to drop into a recovery shell.
...
> Running the v2.28-rc1 release (8c97556ecfc0af358d35dc7a842aaadde5eeafd3)
> I got some compile warnings (appended for reference), but a
> working execurable which could be run as follows:

The problem should be fixed, git pull, make and re-test, please.

Thanks
    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

Re: e2fsck: /dev/hda2 is mounted.

[George Spelvin]

> The problem should be fixed, git pull, make and re-test, please.

After a bit of screwing around with Debian dependencies (the fix ends up
in the libmount package, not the util-linux package, and util-linux only
depends on libmount >= 2.25), I can confirm that it works.

Thank you very much!

LIBMOUNT_DEBUG=all LIBBLKID_DEBUG=all fsck -C -M -A -V

3411: libmount:     INIT: library debug mask: 0xffff
3411: libmount:     INIT: library version: 2.28.
3411: libmount:     INIT:     feature: selinux
3411: libmount:     INIT:     feature: btrfs
3411: libmount:     INIT:     feature: force-mountinfo
3411: libmount:     INIT:     feature: assert
3411: libmount:     INIT:     feature: debug
Available "LIBMOUNT_DEBUG=<name>[,...]|<mask>" debug masks:
   all      [0xffff] : info about all subsystems
   cache    [0x0004] : paths and tags cache
   cxt      [0x0200] : library context (handler)
   diff     [0x0400] : mountinfo changes tracking
   fs       [0x0040] : FS abstraction
   help     [0x0001] : this help
   locks    [0x0010] : mtab and utab locking
   loop     [0x2000] : loop devices routines
   options  [0x0008] : mount options parsing
   tab      [0x0020] : fstab, mtab, mounninfo routines
   update   [0x0080] : mtab, utab updates
   utils    [0x0100] : misc library utils
   monitor  [0x0800] : mount tables monitor
   btrfs    [0x1000] : btrfs specific routines
3411: libmount:    CACHE: [0x8d84048]: alloc
fsck from util-linux 2.28-rc1
3411: libmount:      TAB: [0x8d84068]: alloc
3411: libmount:      TAB: [0x8d84068]: /etc/fstab: start parsing [entries=0, filter=not]
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/hda1 none
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/hda2 /
3411: libmount:      TAB: [0x8d84068]: add entry: proc /proc
3411: libmount:      TAB: [0x8d84068]: add entry: tmpfs /tmp
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/fd0 /a
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/hdc /cd
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/hdd /dvd
3411: libmount:      TAB: [0x8d84068]: add entry: /dev/hda3 /home
3411: libmount:       FS: [0x8d849c8]: free [refcount=0]
3411: libmount:      TAB: [0x8d84068]: /etc/fstab: stop parsing (8 entries)
3411: libmount:      TAB: [0x8d84068]: parsing done [filename=/etc/fstab, rc=0]
3411: libmount:    CACHE: [0x8d84048]: canonicalize path /dev/hda2
3411: libmount:    CACHE: [0x8d84048]: add entry [ 1] (path): /dev/hda2: /dev/hda2
3411: libmount:    CACHE: [0x8d84048]: canonicalize path /dev/hda3
3411: libmount:    CACHE: [0x8d84048]: add entry [ 2] (path): /dev/hda3: /dev/hda3
Checking all file systems.
3411: libmount:      TAB: [0x8d84068]: lookup TARGET: '/'
3411: libmount:      TAB: [0x8d84160]: alloc
3411: libmount:      TAB: [0x8d84160]: mtab parse: ignore mtab
3411: libmount:      TAB: [0x8d84160]: mtab parse: #1 read mountinfo
3411: libmount:      TAB: [0x8d84160]: /proc/self/mountinfo: start parsing [entries=0, filter=not]
3411: libmount:      TAB: [0x8d84160]: add entry: /dev/root /
3411: libmount:    CACHE: canonicalize path /proc/self/mountinfo
3411: libmount:      TAB: TID for /proc/self/mountinfo is 3411
3411: libmount:    UTILS: guessing system root [devno 3:2]
3411: libmount:    UTILS:   devno converted to /dev/hda2
3411: libmount:      TAB: [0x8d84160]: canonical root FS: /dev/hda2
3411: libmount:      TAB: [0x8d84160]: add entry: sysfs /sys
3411: libmount:      TAB: [0x8d84160]: add entry: tmpfs /run
3411: libmount:      TAB: [0x8d84160]: add entry: tmpfs /run/lock
3411: libmount:      TAB: [0x8d84160]: add entry: proc /proc
3411: libmount:      TAB: [0x8d84160]: add entry: devtmpfs /dev
3411: libmount:      TAB: [0x8d84160]: add entry: tmpfs /run/shm
3411: libmount:      TAB: [0x8d84160]: add entry: devpts /dev/pts
3411: libmount:      TAB: [0x8d84160]: add entry: tmpfs /tmp
3411: libmount:      TAB: [0x8d84160]: add entry: /dev/hda3 /home
3411: libmount:       FS: [0x8d87dc8]: free [refcount=0]
3411: libmount:      TAB: [0x8d84160]: /proc/self/mountinfo: stop parsing (10 entries)
3411: libmount:      TAB: [0x8d84160]: parsing done [filename=/proc/self/mountinfo, rc=0]
3411: libmount:      TAB: [0x8d84160]: mtab parse: #2 read utab
3411: libmount:      TAB: [0x8d84160]: lookup SOURCE: '/dev/hda2'
3411: libmount:      TAB: [0x8d84160]: lookup SRCPATH: '/dev/hda2'
/dev/hda2 is mounted
3411: libmount:      TAB: [0x8d84160]: lookup SOURCE: '/dev/hda3'
3411: libmount:      TAB: [0x8d84160]: lookup SRCPATH: '/dev/hda3'
/dev/hda3 is mounted

Re: e2fsck: /dev/hda2 is mounted.

[George Spelvin]

> The problem in this case is /dev/root, libmount tries to convert it to
> the real device name by root= command line option, but the conversion
> is pretty poor and it does not support all relevant use-cases. 

Yes, that is a bit annoying.  It also breaks e4defrag.  I have a kludgey
script on my systems that parses /etc/fstab and creates a /dev/root
symlink, but of course there's no guarantee that fstab matches
the kernel root= command line.

If you feel like creating a makedevroot(8) utility (which creates a
symlink if it can, or a device node if it can't), I suspect it would
be appreciated.