Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Debian 'Stable' not so stable.

1 view
Skip to first unread message

alex...@gmail.com

unread,
Nov 16, 2005, 5:08:32 PM11/16/05
to
I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
do `umount /cdrom`, a Segmentation Fault occurs (Don't tell me to
unmount first, then unplug. I know. But a segfault is always a
programmer's error) Needless to say, nothing of the kind ever happened
to me while using XP on the same hardware.

Peter T. Breuer

unread,
Nov 16, 2005, 5:16:14 PM11/16/05
to
In comp.os.linux.development.system alex...@gmail.com wrote:
> I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
> do `umount /cdrom`, a Segmentation Fault occurs

No it doesn't. You will likely get a kenrel oops, look in your log. You
have just wriecked the programming assumptions - it's like taking a
hammer to your hard disk while the


> (Don't tell me to
> unmount first, then unplug. I know. But a segfault is always a
> programmer's error

No it isn't. This is a userspace manifestaton of a kaput kernel. You have
an ex-kernel. Do not cut the wires to your hard disk!

> Needless to say, nothing of the kind ever happened
> to me while using XP on the same hardware.

Of course it did. There's no avoiding death after disappearing file
system media! If you are going to do that kind of thing you will need to
use the supermunt driver to manage the media for you, and present a
facsimile of the device to the kernel when it has been yanked.

Peter

Keith Keller

unread,
Nov 16, 2005, 5:21:15 PM11/16/05
to
On 2005-11-16, alex...@gmail.com <alex...@gmail.com> wrote:
> I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
> do `umount /cdrom`, a Segmentation Fault occurs (Don't tell me to
> unmount first, then unplug. I know. But a segfault is always a
> programmer's error)

Why don't you file a bug report with Debian and/or the maintainers of
mount? (Or automount, if you're using that for your cdrom; or hotplug,
if you're using that for your cdrom. But it doesn't sound like either
of those are involved.)

Oh, and umount first.

> Needless to say, nothing of the kind ever happened
> to me while using XP on the same hardware.

Needless to say, I don't care what XP does or doesn't do. A bug is
relative to what the software should be doing, not to what some other
piece of software does or doesn't do.

--keith

--
kkeller...@wombat.san-francisco.ca.us
(try just my userid to email me)
AOLSFAQ=http://wombat.san-francisco.ca.us/cgi-bin/fom
see X- headers for PGP signature information

alex...@gmail.com

unread,
Nov 16, 2005, 6:50:20 PM11/16/05
to

Peter T. Breuer wrote:
> In comp.os.linux.development.system alex...@gmail.com wrote:
> > I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
> > do `umount /cdrom`, a Segmentation Fault occurs
>
> No it doesn't. You will likely get a kenrel oops, look in your log.

You can't read? I'm telling you what DOES happen, not what will likely
happen. I don't need to look in dmesg.

$ umount /cdrom
Segmentation Fault
$

Pete Zaitcev (OTID3)

unread,
Nov 16, 2005, 9:15:41 PM11/16/05
to

LOL, Peter, this is what you get for answering trolls. What a moron,
and so uppity, too.

It has to be noted, however, that the word "stable" may mean different
things. It may mean "robust", but also "unchanging". In the later case,
bugs fixes developed long ago are not accepted in order not to jeopardize
the predictability of a stable version. In Fedora Core 4, unplugging USB
CD-ROM while mounted is perfectly ok, for example. However, I see no
value in incorporating disruptive bugfixes into Debian (especially
"stable").

-- Pete

Peter T. Breuer

unread,
Nov 17, 2005, 1:58:20 AM11/17/05
to
In comp.os.linux.development.system alex...@gmail.com wrote:

Read your kernel log.

(if you don't get an ops, then you have survived, and a segfault is a
good result from such a bizarre thing to do).

Peter

Peter T. Breuer

unread,
Nov 17, 2005, 2:04:37 AM11/17/05
to
In comp.os.linux.development.system "Pete Zaitcev (OTID3)" <oted3354a...@comcast.net> wrote:
> It has to be noted, however, that the word "stable" may mean different
> things. It may mean "robust", but also "unchanging". In the later case,
> bugs fixes developed long ago are not accepted in order not to jeopardize
> the predictability of a stable version. In Fedora Core 4, unplugging USB
> CD-ROM while mounted is perfectly ok, for example. However, I see no
> value in incorporating disruptive bugfixes into Debian (especially
> "stable").

You think there is a kernel bugfix here? Surely he CANNOT remove a cd
while it is mounted because the openers count would be positive and
hence the driver would not allow the "open cd door" ioctl. Therefore he
has forced it by "cutting the cables" (yanking them out). That's physical
destruction.

When the medium is destroyed in-situ like that, it's hard to know what to
do. I would imagine a kernel oops would follow. But if it doesn't, then
at least the kernel will error out all requests, and the file system
code would maybe oops (where'd my inode go)? If not, then maybe the
application layer will segfault. The latter seems to me to be a good
result, but I daresay the author of the app will prefer to catch the
segfault and print out "owwwww".


Peter

alex...@gmail.com

unread,
Nov 17, 2005, 3:17:42 AM11/17/05
to
Peter T. Breuer wrote:
> Surely he CANNOT remove a cd while it is mounted

Are you saying that I defied the laws of physics?

> Therefore he has forced it by "cutting the cables" (yanking them out). That's physical
> destruction.

Wow, what an incredible command of logic!

You are saying:

unplugging cables == cutting cables
cutting cables == physical destruction

Therefore, unplugging cables is physical destruction

Analogously,

sitting in a chair == sitting on a toilet
sitting on a toilet == defecation

Therefore sitting in a chair implies defecation.

Do you get it now? Equivalence of one aspect of objects does not imply
equivalence in other aspects. Unplugging a cable is not physical
destruction. Under XP, I plug and unplug this external DVD-RW all the
time. If Linux can't handle this, tough shit, scored one up for XP.

P.S. And by the way, please see a shrink, you just don't sound very
sane to me. Really.

alex...@gmail.com

unread,
Nov 17, 2005, 3:36:45 AM11/17/05
to
Peter T. Breuer wrote:
> In comp.os.linux.development.system alex...@gmail.com wrote:
>
> > Peter T. Breuer wrote:
> >> In comp.os.linux.development.system alex...@gmail.com wrote:
> >> > I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
> >> > do `umount /cdrom`, a Segmentation Fault occurs
> >>
> >> No it doesn't. You will likely get a kenrel oops, look in your log.
>
> > You can't read? I'm telling you what DOES happen, not what will likely
> > happen. I don't need to look in dmesg.
>
> Read your kernel log.
>
> (if you don't get an ops, then you have survived, and a segfault is a
> good result from such a bizarre thing to do).

You said segmentation fault can not occur. But it does. Therefore, you
are wrong. What's in the kernel log will not change that. ¿Comprende?

Peter T. Breuer

unread,
Nov 17, 2005, 3:38:06 AM11/17/05
to
In comp.os.linux.misc alex...@gmail.com wrote:
> Peter T. Breuer wrote:
>> Surely he CANNOT remove a cd while it is mounted

> Are you saying that I defied the laws of physics?

Physics? The laws of O/Ss, yes. You had to physically do something that
is forbidden via the provided interfaces, such as knocking the wall down
with a machanical ball and chain instead of walking in through the door.

>> Therefore he has forced it by "cutting the cables" (yanking them out). That's physical
>> destruction.

> Wow, what an incredible command of logic!

Good.

> unplugging cables == cutting cables
> cutting cables == physical destruction

Yep.

> Therefore, unplugging cables is physical destruction

Yep.

> Analogously,

Nope.

> sitting in a chair == sitting on a toilet

It isn't. Sitting in a chair is the same thing as loading a chair,
however.

> sitting on a toilet == defecation

It doesn't. Loading a toilet is sitting on it.

> Therefore sitting in a chair implies defecation.

No it doesn't.

> Do you get it now? Equivalence of one aspect of objects does not imply
> equivalence in other aspects.

Of course not. I provided proper abstractions both times: unplugging is
an instance of cutting the connection, and cutting the connection is an
instance of destroying the connection, therefore you destroyed the
connection (wires).

> Unplugging a cable is not physical
> destruction.

Of course it is. In what way is it different from cutting them?

Would you mind yanking the cables from your TV a few times?


> Under XP, I plug and unplug this external DVD-RW all the
> time.

CDs are not intrinsically hot-pluggable and neither are hard disks.
That you can physically destroy the wires doesn't excuse you when you do
it. But I must add that you are mistaking what happens - it likely IS
the case that the device driver copes OK with the disappearing device,
but you probably need a FS emulation driver above the device driver if
you are going to be doing that sort of thing in order to keep the fs
code happy with mutating bytes on a readonly medium - use the correct
mount options. You get the same results as yanking a floppy the way you
are ging about it.

> If Linux can't handle this, tough shit, scored one up for XP.

It handles it fine.

> P.S. And by the way, please see a shrink, you just don't sound very
> sane to me. Really.

Keep your foul-mouthed insults to yourself.


Peter

Peter T. Breuer

unread,
Nov 17, 2005, 3:54:26 AM11/17/05
to
In comp.os.linux.misc alex...@gmail.com wrote:
> Peter T. Breuer wrote:
>> In comp.os.linux.development.system alex...@gmail.com wrote:
>> > Peter T. Breuer wrote:
>> >> In comp.os.linux.development.system alex...@gmail.com wrote:
>> >> > I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
>> >> > do `umount /cdrom`, a Segmentation Fault occurs
>> >> No it doesn't. You will likely get a kenrel oops, look in your log.
>> > You can't read? I'm telling you what DOES happen, not what will likely
>> > happen. I don't need to look in dmesg.
>> Read your kernel log.
>>
>> (if you don't get an oops, then you have survived, and a segfault is a

>> good result from such a bizarre thing to do).

> You said segmentation fault can not occur. But it does. Therefore, you
> are wrong.

I cannot be "wrong" about something only you know - it is not in the
domain of discourse. Any time I say "you have chosen the number 7", you
can say "no - you're wrong". I am telling you what has PROBABLY happened,
and if you go and look where I advised, you will be able to tell us
whether that is the case or not, and thus bring something more to the
discussion.

What's in the kernel log will not change that. _Comprende?

Cease insulting behaviour, NOW. If you are interested in learning what is
going on, start again, and make it evident.

Peter

alex...@gmail.com

unread,
Nov 17, 2005, 4:05:08 AM11/17/05
to
Peter T. Breuer wrote:
> In comp.os.linux.misc alex...@gmail.com wrote:
> > Peter T. Breuer wrote:
> >> In comp.os.linux.development.system alex...@gmail.com wrote:
> >> > Peter T. Breuer wrote:
> >> >> In comp.os.linux.development.system alex...@gmail.com wrote:
> >> >> > I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
> >> >> > do `umount /cdrom`, a Segmentation Fault occurs
> >> >> No it doesn't. You will likely get a kenrel oops, look in your log.
> >> > You can't read? I'm telling you what DOES happen, not what will likely
> >> > happen. I don't need to look in dmesg.
> >> Read your kernel log.
> >>
> >> (if you don't get an oops, then you have survived, and a segfault is a
> >> good result from such a bizarre thing to do).
>
> > You said segmentation fault can not occur. But it does. Therefore, you
> > are wrong.
>
> I cannot be "wrong" about something only you know - it is not in the
> domain of discourse.

OMFG. Read the quote above.

I wrote "a Segmentation Fault occurs" and you replied "No, it doesn't".

Now, you are saying that you couldn't have been wrong, BECAUSE you were
talking about something you could NOT know?!

Did they put DSL in the ward?

I'm out of this thread. I want to keep *MY* sanity.

Peter T. Breuer

unread,
Nov 17, 2005, 4:28:15 AM11/17/05
to
In comp.os.linux.misc alex...@gmail.com wrote:
> Peter T. Breuer wrote:
>> In comp.os.linux.misc alex...@gmail.com wrote:
>> > Peter T. Breuer wrote:
>> >> In comp.os.linux.development.system alex...@gmail.com wrote:
>> >> > Peter T. Breuer wrote:
>> >> >> In comp.os.linux.development.system alex...@gmail.com wrote:
>> >> >> > I have a USB DVD-RW. When I unplug it while /cdrom is mounted, and then
>> >> >> > do `umount /cdrom`, a Segmentation Fault occurs
>> >> >> No it doesn't. You will likely get a kenrel oops, look in your log.
>> >> > You can't read? I'm telling you what DOES happen, not what will likely
>> >> > happen. I don't need to look in dmesg.
>> >> Read your kernel log.
>> >>
>> >> (if you don't get an oops, then you have survived, and a segfault is a
>> >> good result from such a bizarre thing to do).
>>
>> > You said segmentation fault can not occur. But it does. Therefore, you
>> > are wrong.
>>
>> I cannot be "wrong" about something only you know - it is not in the
>> domain of discourse.

> OMFG. Read the quote above.

Cease rudeness.

> I wrote "a Segmentation Fault occurs" and you replied "No, it doesn't".

That's my opinion. Check. I told you how.

Are you interested in conversation, or is insults where you get off the
bus at?

Peter

Kasper Dupont

unread,
Nov 17, 2005, 5:32:30 AM11/17/05
to
alex...@gmail.com wrote:
>
> You can't read? I'm telling you what DOES happen, not what will likely
> happen. I don't need to look in dmesg.

If you just want to complain you have come to the wrong group.

If you actually want a problem solved, then for .... sake tell
us what the error messages in the log says. And also tell us
what kernel version you are using (or even better reproduce
the problem with the newest kernel and then tell us what it
says).

--
Kasper Dupont
Note to self: Don't try to allocate
256000 pages with GFP_KERNEL on x86.

Rich Walker

unread,
Nov 17, 2005, 12:04:56 PM11/17/05
to
"Peter T. Breuer" <p...@oboe.it.uc3m.es> writes:

> In comp.os.linux.misc alex...@gmail.com wrote:
>
>> Under XP, I plug and unplug this external DVD-RW all the
>> time.
>
> CDs are not intrinsically hot-pluggable and neither are hard disks.


*BUT HE HAS AN EXTERNAL USB ONE*

Users expect USB devices to be hot-pluggable.

This is a corner case, and it causes umount to fail.

sheesh.

cheers, Rich.

--
rich walker | Shadow Robot Company | r...@shadow.org.uk
technical director 251 Liverpool Road |
need a Hand? London N1 1LX | +UK 20 7700 2487
www.shadow.org.uk/products/newhand.shtml

Peter T. Breuer

unread,
Nov 17, 2005, 12:16:31 PM11/17/05
to
In comp.os.linux.development.system Rich Walker <r...@shadow.org.uk> wrote:
> "Peter T. Breuer" <p...@oboe.it.uc3m.es> writes:
>> In comp.os.linux.misc alex...@gmail.com wrote:
>>> Under XP, I plug and unplug this external DVD-RW all the
>>> time.
>> CDs are not intrinsically hot-pluggable and neither are hard disks.

> *BUT HE HAS AN EXTERNAL USB ONE*

> Users expect USB devices to be hot-pluggable.

But not hot-unpluggable, surely. It's one thing to say "you can put the
clothes in the washing machine at any time", but quite another to say
"you can take the clothes out of the washing machine at any time".

> This is a corner case, and it causes umount to fail.

Well, we know what is going on - it's not precisely that umount will get
stuck (it probably will) but that the cached FS structure now points at
places on the CD that don't exist. The only cure is to cache it all when
you load the cd (owww) or tell your fs driver to expect that anything and
everything may evaporate at any moment - rather difficult since the fs
driver is written supposing that the structure of the fs is GOOD.

Mind you, we don't know if he had a kernel oops or not.

Another possibility is to use a supermount layer over the driver and
under the mount in order to handle the disappearing device.

Peter

Bill Marcum

unread,
Nov 17, 2005, 1:13:04 PM11/17/05
to
["Followup-To:" header set to comp.os.linux.development.system.]
On 16 Nov 2005 14:08:32 -0800, alex...@gmail.com
Unplugging a mounted drive is always a user error. Are you complaining
because you got a segfault instead of a blue screen of death?


--
A platitude is simply a truth repeated till people get tired of hearing it.
-- Stanley Baldwin

M.Kmann

unread,
Nov 17, 2005, 4:51:09 PM11/17/05
to

if you want it the m$ way, just mount it with the sync option.
if you want to use win, then use it.

0 new messages