Re: mke2fs /dev/loop0

Re: mke2fs /dev/loop0

[Arnaldo Carvalho de Melo]

Em Wed, Feb 28, 2001 at 04:07:03PM +0100, Holluby István holluby@interware.hu escreveu:
> 	Hi!
> 
> This command hangs my system. It works for a 100K file, but it hangs my
> system, if the file is 470M. It does not matter, if the disk is SCSI or
> ide.
> 
> linux 2.4.2
> glibc-2.2.2
> gcc-2.95.2.1
> e2fs-1.19
> 
> With kernel 2.2.18, it works fine.

FAQ, try 2.4.2-acLATEST (now its ac6)
 
> ===============
> I also have some problem, with ncpfs. I am not quite sure, because I had to
> hack the source, to compile it, but the same hack works with 2.2.18.
> 
> If you have anything to tell, please mail me. I am not on the list.

I'm interested if this is directly related to IPX, Petr is the guy for
NCPfs, can you please send us more details about this problem? Hangs? Data
corruption? what?

- Arnaldo

mke2fs /dev/loop0

[Holluby]

	Hi!

This command hangs my system. It works for a 100K file, but it hangs my
system, if the file is 470M. It does not matter, if the disk is SCSI or
ide.

linux 2.4.2
glibc-2.2.2
gcc-2.95.2.1
e2fs-1.19

With kernel 2.2.18, it works fine.

===============
I also have some problem, with ncpfs. I am not quite sure, because I had to
hack the source, to compile it, but the same hack works with 2.2.18.


If you have anything to tell, please mail me. I am not on the list.



Holluby Istvan Budapest
(holluby@interware.hu)

Re: mke2fs /dev/loop0

[Richard B. Johnson]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=US-ASCII, Size: 1190 bytes --]

On Wed, 28 Feb 2001, Holluby [Windows-1250] István holluby@interware.hu wrote:

> 	Hi!
> 
> This command hangs my system. It works for a 100K file, but it hangs my
> system, if the file is 470M. It does not matter, if the disk is SCSI or
> ide.
> 
> linux 2.4.2
> glibc-2.2.2
> gcc-2.95.2.1
> e2fs-1.19
> 
> With kernel 2.2.18, it works fine.
> 
> ===============
> I also have some problem, with ncpfs. I am not quite sure, because I had to
> hack the source, to compile it, but the same hack works with 2.2.18.
> 

`mke2fs /dev/loop0`  requires an additional parameter (file size to
create). Otherwise, it will try to use all the RAM in your system, plus...

If it worked before, it was because of luck. FYI, this is not the
way to create a ramdisk. Normally you use the loop device to mount
a file as a file-system, i.e., `mount -o loop filename /mnt`.
So, I don't know what you are trying to do except crash your system.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.

Re: mke2fs /dev/loop0

[Jens Axboe]

On Wed, Feb 28 2001, Richard B. Johnson wrote:
> `mke2fs /dev/loop0`  requires an additional parameter (file size to
> create). Otherwise, it will try to use all the RAM in your system, plus...
> 
> If it worked before, it was because of luck. FYI, this is not the
> way to create a ramdisk. Normally you use the loop device to mount
> a file as a file-system, i.e., `mount -o loop filename /mnt`.
> So, I don't know what you are trying to do except crash your system.

This could not be more wrong. mke2fs will query the size of the
loop device, and make the correct size file system regardless
of whether it's file or block device backed. And it will not
try to use all RAM in the system?! This is loop, not a ramdisk.
Dirty buffers will be flushed to loop like any other block
device in the system, if that doesn't work then we have a mm
bug.

The previous answer was right -- loop has been broken for quite
some time, but use -ac latest on top of 2.4.2 and it should work
flawlessly for you.

-- 
Jens Axboe

Re: mke2fs /dev/loop0

[Richard B. Johnson]

On Wed, 28 Feb 2001, Jens Axboe wrote:

> On Wed, Feb 28 2001, Richard B. Johnson wrote:
> > `mke2fs /dev/loop0`  requires an additional parameter (file size to
> > create). Otherwise, it will try to use all the RAM in your system, plus...
> > 
> > If it worked before, it was because of luck. FYI, this is not the
> > way to create a ramdisk. Normally you use the loop device to mount
> > a file as a file-system, i.e., `mount -o loop filename /mnt`.
> > So, I don't know what you are trying to do except crash your system.
> 
> This could not be more wrong. mke2fs will query the size of the
> loop device, and make the correct size file system regardless
> of whether it's file or block device backed. And it will not
> try to use all RAM in the system?! This is loop, not a ramdisk.
> Dirty buffers will be flushed to loop like any other block
> device in the system, if that doesn't work then we have a mm
> bug.
> 
> The previous answer was right -- loop has been broken for quite
> some time, but use -ac latest on top of 2.4.2 and it should work
> flawlessly for you.
> 

Wrong. The report showed a response to a command. Nothing was reported
to have been mounted through the loop device. Instead, the raw command
`mke2fs /dev/loop0` was reported. Performing such a command on early
linux versions resulted in this:

Script started on Wed Feb 28 11:10:11 2001
mke2fs /dev/loop0
mke2fs 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09
mke2fs: Device size reported to be zero.  Invalid partition specified, or
	partition table wasn't reread after running fdisk, due to
	a modified partition being busy and in use.  You may need to reboot
	to re-read your partition table.

# exit
exit

Script done on Wed Feb 28 11:10:33 2001

This prevented you from killing the system during such a dumb command.

Later versions, resulted in the system halting with no activity. Again,
a pretty good result for such a dumb command.


Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.

Re: mke2fs /dev/loop0

[Jens Axboe]

On Wed, Feb 28 2001, Richard B. Johnson wrote:
> Wrong. The report showed a response to a command. Nothing was reported
> to have been mounted through the loop device. Instead, the raw command
> `mke2fs /dev/loop0` was reported. Performing such a command on early
> linux versions resulted in this:
> 
> Script started on Wed Feb 28 11:10:11 2001
> mke2fs /dev/loop0
> mke2fs 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09
> mke2fs: Device size reported to be zero.  Invalid partition specified, or
> 	partition table wasn't reread after running fdisk, due to
> 	a modified partition being busy and in use.  You may need to reboot
> 	to re-read your partition table.
> 
> # exit
> exit

This was the report:

	[on mke2fs]
	This command hangs my system. It works for a 100K file, but it
	hangs my system, if the file is 470M. It does not matter, if the
	disk is SCSI or ide.

so you pretty much have to assume that he's losetup the files first.
How else could he state that it works for a 100k file, but not for
a 470m one?

BLKGETSIZE will return the device size, nothing unexpected here.

-- 
Jens Axboe