more info on bugs

more info on bugs

[Yoshinori K. Okuji]

This is mainly to Marco. I'm looking at the "Known Bugs" on the wiki, and I 
would like to know more info on some bugs:

  * JFS support does not work on big filesystems.

Is this about the partition size or the block size? To reproduce this, how big 
size do I have to use?

  *  Large files do not work on ext2, this can be tested with loopback and a 
hd image.

How large?

  *  Some filesystems have regressions, only ext2 has been tested a lot...

I do not understand this. Is this a bug?

Okuji

Re: more info on bugs

[Marco Gerards]

"Yoshinori K. Okuji" <okuji@enbug.org> writes:

> This is mainly to Marco. I'm looking at the "Known Bugs" on the wiki, and I 
> would like to know more info on some bugs:
>
>   * JFS support does not work on big filesystems.
>
> Is this about the partition size or the block size? To reproduce this, how big 
> size do I have to use?

It was reported to me, but I was not able to reproduce this yet.  If I
can get a bigger harddisk I will try to fix this bug.

>   *  Large files do not work on ext2, this can be tested with loopback and a 
> hd image.
>
> How large?

A few 100MBs.

>   *  Some filesystems have regressions, only ext2 has been tested a lot...
>
> I do not understand this. Is this a bug?

Yes, but not a specific one.  For example there were some bugs in some
misc.c functions which are used by the filesystems.  They relied on
broken behavior so there might be some bugs in the other (non-ext2)
filesystems.  I should do some extensive testing.

--
Marco

Re: more info on bugs

[Yoshinori K. Okuji]

On Monday 18 July 2005 22:45, Marco Gerards wrote:
> It was reported to me, but I was not able to reproduce this yet.  If I
> can get a bigger harddisk I will try to fix this bug.

OK.

> A few 100MBs.

I tested this, and I got "ext2fs doesn't support tripple indirect blocks". Is 
this what you mean? How difficult is it to fix this bug? Isn't it enough to 
just add one more read?

> Yes, but not a specific one.  For example there were some bugs in some
> misc.c functions which are used by the filesystems.  They relied on
> broken behavior so there might be some bugs in the other (non-ext2)
> filesystems.  I should do some extensive testing.

If you can give me more specific information, I can take a look.

Okuji

Re: more info on bugs

[Marco Gerards]

"Yoshinori K. Okuji" <okuji@enbug.org> writes:

> On Monday 18 July 2005 22:45, Marco Gerards wrote:
>> It was reported to me, but I was not able to reproduce this yet.  If I
>> can get a bigger harddisk I will try to fix this bug.
>
> OK.
>
>> A few 100MBs.
>
> I tested this, and I got "ext2fs doesn't support tripple indirect blocks". Is 
> this what you mean? How difficult is it to fix this bug? Isn't it enough to 
> just add one more read?

No, that was not my problem.  I will try to reproduce the bug and send
in a test for this.


>> Yes, but not a specific one.  For example there were some bugs in some
>> misc.c functions which are used by the filesystems.  They relied on
>> broken behavior so there might be some bugs in the other (non-ext2)
>> filesystems.  I should do some extensive testing.
>
> If you can give me more specific information, I can take a look.

Sorry, I do not remember it correctly.  But I will do some testing
later this week.  So either more information or a patch will follow.

--
Marco