Access beyond end of device

Access beyond end of device

[Luiz Fernando N. Capitulino]

 Hello,

 Since I've started using the pata_via driver in 2.6.22, I'm getting
zillions of these messages:

"""
attempt to access beyond end of device
sda: rw=0, want=156367809, limit=156365903
printk: 22 messages suppressed
Buffer I/O error on device sda1, logical block 78183872
...
"""

 It also does happen with 2.6.23-rc3.

 Controller:

"""
00:0f.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06) (prog-if 8a [Master SecP PriP])
"""

-- 
Luiz Fernando N. Capitulino

Re: Access beyond end of device

[Luiz Fernando N. Capitulino]

Em Tue, 31 Jul 2007 15:04:14 -0300
"Luiz Fernando N. Capitulino" <lcapitulino@mandriva.com.br> escreveu:

|  Since I've started using the pata_via driver in 2.6.22, I'm getting
| zillions of these messages:
| 
| """
| attempt to access beyond end of device
| sda: rw=0, want=156367809, limit=156365903
| printk: 22 messages suppressed
| Buffer I/O error on device sda1, logical block 78183872
| ...
| """
| 
|  It also does happen with 2.6.23-rc3.

 Heh, I'm pretty advanced, I meant 2.6.23-rc1.

-- 
Luiz Fernando N. Capitulino

Re: Access beyond end of device

[Alan Cox]

On Tue, 31 Jul 2007 15:15:46 -0300
"Luiz Fernando N. Capitulino" <lcapitulino@mandriva.com.br> wrote:

> Em Tue, 31 Jul 2007 15:04:14 -0300
> "Luiz Fernando N. Capitulino" <lcapitulino@mandriva.com.br> escreveu:
> 
> |  Since I've started using the pata_via driver in 2.6.22, I'm getting
> | zillions of these messages:
> | 
> | """
> | attempt to access beyond end of device
> | sda: rw=0, want=156367809, limit=156365903
> | printk: 22 messages suppressed
> | Buffer I/O error on device sda1, logical block 78183872

Does your disk have a host protected area set and if so are you telling
libata to disable it ?

Re: Access beyond end of device

[Luiz Fernando N. Capitulino]

Em Wed, 1 Aug 2007 01:12:26 +0100
Alan Cox <alan@lxorguk.ukuu.org.uk> escreveu:

| On Tue, 31 Jul 2007 15:15:46 -0300
| "Luiz Fernando N. Capitulino" <lcapitulino@mandriva.com.br> wrote:
| 
| > Em Tue, 31 Jul 2007 15:04:14 -0300
| > "Luiz Fernando N. Capitulino" <lcapitulino@mandriva.com.br> escreveu:
| > 
| > |  Since I've started using the pata_via driver in 2.6.22, I'm getting
| > | zillions of these messages:
| > | 
| > | """
| > | attempt to access beyond end of device
| > | sda: rw=0, want=156367809, limit=156365903
| > | printk: 22 messages suppressed
| > | Buffer I/O error on device sda1, logical block 78183872
| 
| Does your disk have a host protected area set and if so are you telling
| libata to disable it ?

 Ok, the disk has a host protected area and now I'm telling libata to
disable it.

 The error messages has disappeared.

 But... Is it safe to do it? I've read somewhere that the BIOS and
other softwares may use that area.

 Also, I remember that with the IDE driver I wasn't seeing that
messages, which means either: it was disabling it by default or
was just ignoring it.

-- 
Luiz Fernando N. Capitulino

Re: Access beyond end of device

[Alan Cox]

>  Ok, the disk has a host protected area and now I'm telling libata to
> disable it.
> 
>  The error messages has disappeared.
> 
>  But... Is it safe to do it? I've read somewhere that the BIOS and
> other softwares may use that area.

On some systems. Its also used for different things on others (rounding
off disk sizes nicely, clipping to avoid bios limits etc)


>  Also, I remember that with the IDE driver I wasn't seeing that
> messages, which means either: it was disabling it by default or
> was just ignoring it.

Old IDE disables it by default, libata is more conservative. Personally I
think libata should disable it by default too but thats an argument to
have with Jeff

access beyond end of device

[Kevin]

I checked the mailing list archive, and I didn't see any solutions.
It also seemed the the similar problems happened after a crash, which
is not so in my case.  Recently I was trying to extract a file and I
got read errors.  I checked dmesg and found the following repeated:

  attempt to access beyond end of device
  38:01: rw=0, want=788705992, limit=58633312
  attempt to access beyond end of device
  38:01: rw=0, want=788705992, limit=58633312

Running reiserfsck --check gave me a bunch of things similar to this:
bad_indirect_item: block 16371: item 2014 2043 0x1 IND (1), len 1276,
location 96 entry count 0, fsck need 0, format new has a pointer 131
to the block 9437328 which is in tree already
9437328 which is in tree already

It also found one broken file, and said that there was 1 corruption
that required --rebuild-tree.  I haven't done this yet because I'm a
little worried it might die in the middle of the rebuild.  Anyone have
any idea what I should do?

I am running 2.4.18, and reiserfsprogs 3.x.1c-pre4.

Re: access beyond end of device

[Yury Umanets]

Kevin wrote:

>I checked the mailing list archive, and I didn't see any solutions.
>It also seemed the the similar problems happened after a crash, which
>is not so in my case.  Recently I was trying to extract a file and I
>got read errors.  I checked dmesg and found the following repeated:
>
>  attempt to access beyond end of device
>  38:01: rw=0, want=788705992, limit=58633312
>  attempt to access beyond end of device
>  38:01: rw=0, want=788705992, limit=58633312
>
>Running reiserfsck --check gave me a bunch of things similar to this:
>bad_indirect_item: block 16371: item 2014 2043 0x1 IND (1), len 1276,
>location 96 entry count 0, fsck need 0, format new has a pointer 131
>to the block 9437328 which is in tree already
>9437328 which is in tree already
>
>It also found one broken file, and said that there was 1 corruption
>that required --rebuild-tree.  I haven't done this yet because I'm a
>little worried it might die in the middle of the rebuild.  Anyone have
>any idea what I should do?
>
>I am running 2.4.18, and reiserfsprogs 3.x.1c-pre4.
>
>
>  
>
Anyway you may backup your data and try to rebuild tree by fsck.

>  
>

Re: access beyond end of device

[Kevin]

I'll give it a shot.  So far I've only seen that one file generate the
error, so hopefully I can just tiptoe over it and back everything else
up.  I'll let you know how it goes.  Thanks.

> Anyway you may backup your data and try to rebuild tree by fsck.

Re: access beyond end of device

[Hans Reiser]

Kevin wrote:

>I'll give it a shot.  So far I've only seen that one file generate the
>error, so hopefully I can just tiptoe over it and back everything else
>up.  I'll let you know how it goes.  Thanks.
>
>  
>
>>Anyway you may backup your data and try to rebuild tree by fsck.
>>    
>>
>
>
>
>  
>
You might want to backup using ddrescue on the device.....

-- 
Hans

Re: access beyond end of device

[Kevin]

> I'll give it a shot.  So far I've only seen that one file generate the
> error, so hopefully I can just tiptoe over it and back everything else
> up.  I'll let you know how it goes.  Thanks.

>> Anyway you may backup your data and try to rebuild tree by fsck.

I guess it worked?  I no longer got the error when trying to access
the original file, however it seems that part of the file got lopped
off, as the md5sum changed.  Everything else seems be working alright
though.  Thanks again.
---

reiserfsck --rebuild-tree started at Sat Jun 15 14:27:14 2002
###########

Pass 0:
Loading on-disk bitmap .. ok, 8092548 blocks marked used
Skipping 8658 blocks (super block, journal, bitmaps) 8083890 blocks will be read
0%....20%....40%....60%....80%....100%                         left 0, 579 /sec
        "r5" hash is selected
Flushing..done
        Read blocks (but not data blocks) 8083890
                Leaves among those 8102
                        - corrected leaves 1
                pointers in indirect items to wrong area 36 (zeroed)
                Objectids found 2066

Pass 1 (will try to insert 8102 leaves):
Looking for allocable blocks .. ok
0%....20%....40%....60%....80%....100%                         left 0, 352 /sec
Flushing..done
        8102 leaves read
                8102 inserted
        non-unique pointers in indirect items (zeroed) 36
Flushing..done
Pass 3 (semantic):
Flushing..done                                                                 
        Files found: 1916
        Directories found: 150
        Broken (of files/symlinks/others): 1
Pass 3a (looking for lost dir/files):
Looking for lost directories:
Flushing..done done 1, 0 /sec
Pass 4 - done           done 0, 0 /sec
        Deleted unreachable items 4
Flushing..done
Syncing..done
###########
reiserfsck finished at Sat Jun 15 18:20:39 2002