В Вск, 05.05.2002, в 18:01, Ron Stodden написал:
> Borsenkow Andrej wrote:
> >
> > > A correctly working md5sun is _essential_ to the production of CD-RW
> > > disks.
> > >
> >
> > A correct diagnosis is _essential_ to fixing problem. It has nothing to
> > do with md5sum:
> >
> > {pts/1}% dd if=/dev/scd0 of=test.iso bs=1k
> > dd: ÞÔÅÎÉÅ `/dev/scd0': Input/output error
> > 54856+0 ×ÈÏÄÎÙÈ ÚÁÐÉÓÅÊ
> > 54856+0 ×ÙÈÏÄÎÙÈ ÚÁÐÉÓÅÊ
> >
> > dmesg ->
> >
> > I/O error: dev 0b:00, sector 109712
> > I/O error: dev 0b:00, sector 109712
> >
> > it seems that something evil happened with the way cdrom driver handles
> > EOF.
>
> For the first time on each CD, but not the second? Changing the CD
> resets it to the first time.
>
So what? Is it a proof of md5sum misbehaviour?
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10788+0 records in
10788+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10790+0 records in
10790+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10792+0 records in
10792+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10794+0 records in
10794+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10796+0 records in
10796+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10798+0 records in
10798+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10798+0 records in
10798+0 records out
{pts/0}% dd if=/dev/cdrom of=/dev/null bs=2048
dd: reading `/dev/cdrom': Input/output error
10798+0 records in
10798+0 records out
all of this without changing CD on a otherwise perfectly functioning CD.
As I already said, CD-R/W written in DAO are read correctly; CD-R/Ws
written in TAO show the above problem. I do not know why it reads it
correctly second time in your case; may be it is drive dependent.
> I gave my second diagnosis - the first was defective BURN-free
> firmware inside the drive that was stupidly attempting to read ahead
> after having read the last sector, this diagnosis 'confirmed' by going
> through 3 CD-RW Drives from different manufacturers courtesy of my
> friendly dealer. All exhibited the problem and were returned, but I
> did not try md5sum twice in these cases. There is no way to do this
> test under Windows since the md5sum Windows port is limited to a
> single file only.
>
> --
> Ron. [au]
>
