linux user group brescia

Alle 16:26, giovedì 30 ottobre 2003, Giuseppe Corbelli ha scritto:
> On Wed, Oct 29, 2003 at 07:15:25PM +0100, Luca Coianiz wrote:
> > >Io eviterei mandrake: per la pessima organizzazione, per la pesantezza,
> > >e SOPRATTUTTO per i casini legali in cui naviga. Dai poi un'occhiata a
> > >uno degli ultimi messaggi sulla ml. Si parla di un simpatico giochetto
> > >che fa con i lettori cd. Non ho ben capito cosa faccia ma è meglio non
> > >fidarsi!!!
> >
> >  Mi scuso in anticipo per la poca precisione nel riportare la notizia
> > ma... in questo momento non ho molta voglia di spulciarmi tutti i thread
> > del forum utenti di P.I. ;-)
> >
> >  In ogni caso, leggevo che LG pare abbia utilizzato un comando standard,
> > probabilmente poco utilizzato... non so (flush_buffer), riscrivendolo in
> > modo tale da effettuare l'upgrade del firmware di alcuni suoi lettori CD.
>
> Presumo si tratti di syncronize cache (35h), descritto nello standard MMC.
> Mentre per il caricamento di nuovo firmware c'e' write buffer (3Bh). Non
> voglio pensare che si sia trattato di un typo o di una svista. Anche
> perche' non c'e' ragione per cui l'installer di una distro debba usare uno
> dei 2 cmd. In ogni caso c'e' da insultare l'LG che fa queste cagate, non
> MDK.



Come leggo da:http://forum.firmware-flash.com/search.php?search_author=mac_es,
credo che sia poichè il kernel di Mandrake utilizza la packet-writing patch 
che introduce il comando GPCMD_FLUSH_CACHE in ide-cd.c

da ide-cd.c patchato:

/*
 * Close down the device.  Invalidate all cached blocks.
 */

static
void ide_cdrom_release_real (struct cdrom_device_info *cdi)
{
  struct cdrom_generic_command cgc;

  init_cdrom_command(&cgc, NULL, 0, CGC_DATA_NONE);
  cgc.cmd[0] = GPCMD_FLUSH_CACHE;
  cgc.quiet = 1;
  (void) ide_cdrom_packet(cdi, &cgc);
}

(controllare chi ha Mandrake se esiste il programma pktsetup: allora la 
packet-writing patch è sicura)

difatti dalla packet-writing mailing-list:

On Thu, 30 Oct 2003, Ljubomir Bozic Jr. wrote:

> On Monday 27 October 2003 18:43, Peter Osterlund wrote:
> 
> > - Removed unnecessary FLUSH_CACHE command from ide-cd.c. It is not
> >   necessary, and some broken LG CD-ROM units get damaged by the
> >   FLUSH_CACHE command. (Apparently they interpret the command as a
> >   firmware upgrade command.)
> >
> 
> I would like to ask you if you could explain FLUSH_CACHE a little bit more.
> What exactly is the FLUSH_CACHE function and for what purpose was it used by 
> You/Mandrake?
> Why is it not necessary to have it any more?

The purpose of the FLUSH_CACHE command is to make sure all data has been
written to the disc before the user has a chance to eject the disc. This
is necessary when writing to the disc but doesn't do anything useful if
the disc is only being read.

The old code issued FLUSH_CACHE commands from two places. It did it in the
pktcdvd module when closing the device if it was opened for writing. It
also did it from the generic cdrom code (ide-cd.c), but this generic code
did the flush even if the device was only opened for reading.

The new code simply drops the FLUSH_CACHE command from ide-cd.c, which is
ok because the flush in pktcdvd is still present.

-- 
Peter Osterlund - petero2 a telia.com
http://w1.894.telia.com/~u89404340

e dalla Mandrake mailing-list:


jos> True, strange that this bug was able to tunnel trough all RCs,
jos> but this once again proves that it's better to use the entire
jos> linux community as testers instead of a few beta testers,
jos> i.e. use only stock kernels.

That is good if the functionality of stock kernel is not ok for you.
Examples:

....
- You think that having a CD-RW being able to do more things that only
  copy CD's is a good idea (packet writing).

http://marc.theaimsgroup.com/?l=mandrake-cooker&m=106727100026364&w=2


(che dà un buon indizio)

Carlo