Re: ISO9660 4GB directory structures boundary limit and growisofs

From: Paul B Mahol <onemda_at_gmail.com>
Date: Tue, 27 Apr 2010 21:18:56 +0000
On 4/19/10, Paul B Mahol <onemda_at_gmail.com> wrote:
> On 4/19/10, Tom Evans <tevans.uk_at_googlemail.com> wrote:
>> On Mon, Apr 19, 2010 at 1:48 PM, Paul B Mahol <onemda_at_gmail.com> wrote:
>>> On 4/17/10, Paul B Mahol <onemda_at_gmail.com> wrote:
>>>> On Sat, Apr 17, 2010 at 4:31 PM, Tim Kientzle <kientzle_at_freebsd.org>
>>>> wrote:
>>>>> Paul B Mahol wrote:
>>>>>>
>>>>>> It is apparently not possible to make use of -use-the-force-luke=4gms
>>>>>> on FreeBSD when appending new session after 4GB. Mounted disk
>>>>>> afterwards  show nothing.
>>>>>>
>>>>>> Should we allow it like linux does?
>>>>>
>>>>> Are you claiming there is a problem when FreeBSD reads such
>>>>> images or a problem with creating such images?  What
>>>>> programs are you using?
>>>>
>>>> I burn flac files in multiple sessions, each session have separate
>>>> directory, on DVD+R DL MKM/003
>>>> After I used 4gms switch mounted fs shows nothing. (but there is >5GB
>>>> of
>>>> data)
>>>>
>>>> According to growisofs source BD (bluray) dont need this switch at all
>>>> ...
>>>>
>>>>> This sounds like a pretty unsurprising 32-bit truncation
>>>>> bug:  the filesystem structures in ISO9660 are all sector
>>>>> numbers so 8TB should be the natural limit (4G sectors
>>>>> times 2k bytes/sector).
>>>>
>>>> I did not tested this on FreeBSD amd64 yet.
>>>
>>> Update: Linux shows all sessions and Windows 7 shows only first one.
>>
>>
>> From the source code of groisofs.c:
>>
>>  * - DVD+R Double Layer support;
>>  * - -use-the-force-luke=4gms to allow ISO9660 directory structures
>>  *   to cross 4GB boundary, the option is effective only with DVD+R DL
>>  *   and for data to be accessible under Linux isofs a kernel patch is
>>  *   required;
>>
>> So I'm guessing it does something non standard, particularly if
>> windows also refuses to see the data.
>
> That is pretty old, from 2.4 era, it was added after it was found that
> isofs had bug. Windows at least "try" to show something - only one
> session, but fourth and not second session crossed 4GB limit.
>
> The source also claims that in BD case there is no need for _force_
> switch at all.
>

Mounting with -norrip shows all sessions.
Kernel displays "RRIP without PX field?" if I try to mount "normal" way.
Received on Tue Apr 27 2010 - 19:19:02 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:03 UTC