Re: Linux ISO-9660 Rock Ridge bug needs fix

From: Luca Tettamanti
Date: Tue Oct 17 2006 - 14:02:11 EST


Ismail Donmez <ismail@xxxxxxxxxxxxx> ha scritto:
>>
>> while working on better ISO-9660 support for the Solaris Kernel,
>> I recently enhanced mkisofs to support the Rock Ridge Standard version 1.12
>> from 1994.
>>
>> The difference bewteen version 1.12 and 1.10 (this is what previous
>> mkisofs versions did implement) is that the "PX" field is now 8 Byte
>> bigger than before (44 instead of 36 bytes).
>
> Is there a test iso file somewhere? I think the attached *untested* patch will
> fix it.

I was also looking at this ;) I cannot reproduce the failure even with
images generated with the new version of mkisofs (I actually _see_ that
PX record size is changed, but isofs doesn't seem to care...).

> diff --git a/fs/isofs/rock.c b/fs/isofs/rock.c
> index f3a1db3..061a633 100644
> --- a/fs/isofs/rock.c
> +++ b/fs/isofs/rock.c
> @@ -349,6 +349,7 @@ #endif
> inode->i_nlink = isonum_733(rr->u.PX.n_links);
> inode->i_uid = isonum_733(rr->u.PX.uid);
> inode->i_gid = isonum_733(rr->u.PX.gid);
> + inode->i_ino = isonum_733(rr->u.PX.ino);
> break;

I don't think it's correct. When reading disk with old format i_ino will
be filled with garbage.
Now, who is in charge of isofs?

Signed-off-by: Luca Tettamanti <kronos.it@xxxxxxxxx>

---
fs/isofs/rock.c | 4 ++++
fs/isofs/rock.h | 1 +
2 files changed, 5 insertions(+)

diff --git a/fs/isofs/rock.c b/fs/isofs/rock.c
index f3a1db3..9a87010 100644
--- a/fs/isofs/rock.c
+++ b/fs/isofs/rock.c
@@ -349,6 +349,10 @@ #endif
inode->i_nlink = isonum_733(rr->u.PX.n_links);
inode->i_uid = isonum_733(rr->u.PX.uid);
inode->i_gid = isonum_733(rr->u.PX.gid);
+
+ /* Rock Ridge V1.12, override inode number */
+ if (rr->len == 44)
+ inode->i_ino = isonum_733(rr->u.PX.inode);
break;
case SIG('P', 'N'):
{
diff --git a/fs/isofs/rock.h b/fs/isofs/rock.h
index ed09e2b..df5f2a7 100644
--- a/fs/isofs/rock.h
+++ b/fs/isofs/rock.h
@@ -33,6 +33,7 @@ struct RR_PX_s {
char n_links[8];
char uid[8];
char gid[8];
+ char inode[8];
};

struct RR_PN_s {


Luca
--
Sbagliare è umano, ma per incasinare davvero le cose serve un computer.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/