On Thu, 21 Sep 2006, Jürgen Keil wrote:
The latest on-20060918 source drop included some pcfs fixes:
PSARC case 2005/361 : PCFS timestamp handling cleanup
BUG/RFE: 4401668 RFE: pcfs should support force unmount
BUG/RFE: 6248624 pcfs file timestamp handling needs cleanup
BUG/RFE: 6278733 pcfs statvfs implementation is slow
BUG/RFE: 6351785 pcfs should support cluster sizes over 32Kb for Fat32 despite
violating the spec
BUG/RFE: 6364444 pcfs can still crash in pc_getcluster with one-off overflow
Can anyone running these bits (or newer bits) try the following:
[b]
# mkfile 1440k /tmp/bla
# lofiadm -a /tmp/bla /dev/lofi/9
# mkfs -F pcfs -o nofdisk,fat=12,spc=1,nsect=18,ntrack=2,size=2880 /dev/rlofi/9
# mount -F pcfs /dev/lofi/9 /mnt
# touch /mnt/AbCdEfGh
# touch /mnt/AbCdEfGhIj
# ls -l /mnt
total 0
-rwxrwxrwx 1 root root 0 Sep 21 17:17 ABCDEFGH
-rwxrwxrwx 1 root root 0 Sep 21 17:17 AbCdEfGhIj
[/b]
With S10 6/2006 the mixed upper/lowercase characters of the
short filename were preserved. Does anyone else get uppercase
characters only? A new bug?
Hmm - I've replaced pc_validchar() in pc_subr.c with a (what I think)
equivalent macro in <sys/fs/pc_dir.h>. I don't see where I could've made
the mistake there ? Someone willing to check that ?
The new code:
http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/sys/fs/pc_dir.h#pc_validchar
The previous code:
http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/fs/pcfs/pc_subr.c?r=1.15
(goto line #399)
That's the only filename-parsing related change in the above bunch. I'm
trying to understand how it could cause what you describe and can't quite
come up with it (yet). Will see.
On the other hand, I don't think it's a bug if the short filename doesn't
accept mixed case. That's spec.
Best wishes,
FrankH.
_______________________________________________
ufs-discuss mailing list
[email protected]
