I'm not 100% sure, but I think, that developers already fixed this some time ago. But probably your Image is not new enough to already contain this fix...
I'm not 100% sure too, but I am quite convinced that an image builded with source date 20090612 have to be new enought to be updated.... 
But the situation remains the same... 
Please the following:
hdparm -tT /dev/discs/disc0/disc
hdparm -tT /dev/discs/disc1/disc
and
hdparm -i /dev/discs/disc0/disc
hdparm -i /dev/discs/disc1/disc
and tell us the output
Hi! Here the results:
root@dm7025:~# hdparm -tT /dev/discs/disc0/disc
/dev/discs/disc0/disc:
Timing buffer-cache reads: 128 MB in 4.05 seconds = 31.59 MB/sec
Timing buffered disk reads: 64 MB in 4.76 seconds = 13.44 MB/sec
root@dm7025:~# hdparm -tT /dev/discs/disc1/disc
/dev/discs/disc1/disc:
Timing buffer-cache reads: 128 MB in 4.03 seconds = 31.72 MB/sec
Timing buffered disk reads: 64 MB in 19.50 seconds = 3.28 MB/sec
root@dm7025:~# hdparm -i /dev/discs/disc0/disc
/dev/discs/disc0/disc:
Model=RIDATA CF-ATA, FwRev=1.1, SerialNo=1G02C8R507008A22257E
Config={ HardSect NotMFM DTR>10Mbs }
RawCHS=1985/16/63, TrkSize=33264, SectSize=528, ECCbytes=4
BuffType=1Sect, BuffSize=2kB, MaxMultSect=1, MultSect=off
CurCHS=1985/16/63, CurSects=2000880, LBA=yes, LBAsects=2000880
IORDY=no, tPIO={min:120,w/IORDY:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
AdvancedPM=no
* signifies the current active mode
root@dm7025:~# hdparm -i /dev/discs/disc1/disc
/dev/discs/disc1/disc:
Model=SAMSUNG HM160JC, FwRev=AP100-16, SerialNo=S0CMJD0P501681
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=16383/16/63, TrkSize=34902, SectSize=554, ECCbytes=4
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=off
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 *mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5
AdvancedPM=yes: disabled (255) WriteCache=enabled
Drive conforms to: reserved:
* signifies the current active mode
root@dm7025:~#Have a look:
root@dm7025:~# cd /usr/script/
root@dm7025:/usr/script# hdpinfo.sh
/dev/discs/disc0/disc:
Model=TRANSCEND, FwRev=20080128, SerialNo=20080903 0000B68D
Config={ HardSect NotMFM Fixed DTR>10Mbs }
RawCHS=3884/16/63, TrkSize=0, SectSize=576, ECCbytes=4
BuffType=DualPort, BuffSize=1kB, MaxMultSect=1, MultSect=off
CurCHS=3884/16/63, CurSects=3915072, LBA=yes, LBAsects=3915072
IORDY=no, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2
AdvancedPM=yes: disabled (255)
Drive conforms to: device does not report version: 4
* signifies the current active mode
/dev/discs/disc1/disc:
Model=SAMSUNG SP2014N, FwRev=VC100-41, SerialNo=S088J10L505601
Config={ Fixed }
RawCHS=16383/16/63, TrkSize=34902, SectSize=554, ECCbytes=4
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5
AdvancedPM=no WriteCache=enabled
* signifies the current active mode
root@dm7025:/usr/script# hdpbench.sh
/dev/discs/disc0/disc:
Timing buffer-cache reads: 128 MB in 3.57 seconds = 35.85 MB/sec
Timing buffered disk reads: 64 MB in 3.54 seconds = 18.08 MB/sec
hdparm: memory exhausted
/dev/discs/disc1/disc:
Timing buffer-cache reads: 128 MB in 3.67 seconds = 34.92 MB/sec
Timing buffered disk reads: 64 MB in 4.60 seconds = 13.91 MB/sec
root@dm7025:/usr/script#do you see the difference? Your DMA settings are somehow wrong. Please show us your bootup
Thanks a lot for your help! Of course I can see the difference!
So, as I started speaking about standard DMM configuration, my bootup and fstab now are closely standard.
Bootup:
echo 4 > /proc/progress
[ -e /dev/dbox/oled0 ] && ln -s /dev/dbox/oled0 /dev/dbox/lcd0
mount -t sysfs sysfs /sys
echo 12 > /proc/progress
depmod -Ae
echo 16 > /proc/progress
for mod in cifs autofs4 vfat; do
modprobe $mod
done
echo 18 > /proc/progress
mount -at nonfs,nosmbfs,noncpfs,nocifs
echo 22 > /proc/progress
swapon -a
echo 24 > /proc/progress
tar xzf /etc/var.tar.gz -C /
echo 28 > /proc/progress
# sleep after 10min
hdparm -S 120 /dev/discs/disc0/disc
echo 32 > /proc/progress
# set UDMA66
hdparm -X66 /dev/discs/disc0/disc
echo 34 > /proc/progress
# accustic management
hdparm -M 128 /dev/discs/disc0/disc
echo 44 > /proc/progress
for mod in xilleon dreambox_keyboard fp rt73 zd1211b rfmod; do
modprobe $mod
done
echo 50 > /proc/progress
echo 54 > /proc/progress
[ -e /etc/dropbear/dropbear_rsa_host_key ] && /usr/bin/showiframe /boot/bootlogo.mvi || /usr/bin/showiframe /boot/bootlogo_wait.mvi
echo 62 > /proc/progress
hostname -F /etc/hostname
echo 66 > /proc/progress
/usr/bin/wdog /usr/bin/dccamd -
echo 70 > /proc/progress
echo 78 > /proc/progress
echo 82 > /proc/progress
[ -e "/etc/sysctl.conf" ] && /sbin/sysctl -p "/etc/sysctl.conf"
echo 86 > /proc/progress
ldconfig
echo 90 > /proc/progress
echo 94 > /proc/progress
Fstab:
rootfs / auto defaults 1 1
proc /proc proc defaults 0 0
devpts /dev/pts devpts mode=0620,gid=5 0 0
usbdevfs /proc/bus/usb usbfs defaults 0 0
/dev/root /boot jffs2 ro 0 0
tmpfs /var tmpfs defaults 0 0
tmpfs /tmp tmpfs defaults 0 0
/dev/ide/host0/bus0/target0/lun0/part1 /media/hdd auto defaults 0 0
/dev/ide/host1/bus0/target0/lun0/part1 /media/cf auto defaults 0 0
/dev/ide/host1/bus0/target0/lun0/disc /media/cf auto defaults 0 0
/dev/scsi/host0/bus0/target0/lun0/part1 /media/usb auto defaults 0 0
/dev/scsi/host0/bus0/target0/lun0/disc /media/usb auto defaults 0 0Yes, the default bootup leads to bad performance, when a CF ist used. You should replace the lines 24-31 with just these:
# HDD Tuning
hdparm -X66 -m16 -S120 -M 128 /dev/discs/disc1/disc
echo 32 > /proc/progress
# CF Tuning
hdparm -X66 /dev/discs/disc0/disc
echo 34 > /proc/progress
then everything should be fine.
Yes, the default bootup leads to bad performance, when a CF ist used. You should replace the lines 24-31 with just these:
...
then everything should be fine.
Thanks again a lot. You were very kind.
I understand that is possible to "patch"this problem. And I had no doubt about it 
What I not able to undestand is, if
1) The problem is well known
2) Old images (pre 8k age) didn't suffer of this problem (so we are sure that it isn't an harware problem)
3) The patch is succesfull (so again we are sure that hardware is ok)
4) The solution seems to be quite simple
why are we still speaking about it, instead of having a strong and definitive solution into CVS?
Bye and thanks again! 
The problem is not resolved on 7025 yet! 
