1 |
On Wed, 02 Jun 2004 12:15:00 +0200 |
2 |
Toon Verstraelen <Toon.Verstraelen@×××××.be> wrote: |
3 |
|
4 |
> Afaik most ide hard disks have write caching switched on for performance |
5 |
> |
6 |
> reasons. (At least on my disk this is so. hdparm -i /dev/hda tells me.) |
7 |
> Is their a fast alternative for write caching on ide disks? |
8 |
> |
9 |
The theoretical solution would be TCQ. However this works only on IBM |
10 |
disks and is very unreliable on many controllers. As a result support has |
11 |
been removed from kernel. |
12 |
|
13 |
Hopefully this situation will improve with Serial ATAs Native Command |
14 |
Queuing. The kernel infrastructure (write barriers and libata) is already |
15 |
in place, only the hardware has to appear. |
16 |
|
17 |
Regards |
18 |
|
19 |
-- |
20 |
gentoo-security@g.o mailing list |