Re: hdparm, cdrom, and retaining params
hi ya bill
> > > (see below) I have to use -d1 -c1 -X34 -u1.
> > there is no "reason" you "have to use" those options
> That's my poor choice of words. That's just what I used.
i famous for that
> > -u1 allow the cpu to do other stuff while waiting for the "disk
> > seeks/etc" .. non-critical..
> I think it was the -u1 that fixed my problem of the system clock falling
> way behind while ripping or burning a CD. For a 4 minute burn the clock
> would be almost a minute behind when done.
yeah... that'd give the cpu mroe time for other things,
but the rtc/nmi should be highest priority interrupts ...
no user app should be able to slow the "clock" down
- maybe something else is also whacky that is causing
the slow down
- cdrw burning is a slow process... the cpu should have
plenty of spare cycles
> > > And my config has:
> > >
> > > --- IDE chipset support/bugfixes
> > >
> > > [*] Generic PCI IDE chipset support
> > > [*] Sharing PCI IDE interrupts support
> > > [*] Generic PCI bus-master DMA support
> > > [*] Use PCI DMA by default when available
> > > [*] Intel PIIXn chipsets support
> > > [*] PIIXn Tuning support
> > and what chip set does the dell have for itd ide controller
> > - those bug fixes need to be turned on too
> You mean other than the lspci output I posted?
open your box... find all the big black squares yu can find
and write down all the names of the chipsets
go into the kernel source ....
- click into the IDE/DMA support section
and see if any of the numbers of the big-black-squares
matches anything in the kernel ide/dma options
- turn it on if it does
- leave the others off ( prevent potential problems )
> 00:07.1 IDE interface: Intel Corp. 82371AB/EB/MB PIIX4 IDE (rev 01)
> And I think I've got everything enabled in the kernel config that
> mentionds PIIXn.
i dont know if the above IDE interface is a ide controller or not
and if it is.. yes.. thats the plan...
> > > DMA modes: sdma0 sdma1 sdma2 mdma0 mdma1 *mdma2 udma0 udma1 udma2
notice the list of modes your drive supports
notice the * which indicates your current (very slow) disk xfer speed
> > you should be using udma2 ( -X64 ) instead ... faster/better
> Ok, I'll give that a try. man hdparm said for -X64 I'd need to prepare
> the chipset for UltraDMA beforehand, so I'm not sure what the means. Is
your drive supports udma0, udma1, udma2 ( -X64, -X65, -X66 )
- now get the kernel to support the ide controller
and it in turn will talk to the drive at the right speed