Hi Bill, You can put the following into /boot/loader.conf hw.ata.ata_dma="0"
It disables the dma on ata drives. It works on some of the high speed 100X 2gb CF cards, but not on the Seagate on my systems. Since mine are VIA-based, so it may be different from the Wrap, which is based on the AMD (Geode). On lInux, I put the "ide=nodma" on the kernel parameter line and it works. Otherwise, it just hangs. It looks like the Linux ata driver may have some looser timeouts than FreeBSD, allowing it to recover during the initial start-up. On another note, PIO mode 4 is 16.6 MB/s, about the speed of 100X CF cards. It is plenty fast enough! Umda2 in the Seagate is ata33 or 33MB/s, about twice as fast. But, I think it is more of bust speed, than sustained. Bao on Wed, May 31, 2006 at 12:48:19AM -0700, [EMAIL PROTECTED] wrote: > At 05:27 PM 5/30/2006, Bao C. Ha wrote: > > >I think it is a kernel issue. On Via-based hardware, the CF socket > >cannot handle DMA, just PIO, not until the c7-based systemboard > >arriving later this year. These high capacity and fast CF will > >keep choking on these boards during boot-up. > > The WRAP CF card supports (apparently) DMA33, so it's not quite that > simple. From my boot sequence: > > atapci0: <National Geode SC1100 ATA33 controller> port > 0x1f0-0x1f7,0x3f6,0x170-0 > x177,0x376,0xfc00-0xfc0f at device 18.2 on pci0 > ata0: <ATA channel 0> on atapci0 > ata1: <ATA channel 1> on atapci0 > > The CF device is on ata0-master. > > BUT, as I just found out, reading the WRAP docs: > > "Another possible issue are cards with IDE DMA support ? the DMA signals > are a fairly recent > addition to the CF spec, and are not supported by WRAP. If you experience > problems with IDE > access later in the boot process, try to disable IDE DMA in your operating > system." > > To test this theory, is there a way to tell the boot sequence to use only > PIO for disk access, or do we need to build a new kernel? > > I'm completely bummed out by this discovery, which was just sitting there > waiting for discovery. Yes, if I can get the embedded platform to run on my > Seagate ST1.2 by using PIO, I'll do it, but there goes some of the benefit > of the real hard disk. > > Bill Plein > [EMAIL PROTECTED] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > -- Best Regards. Bao C. Ha Hacom OpenBrick Distributor USA http://www.hacom.net voice: (714) 530-8817 fax: (714) 530-8818 8D66 6672 7A9B 6879 85CD 42E0 9F6C 7908 ED95 6B38 --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
