FreeBSD and serial ata
Rumen Telbizov
altares at e-card.bg
Wed Nov 5 00:44:46 PST 2003
Hi
> If I dont have the BIOS set to SATA only, I can make the system repeatedly
> crash by running bonnie with a -s value twice the size of the RAM on the
> machine. So far it seems to run fine in this mode. I can provide more
> details if you are interested.
Hmm, I tested my sata discs with bonnie++ and rawio (yeah I did not have
information on that disc) and also loaded the system with *heavy*
disc operations myself like copying files (2GB) and making replications
on 3 other systems and I can gladly say that so far the system
is rock solid. The only thing that I still wonder about is if
the discs run as fas as they can in this legacy mode set in the
bios.
I ran bonnie++ with filesize 2GB and -x 100 - took about 10 hours :)
My RAM is 1GB.
Another test on the discs was vim file.sql (file.sql == 7GB).
Biiiiig waiting but that's because of the editor. Vi also
created ~8GB .swp file on this operation. Nevertheless all
operations on this system are stable (thank god).
A simple copy test from one disc to the other shows me
that I have exactly 32MBytes/s transfer. I repeated the test
many times. I do have softupdates enabled on both discs.
bonnie++ shows results about 56MBytes/s in sequential read/write
as far as I remember.
What is the speed that you can suck out from your discs ?
I will be glad to hear from you.
Have a nice day.
Rumen Telbizov
On Tue, Nov 04, 2003 at 03:19:42PM -0500, Mike Tancsa wrote:
>
> Just to follow up on this, I was testing an AOPEN board with
>
> atapci0 at pci0:31:2: class=0x01018a card=0x042ea0a0 chip=0x24d18086
> rev=0x02 hdr=0x00
> vendor = 'Intel Corporation'
> device = '82801EB (ICH5) SATA Controller'
> class = mass storage
> subclass = ATA
> chip0 at pci0:0:0: class=0x060000 card=0x25708086 chip=0x25708086 rev=0x02
> hdr=0x00
> vendor = 'Intel Corporation'
> device = '82865G/PE/P, 82848P DRAM Controller / Host-Hub Interface'
> class = bridge
> subclass = HOST-PCI
>
> using a seagate SATA drive
>
> satatest# atacontrol cap 0 0
> ATA channel 0, Master, device ad0:
>
> ATA/ATAPI revision 6
> device model ST380023AS
> serial number 3KB1EWQR
> firmware revision 3.01
> cylinders 16383
> heads 16
> sectors/track 63
> lba supported 156301488 sectors
> lba48 not supported 156301488 sectors
> dma supported
> overlap not supported
>
> Feature Support Enable Value Vendor
> write cache yes yes
> read ahead yes yes
> dma queued no no 0/00
> SMART yes yes
> microcode download yes yes
> security yes yes
> power management yes yes
> advanced power management no no 65278/FEFE
> automatic acoustic management yes no 0/00 128/80
>
>
> If I dont have the BIOS set to SATA only, I can make the system repeatedly
> crash by running bonnie with a -s value twice the size of the RAM on the
> machine. So far it seems to run fine in this mode. I can provide more
> details if you are interested.
>
> ---Mike
>
>
> At 04:41 AM 03/11/2003, Rumen Telbizov wrote:
> >Thanks for the fast reply Soren.
> >I do appreciate it.
> >
> >On Mon, Nov 03, 2003 at 10:06:22AM +0100, Soren Schmidt wrote:
> >> It seems Rumen Telbizov wrote:
> >> > any answers the first time.
> >> >
> >> > I have P4P800-VM motherboard with serial ata
> >> > controlers (ICH5) and serial ata discs. I managed
> >> > to install FreeBSD 4.9-STABLE on this box when
> >> > setting the IDE controller in legacy mode.
> >> > Everything works normal except when booting,
> >> > the drives are set into UDMA33 mode:
> >> >
> >> > atapci0: <Intel ICH5 SATA150 controller> port
> >> > 0xfc00-0xfc0f,0-0x3,0-0x7,0-0x3,0-
> >> > 0x7 irq 0 at device 31.2 on pci0
> >> > ata0: at 0x1f0 irq 14 on atapci0
> >> > ata1: at 0x170 irq 15 on atapci0
> >> >
> >> > ad0: DMA limited to UDMA33, non-ATA66 cable or device
> >> > ad0: 76319MB <ST380013AS> [155061/16/63] at ata0-master UDMA33
> >> > ad1: DMA limited to UDMA33, non-ATA66 cable or device
> >> > ad1: 76319MB <ST380013AS> [155061/16/63] at ata0-slave UDMA33
> >> >
> >> > I read somewhere that this message is a hoax and that
> >> > the drive should actually work in UDMA150. Is this my case?
> >>
> >> Yes, there are no "modes" on SATA (so far) its just SATA-150 in this
> >> case and the drives (if true SATA drives not PATA with converters)
> >> run SATA-150 over the inteface as soon as youselect any DMA mode.
> >>
> >Great!
> >Yes my drives ARE Serial ata.
> >I tested the speed by copying from one drive to
> >the other and I have a rough estimation of 32MBytes/s
> >transfer. (Softupdates enabled).
> >bonnie++ reports ~ 55MBytes/s sequential read/write
> >speed.
> >
> >> > I also checked the source of ata-dma.c:
> >> > Here is the section.
> >> >
> >> > #if 1
> >> > if (udmamode > 2 && !atadev->param->hwres_cblid) {
> >> > ata_prtdev(atadev,"DMA limited to UDMA33, non-ATA66 cable or
> >device\n");
> >> > udmamode = 2;
> >> > }
> >> > #endif
> >> >
> >> >
> >> > Maybe this should sound like:
> >> > if ( (udmamode > 2) && (udmamode < 5) &&
> >(!atadev->param->hwres_cblid) ) {
> >> > ata_prtdev(atadev,"DMA limited to UDMA33, non-ATA66 cable or
> >device\n");
> >> > udmamode = 2;
> >> > }
> >> >
> >> > All comments are wellcome.
> >>
> >> That wont work..
> >>
> >> The problem here is that -stable's ATA driver doesn't have the
> >infrastructure
> >> to handle SATA properly, so what you have here is a crude hack to get
> >some
> >> controllers working, but you will have to live with the fallout above...
> >
> >Yeah I know this is an ugly patch. I just thought it might be
> >usefull to make the OS stop weeping nothing else. Anyway.
> >
> >What exactly do you mean by saying that -stable's ATA driver doesn't have
> >the infrastructure to handle SATA properly? How do my controlers
> >work now? Because of the legacy support in the BIOS?
> >If the controler is in legacy mode does it work with the full
> >speed as if it were in native (enhanced) mode?
> >
> >
> >Thank you.
> >
> >Rumen Telbizov
> >_______________________________________________
> >freebsd-stable at freebsd.org mailing list
> >http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> >To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
>
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
>
More information about the freebsd-stable
mailing list