Date: Wed, 27 Jan 1999 00:04:55 -0500 From: Christopher Masto <chris@netmonger.net> To: current@FreeBSD.ORG Subject: Re: IDE DMA works, I'll be a... Message-ID: <19990127000455.A6699@netmonger.net> In-Reply-To: <Pine.BSF.4.05.9901252123110.13366-100000@janus.syracuse.net>; from Brian Feldman on Mon, Jan 25, 1999 at 09:25:37PM -0500 References: <Pine.BSF.4.05.9901252106160.12387-100000@janus.syracuse.net> <Pine.BSF.4.05.9901252123110.13366-100000@janus.syracuse.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Well, now I'm plenty confused. I wasn't aware that IDE DMA didn't
work, so this subject line caught me a bit by suprise. Maybe I just
have "supported" hardware everywhere, or maybe I'm missing something
and I don't even know it. Or maybe I'm just posting this for
comparison purposes.
I know that this 2.5-year-old machine has an Asus motherboard of
some sort:
FreeBSD 4.0-CURRENT #1: Sun Jan 24 02:49:52 EST 1999
root@lion-around.at.yiff.net:/usr/local/usr-src/sys/compile/LION-AROUND
Timecounter "i8254" frequency 1193182 Hz
CPU: Pentium/P54C (166.19-MHz 586-class CPU)
Origin = "GenuineIntel" Id = 0x52c Stepping=12
Features=0x1bf<FPU,VME,DE,PSE,TSC,MSR,MCE,CX8>
real memory = 100663296 (98304K bytes)
config> quit
avail memory = 94875648 (92652K bytes)
Preloaded elf kernel "kernel" at 0xf02d9000.
Probing for devices on PCI bus 0:
chip0: <Intel 82439> rev 0x03 on pci0.0.0
chip1: <Intel 82371SB PCI to ISA bridge> rev 0x01 on pci0.7.0
ide_pci0: <Intel PIIX3 Bus-master IDE controller> rev 0x00 on pci0.7.1
[...]
wdc0 at 0x1f0-0x1f7 irq 14 flags 0xa0ffa0ff on isa
wdc0: unit 0 (wd0): <WDC AC32500H>, DMA, 32-bit, multi-block-16
wd0: 2441MB (4999680 sectors), 4960 cyls, 16 heads, 63 S/T, 512 B/S
wdc0: unit 1 (atapi): <CD-ROM CDU311/3.0i>, removable, accel, dma, iordis
acd0: drive speed 1378KB/sec, 128KB cache
[...]
wdc1 at 0x170-0x177 irq 15 flags 0xa0ffa0ff on isa
wdc1: unit 0 (wd2): <Maxtor 91008D7>, DMA, 32-bit, multi-block-16
wd2: 9617MB (19696320 sectors), 19540 cyls, 16 heads, 63 S/T, 512 B/S
On the old wd0 that I got with the machine:
$ time dd if=/dev/zero of=test2 bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 5.177823 secs (6480413 bytes/sec)
dd if=/dev/zero of=test2 bs=32k count=1024 0.03s user 2.08s system 28% cpu 7.458 total
$ time dd if=/dev/rwd0 of=/dev/null bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 4.064697 secs (8255088 bytes/sec)
dd if=/dev/rwd0 of=/dev/null bs=32k count=1024 0.00s user 0.14s system 3% cpu 4.070 total
And on the brand spanking new Maxtor UltraDMA happy happy:
$ time dd if=/dev/zero of=test2 bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 3.073696 secs (10916640 bytes/sec)
dd if=/dev/zero of=test2 bs=32k count=1024 0.01s user 2.26s system 72% cpu 3.124 total
$ time dd if=/dev/rwd2 of=/dev/null bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 2.521986 secs (13304765 bytes/sec)
dd if=/dev/rwd2 of=/dev/null bs=32k count=1024 0.00s user 0.13s system 5% cpu 2.528 total
Just out of curiosity, I compared with a machine with what was at the time
an expensive SCSI controller. Same CPU as mine.
FreeBSD 2.2.7-STABLE #0: Sat Aug 8 22:38:59 EDT 1998
[...]
ahc0 <Adaptec 3940 Ultra SCSI host adapter> rev 0 int a irq 9 on pci1:4:0
ahc0: aic7880 Wide Channel A, SCSI Id=7, 16 SCBs
(ahc0:0:0): "SEAGATE ST34371W 0484" type 0 fixed SCSI 2
sd0(ahc0:0:0): Direct-Access 4148MB (8496884 512 byte sectors)
$ time dd if=/dev/zero of=test2 bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 3.393906 secs (9886671 bytes/sec)
dd if=/dev/zero of=test2 bs=32k count=1024 0.04s user 1.33s system 40% cpu 3.411 total
$ time dd if=/dev/rsd0 of=/dev/null bs=32k count=1024
1024+0 records in
1024+0 records out
33554432 bytes transferred in 3.525452 secs (9517767 bytes/sec)
dd if=/dev/rsd0 of=/dev/null bs=32k count=1024 0.03s user 0.12s system 2% cpu 5.330 total
The bleeding edge and I have a pretty good relationship. I haven't
been bitten by softupdates, IDE DMA, or the new VM (well, I did have
that panic, but the fix had already become available).
Of course, I'm now running afoul of 4.0-related libtool breakage in
the ports collection, but it's not exactly a mystery to fix.
--
Christopher Masto Director of Operations NetMonger Communications
chris@netmonger.net info@netmonger.net http://www.netmonger.net
"Good tools allow users to do stupid things." -- Clay Shirky
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990127000455.A6699>