Date: Wed, 4 Apr 2001 16:04:48 +0700 From: "Pavel Gubin" <pg@ie.tusur.ru> To: <freebsd-hardware@freebsd.org> Subject: ATA problems on MicroStar MS6340M KM-133 SocketA motherboard Message-ID: <000d01c0bce6$4cfb2000$497ac0d4@ie.tusur.ru>
next in thread | raw e-mail | index | archive | help
Hello,
I've encountered a problem running misc FreeBSD releases on MS6340M M/B.
System:
Microstar MS6340M SocketA Athlon/Duron uATX motherboard - KM133 chipset
north bridge=VT8365 (~KT133 + integrated S3 Savage SVGA on chip),
south bridge=VT62C686B UATA100
Award BIOS v6.0; versions tried: 2.8, 4.0
(tried with external S3 Trio3D/2X AGP - no difference)
Athlon 700 SocketA CPU (replacing with Duron 700 makes no difference)
64M PC133 Samsung DIMM (8M used for videobuffer if using integrated video)
Fujitsu MPF3102AT 10Gb UATA66 drive (1st controller, UATA66 cable)
(replacing with Quantum Fireball CR43A013 4.3Gb UATA66 makes no
significant difference)
Philips PCA362CDA 36x CD-ROM (2nd controller, standard cable)
Previous variant of system:
All devices but motherboard are the same;
M/B - Gigabyte GA7ZM SocketA KT133+VT82C686A (UDMA66),
SVGA - S3 Trio3D/2X AGP
FreeBSD versions tried: 4.2-RELEASE, 4.3-RC, 5.0-CURRENT
1. FreeBSD (all versions) doesn't sense if UATA66 drive connected
to UATA66 controller by a non-ATA66 cable (whereas BIOS senses),
which leads to ICRC errors and falling back to PIO mode.
2. Strange transfer rates measured by iozone 2.01 (200Mb file on
700Mb /tmp filesystem mounted async, 8192b chunks) - write/read, Mb/s:
UDMA mode: ---------- MS6340M --------- GA7ZM
BIOS v2.8 BIOS v4.0
4.2 ~22/23(*) ~5/17(*) ~22/23
4.3 ~ 7/23 ~4/14 ~ 7/23
5.0 ~ 7/23 ~4/14 ~ 7/23
PIO mode: all combinations of FreeBSD/BIOS gives xfer rate
~7-8 Mb/s write/read (at 100% CPU load...)
(*) 4.2 was slightly hacked to limit UDMA to UDMA33 mode, because
working in UDMA66 leads to ICRC errors and fallback to PIO;
this problem only appears with 4.2-RELEASE running on MS6340M;
4.3/5.0 on MS6340M and all versions on GA7ZM runs without errors,
(**) Forcing drive to be UATA33 (using Fujitsu's firmware modification
utility) changes nothing; using Quantum drive instead of Fujitsu
changes maximum xfer rates (~14 Mb), but essentially situation
remains the same, including ICRC on 4.2+MS6340M
(***) CDROM drive works with errors when in UDMA mode on MS6340M,
whereas on GA7ZM it works OK.
So, the question: can I (or anybody else) do with this stuff to
make it work properly?
Thanks in advance,
/pg
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000d01c0bce6$4cfb2000$497ac0d4>