Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 2 Mar 1996 19:58:57 +0200 (EET)
From:      Tomi Vainio <tomppa@fidata.fi>
To:        amurai@spec.co.jp
Cc:        nate@sri.MT.net, questions@FreeBSD.org
Subject:   Re: Backup problem
Message-ID:  <199603021758.TAA05633@zeta.fidata.fi>
In-Reply-To: <199603020745.QAA02341@tama.spec.co.jp>
References:  <199602292328.BAA02445@zeta.fidata.fi> <199603020745.QAA02341@tama.spec.co.jp>

next in thread | previous in thread | raw e-mail | index | archive | help
Atsushi Murai writes:
 > > AMI SCSI adapter is EISA bus master and it also can do DMA over 16MB
 > > memory in Adaptec 1542 compability mode  but FreeBSD driver doesn't
 > > support it.
 > 
 > Sounds like you are confusing.. 
 > 
 > 	1. Adpatec 1542 compatible mode is impossible to handling
 >            above 16MB due to 24bit pointer architecutre of firmware.
 >            So even hardware has 32bit addressing like EISA, driver
 >            can tell to firmware as 32bit address.
 >         2. Some old EISA mother board has a limitation to over 24 bit 
 >            addressing phygically.
 >
I have used this card in 1542 mode with 32MB RAM before there was
operating systems which can handle bounce buffers (year was 1991 or 1992).
Dont ask how card works but it worked fine with an old ISC SVR3 V2.2
system. This motherboard works fine with 32MB RAM and EISA bus master
cards. I have used ISC and SCO SVR3, FreeBSD 2.0, 2.0.5 and 2.1.0
without any problems on tish machine.

 > > I'm not using options BOUNCE_BUFFERS when running this
 > > card in Buslogic 742a EISA mode but with Adaptec driver FreeBSD wont
 > > even boot without it.
 > 
 > As far as I know, BOUNCE_BUFFERS is not for driver opiton, it's
 > implement to upper layer (i.e. SCSI system). Addition aha driver is no
 > longer compatible adaptec 1542 well know as industry standard ;-)
 > 
Lets be more specific on this. When I compile kernel with bt driver I
don't use BOUNCE_BUFFERS and when I compile with aha drive I use
BOUNCE_BUFFERS.

My backup story continues. Writing backup was quite fast (~350kb/s)
but reading taked almost four hours (~90kb/s). Should I change block
size or how I can get better performance out of this DAT drive?

  Tomppa



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603021758.TAA05633>