Return to BSD News archive
#! rnews 3366 sserve.cc.adfa.oz.au
Newsgroups: comp.unix.bsd.bsdi.misc
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!simtel!vtc.tacom.army.mil!ulowell.uml.edu!europa.chnt.gtegsc.com!news.mathworks.com!solaris.cc.vt.edu!spcuna!spcvxb!terry
From: terry@spcvxb.spc.edu (Terry Kennedy, Operations Mgr.)
Subject: Re: 10 Mbs FAST SCSI-ii
Nntp-Posting-Host: spcvxa.spc.edu
References: <3rph03$m2e@gold.interlog.com>
Sender: news@spcuna.spc.edu (Network News)
Organization: St. Peter's College, US
Date: Fri, 16 Jun 1995 07:40:06 GMT
Message-ID: <1995Jun16.034006.1@spcvxb.spc.edu>
Lines: 74
In article <3rph03$m2e@gold.interlog.com>, ryan@interlog.com (David Ryan) writes:
> ve tried all the settings for my buslogic 946C card and attached devices
> baracuda and hawk) and the kernel still reports only 5 Mbps through isa0
> during boot up. This is a PCI card. Are there any kernel settings that
> could change/effect this. BSDI 2.0
It's a cosmetic bug. There are 2 places where the displayed message says
5MB/sec:
bha0: synch negotiation, parity, bus transfer 5 Mb/s, bus on/off 7/4 us
^^^^^^
tg0: synch transfer period 200 ns, offset 15 (5 MB/s)
^^^^^^
The first one is an emulation done in the BusLogic - this speed report
value is for compatibility with Adaptec 154x-series drivers (like BSDI's
"aha" driver). The actual speed is dependent on the bus used (EISA/VLB/
PCI) and you can't tell what bus is used from the program (all of the Bus-
Logic C-series boards report the same model type). I suppose we could just
suppress this field if it's a bha.
The next one is a trivial cosmetic error in the display routines. The
BusLogic boards have a sync base value of 100ns, like the newer Adaptec
models. The driver doesn't know this and uses a base of 200 for the dis-
play. This only affects the display, not the actual performance. That same
status line shows another Adaptec compatibility bit, "disconnect off". It
isn't really off, the BusLogic emulation is backwards. Again, this is just
in the display.
The attached patch to 2.0's ab_subr.c fixes both of these, as well as
shortening the model number string to fit on one line. If you have sources,
you can apply this patch to fix your display. Note that I just #ifdef'd
out the disconnect message - a proper fix would be to retain the controller
model info and only suppress the message if the controller is a bha. I'd
expect something like this will make its way back into BSDI's next release.
*** ab_subr.c_orig Fri Feb 3 02:32:57 1995
--- ab_subr.c Fri Mar 31 04:40:33 1995
***************
*** 270,276 ****
if (args->ia_drq != DRQNONE)
id = "542B";
else {
! id = "{445S,747,742,946C}";
}
break;
case 'B':
--- 270,277 ----
if (args->ia_drq != DRQNONE)
id = "542B";
else {
! id = "{445,747,742,946}"; /* tmk */
! ab->ab_sync_base = 100; /* tmk */
}
break;
case 'B':
***************
*** 473,481 ****
--- 474,484 ----
rates[i]);
}
if (in[16] & (1 << targ)) {
+ #if 0 /* tmk */
if (syncoff)
printf(",");
printf(" disconnect off");
+ #endif /* 0 */ /* tmk */
}
printf("\n");
#endif
Terry Kennedy Operations Manager, Academic Computing
terry@spcvxa.spc.edu St. Peter's College, Jersey City, NJ USA
+1 201 915 9381 (voice) +1 201 435-3662 (FAX)