2.11BSD/sys/OTHERS/emulex-cs02/cs02-q22-fixes
Received: from lanai.cs.ucla.edu (lanai.cs.ucla.edu.ARPA) by corwyn.cognet.ucla.edu (4.12/1.1)
id AA02393; Wed, 17 Aug 88 12:45:53 pdt
Return-Path: <bostic@okeeffe.Berkeley.EDU>
Received: from okeeffe.Berkeley.EDU by lanai.cs.ucla.edu (Sendmail 5.58.2/2.10)
id AA02195; Wed, 17 Aug 88 12:43:36 PDT
Received: by okeeffe.Berkeley.EDU (5.60/1.29)
id AA01264; Wed, 17 Aug 88 12:45:18 PDT
Date: Wed, 17 Aug 88 12:45:18 PDT
From: bostic@okeeffe.Berkeley.EDU (Keith Bostic)
Message-Id: <8808171945.AA01264@okeeffe.Berkeley.EDU>
To: casey@cs.ucla.edu
Subject: fix to dh
Status: RO
Not sure if we should put this in -- maybe as part of the man page
for the dh? Or a README? Your choice...
--keith
> From tcdmath!ajudge@jade.berkeley.edu Wed Aug 17 11:02:26 1988
> Date: Wed, 17 Aug 88 18:21:39 BST
> From: Alan Judge <ajudge@maths.tcd.ie>
> Subject: New Release of 2.10 and Ethernetting
> To: bostic@okeeffe.Berkeley.EDU
> X-Fortune: Mix a little foolishness with your serious plans; it's lovely to be
> silly at the right moment. -- Horace
Also when we installed 2.10 we had a problem with port-boards because of a
slight difference between Emulex CS-02 boards in DH-11/DM-11 emulation and
the real thing. The rules for 22-bit QBUS operation are different. Anyway at
the time we fixed the problem locally and since we were not on the net
reliably I don't think we told anybody. In case the problem still remains
unfixed I include patches for 2.10 for the problem. You can then post a fix
to news if necessary.
--
Alan Judge, System Administrator
School of Mathematics, Trinity College, Dublin.
ajudge@maths.tcd.ie
RCS file: RCS/dh.c,v
retrieving revision 1.1
diff -c -r1.1 dh.c
*** /tmp/,RCSt1011296 Wed Aug 17 18:14:33 1988
--- dh.c Sun Nov 8 19:37:11 1987
***************
*** 405,413 ****
--- 405,419 ----
#else
{
ubadr_t car;
+ #if defined(CS02)
+ struct dmdevice *dmaddr = (struct dmdevice *)dmi
nfo[dh].ui_addr;
+ car = (ubadr_t)addr->dhcar;
+ car |= ((ubadr_t)(dmaddr->dmlst_h&077) << 16);
+ #else
car = (ubadr_t) addr->dhcar
| (ubadr_t)(addr->dhsilo & 0300) << 10;
+ #endif
cntr = car - cpaddr(tp->t_outq.c_cf);
ndflush(&tp->t_outq, cntr);
}
***************
*** 492,500 ****
--- 498,514 ----
addr->dhcar = (u_short)cpaddr(tp->t_outq.c_cf);
#else
ubadr_t uba;
+ #if defined(CS02)
+ struct dmdevice *dmaddr = (struct dmdevice *)dminfo[dh].ui_addr;
+ #endif
uba = cpaddr(tp->t_outq.c_cf);
+ #if defined(CS02)
+ addr->un.dhcsrl = (unit&017) | DH_IE;
+ dmaddr->dmlst_h = hiint(uba) & 077;
+ #else
addr->un.dhcsrl = (unit&017) | DH_IE | ((hiint(uba)<<4)&060);
+ #endif
addr->dhcar = loint(uba);
#endif
{ short word = 1 << unit;
===================================================================
RCS file: RCS/dmreg.h,v
retrieving revision 1.1
diff -c -r1.1 dmreg.h
*** /tmp/,RCSt1011296 Wed Aug 17 18:14:39 1988
--- dmreg.h Sun Nov 8 19:37:43 1987
***************
*** 11,19 ****
--- 11,30 ----
*/
struct dmdevice {
short dmcsr; /* control status register */
+ #if defined(CS02)
+ union { /* Need to access high byte independently */
+ short dmlst_w;
+ char dmlst_b[2];
+ } cs02un;
+ #define dmlstat cs02un.dmlst_w
+ #define dmlst_l cs02un.dmlst_b[0] /* low byte */
+ #define dmlst_h cs02un.dmlst_b[1] /* high byte */
+ #else
short dmlstat; /* line status register */
+ #endif
short dmpad1[2];
};
+
/* bits in dm csr */
#define DM_RF 0100000 /* ring flag */
Return-Path: tcdmath!@CUNYVM.CUNY.EDU:ajudge
Return-Path: <tcdmath!@CUNYVM.CUNY.EDU:ajudge>
Received: from CUNYVM.CUNY.EDU (cunyvm.cuny.edu.ARPA) by corwyn.cognet.ucla.edu (4.12/1.1)
id AA00331; Mon, 26 Sep 88 16:46:33 pdt
Received: from CUNYVM.BITNET by CUNYVM.CUNY.EDU (IBM VM SMTP R1.1) with BSMTP id 7572; Mon, 26 Sep 88 18:48:06 EDT
Received: from CUNYVM by CUNYVM.BITNET (Mailer X2.00) with BSMTP id 0180; Mon,
26 Sep 88 18:45:03 EDT
Received: from IRLEARN.BITNET by CUNYVM.CUNY.EDU (IBM VM SMTP R1.1) with BSMTP
id 7550; Mon, 26 Sep 88 18:45:01 EDT
Received: from ccvax.ucd.ie by IRLEARN.BITNET (Mailer X1.24) with BSMTP id
0531; Sat, 24 Sep 88 19:39:42 GMT
Received: from cs.tcd.ie by ccvax.ucd.ie; Sat, 24 Sep 88 19:35 GMT
Received: from csclnb.uucp by cs.tcd.ie; Sat, 24 Sep 88 19:35 GMT
Received: from maths by csclnb.cs.tcd.ie id aa28619; 24 Sep 88 19:33 EDT
Date: Sat, 24 Sep 88 17:49:15 BST
From: Alan Judge <ajudge%maths.tcd.ie%CUNYVM.CUNY.EDU@CUNYVM.CUNY.EDU>
Subject: Re: Your fixes for pdpuba/dh.c
To: casey@cognet.ucla.edu
X-Fortune: If everything seems to be going well, you have obviously overlooked
something.
In-Reply-To: Your message of Mon, 29 Aug 88 00:58:18 -0700 .
Message-Id: <8809241749.aa011619@hamilton.maths.tcd.ie>
> Alan,
> This is my third attempt to get through to you. Keith Bostic forwarded
> your note about the changes you had to make to dh.c to get it to work
> with the Emulex CS-02. I'd like to know if these changes are specific to
> the Emulex, or to *any* DH on a QBUS. In particular, I wasn't aware that
> there was any DH available for the QBUS. Is Emulex's the only such or
> does DEC also supply one? Please answer as soon as possible.
>
> Casey
I am just back from holidays and this message arrived today. It seems that
this message arrived was delayed in IRLEARN.BITNET (a site in another
irish university) for 3 weeks. None of your other
messages have yet arrived. Normally the through-time by this link is less
than an hour. Hopefully mail to ajudge@maths.tcd.ie should only take a matter
of hours from the states. Something must have gone wrong while I was
away.
The manual for the CS02/H1 says that the board can emulate either one
DH11/DM11 or 2 DHV11's. This and the programming section seems to imply that
there is a DEC product. Unfortunately I don't know for sure. The Emulex board
may be emulating the DH on some other bus. However the DEC product probably
does not support 22-bit addressing as Emulex use part of the LSR (line status
register for this). The patches ARE specific to 22-bit addressing on a Q-BUS,
they may be specific to Emulex if DEC does have a 22-bit Q-BUS DH.
--
Alan