X-pop3-spooler:	POP3MAIL 2.1.0 b 4 980420 -bs-
Mail-Copies-To:	never
To:	Shawn Leas <sleas AT ixion DOT honeywell DOT com>
Cc:	Topas <yrwang AT cc DOT Nctu DOT EDU DOT TW>, beastium-list AT Desk DOT nl
Subject:	Re: Rebuild the kernel of Redhat5.1!
References:	<Pine DOT HPP DOT 3 DOT 96 DOT 980903220958 DOT 4488B-100000 AT ixion DOT honeywell DOT com>
X-Face:	(:YAD AT JS'&Kz'M}n7eX7gEvPR6U1mJ-kt;asEc2qAv;h{Yw7ckz<7+X_SYeTNAaPui:e~x$
	,A=gkt*>UPL/}\a/#C~v2%ETiAY_sx;xve0yL??JWTtX_-NUzXyP38UdW#cmN1\4(X!c3m#%IbtB-3
	Z-!xpZi!`E.s{(;aP=b11"!3wQu]1j@^V|;n=B|{l<bZV1.AI`zWV%kPCnUhcgEe\(}/_kNd6,*3ZJ
	Q3o<YQ3^u;7jS=:p0--u3msQO
X-Attribution:	sb
From:	SL Baur <steve AT xemacs DOT org>
In-Reply-To:	Shawn Leas's message of "Thu, 3 Sep 1998 22:11:39 -0500 (CDT)"
Mime-Version:	1.0 (generated by tm-edit 7.108)
Date:	03 Sep 1998 20:45:04 -0700
Message-ID:	<m2n28ga6db.fsf@altair.xemacs.org>
X-Mailer:	Gnus v5.6.42/XEmacs 21.0 - "Irish Goat"
Sender:	Marc Lehmann <pcg AT goof DOT com>
Status:	RO
Lines:	37

Shawn Leas <sleas AT ixion DOT honeywell DOT com> writes in beastium-list AT Desk DOT nl:

> On Fri, 4 Sep 1998, Topas wrote:
>> I rebuild the kernel of Redhat5.1(kernel ver 2.0.35) using pgcc 1.1a,
>> and (1) the sound device can not start (2) the Xwindow can not use too.
>> It works fine when I used pgcc 1.03a, what's wrong?

Building 2.0.x kernels with egcs/pgcc/gcc-2.8 is not supported, so
whether they work or not is apparently a matter of luck.

You might give the 2.1 kernels a shot.  I've run 2.1.117 through
2.1.119 without a single crash and currently have about a week of
uptime.

$ uname -a
Linux altair.xemacs.org 2.1.119 #3 Thu Aug 27 23:29:22 PDT 1998 i586
$ uptime
  8:34pm  up 6 days, 20:39, 14 users,  load average: 0.32, 0.18, 0.20

>> by the way, I'm NOT on the mailing list yet

> Wait, didn't see the last comment about "works fine when I used pgcc
> 1.03a".  Well, out of curiosity, are you taking all the necessary steps?

What "necessary" steps?  I've built kernels 2.0.31 through 2.0.34
with default compiler flags and whatever version of pgcc was current
at the time without having to change any source.  I've heard of the
replacing ioport.c trick, but my understanding is that the problem is
much deeper than that.


There are regression problems with pgcc-1.1a that I haven't tracked
down yet.  Building python with -O6 -mpentium -march=pentium gives new 
`make test' bombs, and building XEmacs `-O6 -mpentium -march=pentium'
results in a binary that drops core at startup.  In the latter case,
adding the flags `-fno-risc -fno-peep-spills' results in what appears
to be a binary of normal stability.

- Raw text -