Subject: Re: bootblocks, VCs, gcc
To: Louis A. Mamakos <louie@TransSys.COM>
From: Dirk Steinberg <steinber@machtnix.ert.rwth-aachen.de>
List: current-users
Date: 02/01/1994 12:17:36
>>>>> "Louis" == Louis A Mamakos <louie@TransSys.COM> writes:

    Louis> The current problem appears to be keyboard input causing
    Louis> the bootstrap to hang.  I hacked boot.c to never perform
    Louis> any input operations, and it succeeds in loading a kernel
    Louis> off the floppy disk and starting it up.

    Louis> The kernel it loaded seems to hang, complaining:

	wdc0: timeout in wdintr WDCS_BUSY
	wdc0: timeout in wdintr WDCS_DRQ
	wdc0: timeout in wdintr WDCS_DRQ

    Louis> and then hanging.  I don't know how "real" a problem this
    Louis> is, as this is a circa Jan 14th kernel built for my other
    Louis> system, and not a generic kernel.

    Louis> More as I expore this further...

    Louis> louie

I also see these timeout messages when booting a -current kernel on my
386DX/40, 16MB, IDE LPS240AT system, i.e., the current wdc.c is
completely broken for me. I hacked the old wd.c from 0.9 to compile
with -current and that works for me. Not that I'm happy with this....
What has changed?

BTW: Is there any good reason why NetBSD does *not* adopt VC's
(virtual consoles)? They are indispensable when working without X,
i.e. on a laptop. Linux even offers mouse-based cut-and-paste between
VC's ala xterm. Nice. Does syscons or pcvt work with NetBSD-current
and XFree-2.0?

Another note: Why in the world is everything compiled with -O? I
remember that -O == -O1, and that many (most) gcc optimizations are
only enabled with -Ox (x >= 2). What about -fomit-frame-pointer?
The 486 has so few registers that this does pay off. Linux compiles
*everything* (kernel, libc, user progs) with -O2 -fomit-frame-pointer.
What about gcc 2.5.8?

Currently I'm working on merging the Intel Pentium/486/386 gcc
extensions into 2.5.8, and I'll try to 'make world' when I succeed.

	Dirk

PS: I patched my sys.mk to use -O2 -fomit-frame-pointer by default,
    but I'm chicken and didn't use that for the kernel so far.

-----------------------------------------------------------------------------
Dirk W. Steinberg - RWTH Aachen - Internet email: steinber@ert.rwth-aachen.de
Aachen University of Technology / IS2-Integrated Systems in Signal Processing
Rhein.Westf.Tech.Hochsch. Aachen / Integrierte Systeme der Signalverarbeitung
Templergraben 55 / D-52056 Aachen / phone:+49 241 807879 / fax:+49 241 807631
Home address: Kleikstr. 63, D-52134 Herzogenrath,Germany/phone: +49 2406 7225

------------------------------------------------------------------------------