bzdisk broken in 2.4.5?

bzdisk broken in 2.4.5?

[D. Stimits]

I've tried on two separate machines to test out 2.4.5 through the "make
bzdisk" boot floppy, and it fails on both (the compile succeeds, but
boot never gets to LILO, it simply gives "400" and a repeating list of
AX, BX, CX, and DX registers). Both are scsi aic7xxx, but use different
controllers, and have scsi directly compiled in. One machine is based on
RH 7.1 beta, the other on RH 7.1. Both are x86 SMP, with motherboard and
all hardware being different. Using the same kernel through a
"mkbootdisk" works, only "make bzdisk" fails. Can anyone here verify
that "make bzdisk" will create a bootable floppy (I did try an entire
box of different floppies) on 2.4.5+? Especially, can anyone verify this
for SMP and/or purely scsi machines? If scsi, do you use aic7xxx?

D. Stimits, stimits@idcomm.com

Re: bzdisk broken in 2.4.5?

[D. Stimits]

"D. Stimits" wrote:
> 
> I've tried on two separate machines to test out 2.4.5 through the "make
> bzdisk" boot floppy, and it fails on both (the compile succeeds, but
> boot never gets to LILO, it simply gives "400" and a repeating list of
> AX, BX, CX, and DX registers). Both are scsi aic7xxx, but use different
> controllers, and have scsi directly compiled in. One machine is based on
> RH 7.1 beta, the other on RH 7.1. Both are x86 SMP, with motherboard and
> all hardware being different. Using the same kernel through a
> "mkbootdisk" works, only "make bzdisk" fails. Can anyone here verify
> that "make bzdisk" will create a bootable floppy (I did try an entire
> box of different floppies) on 2.4.5+? Especially, can anyone verify this
> for SMP and/or purely scsi machines? If scsi, do you use aic7xxx?
> 
> D. Stimits, stimits@idcomm.com

I found some references to bzdisk breaking in 2.3.28, followed by a fix.
Checking /usr/src/linux/arch/i386/boot/bootsect.S, the fix has remained
and has not been lost. See:
http://web.gnu.walfield.org/mail-archive/linux-kernel/1999-November/1818.html

However, something else must have changed since then to cause the image
size to go over its maximum. Has bzdisk being abandoned (or at least
ignored)?

D. Stimits, stimits@idcomm.com