[syslinux] Syslinux 4.03-pre5 - final in a day or two

[Michal Soltys]

On 10-10-13 13:17, Gene Cumm wrote:

> 2) I just got in touch with Dell developers with regards to DRMK.  I
> want to try and verify what they told me and see if there is unneeded
> code in chain.c32.
>

In this context, if you're up for some testing, you could try my version 
of chain module. Optimistically it shouldn't hang on any drmk variation 
(and it doesn't rely on any drmk-specific mangling).

git://git.hasevolq.net/syslinux.git chaindev

build as usual, then

1) for kernel-on-fs version (option setbpb is implied):

chain.c32 hdX Y drmk=<kernel> save

defaults assumed in this mode:

sector @ 0x2000:0 (can be changed with sect=S[:O[:J]])
file @ 0x70:0:0 (can be changed with seg=S[:O[:J]])
no handover, thus ds:si,bp point to the sector

2) for direct chainloading

chain.c32 hdX Y setbpb save

In both cases (when booting from a partition), there's another 
potentially useful option - 'mbrchs' - it walks whole partition layout 
on the selected disks and sets proper CHS values matching current bios 
imaginations (which can be quite surreal - recently testing partitioned 
usb stick, I met geometries with 1 head, 50 spt ... on modern 2009+ 
motherboards). Shouldn't really matter these days, but who knows what 
random dos-ish system build could use or expect (in context of CHS 
values stored here or there).