[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: status grub2 port of grub-legasy map command

From: Vladimir Serbinenko
Subject: Re: status grub2 port of grub-legasy map command
Date: Sat, 18 Apr 2009 01:46:33 +0200

On Sat, Apr 18, 2009 at 1:12 AM, John Stanley <address@hidden> wrote:
Hi Javier and all,
I could probably have a version of the drivemap completed this weekend as well, but since I've basically simply adapted your drivempa.patch.8 to r2106, with little change, I'd be more comfortable going with your new patch since you know the code much better.

In addition, in the last several days,  have started leaning away from it altogether -- because of the use  of an int13 intercept TSR. When I started looking at the drivemap patch about a week ago, I assumed it simply did some sort of bios (internal) table swapping only. I'm not really comfort able with the int13 handler redirect. I worry about (possibly rare) problems downstream while running Windows. On the other hand, I know it "does work," as I and many others have been using map with grub-legasy for several years now. Nevertheless, I'm now a bit apprehensive.
Modifying internal bios tables isn't practicable because they may be anywhere and in any format. Windows follows the memory map supplied by int12/int15/bda which says that the ranges used by hook are unusable. Also even if windows overwrites int13 handle it doesn't matter at all because windows switches to its own disk drivers after initial start

For me, for the time-being, as my laptops are all Thinkpads, which have bios supplied "boot device table"  functionality (via f12 on poweron), I can live without drivemap for those occasions when I need to come down from the hilltop to work on Windows.

What I'd like to do with drivemap, is to do boot device swapping by changing in the bios which device is to be treated as the "first" boot device, which is the second, etc. I know this is doable in principle, but it may not be practical, or wise-- I'm not sure.
There is no public API to do so. So you need to reverse engineer BIOS which is impractical to do for every possible mobo.
 It would also have to be done in a way which does not "pull the rug out on grub" insofar as boot device enumeration is concerned.

Javier Martín wrote:
Hello there. I am the original author of the drivemap command. I was
told by a friend that it was being discussed again on the GRUB list, so
I came back and examined the two new enabling functionalities, preboot
hooks and mmap. I've adapted drivemap to use them both, and I'm
polishing some rough edges. I might send a new version of the patch this
weekend. Should people, however, prefer John's version, I'd be glad to

El mié, 15-04-2009 a las 11:34 +0200, phcoder escribió:
Yes it is. Also it's better to use grub_mmap_iterate instead of basing the location on 0x413 value. How to do it look at mmap/i386/pc/mmap.c
The preboot hooks patch works beautifully. However, mmap causes some
sort of glitch that makes FreeDOS (one of my three drivemap tests,
together with ReactOS and XP) triple-fault. I've been trying to debug it
with GDB-over-QEMU, but it is a real PITA. I haven't found any sort of
bug in the mmap code, so I'm quite dumbfounded right now. Nevertheless,
I'm proceeding to the XP test, which would likely be the most common use
case for drivemap.


Grub-devel mailing list

Grub-devel mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]