Bug 124802
| Summary: | Can not boot in Windows XP from GRUB | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Nicolas Mailhot <nicolas.mailhot> | ||||
| Component: | grub | Assignee: | Jeremy Katz <katzj> | ||||
| Status: | CLOSED WORKSFORME | QA Contact: | Mike McLean <mikem> | ||||
| Severity: | medium | Docs Contact: | |||||
| Priority: | medium | ||||||
| Version: | rawhide | ||||||
| Target Milestone: | --- | ||||||
| Target Release: | --- | ||||||
| Hardware: | All | ||||||
| OS: | Linux | ||||||
| URL: | http://savannah.gnu.org/bugs/?func=detailitem&item_id=9159 | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2004-06-02 20:13:08 UTC | Type: | --- | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
|
Description
Nicolas Mailhot
2004-05-30 13:00:31 UTC
Created attachment 100703 [details]
Various command outputs
Shouldn't it be
title Ouine
debug
map (hd0) (hd2)
map (hd2) (hd0)
rootnoverify (hd0,0)
makeactive
chainloader +1
instead?
Well, this one produces an error 13, invalid executable format The entry I gave is the one I used when the system was FC1 + Windows 2000 (don't remember which software made the partitions back then), so it might have worked :( I'm a bit suspicious about your rootnoverify (hd0,0), I think grub mapping is for external use only (ie grub will always use original disk names regardless of maps) I know the external ATA controler makes some systems mad - half the OSs/disk utilities enumerate in bio order (on-mobo VIA then SII680), and the other half in PCI id order (SII680 then VIA). I'm half of a mind to try all the possible map & root permutations, but surely smart human beings like us should be able to type the right one just by looking at the debug info, right ? (and what if *none* of them work ?) Well, I solved it Not cleanly, I'm afraid, I cheated a bit: 1. replaced all references to (hd0,0) with (hd1,0) (it just happens hda1 and hdc1 are a raid1, and regardless of the drive ordering chosen by BIOS/Windows/Linux (hd1,0) *always* hits hda1 or hdc1 (I shudder just thinking what will happen if I have to remove one of those two drives to get it replaced) 2. created a grub floppy with menu like explained in the gru faq, removing the device map anaconda created 3. changed the bios boot order to have the windows disk before the linux ones (floppy,hd0,scsi instead of floppy,scsi,hd0) 4. booted on the floppy and used it to install grub on the windows disk MBR using the files on (hd1,0) (no idea if it was hitting hda1 or hdc1) And it works. But I pity the next SOB that will have to go through this. I is different from my win2k/fc1 receipe, even though the partitions are ordered mostly the same logicaly It seems the core of the problem is the BIOS and Linux disagree on the drive order, grub chooses either the BIOS or the Linux side depending if you get into it from the floppy or a linux drive, windows old legacy parts believe the bios but new ones behave like linux (the management console do shows drives in the linux order). So it's a big f* mess. And It's not even some exotic harware:(( |