On 2016-04-21 23:51, Dennis Gallien wrote:
On Thu, Apr 21 11:09:43 PM Carlos E. R. wrote:
No, this is not the method I use. I have grub in partition 4. I do not use this method because it causes problems on some Windowss upgrades.
Right, only pay attention to what I mean, not what I actually write. :)
Ok! :-)
Having MBR generic, and grub in a partition, you can change the boot mark to the windows partition to make it temporarily happy during certain upgrades, and is easy to undo.
Are you aware that you can change the partition table boot flag with the grub command "makeactive"? (Not in grub2 though.) Perhaps this would give you the flexibility you need?
No, I had forgotten this. So, you suggest to tell grub to make the windows partition active when booting windows. This is written to disk, I suppose. When is this undone? Because if it remains this way, grub will not boot again, as it is installed in another partition, not in the MBR.
The method Neil posted looks promising. I have already made the changes, but I still have to reboot to test. And then wait for a service pack.
Re the 440 byte "detail", the table actually begins at byte 446. The bootstrap is inside the first 439 bytes, followed by the disk signature, then the table. Neil's code will be written to the code section, and will essentially do what grub (original) does when installed to the mbr with a pointer to partition 4. The only risk I see with this approach is that Windows has been known to rewrite the boot code in an update; I have no idea is that still happens but I wouldn't be surprised (MS sometimes messes around with other bytes in that sector).
Yes, I'm scared of that. Which is one reason why I installed grub on a partition initially. -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar)