On Fri, 03 May 2013 19:34:17 +0200, Martin Sucha <[email protected]> wrote:
Hi,
On 05/03/2013 06:57 PM, Dominik Taborsky wrote:
On Fri, 03 May 2013 17:57:21 +0200, Martin Sucha <[email protected]>
wrote:
Hi Dominik,
On 05/03/2013 05:19 PM, Dominik Taborsky wrote:
I think you may have misunderstood the point of this library. This
library is aimed at HelenOS, not to replace standard tools. It is for
the case when you have blank harddrive so that you can format it and
install HelenOS. It's not designed to fit all needs. If it is really
necessary (for inter-OS compatibility), you can use better tool
instead,
since you are going to be using a different OS anyway.
I tend to disagree, I'd like to be able to install HelenOS alongside
another operating system (or two) from within HelenOS, even on a
notebook computer with recovery partitions and such things. I think the
ultimate goal is for HelenOS to *be* the standard OS you use to run
programs on your computer ;)
I already have triple-boot setup on my laptop (Windows/Linux/HelenOS),
why should I have to purchase a separate hard-disk just to use the
installer on real hardware?
What I meant was that when you do this kind of setup, you partition the
drive from Linux anyway. I also have dual, triple or multiple boot
setups. When you get your hardware with preinstalled Windows, you have
to run Linux so that you can run ntfs-resize, for example. When you
don't plan on using Windows but only sensible OSs like Linux, then you
either partition the drive from Linux so that you can use LVM and other
advanced stuff, or you are just fine with LBA-only MBR from HelenOS.
OK, so if I add say working ext4fsresize (or ntfsresize) to HelenOS,
would the installation work on my computer without using Linux?
(It wasn't clear to me from your previous e-mail, since you mentioned
that it is for the case when you have a blank harddrive).
I don't see why not. CHS has been obsolete since ATA5 or ATA6 (as I've
read) and even Windows have been using LBA since XP or Vista. But that's
not tested (I only know they refuse to work when there's wrong partition
type even though the FS is fine).
So instead of saying the library is implemented this way because it is
aimed at HelenOS, isn't it better to say that the aim was to create a
prototype for your thesis?
Surely, I admit it is for my thesis. But I also sincerely think an
all-or-nothing implementation is not necessary. Or do you have disks
smaller than 8.4GB on top of legacy hardware and use legacy OSs? :-)
Well, I also own a box with a 486-like SoC, who knows how the IDE
emulation for the 1GB SD card that contains an old Linux image is done :)
Oh yes, let's all go back to 486s. I hope you don't mean that as a serious
argument :-)
To clear any doubt:
We have CHS addressing, which has been obsolete for more than 10 years
now, deprecated even longer. It supports drives with capacity up to 8.4GB.
It is awful and brings nothing more than compatibility with configurations
that are better to be forgotten.
On the other hand, we have very simple LBA, that just works and is being
used by everything. It supports drives up to 2TB, but nowadays it it's not
enough so people are starting to adopt GPT with 64bit LBA so that you can
have enough space to backup the Internet.
Since last update of CHS it has been set to support up to 1024 cylinders,
255 heads and 63 sectors. If a partition starts or ends outside this
border, CHS is set to FE FF FF, which means it's invalid and LBA should be
used instead.
Do we need CHS?
What I can do (and probably will) is to set CHS always to FE FF FF. OK?
Dominik
PS: I am all for compatibility, efficiency and I like embedded systems in
general. But this is counterproductive.
Regards,
Martin Sucha
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
