On 2017/10/12 00:18, Rostislav Krasny wrote: > On Wed, Oct 11, 2017 at 7:01 PM, Stuart Henderson <s...@spacehopper.org> > wrote: > > On 2017-10-11, Rostislav Krasny <rosti....@gmail.com> wrote: > >> On Wed, Oct 11, 2017 at 6:28 AM, Eric Furman <ericfur...@fastmail.net> > >> wrote: > >>> On Tue, Oct 10, 2017, at 04:29 PM, Rostislav Krasny wrote: > >>>> I think it's worth to be supported. The RAID mode of storage > >>>> controller seems to be a default BIOS configuration in all modern > >>>> desktop computers. I think most desktop users don't configure any real > >>>> RAID and continue to use their disks as separate devices. If at least > >>>> this RAID configuration is supported it would be a great progress. > >>> > >>> I disagree, but that's just my opinion. > >>> And just because something is "a default BIOS configuration in all > >>> modern desktop computers" doesn't mean it's a good thing. > >> > >> Most desktop users don't change BIOS settings. They just try a > >> software and if it doesn't work they usually throw it away and > >> consider that software as bad or broken. In case of dual-boot they > >> also depend on the previously installed OS. In case of already > >> installed Windows changing the storage controller mode in BIOS from > >> RAID to AHCI leads to BSoD. After all not supporting a common BIOS > >> configuration leads to loss of users. > >> > >> What is not good in RAID mode without actual RAID array, except the > >> fact OpenBSD doesn't run on it? > > > > What is not good is when you do have a RAID array, the controller is > > in RAID mode, but OpenBSD doesn't understand the metadata, so it corrupts > > data on the disk. > > This is not the case.
Yes that *IS* the case. What is going to happen if you, for example, modify the DOS-style partition table on only one component of the array? Writing data out-of-sync between the disks counts as corrupting the array as far as I'm concerned. > Can you identify RAID mode without RAID array > and for now support only this? That could be a good compromise. Yes that might be possible. It might also be possible to support the foreign metadata in softraid. > > This is a difficult area. We don't want to corrupt data, but then some > > BIOS don't allow this option to be changed at all, and on others BIOS > > only offer a choice between IDE and (unsupported) RAID, even though > > it's an AHCI controller. > > Yet another reason to support RAID mode without RAID array. Yes it would be useful for those users who are affected by it. Perhaps one of them will be sufficiently motivated to do the work. (From past experience what actually happens is they build a custom kernel with the PCI ID added and forget about it until upgrade time). > > I don't think we (developers) are particularly bothered about users > > who can't be bothered to change an existing BIOS setting. > > Even if they can it could be too late, as in my case with Windows. > This is lose-lose situation. You just lose users and popularity. http://www.openbsd.org/goals.html says nothing about popularity. It's nice to provide something that is useful to others, but other things are more important to the project.
