svn commit: r216230 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs

[Warner Losh]

From: Erik Trulsson <ertr1013 at student.uu.se>
Subject: Re: svn commit: r216230 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
Date: Tue, 7 Dec 2010 13:10:58 +0100

> On Tue, Dec 07, 2010 at 01:57:04PM +0200, Andriy Gapon wrote:
> > on 07/12/2010 13:51 Bruce Cran said the following:
> > > On Tue, Dec 07, 2010 at 01:31:27PM +0200, Alexander Motin wrote:
> > >> Not necessary. Some places indeed may have some legacy requirements,
> > >> for example, in theory MBR want partition to be aligned to "track
> > >> boundary" (but I've seen many pre-formatted SD cards with MBR
> > >> violating it to align partition to flash sector). Same time for BSD
> > >> label I see no problem to align partitions any way we want. I also
> > >> see no problems to make FAT cluster, UFS block/fragment, etc, to
> > >> match some sizes.
> > > 
> > >>From a new installation of Windows 7 and FreeBSD CURRENT:
> > > 
> > > GEOM: ada0: partition 3 does not start on a track boundary.
> > > GEOM: ada0: partition 3 does not end on a track boundary.
> > > GEOM: ada0: partition 2 does not start on a track boundary.
> > > GEOM: ada0: partition 2 does not end on a track bounary.
> > > 
> > > Partition 2 is the reserved partition while 3 is an NTFS partition, both
> > > created in the Windows setup application.
> > > 
> > > Since Windows isn't bothering to align partitions do we still need to
> > > warn about it?
> > > 
> > 
> > No.
> > And another reason is that modern drives do not actually report any CHS
> > parameters, so I don't even know where we get them and how we (pretend to) know
> > we track boundaries are.
> 
> Modern drives do actually report CHS parameters.  The values reported
> rarely (if ever) have any relationship whatsoever to reality, but they
> are reported.  I guess the reason for them still being reported is to
> satisfy old systems that do not undestand LBA and insist on getting CHS
> parameters.

And even if the CHS parameters are reported, they are molested at many
layers so cannot be trusted at all.  CF Cards will report one thing if
attached via usb adapters and another thing if attached directly to a
pata bus, for example.  Sometimes this is software (cam lying), other
times this is hardware (usb adapter lying).

Warner